Research Conference On Shear Strength of Cohesive Soils (1960 University of Colorado)

Research Conference on Shear Strength of Cohesive Soils.
New York, N.Y. : American Society of Civil Engineers, 1960.
https://hdl.handle.net/2027/mdp.39015009800312
Public Domain, Google-digitized

http://www.hathitrust.org/access_use#pd-google
We have determined this work to be in the public domain,

meaning that it is not subject to copyright. Users are
free to copy, use, and redistribute the work in part or
in whole. It is possible that current copyright holders,
heirs or the estate of the authors of individual portions
of the work, such as illustrations or photographs, assert
copyrights over these portions. Depending on the nature
of subsequent use that is made, additional rights may
need to be obtained independently of anything we can
address. The digital images and OCR of this work were
produced by Google, Inc. (indicated by a watermark
on each page in the PageTurner). Google requests that
the images and OCR not be re-hosted, redistributed
or used commercially. The images are provided for
educational, scholarly, non-commercial purposes.
AMERICAN SOCIETY OF CIVIL ENGINEERS
AMERICAN
ENGIREERS
RESEARCH CONFERENCE
ON
SHEAR STRENGTH OF COHESIVE SOILS
Sponsored by the
SOIL MECHANICS AND FOUNDATIONS DIVISION , ASCE
UNIVERSITY OF COLORADO JUNE , 1960

BOULDER , COLORADO
. &
,
MICHIC
VERSITI
MICHI
UNIV
VERIL
* THE
TE
1270
CENTIE
3
19
ARIES
LIBRAR
AMERICAN
SOCIETY OF
CIVIL
ENGINEERS
FOUNDED
1852
RESEARCH CONFERENCE
ON
Sponsored by the
SOIL MECHANICS AND FOUNDATIONS DIVISION , ASCE
Hosts
UNIVERSITY OF COLORADO
and
COLORADO SECTION , ASCE
UNIVERSITY OF COLORADO JUNE, 1960

BOULDER , COLORADO
Engineering
TA
910
Al
Rh3
1960 AMERICAN SOCIETY OF CIVIL ENGINEERS
SOIL MECHANICS AND FOUNDATIONS DIVISION
BRC EXECUTIVE COMMITTEE

(Members serving at time of Conference )
J. O. Osterberg , Chairman , T . M . Leps , S. J. Johnson ,

R. A . Barron , G . A . Leonards , Secretary
TASK COMMITTEE ON SHEAR STRENGTH OF SOILS

( Responsible for the general planning
and organization of the Conference )
W. J . Turnbull , Chairman , M . J . Hvorslev , Vice - Chairman ,

J. W.Hilf , Secretary , R. A . Barron , A . Casagrande , R . B. Peck ,
H . B . Seed
LOCAL COMMITTEE ON ARRANGEMENTS

( Responsible for planning and supervising all de
tails in connection with registration , housing , con
ference sessions and social activities )
Wesley G . Holtz , Chairman , Milton E . Bender , Jr . , Hugh W .

Hempel, James G . Johnstone , Mrs . James G . Johnstone ,
Warren Raeder , Roland C . Rautenstraus , Edward Sampson , Jr . ,
Jack W. Hilf
Additional copies can be obtained from the Executive Secretary , American Society of
Civil Engineers , 33 West 39th Street , New York 18, New York . Order number 1961 - 18.
Price : $10. 00 per copy .
‫رس‬‫مراد‬
7-23 -89

Research Conference On
PRINCIPAL SECTIONS
Page
Foreword .. . .. .. ... .. . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Program of Conference .... .. .... ..

Photographic Report on Conference ....
Opening Address .. .. .. .. . .. . .. .. 17
Conference Papers . . . . . . . . . .
Conference Sessions - Written Discussions and Closures . . . . . . . . . . 985
Reports by Moderators of Panel Sessions . . . . . . . . . . . . . . . . . . . . 1121
Table of Contents - Subject Index ......... . . . . .. 1157
Index of Authors and Discussors .... ... . . .. .. .. 1163
Copyright 1961 by the American Society of Civil Engineers .

FOREWORD
PURPOSE AND ORGANIZATION OF CONFERENCE
By W . J. Turnbull , F . ASCE
The procedure of the ASCE Research Conference on Shear Strength of Co

hesive Soils departed to an appreciable extent from the general pattern of
previous soil mechanics conferences ; and a summary account of the purpose ,
organization , and conduct of the conference may assist the reader in under
standing and judging the proceedings .
After considerable preliminary investigations , the Executive Committee of
the Soil Mechanics and Foundations Division , ASCE , decided in February 1958
to sponsor a national conference on the shear strength of cohesive soils , and
it appointed a Task Committee to plan and organize such a conference . It was
agreed that the purpose of the conference should be to assemble , summarize ,
and discuss present knowledge , or lack of knowledge , of factors which govern
the shear strength or failure conditions of cohesive soils . It was also sug
gested that the conference should be of the closed type and should deal with
research and fundamental relations rather than the determination and applica
tion of shear strength in practical problems . However , it became apparent
that there was much greater interest in attending the conference than initially
estimated , and it was finally decided to hold a conference which would be open
to all , and that anyone who desired to do so would be at liberty to submit dis
cussions of the conference papers or of the topics of individual sessions of
the conference . The original purpose of the conference was not changed , but
the scope of the conference was enlarged to include consideration of and one
full session dealing with the application of shear strength data to practical
problems . The Colorado Section , ASCE , and the University of Colorado offer
ed to act as hosts to the conference . This invitation was gratefully accepted ,
and a Committee on Local Arrangements was appointed .
Ten educational, governmental , or private engineering organizations in the
United States and one each in Canada , Mexico , England , and Norway , were in
vited to prepare papers for the conference These invitations were issued on
the basis of past and current research on shear strength of soils performed
by members of the organizations . The Task Committee realized that engineer
ing organizations in other countries also could have made valuable contribu
tions to the conference , but the Committee was reluctant to change the char
acter of the conference from that of an essentially regional to a more or less
international conference . The invitations were issued to organizations in
stead of individual engineers in order to facilitate the preparation of papers
and compliance with the requirement that an adequate number of preprints of
each paper be furnished for distribution to registrants before or at the con
ference.
1. Chf . , Soils Div . , Corps . of Engrs . , Waterways Experiment Sta . , Vicks

burg , Miss .
SHEAR STRENGTH CONFERENCE
It is noted again that the objectives of this conference were not only to pre
sent new ideas and results of recent investigations but also and more impor
tantly to summarize current knowledge and suggest needed research . There
fore , definite restrictions were not placed on the length of the papers except
in regard to pertinence and conciseness , and the organizations and authors
were given the choice of presenting available data and experience as an inte
grated report on several subjects or as individual papers each dealing with a
specific topic . A total of 27 papers were submitted and represent the work of
44 authors . These papers deal with the following general topics : ( 1) hypothe
ses and concepts concerning shear strength and failure criteria ; (2) testing
equipment , techniques , and errors ; (3) summaries and critical reviews of
previously published research data ; (4) results of recent research ; (5) field
observations and properties of special cohesive soils ; and (6) application of
strength data to practical problems .
As shown in the program , the conference consisted of seven technical ses
sions , social activities and excursions , and a very informative visit to the
laboratories of the U. S. Bureau of Reclamation . Professor N . M . Newmark
opened the conference with a stimulating address on failure hypotheses , which
emphasized the research character of the conference . This address was the
only paper presented in full or in abstract at the conference . Technical Ses
sions 2 through 6 were conducted as panel sessions on specific topics as list
ed in the program . Many of the papers submitted dealt with several topics
and were used as a basis for panel discussions in more than one of the tech
nical sessions .
Each of the five panel sessions was conducted by a moderator , assisted by
an associate moderator . Each panel had six members , so chosen that differ
ent viewpoints were represented . The moderators were allowed considerable
liberty in presenting the general topic of their individual sessions , and they
adopted slightly different procedures . In general , the moderator summarized
and explained the topic of the session and then guided the discussion of the
problems by the panel members . Each member had been asked to initiate the
discussion of one or two specific problems which thereafter were discussed
by the entire panel. This more or less formal discussion was followed by a
general discussion and replies to questions submitted by the audience . Mem
bers of the audience were encouraged to submit written discussions of the
general topic of the session or of individual papers discussed during the ses
sion .
During the seventh and closing session , preliminary reports were present
ed by the moderators of the five panel sessions . The principal feature of this
session was a tape - recorded discussion of the conference , the preprinted
papers, and needed research by Dr . Karl Terzaghi , who was unable to attend
the conference because of illness . Final comments were then made by repre
sentatives from Canada , Mexico , England , and Norway , and the conference
was closed by announcements concerning written discussions ar. d the proceed
ings of the conference .
The arrangement of the material in the proceedings is shown in the pre
ceding short table of contents . The papers submitted to the conference are
assembled in groups representing the contributions of individual organizations .
Each of the sections dealing with technical sessions begins with a summary
of the problems discussed during the particular session , followed by discus
sions of the topic of the session or pertinent papers , and terminated with clo
sures by the moderators and authors involved . Further details on the panel
FOREWORD
sessions are contained in the reports by the moderators , which are based not
only on the papers submitted and discussions in the panel sessions but also
on the subsequent written discussions and closures .
Approximately 400 engineers , representing 18 countries , registered for
the conference. About 350 of the registrants actually attended the conference ,
and many were accompanied by relatives and friends so that the total attend
ance was close to 600 persons . It is believed that the panel sessions in gen
eral were successful in presenting a great amount of data and controversial
interpretations in an orderly , informative , and interesting manner . The con
ference papers , panel discussions , and written discussions demonstrate that
there are many voids in the knowledge of the strength and deformation prop
erties of clays and that these voids currently are covered by opinions and ac
cepted practice but should be filled by additional laboratory and field investi
gations . One of the objectives of the conference was to call attention to these
voids and to stimulate needed research , and it is hoped that this objective has
been partly attained .
When the plans were changed from a closed - to an open - type conference ,
it was expected that a majority of the audience primarily would be interested
in problems of practical determination and application of shear strength data .
Therefore , these problems were designated as the topic of the last panel ses
sion , but it may be noted that the discussions in other panel sessions , as well
as the questions asked and the written discussions , often dealt with practical
applications rather than fundamental relations . The improvement of practical
methods for soil testing and for investigation and design of foundations and
earth structures should be the final objectives of all research , but these ob
jectives in many cases cannot be attained without further clarification of basic
properties and relations . There is a great need of fundamental research in
soil mechanics , but there is an equally great need of demonstrating how the
results of this research and of already available knowledge can be used to
advantage in the solution of practical problems .
The organization of a technical conference requires a large amount of work
by , and the cooperation of, many persons and groups of persons . The Task
Committee on Shear Strength of Soils extends the most sincere thanks to the
hosts of the conference , the Colorado Section of the American Society of Civil
Engineers and the University of Colorado , for their generous and cordial
hospitality , and to the Committee on Local Arrangements for its perfect or
ganization and scheduling of all details , from the initial reception and trans
portation through the technical and nontechnical sessions and activities , to
the final departure of the conference members . The actual conduct of the
technical sessions was in the hands of the moderators , associate moderators ,
and the panel members ; the Task Committee is grateful for their willingness
to accept this responsibility and additional work , and their success in creat
ing interesting and informative sessions was recognized and appreciated by
practically all who attended the conference .
The permanent results of the conference are embodied in the proceedings ,
and the Task Committee expresses its sincere appreciation to all the authors
of papers , discussions , and reports . Special recognition and thanks are due
to representatives of Canada , Mexico , England , and Norway , who not only
pre
pared papers for the conference but also served on the panels , travelled long
distances , and made outstanding contributions by explaining and discussing
the results of recent laboratory and field investigations in their respective
countries . Finally , the Task Committee extends its thanks and appreciation
to the Executive Committees of the Soil Mechanics and Foundations Division ,
ASCE , for solid support throughout the organization and conduct of the con
ference and to the American Society of Civil Engineers for sponsoring the
conference , assisting in its organization , and undertaking the printing and
distribution of the proceedings of the conference .
W. J . Turnbull , Chairman
Task Committee on Shear
Strength of Soils
PROGRAM OF CONFERENCE
TECHNICAL ACTIVITIES
MONDAY AFTERNOON , JUNE 13
SESSION 1, 2 :00 to 4 : 30 p . m .
Chairmen : S . Mark Davidson , President , Colorado Section , ASCE , and

Chief Engineer , Thompson Pipe & Steel Co . , Denver , Colo . ; Willard
J . Turnbull , Chairman , Task Committee on Shear Strength of Soils ,
ASCE , and Chief , Soils Divisions , USAE Waterways Experiment Sta
tion , Vicksburg , Miss .
Assistants : M. J. Hvorslev and J. W. Hilf

Opening Address : Eugene H . Wilson , Vice President , University of
Colorado , Boulder , Colo .
Lecture on Failure Hypotheses: Nathan M . Newmark , Head , Department of
Civil Engineering , University of Illinois , Urbana ,
Ill
.
Ralph Engineering Universi

of
Discussion Peck Professor Foundation

,
,
B
:
Illinois Ill
of
ty
Urbana
,
,
14
.
TUESDAY MORNING JUNE

,
00
11
40
to
SESSION
,
2
a
:
.
.
Topic Testing Equipment Techniques and Errors

,
.
:
Moderator summary panel discussion and general discussion

,
,
's
Moderator Arthur Casagrande Professor Soil Mechanics Founda

of
and
,
:
tion Engineering Harvard University Cambridge Mass

,
,
,
Associate Moderator Stanley Wilson Partner Shannon and Wilson

,
,
D
.
:
Seattle Wash
,
PANEL
:
Bishop Reader Soil Mechanics Imperial College

of
Science and
W
Alan
in
,
,
.
University
of
Technology London England

,
Wesley Chief Earth Laboratory Reclamation Denver

of
Holtz Bureau
G
,
.
Colo
.
Woodland Shockley Assistant Chief Soils Division USAE Waterways

G
, ,
,
.
Experiment Station Vicksburg Miss

,
Consulting Engineer Hastings

on
Warlam Hudson
,
,
N
A
A
.Y
-
-
.
.
Associate Professor Soil Engineering

of
Robert Whitman
,
,
V
.
Technology Cambridge
of
Massachusetts Institute Mass

,
.
TUESDAY AFTERNOON , JUNE 14
SESSION 3 , 2:00 :
to 4 40 p m . .
Topic : Shear Strength of Saturated , Remolded Clays
Moderator 's summary , panel discussion , and general discussion
Moderator : Stanley J. Johnson , Associate , Moran , Proctor , Mueser &
Rutledge , New York , N . Y .
Associate Moderator : Robert V. Whitman , Associate Professor of Soil

Engineering , Massachusetts Institute of Technology , Cambridge , Mass .
PANEL :
Alan W . Bishop , Reader in Soil Mechanics , Imperial College of Science
and Technology , University of London , England .
Laurits Bjerrum , Director , Norwegian Geotechnical Institute , Oslo ,

Norway .
M . Juul Hvorslev , Consultant , USAE Waterways Experiment Station ,

Vicksburg , Miss .
T . William Lambe , Prof . and Head , Soil Engineering , Mass . Inst . of Tech . ,
Cambridge , Mass .
Gerald A. Leonards , Prof . of Soil Mechanics , Dept . of Civil Eng . , Purdue

Univ ., Layfayette ,
Ind
.
Jorj Osterberg Professor Civil Engineering Northwestern Uni

of
O
,
.
versity
Ill
Evanston
,
JUNE
15
WESNESDAY MORNING
,
00
11
40
SESSION
to
.m
,
4
a
:
Topic Strength
of
Shear Undisturbed Cohesive Soils

:
Moderator summary panel discussion and general discussion

,
,
's
Engineering Uni
of
Moderator Ralph Peck Professor Foundation

,
,
B
.
:
versity of Illinois Urbana Ill

,
.Y III
Associate Moderator John Lowe Associate Partner Tippetts Abbett

, ,
. ,
,
, :
McCarthy Stratton New York

N
-
PANEL
:
Laurits Bjerrum Director Norwegian Geotechnical Institute Oslo

,
Norway
.
Harold Gibbs Supervisory Civil Engineer Soil Mechanics Earth Labo

,
,
, .
J
ratory Bureau
of
Reclamation Denver Colo

,
James Gould Associate Moran Proctor Mueser Rutledge New York

&
,
,
P
.
N
Y
.
.
Ronald Hirschfeld Instructor Soil Mechanics Harvard University

in
,
,
C
.
Cambridge Mass
,
.
CONFERENCE PROGRAM
Raul J . Marsal , Research Engr ., Instituto de Ingenieria , Universidad
Nacional de México , Mexico .
Stanley D . Wilson , Partner , Shannon and Wilson , Seattle , Wash .
THURSDAY MORNING , JUNE 16
SESSION 5 , 9: 00 to 11 : 40 a . m .
Topic : Shear Strength of Compacted Cohesive Soils
Moderator 's, panel discussion , and general discussion
summary
Moderator : H . Bolton Seed , Associate Professor of Civil Engineering ,

University of California , Berkeley , Calif .
Associate Moderator : Ronald C . Hirschfeld , Instructor in Soil Mechanics ,

Harvard University , Cambridge , Mass .
PANEL :
Reginald A . Barron , Engineer , Civil Works , Office of the Chief of Engi
neers , Department of the Army , Washington , D . C .
Alan W . Bishop , Reader in Soil Mechanics , Imperial College of Science
and Technology , University of London , England .
Jack W . Hilf , Supervisory Civil Engineer , Soil Mechanics , Earth Dams ,

Bureau of Reclamation , Denver , Colo .
T. William Lambe , Professor and Head , Soil Engineering , Massachusetts
Institute of Technology , Cambridge , Mass .
Gerald A . Leonards , Prof. of Soil Mechanics , Dept . of Civil Eng . , Purdue

Univ . , Lafayette , Ind .
Jorj O. Osterberg , Professor of Civil Engineering , Northwestern Uni
versity , Evanston , nl .
THURSDAY AFTERNOON , JUNE 16
SESSION 6 , 2 : 00 to 4:40 p. m .
Topic : Some Problems Associated with Practical Applications of Shear
Strength Data
Moderator 's summary , panel

, and general discussion
discussion
Moderator : Philip C . Rutledge, Partner , Moran , Proctor , Mueser &

Rutledge , New York , N . Y .
Associate Moderator : Fred C . Walker , Head , Earth Dams Section , Bureau

of Reclamation , Denver , Colo .
PANEL :
Reginald A . Barron , Engineer , Civil Works , Office of the Chief of Engi

neers , Department of the Army , Washington , D . C .
Laurits Bjerrum , Director , Norwegian Geotechnical Institute , Oslo ,
Norway .
Arthur Casagrande , Professor of Soil Mechanics and Foundation Engi

neering , Harvard University , Cambridge , Mass .
Raymond F . Dawson , Professor of Civil Engineering , University of Texas ,
Austin , Texas
John Lowe , Abbett McCarthy
III
Associate Partner Tippetts Stratton
,
-
-
New York
N
,
Y
.
.
Robert Peterson Chief Soils Mechanics and Materials Engineer Prairie
,
Farm Rehabilitation Admin Saskatoon Saskatchewan
.,
.
FRIDAY MORNING
17
JUNE
,
00
11
40
SESSION
to
m
7
9
a
,
:
.
Presiding .
on
Willard Turnbull Chairman ASCE Task Committee
,
J.
:
Shear Strength Chief Soils Division USAE Waterways Exp

of
Soils
;
.
Station Vicksburg Miss
,
.
and
Assistants Hvorslev W Hilf

M
.
J.
J.
.
:
by
Reports panel sessions

on
moderators
Discussion Terzaghi Professor Emeritus Harvard University
,
,
, K.
:
Cambridge Mass
.
General discussion
Closing
of
Conference
EXCURSIONS AND SOCIAL ACTIVITIES
GET ACQUAINTED HOUR

-
Monday Evening
13
00
30
June
to
.m
p
,
,
6
.
:
Get acquainted hour for registrants and wives Banquet Hall 5th floor
,
,
-
.
of
Harvest House Boulder

.
FAMILY WESTERN BARBECUE

Tuesday Evening
14
30
June
.m
p
6
,
,
:
Family Western Barbecue Memorial Terrace University

of
Colorado
at
Informal barbecue for delegates wives and families Entertainment

,
and square dancing will follow

.
WEDNESDAY TOURS
15
Wednesday Afternoon June

,
Colorado Big Thompson Project

30
30
to
Tour No
.m
p
1,
,
1
7
-
.
.
:
Study tour of Bureau of Reclamation earth dams power plants pumping

,
plants canals and siphons the foothills the Rocky Mountains

of
in
;
,
,
dinner Colorado State University Fort Collins

at
.
CONFERENCE PROGRAM
Tour No . 2, U . S. Air Force Academy, 1: 30 to 9: 30 p . m .
Guided tour of the new Academy near Colorado Springs , including
buffet dinner at the Officers ' Open Mess .
CONFERENCE BANQUET
Thursday Evening , June 16
Jointly with Colorado Section , ASCE , at Harvest House of Boulder
6 :00 to 7:00 p . m . Informal cocktail hour .

:
7 00 to 9 00 p m : . . Banquet , 5th floor .
Speaker : Ellis L . Armstrong , Commissioner , Bureau of Public Roads,

Washington , D . C .
Subject : Cohesion - Soils and People .

Toastmaster : Arthur Casagrande , Professor of Soil Mechanics and
Foundation Engineering , Harvard University , Cambridge , Massachusetts .
9: 30 p . m . to 1:00 a . m . Dancing
OPEN HOUSE
At Bureau of Reclamation Laboratories
Friday Afternoon , June 17
1: 15 to 5 : 30 p . m . Open House at Bureau of Reclamation Laboratories

near Denver .
Guided Tours of Soils , Concrete , Hydraulics , and Chemical Labora

tories .
LADIES ' ACTIVITIES

Monday Afternoon , June 13
3 : 00 p . m . Get - acquainted reception , University Memorial Center .
Tuesday , June 14
9: 00 a . m . to 3:00 p . m .
Tour of historic Central City in the Rocky
Mountains . Lunch in Central City .
Thursday Morning , June 16
11: 00 a . m . Tour of University Campus.
12: 30 p .m . Wives ' luncheon and ,
bridge if desired , University Faculty
Club .
PHOTOGRAPHIC REPORT ON CONFERENCE
Many of the engineers attending the Conference will no doubt remember

special events or discussions having particular significance to their under
standing and enjoyment of the proceedings . To aid in recalling the arrange
ment and conduct of the Conference a photographic record of some of the
more significant events and the main participants is presented in the follow
ing pages . a
Fig . 1. Setting for the Conference - University of Colorado
a Photographs were taken by the Photographic Dept . of the Univ . of Colorado ,

Boulder .
. Fig
Executive Committee Soil Mechanics and Foundations Division

,
. 2
A .
,:
Barron Leonards Secretary Osterberg Chairman
,
,
- O
S G
R
A
.
.
J.
.
. (
)
Johnson Leps Vice chairman
M
,
T
.
J.
Turnbull Peck Casagrande Barrom

Seed Hilf Hvarsler
Fig Shear Strength

on
Hilf
of
Task Committee Soils

W
Seed
,
H
B
3
.
.
.
.
J.
.
J. :
Secretary Hvorslev Vice chairman Turnbull Chairman

A W
M
,
. ),
,
.
. .
. .
-
R J
(
Peck Casagrande Barron

,
,
R
B
A
.
.
PHOTOGRAPHIC REPORT
Fig . 4. Local Committee on Arrangements : J. W . Hilf , E . Sampson , Jr .,

W. G . Holtz (Chairman ) , H . W. Hempel.
Fig . 5 . The opening address at the Conference was presented by Professor

N. M. Newmark shown here (second from left ) chatting with
W . J . Turnbull , s . M . Davidson and R . B . Peck .
AM
SO
Fig Toastmaster Arthur Casagrande introducing the speaker
at
the
6
.
.
Armstrong Commissioner
of
of
Banquet Mr Ellis the Bureau
,
,
L
.
Public Roads
.
Fig Moderators and Panel Members for Session Testing Equipment

,
7
2
-
.
.
Techniques and Errors Standing Wilson Associate Moderator

D
,
S
)
(
:
.
.
.
Casagrande Moderator Seated Warlam Whitman

,
R
,
V
A
W A
A
. .
.
.
W .
, .
:
, )
(
.
Bishop Holtz Shockley

W
G
A
.
.
.
.
.
PHOTOGRAPHIC REPORT
NoSmok
Fig . 8. Moderators and Panel Members for Session 3 - Shear Strength of

Saturated , Remolded Clays . Standing : S . J. Johnson (Moderator ) ,
R. V . Whitman (Associate Moderator ) , Seated : A . W . Bishop , T. W . Lambe ,
G . A . Leonards , M . J. Hvorslev , J . Bjerrum , J. O . Osterberg .
NoSmok
Fig . 9. Moderators and Panel Members for Session 4 - Shear Strength of

Undisturbed Cohesive Soils . Standing : R . B . Peck (Moderator ) , J . Lowe
III
Associate Moderator Seated Wilson Marsal

D
,
R
,
J. S
.
J.
.
.
. :
)
(
Hirschfeld Gould Gibbs Bjerrum

H
R
,
.
J.
J.
.
16 SHEAR STRENGTH CONFERENCE
Sul
No
Fig
of
10
Moderators and Panel Members for Session Shear Strength

5
.
.
-
Compacted
Cohesive Soils Standing Hirschfeld Associate
R
.
.
:
(
Moderator Seed Moderator Seated Osterberg
O
, ,
, ,
,
G
H
W B
. A
. .
. .
J.
, .
.
.
)
Leonards Lambe Hilf Bishop Barron

W
R
,
A
T
J.
.
.
.
.
Fig
11
Moderators and Panel Members for Session Problems Associated

6
.
Practical Applications
. of
with Shear Strength Data Standing Walker

F C
. R F
.
. .
. .
. : :
Associate Moderator Rutledge Moderator Seated Dawson

),
,
III C
P
.
.
(
Peterson Lowe Casagrande Bjerrum Barron

,
,
R
R
A
A
L
.
J.
.
OPENING ADDRESS
FAILURE HYPOTHESES FOR SOILS
By Nathan M . Newmark , 1 F . ASCE
INTRODUCTION
My principal qualification for being selected to give this address , I sus

fact that I know so little about the subject . However , in reviewing
pect , is the
what has been done on this topic for the past 20 years , I find that my ignor
ance is not unique . There have been many concepts and ideas advanced on
the subject of failure of cohesive materials . I shall describe some views
which appear to be of use in considering the failure of materials other than
soils . In presenting this material, I may use some terms in a sense different
from that which some of you have adopted . If I do so , it is because these
terms have had a long background with another meaning, and I am possibly
unfamiliar with the meanings which you have assigned . For example , I find
myself in disagreement with your general use of the term " deviator " stress
when you really mean " stress difference ” or possibly “ principal stress dif
ference " .
An early paper on the subject of failure of soil in shear , and still a very
good one , is the one given by Dr . Rutledge at the Purdue Conference in 1940
( Rutledge - - 1940 ) . 2 Very little has changed in the past 20 years with regard
to theknowledge of failure conditions ; possibly some new theories have been
developed , but they are still as inapplicable to actual conditions as those pre
sented 20 years ago .
In part , the difficulties arise from the lack of a clear definition of failure .
This is not a situation which is peculiar to soil mechanics , however . We have
the same difficulties in dealing with steel or concrete . Your problems are
perhaps more difficult because your materials vary more widely in their
properties , although when I read about your concern over differences of as
little as 10 percent , as determined by different kinds of test techniques, I be
gin to wonder whether you have not attempted to go beyond what we normally
feel is reasonable even with such materials as steel and concrete .
I propose to discuss some of the general concepts of stress and strain , and
then to consider the various types of simple failure hypotheses .
CONCEPTS OF STRESS
Where the stress distribution is uniform , or in the neighborhood of a point ,

the manner in which the normal and shearing stresses vary with the direction
of a plane passing through the point considered is described by the general
Prof . and Head , Dept . of Civ . Engrg . , Univ . of Illinois , Urbana ,

Ill
1.
.
Items indicated thus Rutledge corresponding entries

ar
refer
to
1940
2
,
.
-
-
ranged alphabetically the Appendix References

in
,
17
(a) Stresses on surface of

a sphere
2
(6 ) Mohr ' s circle construction
Fig . 1. Mohr ' s Construction for Stress

OPENING ADDRESS 19
transformation relationships for a “tensor ” . A simple picture of the various

possible directions of a plane is shown in Fig . 1 ( a ) by considering all of the
tangent planes to the octant of the sphere shown therein . The principal
stresses are in the coordinate directions . The normal stress o and the
shearing stress T are indicated at a particular point . If the magnitudes of the
principal normal stresses are known , the normal and shearing stresses on
any plane can be determined by a simple construction due to Otto Mohr
( Westergaard - - 1924 ) . The graphical construction with which all of you are
familiar , at least in its two - dimensional aspects , is represented , for the
three - dimensional condition , in Fig . 1 ( b ) . All of the possible states of stress
are indicated by the area marked out by the intersecting circle arcs which
can be considered as a sort of transformation of the surface of the sphere
shown in Fig . 1 ( a ) . The absolute maximum shear is ( 01 - 03 ) / 2 . The cen
troid of the shaded area can be considered as denoting some " average ” shear .
For various magnitudes of the intermediate principal stress it varies over
only a small range for a given maximum shear .
In general , it requires six quantities to define the state of stress at a point ,
namely three principal directions and the three principal stresses in those
directions . It is equally valid to have the normal and shearing stresses on
three planes , preferably in rectangular directions . However , with certain
restrictions , it is sufficient to know merely six values of normal stress and
their directions . If three mutually perpendicular directions are selected , the
normal and shearing stresses in those directions give the necessary six in
dependent quantities from which the directions and magnitudes of the princi
pal stresses can be computed . In this case the magnitudes of the principal
stresses are determinable from a cubic equation , the coefficients of which
are independent of the particular coordinate directions used . These coeffi
cients are , therefore , called " invariants " . These invariants designated as
J1, J2 , and J3 are given in terms of the normal stresses and shears in the
coordinate directions by the following relations :
= Ox + o , + oz ( 1)
-ty
xy + 0,92 + 0,
ŽX
20x
(2)
dy
-
xyzxyz yzxz
OTxyxy
2T
Uz
zx
+
XyTyz
yz
-
zx
)
?
(3
-
'
by
be
of
The invariants can also stated terms the principal stresses noting
in
the shears the planes the principal directions are zero

on
that This gives

in
the simple results

:
(4
)
02
03
J,
,
=
+
0
0301
Jy
)
(5
02
,
0203
0
=
+
dz
)
(6
,
0203
=
0
any function quantities J1

of
Of course
J2
an
J3
the and also invariant

is
,
,
,
.
It is of interest to compute the normal and shearing stresses on the planes
whose normal makes equal angles with the three principal directions . Such a
plane is called the " octahedral ” plane . The cosine of the angle between the
normal to the octahedral plane and any one of the principal directions has the
magnitude 1 / V3 . The octahedral plane direction and the stresses on it are
sketched in Fig . 4. The magnitudes of the octahedral normal stress and the
octahedral shearing stress , called herein octahedral stress , Coct , and octa
hedral shear , Toct , are the following :
Poct = } (0, + oz + 03 ) ( 7)
Foct
Vio
102
oz
03
,
,
03
12
-
-
)2
-0
į
(
)?
+
8
)
(
seen from Eq
be
can that the octahedral normal stress equal the
to
It
is
7
.
(
)
hydrostatic stress that would produce the same volume stress as the actual
"
"
the octahedral planes can be
of
on
state stress considered The stresses
.
consequently
of
of
written terms the three invariants terms the stresses
in
in
,
by
the general coordinate directions the following relations

of
means
in
:
,
Poct
-
9
}
(
)
tóct
10
da
-
-3
(
?
?
ž
(
)
seen from Eqs hydrostatic
be
that the addition

of
can and
It
(8
a
.
(
)
in )
stress consisting equal components the three directions will change the
of
octahedral normal stress by the magnitude of the additional hydrostatic

stress but will not change the octahedral shearing stress all Consequent
at
,
the total octa

is as
octahedral shearing stress

ly
the effective the same

is
,
”
"
hedral shear and the effective octahedral normal stress equal

, to
the
,
total octahedral stress minus the added hydrostatic stress component or

minus the pore pressure
to .
general state stress into two compon

of
often convenient divide

; ( is
It
ents state of hydrostatic stress accounting for the entire volume

a
a
)
:
stress and deviator stress more correctly deviator stress ten

b
a
,
"
"
(
)
sor accounting for the remainder the hydrostatic stress components are
If
,
)
designated by prime and the deviator components by double prime one

a
finds the following results

:
oct
oź
;
;
=
=
0
Jy
;
°oct
=
=
3
v
J₂
act
3
=
o
OPENING ADDRESS
Poct
Poct
-
Foct
0
=
and
Poct
,
=
o
-
di
Oct
02
=
-
on
Oct
03
=
-
0
=
J sa
oct
Oct
J; Jį
vz
-
-
=
Ž
vz
-
y
-
*
Jv
0
=
Oct
OVI
0
=
Oct poct
=
may also be desirable have third invariant which has the same
to
It
dimensions as Ooct or Toct Such quantity may be defined as

a
.
11
(
)
OC
largest normal stress

by
always designated
oi
and the smallest

is
the
If
possible shearing stress

of
given as one half

by
the dif
03
is
the maximum
,
-
or
ference these stresses one half the stress difference

in
"
"
:
max
03
,
=
Ź
*
-
(
)
be
given state
of
can shown that for stress the octahedral shear can vary
It
a
of
only between the limits 943 times the maximum shear has
to
816
It
0
0
.
the larger these two values when the intermediate principal stress equal
of
is
magnitude either the largest the smallest principal stress and

or
has
to
in
it
,
smaller the two values when the intermediate principal stress

of
the
is
the
average the largest and smallest principal stress consequently
of
is
There
,
,
.
percent change the relationship between the octahedral

15
only about
in
a
shear and the maximum shear , over the entire range of possible values . In
many cases , therefore , one may be substituted for the other .
In a natural mass of material where stresses vary from point to point , at
a point or in a small region near a particular point the relationships given
above are applicable . However , where the stresses vary , the variation in
stress from point to point must obey certain statical relationships . These
are not given here . Texts in the theory of elasticity give all of the details
( Timoshenko and Goodier - - 1951 ) .
CONCEPTS OF STRAIN
Where strains are uniform , or in the neighborhood of a point, no matter

what the properties of the material are , the variation of strain with direction
obeys precisely the same relationships as the variation of stress with direc
tion , provided that one uses in place of normal stress o the linear strain e ,
and in place of the shearing stress T, one - half the shearing strain or detru
sion , One can therefore write relationships similar to Eqs . ( 1 ) through

( 11) in terms of strain . Only two of these equations are written here , in
terms of the octahedral linear strain and the octahedral shearing strain :
( 12 )
1 + 62 + 63?
yoct
yle
13
67
,
eq
ey
Ź
-
ez
į
)
ez
-
+
)2
(
-
+
)2
(
-
)
2
Eq appears one third the magni
12
From that the octahedral strain

it
is
,
.
(
-
.Eq
be
quantity similar may also
11
to
of
tude the volume strain

in
A
o
.
(
)
as
defined for example

,
ea
ei
The octahedral strain the change

of
therefore measure volume and

is
in
,
,
or
change shape
, of
the octahedral detrusion measure the distortion

is
in
It
a
might reasonably be expected therefore that the octahedral strain general

is is
,
general
ly ly
the octahedral stress and the octahedral detrusion

of
function
a a
function of the octahedral shear many materials These relationships

in
,
are sometimes used even the inelastic range give approximate stress
to
in
strain relationships They cannot be completely general however because

,
,
.
they neglect the third invariant each case and furthermore they do not
in
take account the fact that the principal directions of stress and strain may
of
not be the same

in .
general an actual mass material the strain varies from point

of
to
In
point and the variation strain must satisfy certain geometrical relation
in
,
ships These relations designated as the equations compatibility

of
are
,
,
“
”
.
stated texts the theory elasticity Timoshenko and Goodier 1951

of
in
in
.
-
-
(
)
OPENING ADDRESS
RELATIONS BETWEEN STRESS AND STRAIN
Before considering some of the concepts of failure , it is desirable to con

sider possible relationships between the stresses and strains in a material .
For elastic materials , where the material is isotropic , the general relation
ships between stress and strain involve only two independent elastic con
stants , generally stated as the modulus of elasticity E , and Poisson ' s ratio ,
M. For non - isotropic material, additional elastic constants are required .
However , for non - linear materials the relationships become much more com
plex . Simple approximations which are adequate in many cases , including the
plastic range , involve the concept of a functional relationship between octa
hedral shear and octahedral normal stress , and between octahedral shearing
strain and octahedral detrusion , as mentioned above . Each of these relations
is , in many cases , nearly independent of the other . However , in soils , the
relations become much more complex than in materials such as metals and
concrete . The complexity arises from the fact that soil consists of solid
particles , water , and air . When soil is stressed , it deforms , and in the pro
cess of deformation , the hydrostatic stress in the water and air changes .
This change is a function of the change in volume . However , a stress situa
tion which produces large shearing deformations may cause either an in
crease or a decrease in the volume with a consequent increase or decrease
in the hydrostatic stress . Consequently the simpler relationships that might
be applicable to many metals , in which the volume stress is a function only of
the volume strain and not of the octahedral detrusion , are not applicable in
general to soils . In other words , there will in general be some coupling in
the relationships connecting the four quantities , in Eqs. ( 7 ) , ( 8 ) , ( 12 ) and
( 13 ) . Moreover , the other invariants may also be involved .
Furthermore , any relationship between stress and strain in soil, as in
other materials , involves the influence of time and speed of loading , the test
ing techniques used , and in some cases the manner in which measurements
are made.
Finally , one must consider the situation in a nonuniform mass of material
where the change in deformation and in stress from point to point may in
volve a considerable change in the stress - strain relationship arising both
from the differences in the conditions at neighboring points , and from the
fact that failure for large deformations may proceed in a progressive manner
from point to point .
In spite of these various limitations , it is suggested that consideration be
given to the possibility of defining a relationship between stress and strain in
cohesive soils by use of relationships , even in the non - linear range , of the
form indicated in the following :
f ,lock )
oct
14
+ f2 ( 7
fz
Eoct
(m
=
+
(
)
)
15
f5
Yoct f46ock tock

Ź
)
+
fo
)
+
)
(
) a
oct
fol
oct
fg
16
=
0
+
(
f
+
)
(
)
(
)
these equations are arbitrary functions the near

fg
the functions
to
f1
In
ly In
linear range these may be defined by constants times the variables

.
,
Possibly the in volume , which is proportional to the octahedral linear
change
strain , ( 14 ) , can be written as a simple function of the octahedral
in Eq .
stress or the volume stress , and the octahedral shear . A relation similar to
Eq . ( 14 ) , giving the change in pore pressure , which is related to the change
in volume , has been given ( Henkel - - 1960 ) .
It can be noted that it would be possible to invert the form of these equa
tions to state the octahedral stress and the octahedral shear in terms of func
tions of the octahedral strain or volume strain , the octahedral detrusion , and
the third strain invariant 0 .
CONCEPTS OF FAILURE
The definition of failure has not been given in general terms . We may
mean the beginning of inelastic action or we may mean actual rupture of the
material . In cohesive soils the failure situation can be stated in various
ways , some of them relatively arbitrary but in general accounting for the
situation at the beginning of loss of shearing resistance or at a relatively ad
vanced state in the loss of shearing resistance . It is difficult to be precise
about the particular theories of failure which might be used if one is not pre
cise about the description of the failure condition . The situation is further
complicated by the fact that failure as measured by any of the commonly
used concepts for cohesive soils may involve relatively large differences in
the strains at which such failure occurs .
Possibly a clearer picture might be obtained if a relationship similar to
Eq . ( 15 ) , but with f3 zero , is reasonably applicable . If this can be considered
to be the case , the octahedral shear and the octahedral detrusion can be
plotted in a simple relationship against one another to give a curve of the
type that one usually considers applicable for the relationship between linear
stress and strain in a particular direction in a material. Possible relation
ships might be like those in Fig . 6 (b ) . In the case of the " unstable ” curve in
this figure , the position of the point of failure is relatively unequivocal , at the
high point of the curve . However , in the “ stable " situation , the condition at
failure is more arbitrary , and also corresponds to a considerably higher de
formation .
THEORIES OF FAILURE IN TERMS OF STRESS
One - Parameter Theories

Simple theories of failure involving only one parameter may be stated in
terms of the condition that failure occurs when the maximum stress (maxi
mum shear , or maximum octahedral shear , etc. ) reaches a certain limiting
value . Such theories as these which involve maximum shearing stress or
octahedral shearing stress imply that the strength in tension and compres
sion are equal . Further discussion of these theories will be included in a
more generalized form in the next section .
Two - Parameter Theories

Thebest known and most widely used of these theories is Mohr ' s theory of
failure . This is sketched in Fig . 2 . It is of interest to indicate the complete
curve including the Mohr ' s circles for simple tension , pure shear , and simple
:-'
-
-
constant
tan
o
+
C
=
T
Triaxial
-|
Tension
OPENING ADDRESS
Lsimple
Tension
Pure Simple
Shear Compression
.2.
of
Fig Mohr
's Theory Failure
compression . The theory involves an envelope or limiting line to all of the
possible Mohr circles that can be drawn for states of stress . However , from
reference to Fig . 1, it can be seen that this whole class of theories of failure
is independent of the intermediate principal stress . Moreover , the theories
imply that the strength in simple tension is greater than half of the strength
in pure shear .
Mohr ' s theory has the advantage that it is so widely used and well under
stood by soil engineers . Moreover , it gives the plane on which failure occurs
and the angle such plane makes with the principal directions .
Another and possibly more useful set of theories of failure involve the re
lationship between the octahedral shearing stress and the octahedral normal
stress as indicated in Fig . 3 (Nadai - - 1933 ) . This is in many respects simi
lar to Mohr ' s theory , but the construction is somewhat different . With this
theory there is some effect of the intermediate principal stress , but not a
large one .
Incidentally , the maximum shear theory of failure is defined by a horizon
tal line on the Mohr diagram . The one - parameter theory involving the con
cept of a limiting octahedral shearing stress involves merely a horizontal
line on the octahedral diagram .
This group of failure theories , or any failure theory which can be repre
sented on a planar diagram with a single curve , can take into account only
two of the three invariants of the stress tensor , and therefore is not of com
plete generality . This is not a serious criticism because of the fact that
there is fairly good evidence that simple theories of the form of Mohr ' s
theory apply quite well in describing the conditions of failure of soil , pro
vided that one takes proper account of the hydrostatic stress or pore pres
sure .
Three - Parameter Theories

Athree - parameter theory can be developed in general from theories
which can be represented on a plane by adding a third dimension or by using
contour lines on the planar representation . This is not possible with Mohr ' s
theory but the octahedral stress theory can be readily generalized .
For example , consider a plot projected on an octahedral plane , as shown
in Fig . 5 . In this figure, the normal to the octahedral plane shown in Fig . 4
is perpendicular to the plane of the paper in Fig . 5 , through Point 0 . The
projection of the three principal stress axes are indicated . For a particular
value of octahedral stress , a plane parallel to the plane of the paper is de
termined . The contour on this plane of the failure conditions represents the
various combinations of stresses , all with the same octahedral or volume
stress , for which failure occurs . It is necessary to consider only one 30
degree arc in the figure . For example , consider the points near the 01 axis
projection . In this neighborhood , the plane perpendicular to the octahedral
plane which passes through the 01 axis defines all those states of stress for
which 09 and 03 are equal . In the region of the positive projection of the 01
axis , the values of oz and oz are greater than 01 , or the situation is one of a
reduction from a hydrostatic stress condition , which can be designated by the
term “ tension " . On the other side of the axis , as in the region from the ori
gin to Point b, the magnitude of oj is greater than that of the equal stresses
02 and 02 , and this is therefore a situation of “ compression " .
Because of symmetry , the same intercepts are involved along the three
OPENING ADDRESS
oct
Foct = const .
Ioct
* flooced
oct
Fig Octahedral Stress Theory Failure
of
3
.
.
oot
Ox
xo
,
=
0
xo
No
oct
3
Fig Octahedral plane direction

4
.
.
Projection of
-
02 axis
Max . Octahedral Shoar =

Const .
+Max . Shear = Const .
L ' Possible
Failure Theory
Projection of Projection
Oz - axis of 0, - axis
Fig . 5 . Normal View of Octahedral Planes
principal axes and also at their opposite extensions . Consequently , if one de

fines the failure curve between the 01 axis projection and Point a , the same
curve is reproduced by multiple reflections for all of the other regions.
On such a plot , the curve for maximum octahedral shear equals a constant
is represented by the outer circle shown . The curve for maximum shear
equals a constant is represented by the hexagon indicated in the figure . The
inner roughly triangular curved line represents a failure hypothesis in which
the " compressive " strength is greater than the " tensile ” strength ( Topping - -
1955 , Kirkpatrick - - 1957 ) .
If the distance from Point 0 to the failure line is measured in the radial
direction , the intercept in Fig . 5 is precisely the octahedral shear multiplied
by the quantity ✓ 3. Therefore , Fig . 5 represents a radial plot of octahedral
shearing stress for a given value of octahedral normal stress , as a function
of the relative magnitudes of the principal stresses . A third parameter here
can be defined as the angle from one of the axes . The curves for other values
of octahedral stress may be drawn to show contours of the intersection of the
OPENING ADDRESS 29
octahedral plane with the three - dimensional surface. The plot in Fig . 3 can
be considered as a trace through the surface showing the intersection of the
surface with a plane containing the octahedral axis and the 01 axis . There
fore , by showing several different lines in Fig . 3, representing the traces of
the failure surface with other planes passing through the octahedral axis ,
oblique to the 01 axis , one can represent the entire surface conveniently .
It is of some interest to compare the stress conditions for three stress
situations which correspond to the same magnitudes of octahedral normal
stress and octahedral shearing stress . These are shown in Table 1. The
third state is one which represents plane strain for a value of Poisson ' s ratio
of 0 .28 . The maximum shears are not quite equal for the three stress states .
The magnitude of the third invariant for the stress situation is indicated . It
is of some interest to note that only a four percent difference in the magni
tude of this invariant accounts for a difference of nearly 17 percent in the
magnitude of the maximum principal stress . This indicates a large sensi
tivity of the stress situation to the third stress invariant stated in the form
defined in Eq . (6 ) . However when the third invariant is stated in terms of
Jh' or in terms of ø from Eq . ( 11 ) , the numbers are much different for the
BC
three stress states considered , as shown in Table 1.
TABLE 1. COMPARISON OF STRESS STATES
Stress State
Quantity
А
5 + 3 = 6 . 73
5 - 1 3 = 3 27.
1.414 1. 414 1.414
Max Shear 1. 5 1. 5 1 . 73
.
112
112
108
Jz
,
0203
=
0
most
the
un
of
=
Eq
11
0
.
(
)
COMMENTS ON SIMPLE TYPES OF FAILURE THEORIES
A group of theories involving strains rather than stresses can be developed

but it is not believed that these are particularly useful . The stresses are
more readily determined than the strains , and there seems to be no particu
lar advantage in restating the theories in terms of strain .
Concepts involving energy have been developed . As a matter of fact, the
octahedral shear theory of failure , involving Toct = constant , was originally
developed as a theory in terms of the maximum energy of distortion , in the
elastic range . However , the octahedral stress concept seems to be capable
ofmore convenient generalization .
Ithas been pointed out a number of times , as early as 1940 ( Rutledge - -
1940 ) that the effective stress rather than the applied stress determines the
failure condition . To determine the effective stress one must determine the
pore pressure . However , this is a function of the change in volume. Unless
the volume change can be determined from a relationship similar to Eq . ( 14 )
or in other ways , the magnitude of the effective stress cannot be readily de
termined in general . Consequently , much of the discussion of failure theory
in terms of stress may be entirely academic because the stresses may not be
readily definable . Much of your work at this conference will be related to the
matter of developing simple ways of finding the volume change , or pore pres
sure change , when shear takes place .
No consideration has been given in this discussion to anisotropy . Certain
ly soils are not homogeneous and isotropic , and their failure to be so causes
difficulties in the application of simple theories of failure . Moreover ,
anisotropy is induced as a result of strains . Even though isotropic situa
tions exist for small deformations , when large deformations approaching
failure are encountered , the simple relationships derived herein may require
amendment because of the induced anisotropy .
Finally , when stress - strain relationships of the sort indicated in Fig . 6 ( b )
are encountered , in the unstable range, progressive failures may take place .
Consider for example the slope shown in Fig . 6 ( a ) . Before sliding takes
place , the shearing stress distribution along a potential sliding surface is not
uniform . Consequently , conditions approaching failure may be reached at
some particular point along the curve such as indicated near the toe of the
slope . When failure occurs here , if the resistance remains nearly constant ,
then a failure condition at neighboring points may progress until a nearly
constant resistance is reached along the entire curve . However , if a stress
strain relationship is involved in which the strength drops off with progres
sive movement , then it is not possible to mobilize the maximum strength
along the entire curve , and the progress of the failure must be considered in
estimating the stability . Consequently , one must always take into account the
practical implications of the failure hypotheses used , in considering the appli
cation to an actual problem .
SUGGESTIONS FOR RESEARCH
I would hesitate to suggest that you should discard your simple procedures
involving Mohr ' s envelope in favor of the more complicated considerations
involved in the use of the octahedral theories , especially when adapted for the
most general case as indicated in Fig . 5 . However , it might be worth some
OPENING ADDRESS
(a) Failure of
a slope
stable
unstable
(b ) Shear Stress - strain relations
Fig . 6. " Progressive ” Failure Conditions

study to determine whether some of the minor discrepancies that have been
observed are more nearly eliminated by these more complicated considera
tions .
The possibility of using general stress - strain relationships of the type in
Eqs . ( 14 ) and ( 15 ) might lead to a better understanding of the possible
changes in pore pressure accompanying deformation and thereby permit de
termination of actual effective state of stress in more general terms . One
possible difficulty in investigating the applicability of these concepts , how
ever , is in the fact that ordinarily measurements of all of the quantities
needed are not available . The problem of measuring the various stresses and
deformations simultaneously is a difficult one, but it has been done in some
cases , and it would be helpful if values could be recorded in as many in
stances as possible .
APPENDIX - REFERENCES
1. D. J. Henkel, “ The Shear Strength of Saturated Remoulded Clays ,” ASCE

Research Conference on Shear Strength of Cohesive Soils , Boulder ,
Colorado , June 1960 .
2. W. M . Kirkpatrick , “ The Condition of Failure for Sands , " Proceedings

Fourth International Conference on Soil Mechanics and Foundation Engi
neering , London , 1957 , Vol. 1, pp . 172 - 178 .
3. A . Nadai , « Theories of Strength ,” Journal Applied Mechanics , Vol. 1, 1933 ,

pp . 111 - 129 .
4. P . C . Rutledge , “ Theories of Failure of Materials Applied to the Shearing

Resistance of Soils , ” Proceedings Purdue Conference on Soil Mechanics
Applications 1940
pp
Its
and 191 204

,
-
.
Goodier Theory Elasticity McGraw Hill

of
Timoshenko and
,
”,
N
5
-
.
.
.
.
., J
Book Company Inc New York 2nd Edition 1951

,
Topping Experimental Constants the Application

of
of
The Use
In
,
D
A
6
“
.
.
.
Strength Proceedings Second Midwest Conference

of
on
Theories Rock
to
",
.pp
Solid Mechanics Purdue University 1955 178 192

,
Westergaard Einfache Ableitung der von Mohr gegebenen graphis

M
,
H
7
.
.
.
"
chen Darstellung des dreiachsigen Spannungszustandes ZAMM Vol

",
4
,
.
1924 520
p
,
.
.
CONFERENCE PAPERS
By H . J . Gibbs , 1 M . ASCE , J . W . Hilf ,2 F . ASCE , W . G. Holtz , 3 F . ASCE ,

and F . C. Walker , 4 F . ASCE
1. INTRODUCTION
A. Development of Ideas and Equipment

To evaluate the practices of the Bureau of Reclamation
.
with respect to the
shear strength of soil it appears worth while to examine the historical de
velopment of these practices . The Bureau of Reclamation is engaged primari
ly in a program of investigation , design , and construction of water resources
facilities rather than basic scientific research . The written record consists ,
therefore , primarily of reasons why any investigation is required and justified
rather than a documentation of sources from which basic information was ac
quired . We cannot therefore separate the ideas which were original with the
Bureau of Reclamation engineers from those which they secured elsewhere .
The Bureau of Reclamation originated in 1902 as the Reclamation Service .
present title was introduced carrying out its responsibilities
Its
to
1923
In
in
the arid western part

of
conserve the limited water supplies
of
the United
immediately initiated construction reservoirs and canals for the
of
States
it
,
Many necessity were

of
storage and conveyance these structures

of
water
of
.
soils The engineers charged with the creation these struc

of
of
on
built and
.
tures were initially assigned projects with the responsibility investigat

of
to
ing designing and constructing all the parts of project These engineers
a
,
having very little centralized direction built their structures

of
on
the basis
,
their individual education and experience Since they had varied backgrounds ,
.
the structures they designed and built also varied considerably The gradual
.
of
development centralized control organization the interchange

of
ideas
,
a
among individual engineers and information from the performance records

of
,
structures under Bureau control were combined the cur

of
the formulation
in
rent earthwork practices

.
Soil Testing Laboratory and

of
of
Establishment the Bureau Reclamation

1
.
Initial Concepts Specialization soil problems was initi

of
Shear Strength
in
.
--
Prior this time soil problems were solved on the basis of em

to
ated
in
1933
.
pirical rules which related soil descriptions stable slopes for cuts and fills
to
The use of soils which did not fit these empirical rules were avoided the
in
construction of fills and removed from foundations Where they occurred

in
.
Special Investigations and Research Sect Earth Lab Dept

of
Head the
,
.,
.,
1
.
.
of
Interior Bur Reclamation Denver Colo

,
,
.
Soils Problems Unit Earth Dam Sect Dept

of
of
Head the Interior Bur

,
2
.,
,
.
Reclamation Denver Colo

,
Dept the Interior Bur

of
of
Chf Earth Lab Reclamation Denver

.,
.,
,
3
.
Colo
.
Dept
of
Dams the Interior Bur

of
Head Earth Sect Reclamation

,
.,
,
4
.
Denver Colo
,
33
cut slopes , frequent maintenance was assumed to provide an acceptable so
lution . Testing of soils was expected to provide refinements in the solutions
where soils with an experience record would be used and a means for dealing
with soils previously rejected .
At the time the earth laboratory was organized an extensive search of engi
neering literature was made to determine what tests should be made and how
to make them . The basic literature included Terzaghi , 1925 ; 5 R . R . Proctor ,
1933 ; and the Proceedings of the First High Dams Congress , 1933 . For the
determination of the shear strength of soil , a direct shear machine was con
structed which would handle cylindrical samples from drill cores to provide
undisturbed foundation tests . Later another direct shear machine was con
structed which would handle specimens 1 foot square and from 1/ 2 to 8 inches
thick , Fig . 1- 1. In formulating the design for this machine every effort was
made to reduce strains , stresses and friction in the machine to negligible
values . A large number of samples of soil from several dam projects had
been collected while this machine was being constructed . Based on the best
procedures that could be determined at that time , the routine testing of these
samples was started . Concurrently , bureau engineers directed their attention
to the determination of procedures for analyzing the adequacy of embankments .
Also , studies were initiated for the instrumentation of structures so that per
formance could be measured .
PX - D - 20434H
Mo -F. EDIT
Fig . 1 - 1 - Direct Shear Machine
5. Items indicated thus , Terzaghi , 1925 , refer to corresponding entries listed

alphabetically in the Appendix Bibliography .
COHESIVE SOILS 35
The development of theories of soil behavior , procedures and equipment for

testing soils , design practices , and the measurement of performance in em
bankments has been carried on simultaneously . But the rate of development
is to a large extent controlled by the types of situations encountered , and the
necessity or practicability for modification of previously established pro
cedures . Furthermore , a considerable period of time is required between the
time a theory is conceived and the proof of it on the basis of actual per
formance records is secured . This situation seems to be peculiar to Soil Me
chanics . It explains why there are still controversial areas in a general theo
ry of soil behavior , and why there are differences between theoretical con
cepts , testing procedures and practical applications .
One of the justifications for the establishment of a testing laboratory was
the growing recognition that the shear strength of soil could not be determined
from the angle of repose . Another factor which made a laboratory desirable
was a theory that the shear strength of unsaturated and saturated soils differed
greatly although there did not appear to be any difference in slopes above and
below still water . Coulomb ' s equation s = c + o tan Ø ( I - 1) where :
S = shear strength
c = cohesive strength
o = stress normal to the shear plane
tan 0 = coefficient of friction
was considered to define adequately the strength of a soil but the mechanism
by which the strength of a soil could be reduced by saturation was not ap
parent . One idea was that water acts as a lubricant . Another idea was predi
cated on the assumption that if sufficient densification could be accomplished
before saturation , then strength loss upon saturation would be negligible . The
literature available at the time was rather vague in providing answers for
these questions , and it was considered necessary to resolve them by labora
tory tests and further field observations .
The Bureau approach differed in several ways from that being followed by
engineers in the universities . For example , the first types of soils studied
were those which empirical experience had indicated to be suitable for em
bankment construction . It was the reworked soil rather than in - place soil that
was studied , and it was the unsaturated rather than the saturated condition
that received the most attention . These differences in concepts often inter
fered with the exchange of ideas with engineers whose experience was related
to undisturbed , saturated soils .
The first difficulty encountered in developing an understanding of the be
havior of soils was in the performance of actual structures . The first efforts
at measuring performance were directed to locating the phreatic line , that is ,
the line at which the hydrostatic pressure is zero . In constructing wells in
embankments to locate this line it was discovered that the water level stood
higher in deep wells than in shallow wells in some instances and in other
instances the difference was in the opposite direction . Since at the time no
rational explanation for such behavior could be found , it was ascribed to faulty
instrument performance . Various modifications were introduced to assure
that the process of instrumentation did not interfere with structure per
formance . When the use of no - flow cells failed to change observations of ab
normal behavior , further consideration was given to finding a theoretical
explanation .
In the meantime , some unanticipated behavior was being observed in the
laboratory . For example , it was noted in the routine testing of cohesive soils
that the failure envelope was curved rather than straight , as was expected
from Coulomb ' s equation . In attempting to account for this behavior it was
found that this curvature was influenced by the testing rate , the length of time
prior to testing that the preconsolidating load was applied , and by the thickness
of the sample . Another set of tests , designed to determine the most favorable
moisture condition for embankment placement to secure maximum strength ,
showed that cohesion changed noticeably with moisture content but that the
friction coefficient did not. In this suite of tests it was noticed that very dry
soils , that is , soils that were placed dry in the sample chamber , behaved as
noncohesive soils in that the failure line was straight .
2 . Recognition of the Pore Pressure Concept . - On the basis of these varied
behavior patterns it became apparent that a far more thorough knowledge of
the factors influencing soil strength would have to be developed than had origi
nally been anticipated . In a series of conferences in the spring of 1936 , the
idea was advanced that a pressure developed in the pores of a soil during con
solidation might somehow be responsible for some of the observed behavior .
In these conferences a number of other variable factors were proposed for
consideration . The first test to measure internal pressure was made July 10 ,
1936 . The first report formulating a hypothesis concerning internal pressure
and its effect on the stability of soils was issued July 23 , 1937 . This was
followed by a special report on internal pressure studies dated October 24 ,
1938 , which reported on the difficulties and development of procedures for the
mea surement of these internal pressures .
At first, these internal pressures or " pore pressures " were considered
purely a phenomenon of laboratory testing . If , therefore , they could be elimi
nated from influencing the laboratory test , then the results could be used di
rectly in stability analyses for design . A soil sample representative of the
extreme conditions anticipated in practice was exhaustively tested . This soil
sample consisted of 25 per cent silt and 75 per cent clay size particles with a
liquid limit of 49 per cent and a plasticity index of 22 per cent . The placement
condition considered was 28 per cent moisture and a dry density of 92 . 5 pounds
per cubic foot . These tests indicated that for preconsolidation periods in ex
cess of 30 minutes when a 1 - inch - thick direct shear sample was used there
would be little change in indicated strength . However , unless this preloading
was followed by a slow rate of strain , erratic results were obtained . Only
when loading and shearing together extended for 6 or more hours were shear
ing strength determinations consistent . This series of tests was reported
December 13, 1938 .
Further consideration of the test results indicated that if the effect of pore
pressure could be completely eliminated , the shear strength of clays would
approximate that found for sands , and cohesion as a source of strength might
disappear . Although this hypothesis was not taken very seriously , it did serve
to influence the organization to concentrate efforts on finding a way to de
termine and control the development of pore pressure . It also came into con
sideration in the formulation of the various theories proposed for the appraisal
of strength available in embankments . The result of these experiments was
reported by Hamilton (1939 ) .
While the above - described difficulties in formulating a testing procedure
were being explored , the results from measurements with hydrostatic pressure
indicators in actual dams were beginning to be available . Surprisingly , tests
COHESIVE SOILS 37
on some structures indicated that percolating water had moved very rapidly
through the embankment notwithstanding the low permeability of the materials .
On other dams where the permeability of the soil was higher , the development
of a flow pattern was very slow . The hydrostatic pressure indicators were
installed after the dams were completed and some of these peculiar effects
were considered the result of disturbance produced by instrument installation .
However , the mechanism of disturbance could not be determined and an expla
nation was sought , assuming the performance data were reliable .
On the basis of the pore - pressure theory being developed , it seemed possi
ble that the compression that occurred during construction was producing pore
pressures . For the more pervious soils , these pressures could drain away
before the hydrostatic pressure indicators were installed . In less pervious
soils some pressure still remained at the time the apparatus was installed .
Since very large pore pressures could be developed in the laboratory , there
was a possibility they could also be developed in the embankment . Therefore ,
plans were formulated to install the hydrostatic pressure indicators during
construction . It also appeared necessary to determine to what extent consoli
dation occurred during and following construction , and apparatus to accomplish
this was developed .
Before any more hydrostatic pressure indicators were installed , the idea
was advanced that a simple piezometer system would be both more effective
and economical . The advantages and drawbacks of such a system were thor
oughly explored , and it was decided to adopt this mechanism for future instal
lations .
In an effort to overcome the shortcomings of the direct shear test for
mea surement of pore pressure , the triaxial test procedure was adopted in
1939 , and subsequent testing , particularly further development in trying to de
termine the strength of cohesive soils , has been performed with this later
apparatus . This is explained in Part II .
3. Introduction of the Effective Stress Concept . -- The observation that
greatly enhanced strengths would result if pore pressure did not occur in im
pervious soils , led to further searches for ways to eliminate pore pressures .
If , on the other hand , pore pressures could not be eliminated , then they would
have to be considered in stability analysis . For this purpose a correction to
Coulomb ' s equation was required to s = c + (o - u) tan Ø (1- 2) where u = pore
pressure . To use this equation required further knowledge as to how pore
pressure developed . Initial attempts at direct measurement of pore pressures
were inconclusive and it seemed desirable to try and find an indirect approach
that would provide confirmation of these results . The recognition that pore
pressure was related to compression and consolidation of the soil appeared to
offer a means by which the magnitude of pore pressures at any load and at any
time could be determined .
In 1940 consolidometers were constructed on the basis of some designs
prepared by A . Casagrande with which the time and load compression charac
teristics of a soil could be measured . Since crossarms for measuring the
compression that occurs in an embankment had already been installed in
several embankments , it appeared easy to compare these two types of
measurements . At the time, it was hoped that through this indirect approach
a mechanism for relating pore pressures measured in the laboratory to those
occurring in the embankment could be found . A direct correlation between
compression measurements in the field and laboratory could not be
accomplished .
At this point two events occurred that had a major influence on subsequent
developments . The first , the start of World War II resulted in the dissipation
of much of the personnel engaged in these investigations so that progress for
a time was brought to a standstill . The second , such high pore pressures were
being measured in the 309 - foot - high Green Mountain Dam that apprehension
for its safety was felt . At the same time plans for the 456 - foot - high Anderson
Ranch Dam were being formulated . Although procedures for accurately
mea suring pore pressure in the laboratory had not been established , sufficient
evidence had been accumulated and thought given to the pore - pressure concept
to develop a faith in its existence and mechanism of operation .
If the amount of air remaining in a compacted soil was appreciably greater
than the amount of consolidation that could occur under loading then the de
stabilizing effects of pore pressure resulting from consolidation could be
avoided . This could equally well be accomplished by increasing compaction
and thus reducing consolidation , but extensive efforts in this direction had al
ready been made in an effort to reduce the permeability of soils and it was
felt that this line of attack had already been pushed to its feasible limit .
A series of consolidation tests were performed on the soils proposed for
Anderson Ranch Dam at various moisture contents . On the basis of the change
in air content , computations were made of the probable pore pressures and it
was found that a series of curves , approximately parallel to the zero air voids
curve on the Proctor compaction plot , could be drawn representing pore
pressures induced by various applied stresses . It was also noted that when
the consolidation was complete , samples compacted appreciably dry of opti
mum and wetted would consolidate further , whereas those compacted wet of
optimum would not . This " saturation collapse ” phenomenon appeared to de
scribe a limiting moisture condition below which it would be undesirable to
place embankment . Fortunately , between this lower limiting condition and an
upper limiting condition in which high pore pressure would result from con
solidation there was a sufficient range of moisture content that it appeared
possible to construct embankments without incurring either saturation collapse
or high pore pressure . This range was quite wide at low loads but became
very narrow at high loads for this particular soil. This test program was
called the limiting moisture control test .
During the war , while results of attempting the control of pore pressure
through limiting moisture control were being awaited , further attempts were
made spasmodically to find a relationship between laboratory test results and
behavior in the field . A recognition that coarse particles in the field material
had been removed for laboratory tests resulted in bringing the data into closer
agreement but insufficiently so to permit the development of confidence in the
tests . Following the war , J . W . Hilf (1947 ) pointed out that if pore -pressure
corrections were applied to the total load consolidation curves measured in
the field they would approximate the laboratory load - volume change curves .
This effective stress principle had already been recognized in 1936 in the ex
pressions explaining the use of the pore - pressure concept , but its importance
had been lost in the multitude of problems being investigated .
With the passage of time it has become apparent that the development of
pore pressure can be anticipated from laboratory tests and in the field , can
to some extent be regulated . This regulation is accomplished by selection of
materials , control of moisture , compaction , rate of loading , and design . Tests
made on the basis of effective stress can then be used with the inclusion of
proper values of pore pressure to arrive at the knowledge of a factor of safety .
COHESIVE SOILS 39
Improvements in techniques indicate that pore pressures can be measured

both in the structure and in the laboratory which agree closely with the theory .
But , in soils placed appreciably wet of the Proctor optimum the change of
pore pressure is so great with minor changes in moisture that precise control
cannot be achieved . If it becomes necessary to work with materials in this
range , it must be recognized that a saturated condition is being approached .
4. Recent Developments . - Further improvements in laboratory testing
have been introduced by the construction of a large size triaxial machine
which permits evaluating the influence of large - sized particles in the soil .
Other innovations involve making tests under constant effective stress and at
the same void ratio at the start of the shear process . These procedures at
tempt to bring the relationship between behavior in the laboratory and in the
field into closer agreement and should result in better estimates of strength .
The influence of negative pore pressures due to capillarity is being explored
with an indication that it will provide a better understanding of why cohesive
forces vary the way they do .
With shear tests based on the effective stress principle and the gradual
accumulation of test results from a variety of soils it is possible that a satis
factory correlation between visual description , or index tests and shear
strength , can be found . This should permit an improvement or extension of
present methods for evaluating soils for preliminary designs . This aspect is
being investigated as time permits .
It has been found that the pore - pressure susceptibility of soils can be de
termined by means of the consolidometer test and a working range of moisture
for compaction established that will permit the use of many soils with confi
dence . Soil selection on the basis of these tests is therefore a possible method
of control . Where saturated foundations are involved , procedures being used
include foundation unwatering , selective removal , and controlled rate of load

ing . Conversely dry , low - density foundations are being wetted to induce con
solidation under loading during construction so that sudden soil structure col
lapse by percolating water during operation will not occur . Procedures to
prevent the loss of strength by cracking induced by drying are being studied
to provide suitable design and construction procedures . Laboratory tests are
being developed to provide a better quantitative measure of cracking potenti
ality .
B. Organization of the Paper.

The terms and symbols used in this paper are those prepared jointly by the
American Society for Testing Materials and the American Society of Civil
Engineers , ASTM Designation D - 653 - 58T . The parameters : cohesion , co
hesive strength , or c ; and coefficient of internal friction , friction , or tan 0 ;
are used in the sense of being the no - load indicated intercept, and the slope of
the Mohr ' s envelope developed on the basis of effective stress , respectively ,
unless specifically described otherwise in the narrative . In unsaturated soil
tests , these parameters are obtained without drainage of the sample during
the test and pore - pressure corrections are applied . In the saturated soil tests
drainage is permitted under the application of all - around stress until further
changes in pore pressure do not occur . The sample is then sealed and tested
in the same manner as an unsaturated soil , unless otherwise noted . No differ
entiation is made in the two types of parameters , although the interpretation
of the Mohr ' s strength envelope on an effective stress basis may be different
for
saturated soils than unsaturated soils

.
The equipment and test procedures presently in use by the Bureau of
Reclamation are described in Part II . The accuracy and reliability of the
tests are also discussed .
Part III of the paper develops the pore - pressure concept , explains its ef
fect on shear strength , and describes the methods used for its measurement
in the laboratory and the field .
Parts IV and V discuss the shear strength of unsaturated and saturated
soils , respectively . Each part is further divided into disturbed and undis
turbed soils .
Part VI describes Bureau experience with special soil types which have
peculiar behavior characteristics . These may be defined as those soils whose
soil structure or chemical make - up are such that they may change during the
life of the structure with accompanying changes in strength characteristics .
To a large extent these are the soils that past prudent engineering practice
indicated should be avoided and whose performance is in some cases still im
perfectly understood .
Part VII describes Bureau experience with measuring the strength of a soil
in - situ using the vane shear test . A number of innovations have been made in
the apparatus that permits its use in an investigation with good results and
permits the elimination of a number of sources of error .
II. LABORATORY TRIAXIAL TESTING EQUIPMENT AND PROCEDURES
A. Equipment
1.
.
Size of Testing Machines and Specimens . - The shear testing equipment
and procedures used by the Bureau of Reclamation are considerably different
from those found in most soils laboratories in the United States . This differ
ence is the result of shear testing evaluation and basic principles of interpre
tation and design developed during the past two decades , as discussed in
Part I . During this period , research was being performed to study the effects
of excess - hydrostatic (pore ) pressure on the shear strength of soils in an ef
fort to explain some of the apparently anomalous results obtained in the direct
shear tests .
As result of the initial pore pressure research studies , it was decided to
a
utilize triaxial method of shear testing so that drainage could be controlled
the
and pore pressures measured . After studying triaxial equipment available at
that time, the first Bureau apparatus was designed and constructed in the
Bureau shops . This apparatus was built to test compacted soil specimens
3 - 1/ 4 inches in diameter by 9 inches high . Although the first machine has
been improved several times , this machine and the two other machines now in
use have the same basic design . Other sizes of specimens are now in use.
All specimens tested have a height: diameter ratio of 2. 25 or greater .
The three triaxial machines for soil testing available in the Bureau labora
tory are referred to as small , medium , and large machines . Several features
of the large machine were originally modeled in 1950 after the large triaxial
machine of the Corps of Engineers ' laboratory in Sausalito , California . The
Bureau also has several individual triaxial test chambers in which axial loads
may be applied by the compression machine or loading scales . Figs . II - 1,
II - 2 , and II - 3 are photographs and diagrammatic sketches of the small , medi
um , and large machines , respectively . The following tabulation is a summary
of information on the various types of specimens tested in the three machines .
COHESIVE SOILS
Vent- Vocuum
Lateralpressure
goge
-
SPECIMEN
Axiotloodgoge
-
--
SOIL
goges
Droinge soturation
goge
Strain Air
change
goges Pose Supo

UDoly
||
or
Volume
Motor Lood
weighing
Capsule
Airtonk
pressure
Loodpiston goge
TRANSMISSION
TO
no
flow
-
pressure
Drain Cd cells
or transducers
--
-
Fig Small Triaxial Machine

II
.1
-
-
.
PX - D - 16604
FLIEZS6
Vocuum
line
goge
Deviator
air
Toconnection
A Compressed line
forlowloads forBandCla
Similar No- flowpressure goge
Pressure
-
cell
Lateral
Loodweighing pressure
goge
-
Consule
-
goge
Deviator
forhighloods
105
335
- Pressure
balance
Drcinogesaturation
indicotor
Milliometer
goges
)
change
tube
Strain
Vacuum
Constant goges Lood Airfonk or

Volume
rotedevice pressure
J'
piston
goge
weighing
Lood
capsule
LOOD
Jock
Oildrain
Pressure supply
or To
No- Flow
pressure cells
STORAGE pressure
transducers
oil
Fig Medium Triaxial Machine

II
2
.
.
-
-
COHESIVE SOILS
Volume Change , and

Drainage or Satur Testing Machine
ation Gages 60 ,000 Lb . Cap .
- - -
Electronic Recorder
For Measuring Lateral
Load & Pore Pressure
Dial Indicator
Bor
12 Piston Cop
H -Gage
Piston
Ring
-B "O
RetoinerRing ushing
TopPlatea
Specimen
Endplate
Pervious
Disk Soil IL tie Rod
-
Specino
Pressure
Cylinder
Hard Plastic
Tubing
- -
-
6
Water
no
To orTo
Water
Rubber flowcells
Membrane transducers
o drainage
or
E D
saturation gages
'
Brass Bottom Plate

Cock
- -
Gasket-
To
VolumechangeGoge
Fig Large Triaxial Machine

II
3
-
-
.
.
Internal
size of Maximum
pressure Size of particle
chamber specimens size
Machine ( inches )
( inches ) ( inches ) Type of samples tested
( 1 ) Small 3- 1 / 2 x 5 - 1 / 2 1 x 2 - 1 /4 Fine sand Undisturbed specimens cut
from core samples of
3- inch diameter
( 2 ) Small 3 - 1/ 2 x 5 - 1 / 2 1- 3/ 8 x 3 Fine sand Undisturbed specimens cut
from core samples of 5
inch diameter or larger .
( 1) and ( 2) advantageous
for clay soils .
( 3 ) Medium 5 - 5 / 8 x 12 /
2- 5 8 x /
7- 1 4 /
3 16 ( 3) and (4 ) - - Undisturbed
specimens cut from large
(4 ) Medium 5 - 5 / 8 x 12 3- 1/ 4 x 9 3 / 16 blocks . Compacted fine
grained soils
or finer
fraction of coarse
grained soils .
( 5 ) Large 14 - 1 / 4 x 25 - 7 / 8 6 x 15 3/ 16 - 3/ 4 Total 6 - inch diameter
undisturbed cores
( particularly shales ) .
( 6 ) Large 14- 1 / 4 x 25- 7 / 8 9 x 22 - 1 / 2 3/ 4 - 3 ( 5 ) and (6 ) compacted
gravelly soils .
(7 ) - /
5 3 4 x 11 3- 1 /4 x 9 3 / 16 Special chambers available
for special tests such
as long - time drainage or
saturation under load .
All
of the Bureau project testing is performed with these three machines
alone . This is accomplished by keeping the number of shear tests at a mini
mum through careful selection of a few representative samples for testing ,
thus stressing quality testing rather than quantity testing . Secondly , by
measuring pore pressures during the triaxial test and analyzing the tests on
the basis of effective stresses so obtained , there is no need for lengthy drain
age periods and the triaxial test can be performed in a few hours .
2. Basic features . - - The axial load in the small machine is applied by an
electric motor drive through a gear - reducer box which may be set for axial
deformation speeds of 0 . 05 , 0 . 025 , 0 .01 , 0 . 005 , and 0. 0025 inch per minute ; the
last speed corresponds to a rate of strain of only 0. 08 per cent per minute for
a 3 - inch long specimen . The axial load in the medium - size machine is applied
hydraulically by means of high pressure nitrogen over oil . A constant speed
( clock ) device can be set for axial deformation speeds of 0. 1, 0. 05 , 0 . 025 , 0.01 ,
0. 005 , 0. 0025 , and 0. 001 inch per minute . Thus , for a 9 - inch long specimen ,
the rate of strain can be made as small as 0. 011 per cent per minute . The
constant speed device is connected to a hand that is installed on the outside of
the dial gage which measures axial deformation . A hand valve is used to regu
late the flow of oil to the hydraulic axial load piston . By regulating this valve
so that the regular dial gage hand (which measures the movement of the axial
load piston ) is synchronized with the outside dial gage hand (which is con
trolled by the constant speed device ) , the desired rate of strain is achieved .
COHESIVE SOILS 45
Other rates of strain may be obtained by use of a stopwatch to time the axial
movement .
The axial loads for both the small and medium machines are applied from
the bottom of the pressure chamber and are measured by pressure capsules
which are placed between the lower and upper parts of the axial load pistons .
The pressure capsules actuate Bourdon gages on which the applied loads are
measured . The axial load for the large machine is applied at the top of the
chamber by a universal testing machine of 60 , 000 pounds capacity . Axial
loads can be measured by a four - stage dial with total dial capacities of 200 ,
1, 000 , 5 , 000 , and 20 , 000 pounds or by a separate Bourdon gage of 40 , 000
pounds capacity . The Universal testing machine is equipped with stress and
strain rate control devices , stress - strain plotter , and a load maintainer de
vice . In all three machines , the axial load pistons enter the pressure cham
bers through closely machined and greased bushings fitted with O- ring seals .
Sylphon bellows seals and oil pressure seals were formerly used in the small
and medium machines , respectively .
Water is used in the pressure chambers to apply the all - around pressures
to the triaxial shear specimens . The water pressure is controlled by air
pressure applied on the water in calibrated constant - bore glass tubes , which
are connected by copper tubing with the chambers . These tubes are used to
mea sure specimen volume changes during the tests , which is a unique feature
of these machines . Accurate Vernier gages are used to measure the drop
( specimen volume decrease ) or rise ( specimen volume increase ) of water in
the volume - change measuring tubes . Large air tanks connected by copper
tubing to these glass tubes are charged with the desired chamber pressures
prior to the start of the test so that a minimum amount of adjustment of the
chamber pressures is required during testing .
The chambers for the small and medium machines are identical except for
size . The pressure chambers consist of a brass ba se plate , a brass top cap ,
and a brass or plastic cylinder . The bases are an integral part of the body of
the machines , and are fitted with recessed rubber seals into which the cylin
ders fit . Thick , brass cylinders are normally used . The top caps are fitted
with recessed rubber seals , and are equipped with serrated blocks against
which the top specimen end plates bear during the tests . The serrations per
mit the chamber pressures to act on the end plates . In these two machines ,
the caps are held in place by yokes fastened down by turnbuckles. The medi
um size machine is also equipped for a second chamber cap , containing a
pressure capsule and gage for measuring low axial loadings , if desired .
The chamber parts for the large machine (base plate , cylinder , and top
cap ) are made of aluminum to reduce the weight of the assembly . Recessed
rubber seals are used in the base plate and cap in a manner similar to that for
the smaller machines . The assembly is held together with six stainless steel
rods which are threaded into the base and are used to hold the top plate in
place by nuts tightened against the top plate . For all machines , chamber
cylinders of heavy clear plastic are available for use when it is desired to
view the specimens during testing . The metal cylinders are preferred for
general use , however , because the plastic cylinders change volume with re
peated use .
Each container for the triaxial shear specimens consists of two end plates ,
a cylindrical rubber sleeve , and special metal bands , or rubber bands , for
fastening the sleeve to the end plates . The end plates may be solid or of a
type used for pore - pressure measurements . The latter may consist of a water
chamber and a disk perforated with numerous holes of 1 / 32 - inch diameter , or
may be fitted with very fine ceramic disks which have pore diameters of ap
proximately 18 - microns to 0 . 6 - micron , having bubbling pressures of from 2. 5
to 75 psi, respectively . Drainage and saturation may be accomplished through
these end plates . The ceramic end plates are required when negative pore
pressures are to be determined .
Pore pressures may be measured at one or both ends of the specimen , and
facilities are available for inserting a small tube , fitted with porous or ce
ramic stones , into the interior of the specimen . Small bore copper tubing
from the disk end plates (or inserts ) are connected to receptacles in the base
of the machine . These receptacles are , in turn , connected by similar tubing
to no -flow pressure cells or pressure transducers for measuring the pore
pressures occurring within the specimens . The no - flow cells , used since
1939 , consist of a water chamber attached to the pore - pressure lines from the
machine , and an air chamber which is attached by copper tubing to an air sup
ply and accurate Bourdon test gages . A very thin brass diaphragm separates
these two chambers . An electrical contact point is built into the air chamber
directly below the diaphragm . As the pore pressure builds up within the cell ,
the diaphragm deflects to make contact which is recorded by an ammeter .
Air pressure is then applied to the air chamber side of the cell until the con
tact is just broken . Thus , the pressures are balanced and the balancing air
pressure can be read as pore pressure on the Bourdon gage without removal
of a significant amount of pore fluid from the soil specimen . A more recent
and advanced type of equipment for measuring pore pressure is the pressure
transducer . With this equipment , the pore - pressure lines are connected into
the fluid chamber of the transducer . As the pore pressure acts on the trans
ducer , resistance wires are stressed and the change in wire resistance is
mea sured in terms of the pressure applied . A three - stage electrical recorder
is available for measuring pore pressures at 1, 2, or 3 locations within a
shear specimen at 2 - second intervals .
De - aired water is used as the pressure transmission medium from the
specimens to the pore - pressure measuring devices . Unless all entrapped air
is removed from the perforated or porous ceramic end plates , connecting
lines , and water chambers , the measurements will be erroneous . For this
reason , techniques have been developed to eliminate entrapped air and facili
ties are provided for attaching the end plate or insert devices to the pore
pressure line receptacles under water .
Auxiliary facilities to the machines include saturation and drainage volume
tubes which may be attached to the receptacles in the bases of the machines .
These tubes are of constant bore and are equipped with Vernier gages for
accurate reading .
All specimen preparation and testing is performed in temperature
controlled rooms (72° F + 2º) .
B. Test Procedures
1.
.
Basic Procedures . - A minimum of three cylindrical specimens are test
ed at different chamber pressures for each complete triaxial shear test . The
chamber pressures are selected to provide a wide - range of effective stresses .
When partially - saturated embankment soils are being tested , additional satu
rated specimens tested at low or zero chamber pressures are often included
in the series to obtain saturated strength data for the soil .
COHESIVE SOILS
The procedure for preparing undisturbed soil specimens for the small and
medium machine tests is similar . The specimens are cut from undisturbed
samples , as shown in Fig . II - 4, by means of a cutting bit on the lower end of a
slightly larger specimen container . The container and bit are supported by a
frame and a rod to which slight pressure is applied manually . The excess soil
is trimmed with a knife leaving very little soil to be shaved off by the cutting
bit as it passes down and over the specimen . When sufficient length is cut ,
the specimen is placed in a 3 - section mold which has an inside diameter of
exactly the required specimen diameter and height. The ends are trimmed to
provide the correct specimen height. By following these procedures with
great care , disturbance to the specimen during preparation is minimized .
Care is always taken to remove disturbed soil from the surfaces of samples
prior to cutting specimens from them . Because of the disturbance commonly
found on core samples , the normal practice in the Bureau is to use specimens
smaller in diameter than the core sample . After the specimen ends have been
trimmed , a solid end plate is placed on each end to protect the edges of the
specimen and the rubber sleeve is placed around the specimen and end plates ,
using a vacuum expander to enlarge the sleeve .
Perforated end plates are stored in a water bath and with the cock open ,
water is forced through the tube from the cock end by means of a syringe .
After complete filling is assured , the cock is closed so that the water cannot
escape . When end plates with ceramic disks are used , the procedure is the
same except that the assembly is boiled in water to remove air . A perforated
or ceramic - type end plate is removed from the water bath , free water is
carefully removed and the specimen immediately placed on the perforated or
ceramic end plate just as the solid plate is removed . Extreme care must be
exercised when placing the perforated or ceramic end plate on the specimen
to assure that an intimate contact , without free water , is obtained at the two
surfaces .
When a ceramic end plate is used to measure negative pore pressures , a
complete and extremely intimate contact is necessary . This requires that the
special techniques that have been developed be followed closely . After the
end plates are in place , the rubber sleeve is sealed to them with metal clamps
or rubber bands . In some studies , perforated or ceramic end plates are
placed at both ends , or special ceramic or pervious inserts are placed at in
terior points within the specimens . Two perforated end plates are used when
specimen saturation or drainage is required . Filter paper strips placed
lengthwise between the specimen and rubber sleeve are sometimes used to
speed up pore pressure equalization (or drainage ) when testing very plastic
clays .
Six - inch cores of very hard soil formations , such as shales , may be tested
as full 6 - inch - diameter specimens in the large machine . No trimming is done
to the sides of these specimens and only the ends are trimmed to proper
length . Otherwise , the procedure is essentially the same as described for the
smaller undi sturbed specimens .
In preparing compacted specimens for the medium and large machines , a
hinged 3 - section specimen container is assembled with an oiled - paper or
plastic liner and clamped around a solid bottom end plate . The soil to be used
is wetted to the desired moisture and allowed to sea son until the moisture is
completely distributed throughout all the particles . Sufficient quantity of ma
terial is accurately weighed to provide the wet density for the volume of the
specimen being prepared . The weighed sample is then divided into nine equal
E- 945- 7 PX - D - 20333
Fig . II - 4 . -- Cutting Undisturbed Soil Specimens
Sleeve Compaction - Compaction

Rubber expander mold - hammer
membrane
Porous - Moldau - Collar i

end plate liner
Fig . II - 5 . -- Equipment for Compacting Specimens

COHESIVE SOILS 49
parts , care being taken to avoid any particle segregation . Each part is tamped
into the mold by a drophammer to a depth one -ninth of the total height , which
will give the desired density . After the mold has been tamped full and
trimmed flush with the top of the mold , the specimen container , (rubber
sleeve , solid bottom end plate , and top perforated or ceramic end plate ) is
placed on the specimen in the same manner as described for the undisturbed
specimens . Fig . II - 5 shows the equipment used for the specimen preparation
and container assembly for 1- inch and 3 - 1 / 4 - inch - diameter specimens .
When cohesionless soils are compacted for shear testing , facilities are
available for placing the rubber sleeve directly in the mold and compacting the
soil into the sleeve . The sleeve is clamped to the end plates while the speci
men is still in the containerand a vacuum is applied through one perforated
end plate to support the specimen while it is being handled during volume
measurements and setting up in the testing machine .
After the assembly of the specimen container on either an undisturbed or a
compacted specimen , it is weighed in air and in water for a total volume de
termination . Since the volume of the container assembly has been previously
determined , the volume of the soil specimen can be determined . Wire
carriages are provided for facilitating weighing in air and in water , as shown
in Fig . II - 6. During the weighing - in - water period , all air bubbles are care
fully removed from the assembly by using a syringe to force the air from the
parts . A period of 15 minutes in the water is provided so that small volume
changes , which occur due to the small water load , can be properly evaluated .
The following description of test procedures for installation and load appli
cation applies to both the small and medium machines and covers the most
commonly used method of performing the test . Exceptions are noted . After
weighing the specimen , the water pan at the base of the machine is filled and
the lower cylinder gasket is removed and replaced in a manner so that all en
trapped air is removed . The specimen is then centered on the axial load
piston and the end - plate tube cock connected under water to the pore - pressure
tube receptacles in the machine base . The cock is opened and the sealing cap
replaced .
At this point , the specimen is in a completely sealed condition . Next, the
pressure chamber cylinder is placed on the base plate gasket and completely
filled with de - aired water at room temperature , care being taken to remove
any entrapped air during the filling . To accomplish this , a syringe is used to
force air bubbles from the specimen container and sides of the cylinder . The
chamber cap , which is stored in a water bath , is then placed on top of the
cylinder with the axial load piston set so that about 0.025 - inch clearance exists
between the specimen end plate and the cap when clamped tight . As the yoke
clamp is tightened , excess water is drained from the volume - change gage tube
to a reading of 1. 00 (near the top of the tube ) when the clamp is tight . After
this operation is completed , the pan at the base of the machine is drained and
the outside of the pressure chamber is wiped dry to prevent cooling by evapo
ration .
After making and recording initial readings of all gages and the pore
pressure cell, or transducer , the desired chamber pressure is applied through
the volume - change tubes . The specimen is allowed to change in volume for at
least 15 minutes or until the volume and pore - pressure readings stabilize .
Volume and pore - pressure readings are taken at periodic intervals . When
volume and pore - pressure equilibrium have been reached , the piston is care
fully raised until the specimen top end plate is barely touching the cylinder
cap . As the piston is raised , the axial load gage is read . The load measured
09
SHEAR STRENGTH
Dort
CONFERENCE
Die
11
Fig
–
.
.6-
Facilities for Determining Specimen Weights and Volumes
COHESIVE SOILS 51
is the chamber - pressure load on the piston plus any piston friction under the
particular chamber pressure being used . This reading is used as a zero point
for subsequent deviator load readings . Contact between the top end plate and
chamber cap is detected by a slight increase in the axial load gage . Readings
of the axial strain gage , volume gage , and pore - pressure measuring device
are taken , after which the water is drained from the volume gage to give a
reading of 30 . 00 (at the bottom of the tube ) . Air is admitted to the tube to
maintain the chamber pressure .
A rate of strain , for the type of soil being tested , is estimated and the axial
loading is initially started at this rate of strain . The rate of strain may be
changed during the early part of the test if the pore - pressure and volume
change measurements are not compatible , as explained later . As the axial
loading is applied , readings of the axial load gage , axial strain gage , volume
gage , and pore - pressure devices are taken at periodic intervals , usually at
about 1 per cent axial strain . A buzzer mounted on the machine, actuated
through contacts on one of the strain dial gages, provides a signal for reading
all of the gages at the predetermined interval . If the first volume gage tube
becomes full during the test , it is closed and a second gage tube is opened for
continued measurements , with the chamber pressure held constant .
Failure of the specimen , based on the maximum deviator stress (01 - 03 ) ,
may be indicated by an axial load gage reading which is less , equal to , or even
greater than the preceding reading . The first two states are easily recogniz
able ; the third is interpreted by means of a prepared table of the minimum
axial gage increase necessary to maintain a constant axial unit stress in the
specimen for a given load and deformation . Failure of the specimen on the
basis of the maximum effective principle stress ratio (01 / 03 ) is determined
by making computations of the unit effective stress values and computing the
ratios as the test is being performed . This is usually recognized by a drop in
pore pressure. Following the failure of the specimen , the load piston is
lowered and the chamber pressure is reduced to atmospheric to permit ex
pansion of the specimen , after which final axial load gage , volume gage , and
pore - pressure readings are taken . The failed specimen is then weighed in
water and air for final weight and volume measurements . The final weight de
termination provides a check on leakage of sealed specimens during the test .
Shear specimens are tested in the large machine in a manner similar to
that described above for the smaller machines ; however , differences in
equipment require some procedural differences which are briefly described
as follows . Each shear specimen is assembled in the same manner and initial
weight and volume measurements are made , as described for the smaller
specimens . The specimen , in its container , is then centered on the bottom
base plate of the pressure chamber which has been placed in a water - filled
container on a work bench near the machine . There , the pore - pressure or
drainage - saturation tube connections are made to the proper receptacle in the
base . The pressure cylinder is then placed on the base ga sket , and the top
plate placed on the cylinder and the nuts are tightened on the rods over the top
plate to hold the chamber together . A torque wrench is used to tighten the
nuts to a constant torque so that a constant chamber volume will be obtained .
The chamber is then filled with water allowing air bubbles to escape through
the piston bushing . The axial load piston is greased and placed through the
piston bushing in the top plate against a bearing block previously centered on
the top specimen end plate . The piston has a convex point on the lower end
which fits into a concave surface on the bearing block to allow some rotation
of the specimen top with relation to the piston . The entire assembly is then
placed in the Universal compression machine .
2 . Data and Computations. -- The data and computation sheets used for the
triaxial shear tests are given on Figs . II - 7, II - 8 , and II - 9. The example data
shown are for tests conducted with the medium - size machine . Fig . II - 7 is the
Specimen Placement Data Sheet which is used for recording the initial and
final weights and volumes of the specimen from which the initial and final
densities , void - ratios , moistures , degrees of saturation , void volumes , air
volumes , water volumes , soil volumes, and related data can be determined on
the form . Columns 1, 2, 4, 6, 7, 8, and 10 of Fig . II - 8 are used for recording
test measurements of time, axial load gage, axial strain dial, volume change
gage , drain tube gage ( if used ) , and pore pressure . Columns 3, 5, and 9 of
Fig . II - 8 and the columns of Fig . II - 9 are used for computing all of the gage
readings in terms of desired values. The use of these data and computation
sheets is explained in the Bureau ' s Earth Manual . By recording the test
values and computing the data on these forms , the computations required can
be easily and accurately performed in a logical manner and all correction
factors included .
3. Variations and flexibility of procedures . - Many variations in the basic
test procedure are required to obtain data which will be of value for appli
cation to specific stability problems . Like other soil testing work , consider
ation must be given to the effect of natural , construction , and operation con
ditions upon the soil property under study . For instance , the effects of future
saturation and loading are important considerations . The sequence of loading ,
and the amount of drainage that takes place at any time in a foundation or
earth structure must also be given consideration in the test procedure and
data evaluation . As it is not possible to duplicate the prototype conditions with
respect to time , it has been the practice of the Bureau to obtain shear test
data on the basis of effective stresses , rather than upon undrained and drained
tests wherein only applied stresses , or completely drained conditions , are
considered . The shear data , based on effective stresses , are then used with
contemplated effective stresses that will exist in a foundation or earth
structure at important or critical times during the construction or operation
of the structure .
The basic procedure described in detail previously is that in which
partially - saturated specimens are tested at constant chamber pressure
( constant 03 ) in a sealed condition with pore - pressure measurements being
taken at the end of the specimens. This type of test is made for shear data on
soils proposed for embankment construction purposes , and is called a “ sealed "
test . The initial density and moisture conditions for a suite of compacted
specimens are usually those similar to anticipated field placement conditions .
A suite of at least three , and up to six specimens are compacted as near
identical conditions as possible . The specimens in a suite are normally tested
under 3 . 1, 6 . 2 , 12 . 5 , 25 . 0 , 50 . 0 , 75 . 0 , or 100 . 0 pounds per square inch chamber
pressure (other pressures up to 200 psi may be used ) . The pressures used
are determined to give the effective stress range desired . The number of
specimens used depends upon the detail which is desired from the failure en
velope . When details of the saturated - cohesion value are desired , one or two
specimens are saturated after compaction by submersion in water while in a
porous container or by permeating water from the bottom to top perforated
end plates , after which they are sheared at low ( 0 to 3. 1 psi) chamber
pressures .
COHESIVE SOILS
EARTH TESTING
SPECIMEN PLACEMENT - DATA SHEET I
Sweet
. )
die
(FOR TRIAXIAL SHEAR AND THREE DIMENSIONAL TESTS ) SHEET_
op
L
1 TESTS
COMPACTION
98
00
ND
ONO
T8
Vollecito Dom
%
SAMPLE FEATURE
_
-
. . ._
35 16
SPECINENN 114. Pct DATE Nov 1947
_
2
,
.
NO
15
%
CONTAINER MOISTURE
_
_
/
.
INITIAL FINAL WETTED
3856 3856
IN
WEIGHT
OFSPECINENCONTAINER AIR
.1
1
(1
.
)
1279
OF
1280
3
N
WEIGHT CONTAINERI AIR
.0
.
(2
_
)
-
INITIAL FINAL
RUBBER 60 60
4
.6
.
END
PLATE 754 754

.6
.6
OF
(3 11
) 12
WETWEIGHT SPECIMEN 2576 2576

=
8
(3
.
./
)
)
-
OF
ORYWEIGHT SPECIMEN 2236 2236

.2
.2
=
+
(4
1
)
-
-
-
-
-
-
-
-
-
F
23779
IN
WEIGHTO SPECIMENCONT WATER- 2376

+
.8
(5
.
)
OF
VT
61
11
51
1479 1478
.3
LOSS SPEC CONT

2
+
.=
.
(
)
(
-
.
257 257
OF OF
5,
.5
WEIGHTLOSS CONTAINER
(7
)
SPECIMEN 171 1221. 1220

=
LOSS
,
8
WEIGHT
(6
__
(3
)
-
)
728
74
74
SPECIMEN
VOLUME 661
cu
=
.
(9
in
16
)
350
(
.)
.
.
-
-
-
-
-
-
-
15 114 -
20 01 -
-
-
-
-
15 114-
11 -
-
-
29
) 62
LO DRYDENSITY=
(4
( x
.
)
.
)
100
23
.
10
x 13
.
, %
MOISTURE DRYWT
G =
1
.
-
x
(
)
)
]
Sp
12) ABS SOILDENS 62 43 171058

.=
. .
(
SOIL .
. 66
53 82 65
115 )14 113
100
10
12
,
VOLUME
)
= ((
100 10 ))
27
VATER
VOLUME
(
62 29
- .
VOL
113
114
AIR
5
.
FREE
=
-
.
)
[
)
500
(16
VOIDRATIO 100
=
)
) 100
42
83
18
OF
17
.
=
DEGREE SATURATION 100-

13
__
(x
)
)
(
(
)
(
-
-
-
pct -
-
-
-
-
-
-
-
-
-
--
WEIGHED MEASURED
02
cu
07
19
18
in
CHANGE
0
DENSITY VOLUNE
.
CONSOLIDATED
-
.6
.)
(
.
/
)
(
19
BY 19
cu
pcr DRAINAGE
in
DENSITY
FAILURE
.8
(
)
.)
.
KC
Y
BY
PREPARED COMPUTED CHECKEDB

N
.A
.C
.0
K
_
_
_
.
_
Fig Data Sheet No Specimen Placement Data

II
7
1
-
-
-
.
.
EARTH TESTING
TRIAXIAL SHEAR TEST , DATA SHEET 2
SHEET2 or 3
PROJECT_ Pine River - FEATURE Vollecito Dam

- NO
SAMPLE . 00 - 18
NO.
SPECIMEN 1 BY__
RECORDED K. C. _
,
DATE_ Nov . 2 1947
100. 0 CONSTANT
APPLIEDLATERALPRESSURE .050 IN./NIN
STRAINRATE . 720F, FINALTENP
INITIALTEMP , 7170F
TIME AXIALLOAD CORRECTED AXIALSTRAIN VOLUME

CHANGE DRAINTUBE PORE
GAGE
ACTUAL READING DEVIATOR PRESSURE
MIN. READING DIAL STRAIN TUBE TUBE2 READING VOLUME
1
)
'
(
psi psi psi
(9 cu
inches inches inches
in
1
,
i
10
(4)
(5
(8
(3
(2
(6
)
)
(
)
)
)
)
O 075 00
0
6
24 26 27 29 30 10 12 12 11 1
12 7
.4
.
. . . . . . . . . .
.
29 95
.075 48
0 .8 6
.6 6
. 52 .
.
Form EL 562
. 12
075
18 16 15 5
561
-
.5 55
12
075 43
. .2 .13 6
561
.
00 06
. 55
>
El
5
*
224 149
2
. 2..
13
55
224
Form
El
6
5
-
19
. 17
10
076 69
3
3
9
Form
.7
5
. . 0
.
.
Col
22
89 36 80
-
76
19
–
100
6
4
4
3
5
2 5
i
.
.
Col
Col
. 81
210 78 Oz
83
21
=
8
7
5
9
16
.
og
.
-1
22 23 39 84
6 .7 3
.6
10 10 9
6 0 4
15 17 18 20 21 .
84 .
. . .
16 15 15 15
56
82 35 89
23 706
3
3
2
. .5 .61 .4 .7
|
.
|
. . .4 .
| || | || ||
. . . . .
.
.84
23 59 857
3
.8
.
243 862 84
4
5 9
8 5 1
**
*
.
!
.0 . . .
11
26 26 26 25 25 24
27 63 848 83
.6
3
3 8 5 1 .6 2 .8
82 82
78
120 64 172 767

3
8
.
.
12 12
. . 14
25
4 3 1 8 5 2 / .6 .3 9 4
.
at
Point which
12
69 81
*
*
2
.
. .
.
.14 . . . 13 13 .
of
maximum value
11
14
80
Õlo occurs
4
.
,
57 79
7 9 9 .8
9
. .
.78 78
27
14 14 14
797
.
79 39
274
6
.
27
.77
7
. .4
,
28 19 77
3
0
0 7 4 0
. . .
.
.6 . . 15
00
28 30
60
28 76
3
. .
.
.
.75 75
15 15 15
28 44
5
.
28 26 88
.6
25 .
21 34
28 73
8
8
.
.
.
21
.4
3
1
.30
-
.
which piston lowered ofter failure

to
Point
*
*
*
SPECIMEN
TOPENDPLATE JUST TOUCHING
CYLINDERCAP
*
No
Fig Test Readings

of
Data Sheet Record

II
8
2
-
-
--
.
.
6 1 9
7 6 5 2 9 8 4 8 7
3 2 o s 4
ol
! 51 1 :
12 5 2 1 +
- :.
. 11
, 10 9
. 89 6 6 4 3 3 3 . 3
. . 8 .
75
. . . . . . . 3 . 3. 2. . 1. . . . . .
1155
02
777
99 47 74 00
PROJECT
722
91 86
67
_
93 39 84 13 59 06 53 98 44 89 35 82 30 78 32 00 09
301
SPECIMEN
|
NO.
9 10 10 9 9 8 7 6 6 5 5 4 4 3 3 2 2 - - (
. 11 . .
. 11 . . . . . . . . . . . . . . . 1
. .
Fig
12
)
69
95 41
744
82
15 61 07 90 36
49
. 07 77 23 53 99 28 74 20 66 12 58 03
Pine
GAGE PISTON
811
| |
II + + + + +
- -. . . . . . . . . . . (
9 . . . . . . . . . .
02
02
.
03
02
03
03
03
01 32 32 32 32 .
131
03
32 03
03
03 03 03 03 03 03 03 03 02 02 02
River
-
| |
+ -
cubicinches
-
-
3
. . . . . . . . 3
. 3. 3
. 3
.
. . . . . .
59
54
60 19 18 19 19 20 20 19 19 20 20 20
19 18 19 21 21 19 17 15 08 58 42
201
| | | | | |
CHANGE CHANGE CORR CHANGE
TRIAXIAL
1970
70 70 70 70 70 70 70 70 70 70 70 70 70 70 71 71 71 74 (
74 70 70 70 .
. . . . . . . 70 . 70
. . . . . . . . . . . . . . . 71
. .71 71
. . 15
96 )
00
73
07
75 96 97 96 96 96 95 95 96 96 95 95 96 97 95 94 94 96 98 15 14 19 31
951
Data Sheet No
. 6 . 7 . 8 8
. . .6 . . .6 .7 .8 . (
9 .9 .4 .5
8
.7 .2 . 4
.5 8
3
70 /
.80 2 3 16 in
VOLUME SPECIMEN
3 ) .
FEATURE
_
876
-
| | | | | | |
VOLUME LENGTH
110
. 9 9 9 8 8 8 8 8 8 8 ( sq
10 .10 .10 9
. 9. . . 9 . 9. 9. . 8. . 8 . 8. . . . . . . 17 .
) in
65
EARTH TESTING
59 44 28 14 99 85 72 59 46 34 21 10 98 87 76 55 45 35 25 16 08 .
COHESIVE SOILS
AREA
DATE
MEAN
| | | | | | __
FEATURE Vollecito
626
689
538
582
422
18 Ib
485
456
732
684
703
645
563
601
718
660
742
189
378
)
514.
310.
757
747
. . . . . . . . . . . . . . . . .. .
Computation
| | | 1 T | / | Dom
SHEAR TEST , DATA SHEET
Nor
64
_
72
68
56 38
69
45
psi
71
3
of . 71
.6 . .72 .71 71 . .
. .70 71 .69 . .66 . .63 61 .59 . .54 50 .
5 2 9 4 9 3 8 1 0 0 8 5 .4 0 5. .8 . .23
4 2
41
71
1 _
01
11911
|
LOAD STRESS
222
24 18 16 15 14 13 12 11 9 8 6 5 4 3 1 (
. . 21
. 23 . 20
. . 17. . . . . . . . . . . . . . . . 20 %
)
49
1947
00
63
52 39 26 14 01 88 76 50 38 25 12 87 74 36 24 11 98 86 00
7621
! | ! |
SAMPLE
-
SAMPLE
Test Data
5 5 5 5
5 5 5 5 4 4 - ( NO
NO
. . 5 5 5 5 5 5 5 5 5 %
. . . . . . . . . 5 . . . 5
. . . . . . . .4 21 .
)
503
07
04
07
06
04
04
99 90
02
04 04 06 06 04 04 06 06 06 03 04 04 79 00
DEVIATOR AXIAL VOLUME
( -
18
22)
SHEET3 of 3
STRAIN1 CHANGEDRAINAGE
Since 1954 , many of the triaxial shear tests have been performed on im
pervious materials proposed for large earth dams ( 1) with a constant ef
fective lateral pressure (03 ) , and (2) with the shear action starting with all
specimens in any one suite being at the same density (Gibbs and Hilf , 1957 ) .
This type of effective lateral pressure control ( 1) eliminates prestressing of
the shear specimens and the failure point based on maximum 01 / 03 or maxi -
mum 01 - 03 are identical . The tests with uniform density at the start of
shear stress application (2 ) provide somewhat lower values of 0 than those ob
tained in the constant placement density test . This occurs because of the
difference in the range of void ratios at the start of the test and at failure for
each specimen .
The constant effective lateral pressure test ( 1) is achieved by keeping the
chamber pressure minus the measured pore pressure at a constant value . Im
mediate , accurate pore - pressure measurements are extremely important .
The test is conducted by connecting one pressure transducer into the chamber
water system and one pressure transducer to a ceramic or perforated end
plate . Thus , the electrical recorder registers the chamber pressure and the
pore pressure on a graph throughout the shear test. As the pore pressure in
creases or decreases during shear action , the chamber pressure is increased
or decreased manually to keep the difference between the two constant . A set
of dividers with points fixed at this constant distance on the graph is used to
facilitate holding the distance constant .
Constant density at start of shear loading is achieved by the following pro
cedure . The specimens for any one shear test are all compacted at the antici
pated field placement moisture content and at various densities . The high
density specimen is compacted to just slightly below the anticipated field
placement density and the low density specimen considerably below this , with
the remaining specimens compacted to intermediate densities . The high
density specimen is placed in the shear machine in a sealed condition and suf
ficient chamber pressure is applied to produce a specimen density equal to
the anticipated field placement density . As the specimen placement density is
only slightly lower than the anticipated field placement density , only a low
chamber pressure is required to produce the desired densification . The
chamber and pore pressures are measured and the shear action is started by
applying the axial load . At the same time, the effective lateral pressure is
kept constant in the manner previously described . The next lower density
specimen is then tested and an applied chamber pressure higher than that used
for the first specimen is required to produce a density equal to the anticipated
field placement density . The shear test is again performed , keeping the ef
fective lateral pressure so developed at a constant value . By repeating the
test procedure for all of the specimens which were compacted to consecutively
lower densities , tests with a range of constant effective lateral pressures are
obtained . The above procedure has been found satisfactory for production
type work . Experienced , high - caliber technicians can readily estimate the
specimen placement densities required to produce the lateral pressure spread
desired and re - runs are seldom necessary .
With undisturbed specimen tests , the initial specimen densities may vary
somewhat . When the undisturbed specimens have a low degree of saturation ,
the spread of effective stresses , as obtained by tests made at low to high
chamber pressures on sealed specimens are sufficient to provide adequate
shear test data . If the degree of saturation of the undisturbed specimens is
high , the spread of effective stresses using sealed specimens is so small , even
COHESIVE SOILS 57
with a large
range of chamber pressures , that the shear test data is of limited
value . In case , drainage is permitted during the consolidation
this period .
After this is completed , the drainage is discontinued ( the specimen is sealed ) ,
and the lines de - aired , prior to the application of the axial load . Thus , the
specimens in any one suite are drained under different chamber pressures
and a satisfactory spread of effective stresses is achieved . For highly im
pervious clay soils , the smallest size specimens are used to reduce the drain
age time . Drainage usually is permitted at both ends of the specimen through
the perforated end plates and any remaining pore pressure is given time to
equalize throughout the specimen prior to applying the axial load . This type
of test is called the " consolidated - sealed ” test .
The “ sealed ” and “ consolidated - sealed " tests are similar to the commonly
known " unconsolidated - undrained " and " consolidated - undrained " tests , re
spectively , (ASTM Designation 653 - 58T ) , except that the speed of testing may
not be as " quick " as that normally associated with the latter definitions . In
addition , the “ sealed " test specimens are allowed fully to decrease in volume
in their sealed condition under the chamber pressure prior to the application
of shear stresses .
The rate of strain for the various types of shear tests conducted is not
the time required for pore pressure equaliz
fit
constant but is regulated to
-
ation throughout the shear specimen The pore pressures developed within
a a
.
sandy soil equalize throughout the specimen quickly but pore pressures
in
,
very fine grained highly plastic soil require additional time for
of
specimen
,
a
equalization pore pressures large specimen

of
Also the equalization

in
a
,
.
takes longer time than for that small specimen However even fine
in
a
,
.
soils and large specimens may be tested relatively rapid rates within
at
a
,
working day using the sealed test with effective stress measurements be
,
,
of
appreciable flow
re
pressure with fluid
no
cause only equalization
is
an
of
quired
.
axial strain which may be ap

of
method determine the maximum rate

to
A
plied soil specimen with assurance accurate pore pressure measure

of
to
a
continuous plot the pore pressures

is
vs
of
ments maintain the volume

to
changes from observations taken periodic intervals throughout the test

at
Ellis and Holtz 1959 comparison can then be made continuously with
A
a
,
(
)
.
family curves which show the basic theoretical relationships volume

of
of
change and pore pressure

.
The relationship between pore pressure and volume change soil mass
in
a
under stress Part this paper The values for the factors
is
III
of
described
in
of Equation III 11 Part III are obtained from very accurate measurements of
,
,
-
weight volume soil specific gravity and moisture content for each specimen
,
prior starting the shear action and the accurate volume change measure
to
ments taken during the application of all stresses Procedures for these
.
measurements were previously described

.
The family pore air pressure vs volume change curves which are com
of
puted from Equation for various initial air contents are shown as dashed
11
III
-
Figs The paths of pore pressure

10
12
on
lines Part
of
11 and
II
II
,
II
c
(
)
.
-
vs volume change values obtained during shear testing are shown as solid
.
of
these plots there were no surface tension effect the path

on
lines
If
,
.
measured pore pressure vs volume change values should fall along the chart
.
pattern for the true air content the specimen being tested These paths
of
of
.
mea surements solid lines are generally good agreement with the theoreti
in
(
cal curves da shed lines allowing for surface tension except when the rate
of
),
(
. Foilure Max ailos Shear Values
Specimen
Initial Specimen Dato
at
Test Values
.
Corrected for
Dry Degree Applied Effective Volume Deviator
Pore Pressure
No
Density Water Satu- Loterat Loteral Change Stress
of
| |
per content ration Press si Presspsi ercent psi
(Ib
Cohesion
(
.)
.)
.)
per cent percent)
.7 o (.
.
.(p
(p
si
of
Initiall
)
(p
76 76
99. 53
.3 2
.0 2 .3
52
.15.9 .7
6
.5 .2
. 12
.4 15
-
.1 45
76. 99. 22
.8 4
.5
6311
45 .0
26
-
99. 92
2
5
0
.7
.
, WILLARD DAM
STRESS PSI
Lean Clay CL
. (
)
: : Og
No
Constont Sample 162- 167
Max gro Undisturbed
Yea
01
03
To Max. Specimen Size
+
-
1
2
"
"
PercentGravelo
AR PercentSand
2
SHE
LiquidLimit 45
PlasticityIndex22
ConsolidatedSealed
-
12
16
20
24 32
:
NORMAL STRESS PSI

,
,)
psi ST
1
%
Mayon
av-
AV
:(
de
.V
2AV
Sad
P
van
12
oh
--+
PORE PRESSURE
PORE PRESSUR
Strain 002s per min

or
% of
Rate
in
0
,
:
.
11
InitialSpecimen
lengthpermin
of
0
.
VOLUME STRAIN CONSOLIDATION

AVI PER CENT
,
,
(
Fig Typical Shear Test Data Lean Clay Soil

10
Undisturbed
,
II
-
-
.
Small Specimens
03
Constant
,
COHESIVE SOILS
03
Initial Specimen Data Shear Values
Specimen
Foilure Mox 011o3Max
O
of
Test Value
)
-
.
(
Corrected for
i
Dry Density Degree Pore Elfective Volume Deviotor Pore Pressure
per cu ft
i
Ib
Waterlof Sotu- Pressure Lateral Change Stress
(
)
|
Content ration psi Press. si er cent psi
į
Place Consoli percent) per Tan Cohesion
(
1 7 .8 9 .)
(
.)
psi
(p
(p
59 01
-
- - -5 - of
cent) Initial)
1-
(
ment dated 03
(
)
-
21
95. 959 25 08 49.
11.7
3
.9 .0 2
6
91
.4
. 4 7 .3
.
94. 96. 247 85 41.
.8
32
943 96. 24 84 173 40. 59. 15L 18.
06
.8 5
0
ON
.
95 25. 86 31 55.
2
3
.6
.8
L
2
.
..
SO
og
Constant
03
,
MaxovozMax
=
,
X
-
EMIGRANT DAM
STRESS PSI
Sample No. 18P X156
-
CH
Fat Clay
(
)
Compacted
SpecimenSize
Per cent Grovelo
AR Per cent Sond68
SHE
Liquid Limit 69
Picsticity Index48
Sealed
10
20
30
40
50
60
70
80
NORMAL STRESS PSI
,
.5
,)
Voth
12
%
Nothvwl
PSI
TV
-a
:D(
de
TO
A
2
VNU
von
PORE PRESSURE
94
.
c12e
13
140
1150
Stroin 025 per min
of
Rote
orin
1 0
:
offor
&.
28
Specimene
SE
%
2
0
.
InitialSpecimenlengthper
Etter min. Specimen 0025i per

n
=
of6
0
.
Initial Spec
or
min 03
0
.
imenlengthpermin.
12 1
10
11
13
9
AV
VOLUMESTRAIN CONSOLIDATION PER CENT

,
),
(
Fig Typical Shear Test Data Compacted Fat Clay Soil

11
II
,
-
--
.
Medium Size Specimens

03
Constant
. Initial Specimen Dato Moxies
Specimen
Shear Values
at
Test Values Failure Max
.7
0
,
,
,
-
(
Dry Density Corrected for
Degree Pore Effective Volume Deviator Pore Pressure
Water
No
cu
er Satu- Pressure Lateral Change Stress
of
ib
ft
Content
.
).
Press. si er cent ipsi
.p
Place Consoli- ercent) ration si Cohesion
03)
Tan
)
)
01
si
(p
3 (p
er cent)
(p
-
of
Initia
(p
ment doted
)
)
(p
(p
98. 19 82
.4 1 .2
.0 0
.9
.5 8 .0
-
6 85
.9 10
98. 100 18. 52. SO 194
.9 .4
02
.8
.6 0
- -
36
98. 110011 18 79 147
2
2
L
_
I
, TWITCHELL VAQUERO DAM
)
(
SHEAR STRESS PSI
Silty Grovel GM
(
)
SampleNo IOF X116
-
.
Compacted
oz
Constant Large Scale TriaxialShear
83
73
Mox Mox SpecimenSize x22
,
7,
X
.0
1
/2
32 9
"
-
"
/
%
Per centGrovel34
%
PercentSand
30
LiquidLimit
PlasticityIndex
5
Sealed
120
160
20
40
60
80
100 140 180

o
NORMAL STRESS PSI

,
)%(=
F2
) .
5
2 mauront
PSI
PS
=
(
.-n+
T
AV
%
VorhVw
1
401
Vw 30
:P(
12
Va
--
PORE PRESSURE
.
24
out
TH 12
'O
D1
15
2 .
IV
x
Stroin 050 perminor

of
in
Rote
=
0
.
22
Initial Specimen
lengthpermin
of
%
0
.
13
10
11
12
5
,8
9
2
3
1
AV
'),
VOLUME STRAIN CONSOLIDATION PER CENT

(
Fig Typical Shear Compacted Silty Gravel Soil

12
Test Data
,
II
-
.-
-
-
.
Large Specimen
73
Constant
,
COHESIVE SOILS
strain is too great or when leaks in the specimen container occur . For
instance , Specimen 5 of Sample 18P - X156 ( Fig . II - 11) was sheared too rapidly
as is readily apparent on Part ( c ) of the figure . The inconsistent shear value
for this specimen is also apparent on the shear stress - normal stress plot,
Part (b ) . The test data for Specimens 1 and 2 show good agreement with theo
retical values ( Part ( c)) but it is evident that Specimen 6 was tested at slightly
too great a speed during the latter part of the test . Specimens 1, 2 , and 5
were tested at a constant axial strain rate of 0. 025 inch per minute (0 . 28 per
cent axial strain per minute - based on initial specimen length ) and Specimen 6
was tested at a constant rate of 0 . 0025 inch per minute (0 . 03 per cent per
minute ) . It has been noted that impervious specimens sheared at low effective
lateral pressures , such as Specimens 5 and 6 , usually require lower strain
rates for accurate pore -pressure measurements because total consolidation
is usually low and some expansion often occurs , with a lowering of pore
pressure , as failure begins to take place .
One might inquire why it is necessary to measure pore pressures if they
can be computed from Equation III - 11 on the basis of initial soil specimen
data . It has not been possible to make the initial moisture , density , volume ,
and particularly specific gravity measurements with sufficient accuracy to
make precise determinations of pore pressure by computation . This is par
ticularly true when the specimens have high degrees of saturation and corre
spondingly low air contents because the air content value becomes critical.
Negative pore pressures from capillary and other tensile forces exist in parti
ally saturated specimens at the start of the shear tests . As the values of
these negative forces are usually unknown , they cannot be accounted for in the
computation .
conditions , however , do not affect the usefulness of the charts
These
Part (c) of Figs .
II - 9, II - 10 , and II - 11 for checking pore pressures measured
during triaxial shear testing . The measured initial pore pressure values can
be recorded on the chart, and subsequent values should then fall approximate
be
air
ly along the chart pattern for the

of
content the specimen being tested

,
ing parallel curve computed for the air content nearest that the
to
of
to
a
specimen
.
Reporting Data The test values obtained during the shear test may
of
4
.
.
-
be reported several ways provide the data needed for the particular
to
in
problem being studied As the purpose provide data

on
shear test
of
is
to
a
.
the strength properties for use necessary

of
the design structure

is
to
in
it
,
a
report the data that they will be usable for this purpose This does not
so
merely mean report and value for each soil tested but furnish
to
to
,
o
a
“
c
“
”
"
all the data required This in

of
understand the soil strength completely

to
initial and final specimen conditions how the test was

on
cludes information
,
on
performed comments the soil behavior and information stress strain

on
,
,
,
volume change and pore pressure conditions throughout the test particularly
,
,
The data presentations shown Figs

11
10
on
as failure approached
is
II
II
,
,
-
.
-
.
show how these data are often presented Plots axial strain vs
12
of
and
II
on .
.
-
axial stress pore pressure and volume change for these tests are shown
,
,
Fig The soil behavior during shear terms the other

is
of
13 discussed
in
II
,
.
.
soil properties the narrative portion the report which includes the data
in
of
,
plots
.
Reliability of Results
--
.
C
.
Variations Soil Specimens With the techniques developed compact

in
,
1
.
.
specimens can be prepared with small variations initial density and

ed
in
.
62
T
Failure
Point WILLARD DAM EMIGRANTDAM VAQUERODAM
.
-
.
Sample162167 Sample18P
X156 .
No
Somple10FX116
No
No
o
-.
Specimen DomSpecimen Noi
Specimen
2
.D
DoelSpecimen Specimen No
Specimen
3
- .
NoNO
De Specimen
2 34
No
.
.
Specimen
for
No
Specimen
1 265
per
Specimen
NoNoNoNo
De
oo
CHANGE
VOLUME
%
24
-
fondato
VAQUERODAM
Sample
Na10FX116
--
--
Co Specimen
- No
-D
WILLARD DAM EMIGRANTDAM Specimen

No
.1 .2 3.
.
- .
-
Sample162167 Sample
No X156 Specimen
No
No
18
SHEAR STRENGTH
Specimen21 Specimer
No
Specimen
No bottest toplama
Specimen
NoNo
PSI PRESSURE PORE

Specimen
34
Specimen
No
D
- - ..
.
1 2 65
Specimen
-P ..NoNo
testostoga
PSI
CONFERENCE
WILLARD DAM EMIGRANTDAM VAQUERODAM
.
-
-
.
-
Sample162167 Sample
NoIBPX156 Sample
NQ10FX116
No
2
Specimen
No Specimen
No Specimen
NOI
o
Gastr
-D
nomy Specimen Specimen

No Specimen
No
3+.
23
De Specimen Specimen
No Specimen
No
NoNo
STRESS DEVIATOR
O
..
. 26 5
Specimen
No
8
14 16 20
9
13 11 12
10
12
16
22
10
11
12
10
%
AXIALSTRAIN AXIALSTRAIN
%
AXIAL STRAIN
13
II
.
.--
of
,
Fig Typical Plots Axial Strain versus Volume Change

Pore Pressure and Deviator Stress
COHESIVE SOILS 63
moisture for either the constant placement density tests or the constant densi
ty at start of deviator stressing tests . In these cases , initial densities can be
controlled to about 0 . 5 pound per cubic foot and moistures to about 0. 5 per
cent . These density and moisture conditions are considered to provide results
well within the accuracies of other test measurements and construction place
ment conditions .
When testing undisturbed soils , it is not possible to control these factors
as closely because the natural conditions within a soil sample may vary .
These variations can be minimized by using small shear specimens which can
be obtained from the same horizon of a core sample . For the shear test shown
in Fig . 10, the dry density varied only by 0 . 3 pound per cubic foot and the initi
al moisture content only 0. 3 per cent . Such close specimen uniformity is not
always obtained , however , and undisturbed specimens often vary by as much
as 5 pounds per cubic foot dry density and as much as 5 per cent moisture
content . When variations of this magnitude are encountered they must be taken
into account in the interpretation of the shear strength data .
2. Accuracy of Stresses Measured . - For each of the machines , the load
measuring capsules and gages are capable of measuring stresses which can
be read to nearly one - half per cent accuracy . Calibration charts are developed
from dead load calibrations of the small and medium equipment and from a
calibrated proving ring for the large equipment . For the small 1- inch and
medium 3 - 1 / 4 - inch - diameter specimens , the minimum accuracy of the piston
capsule devices is 0 . 4 and 0 . 6 psi of deviator stress , respectively . For total
axial loads under 500 pounds (60 - psi deviator stress on 3 - 1 / 4 - inch - diameter
sample ) , the pressure capsule cap is used for the medium machine because it
has greater accuracy than the pressure capsule in the load piston for that
machine . All measuring devices are considered to be sufficiently accurate so
that the final shear values will not be affected appreciably . This has been
demonstrated by drawing Mohr ' s stress circles in pairs which include the
range of variations .
The pressure capsule cap measuring device for low loadings on the medium
size machine measures only the load on the specimen top end plate and , thus ,
no piston friction is involved . The regular piston pressure capsule for that
machine , the small machine , and the large testing machine mea sure load
values which include any piston friction developed during the test . The zero
point for the " no specimen load " is obtained on the piston pressure capsules
of the small and medium machines while moving the pistons and specimens up
to the cap bearing surfaces . Thus , the effect of chamber pressures on the
pistons , and the effect of piston friction from the chamber pressure action on
the lubricated O- ring seals , are measured simultaneously . The piston friction
amounts to about 0. 6 - and 0. 9 - psi deviator stress on a 3 - 1/ 4 - inch - diameter
specimen at 100 - and 25 - psi chamber pressures , respectively , and the top
capsule is usually used for lesser chamber pressures . It has been determined
that non - concentric loads do not affect the friction value (no difference be
tween top and piston capsule readings ) . The effect of horizontal forces which
might develop during the shear test has not been accurately determined be
cause it would be extremely difficult to do so . However , in the medium size
machine we have observed that , when tests with approximately 500 pounds
maximum axial load are made , the lower capsule indicates about 5 pounds
more load than the upper capsule . This value , which can be interpreted as an
over - reading of about 0. 6 psi in the deviator stress on a 3 - 1/ 4 - inch -diameter
specimen was interpreted as being of the same order as the piston friction
alone . Thus friction from horizontal forces was not apparent.
The large machine also has 0 - ring seals . As the weight of the piston , 9
pounds , is sufficient to readily let the piston drop through the seals at zero
chamber pressure , it is known that the friction is less than that amount at that
chamber condition . This friction then has an effect of less than 0. 2 - psi devi
ator stress on the 9- inch - diameter specimen .
As the normal Bureau shear test is performed with continuous axial strain ,
the friction effects of a static piston condition do not occur as they might in
so - called " drained " tests where long periods of no - movement are required to
produce the drainage desired .
Chamber pressures can be read to an accuracy of 0 . 1, 0 . 2 , and 0. 4 psi on
the 0 - 30 , 30 - 100 , and 100 - 300 psi capacity Bourdon test gages , respectively .
Specimen volume measurements can be determined by weighing in air and
water to an accuracy of 0 .01 cubic inch . Accurate volume change measure
ments require the use of correction factors for the chamber and axial
pressures used . These correction factors take into account the expansion of
the chamber , connecting lines , and volume reading tubes , under the chamber
pressure and the upward thrust on the chamber cap by the specimen under
axial load . The constant bore volume - change tubes are accurately calibrated .
The accuracy of the volume - change measuring system is 0 .002 cubic inch .
Specimen length changes while in the testing machines can be measured to an
accuracy of 0 .001 inch . Translated in terms of the mean specimen area , the
area can be computed to an accuracy of 0.01 square inch , which is used for
stress computations . Error may occur in computing the unit deviator stress
( Volume at failure ) s
at failure on the basis of the mean area Height at failure of the specimen .
* (
The mean area method presumes that the specimen is cylindrical in shape .
Normally , there is a tendency for some barrel shaping to take place , the
amount depending upon the axial strain , volume change , and plasticity of the
material involved in a particular test . A very hard specimen will maintain a
near - cylindrical shape , a medium - hard specimen will have a slight barrel
shape , and a soft , plastic specimen will have a definite barrel shape . To study
the effect of barrel shaping upon errors introduced by assuming a cylindrical
shape of average diameter , the case of extreme barrel shaping was studied by
computing mathematically the difference in the elliptical failure surface area
of cylindrical specimens of average diameter and barrel shaped specimens
with circular arc sides , assuming the failure plane at a 60° angle with the
horizontal . The results of these calculations are on the top of page 33 .
3. Pore - Pressure Measurements . -- Pore - pressure measuring equipment
has an accuracy of 0. 05 psi for the no - flow pressure cell - an accuracy within
one - half per cent . The test gages have a one - half per cent accuracy and can
be read to 0. 1, 0. 2 , and 0. 4 for 0 - 30 , 30 - 100 , 100 - 300 psi capacities , re
spectively . Readings are made to the nearest 0 . 1 psi . The accuracy of the
pressure transducer and electric recording system is 0. 3 psi. The pore
pressure readings are algebraically subtracted from the applied stresses to
determine the effective stresses and , therefore , any measurement errors di
rectly affect the stress measurements . The no - flow cells only require
0 . 0000042 and 0. 00018 cubic inch of fluid for normal balanced and extreme un
balanced operation , respectively , and the pressure transducer only 0 . 0007
cubic inch at 100 - psi pressure , which are considered negligible . The method
for the measurement of pore pressures has been described previously . Al
though every procedural precaution is taken , there is more chance of error
occurring in this measurement than in any other stress measurement of the
test . For example , excess water on the contact plane between the end plate
COHESIVE SOILS 65
on
Axial Volume Area error 60°
strain - - change * - - % Failure Plane
%
*
*
.
-
90 11 16 00 94 54 00 0076
+
0 0 0 0 0 1 1 0 0 1 3 0 0
. . . . . . . . . .
nou
+ + -
nou
+ + +
brou
00 4563 26 00
+ +
. . . . .
41
0 0 +
torou
+
svell es consolidation
+
+
=
,
(
Area Error Mean cylinder area Barrel shape area 100

#
x
-
Mean cylinder area
and the specimen will produce excessively high pore pressure readings during
-
all ceramic end plates

of
or
or
part the test period Poor contact insert
of
.
devices will produce erratic results Therefore alert and experienced

,
.
personnel are needed watch for possibility error The importance

of
of
to
strain rate control and other pore pressure measuring requirements have
-
been discussed previously The possibility that pore pressure measurements

.
may not always be 100 per cent correct recognized and there still room
is
is
for improvements equipment and techniques

of
However felt that test

is
it
,
.
effective stresses and pore pressure measure

on
results obtained the basis

of
ments are greatly superior the basis of applied

on
test results obtained

to
stresses
.
For many years controversy has existed regarding the use end plate
of
,
pore pressure measurement as compared with small measuring devices in

-
serted into the shear specimen Innumerable measurements have been made
.
using perforated metal end plates which pore pressure measurements

in
agreed closely with computed values Porous stone end plates except fine
.
ceramic disks were never satisfactory Since 1936 we have experimented

,
.
)
with many types measuring devices the early work short and long
of
In
,
.
vertical rigid inserts flexible vertical inserts horizontal rigid inserts small
,
vertical central sand cores and small resistance transducers were tried on
inch diameter specimens plastic lean clay soil All
of
of
these had
,
3
1
/4
a
-
some disadvantages Even though the inserts were fine difficulty was experi
,
.
keeping air out avoiding saturating the area around the

of
or or
enced
in
them
insert thus obtaining overly low overly high readings
on
nonsaturated
,
specimens similar disadvantage was experienced with the sand cores

A
.
Readings were often extremely erratic Some flexible tubes and cores were
.
completely shut off by shear failure movements great many instances

In
a
,
.
the inserts influenced the strength the specimen and the position
of
the
of
failure plane
.
The most successful inserts have been small , flexible plastic tubes , with
fine porous or ceramic stones inserted into the ends , which were used in large
9 -inch - diameter compacted specimens . With specimens of this size , insert
ing a relatively small foreign object has no apparent effect on the strength . In
a recent study , perforated metal top end plates , fine ceramic disks in bottom
end plates , and fine porous stone flexible inserts placed at mid - length and 3
inches into the large specimens , were used . Pore - pressure measurements
were read at the three locations simultaneously by means of the pressure
transducers and the stage recorder . Pore pressures were generally in agree
ment at the three locations , but there were several instances where one did
not agree with the other two . Installation of the insert and ceramic devices
are extremely difficult and installation techniques and experience are still in
the development stage . A slightly tipped insert , an insert against a large void ,
or a poor soil contact against a ceramic plate will cause erroneous results .
We have made tests wherein all three measurements were almost identical .
Fig . II - 14 shows the data obtained in such a test on a compacted SC - CL soil ,
wherein the values by the three methods were so close that only one line could
be drawn through the points on the axial strain vs . pore pressure plot . There
fore , it is possible to obtain similar end and insert measurements when
perfect placement of the insert measuring devices are obtained . Of interest
in the axial strain vs . pore pressure plot of Fig . II - 14 is the negative pore
pressures which were measured through the fine ceramic disk before appli
cation of any load . Later , as consolidation occurred from application of the
chamber pressure , the pore pressures increased to the positive side and were
.
then measured equally by all three devices .
4 . Errors in Procedure
a . General . - Insofar as possible , an attempt has been made to eliminate
procedural errors . Of first importance is the requirement for well - trained ,
conscientious laboratory technicians . In our laboratory , only technicians with
several years of laboratory training and an aptitude for very accurate test
work are allowed to perform shear tests . The general procedures have been
included in the Earth Manual so that uniformity can be achieved within the par
ticular testing program being executed . Secondly , an attempt has been made
to provide accurate correction factors . Frequent equipment calibrations are
made to assure that measurements are accurate . Temperature - controlled
rooms are provided to eliminate temperature effects . Probably , the greatest
deviation occurs in the variation between the specimens of any triaxial test or
within an individual specimen from structure , density , and moisture stand
points . As previously stated , density and moisture can be controlled to
reasonable limits in tests on compacted specimens , but such variations cannot
be precisely controlled for tests on natural soils . There may be considerable
structure variation in natural soils , although care is exercised to secure
specimens which are as uniform as possible . Some structure variation may
occur in well - controlled compacted specimens, although this is not normally
the case .
b . Sleeve Effect . - The restraining effect of rubber sleeves on triaxial
shear specimens produces a possible source of error which should be under
stood . Data obtained during early triaxial development studies as to the effect
of rubber sleeves on specimen strength are no longer available but the con
clusions were that the effect is not extremely large for the procedures used .
Somemore recent data of this type are given on the following page . The first
study involved the testing of three rubber thicknesses in the medium - size
COHESIVE SOILS
TIME, HOURS AXIAL STRAIN, PER CENT
20 4 12
AO
Perforated end plate
,
30 Ceramic end plate
Porous insert
,
PSI
PSI
PRESSURE
PORE PRESSURE
Consolidation period
PORE
Lateral
o
stress Shear stresses applied
Lateral
stress applied
Ceramic end plate

--
Stress LABORATORY RESEARCH

-
SC
CL
50 Sandy Clay Silty Clay
(
)
-
Sample No 27C
2
-
.
Compacted
,
Large Scale Triaxial Shear
PER CENT
22
Specimen size
50 o 9
"x
Psi
"
Per cent Gravel
,
Max stress ratio
.
Per cent Sand

STRESS
Liquid Limit 49
CHANGE
Plasticity Index 28
SPECIMENNo
4
.
DEVIATOR
07 50 psi constant
(
)
:
VOLUME
Volume change
.
12
AXIAL STRAIN PER CENT

,
vs
Fig Plot Axial

14
Deviator Stress and Pore Pressure

of
Strain
II
-
-
.
Measured Three Locations Specimen

at
in
machine with 3 - 1 / 4 - inch -diameter by 9- inch - high specimens. The soil used
was size 20 - 30 Ottawa sand and each specimen was placed close to an average
density of 107 . 4 pounds per cubic foot.
Sleeve 03 Max . (°1 - 93 )

thickness ( in .) (psi ) ( psi ) Tan Ø
0 .015 13 . 1 34 . 9
( 1 / 2 Standard ) 13 . 1 34. 3
100 . 4 259 . 5
100 . 4 259 . 1 0. 68 0.2
0 .03 13 . 0 34 . 7
( Standard ) 13 . 0 34. 3
100 . 5 271 . 2
100 . 6 274 . 7 0 . 72 -0. 2
0 . 06 13 . 1 34 . 9
( 2 x Standard ) 13 . 1 35 . 4
100 . 5 276 . 2
100 . 6 273 . 4 0 . 71 -0. 2
The second set of available data were made using the same soil with large , 9
inch - diameter by 22 - 1/ 2 - inch long , specimens placed close to the average
density of 104 . 9 pounds per cubic foot.
Sleeve Max . (°1 - 03 )

thickness ( in . ) ( psi ) ( psi ) Tan
Tanø C
0 . 04 12 . 5 31 . 8
( 2/ 3 x Standard ) 12 . 5 34. 1
5 0. 191 . 5
75 . 0 189 . 9 .
0 67 0 .4
0 . 08 12 . 5 34 . 1
( 1 - 1/ 3 x Standard ) 12 . 5 32 .5
75 . 0 190 . 5
75 . 0 188. 9 .
0 67 0 5 .
No data are available for the effect of the standard rubbers (0 . 007 to 0 .01 inch
thickness ) used for the small size specimens.
From the above data it can be seen that some strength gain for the high
chamber pressures is apparently given by the 0. 03 - and 0 . 06 - inch thick
sleeves , as compared with the 0 .015 - inch thick sleeves for 3 - 1/ 4 by 9 - inch
specimens , and that very little difference is apparent between the 0 . 04 and
0 . 08 - inch thick sleeves for the 9- by 22 - 1 / 2 - inch specimens . In fact , the
latter produce results very similar to the thin sleeve on the 3 - 1/ 4 - by 9 - inch
specimens and show a very close agreement between the large and medium
specimens . It is interesting to note that the strength of all specimens , large
and small , and all rubber thicknesses , tested at the lower lateral pressures ,
show extremely close agreement . This would indicate that the rubber thick
ness has a greater effect on the specimen strength at high chamber pressures
as the rubber is forced tighter to the specimen . These data show that there is
a small effect of the rubber sleeves increasing the c and 0 values measured
during the shear tests . Henkel and Gilbert ( 1952 ) point out that rubber sleeves
of 0. 0039 , 0. 0079 , and 0.0197 inch thickness cause strength increases for clay
specimens , 1- 1/ 2 - inch diameter and 3 inches high , which are proportional to
the stiffness of the sleeve , in the order of 0 . 3, 0 . 6 , and 1. 4 psi, respectively .
COHESIVE SOILS 69
c.
Size Effect . The effect of specimen size on the strength values ob
tained in the triaxial shear test is another possible source of error that should
be considered . An extensive series of tests were made (Holtz and Gibbs ,
1956 ) to study the effect of specimen size for free -draining sand and sand
gravel mixtures . These tests showed very good agreement of results for the
small , medium , and large size specimens , provided the significant maximum
particle sizes did not exceed certain limits (See also sleeve effect ) . The
for
recommended significant maximum particle size various specimen sizes
is
:
Specimen size
10
Maximum particle size
(
.)
Fine sand
1
3
/8
3
x
-
16 16
inch
3
1
9 6 4
3 3 3
x x x
/
-
/4 / /
23 15
inch
3 3
4
- -
/
inch
Summary summarizing the available information reliability
on
of
In
5
-
.
results can be stated that primary importance
of
it
is
testing
of
the use
,
personnel who have the experience and ability perform tests an accurate
to
in
manner and the insight observe conditions that may lead erroneous re
to
to
reduce measurement errors
of
sults Accurate calibration the equipment

to
is
.
also of prime importance Even with these precautions must be recognized it

,
.
that certain errors exist the final determination of the shear strength
in
possible that errors as much as per cent or even

is
per
It of
10
It
values
5
.
cent might occur for some soils also probable that these testing errors
is
)
larger than the errors which could occur from variations

no
are natural
in
or
or
of
soil conditions the case foundation borrow areas the variations

in
that may occur between the presumed conditions duplicated the tests and the
in
of
actual conditions completed structure

a
PORE PRESSURE COHESIVE SOILS

III
IN
.
Theory
.
A
.
solid particles
of
of
The solid constituents soil soil mass consists

A
1
.
.
-
and pore fluids The solid particles generally are mineral grains various
a of
,
,
.
sizes and shapes occurring every conceivable arrangement well

is
in
It
,
known fact that important engineering properties soils such as permeabili

of
, by
compressibility and shear strength are determined the fine particles

to ty
,
degree out proportion with their percentage the mass These fines
of
in
a
are characterized by large specific surfaces surface area per unit volume
)
(
which for particles smaller than about 001 mm produce what known as
is
,
0
,
Specific surface increases as the shape particle devi

of
colloidal behavior
a
.
crystals
or
sphere hence sheets needles

of
ates from that rod shaped

,
,
;
a
to -
characteristic of many the clay minerals are subject surface phenome

of
On the other hand silt and sand particles which may be the major con
na
,
,
.
cohesive soil are both larger size and more nearly spherical
of
stituents
in
a
shape than the clay minerals They normally exhibit surface phenomena
to
in
smaller degree than do clays

a
Depending on the geological processes which determined its occurrence

a
or ,
soil its natural state the ground may have single grained structure
in
in
by
compound structure the former type each particle supported contact

is
In
,
.
with several of the grains . In the latter type , large voids are enclosed in a
skeleton of arches of individual fine grains ( honeycomb structure ) or of aggre
gations of colloidal - sized particles into chains or rings (flocculent structure ) .
(Casagrande , 1932 ) Compound structure is the result of sedimentation of
particles which are small enough to exhibit appreciable surface activity . Soils
with compound structure are usually of low density , but may have developed
considerable strength due to compression of the arches in the soil skeleton .
When these soils are remolded , their structure is changed and it approaches
the single - grained structure , depending on the thoroughness of remolding .
The solid particles of a soil mass can be considered to be a skeleton
through which forces may be transmitted by grain - to - grain contact . The per
centage of the surface area of a particle which is in contact with other parti
cles of the mass is known to be small . For granular soils , it is less than 1
per cent . Terzaghi (1936 ) reported that even for clays it is small enough to
be neglected in soil mechanics computations .
Consider the concept of stress in a soil skeleton with that in an ideal,
homogeneous , isotropic material . Stress at a point in ideal material is force
per - unit area of a plane surface containing the point . For each of the infinite
number of planes containing the point, the stress (which , in general , is in
clined to the plane ) can be resolved into a component stress at right angles to
the plane and a component stress parallel to the plane . The former component
is called normal stress , and the latter is shear stress . Glover ( 1943 ) has de
fined stress in a soil skeleton in a way such that the stress computation made
for a point , as designated by coordinates , will represent an average stress in
the immediate vicinity of the point rather than stress at a point. He shows
that the equations of static equilibrium relating the normal and shear stresses
at a point on a plane to the principal stresses at that point are valid for soils .
The normal and shear stresses referred to above are known in soil me
chanics as effective stresses : ō and T, respectively . The equilibrium
equations for the stress on the soil skeleton are
Ở = cose + o , sino (III - 1 )

sin
20
III
2
-
(
)
plane
on
where the effective normal stress

is is is is is
T ő
the plane
on
the shear stress

effective major principal stress point
at at
01
the the
on the effective minor principal stress the point
the angle between the plane act and the one on
on
which and
,
ō
major principal plane

71
which acts the

(
, ).
effective stresses the soil skeleton generally under

of
Under the action

goes elastic deformation and its structure altered by particle rearrange
is
,
The relationship of the volume soil mass the effective normal

of
to
ment
a
.
stresses applied the soil skeleton known as the compressibility Simi

is
to
the shear strength of soil mass depends the ability of the soil skele
on
larly
a
to ,
resist shear stresses Hence the mechanical properties soils are

of
ton
on ,
.
controlled entirely by the stresses the soil skeleton

.
The Pore Fluid Two Phase System Completely surrounding the

in
2
.-
-
, -
.
grain grain contact

of
particles the soil skeleton except the points

at
of
to
,
-
-
COHESIVE SOILS 71
is the pore fluid which may consist of air , water vapor , and water . When the
pore fluid consists of air only or water only , the soil mass is a two - phase
system of soil soil grains and gas or water , respectively .
The Perfectly Dry Soil Mass. - Fig . represents soil mass which
III
is
or la
a
-
water such as very fine sand silt which has been oven dried
of
devoid
a
,
,
-
disintegrated and compacted sample this soil
is
of
sealed thin
in
If
,
to a
a
.
rubber membrane and subjected external all around compressive stress
an
on
clear that portion the applied load will act
is
of
the skeleton and
it
,
0
resulting
of
cause undergo deformation decrease volume the soil
in
it
to
in
a
,
mass the sealed sample the air the voids would be compressed Let
in
In
,
e .
.
let Pa equal atmospheric
of
volume air soil mass

ea
volume
of
in
=
+
1
;
e
a
pressure let the subscripts and denote initial and final conditions re
,
0
1
;
spectively Assume that the temperature remains constant during the com
.
pression and that the compressibility the individual soil particles
is
of
VOLUME WEIGHT
eee Air
Solid soil
groins
Dry Soil
0
.
1
woter WGY
-
Solid soil
groins
GY
Fully Saturated Soil

b
.
09
101
Air
Water
-
-
-
Solid soil
groins
GY
Partially Saturated Soil

C
.
Fig Weights and Volume Soil Masses

III
of
1
-
-
.
.
negligible compared to the compressibility of the soil skeleton . Then , by
Boyle ' s law , the final air pressure ( above atmospheric )
Pose
III
(
u
.
3
-
De
-
)
-
This equation Hilf gives the pore air pressure air
of
1956 terms initial
in
,
(
-
)
void ratio and change void ratio Ae for perfectly dry sealed soil mass
in
,
a
(
.
or
The average contact stress effective stress normal every plane this
to
in
,
,
Terzaghi
be
compressed soil mass can found from equation 1936
's
):
Ug III
Ở
=
4
0
)
-
-
for the dry soil mass Hence
,
.
Ae
III
=
Ō
0
Ae
5
(
-
)
-
general
In
De
a.P
cos
sin
III
©
Ae
=
+
e
6
-
-
)
which reduces Equation III for
to
03
=
=
5
0
an (
)
-
For soils subjected all
no
around external stress internal shear
to
,
-
03 by
sin
On
stresses are induced since
III
Equation
20
the other
,
,
=
=
-2
0
T
.
01
hand for the condition internal shear stresses are induced

,
,
7
0
-
.
the dry soil mass the development shear stress may accompanied
be
of
In
by volume decrease volume increase or no volume change clear how

is
It
,
,
.
ever that the pore air pressure sealed specimen will be
of
function the
in
,
a
-
volume change regardless

of
how the shear stress applied This true

is
is
,
,
.
regardless the stress history prior

of
and beyond the condition

at
of
to
,
,
failure that pore air pressure volume change reversi

is
the relation
of
to
is
;
ble
.
The Perfectly Saturated Soil Mass

III
Fig two phase system

1b
but
is
,
a
–
.
.
–
the voids this case are completely filled with water instead
of
air The
in
saturated conditions represented here can be found can be pro

or
nature
in
the laboratory specimen soil particles

of
two phase system

of
duced
in
If
a
a
a .
subjected all around

is
and water sealed by thin rubber membrane and

is
to
external stress of moderate intensity say 100 psi there will be no mea sur
),
(
by
pore
be
able volume change the mass and the entire stress will carried
in
water pressure This follows from the fact that water virtually incom
is
u
,
.
pressible bulk modulus equals 300 000 psi Under these circumstances the
,
,
(
).
only stress that could carried by the soil skeleton would be the magnitude
be
necessary compress the skeleton an amount equal the volume change

to
to
in
the water This volume change exceedingly small for the stresses en
is
.
soil mechanics For applied stress

of
an
of
countered void ratio and

in
.10
.
100 psi the volume change would be approximately 02 per cent Thus for
0
,
,
, .
all practical purposes completely saturated soil the effective compres

in
a
,
regardless the compressibility characteristics

of
sive stress equals

of
,
o,
0
,
the soil skeleton

.
COHESIVE SOILS 73
The foregoing statement is true for all - around compressive stresses only .
When compressive stresses of unequal magnitude are applied to a soil mass ,
shear stresses are induced , with corresponding shear strains which may re
sult in the tendency of the soil to expand . A fully saturated soil cannot expand
without putting tension in the pore water . Thus, many saturated soils when
subjected to shear stresses develop negative pore pressures ; for example ,
heavily overconsolidated clays near the point of failure . Hence , the complete
ly saturated soil can exhibit either negative , 0, or positive pore - water
pressures without measurable change in volume of the soil mass .
3. The Pore Fluid in the Three - Phase System . - It remains to consider the
the general case which includes all
III
three - phase system of Fig . This
1c
is
.
-
compacted cohesive soils and all partly saturated undisturbed soils Here
,
.
the soil skeleton surrounded by pore fluid consisting
of
water water vapor
is
,
and air The liquid phase or soil water will be considered first
,
,
.
.
Soil Water Water unsaturated porous material such as soil held
is
in
,
,
.
-
by
the soil
of
the phenomenon surface tension which occurs
at
the bounda
in
ries separating the water and the pore air known that surface tension
of is
is
It
.
virtually independent
of
of
of
the form the surface contact but function
of is
a
,
temperature The rise of liquids fine bore tubes result surface
is
in
a
.
called capillarity
is
tension and
.
capillary tubes spherical

of
The surface the water curved the

is
is
if
in
it
;
cylindrical The surface
is
the water intersects the wall

of
at
of
tube the tube
.
angle This angle depends the properties the solid and the impuri
an
on
In of
a
,
.
clean glass general the contact

on
on
ties the wall For water

=
,
,
0
a
.
by
angle between gas liquid surface and any solid covered
of
film the
a
a
liquid approaches The height capillary rise

of
,
0
.
21
III
he
7
rya (
)
-
-
where Tg yw the unit weight

of
the surface tension water and

is
is
the
is
,
capillary tube Since the capillary tube atmospheric air

of
is
radius open
to
,
a
the free water surface the pressure

at
on
which also acts the meniscus must

,
be equal minus whe The inverse relationship between the pressure at

to
the meniscus and its curvature holds regardless verti

of
whether the tube

is
,
cal Water within horizontal capillary tube the same radius would be
. of
a
.
bounded by two menisci The pressure

of
the same curvature

at
the water
in
each meniscus and throughout the horizontal tube would be minus ywhe
:
general the pressure

on
any liquid film contact with the atmosphere

In
in
,
by
at
any point can be expressed the equation

:
III
8
(
)
-
To
]
[G
-
*
W
r1 Ts uc
where capillary pressure taking atmospheric pressure

is is
=
0
(
surface tension
by
r2
are radii curvature the two sections formed passing

of
of
and
planes normal the tangent plane through the point and right angles
at
to
to
two
unsaturated soil the water wets the soil grains and
an
each other held

is
In
,
.
the capillary spaces between them Since the pore spaces soil mass
in
in
a
.
are interconnected at equilibrium all the menisci will have the same curvature
,
Lr1
El
r2 ]
, and all the water will have the same capillary pressure , uc ; pro
vided the mass of soil is small enough for gravity to be neglected .
Richards ( 1928 ) performed experiments to measure capillary pressure by
means of field tensiometers . He considered an isolated mass of moist soil
and concluded that water content , size of soil particles , and state of packing
will affect the capillary pressure through their effect on the curvature of the
menisci . If , for a given soil skeleton , the amount of water in the voids is de
creased , there will be an increase in the curvature of the water surface ;
hence , the more negative the capillary potential will be . Also , if equal weights
of a silt and a sand have the same water content , the silt will have more sur
face area , more contact points between particles , less water for each contact
point; hence , lower pressure for the final soil than for the coarse one . The
compactness of a soil also influences capillary pressure . If the soil particles
are pushed closer together , the curvature of the water surface decreases . If
the soil is sufficiently compressed , the water will run out of the moist soil by
gravity . Up to that point , the water content remains the same , but the capil
lary pressure increases to 0 .
The magnitudes of capillary pressure have a very large range , from 0 for
all saturated soils , to negative pressures of more than 100 atmospheres in dry
clays . The ability of water to withstand high tensile stresses has been proved
by experiment . Dorsey ( 1940 ) reports that water containing dissolved air and
fibers of wood was subject to a tension of nearly 160 atmospheres before rup
ture occurred . Puri (1949 ) gives a table showing the relation of the mean di
ameter of the soil particle , the maximum height of capillary rise , the pF
(logarithm of the reciprocal of the height of capillary rise in centimeters ) , the
relative humidity of the capillary water , and the negative pressure in atmos
pheres , all of which are related . Based on the results of direct and indirect
capillary pressure measurements , he concludes that all water in soil is " capil
lary ” water , and the so - called differences are merely those due to size of
pores and magnitude of free energy . There are no gaps or breaks within the
entire range of relations between per cent saturation and free energy or
pressure deficiency , only points of inflection .
In considering the engineering aspects of unsaturated soils , air pressures
different from atmospheric are encountered . Their effect on capillary
pressure in the water must be evaluated in a general pore - pressure theory .
It has already been shown by Equation that the height capillary rise
of
III
in
7
-
vertical cylindrical tube depends only on the surface tension the radius of
a
the tube and the unit weight the magnitude

of
water independent
of
of
It
is
,
atmospheric pressure that the capillary tube free water surface system
is
if
;
were placed container and the air pressure were changed from
ua
in
to
a
the height capillary rise hc and the curvature the meniscus would re
of
of
,
,
at
main the same However the pressure every point the capillary column
in
,
.
would be changed by precisely the change atmospheric pressure The

in
pressure
ug
the water the meniscus would be general

at
in
Ywhc
In
=
u
,
-
Tsl
r2 1
of 71 L
the capillary pressure

ua
uc
uc
+.
where
is
then
=
=
,
;
u
ems
-
'
]
|
Soil Air The presence of both air and water the voids soil mass
in
,
.
-
which sealed and subjected volume changes as the triaxial shear test
is
to
in
,
requires consideration
of
the interaction between the two fluids Air con

in
.
tact with water contains water vapor

on
the other hand the water contains

,
of ;
air
these and the time required

of
dissolved The amount each obtain

to
.
equilibrium when the pressure changed requires examination

is
.
COHESIVE SOILS 75
Thepressure of water vapor in equilibrium with water depends primarily

on the temperature , but it is also affected by the partial pressure of the air
with which the vapor is mixed and by the shape of the liquid surface . Hilf
( 1956 ) points out that the effect of air pressure on the vapor pressure of an
air - vapor pressure mixture is small for pressures encountered in soil me
chanics . Also , it is shown that the effect of curvature of the water surface on
the vapor pressure is very small . Hence , vapor pressure can be considered
constant for a given temperature , regardless of the variation of the soil air
pressure and the variation of the curvature of the capillary menisci .
Air is soluble in water to an extent that must be considered in soil . Ac
cording to Henry ' s law , the volume of air dissolved in a given volume of water
is equal to the coefficient of solubility of air in water by volume , h , which
varies with the temperature but is independent of pressure . Values of h are
as follows:
Degrees C h_
0 . 0288
0 . 0260
0 0 . 0235
0 . 0216
20 0 . 0201
0 . 0188 .
Hilf ( 1956 ) concludes that in a compacted cohesive soil where the water can
be considered to be in the form of thin films , the time involved in reaching
equilibrium for solution of air in the water film is not of great importance,
even under laboratory conditions , when pressures are applied relatively
rapidly .
The forces caused by surface tension must be taken into account in the the
ory of pore pressure in unsaturated soils . The soil water is considered to
exist in the form of a continuous film covering the grains , and menisci near
the contacts between grains . Neglecting gravity , the menisci have a particu
lar curvature which corresponds to a particular capillary pressure for a given
temperature , water content , and state of packing for a given soil . The re
mainder of the void space surrounding the wetted soil skeleton is occupied by
air . In soils of low degree of saturation , the air voids are interconnected , but
Gilbert (1959 ) reports that at about optimum water content the air voids in a
compacted cohesive soil are no longer interconnected . For equilibrium con
di tions , however , it does not matter whether the air voids are interconnected ;
the air pressure would be the same throughout the soil mass , because of
Henry ' s law . Under these circumstances , a change of pressure in the air ,
from whatever cause , will have no effect on the curvature of the menisci , but
will result in a corresponding change in pore - water pressure .
Pore - Air Pressure . - The pressure in the air of the voids of a soil mass ,
which has been compressed without permitting escape of the pore fluid , can
be calculated by combining Boyle' s law of compressibility of air with Henry ' s
law of solubility of air in water . It has been established that the curvature of
the menisci has a negligible effect on the interaction of air and water . Also ,
the time involved in obtaining equilibrium for Henry ' s law , as the pressure is
increased , has been shown to be very small . Hence , the water can be assumed
to be saturated with air for the pressure considered . At temperature ranges
involved in soil mechanics , the magnitude of the vapor pressure of water is
less than 4 per cent of 1 atmosphere and is virtually independent of pressures .
Hence , it can be neglected , excepting for laboratory research conditions where
less than atmospheric air pressures may be used .
Consider a compacted soil mass of volume 1 + eo , as in Fig .
1c
III
with
,
-
air volume at atmospheric pressure Pa then compressed at constant
ea
is
It
,
.
temperature without drainage void ratio ej For pressures used
in to
in
T
a
,
- .
soil mechanics the entire change void ratio ey can be considered as
eo
,
,
occurring the air since the bulk modulus of water about 300 000 psi and
is
in
,
of ,
the bulk modulus the mineral grains 000 000 psi The soil skele
is
about
5
,
.
considered be reduced volume without change its solid volume
to
is
ton
in
in
,
rearrangement particles the grains
by
elastic bending
of
ew of
of
by means
or
.
The soil water volume for conditions contains more
of eo
of
and es but
is
it
,
air solution the latter condition because Henry law By considering
in
in
.
'
the weight of air free and dissolved prior and subsequent compression
to
to
,
(
)
the pressure the air after compression equals
in
:
Pose De
P
.
ua
III
hey
e
-
de
her
9
-
(
)
+
-
The quantity total volume air free and dissolved soil
ea
of
in
is
hew the
+
a
,
)
(
prior compression When ear the air pressure becomes
to
,
e
.
III
10
-
(
)
which gives the pressure required dissolve the air completely accordance
to
in
whichhenry
9law
'
can
also
s
.
as
also be expressed
III
Equation follows Hamilton 1939 and Hilf

be
can
,
-
(
e
1948
:)
ΔΥ
Ρ
8.
III 11
ov
AV
)
Va
-
+
volume change per cent

of
where AV total volume

in
=
Va
air volume per cent total volume

of
in
= =
water volume per cent total volume

of
in
V
.
can
Similarly
be
Equation
10
written
III
,
aa
12
III
un
-
(
)
hv
-
The pressure the pore water

of
an
Pore Water Pressure unsaturated

in
.
-
-
soil which the pressure contact with the soil skeleton

is
the fluid
is
in
in
,
,
sir11L
by
given the formula

:
uz 27 de
'a8
III
TS
he
u
13
se
.
-
-
(
or more simply
,
III
14
ua
uc
=
-
(
)
u
COHESIVE SOILS
B. Relation of Pore -Water Pressure to Shear Strength
1.Positive Pore - Water Pressure . - In Part A , the value of pore -water
pressure for the general case of a 3- phase system of soil grains , gas (air and
water vapor ) , and water was shown to consist of 2 independent components :
the pore - air pressure , Ua (relative pressure ) , which is zero for atmospheric
pressure in the laboratory or at the job site ; and capillary pressure , uc , which
is the pressure due to the curvature of the menisci . The value of uc can never
be positive ; that is , if curvature of the menisci exists , uc is always smaller
in magnitude, on the absolute pressure scale , than atmospheric pressure .
When the soil is saturated , meniscus curvature is zero , and uc = 0.
In order for the pore -water pressure in an unsaturated soil to have a posi
tive value on a Bourdon pressure gage , the pore - air pressure , uz, must be
positive and must exceed the value of uc , since , from Equation
III
14
,
-
uc This frequently occurs during the sealed portion
of
ua
triaxial shear
=
on +
u
tests un saturated soils and believed occur

is
earth embankments con
to
in
of
structed cohesive soils

.
should be recognized that the effect positive pore water pressures

of
on
It
-
the shear strength soil the same whether the water completely fills the
of
is
voids exists thin films surrounding the soil grains with the remainder
or
in
,
of
the voids filled with gas
on
The effective normal stresses the soil skeleton
.
are equal magnitude the difference between applied stress
or
total
to
in
,
stress and pore pressure the fluid surrounding the grains which always
is
in
,
,
the pore water dependent the magnitude
of
on
The shear resistance the soil
is
.
on
these effective normal stresses which act Derived

of
the soil skeleton
.
from the equations of static equilibrium Equation III shows that at any point
,
1
-
any plane
on
the soil mass the effective normal stress function the

in
= is
of
0 a
principal effective normal stresses at that point Since Equation III

U
0
(
-
.
positive values will result

of
effective normal stress being smaller

in
4
),
7
,
than the total normal stress

o
.
The companion equation of equilibrium for the soil skeleton

is
static
:
sensin
III
26
2
(
)
-
This can also be written

:
III
15
20
sin
3
2
-
(
by considering Equation III

This shows that the shear stress on the soil
4
.
-
the magnitude
pore pressure
of
independent follows that the

is
of
skeleton
It
.
development positive pore water pressure soil mass for given total
, of
in
a
a
,
-
will reduce the ability resist shear while having

of
stress system the mass

to
no effect on shear stress On the Mohr plot positive pore water pressure has
-
.
translating the applied stress circles toward the origin

of
the effect make

to
them effective stress circles

.
Negative Pore Water Pressure The pore water pressure unsaturat

in
2
–
.
, .
-
can be negative whenever the positive

Ug
soils uc Equation III

14
ed
+
ug u
,
)
-
(
of smaller than the negative value

of
un
value uc soils stressed ex

is
In
:
ternally will be the pore water pressure always

ua
at
where equilibrium
is
, ,
,
0
also possible for

ua
be negative as for example

uc
negative
is
, to
in
=
It
,
,
,
u
sealed specimen of unsaturated soil which during shear test expands

to
,
a
a
volume greater than its volume when it was first sealed . In this case , ua and
uc are both negative .
The effect of negative pore - water pressure on the shear strength of soils
stems from the equations of equilibrium , Normal effective
III
III
and
2
.
-
-
stress increased and shear stress
is
On the Mohr plot negative
is
unaffected
.
pore water pressure has the effect of translating applied stress circles away
-
the origin make them effective stress circles For stress circles
to
from
.
near the origin this will always result smaller vertical intercept the
of
in
,
a
strength envelope on an effective stress basis than on total stress basis
.
unsaturated cohesive soil unstressed externally and the pressure
an
is
If
atmospheric then from Equations
14
the air voids for
III
III
and
is
Ua in
=
4
0
o
,
,
-
-
=
0
;
III
16
=
u
ở
-
(
)
Since capillary pressures are isotropic the mechanical effect
of
these
,
pressures on the soil skeleton identical the effect of
is grain grain
to
to
a
-
compressive stress equal magnitude acting all directions Hence un
an
of
of in
,
.
saturated soil mass unless completely devoid moisture can never be con
,
the capillary stress
of
on
be unstressed and the mechanical effect
to
sidered
,
compressibility and shear strength must be taken into account

on
.
This concept suggests that the intercept of Mohr strength envelope
a
c
,
of
drawn on the basis total stresses not the no load shear strength
of
an
is
,
-
Also points up the necessity providing means of accu
of
unsaturated soil
it
,
.
rately measuring negative pore water pressures during triaxial shear tests
-
obtain the correct effective stress envelope For example pore

to
if
,
a
.
pressure measuring device not capable of measuring negative pressures

is
below given value say atmosphere the lower portion the Mohr envelope of
,
,
1
a
will be incorrect
of
the extent the difference between the true and the

to
Many pore pressure devices being used are not capable
of
mea sured values
.
mea suring negative values value zero for initial pore water pressure
of
A
.
-
of
unsaturated soils frequently obtained The Mohr envelopes for these

is
soils may therefore be incorrect the vicinity the origin The magnitude
of
in
,
the slope the Mohr envelope tan best determined when negative
of
of
is
,
o
,
pore water pressures are accurately measured

.
-
Pore Water Pressure

of
Measurement
C
.
Laboratory Measurements The laboratory equipment for measuring

1
.
.
--
pore water pressures described Part which also contains data Figs
is
in
II
.
-
showing pore pressure measurements obtained the laboratory

10
14
to
in
II
-
This section will describe attempts measure capillary pressures

un
to
in
saturated soils
.
capillary pressures
of
of
The need for renewed consideration the influence

by
of
soil mechanics was pointed out Hilf 1956 who also suggested the use
in
fine porous stones attached the measuring device contact the unsaturated
to
to
soil Reclamation continuing research this field with the

of
The Bureau
is
in
,
.
objective developing equipment and procedures for the measurement

of
of
negative pore water pressures soils situ after compaction and during
in
of
,
-
compressibility strength Subparagraph describes the

' of
tests and shear

a
.
experience negative pore water

of
Bureau the direct measurement

in
s
pressures Subparagraph discusses indirect methods used extend the

to
b
.
range negative pore pressure reading beyond atmosphere negative Sub

of
paragraph capillary pressure with void

of
on
contains data the variation

c
ratio and with water content

.
COHESIVE SOILS 79
a. Direct Measurement . - Negative pore -water pressures are measured

with equipment consisting of contact devices connected by small copper tubing
to no - flow pressure measuring instruments . The contact devices consist of
very fine porous tips, or disks placed in the end plates . The no - flow pressure
measuring instruments are the same as those used for measuring positive
pore pressures described in Part II . If the pressure in the water contacted is
negative , water from the tip or disk , which is at atmospheric pressure , will
attempt to flow into the soil . If it is restrained from doing so by the no - flow
device , tension will develop in the de - aired water which , when equilibrium is
reached , will be a measure of the tension in the soil water . Unless measures
are taken to prevent it , air from the soil will enter the negatively stressed
pressure measuring system through the tip or disk , and very quickly will
break the column of water .
To preclude this , the tip or disk should have a sufficiently small pore
diameter so that no air from the soil can penetrate the system . The bubbling
pressure of a saturated porous material , such as is used in a tip or disk , is a
function of the largest pore diameter , the surface tension of the liquid - gas
contact , and the pressure of the surrounding fluid . In order for air to pass
through the saturated tip or disk ( see Fig . III - 2) , it must form into a bubble of
diameter d . Then
17
III
(
-
)
where ug the pressure the air bubble
is is
in in
the pressure the water saturating the porous tip

or
disk
u
Interior
of
porous tip
Air bubble
about to
form
,
Air
at
pressure
VA
y
UO
,
4 ;
:
Water
at
pressure
u
Fig
III
Bubbling Pressure Test

2
-
-
.
.
To is the surface tension of water
dº is the largest pore diameter of the porous material .
For u = 0 ( gage pressure ) , the bubbling pressure is related to the pore diame
ter , as shown in Fig . This chart similar those used by manufactur
III
is
to
3
.
-
ers of bacteriological filters determine the pore diameter the filter
of
of to
P .
Fig III photograph filter cylinder The filters are graded
is
4
1
a
,
-
-
.
.
and The filter cylinders are composed ceramic material
of
,
,
2
5
P
.
-
55 -
approximately average diameter The porous
10
an
mm long with
of
mm
.
filters are hollow with one closed end The open end has an impervious
,
.
glazed section about Before use these cylinders are satu
20
mm length
in
,
.
rated by boiling atmos
15
distilled water for least minutes either
at
at
in
,
pheric or reduced pressure For the bubbling pressure test they are connect
,
air pressure which .
applied while the filter cylinder
of
ed
source
is
is
to
a
atmospheric pressure The applied pressure in

at
submerged water
in
is
.
creased until visible bubbling occurs and the water temperature and baro
,
metric pressure as well as the applied pressure are recorded
,
.
The permeability of the cylinders was measured by using filtered de aired
-
water under low heads maximum feet of water Table III shows the
of
1
7
).
-
(
bubbling pressure and permeability tests typical filters used
on
results
of
in
the laboratory Disks for use end plates are also manufactured the same
in
in
.
porosity grades as the filter cylinders inch

III
Fig shows
5
3
1
4
a
.
-
/
diameter bottom end plate fitted with grade filter disk rings are used
O
3
a
P
of .
-
-
seal the filter disks the end plates The arrangement the capillary
to
in
pressure test apparatus for simultaneous measurement insert and end plate
by
shown Fig III
is
in
6
.
.
-
TABLE CHARACTERISTICS OF FILTER CYLINDERS

III
1
-
--
.
Ceramic filter Bubbling pressure Pore diameter Permeability

_
cylinders psi microns

yr
ft
/
24
1
8
P
-
l
.
2
P
4
.4
-
9
5
.
. 23
-3
P
.3
1
8
.
39
1
1
.
0
5
P
-
Prior placing the tip or disk contact with the soil kept submerged
is
to
in
it
,
water Its ability measure negative pore water pressures checked by

to
is
in
taking out the water and permitting the water évaporate from the sur
of
to
If it
face the apparatus working properly minute or two the pore

is
in
a
,
.
pressure gage will read negative pore water pressures appreciable magni
of
-
tude the tip or disk immersed water and the reading goes back
is
Then
in
,
,
.
zero should be obvious that evaporation cannot take place from the sur
to
It
of .
or
tip disk while contact with the soil However as

is
face check
in
it
,
a
a
.
experiments were made with the contacting tip

or
disk inserted sealed

in
a
air which was saturated with water by having small amount of

of
container
a
free water at the bottom these tests the pore pressure

of
the container
In
,
.
remained at zero
.
)
(C
IN
.
BUBBLING PRESSURE PSI AT 20° ug
Fig
.
III
-
3
.
-
2
Relation
to
PORE DIAMETER
of
, 4
d
,
IN CORRECTION FACTOR
5 0 PRESSURE
OF °C . 0
.
970
COHESIVE
9604
6 32 0 H
MICRONS
( 5
0 7
Bubbling Pressure
. 41
SOILS
001
of
a m 8 50 10
.m
its Maximum Pore Diameter

.)
9 59 15
TEMPERATURE
Saturated
10 °C 68 20
Ts
-
AND
77 25
OF
tension
86 30
Surface
Porous Stone
95 35
OBECLIO
PX 20439
D
-
- -
1835
6
E
-
Fig Filter Cylinders Capillary Pressure Measurements

III
used
in
4
-
-
.
.
Fig inch Diameter Bottom grade

III
End Plate with

5
3
1
4
3
P
.
/
.
-
-
-
Filter Disk
COHESIVE SOILS
Sovac
ONGES
PX - D - 20435
F21835 - 3
Fig . III - 6 . - Apparatus for Simultaneous Measurement of Capillary

Pressures by Insert and End Plate
A 9000
MD- 20437
Fig . Apparatus for Measuring Capillary Pressure by Reductions

III
in
7
.
-
-
Bubbling Pressures
The possibility of the porous surface of the tip or disk acting as a semi
permeable membrane and inducing osmotic pressures was investigated . For
this purpose , a P - 3 cylinder was placed in water containing varying per
centages of sugar solution and salt solution of known osmotic pressures . No
negative pore pressures were observed in tests lasting several days . Chemi
cal analyses showed that the sugar solutions and the salt solutions diffused
into the filter cylinder .
Hence , it is concluded that the negative pressure measured when the tip or
di sk is in contact with unsaturated soil is the pressure of the water in the soil .
An indication of the reproducibility of the results of direct measurements of
capillary pressures by use of the filter cylinders , including one with porous
end plates , is given in Table for variety of soils whose index properties
III
2
a
-
are given
.
The possibility negative pressures developing saturated clays by the

of
in
phenomenon osmosis or other causes was also investigated filter
of
in A
3
P
.
-
tip was placed core sample saturated silty clay inches diameter
of
in
2
a
and inches long which was placed container Evaporation was pre
in
3
/1
2
-
.
vented and temperature maintained fairly constant Readings were taken peri
.
odically for days during which time appreciable negative pore pressures
no
,
9
developed Slight temperature changes caused fluctuations but the maximum

.
tension measured was psi

1
.7
capillary
of
Indirect Mea surements The direct measurement
b
.
.
-
pressure unsaturated soils satisfactory for pressures not smaller than

is
in
atmosphere Very dry soils or relatively dry clayey soils where the capil
1
lary tension exceeds atmosphere require special procedures the
In
1
.
Bureau laboratories near Denver Colorado elevation 5400 feet where the
,
,
's
atmospheric pressure psi the maximum direct readings are

12
about
is
,
.
psi that point the column

is At
of
11
about water which tension breaks

is
, in
,
,
.
and the space filled with water vapor As consequence two indirect
a
.
methods of measuring pore water pressure smaller than atmosphere nega

1
-
tive have been attempted the Bureau laboratories The method suggest
in
1
:
(
)
Hilf 1956 for transferring the origin which has been used successfully
by
ed
;
(
obtaining the pressure indirectly by bubbling pressure

of
and method
2
a
a
(
)
test using the soil water reduce the bubbling pressure given porous tip
of
to
a
or
disk
.
Origin The capillary pressure the pore water

of
Transfer
is
1
(
)
.
-
pressure contact with the water films

the air atmospheric How
is
when
in
the air higher pressure the capillary pressure will be the

is
at
if
ever
,
a
,
difference between the applied air pressure and the measured water pressure
.
Transferring the origin accomplished by increasing the atmospheric air

is
pressure which the soil subjected The apparatus similar Fig III
is
is
to
to
6
-
.
excepting that only the disk the end plate used and the specimen not
is
is
in
sealed but placed pressure chamber Evaporation prevented by keeping

is
in
a
the pressure chamber The test data given

of
free water
in
the bottom
in
Table III show that negative pore water pressures smaller than atmos
1
3
-
-
phere can be measured accurately by varying the atmospheric air pressure

”
"
unsaturated soil
an
in
Bubbling Pressure Method Attempts have been made measure

to
2
(
)
.
--
negative pore water pressures great negative value by finding the difference
of
-
the bubbling pressure of porous tip or disk when contact with the soil
in
in
,
in a
and its bubbling pressure atmospheric pressure The principle

of
at
water
, .
bubbling pressure test using stone of

on
the method based the fact that

is
a
TABLE CAPILLARY PRESSURE TESTS ON REMOLDED SOILS
in
-
Classifi Pore water pressure psi 3
yd
2
LL
-III :
Lab Sili and clay cation Spec Trial Trial Trial
PIle
NO • pcf Q
No
sert Bottom Insert Bottom Insert Bottom
_
Sample Gravel Sard symbol
.. .
.
10
: :0.
::* :
0
:. :
1
.4
: :: :0
1
3
-.2 :: :
;- :: :
::. .: ::
ML
-
.. .. ..
.
3
16
5
5
0
23J
.. .. • . .. ..
• .
107107
..
_: - -
.
..
..
.. .. • 55 .. .
.. .. • . 5 ..
.
3.
15
5-
3
.. .. • .. .. ..
•
:
76
54
28
2
..
.0
240
:
:0
3
.
.
:0.
..
CH
-1
:: : .4 :
.5- .9-
: :5 3
22L .. ..
.. .. • .. ..
..
: :5 -:8 :6 :)8
_:: : 5-: 5-: 8-: 9-(:
.. ..
....
.. 104
.. .. • .. . . ..
• .. ..
21
..
.: 0. 0.:
6
.. ..
90
)) 4
..
21
COHESIVE
.. .. • .. .. ..
•
10
.. .. ..
I: : : 9.4-: : : :
7
)3.
::
: 5: - - -(: -:(
.. ..
? .. .. ..
.. .. •
.. .. • .. . ..
?
SOILS
.. .. • .. . ..
18
23
7
SC
• ..
:0
.4
:0
.. ..
:
0
.
-
13
24
53
SM
. .-:
94 91 122
.. ..
26PX
.. .. • .. .. ..
:
1
14
34
25
SC
57
0 0
:0
.7
:0
.. .. ..
. .7
..
.
•
::
-
:0 0:
SM
26PX .. ..
.. .. • .. .. ..
.. • . .. ..
.. .. . .. ..
.. .. .. ..
:
26PX
• . ..
.. ..
15 16
33
63 49
24 13
20 30
118 117 118
12
SM SM
SC SC
00
79
6
- - --
:0 :0.
:0. :0.
..
:0. :0.
:0. :0.
. . 4. 0. 1. 3. 7.
:
.:
2PX
:2 :1 :i ..
cf
in
.
The gravel fraction the soil was excluded the specimen tested
to
.
-- i
in
.3 : * :'
ote Values perentheses are too close atmosphere for confidence
85
known bubbling pressure , say 35 psi , will result in bubbling at 45 psi when the
tip or disk is placed in water at +10 psi pressure . Similarly , if a pressure of
- 10 psi is subjected to the reservoir in which the tip or disk is inserted ,
bubbling will occur at precisely 25 psi . A few tests made with this method
indicate that it is correct in principle , but there are difficulties which have
yet to be surmounted . The main difficulty is that once the tip or disk is in
contact with the soil , the determination of when bubbling occurs must be done
indirectly , since the bubbles cannot be seen . Fig .
of
photograph
III
the
is
7
in a
-
pressure
tip
apparatus measure bubbling soil The air
of
inserted
to
a
a
.
pressure was applied through inch outside diameter Saran tubing
on
a
1
8
-
/
-
-
the tubing
of
of
column colored water The location the air water interface
in
.
-
was plotted the graph paper shown the figure as pressure was increased
on
in
.
air into the soil was indicated by significant increase
of
breakthrough
in
a
A
Capillary pressure of psi was
17
rate of movement the interface
of
-
.
measured by this method another test the method gave result psi
of
In
5
a
-
.
which was checked by direct reading

a
.
Capillary Pressure with Water Content and Void Ratio
of
Variation
c
.
III .
Fig shows the apparatus for measuring capillary pressures unsatu
in
8
.
-
rated soils by floating ring consolidometer the lower piston
of
is
which
a
of
equipped with filter disk this test specimen soil inches
In
4
4
1
a
5
a
P
/
-
.
-
diameter and inches thick very slowly avoid drainage

is
of
loaded
to
in
1
1
4
-
/
by
pore water while permitting consolidation
of
extrusion air
.
of
Three specimens remolded clayey sand two per cent less than
at
,
0
.5
a
per cent dry

of
water content and one were tested
at
optimum optimum
in
2
.7
- PX
20436
D
-
.
1905
7
E
Fig Floating Ring Consolidometer Capillary Pressures

III
which
in
8
–
-
.
are Mea sured

COHESIVE SOILS 87
this manner to obtain data on the relation between capillary pressure and void
ratio during compression . The results are plotted in Fig . III - 9 which also
contains data on the soil properties and placement conditions. Note that the
water contents decreased slightly during the test but the measured losses of
0 . 22 per cent, 0. 45 per cent, and 0. 39 per cent are considered negligible .
The upper plots show the relation of volume change with axial stress . The
solid lines show the way this test is plotted in the conventional consolidometer
where pore pressures are assumed to be zero at equilibrium under each in
crement of axial stress . When the effect of measured negative pore pressure
is included , the dashed curves result . The latter curves are to our knowledge
the first stress - strain curves for a soil skeleton of a compacted cohesive soil
that take into account the effects of capillary pressure .
The lower curves show the relation between capillary pressure and degree
of saturation at virtually constant water content . The curves reveal that the
capillary pressure gradually increases to zero as saturation by compression
is approached . The zero value at less than 100 per cent saturation is proba -
bly due to the fact that at high degrees of saturation , the pore - air pressure is
greater than the zero value assumed in this test .
The data presented in Fig . III - 9 are the first of their kind obtained and ad
mittedly only scratch the surface in exploring the relationships . Additional
research of this type is needed on a variety of soils and placement conditions
to evaluate fully the stress - strain relationships of compacted cohesive soils .
show the relation
III
The data given in Table are plotted Fig

10
III
at to
of in
3
.
-
ship between capillary pressure and degree saturation constant void

ratio for an undisturbed sample of very dry sandy clay This was done by
.
TABLE CAPILLARY PRESSURE TESTS BY TRANSLATION

III
3
.
-
OF ORIGIN
244
074
29F
ma
58
42
Lab sample percent percent sand

<
,
0
-
.
PI
ML
percent
17
CL
percent
LL
721 undisturbed
=
=
G
6
2
-
-
.
Applied air
ug
Measured pore Degree

of
u
2
,
in :
pressure
ua
Specimen pressure saturation

psi
u
,
,
,
,
in psi
10
psi ft3
No
%
in
S
2
/
.
12
-3 -1 12 22
32
27 27
111
5
2
.4
.6 .6
.0 0. 4. 4.
.
- -
.
28 31
.0 0
- -
.
114 114
41
11
3
1
9
-
.
.
13 13 13 14
34
.9 .4 .8 1
6
0
-
.
.
16
.6 .2
-
.36 26
-
1
114 224 114
46
. .
. 80.
-1 -2
. .4
20
30 57
8
0
17
1
5
3
4
-
.
.
%
%
%
%
%
=
=
=
=
=
.
,
,G
,
.
,
.0
1
9
:2- 48
52
28
18
SC
LAB SAMPLE 276 074 SANDLL 49 270 PROCTOR mox 105 pof SOIL
mm
,,
.," 5
5,
EFFECTIVE STRESS EFFECTIVE STRESS EFFECTIVE STRESS
O
_
_
25
1
_
25 50 100
.P,. _
_
75 125 150
50
50
25
,-
72
--
uc
Including
Neglectinguc
% %
!u in
teo
teo
de
NOTE
.
Floating ring consolidometer used Lower
-P 5
piston fitted with counterbored
of
ceramic disk PSI bubblingpressure
a 30
+
HO
SHEAR STRENGTH
10
,
psi
psi
CONFERENCE
-
5
4
in in
uc
uc
pressureCapillary
100 105
70
85
90
95
80
100 80
80
100
90
70
75
70
90
%
%
SATURATION SATURATION SATURATION
22
23
25
SPECIMEN SPECIMEN SPECIMEN

Ya ya Yo
.95
80
INITIAL 106 pcf INITIAL 107 pcf INITIAL 105 pcf

%%
..
..
..
%%
..
%%
FINAL 117 pcf 41

1817
0257
FINAL pcf FINAL 11 82 pcf
9474
1817
w1189
4127
1515
110
.. 9
–
III
of
.
.-
Fig Results Confined Compression Tests with Pore Pressure Measurements
COHESIVE SOILS
9
%
%
001
PSI
80
=
W M
,
IN
ug
%
PRESSURE
60
'
1 =
% W
£
9
=
M
CAPILLARY
.42
%
=
w
%
-M
61
7080
90
20 30 40 50 60
IN
DEGREE OF SATURATION PERCENT
at
Fig Capillary Pressure
III
10
Relation
of
Water Content
to
-
-
.
Constant Void Ratio
The resulting curve indicates

an
adding water the specimen increase
in
to
do
capillary pressure as degree
of
saturation increases although the data not

,
plot smooth curve

in
a
years the Bureau

in 20
Field Measurements Reclamation has

of
For
,
2
-
.
used closed system pizeometer which water filled tubes lead from cells
a
-
-
located the embankment accessible Bourdon gages These instruments

in
to
were described by Walker and Daehn 1948 and details concerning them may
(
be found Designation 24 Earth Manual 1958

in
)
.
-
:
Technical Memorandum No 650 Gould 1959 the pore pressures ob

In
),
.
Reclamation earth dams during their construction

of
26
served Bureau
in
in
period analy
of
from 1936 1951 were summarized evaluate the methods

of to
the
,
-
sis developed by the Bureau Brief histories the construction each dam
of
.
are presented this memorandum illustrate those factors which influence

to
in
the buildup pore pressures

of
embankment test apparatus

by
Gould concluded that the evidence provided

generally substantiates the Bureau theoretical analysis of construction pore
s
'
pressures reported TM 650 that number embankments the

is
of
in
in
It
,
a
.
observations agreed closely with the theoretical pressures Discrepancies

.
between them could be reasonably explained the basis of material properties

on
study histories Drainage relief and surface tension

of
construction
in
and
a
pore water caused some observed values fall below the theoretical
to
the
.
Pore pressures exceeding those predicted for the average placement condition
the fill were developed wet and compressible soil
of
concentrations
of
in
Recent installations piezometers earth dams have included piezome

in
of
ter tips with fairly high bubbling pressures small pore diameters The
).
(
Truetotalstro
-
PLACEMENT VOLVOLVOLA AT Standard
CONDITION DEVOPT wat
deviation Average
1
plocement
of
MATERIAL SOLIDS
WATER PSI
AIR 100 overoge
placement
.69:
.6
-
%
%
Ws10
0
16
1
1
4%
%
%
%
STANDARD
DEVIATION
WET Placement
at
10
.8
7
%
/- (
OPTIMUM 82 optimum w29
10
1616
. 0
w
PREDICTED
32 42
9
06 ..
%
...
"4 ww =
)%)%
10
200 AVERAGE
81 82
61 2 3
%9 2 3
+ 8
6 6
+ 01
044
50 4 8
91 4 9
POREPRESSURES
....
% %
. .-
%%
W W
STANDARD
DEVIATION
DRY 82
1513
.... ,
.... I
INITIAL PLACE CHARACTERISTICSZONE MATERIAL
-
P57
190
( IN w
/- . .-.- "-
of
Watercontent Material
V4
P56 P58
WATER
-
P64 P60
OFFEET
,
A
SHEAR STRENGTH
-
4
w
.9:%
P61
PRESSUREPORE
CONFERENCE
..
9
0 %
170 180 190

#
TATE
UNITED
INTERIOR
-
Standard
deviation
dry OFTHE
PARTNENT
OUREAURECLAMATION
DE
overoge
plocement
OF
8
6
*
%
.
ofW
GRANBY DAM
ANDOBSERVED
PREDICTED
I
60 100 130 140 POREPRESSURES

150CONSTRUCTION
90
80
70
50
40
30
20
O
10
,,
--
120
110
TOTALVERTICALSTRESS PSI WR
DRAWN
,
,
11
-. Dam Colorado
-III
.
Fig Construction Pore Pressure Granby
COHESIVE SOILS 91
purposes of these installations are twofold : initial ( 1) To assure obtaining

readings of negative pore -water pressures in the fill , ( 2) it is believed that
when properly installed , it may be possible to exclude air permanently from
the closed piezometer system using the fine porous disks .
In some instances , the Bureau of Reclamation has also used Casagrande
type porous - tube piezometers in foundations of dams and other hydraulic
structures . The details of the type of installation used , are given in Desig
nation : E - 25 , Earth Manual ( 1958 ) .
On several occasions where the construction of the terminal facilities for
the Bourdon gages was impracticable and a no - flow measuring system was
needed , installations were made using the hydrostatic pressure indicator ,
based on the principle of the no - flow pressure cell used in the triaxial shear
test . A portable test set was used to measure pore pressures developed in
the fills .
Fig .
11 shows the predicted and observed construction pore pressures
III
-
for Piezometer Tips Granby Dam Colorado
of
58
60
, 56
57 59 61 and 64
,
,
,
,
Big Thompson Project Colorado Fig III these tips
of
12
shows the location
.
-
.
Drainage relief evident during the entire

of
cross section
is
the dam
in
a
at
the stripping line while
of
season the trend the curve for Piezometer 61

in
,
at
the later stages
of
Piezometer 60 became effective construction The
in
it
.
and 64 probably results
56
erratic fluctuations the curve for Piezometers
of in
from the presence the lines followed by pumping attempt fill
an
air
to
in
in
them
.
Fill completed 1949

OF
LOCATION PIEZOMETERS
-
-
-
14
10
STATION
+
250
$
Rockfilla
1948
-
200
#
ELEVATION
30
8100 Piezometers
8050
1947
oom
r
SOO
200
100
200
soo 200 100 10

-
--
-
-
Fig
Piezometer Tips Granby

of
12
III
at
Location Dam Colorado

,
-
-
.
According the practice used plotting these curves the total vertical
to
in
taken as the weight overburden immediately above the piezome

of
stress
is
beneath the upstream slope pore

58
56
57
ter At Piezometers and

,
,
,
.
pressures rose after the completion the embankment directly over the tips
of
the observed curves this pore pressure rise appears occur without any
to
In
The ultimate vertical stress

at
increase the total vertical stress these

in
tips more likely equal the intersection of the pore

of
the abscissa
to
is
pressures with the extension the curve observed before construction was
of
completed According this approximate rule the true total vertical stress
to
,
.
165 psi per cent greater than the

12
be
Piezometer 57
at
estimated
is
to
weight of overburden
.
In general , the position of the observed curves in the theoretical family
substantiates the analyses given in Part According
III
the control tests
to
A
,
.
the material placed elevation 8060 and directly above averaged per cent
at
9
.6
The pore pressures observed the early stages
of
moisture construction
in
.
piezometers this area bracket the theoretical curve for material at
at
in
a
per cent Later drainage relief and the
of
water content the season
in
6
,
of 9.
.
lateral transfer stress have caused deviations from the theoretical trend
.
While drainage favors stability
of
the lateral transfer stress increased the
,
raising pore pressure beneath the upstream slope
by
danger of sliding
.
STRENGTH OF UNSATURATED
IV
SHEAR COHESIVE SOILS

.
There are two important types unsaturated cohesive soils The com
of
1
:
(
)
pacted soils used as structural materials for earth dams highway and canal
,
and for backfill structures and natural deposits fine
of
of
embankments
2
;
,
(
)
grained soils that have either never been saturated or once were saturated
,
but have partially dried appreciable amount
of
an
both cases air
is
In
,
is . .
present the soil mass An important factor that influences the shear
in
strength of these soils the pore water pressures that develop phase
in
3
a
-
-
system as discussed Part III Moreover the presence air permits
of
in
,
,
.
specimens the laboratory without drainage both prior and
to
be tested
to
in
,
,
during application
of
shear stresses
.
Compacted Cohesive Soils

.
A
.
The concept of shear strength of unsaturated cohesive soils will be illus

rolled fill
by
compacted soils used fills During
of
trated construction
of in
at ,
a
.
the objective for each layer soil be identical and be compacted

is
to
to
,
the same water content and the same unit weight Immediately after com
to
,
paction can be assumed that the soil virtually unstressed externally As

is
it
,
.
was pointed out Part III however capillary pressures negative pore water
in
-
pressures exist the soil after placement These stresses are accompanied
in
.
)
by normal effective stresses which essentially are equal all directions

in
within the soil layer Hence virtually shear stresses exist As con
no
,
.
struction proceeds the load superimposed layers fill simultaneously ap

of
of
,
plies normal and shear stresses the soil causes change volume
to
to
in
it
,
and induces pore air pressures and changes capillary pressures

in
apparent then that the fill the shear process for each layer
of
soil
is
in
It
starts at the same void ratio the laboratory process of triaxial compression
;
without shear does not have counterpart the fill appears desirable
in
It
,
,
a
conduct the laboratory triaxial test compacted cohesive soils

on
therefore
to
,
such manner that each of the specimens whose strength will determine
in
the Mohr envelope subjected shear stresses while at the anticipated

to
is
s
'
placement void ratio and water content

.
Two procedures can be used accomplish this Compacting each

of
to
1
:
(
)
the specimens the anticipated placement water content and void ratio and
to
varying the chamber pressure and deviator stress simultaneously for each
specimen as approach failure along different vector paths all starting
so
( to
from the origin specimens can be prepared all at the same water content
2
;
larger void ratios than the contemplated placement condition with each
at
but
,
specimen at different void ratio specimen then placed the triaxial

is
in
A
a
machine and compressed with all around pressure without drainage

an
the
to
,
volume change the specimen

of
desired void ratio measured accurate

is
,
If
.
COHESIVE SOILS 93
control of void ratio can be accomplished . In applying chamber pressures to

bring all sealed specimens to the same void ratio , pore -water pressures will
be developed . If the initial void ratios of the specimens are properly chosen ,
the effective lateral stress (applied stress minus pore - water pressure ) will
be different for each specimen . In this manner , a strength envelope on the
basis of effective stresses can be obtained for an unsaturated sample of soil ,
which is comparable to fill conditions to the extent that shear started at
placement void ratio and water content . Method (1) is believed to be not as
practicable as Method ( 2) , which is used by the Bureau of Reclamation .
Casagrande ( 1953 ) pointed out that in consolidated quick triaxial shear tests
on saturated clays , the effective principal stress ratio at failure was larger
than in slow (drained ) tests on the same soil . This was attributed to pre
stress , caused by steadily falling effective lateral stresses during tests in
which pore pressures are present . Loci of computed points , representing ef
fective normal stress and shear stress on the failure plane , which are called
vector curves , were used to illustrate the phenomenon of prestress . Similar
ly , it is recognized that sealed triaxial tests on unsaturated embankment soils
may induce prestress in the specimen as its volume decreases under appli
cation of the deviator stresses . The magnitude of prestress in tests on un
saturated soils will depend on the amount of pore pressure developed during
the test . Although a small amount of prestress may occur in an embankment
on the verge of failure , the additional strength provided by prestress cannot
be counted on in a large fill where the applied lateral pressures are continu
ally increasing . The test procedure developed by the Bureau of Reclamation
to avoid prestress effects in determining the strength of unsaturated cohesive
soils is to use the constant effective lateral stress method . A description of
this method is given in Part II B 3 .
This method of shear testing avoids controversy over which criterion of
failure to use in the test . In the laboratory procedure for triaxial shear test
ing of compacted cohesive soils , using constant applied lateral pressure , the
criterion of failure used by the Bureau is the maximum principal effective
stress ratio (o 1/ 03 ) . Taylor ( 1950 ) expressed a preference for the criterion
of maximum deviator stress . In the method of test using constant effective
lateral stress , the maximum principal effective stress ratio and the maximum
deviator stress occur at the same point in the test .
The combination of starting the shear process on sealed , compacted soil
specimens , all at the same water content and void ratio , conducting the test so
that no prestress occurs , and measuring pore - water pressures , results in
strength envelopes shown in Fig . IV - 1. Two envelopes are shown for the same
data : ( 1) The solid circles are effective stress circles , based on pore
pressures measured at the top of the specimens using the perforated end plate
( 1 / 32 - inch - diameter holes ) to contact the soil ; and ( 2 ) the da shed circles are
effective stress circles based on pore pressures measured at the bottom of
the specimens , using a P - 3 filter disk . The two different values of pore
pressure , effective stress , c intercept , and tan 6, are shown paired in the
figure . The upper value corresponds to the top measurement .
A plot of the test data for one of the specimens (No . 4) is given in Fig .
IV - 2. The plots for the other specimens show similar characteristics . For
Specimen No . 4, the measurement of pore pressure at the top of the specimen
was initially close to 0 (plus 0. 7 psi) , while at the bottom of the specimen , the
ceramic disk contact measured minus 9. 4 psi . Also , compare the measured
pore pressures , top and bottom , during the test with the computed value of
pore air pressure , ua , obtained from Equation III - 9.
DEGREE
OF
MOISTURE
P(
DRYDENSITY SATURATION TESTVALUES FAILURE SHEARVALUES
AT
%
)(
PCF
)(%
CONTENT
DAY
CORRECTED
FOR
)CF
DENSITY
POREPRESSURE
g
PORE EFFECTIVE VOLUNE AXIAL VIATOR
DE
PLACE WETTED CONSOLI PLACE WETTEDPLACE WETTED LATERAL
OF STRAIN STRESS
(%
MENT PRESSURE PRESSURECHANGE COHESION
NENT DATED MENT
(
)
)
(
)
)
PSI PSI INITIAL PSI TAN PSI
SPECIMEN
3
6
.8
5
.2
116 118 81 00 17
7.733
( 13
8
0.0
17
( 01
.
.
0151
129
4
1
) 41. .9
.
.
5. 5
- 4-
118
19
) )0.7
3
.((
.3-))
116 79 76
+
13 81 72
3
13
( -
-
.
.
118 305
19
st16
91 150
.9 2
.
/6
7
. . .4
00. 2.
08
58
.. .
)
):
.
155 .1
.7-
115 118 75 57 23 09 115 668
14
PSI
–
STRESS SHEAR
40 80 120 160 200 240
EFFECTIVE NORMAL STRESS PSI
Borrow Area
15
AP
?
__
.
--.
0
0
--
-
--
A-
- -
-
-
--
.-
-
NO
NO
LABORATORY
SAMPLE 29F EXCAVATION 217 DEPTH FL
X16
Fig
of
IV
.
-.1-
Shear Strength Compacted Cohesive Soil
:
1
All aroundpressure Perforated and plate
S
! !" PP
Deviator
1
-
(
)
-+
of
Hour Topmeasurement
!0. :
%
20 Rate axialstrain 005
air
ug
=
Calculated Pore pressure
.I
3
.S
V Ceramic disc
.
-P- (
P
)
1 Bottom measurement
IN
Brojbrigadeiro
PRESSURE
PORE
COHESIVE SOILS
-
=
-
Calculated uc Bottom top

measurements
- 10
3
-
VOLUME CHANGE AV =
IV
.
vs
-.2-
.
in
on
Fig
a
Pore Pressures Volume Change Triaxial Test

96
Unsaturated Soil
A possible explanationfor the difference in the readings, from the perfor
ated top end plate and the bottom
ceramic end plate is that , in this case , the
top measurement is of pore air pressure rather than pore water pressure .
This might occur if air entered the 1 / 32 - inch holes of the top end plate during
the 16 - hour period allowed for initial readings , and disrupted the continuity of
water between soil and measuring device . This assumption is supported by
the observation that the pore pressure measured by the top end plate agrees
closely with the theoretical values for ua .
On the other hand , the pore -pressure measurements at the bottom start
near 1 atmosphere negative and increase with decrease in volume, passing
through O at 2. 4 per cent decrease in initial volume. The P - 3 ceramic disk
prevents air from entering the system ; hence , these readings are assumed to
include measurement of capillary pore water pressure . It is recognized , how
ever , that the speed of testing was not regulated by considerations of pore
pressure equilibrium . The rate of axial strain (0 . 055 per cent) was slow but ,
perhaps , not sufficiently slow for precise measurements of pore water
pressure .
By use of Equation III - 14, the capillary pressure , uc, for each test point of
Specimen 4 has been determined , using the bottom measurement for pore
water pressure as u , and the top measurement for ua : The plot of uc is shown
in Fig . IV - 2.
Since the top pore - pressure measurement which was used to adjust the ap
plied chamber pressure in this test apparently measured only ua , a constant
effective lateral stress did not result . However , there was no prestress be
cause the lateral effective stress increased at every step during the test .
These data indicate that the effective strength envelope of unsaturated soils
requires the measurement of pore - water pressure , including capillary
pressure , to secure the best result .
At the time of this writing , the Bureau of Reclamation Earth Laboratory
has had the opportunity to perform only a limited number of tests which in
volve pore - pressure readings at the top and bottom of specimens with perfor
ated and ceramic end plates . Plans are being made to fabricate devices to
secure readings through the perforated disks and ceramic disks at the same
end . Because of the limited examples , there are some anomalies which are
not clear to us at this time. For instance , the test reported in Part II , Fig .
II - 14, showed the same positive pore pressure at a perforated end plate , a
ceramic end plate , and a fine , porous central insert . This is different than the
results reported above ( Fig . IV - 2). It is conjectured that three possible con
ditions could exist which will be investigated further . They are as follows :
1. If the moisture content of the soil is sufficiently low , the water films
may be in discontinuous contact with the perforated end plate and that
device may be measuring only ua ; at the same time the ceramic device
could be measuring both u , and us ( Fig . IV - 2) .
2 . If the moisture content is sufficient for water films to contact both the
perforated and ceramic devices in the same manner , both devices will
measure ua + Uc , and give the same reading in the positive range ( Fig .
II - 14 ) .
3. If the moisture content if very low the ceramic end plate may not be
contacting the moisture of the soil , the pore - pressure reading of this
device could then be the capillary pressure of the ceramic material ,
plus ua rather than ua + Uc , of the soil.
COHESIVE SOILS 97
Tests made on un saturated soils , in which both correct positive pore

pressures and correct negative pore pressures are obtained and used , would
result in strength characteristics most applicable to the soils being investi
gated . The angle so obtained is most likely to have the proper relation to the
failure plane of all of the many o angles obtained in soil testing . It is believed
that this type of testing will be found fruitful in clarifying the fundamental
strength characteristics of unsaturated soils .
B. Unsaturated Soils . -
- undisturbed
These are natural deposits of soil which have either never been saturated ,
such as loess , or are soils which once were saturated but have become partly
dried out . In either case , the soil , if not perfectly dry , is affected by capillary
pressures which create an all - around compressive stress on the mass , as ex
plained in Part III .
Inthe ground , vertical stresses due to overburden loads and lateral stress
es are present in addition to the capillary pressures . Although there are only
meager data on this subject , it is reasonable to assume that the effective
stresses on an earth mass in the ground are not equal in all directions .
Hence, shear stresses must exist on all planes except the three principal
planes . The removal of overburden and lateral stresses when the soil is
sampled results in a tendency for expansion which is resisted , to some ex
tent , by the capillary pressure . A new equilibrium is established , depending
on the expansibility of the soil and on the magnitude of the capillary pressure .
Shear stresses will be different or may no longer exist in the laboratory
specimen .
Because of the existence of capillary stresses of appreciable magnitude in
the ground , removal of samples constitutes only a partial reduction in stress .
In soils of low degree of saturation at shallow depths, the reduction in confin
ing stress may be small . Hence , the void ratio and water content of
unsaturated - undisturbed soils are generally not subject to great variations be
tween in - situ and laboratory states .
These soils can be tested by cutting specimens from an undisturbed cube
or large - diameter core sample . Variations will be found in density and water
content among the specimens even on very carefully obtained samples . Obvi
ously, the Method ( 1) , used for preparing specimens of compacted soils , A
above , would be satisfactory but is very time - consuming , Method ( 2) cannot be
used . The sealed triaxial tests , using a different constant effective lateral
stress on each specimen , can be used for these soils .
C. Use
1.
of Test Results .
Embankments . In the design of embankments there are a variety of
procedures available to the engineer . These include selection of materials ,
use of two or more materials with differing characteristics to form a com
posite section , and applying various treatments to the soils to improve their
characteristics . Since , in the construction of all embankments , the soil re
ceives some kind of treatment , it is necessary to consider the kinds of treat
ment available first before any decision is made concerning selection of ma
terials for either a homogenous or composite section .
The kinds of treatment available are of two types : Physical treatment and
chemical treatment . The addition of chemicals to improve the properties of
soils is costly and complicated . It has been used only to a limited extent by
the Bureau of Reclamation and will not be discussed further in this paper .
Physical treatment includes removal of some of the components , mixing
differing varieties of soil together , moisture control , and compaction . Re
moval of some of the components is , in practice , confined to removal of over
size to facilitate compaction and removal of fines to improve permeability .
Frequently , the primary purpose for either of these operations is to improve
the shear strength of the end product . Mixing differing soils together is per
formed primarily to provide a homogeneous product with predictable charac
teristics , but it may also be done to produce a soil with specific character
istics , or to extend a limited supply of a desirable material . The control of
moisture will facilitate the densification of a soil either by compaction or
loading and so indirectly affects Compaction improves strength
its
strength
.
and makes more uniform and dependable
it
.
The extent which these various physical treatment procedures are ap
to
plied soil depends upon the objective cost associated with
is
There
to
a
a
.
each process and economical design requires the selection the cheapest
of
combination that will satisfy the requirements high
of
the construction
In
of .
embankments the optimum application these procedures usually
of
all
is
worthwhile Laboratory tests of soils are therefore based the assumption
on
,
,
.
that the best procedures will be used and samples are prepared match
to
prototype conditions
.
Under these conditions adequate shear strength along with such other
,
properties as may be required can be achieved by selecting soil from among
,
a
More commonly
is
those available that has the desired characteristics
it
.
preferable place the various types
of
soils available the embankment

to
in
section manner that their various characteristics are exploited
to
such
in
the utmost such procedure cohesive soils are exploited for their low
In
,
.
permeability rather than their shear strength and their deficiency shear
in
,
up
of
strength made through the effective use cohesionless soil

is
.
The outside slopes and hence the volume of embankment required appre
is
by
ciably influenced the shear strength the soil When two

of
more soils of or
.
with differing strength characteristics are used the strength the composite
,
determines the slopes that may be selected The mechanics by which slopes
on .
are determined most commonly based the assumption that failure will
is
occur along cylindrical path composite section the failure line may
In
a
a
deviate appreciably from such cylindrical path and the process making
of
a
stability analysis consists primarily trial and error search for the line
of
a
a
along which failure most likely occur

is
to
The procedure for making stability analysis simple mechani

on
is
based
a
cal principles but the proper application all the forces acting produce
of
to
failure and the application all the forces resisting failure can be both com
of
plex and time consuming Accordingly many engineers use simplifying as

,
.
sumptions and the loss of accuracy accepted as necessary hazard Among

is
the arguments justifying the use of such assumptions are That the critical
:
that the strength

of
the soil
of
strength soil cannot be determined the

in in
;
a
prototype differs from that tested the laboratory and that the soil the
in
too variable permit predicting its strength with any degree

is
embankment
to
of precision Much this paper devoted demonstrating that these argu

is
to
of
.
by
ments are not valid and accordingly the procedures used

of
the Bureau
Reclamation attempt apply all the known forces and strength factors each
to
to
stability analysis However much the drudgery the operation can

be
, of
of
,
.
eliminated with practice experience and the introduction of labor saving de

,
vices Presently being developed are procedures for performing the work by
.
electronic computers For this purpose cohesive soil subject pore

to
,
.
COHESIVE SOILS 99
pressure phenomena , both positive and negative phenomena is considered as

the basic material. Stresses induced by loading , compaction , changes in
moisture content , seepage , and earthquake may all be involved . Critical con
ditions include those prevailing at the end of construction , steady state with a
full reservoir , and following a rapid drawdown . In some instances there may
also be a critical stage during construction that requires evaluation .
In making a stability analysis the various factors required include : The
weight of material above the failure surface , the effective strength of the ma
terial along the failure plane , and the pore - pressure conditions anticipated .
All of the above factors must be predicated on the moisture conditions existing
in the materials for the condition analyzed . Densities and effective strength
are determined in the laboratory . Pore - pressure conditions are dependent
upon the situation being analyzed . A procedure for determining pore
pressures at the end of construction , assuming that no dissipation occurs , has
been developed by Hilf ( 1948 ) . This procedure is based on determining the
load - compression characteristics of the soils along the failure line by labora
permits the determination the pore
of
tory tests , which with Equation
on 11
is III
-
pressure
of
However this result the basis effective stresses and must
,
.
by
be
corrected Equation total stress basis for use stability

III
to
in
4
a
-
analysis Steady state and drawdown pore pressure determination involves

.
saturation conditions and discussed Part The stability analysis pro

is
on in
V
.
the Swedish slip

by
cedures used the Bureau are based modifications

to
circle analysis developed
by
on
May 1936 Procedures given various texts
in
(
).
Soil Mechanics are equally effective provided the proper factors describing
,
the situation are used

.
The foundations under structures embankments must

or
Foundations
2
.
-
.
commonly determine their adequacy support the structures

be
examined
to
to
some instances such as canals road cuts portions

or
placed upon them
of
In
,
.
the foundations are removed and stable slopes for the excavations are re
quired Solutions may involve the removal material bypassing with piles
of
,
.
material with unfavorable characteristics modification the questionable

of
,
of
material designing the structure tolerate the conditions the foundation

or
to
,
Prior making any decision treatment required the materi

as
material
to
to
,
.
of
als foundation within the zone influence must be classified into bodies
of
a
of
material with similar characteristics the boundaries these bodies must

of
of ,
established and the characteristics the materials within each

be
these
of
,
bodies must be determined

.
of
The division material with similar charac

of
foundation into bodies

a
primarily the basis of visual examination

on
accomplished
of of
teristics
is
surficial geological conditions and cores from drill holes the basis
on
a
,
general knowledge soils and the occurrence geological

of
of
of
the behavior
,
The cores from drill holes made for the purpose defining geologi
of
features
-
.
cal features may provide suitable specimens for testing but more commonly
,
special test holes are required

.
condition often experienced with unsaturated cohesive soils the arid

in
A
potential source low density or

of
of
west that
is
trouble combination
is
,
a
,
a
,
their geological
of
high void ratio and low moisture content On the basis

a
history these soils would be classed as normally consolidated except as dry

,
ing may have produced over consolidation effect but their load compression
an
,
-
characteristics determined the laboratory at natural moisture conditions

in
are those of an over consolidated soil However these soils are wetted
if
,
.
-
of
the load compression curve changes the form characteristic

to
sufficiently
,
100
a normally consolidated soil . Fortunately , this condition is usually only
superficial and can be corrected by removal of the low density material. If
the condition extends to appreciable depth , it can more economically be
corrected by a thorough wetting so that densification occurs as the soil is
loaded . Some examples of this phenomenon are described in Part VI. Experi
ence has shown that complete saturation is not necessary and is actually
detrimental in using this procedure . The addition of moisture softens the soil
and permits it to compress, but the shear strength is also reduced temporari
ly and if too much water is added prior to loading , or the loading is too rapid ,
the temporary pore pressures developed cause a reduction in shear strength .
If laboratory shear tests are made on the basis of effective stress , the rela
tively simple load consolidation tests and pore - pressure computations can be
used to provide the necessary supplemental information for the evaluation of
this kind of a condition .
The procedures for dealing with unsaturated cohesive soils in the moisture
and density ranges comparable or superior to that of disturbed compacted
soils is similar to the procedures already described for embankments , ex
cept , for recognizing their natural moisture and density conditions instead of
the Proctor standards . Those soils found in a nearly saturated condition will
be discussed in Part V on Saturated Cohesive Soils .
V. SHEAR STRENGTH OF SATURATED COHESIVE SOILS
A. General .
Shear testing of saturated cohesive soil is one of the most difficult oper
ations to perform in the laboratory . However , it has important practical
application because the shear strength of these soils is usually critical in de
sign . For very soft undi sturbed clays , the testing is particularly difficult and
the field vane test , described in Part VII , is showing increasing usefulness .
For more dense and firm undisturbed saturated clays , such as preconsolidated
clays , laboratory test results are more easily obtained .
Many theories have been developed on the basis of artificially prepared
samples , such as those remolded at near the liquid limit and artificially con
solidated in the laboratory to represent a preconsolidated condition . This pro
cedure of study has an advantage of a better control to develop theories ;
whereas , the usual undisturbed samples show irregularities . For the princi
pal discussion of this part of the paper , however , it is intended to describe a
test series on undisturbed samples of a natural soil deposit. Although the
soils used are in the category of a very soft clay , these tests were conducted
with extreme care , and significant results were obtained .
The problem of subsequent saturation of compacted soils is one which also
must be considered in earth dam construction . In this case , compacted soil
specimens are wetted in the laboratory to correspond to the condition of satu
ration by the reservoir . It is believed that the most pronounced change in the
shear characteristics at that time is one of reduced cohesion because of wet
ting .
B. Conditions Tests. --
of
The testing facilities of the Bureau laboratory are described under Part II ,
and the flexibility of the equipment to perform the various methods of tests is
discussed . The discussion here of testing conditions is pointed only toward
specific items related to the details of the various forms of testing saturated
cohesive soils .
COHESIVE SOILS 101
In shear tests on compacted cohesive soils , 1 or 2 additional speci

triaxial
mens are usually tested under wetted conditions to represent the situation of
subsequent wetting of the embankment . Truly saturated conditions are seldom
achieved in the laboratory , but it is believed that a thoroughly wetted speci
men can show the effect of high water content on the cohesion characteristics
of various soils . The degree to which thorough wetting is accomplished can
be determined to evaluate the quality of the test .
Undisturbed - saturated soils involve a separate test procedure , which is
also discussed in Part II . The tests are associated with foundations or other
in - place soil problems and are performed on undisturbed samples to evaluate
their shear characteristics with respect to their natural in - place soil con
ditions . Of utmost importance , is the care that must be given to the saturated
samples prior to testing . In the case of undisturbed samples of natural de
posits , the sampling technique must be the best possible and of a type most
fitting to the soil condition . For the examples discussed in this part of the
paper , a fixed - piston , thin - walled , 3 - inch - diameter sampler was used . The
specimens were prepared by cutting the sampling tubes into lengths just suf
ficient to obtain the required specimens . The tube walls were then cut length
wise and taken away from the sample rather than pressing the sample out of
the tubes. The test specimens , 1- 3 / 8 - inch diameter by 3 inches long , were
carefully trimmed to size in the manner described in Part II .
Saturated undisturbed specimens are tested in the triaxial shear machine
by partially draining them , that is , by the consolidated - sealed method , as
previously explained in Part II . This is necessary to obtain a spread of stress
circles and a relationship of shear stress to effective normal stress . Other
vise , saturated specimens tested under sealed conditions cannot develop in
creasing effective normal stress . To facilitate drainage , filter - paper strips
were placed along the sides of each specimen , prior to installing the rubber
membranes . Long periods of drainage were permitted during the consolidation
periods , to permit the pore pressure to stabilize at low values. This usually
was done overnight and for some specimens, mainly those tested under low
pressures , longer drainage periods were used . The valves were then closed
to seal the end plates . Since the specimens were near saturation , the possi
bility of air getting from the specimens into the drainage lines was small .
Nevertheless , an additional precaution was taken to evacuate the lines for a
few minutes prior to beginning the shear stress application .
In testing undisturbed saturated soils , it is particularly important to select
lateral pressures that will provide stress conditions at failure which describe
the critical features of a problem . These may be , ( 1) to evaluate in - place
conditions and characteristics for additional loading , or ( 2 ) may be only to es
tablish the shear strength in a certain range of loading . In any event , the
actual stress development for the test specimen cannot be predicted exactly .
Therefore , the selection of loadings becomes that of assuring a bracketing of
test results on each side of the range for which investigation is desired .
C. Tests on Compacted Wetted Soils . --

The shear test performed on compacted soil specimens which have been
thoroughly wetted after placement , is a part of the overall triaxial shear test
on compacted soil and only involves one or two specimens. It will be de
scribed only briefly .
The procedure usually followed in preparing a wetted specimen is to ( 1 )
compact the specimen to the same moisture and density conditions as the
other triaxial shear test specimens , ( 2) determine the weight and volume of
the specimen as compacted , (3) wrap the specimen in two layers of cheese
cloth and immerse it in a water bath for a minimum of 48 hours (7 days are
generally used for very plastic clays ) , ( 4) determine the weight and volume of
the wetted specimen , (5 ) test the specimen in a sealed condition , and (6 ) de
termine the moisture content of the specimen after testing . Variations of this
procedure have been used . For example , wetting of large ( 9 by 22 - 1 / 2 inches )
specimens has been accomplished by application of a head of percolating
water . Also , very expansive clays have been wetted in the water bath while
partially restraining the test specimens with a 1 / 8 - inch wire mesh container .
Specimens can be wetted by application of water while they are in the triaxial
chamber under load .
This procedure of testing for compacted soils does not result in a fully
saturated specimen . However , by determining the volume and weight of the
specimen at various stages of the test, the degree of saturation resulting from
the wetting process can be determined . The soils that exhibit greatest loss of
strength on wetting readily absorb moisture and reach high degrees of satu
ration .
It is believed that this test demonstrates the change in cohesive charac
teristics caused by wetting when the specimens are tested under low lateral
pressure . The test is conducted under sealed test conditions in the same
manner as the other triaxial shear tests on compacted cohesive soils . The
process of wetting the specimen may cause some change in the specimen , such
as swelling, but this is minimized by the wrappings while the specimen is sub
merged . The adequacy of the procedure and the suitability of the specimen to
represent the condition are checked by latter mea surements of the volume
change and moisture content .
D . Theoretical Comparisons and Test Results for Undisturbed Saturated

Soils . --
1. Theoretical considerations . It is well understood that so - called undis
turbed specimens are always disturbed to the extent of removing the over
burden loading . Also , it is well known that the normal consolidation develop
ment of a saturated cohesive soil deposit with increasing pressures , such as
increasing overburden loading , resembles in characteristics the Curve ABC
in Fig . V - 1. If it were possible to sample the soil with in - place condition at
Point B so perfectly that no strain ( expansion ) would take place , then all of the
effective stress existing in place would have to be replaced by internal tensile
stress , and the rebound and reloading curve would remain at Point B in Fig .
V - la . However , it is quite certain that this rebound during sampling does re
sult in expansion and a reloading curve as shown in Fig . V - lb . Forces tend
ing to retain the perfect rebound condition become complex , involving con
sideration of negative pore pressures .
There are several features tending to prevent the perfect condition from
being realized . These include the dissipation of the pore pressure during
sampling , and the usual inability of the water in the saturated soil to pull in
tension more than one atmosphere of pressure . Therefore , the actual sample ,
although excellently obtained , is certain to expand .
In this imaginary example , the shear characteristics that normally loaded
saturated clays develop are represented by a curve of increasing shearing
resistance with increasing effective pressure , resembling Curve ABC in Fig .
V - 2. A test on the perfectly obtained sample would be expected to resemble
COHESIVE SOILS 103
VOID RATIO
VOIDRATIO
EFFECTIVE PRESSURE EFFECTIVE PRESSURE

LOGSCALE
)
pothetical Condition of Perfect Undisturbance
VOID RATIO
VOID RATIO
O11
EFFECTIVE PRESSURE EFFECTIVE PRESSURE

LOGSCALE
(
Hypothetical High Degree

of
condition
of
b
a
(
)
Undisturbance Actual Sampling

in
Fig Hypothetical Considerations Consolidation Characteristics

of
V
1
.
-
.
-
of
Undisturbed Saturated Soils
Shear curve expected

SHEAR STRESS
for high degree

of
•
Cis
to
undisturbance Line from

B
-
.
shear curve expectedfor

perfect undisturbance
18
-
Shear curve expected

-
-
for normal consolidation
EFFECTIVE NORMAL STRESS

for
Fig Hypothetical Considerations

of
Shear Characteristics
V
2
-
-
.
Undisturbed Saturated Soils

Curve BC . However , since it is certain that some expansion will take place ,
the best undi sturbed sample would have a curve something like Curve DBC .
2 . Test Results of Undisturbed Samples . - With these simplified principles
in mind , the undisturbed samples of saturated soil are analyzed . Samples of
highly stressed soils have sufficiently pronounced characteristics that testing
will display these principles reasonably well . However , soft saturated soils
and sensitive soils are so critical under low pressure and are so easily dis
turbed by handling that the analysis of tests becomes very difficult . The test
series which are discussed in the following paragraphs , are on samples of a
soft normally loaded lake deposit from Willard Bay , in the Great Salt Lake .
The log of the test hole and the location of the samples tested are shown in
Fig . V - 3.
The test results of the two samples , Nos . 28K - 52 and 61 , indicated on the
log in Fig . V - 3 , are shown in Figs . V - 4 and V - 5 , respectively . Each of these
specimens with applied lateral pressures
of
six
12
tests involves
,
3
.1
6
.2
.5
HOLE NO DH 521
-
.
Ground Surface
-
-
0
SILT sandy some clay Water table

, , , , , ,
, ,
;
-
-
-
CLAY lean firm
- - - - -- - -
SILT some fine sand some clay

,
SAND fine silty

,
-
20
SILT firm some sand

,
–
SILT and LEAN CLAY soft

,
- -- -
CLAY lean soft

, ,
.
SAND fine clean
,
TEST SAMPLE
-
52
28K
-
-
FT
CLAY lean soft medium

to
firm
,
,
-
-
-
-
-
DEPTH
61
TEST SAMPLE 28K

-
-
-
-
CLAY lean with interstratified fat

, ,
80
-
-
-
-
ciay soft
to
firm
CLAY lean some sandy silt

,
,
-
-
-
100
by
Fig Conditions Represented the Undisturbed Samples for

V
3
-
--
.
Illustrating the Saturated Undisturbed Test

COHESIVE SOILS 105
.I.
S
-
P Note: Specimen4was
unsuitablefor test
STRESS
SHEAR
30
20 40 50
m
EFFECTIVE NORMAL STRESS PS-
.
.I
160 Tn
Spec
5
Spec
50
3
1
.
Spec
Spec24
?
Soer
40
1
1A
.
Spec
1
Spec
6
130
Spec
-
e Spec
3
VOID RATIO
One dimensional
consolidationtest
Spec
A
Spec
5
Spec
6
A
80
.0
10
100 1000
PRESSURE
.S
P
-
.
.I
failure eoch specimen

at
Ozond
=
=
A
A
0
-,
0,
Fig Laboratory Shear Test

on
Saturated Undisturbed Sample

V
-4
-
.
No 28K 52 from Depth

of
25
26
Feet
to
.
-
.P-
SI
STRESS
SHEAR
20 30 40 50 60
S
P
-
.
.
Spec .I
2
may
- -
Spec 387
- -
100
.
FE
Spec
6 4 5
Spec H Spec
2
Spec Spec
=
3
4 -
-
-
-
.e-
Spec
-
it
One dimensional
Po spec
-
consolidation test
5
.
-
+ Spec
6
VOID RATIO
--
Spec
7
–
--
10 100 1000
PRESSURE
.1
.S
P
.
-
Oz
failure each specimen

at
and
=
0,
,
0
7
•
-
-
Fig Laboratory Shear Test Saturated Undisturbed Sample

on
V
5
-
-
.
of
No 28K 61 from Depth Feet

71
72
to
-
.
COHESIVE SOILS 107
25 , 50 , and 100 psi. The tests were performed by the consolidated - sealed
method . The specimens had the usual irregularities , such as variations in
density and moisture , minor sampling disturbance , and slight variations in
material . The Mohr ' s plot is normal in its general appearance , but the full
value of the test is not obtained until an analysis with respect to volume change
is included . Sample 28K -52 was very soft with high void ratio . This sample
was difficult to test and was ea sily disturbed when handled . Sample 28K -61
was denser and had fewer irregularities and thus more systematic test re
sults .
The void ratio - pressure relationships are displayed in Figs . V - 4 and V - 5
in several different forms . First , there is the relationship shown by the one
dimensional consolidation ( confined compression ) test . This shows the soft
ness and higher compressibility of the upper material, Sample 28K - 52 , when
compared to the denser material from a greater depth , Sample 28K - 61 . The
initial void ratio for each shear specimen is shown on the left border of the
graph . The void ratio at failure for each of these specimens is also shown on
the graph plotted against effective axial pressure , 01 , and the deviator stress
( 01 - 03 ) . These void - ratio pressure relationships when plotted together are
similar to the one - dimensional consolidation curve , but offset on one side or
the other .
If the undisturbed sample and subsequent shear testing had perfect charac
teristics , a specimen , which fails under stresses less than the in - place stress
condition , should not change in void ratio . However , with the procedure used
unavoidable changes occur in the sample , such as drainage during reconsoli
dation and disturbance in sampling and testing . Therefore , a specimen failing
under stresses at near the preconsolidation (or overburden stress for normal
ly loaded clays ) should have a void ratio change approximately similar to that
displayed by the reconsolidation portion of the one -dimensional consolidation
curve . This is approximately true for Specimens 2 and 3 in Fig . V - 4, and
Specimens 4 and 5 in Fig . V - 5 , as shown by da shed lines between the initial
and failure void ratios . Therefore , the in - place stress conditions are near
that amount and the failure circles of these specimens indicate the in - place
shear resistance . Specimens tested at lower stress consolidate a lesser
amount and can even expand . Specimens tested at higher stress have greater
consolidation and their points are farther down the normal consolidation
curve .
As can be seen in these tests on actual field samples , the characteristics
discussed in relation to the hypothetical examples in Figs . V - 1 and V - 2 are
demonstrated but not as well defined . The characteristics shown by the upper
sample , Fig . V - 4 , are not as well defined as those shown by the lower sample ,
Fig . V - 5 . It is of interest to plot the void ratio change characteristics of both
samples on one graph as shown in Fig . V - 6 . The relationship shows that the
upper sample which has lower initial densities has had less overburden or
preconsolidation stress . Thus , it has lower shear strength in - place . The
lower sample shows characteristics in a higher category of strength , higher
range of densities , and higher preconsolidation pressures but in a comparable
pattern of void ratio versus pressure relationships . (Also , shown in Fig . V - 6
are the results of a third series of tests , 162 - 8 , on remolded soil which will
be discussed later . )
The shear test results alone do not indicate the in - place conditions on these
samples . The soils are such that inherent strength of the natural soil is al
most lost or obscured upon sampling and is difficult to recognize . The
1503
El Initial e for
to specimens of
Sample 28K - 52
[
1.40
1.30
--Consolidation test
sample 28K - 52
-
e
A
RATIO
387 Initial for

e
specimens
- of
45
FF
sample 28K 61
-
VOID
90
.0
Consolidation test
Initial for sample 28K 61
e
"
-
specimens
- of
80
C
0
FjSample 167
.
8
70
0
.
60
0
.
50
.0
10
100 1000
.7
PRESSURE
S
P
.I
.
.
-
EXPLANATION
52
Sample 28K 11 failure each specimen

at at at
03
= = =
Ö
A
A
, - 0,
, , ,
,
- - -
- 11= -
Sample 28K 61 01 03 failure each specimen

Sample 162 03 failure each specimen
8
0
0
•
,
-
Fig Comparisons One dimensional Consolidation Tests and Volume

of
V
6
of -
--
.
.
-
Change Characteristics Shear Test Specimens for the Examples

Illustrating Saturated Soil Tests
COHESIVE SOILS 109
consolidation characteristics give much more information and value to the

shear test . This method of analysis would be particularly suited for stiff and
preconsolidated clays in which shear resistance more nearly resembles Curve
DBC in Fig . V - 2.
test results on a typical sample . - In the tests on Samples No .
3. Detailed
28K - 52 and 12 specimens were tested . Each specimen test was an ex
61 ,
tensive study in itself . The specimens contained end plates for pore - pressure
measurements at each end , a perforated end plate at the top , and a ceramic
end plate at the bottom . The equipment and test procedure described in Part
II were used , and particular care was given to pore - pressure readings and
other observations of stress and volume change . It is not possible to report
these data here but detailed characteristics are demonstrated by the
all
of
representative specimens
of
results .
Fig
61
shows detailed plotting the shear test for Sample No 28K
of
, V
7
-
-
.
.
Initially the failure points for each specimen were considered
on
, of
the basis
maximum deviator stress and maximum principal stress ratio most tests
In
.
these were quite similar but the maximum stress ratio was considered most
,
appropriate because of the pronounced pore pressure development

of
effect
.
first approximation the average angle
of
slope of
of
Using the envelope
in a
,
which this case had tangent 52 the vector curves Taylor 1943 shown
of
a
,
.
)
the
part
of
Fig were drawn for point observation made

of
lower each
in in
V
7
-
.
the test
.
of
44
Specimen effective normal stress
an
For example started with
3
6
,
.
around pressure was held constant
all psi Although
at
The applied
50
psi
.
.
-
drainage was permitted during the consolidation period the specimen still had
,
pore pressure psi when

of
was sealed begin the deviator stress

to
to it
a
5
7
.
.
As deviator stress applied the specimen under sealed conditions the
is
effective lateral stress reduced as pore pressure develops For each point ,
is
by
an
effective stress circle can be drawn as illustrated

of
observation one
,
example for the third point of observation on Specimen The point tangen
of
on 6
.
the point
cy
slope
52
on
the stress circle for the vector curve

is
of
0
a
For each succeeding test observation the vector curve progresses until the
of ,
failure circle reached the vector curve gives confidence and

is
The use
.
convincing evidence for the selection failure points Such characteristics

of
distinct hook back of the vector curve several points of observation very
as
a
,
-
close together or top and several points observation tak

on
of
of of
one another
,
ing path along the failure envelope curve are all evidence failure
In
a
Specimen appears that the maximum deviator stress point would be un

it
7
,
satisfactory since there are several points beyond which are closer
it
to
the
probable envelope where the other evidence
of
failure exists
.
interesting note that the vector curves for Specimens

is
to
and
It
2
3
4
,
,
show minor development pore pressure whereas curves for Specimens

of
,
;
and slope the other direction because of considerable pore pressure

in
6
7
,
The probable cause that the stresses ap

of
development this occurrence

is
.
Specimens within the range

of
are less than stress con

or
plied and
in
2
3
4
,
,
ditions existing place and the specimens are resisting consolidation re

or
in
versing the stresses developed when the soil rebounded during sampling This
.
further evidence addition that shown by Fig that the break point
is
in
to
,
,
V
5
.
-
.5 the
shear characteristics somewhere between that shown by Specimens

is
in
and
Figs show the full history the testing and the stresses and
of
and
V
V
8
9
.
-
volume changes that occurred for Specimens and respectively Sample

of
,
,
3
6
Specimen
at
InitialSpecimen
Doto TestValuesFailure ShearValves
No Correctedfor
Asprepared Failure
Condition AppliedEffectivelvolume PorePressure
Dry Moisture Dry Lateral Laterai Change Deviotor
Degree Degree Pressure
Moisture
of
of
.
DensityContent DensityContent P ressure Stress Cohesion
Soturotion Saturation si si Initial si Tan
I
cf et
2 .2 .4 )
34343434 (%
(%
(%
1513 (p
)
(p
(p
11
7 . .2 .6 .7 4 )
6, .7 . .2 .4 .7 )
1 )
.8 3, . 1 .0 . )
)
91 41
(
)
(p
86 (p
71
3435
85. 95. 100
6 3
2
. .9
. . .1
.5 .2 .
|
99
85 94. 88 66
2
2 7 2 7 3
.7
.1 .9 4
0
5 0 967
3
.4 .3 9 .19
.0 12
9 8686
33
89. 100 801
. 2
. . .5
496
37 17
.0 25
90 32 100 23.
.2 9
. 5
5
.8 15
.0 50
29
8734
.3 95
86 33 94 100
100
. 7
6
61
28
12
97
.27
99
.85
34
95
52
0
*
.7
T
.
Averoge
of
value Ton used
*
for
drowing
vectorcurves
-
SHEAR STRESS PSI
1020
30
60
70
80
90
40
50
oo 100
-
Mohr circle plot with tabulation

of
shear characteristics
a
(a
,
)
data for each specimen

at
of
the final selection failure
of
EXPLANATION
-
SHEAR STRESS PSI
Specimen Mox 03
0,
..
.
/oOj -
Mor.
,
-
-
by
Foilureselected
pec observingvector
3
.
'S
curves
15
10
20 25
-
-
SHEAR STRESS PSI
of
Seeinsert
Ton
52
lorgerscole
:0
.
lo
30
60
70
50
40 BO
-
for
Vector curves each test specimen

b
(
)
for
61
Fig
No
Detailed Triaxial Shear Test Data Sample 28K

V
7
.
-
-
.
-
COHESIVE SOILS 111
TIME - HOURS
1 18 20 22 24
Consolidation
ku Sealed
periodº
Tv
periodi
717
-
period
PSI
| Lateral stress applied Axial stress applied
It croire
<- - 1- - <- - - - - -
; Bottom 'ceramic
- Top perforated |
PRESSURE
idend plate i disc end plate

PORE
Top perforated
Bottom ceramic
LL
'
disc end plote end plote
10
20
AXIAL STRAIN
%
-
b
(
)
Specimen No 30
K 3
28 .
Sample No 61
-
.
Initial conditions
- -
%
- - :
pcf
PSI
Density 85
34 .7
-
-
-
-
-
- -
Moist content
%
.4
_
.
Degree
of
saturation
CHANGE
94
STRESS
indicated
%
Deviator stress
7
-
-
-
-
.
-
-
Void ratio 996

0
-
-
-
-
-
-
.
Applied loterol pressure

psi
VOLUME
6
.2
DEVIATOR
Failure conditions
:
pcf Volume change

88
Density
34 .0
-
-
-
-
-
. -
- -
Moist content
%
3
-
-
.
Degree
of
saturation
. 99
indicated
.7
- -
- -
- -
-
- -
Void ratio 943

0
-
10
AXIAL STRAIN
%
-
)
(c
Fig Example Specimen Sample 28K 61

of
Detailed Data
,
V
8
3
-
-
-
.
.
6 No
61 Specimen relatively low applied lateral stress

at
28K was tested

6 3
a
.
-
.
psi and Specimen was tested at higher applied lateral stress 50 psi
of
In of
2
.
these figures pore pressure observations are shown just after

of
Parts
,
(a
-
)
installation and during the consolidation period under the all around pressure
-
.
.Fig
Specimen development negative pore pressure

of
of
8a
shows
. In
,
V
a
on -
psi the ceramic end plate after allowing the specimen stand over
to
2
5
The negative pressure

of
night evidence some retained stress the soil

is
in
.
after sampling
.
Following this observation lateral pressure psi was applied and

of
,
6
.2
a
pore pressure values immediately positive values near psi When

of
rose
to
.
-
the
plates were opened drainage this pressure gradually dissipated

to
to
end
,
TIME - HOURS
16 18 20 22 24 40
41 . Consolidation
periodi
30 |
-
1 / Sealed period - - >
PSI
Axiol stress applied

Loterol stress
applied
PRESSURE
Top perforated end plate
-- Top perforated
+
+
end plate
-
PORE
Bottom ceramic disc end plote
-
t
-
- Bottom ceramic
k
- end plote
-
-
-
- 10
CH
5
AXIAL STRAIN
%
.
b
(
)
a
(
)
Specimen No
28 .6
No .
61
Sample
K
.
Initial conditions
- -
:
Density 86 pcf
.7 9
PSI
%
-
-
-
-
-
-
Moist content 33
%
Volume change
.
-
-
-
Degree saturation
of
indicated 95
5
-
-
-
-
-
STRESS
CHANGE
Void ratio 969
0
-
-
-
-
-
-
.
Applied lateral pressure

To
50 psi Deviator stress

-
.
VOLUME
DEVIATOR
Failure conditions
:
pcf
94
Density
29.3
-
-
-
-
- -
- -
Moist content
%
lo
.7
.
2
.5
Degree
of
soturation
%
indicated 100
- .
- -
- -
-
- -
- -
Void ratio 818

0
-
10
5
AXIAL STRAIN
%
-
C
(
)
Fig Example Specimen Sample 28K 61

of
Detailed Data
,
V
9
6
-
.-
-
--
-
.
COHESIVE SOILS 113
near zero . The pore pressure was periodically checked by briefly closing the
drainage valves and taking a reading . In this specimen , nearly complete
drainage occurred over the weekend . The specimen was then sealed and devi
ator stress was applied as shown in Fig . V - 8c . Also shown in this graph is
the volume change curve which remained nearly constant as would be expected
for the saturated condition . Pore - pressure measurements on the two end
plates during this time increased to near 2 . 5 and 3. 0 psi and then dropped in
value after failure . It appears that the ceramic end plate registered a reason
able value of initial negative pressure , and both end plates satisfactorily
registered positive pressures .
The same characteristics are shown for Specimen 6 with 50 -psi lateral
pressure , except that higher stresses and volume changes were involved .
Here again , a negative pressure of 2 . 5 psi was initially shown . At the end of
42 hours of drainage , when the specimen was sealed and deviator stress was
applied , a pore pressure of 5. 7 psi still remained , making the starting ef
fective stress 44 . 3 psi . During the sealed part of the test , the volume re
mained constant and increasing pore pressures were observed on the two end
plates .
During the tests of all specimens , the slowest available rate of testing was
used (rate of axial strain = 0. 0025 inch or 0 . 08 per cent per minute ) to assure
good pore - pressure readings .
E. Tests on Remolded Saturated Soil .

In 1951 , a series of shear tests were performed on a remolded saturated
soil , which also was from Willard Bay, Utah . The sample was from a differ
ent location in the Bay , but in general was quite similar in characteristics to
the two undisturbed samples just discussed . The soil was placed in an 8 - inch
diameter cylinder at a density of 89 . 3 pcf and moisture content of 33 per cent ,
and was loaded and allowed to consolidate by confined compression under a
load of 50 psi to produce a condition of preconsolidation . The sample was re
moved from the cylinder and cut into 10 shear specimens of which 9 specimens
were suitable for testing in triaxial shear .
Fig . V - 10 gives the shear test results of this sample . The data are simi
lar to those shown by the undisturbed sample , but remolding tends to obscure
the effects of preconsolidation . Here , consideration of the volume change data
is required to obtain the full significance . As can be seen in the figure , all
specimens were very uniform in initial void ratio which should be expected in
a remolded specimen . Since the void ratio changes were small for Specimens
1, 2, 3, 4 , and 6 from initial condition to failure condition , these specimens
represent shear strength in the reloading range for the preconsolidation load
ing of 50 psi , which was applied to the soil in the laboratory . Specimens 7, 8 ,
9, and 10 represent shear strength for stresses beyond this preconsolidation
loading .
Specimens 1, 2, 3, and 4 were tested under sealed conditions throughout the
test. Therefore , their volume changes and effective stresses were limited to
that causing saturation . Specimens 6 , 7, 8, 9, and 10 were tested under
consolidated - sealed conditions with the lowest applied lateral load being 12 . 5
psi . However , this load was sufficiently small to permit Specimen 6 to show
the reloading condition along with the four sealed specimens .
When the void ratio data of this sample , No . 16 Z - 8 , is plotted in Fig . V - 6 ,
along with the data of the undisturbed samples , the results are generally com
patible . It represents a denser condition because it has been remolded and
100
.P-
STRESS
NOTE
Specimens were tested
SHEAR
2
3
4
1,
,
,
under sealed conditions
10 .
Specimens and were
6,
7,
8
9
,
tested under consolidated sealed
-
conditions
.
Specimen was unsuitable for
5
testing
.
40 60 80 100 120 140 160
S
P
-
.
.
I.
-e Initial Void Spec Spec

i
. 2
,
.
.
800
VOID RATIO 1
.
Ratio Spec
✓
-
3
-
-
-
Spec
4
* -
-
--
--
-
Spec
cot
-- 6
- .
-
-
-
-
-
Spec
-
-
• -
o -
7
.
Spec
8
.
tat
Spec
9
.
.
oo Spec
10
t
10 100 1000
PRESSURE
.S
.I.
-P
03
ot
and failure each specimen

0
0
o
•
-
:
Laboratory Shear
on
Fig Sample
10
Test Remolded Saturated 162

V
8
-
-
.
.
COHESIVE SOILS 115
densified by the 50 - psi loading in the laboratory ; whereas , the undisturbed

samples are loaded a lesser amount by their overburden .
F. Summary of Strength of Saturated Soils . --

The shear strength of saturated soil is dependent on the effective stress
acting on the soil particles , the soil type , and the soil structure which is at
tained either through natural deposition processes or recompaction process
es . On this basis , each soil is analyzed by determining the shear strength
relation with changing effective stress , and the individual characteristics of
the different soil types or soil structures are shown by different shear
strength and effective stress relationships .
For undisturbed , saturated soils , the test sample must always go through
the sequence of unloading during sampling and reapplication of loads during
testing . Therefore , the important related consolidation theories described by
Casagrande ( 1932 ) must be utilized . A soil specimen tested in the laboratory
shows inherent strength that the soil structure has attained in the field . This
characteristic is known to be more pronounced in undisturbed samples of
natural deposits than in recompacted and remolded soils .
In all tests involving reloading of saturated soils , some drainage must be
permitted to obtain a spread of effective stresses. If it were possible merely
to reapply stresses occurring in place , it should theoretically be possible to
evaluate directly the shear strength in place. Some investigators have pro
posed making a simple , quick , unconfined compression test on saturated clays
to do this . However , the soil tends to expand and an appreciable portion of the
in - place stresses have been released when the sample was obtained . Only the
structure of the soil is retained . Reloading a specimen in the laboratory , if
it is saturated or even near saturated , develops pore pressure , which must be
drained in order to reapply the effective stresses which were lost through
sampling . Therefore , it appears that , under present testing conditions , the
unloading during sampling can not be avoided , and the reapplication of loads
with drainage must be considered in testing saturated soil to determine in
place shear strength . The resistance to consolidation in shear testing should
be similar to that expected in the consolidation test described by Ca sa grande .
Previous investigators (Waterways Experiment Station , 1947 ) have related
the compressibility characteristics to changes in water content since the
drainage of the saturated shear specimens were measured . However , since
the Bureau of Reclamation ' s instrument has the facilities for measuring
volume change during the test , a direct relationship of void ratio versus
pressure is possible and a direct comparison to the previously known princi
ples of consolidation can then be made .
The envelope lines of shear resistance in the examples shown have been
drawn tangent to the stress circle of each specimen . For these soils, the
distinct break in the curvature of the shear envelope is obscured , but the ad
ditional characteristics shown by the void ratio data are of great assistance
in showing where it probably occurs . It is probable , because of expansion and
disturbance in sampling , that the reloading portion of the shear test shows
somewhat lower values of shear resistance than the field deposit would have
in the process of unloading and reloading . However , for effective stress in
creases greater than in - place conditions , the laboratory test gives a reason
able representation of shear resistance . In either case , inherent strengths
resulting from preconsolidation pressures should be considered , and method
described in this part of the paper is a practical way of evaluating this
characteristic .
G. Use of Test Results
1. Embankment subject to saturation . - Saturated embankment can occur in
three ways : ( 1) By using hydraulic fill or puddled fill construction , ( 2) by a
combination of selected moisture content and loading which will result in
by
saturation as described in Part percolating water during oper
III
and
3
,
(
)
Conditions requiring hydraulic fill construction have not been
en
ation
.
the soil
of
countered by the Bureau
of
Reclamation since the establishment
testing laboratory and will not therefore be discussed further Puddled fill
.
extensively used situations where the plastic nature the material per
of
is
in
not detri
of
mits the development intimate contact and its low strength
an
is
mental Saturation from loading was experienced
of
some the structures
in
.
built during the period the pore pressure theory was being developed Fortu
.
-
nately the portions of the embankment where this occurred were sufficiently
,
reinforced that the loss of strength these portions was not detrimental
to
in
the total structure Where such conditions are anticipated modifications of
,
.
the practices recommended for hydraulic fill construction offer possible so
a
Saturation by percolating water during performance
of
structure
is
lution
a
.
common condition requiring analysis for steady state and drawdown con
a
ditions
.
For steady state and drawdown conditions the stability analysis made
is
if
,
on the basis of effective stresses the principal factor determine the pore
to
is
,
pressure effective under these conditions Procedures for establishing
a
.
steady state flow net from which pore pressures can be determined have been
,
Hypotheses for the determination draw
by
Casagrande
of
described 1937
A
(
.
).
down conditions have been advanced by many investigators including Cedargren

1940 Skempton 1954 and Lowe 1959 the experience record of the
In
),
),
(
)
.
Reclamation situations have not been encountered which support

of
Bureau
,
a
need for criteria as critical as these hypotheses require Therefore the
,
.
Reclamation uses its experience record for the determination of

of
Bureau
pore pressures following drawdown
.
the analysis steady state and drawdown conditions also necessary

of
it
is
In
consider the changes that occur cohesion during saturation The lightly
to
to
loaded wetted tests previously described show that for most soils sub
a
,
stantial part of the cohesion measured unsaturated samples lost upon

is
in
When complete saturation not achieved

is
saturation these tests some

in
,
.
judgment required estimating the saturated cohesive strength component

is
in
.
Since the contribution stability usually small saturated cohesion can be
is
to
neglected most steady state and drawdown analyses

in
Foundations There are foundation situations involving saturated co

2
.
.
-
hesive soils that can be solved without recourse laboratory tests These
to
are those situations where the beds saturated cohesive soils are obvious
of
of
low strength and satisfactory foundation can often be

ly
limited extent
of
A
.
secured economically by removal by bypassing the saturated cohesive soils

,
with piles caissons or other similar devices by drainage

or
,
Those situations where the bodies saturated cohesive soils are suffici
of
ently extensive that investigations and analysis are required may be further
,
subdivided into normally consolidated soils and over consolidated soils The
,
.
-
has built several embankments on foundations contain

of
Bureau Reclamation
ing normally consolidated sediments including cohesive strata which were of
limited thickness but too extensive permit economical removal For these
to
situations spread structure has been satisfactory determine the

To
a
strength of the foundations laboratory tests as previously described have

,
,
COHESIVE SOILS 117
been used in stability analyses to determine the extent of spreading necessary .

With this type of design , construction limitations to prevent differential load
ing more severe than that provided by the design has been sufficient . In the
case of Willard Dam , however , where the conditions described earlier in this
part occur , additional restrictions involving the rate of loading have been re
quired .
Willard Dam has presented the greatest challenge thus far to the theories
of soil mechanics the Bureau of Reclamation has adopted . Initial explorations
on the basis of penetration resistance tests , indicated that the Willard Dam
foundations were virtually without strength to a depth of 200 feet or more , ex
cept adjacent to the abutments . It has been stated that the slopes of embank
ments built on purely cohesive soils extending to great depths without increase
in strength is unimportant, and the height to which an embankment can be built
is limited by the amount of cohesion available in the foundation , Terzaghi and
Peck ( 1948 ) . This statement is based on the fact that the minimum failure
conditions are defined by the deepest failure circle that the geometry of the
structure will permit . Unless there is a limiting boundary restricting the
depth to which a circle can penetrate , increasing the size of the footing pro
duces no benefical results .
However , in the belief that improvements in strength with depth would be
found in the Willard Dam foundation , investigations were continued to de
termine as accurately as possible the amount of strength actually available .
The suite of tests previously described and the vane tests shown in Fig . VII - 8
demonstrate that there is an increase in strength with depth and also permit
an approximate determination of the loading conditions that will produce vari
ous conditions of strength .
Testing has also shown that the undisturbed strength is substantially great
er than the disturbed strength . In considering earthquake effects , it is desir
able to assume that the disturbed strength will prevail , but in the process of
improving strength of the foundation by additional loading , utilization of the
undi sturbed strength is considered permissible .
To develop the dam , stage construction has been adopted . In each stage an
amount of embankment is placed which will not over stress the foundation , as
determined by the tests described in this part , but will permit consolidation
and , therefore , improvement with strength as drainage occurs . Piezometers
are installed in the foundation at various depths so that it may be determined
that sufficient consolidation has occurred to permit the next stage of con
struction to proceed . Vane tests of the gain in strength are also made for
verification ; Fig . VII - 9 shows the results of such tests on the test embank
ment . The second stage of construction is presently being accomplished .
Over - consolidated soils should have greater strength than normally con
solidated soils and therefore may offer less of a problem . However , complete
stress relief in over - consolidated soils may not have occurred , and so further
loss in strength is possible . Excavations change the stress conditions , and
negative pore pressures in saturated foundations can develop . These negative
pore pressures will eventually be lost through the movement of pore fluid and
may be followed by heaving or swelling . It may require years for sufficient
stress relief to occur to permit a deep - seated failure , but superficial ravel
ling or slumping may follow excavation rather rapidly . In temporary exca
vations superficial loss of strength from drying and checking has been suc
cessfully prevented by protecting the surfaces of excavations with an asphalt
emulsion or gunite applied immediately after exposure . The deep - seated
failure which may occur on permanent slopes , is more difficult to analyze and
requires flattened slopes or overexcavation and reloading for a permanent
solution .
3. Sensitive clay foundations . -- It is frequently possible to construct
structures on sensitive clay foundations which stress the soil beyond its re
worked strength . However , under adverse conditions , such as an external
shock , these materials may suddenly change from their natural to their re
worked strength and sudden failure occurs . This situation is particularly
dangerous when it is not recognized in advance . However , it is often advan
tageous for temporary construction to rely on at least a part of this strength
for economic reasons , if the final structure does not over stress the foundation
beyond its reworked strength . Under these conditions , trenches for pipelines
or conduits and cutoff trenches under dams have been successfully accom
plished .
Where permanent construction is involved if its cost of replacement is not
great, a calculated risk is often accepted . Such examples include minor canals
and laterals , minor road construction , and small storage buildings . In areas
where sensitivity is not great and is often unrecognized , even substantial
structures are found on such foundations . Since there are existing structures
on such foundations it is often difficult to convince the owners of the hazard
involved . For this type of foundation the Bureau of Reclamation has not es
tablished guiding principles , rather each problem is considered on its merits .
The problem is one of deciding what strength parameters to use rather than a
selection of an analytical procedure .
VI . SOILS WITH SPECIAL PROBLEMS IN SHEAR
A. Expansive Clays
Throughout
.
almost all of the western part of the United States , soils and
bedrock materials containing expansive montmorillonite minerals often occur .
These minerals have been developed or have been transported throughout
many sections of the area . Expansive soils have created problems of uplift
and instability on Bureau of Reclamation hydraulic structures , ( Holtz , 1959 ,
Holtz and Gibbs , 1956 ) , because there is an opportunity for water to become
available and thus facilitate the expansion of the clay minerals .
The shear strengths of all clay soils are influenced by moisture changes ,
but expansive clays are often subject to extreme changes in shear strength
because of extreme moisture changes . In addition to the strength factors re
lated to the minerals involved , the interrelation of moisture , density , and load
plays an important part in the strength .
Triaxial shear tests can be used to determine the shear strength . How
ever , the sequence of loading and wetting (or drying ) greatly affects the volume
changes and strength . Therefore , it is important to duplicate prototype con
ditions closely . Figs . VI - 1 and VI - 2 are typical shear test data plots . These
tests were made on compacted and undisturbed Beaumont (Ca - Beidelite )
clays , respectively , from the Texas Basins Project, Texas. The soils were
classified as CH (Unified Soil Classification System ) with an average liquid
limit of 70 . A most important fact to be noted from these tests is the loss of
cohesion when each soil was wetted and dried and rewetted from its initial
conditions . The soil which was compacted to 90 per cent of Proctor maximum
density at 2 - 1 / 2 per cent less than optimum moisture content , had a cohesion
01
at
)7
(
Initial Specimen Dota
/.
Test Values Failure Max ShearValues
As prepored Wetted
Constant Correctedfor
PorePressure
Applied Effective Volume Deviator
of
of
Moisture Degree Dry Moisture Degree Lateral Lateral Change
Stress
of
%
Density ContentSaturation Density ContentSaturation Pressure Pressure Cohesion
(
psi Ton
%
)%(
(
)
(
)
psi Initial
(
)
pcf pcf psi psi
No Specimen
-
( 6
64
25
80 58 34
( -
80 25 64 Sealed Tests
12
95
37
WN
80 63 25 28 40
11
) 6. 6. 3. 5.
14
) 3.1 7.4 .2 .
.5- .7- . .
--
.0
) 2. 5. 0. O.
80 64 12
25 25
50
27 41 85 48 20
1114
5
79 27 66 72 43 89
98
. .8
)%1 7. 8. 6. 7. 4. .8
)%( )3. 5. 5. . 6. .0
B
..
1
0
.. 28
.. 77
) 9. .5 .5 .5 61. .
.70 .40
82 22 60 88
74
140
(
Fat Clay
) CH
Organic black
88
71
LL
.,
7
SL
Effective
stress
56
PI%
%
Sand 002
mm
PSI
Applied
.
- stress Specimen wet ted

,
Specimen wetted
air
.,,0. "A""B",
.6 -
dried and rewetted
COHESIVE SOILS
UNCONFINED COMPRESSION
20
STRESS SHEAR
20 40 60 80 120
100

119
-.1
of
-
-
VI
.
,
Fig Shear Test Compacted Expansive Clay Texas Basin Project Texas
at
.
/5.
(
)
Initial Specimen Data Test Values Failure Max oz ShearValues
Corrected for
120
As
prepared Constant Wetted
Applied Effective Volume Pore Pressure
Deviator
|
.
of
Dry Moisture Degree Dry Moisture Degreeof Lateral Lateral Change Stress
%
Density Content Saturation Density Content Saturation Pressure Pressure of Cohesion
)
(
psi Tan
(
(
)
)%
%
(
)
(
(
)
)%
pcf pcf psi psi Initial psi
No Specimen
7
86 32 91 93 12
2 9
( 36
1 3
9
87 31 91 Sealed Tests 34
( 8
85 33 91 05
12
13 17
8 767
( . .
87 31 90 Consolidated 25 21 12 27
68
8 1
87 31 90 Sealed Tests 50 38 64 32
6
.0
) . . 2. . . 0.
) 1. 2. 5. 0. 0. 0.
.0- .1- .4- .- .- .-
31
11
15
87 92 100 94 53
16
1
0
88 33 98 85 35 100
78
.9 0.
..
8
L
. 0.
7642
) 5. 0 4. 4 3. 7. 1. 7.
)% 0. 7. 1. 0. . 4. . 3.
. . . .....
. ..
4 20
529
52
856452
87 32 95 79 41 100
Fat Clay CH
Organic black
61
47
LL
8
SL
( .,
.,
Effective
46
0
Pl%
%
) . ,
401 stress Sond 002 mm
-
- Applied Specimen wetted
SHEAR STRENGTH
B "A" ""
.,
stress Specimen wetted

air
. . . . . . . . .9 -
dried and rewetted
PSI STRESS
4
0
UN CONFINED COMPRESSION
20
SHEAR
CONFERENCE
20 40 60 80 100 120
-
UNDISTURBED EFFECTIVE NORMAL STRESS PSI

of
VI
.
-.2-
-
,
Fig
Shear Test Undisturbed Expansive Clay Texas Basin Project Texas
COHESIVE SOILS 121
of 14 psi .
Upon wetting , the density decreased and the cohesion was reduced
to 0 . 7 psi .
When the soil was air - dried and rewetted , further density decrease
occurred , and the cohesion was reduced to 0 . 4 psi . The undisturbed soil
showed similar strength loss . Under the same sequence of testing , the co
hesion varied from 4 . 5 psi at natural moisture to 2 . 2 psi to 0 . 9 psi , respective
ly .
Fig . VI- 3 is a photograph of a series of slides in an earth section on the
Friant - Kern Canal . The soil is Porterville clay of the Ca - Beidelite type and
was classified as CH ( average L . L . = 63 ) . The canal section is 23 feet deep
and has 1 - 1 / 2 horizontal to 1 vertical side slopes . A short section of slopes
rebuilt experimentally on 2: 1 were also unstable .
Fig . VI - 4 is a typical slope condition for a canal section in expansive clay .
As shown on the figure , the cohesion at the base of the sliding arc (C1) will be
influenced by saturation , low loading , and possible drying and resaturation .
High volume changes and resulting low cohesion can , therefore , be anticipated .
The soil above this segment and below the saturation line will be saturated ,
will be moderately loaded , and will be less likely to be subject to cycles of
wetting and drying . Therefore , the cohesion (C2 ) of this segment may be
greater than Ci. Above the line of saturation , free water conditions do not
exist , and the cohesion of the soil (C3 ) is influenced by capillary and other
tension forces , and may be quite high . As shrinkage cracks often open to
depths of several feet when expansive - type clays become very dry , the co
hesive (CA ) and frictional strength at the uppermost part of the arc cannot be
relied upon . The above factors must be considered in any stability study .
Fig . VI - 5 is a photograph of a slide of a 1- 1 / 4 : 1 slope of a concrete - lined
section of the Friant - Kern Canal . Deep , longitudinal shrinkage cracks oc
curred extensively along the banks , and the clays at the base of the slopes be
came soft . The extent of shrinkage cracking can be seen readily at the upper
part of the slide in the figure . Slides of this type have been occurring from 2
to 10 years after the canal was placed in operation and point up the need to
consider time effects in the strength of expansive clays .
B. Density Soils . -
Low
1. Loessial soils . - The Bureau of Reclamation has encountered loessial
soils in many sections of the western part of the United States . The largest
area being in the Kansas and Nebraska region . Loessial - type soils have been
encountered in the states of Washington , Idaho , Montana , North and South
Dakota , and Colorado . Because of the low density which may be associated
with these aeolian silt deposits the problems involving consolidation and sta
bility are frequent , particularly with hydraulic structures in which wetting is
common . Loessial deposits in Kansas and Nebraska cover an area in which
the Bureau has many irrigation works ; for this reason we have conducted a
large amount of research on the properties of loessial soils in this area (Holtz
and Gibbs, 1951 ) .
By photomicrograph observations of undisturbed loess specimens , the loose
arrangement of the silt particles with typical numerous voids and root - like
channels can easily be seen . By petrographic methods it has been determined
that a majority of the grains are coated with very thin films of
montmorillonite - type clay , which are responsible for intergranular braces in
the structure . Upon wetting , the clay bond is readily loosened causing great
loss of strength . Fig . VI - 6 shows the gradation and plasticity characteristics
of loessial soils in the Kansas - Nebraska area . As shown in this figure , these
orabnego
T- PX - D - 16023
Fig . VI - 3. - Slide in Earth Section - Friant - Kern Canal , California
390 Shear strength may be lost Al l

from shrinkage cracks - k - - 24" - - -
S=3: 1
îF 10.0' E
Water surface One Compacted in U S = 2 :1
soil 2
- 16. 89
Line of Saturation
AVAVI Natural soil LLLLLLLY _
H
- - C2
K - Ci - Cohesion of partially saturated
Cohesion of saturated soils influenced by negative
soil under moderate forces
Cohesion influenced by light
hº load and possible alternate bon1oodings . TOT U
wetting and drying
Fig . VI - 4. - Design of Slopes in Expansive Clays

COHESIVE SOILS 123
T-PA -D - 16021
Fig . VI - 5 . - Slide in Concrete - Lined Section - Friant - Kern Canal, California
soils have been divided into three groups - clayey , silty , and sandy loess .
Based on many samples tested from this area , about 76 per cent were silty
loess , 18 per cent clayey loess , and 6 per cent sandy loess .
The shear resistance is variable and is dependent upon the initial density ,
initial moisture content , and plasticity of the loess . Fig . VI - 7 summarizes
some of the variations in the shear resistance of natural loessial soil . The
solid lines represent soils at natural moisture content , and the da shed lines
soils which had been prewetted . Although the tests included both sealed and
drained specimens , the results have been adjusted for pore pressure ; hence ,
the shear resistance curves represent strengths for effective (grain - to - grain )
stresses .
It can be seen that the shear strength curves have nearly similar slopes .
This would be expected because the soils have very similar basic physical
properties . However , there is considerable difference in the position of the
shear resistance curves . This is caused by a variation in cohesion which is
related to the density , moisture content , and clay content . The several curves
which are bracketed and located high on the graph are those showing high co
hesive characteristics . It may be noted on the figure that these are soils
which are clayey , of very high density or of very low moisture content , or
soils with combinations of these properties . The curves bracketed at the low
er levels on the graph are those of typical silty loess, generally of relatively
low density . This lower group of curves contains test results for two test
conditions , natural moisture and the material wetted prior to testing . It may
be seen that the low density loess which has been thoroughly wetted has a very
low shear resistance . Under low normal stresses , the shear resistance line
for this condition may be very flat and for some soil conditions there is
practically no shear resistance . This would be anticipated from the plasticity
124
HYDROMETER ANALYSIS SIEVE ANALYSIS

CLEAR SQUAREOPENINGS
U
..
TIME READINGS STANDARDSERIES
.
.
3
"
5
"6
'
"8
7HR15MIN
.
.
.
..
.
60MIN MIN MIN ININ
S50
25HR45MIN
100
200
*
19
*
*
IMIN
TOO
CLAYEY LOESSH
Line
-
ZONE
SILTY LOESS
INDEX
ZONE
LOESS
ZONE
SANDY LOESS
PASSINGPERCENT
LOESS
RETAINED PERCENT
LOESS
CLAYÉY
\
PLASTICITY
MD
SANDY
SILTY
50
60
20
30
LIQUID LIMIT
. 2
7
6
.1
.
1
38
.
.
.
..
49
0 42
19
19
.9 52
002
.4 76
005 009 019 037 074 297 590

38
.1"
127
me 152
IN
DIAMETER PARTICLE MILLIMETERS

OF
SAND GRAVEL
COBBLES
(
. )
. -(
)
.02.2 I
TO
CLAY PLASTIC SILT NON PLASTIC FINE MEDIUM COARSE FINE COARSE
:
NOTES GRADATION TEST
V1
of
for
.
Fig
-.6- Trends Gradation and Plasticity Loess
.
LAB TYPE DENSITY
MOIST
,
SAMPLE TESTI TIME TIME
OF
ARKS
NO CONDITIONS
LOESS TEST TEST
-
ATOF
120202
2
. 1.
120202NATURALCLAYEY
10
150
12F44 SILTY
12D200 CLAYEV
142478 SILTY
ATOF 1009484 90
12D183 SILTY 84 SExtremely
cloyey
LoessondLoess
of
0
12F143 SILTY103
of
Extremelycloyey
Loess andLoess highdensity
.
highdensityandlowmoisturecontent 142526 CLAYEY 83 ondlowmois
1813
turecontent
12D181 SILTY
79
6
9
-
-
MZ527 SILTY 89
120
200
142476 SILTY 83
12F
-
0
- 135106 139106 SILTY 09
139 SILTY 81
111
-
12F143 13S109 SILTY 83
0 5 3 5
135112 SILTY 81
-
-
142527 13S
-------- --- -- --
142524 SILTY 82
111
-
>
. .1 .9 .2 .2 .5 .1 .1 .5 .8 . .7 .8 .3 .0 .6
. 3. 28. . 68. . . . . 6. 3.
I
12D191 CLAYEY Typicolsilty

78
6 6 7 9 1020
COHESIVE
PSI STRESS SHEAR

12D181
of
Loess low
density
- +
476
14
PRE
-
+
1
-
13S109AND
135112 135
105 WETTED SILTY 81 29
-Z 4-
Z
3 4
SOILS
478 13S SILTY 82
14
30
-
-
--
|
1201837 135111 142476 SILTY 81

9
142
521 SILTY
-
-
+
. . 34
142526 13S105
142
524 SILTY 78 33
- -111- - - -
. .9 9.
139108 SILTY 28
,-
. .5 .6 .278 0. .478
-
of
142524 TypicalsiltyLoess density
low
of
TypicolsiltyLoess EXPLANATION
-
120191 naturalmoisturecontent Tests naturalmoisturecondition
of
of of
--
Typicalsilty Loess Tests wettedcondition

highlywettedcondition
&
-
,
142476521 524
- 35
E
-
t
108
100 150
-
VI
of
.
-.7-
Fig
125
Shear Tests Representative Loessial Soils

testdata because , after wetting , these low density soils have moisture contents
equal to or higher than their liquid limits at which near zero shear strength is
presumed . However , under higher normal stresses the strengths increased
because frictional resistance was developed after consolidation took place .
As a further explanation of the difference between wet and dry conditions ,
which are most important when building hydraulic structures on these soils ,
the curves in Fig . VI - 8 are shown . These curves are for a typical silty loess
of relatively low density . The solid curves are for undisturbed loess at natur
al moisture conditions , and the dashed curves are for the same material which
was wetted prior to testing . The shear resistance curves in Part (a ) of the
figure show distinct differences in shear resistance between wet and dry con
ditions . From these shear strength curves the great difference in cohesion at
dry and wet conditions is apparent. Graph (b ) shows the consolidation which
occurred in the shear specimens. It will be noted that the wetted loess con
solidated under much lower stress conditions than the dry loess . It is inter
esting to note that wetting has caused some expansion in the loess , since the
initial density of the wetted loess is considerably less than the initial density
of the dry loess . These specimens were obtained from the same field sample
and had similar densities prior to testing .
Certain generalized shear characteristics have been established for
loessial - type soils :
a . " Dry ” loess , at natural moisture contents (less than 10 per cent) gener
ally has inherent shear strengths resembling cohesion , which permits
natural loess to stand on very steep slopes for considerable heights, al
though it may be of low density . This cohesive strength is apparently
due to the small amount of clay which gives the loess a particle - to
particle bond . We have observed this strength to be as much as 15 psi .
Generally , it is about 5 to 10 psi for the Nebraska - Kansas loess . The
tan o is about 0 . 60 to 0.65 . Such values of c and tan o (effective stress
INITIAL CONDITIONS NATURAL PONDED

DRY DENSITY(pcp) . . . . . . .
MOISTURECONTENT(%) . .
DEGREEOF SATURATION
- - . - - .82.02 .
10.57 .
- - - - - 78.02
- . . 33. 96
(%) - - - - - - - - 27.50- - - - - - - - .80.23
SPECIFIC GRAVITY 2.660
FAILURE(MAX. 3 / 7 ) -75
Sealed Tests
-
e
-
SHEARSTRESS PSI
-
DENSITYpcf
VOIDRATIO
+
+
1
-
-
100
120
20
40
60
80
100
0
1
EFFECTIVE NORMALSTRESS- PSI DEVIATORSTRESS- PSI

(a
)
(b
)
Fig Shear and Volume Change Characteristics Dry and Wet Loess
VI
of
8
-
--
.
.
COHESIVE SOILS 127
basis ) permit loess to stand on steep or nearly vertical slopes 50 to 80

feet high .
Wetted loess is greatly reduced in strength . Experience shows that co
hesion is generally reduced to less than 1 psi when wetted , and even for
initially dense loess , it becomes less than 4 psi when wetted . The
breakdown in particle bond can occur at 20 to 25 per cent moisture
content, which is in the order of 50 to 60 per cent saturation . An im
portant finding in our studies is that high pore pressure develops in
highly wetted loess as consolidation occurs , and therefore , the develop
ment of additional shear resistance may be restricted until drainage
takes place . Experience has shown that drainage may be sufficiently
slow to cause lateral movement of embankment foundations if con
struction is too rapid .
2 . Colluvial soils . - Certain colluvial soils , particularly those which have
a silty soil matrix and are of low density , may cause instability problems in
engineering structures . Again , this is particularly true with hydraulic struc
tures where the soils may become wetted during operation . Soils of this type
may also become unstable on natural slopes when wetted by high rainfall or
floods . As an example , a non - Bureau dam and reservoir subsidence problem
in Nevada is cited . In this structure , the foundation soils were deposited by
flash floods and slumping movements from nearby hills . The material was a
well - graded sandy gravel with appreciable silt , having about 60 per cent
gravel, 28 per cent sand , and 12 per cent soil finer than the No . 200 size .
With materials containing this amount of gravel ( larger than No . 4 sieve size )
the high consolidation and instability which was experienced upon wetting the
reservoir would not be expected . It was found upon examination that the dry
density of the soil matrix (material finer than the No . 4 sieve size ) was low ,
being 106 . 6 pounds per cubic foot . The dry density of the total material was
126 . 5 pounds per cubic foot .
Triaxial shear tests were made with 3 - 1/ 4 - by 9 - inch specimens on the
soil matrix and with the large 9- by 22 - 1 / 2 - inch specimens on material up to
1 - 1/ 2 - inch maximum size . The results of the first test are shown on the
shear strength plot, Fig . VI - 9 . On this figure the large volume changes which
occurred under high stresses are readily apparent. Even though the cohesion
for this material was low at its natural moisture content , the cohesion was
completely lost upon wetting the specimens , as evidenced by the fact that both
medium and large specimens slumped under their own weight upon wetting .
The large specimens when dumped into a pan appeared like wet concrete with
a 4 - or 5 - inch slump , as shown by the photograph included on the figure .
The purpose of this example is to point up the fact that colluvial soils which
contain a low density soil matrix , may be subject to severe settlements and
loss of stability even though the material may have a high gravel content .
C. Residual Soils . -
In our workat the Bureau , we have encountered residual soils at several
worldwide locations , including a few locations in the United States . These
soils have been principally disintegrated granites , decomposed granites , kao
linitic clays , and halloy site clays . Residual soils , as discussed herein , are
considered to be soils developed by various stages of rock deterioration in
place . These soils can result from mechanical rupture of rock caused by
temperature changes , chemical alteration of the rock materials , or by a com
bination of the two processes . The use of residual soils in earth and rockfill
128
..
(
at
Initial SpecimenData Tost Values Failure
lure Max
Max TJ ShearValue
No Corrected for
Constant Pore Pressure
1
Specific Dry of
Molsture Degree Applied Pore Effectivel Volume Deviator
Gravity Density Content Saturation Loteral Pressure Lateral Change Stress Cohesion
%
of
$
Pressure
(
Pressure
)%
(
(
psi
%
pcf
(
psi
Ton
(
)
)
(
psi psi Initial psi
Specimen
+
.2
4A 106 34
73
0.7
( 07
28
34
12
106
.80
.0
50
) 8. 4. 2.
..
) 0.4 9. .3
.. .
( .0- .5 .
45 6
) 818
. 7. 5 1.
37 78
16
2A 106 34 60 22 122
7
4
IA 106
) 4. 4. 7. 6. 8.
) .67 .67 5.7 . 7.5
|
35
117
FOUNDATION
Prewetted
-
Specimens PX0 20370
IA
Minus material
"4"lif
--|
Minus material
-
Sealed Tests
PSI STRESS SHEAR

20 40
-

a
,
of
VI
.
-.9-
Fig Shear Test Colluvial Soil Nevada
.
.
Rio Initial specimenDato Test Values of Failure Mox 09 Sheor Value
Constant Corrected for
Specific Effectivel Volume Pore pressure
of
Dry Molature Degree Applied Por Deviator
Lateral Change
|
Gravity Density Content Isaturation Lateral Pressure Stress Cohesion
Pressure
1
Pressure Ton
(
)
(
%
pcf
(
(
psi psi
Specimen
(|)
pal psi psi
Initial
13
118
16
01
) 45
123
) 37
12 12 77 74
12
4
( .. .
) 9. . .
07
118
2
/7 .%of .0 1. 3.
( 3. . .
) 3. 3. .8
13
( .3 . .
) 150
25
25 80 115
.0
94
-
COHESIVE
P81 STRESS
SOILS
SHEAR
/F
-
T
31
20
30
40
50
.O
70
60
EFFECTIVE NORMAL STRESS - PSI
10
.
-
-.
of
a
Fig VI Shear Test
-
Residual
129
Disintegrated Clayey Granite Puerto Rico

dams has been discussed by Walker ( 1959). A few examples of residual soils
which we have tested for shear properties are given below to provide some
shear information on a range of materials .
During the design of Caonillas Dam in Puerto Rico , it was necessary to de
termine the stability of a disintegrated granite deposit which formed a ridge
between the proposed reservoir and the river channel . Large hand - cut undis
turbed samples were submitted to the laboratory for shear testing . The soil
was very friable and exhibited numerous fracture planes which were filled
with clay . The samples were classified petrographically as residual disinte
grated granite , containing several seams of very fine white clay at a small
angle with the vertical . The material was further classified as a well - graded
silty sand with no particles larger than the No . 4 size and containing 17 per
cent fines .
Permeability tests were made on each of the 3- 1 / 4 - by 9 - inch specimens
which were prepared for shear testing by cutting from 1 -cubic foot blocks .
After saturation and permeability testing , the specimens were sheared under
different applied lateral pressures with drainage permitted during the tests .
Fig . VI - 10 shows the results of one of the shear tests . Included on this figure ,
is a photograph of one of the shear test specimens after failure . The test re
sults indicated a high tan 0 of 0 . 94 and a relatively low c value . The high
friction appeared to be related to the angularity , the intimate interlocking of
the particles , and the high in - place density of the material.
Another problem involving residual soils concerned the stability of the
foundation and embankment for the Pierce Dam in Singapore . Two types of
residual soils were encountered in the dam area . The first was a decomposed
granite . The samples received were soft , white and green , and were plastic
when moist . Petrographic analyses showed the material to be granites which
were altered in place . The fabric of the original granite was readily apparent .
The samples were composed of 40 per cent kaolinite , 30 per cent quartz , 25
per cent hydrobiotite vermiculite , and small amounts of orthoclase . These
soils were classified as ML which , in some cases , were borderline with SC ,
CL , and MH . The gradation and plasticity characteristics of this soil are
shown on Fig . VI - 11 .
Both triaxial shear tests and unconfined compression tests were made on
the natural decomposed granite . The triaxial shear test indicated tan o values
of 0. 72 and cohesion values of 1. 0 psi . These data are shown on Fig . VI - 12 .
The unconfined compression test gave c values of from 3. 4 to 34 . 4. The co
low
hesive strength of this material was when the moisture content was high
.
The second residual material tested from this area was kaolin clay which
a
was classified as MH The gradation and plasticity data for representative

a
.
sample are shown Fig VI 11 was not possible test undisturbed foun
on
to
It
.
.
-
dation samples of this material because of the small size and limited number
core samples received However four triaxial shear test specimens were
of
,
.
prepared from similar materials from proposed borrow area These were
a
re
by
at
compacted standard Proctor compactive effort field moisture The

.
Fig VI
of
the shear tests are shown

.13
on
sults This test indicated tan

of a
o
.
To -
psi
of
48
of
this moisture wider range

at
and secure
ef
0
3
4
a
c
a
.
fective lateral pressures and shear strength

on
secure data lower

at
to
,
moisture conditions additional specimens were prepared the same density

at
,
but with per cent lower moisture content The tan for this condition was
48 4
o
-
and the was psi One additional test specimen was prepared
at
also
.0
8
.5
c
.
HYDROMETER ANALYSIS
SIEVE ANALYSIS
7
.
..SU
NR MIN TIME READINGS STANDARDSERIES CLEAR SQUAREOPENINGS
15
.
I
,
4
25WR 45MIN
._
0
.
GONIN MIN MIN
19
IMIN 100 504
30
100
16
20
KAOLIN CLAY
--
KAOLIN CLAY
-
.
GRANITE CLAYT
PASSINGPERCENT
RETAINEDPERCENT
INDEX PLASTICITY
GRANITE CLAY
COHESIVE SOILS
20
30
40
50
60
LIQUID LIMIT
.
.
.
002
.
.0
.
005009 019
1
2
0
037 .
4
074
.9
149
.
.1
1
297 42 590
7
6
2
19
.
.
76
52
19
38 2152
.. 2
38
IN
DIAMETER OF PARTICLE MILLIMETERS
SAND GRAVEL
)
(. -
)
TO
CLAY PLASTIC SILT NON PLASTIC FINE MEDIUM COARSE FINE COARSE COBBLES
(:
NOTES
GRADATION TEST
-
VI
Fig
11
.
of
-
-.
-
Gradation and Plasticity Residual Clays Singapore

131
.
at
(
)
Initial SpecimenDato TestValues Foilure Mor Ölos
132
Sheor
Values
As prepared Wetted Constant Correctedfor
Applied Effective Volume Deviator PorePressure
of
Dry MoistureDegree Dry MoistureDegree Lateral Lateral Change
Stress
%
of
Density ContentSaturationDensity ContentSafurationPressure Pressure Ton Cohesion
(
psi
(
(
(
(
)
(
(
of )
(
(
)%
)%
%
(
)
pcf pet psi psi Initial psi
NaSpecimen
20W
2004 88 25 2501 9T 36
10
L
2
| 1
) 1. 3.
)% .6 6.
) . 0.
92 22 100 28 54
11
0
4
1
211 93 22 12
8
1
14
.5 .0
92 23 50
1939
11
2
0
245 77 141
12
15
| T
50
46
21 73 47 50 21
1
286 94 21 100 34 97
-2 .0 -1 2-
L
9
350 87 30 25 12
. .7 9. .5 7. 4. .
1
6
) . 7. 0. 3. 2. 3. 1. 6. .
3
3
. . 4, 0. .5 6. 2.
) 1. 3. 3. 8. 0. 3. 4. 2. 3.
.37- 5.- 5.5- -. -. 2.3- 9.9- .53- 4.4-

1
. . 0. 0. 0.
86 100 13
31
0
.
.
1
5
70
Sealed Tests
PSI
-
STRESS SHEAR
30 40
50
60
10
70
80
90
100
110
120
Fig
of
VI
a
.
-. 12
-
-
Shear Test Residual Decomposed Granite Clay Singapore

COHESIVE SOILS 133
Specimen
No Initial Specimen
Data
f
TestValueso FailureMax / Shear
Values
.
As
prepared Wetted Constant for
Corrected
AppliedEffectiveVolumeDeviator Pore Pressure
Dry Moisture Degree Dry Moisture Degree Lateral Lateral Change
|
of
of
DensityContent DensityContent
Saturation Sofuration Stress Cohesion
of
%
et Pressure
P ressure si Ton
ct psi si
(%
(%
Initial
(%
)
17 (p
1
)
)
)
5 )
12 (
.0 0 .5 .0 0 .5 )
(p
(p
28.91 88.
(p
89.
91
.6 8 6
.4 3
.3
.5 .4 .9
- -4 -2
.5 .0 .3
89. 28. 88. NaturalMoisture 50. 20
9
2
.2 5
15
89 29. 88
0 100 10 48
.8
3
.4
4
2
.
12
88. 25. 75. 33.
- -8 -4
. .5 .1
.0 8
2 26
8 21
5
497
7 5
14
.13.
%
71890 25. 176 Bolo Natural 50. 50.
. 51
89
10
24. 75. Molstura 100 21. 58. 48
0
2
i
.
89. 28. 86 89. 30 002
.0
99
. 6
2 .4
.3
.0 0
8
.6
51 30.61 93.
28
89. 88. 89. 27 61

6
0
9
5
2
.5
.
Sealed Tests
-
SMEARSTRESSP81
BelowNatural Moisture
4
Natural Moisture
-
Saturated
10
20
30
40
50
60
70
80
90
100
EFFECTIVE NORMALSTRESS- PSI
Fig Singapore
V1
Clay
of
13
Shear Test Residual Kaolin

-
-
.
.
Specimen
Initial Specimen
Data
f
No
TestValueso FailureMon Shear

Values
(
Asprepored Wetted Constant Corrected

for
AppliedEffectiveVolume Pore
Pressure
Dry Moisture Degree Dry Moisture Degree Lateral Lateral ChongoDeviator
of
of
DensityContent DensityContent Stress Cohesion

of
%
Safuration Sofuration
Pressure
Pressure psi Tan
ef c pai si si
(%
(%
(%
(%
Initial
(
)
DE
)
6 .8 .0 .8 )
12(
.0 0 0 .5 )
587 )
)
)
(p
(p
(p
(p
54J82
10
25- 61. 86.

.1 1
.7
.2 .2 .8 6
1
. 4
362
-9 -6 - -
.
54. 84 25. 20. 100

.6 .7 3
21
2 .4 0
61 54. 82 50. 37. 119

24
61. 54.31 82. 100 61 135

.7
34 28.
0
5
.
160 954682 31
05
06
46
25- 17.
0
+
.5
-0
50 .
.
25
61
52. 81.
35
.1 1
-2
200133 11
+
62. 59. 93
+
10
.0
+
5
.0
.9
.0
0.
Sealed Tests
-
PSI
88
SNEARSTRE
50 100 180
EFFECTIVE NORMALSTRESS 731
-
Residual Halloysite Clay

VI
14
Fig
of
Shear Test Guam

-
--
--
.
.
natural moisture and saturated for one week . The c , as determined by this
test, was 2. 5 psi, showing an appreciable drop in cohesive strength upon in
creasing the moisture .
A third problem involving residual soils concerned the stability of proposed
embankment soil for the Fena River Dam in Guam . This material was a
clayey , compressible silt of low plasticity and was classified as MH . The
Petrographic Laboratory described the material as a probable altered tuff ,
composed mainly of a soft , massive halloysite clay matrix intermixed with red
iron oxides and some montmorillonite clay minerals which were embedded
relict rock fragments . A photograph of this material is included on Fig . VI
14. Halloysite clays have some unique properties . Changes in plasticity and
strength may result from changes in the hydrated form of the halloy site clay .
Materials in the transition state of hydration frequently are highly plastic ;
whereas , the material in either the low or high hydrate form is relatively non
plastic . Also , the material may harden somewhat on exposure to air under
constant moisture conditions . Considerable difficulty was experienced in test
ing this material and it was necessary to minimize adjustments in moisture
content and excessive handling and reworking in the laboratory . Fig . VI- 14
shows the results of one of the shear tests made on this material . In addition
to the standard triaxial specimens at optimum moisture and Proctor maximum
density conditions , an unconfined compression test was made on one saturated
specimen to determine the saturated compressive strength . A tan o de
termined by the test was 0 . 34 , and the c was 28 . 5 psi for the soil at optimum
moisture , and 11 psi for the lowest strength saturated specimen , showing a
considerable loss of cohesion upon wetting .
A fourth study involving residual soils was that conducted for the Trinity
Dam in California (Walker , 1959 ). These soils which were being studied for
possible borrow for the dam were generally highly weathered decomposed
gravel - size particles , which were soft and when wetted broke down readily to
organic elastic silt . While there was some variation in the many samples test
ed from several locations , the materials were described generally from a
petrographic standpoint as being composed of sandy and pebbly ferruginous
clayey silt containing up to 50 per cent kaolinite and halloysite , and in some
ca ses 15 per cent vermiculite . Quartz , felspars , schists , and granites make
up the majority of the remaining materials .
Prior to starting the shear testing program , a laboratory study was under
taken to determine the effect of drying on the soil properties and to determine
whether there would be irreversible changes which would invalidate the usual
type of testing procedures . Eleven undisturbed drive samples were obtained .
Three of the samples were split and each of two portions subjected to standard
properties tests , one - half of the material being maintained and tested at natur
al moisture and the second half air - dried prior to testing . The drying pro
cedures for moisture tests were also studied . It was found that the major
change in properties occurred in the compaction test , where differences in the
Proctor maximum density varied from 1. 2 to 4 . 4 pounds per cubic foot for the
three samples tested . In each case , the material which was air - dried and
brought back to the desired moisture content showed higher densities than the
material at natural moisture which was wetted or dried to the desired moisture
content . One of the sets of compaction test curves for this study is shown on
Fig . VI - 15 .
Fig . VI - 16 shows the results of triaxial shear tests performed on one of the
borrow material samples prepared without preliminary drying to simulate
COHESIVE SOILS 135
MOISTURE
- DRYDENSITYCURVES
..
FT
AIR DRIED BEFORE TESTING.
LIQUID LIMIT 55 , PLASTICITY
INDEX 13
.
CU
LOS PER
OPT. MOISTURE 28. 8
-
MAX . DENSITY 88 . 5 i
DENSITY
OPT MOISTURE 33
3
.
8.
DRY
MAX DENSITY 85
.
TESTED AT NATURAL MOISTURE
-
.
LIQUID LIMIT 55 PLASTICITY
,
17
INDEX
20 25 30 35
OF
MOISTURE PERCENT DRY WEIGHT

-
COMPACTION TEST CURVES

VI
Fig Wetting and Drying Compaction Halloysite type

of
of
15
on
Effect
-
-
.
Soils Trinity Dam California

,
-
prototype conditions Several tests were performed material of

on
this type
.
under constant applied lateral pressure and constant effective lateral pressure
conditions all cases the constant applied lateral pressures showed some
In
,
.
what higher values and lower values than the constant effective lateral
o
pressure tests the case illustrated by Fig VI

16
the tan values were

In
0
,
.
-
15 .
psi and
47
38
and and the cohesion values were psi respectively

.19
0
.0
.
,
0
.
.
As indicated by two unconfined compression tests wetted condition there
in
a
was considerable loss cohesive strength upon wetting the soil The differ
of
ence values between the constant applied and constant effective later
shear
in
pressure
tests would normally be expected because the range
of
void ratios
at al
the start of the test and at failure different

is
From our limited experience with the testing and use

of
residual soils
is
it
our opinion that they problems when used for em

do
not offer insurmountable

Testing should be used
or
bankment foundation construction establish the

to
.
basic properties
of
the materials and irreversible changes that take place

,
should be properly recognized For instance drying halloy site clays af

of
if
,
.
fect the properties through dehydration of the clay and anticipated that the
is
it
soils will not dry during the construction operation the tests should be made
,
samples without drying from their natural conditions The loss

on
of
cohesive
.
strength upon wetting usually pronounced

is
Shales
.
D
.
Examples of shear testing which are used the following discussions re

in
fer highly consolidated claystones and siltstones which may be fractured

to
and which may contain seams Shales are commonly

of
bentonitic materials
.
by
sampled hand methods wherein blocks least cubic foot size are
at at
of
,
;
. -
cut from test pits when the explorations are shallow depths When deep
Moisture Degreeof Shear Values
(
)
Dry Density pcf Saturation Test Values at Failure
%
.
(
)
Content Ole Corrected for
136
No
Applied Effective Pore Pressure
Volume Axial Deviator
I
Consoli Place Place edl Lateral Lateral
Placement Wetted Change Strain Stress
of
%
dated ment Pressure Pressure
Wetted ment Wettedly
Tono Coheson
)
%
(
(
)
(
Specimen
psi psi Initial psi psi
0
80 84 86 100 47 117
81 84 87 50 97
3636
9
5
( 369
7 51
0 42
957
. . 0. .
.
) .4 9. 7.
- 7. 3. 0.
. ..
...
81 82 35 86 61
12
QUIN
.
.
0
0
47
15
)(1
79 42 97
10
44
..
..
79 42 97
) 4. 9. ,2 7.7 2.
12
5 .
3
Apparent saturated cohesion
7
9
82 83 86 11 43 63
13
10
81 83 84 60 42 44 99
35 35
(2
)
ON 8
...
...
. 9. 3.
103
251
... 892
... 207
. .6 .4
. .4 4.
83 35 85 34
25
81 45 83
.
.
1
0
38
19
Sealed tests
H
80
PSI
SHEAR STRENGTH
oz
- Constant
()2
Constant og
STRESS
SHEAR
CONFERENCE
20
0
60
80
40 140 160
100
-

,
Dam California
16
of
Halloy site Trinity
-.
VI
-
.
Fig Shear Test Residual Soil Containing
COHESIVE SOILS 137
sampling is required , Denison double -barrelled core samplers are normally

used . Samples obtained by this latter method are 6 - inches in diameter and
from 24 - to 30 - inches long . Shear tests in the laboratory may be made on
small specimens 1 - 3 / 8 - inches in diameter when testing core samples , or the
medium size specimens , 3 - 1 / 4 - inches in diameter , are often prepared from
large hand - cut samples . In many cases , when the materials appear to be uni
form , entire 6 - inch - diameter cores are tested in the large shear machine .
One of our projects required the testing of Carlile shale from Nebraska .
This material is an olive -black , shaley clay and is highly fractured . The ma
terial was classified as a CH clay . All of the particles were smaller than the
No . 4 sieve size with 77 to 98 per cent finer than the No . 200 sieve size , and
from 35 to 58 per cent finer than 5 microns . Standard sealed triaxial shear
tests with pore - pressure measurements were made on eight samples of this
shale by cutting 1- 3 / 8 - inch - diameter specimens from either handcut blocks
or Denison samples . A considerable variety of shear strength values were
obtained from the different field samples . This was caused by the variations
in density , degree of saturation and the amount of fracturing in the shale ma
terial . Tan o values varied from 1. 17 to 0. 49 with c values at natural moisture
varying from 37 to 7. 4 psi . The degree of saturation varied from 73 to 94 per
cent . Fig . VI - 17 shows the results of shear tests for two of the samples . In
cluded on this figure is a photograph showing the fractured nature of the shale .
Sample No . 7S - 119 had a high degree of saturation . The effective stresses
developed during the test were low and the stress circles for the various
specimens tested were close together . For Sample No . 75 - 130 , the degree of
saturation was less and the stress circles had a greater spread . Because of
the severe fracturing , it is thought that the shale acted like separate blocks in
shear . The significance of the blocky structure is apparent at low lateral
stresses but is obscured somewhat at higher stresses as the material is com
pressed into a more massive structure . This would tend to show a flatter
angle for shear tests carried out over a wide range of stresses and a steeper
angle for shear tests carried out over a narrow range of stresses . Caution ,
therefore , needs to be exercised in interpreting data of this type .
Another project involved the testing of Bearpaw shales from Montana to de
termine suitability for a powerhouse foundation and the stability of canal
slopes . This material is a shaley clay and was found to contain 50 to 60 per
cent montmorillonite . Liquid limits for the shale ranged from 83 to 191 and
plasticity indexes from 49 to 141 . Bentonite seams were visible in the ma
terial from the powerhouse foundation . Some fracturing was also apparent.
These two conditions made the testing of the shale in shear extremely difficult .
Most of the shear tests made were the unconfined compression type , although
some direct shear tests were performed to determine the shear strength
along bedding planes . These latter tests showed a tan o of 0.65 for normal
stresses ranging from 0 to 40 psi , and a c of 13 psi. Sliding factor tests to
determine the resistance to sliding of concrete on the shale were also per
formed as a direct shear test . The average sliding factor for a mortar block
on saturated Bearpaw shale was 0 . 42 . The unconfined compressive strengths
were quite variable , ranging from 53 to 402 psi , averaging 156 psi , in the
foundation area .
Tests were made on the bentonite seams from shale samples from the
canal alinement . A number of tests of both triaxial shear and unconfined com
pression were made on this shale . The tests were made on samples which
were intact and sufficiently free from fractures , so that suitable test
.
Initial
of
Specimen Data
(
.
7
)
Test Values Failure Max oz ShearValues
138
As prepared Wetted
Constant Corrected for
|
Deviator
|
Dry Moisture Degreeof Moisture Degreeof Lateral Lateral Change
Stress
%
of
Density Content Saturation Density Content SaturationPressure Pressure
Cohesion
%
(
Tan01
No Specimen
%
(
)
%
1
(
)
( %
)
pcf psi
(
pcf
)
(
75 psi psi Initial
) psi
- 1
019 110 17 910
12
67
21 108 17 87 25 107
( .3 .8 .9
108 83
31
50 116
) 5. .0 0. 0.
41117
.1- 2.- 3.- 0.-

( 00. 7.8 26.
91
16 17
100
) 591
( 5. . . .
4541 88 96
92
.1
. 4
17 13
1
130 104
41
5
73
12
63 70
1
108
16 17
85 25
16
107
18
) .8 .7 9. 2. 7. 6. 0.
- 2
. .. 669
( . . . . ..
) 3 8 5 348
00
...
67 9
17
.. .
109 85 ...542
2 22
.- .- .-
50 29 65 168
.0
95
-
7S 119
-
-
7s 130
Sealed Tests
SHEAR STRENGTH
-
ATT
11
V
PSI STRESS SHEAR

CONFERENCE
80 120 160 200

-
17
.
of
.--
-
,
Fig VI Shear Tests Carlisle Shale Cedar Bluffs Dam Nebraska

COHESIVE SOILS 139
specimens were obtained . Under the standard testing procedures for triaxial
shear , high pore pressures occurred because of the high degree of saturation
in the specimens and o values could not be obtained . Inasmuch as sufficiently
high c values were obtained , as represented by the test samples , triaxial
shear tests were not made on consolidated and drained specimens . The tri
axial tests showed the natural material to have c values varying from 57 to
100 psi and averaging 82 psi. The unconfined compression tests , which were
made on full - sized core samples 5 - to 6 - inches in diameter , showed com
pressive strengths close to an average value of 323 psi . These specimens had
degrees of saturation varying from 88 to 100 per cent and all of the samples
were rather dense , being from 110 to 118 pounds per cubic foot .
Tests have also been made on Pierre shale from South Dakota . This shale
is a massive clayey shale containing numerous fracture planes and often con
taining numerous bentonite seams. The shale itself contains a high montmoril
lonite mineral content . Difficulties also occurred in obtaining valid shear
values for this material because of the fracturing of the shale and the benton
ite seams . Tests were recently completed in which entire 6 - inch - diameter
Fig . VI - 18 . - Shear Test Specimen Showing Bentonite Seam , Bearpaw Shale

Montana
cores were tested in the large shear machine . In making these tests in which
an entire piece of core was used , it was difficult to control the uniformity of
specimens . The shear test data for this study are shown in Fig . VI - 19 . It
will be seen that some of the specimens are of disproportionately high or low
strength for the test conditions used . The tan o on the basis of these tests
might be considered to be about 0 . 20 , and the c about 13 psi. Specimen 69
contained a soft layer which deformed to a great extent during the shear tests .
While there was no apparent differences between Specimen 70 and the remain
ing specimens , it is thought that Specimen 70 was more massive and contained
less fracture planes than the others.
There have been instances where shale materials were removed from re
quired excavation and used in an earth dam . The use of required excavation
material has been described by Hilf (1957 ) . One such material , which was
tested in the laboratory , was from the Monterey shale formation of California .
This material had liquid limits between 40 and 50 and plasticity indexes be
tween 8 and 12. After excavation and compaction in a test fill , the dry - unit
weight was 72 pounds per cubic foot at a natural water content of 33 per cent.
Triaxial shear tests were performed on 3- 1/ 4 - by 9- inch specimens prepared
from material obtained by crushing , in the large compaction machine , cores
of the silt stone and removing the plus 3 / 8 - inch particles . The remaining ma
terial used in the shear test had 70 per cent of particles smaller than the
No . 4 size and 16 per cent smaller than the No. 200 size . As shown on Fig .
VI - 20 ( dashed lines ) , a tan 6 of 0 . 78 and a c value of 7. 8 psi was obtained on
material compacted under standard Proctor compactive effort and natural
moisture . The material was then recompacted under double standard com
pactive effort and retested in shear . By this process , the density was in
creased from 65 to 69 pounds per cubic foot , the material passing the No . 4
sieve size was increased to 78 per cent , and the material finer than No . 200
was increased to 21 per cent. The resulting shear values were tan Q = 0. 76
and c = 12 . 0 psi ( Fig . VI - 20 , solid lines ) . Thus , reworking the material did
not have detrimental effects on the shear strength of the soil . It was decided
to use about 700 , 000 cubic yards of this formation material within the pervious
zone 2 of Cachuma Dam .
Of allthe soils which present special problems in shear , shales probably
present most difficult ones , both from the standpoint of testing and inter
the
preting the test data . Good shear tests can be performed on small specimens
cut from massive shale samples which are free from fractures . The values
obtained , however , represent those of unfractured material and , thus , may be
overly high if applied to the shale in a fractured condition . When large speci
mens are tested in order to include fracturing , the nonuniformity of specimen
structure often makes the interpretation of the test data impossible . Further ,
with this type of condition , the specimen fracturing may affect the strength of
a specimen more than it would affect a large soil mass . Seams of weak ma
terial such as bentonite also complicate testing and data interpretation for the
same reason . Weathering and loss of strength may occur with time . Removal
of as in cuts may provide more severe shear conditions than when the
load
material is loaded as in an embankment foundation , because unloading may
open up natural fissures . Triaxial tests have been tried on large shale speci
mens , 6 inches in diameter , in which overburden axial loads were reapplied
at chamber pressures which were sufficiently large to prevent failure . Failure
was then effected by reducing the chamber pressures while the axial loads
were held constant . This type of test requires further study . Interpretation of
.
of
(
)
No Initial Specimen Dota Test Values Failure Max O17 ShearValues
Corrected
for
As
prepared Constant Wetted PorePressure
Applied Effectivel volume Daviator
Loterol
of
Dry Moisture Degreeof Dry Moisture Degree Lateral Change Stress
of
Cohesion
%
Density Content Saturation Density ContentSaturation Pressure Pressure
Tan
Specimen
psi
(
(
)
(
%
)%
1
Initial
(
)
pc
(
)
psi
(
)
pol 11 psi psi
3ICL
:1
0
98
62 100 24 02 48
L
51
12 50
3343
92 15
...
69 68 41 12 89
78
50
( .4 .17 3
1242
)% .4 .7 2. 0.
100 100 28 51 36
5 0 0 0. O.
|
) 2 3 0. .3 9.
.. ...
) .. . . .
) . .47. |7. .
)% 5. 2. 5. 5. 2.
100 47 12 75 97
56
9291 85 92
2824 39 28
21 901
|
.
.0
0
21 13
50
Sealed Tests
40
-
PSI STRESS
COHESIVE SOILS
SHEAR
10
20
30
40
50
60
70
80
90
100 120
110

141
19
of
VI
.
-.
,
-
-
Fig Shear Test Pierre Shale Pierre Canal South Dakota

142
M
(
Initial Specimen Data Test Values at Failure
ShearValues
Corrected
for
As prepared Constant Wetted
Deviator
of
of
(
Dry Moisture Degree Dry Moisture Degree Lateral Lateral Change
Stress
|
Density Content Saturation Density ContentSaturation Pressure Pressure of Cohesion
(
)
psi Tony
(
)%
(
|)
%
)
(
)
(
Initial
(
(
)%
psi
No Specimen
pct pct psi psi
1AM
.
3
66 65 32 58 12 73 68
( 11
.
63 31 55 22 97
( 25
70
7
56 50 44 90 168
31
65
.0
.
8
5057
7
2 9 67
0 64 31 56 100 89 66 313 78
11
9
.
31 62 34
12
68 78
10
0
SHEAR
)(2
68 31 62 50 43 29 184
10
0
) 5. 0. 0. 0. 5. 0. .0
.0
.
)%( .5 .4 .8 .6 8. 9. 0.
) 2. .1 2. .7 5. .7 9.
) 4. 7. .0 .7 .7 .6 .
. . . . 7. 8. 7.
. . . . ,7 . 7.
168 3131 62 100 90 98 329 76 112
200
-
Standard Proctor
Sealed Tests Compactive Effort
—
( )(
. 4 . . 2 .6 .8 -1) 2
Double Compactive Effort
STRENGTH
PSI STRESS
SHEAR
CONFERENCE
40
80
120 160 200 240280 320 360 400 440 480
-

,
Dam California
20
of
Shale Cachuma
VI
-
-.
.
Fig Shear Tests Compacted Monterey
COHESIVE SOILS 143
failure conditions has been impossible as the material deformed axially to

great amounts without producing either maximum deviator stress or maximum
principal stress ratio conditions .
Determining and interpreting the shear strength of massive shale for
mations must be considered difficult . At the present time , laboratory values
are not precise and definitive and should be used only as indicators along with
judgment gained from experience in designing and operating structures in each
type of shale formation material . Considerable further research on the true
strength of shales , effects of time on strength , and methods of analyzing
natural shales for structures is needed .
VII. IN - PLACE VANE - TYPE SHEAR TEST
A. General .
The vane - type shear test is a useful method for determining the in - place
shearing resistance of soil foundations consisting of soft , saturated clays .
This test gives a shearing strength value of the soil as it exists in place with
natural overburden pressures acting . The value obtained is based on the as
sumption that the shearing occurs on a cylindrical surface without changing
volume or soil structure . The geometry of the sheared surface and its orien
tation with respect to the normal pressures makes the theoretical interpre
tation of shear strength complex . However , the instrument has been used to
considerable advantage to determine the different degrees of in - place strength
occurring in soft , saturated clay and silty clay foundations . These strength
values have been found to be comparable to those interpreted from results of
other testing methods .
B. Equipment .
The vane shear test apparatus used by the Bureau of Reclamation was de
veloped in 1954 (Gibbs 1956 ) . A review of existing literature ( Cadling and
Odenstad , 1950 ; Skempton , 1948 ; Bennett and Mecham , 1953 ) was made and
basic requirements were established to provide the most desirable instrument.
These were as follows : ( 1) Capable of operating at controlled speed of ro
tation and permitting frequent readings for plotting a complete curve of test
results , ( 2) sufficiently sturdy and simple in its operation for a variety of field
applications by field personnel , ( 3) adaptable to a variety of field surface con
ditions , (4) adaptable to standard drilling equipment , and (5 ) the overall instru
ment should be an accurate measuring device which permits the elimination
of frictional effects which would otherwise be registered as shearing strength .
The Bureau of Reclamation instrument is shown in Fig . VII - 1. The im
portant features of this apparatus are :
.
1 The instrument applies a balanced -moment torsional force on a free
standing rod and therefore does not require a thrust bearing that would
consume variable amounts of force in friction as applied forces change .
2. This torsional force is measured by very small strain of a resilient
ring which does not involve friction of moving parts , such as pulleys ,
cables , wheels , etc . , and does not interfere with the rotational obser
vations and speed control.
3. The instrument can operate at variable depths depending upon the shaft
extensions used , which consist of standard drill rod and an upper square
rod passing through the resilient member .
4. The instrument is a compact portable unit simply supported by a piece
of8 - inch pipe embedded a short distance in the surface of the ground .
5. The instrument is fitted with suitable gearing to permit the operator to
control the speed of rotation .
By having the resilient measuring device directly connected to the free
standing rod , the effects of friction in the gears are of no concern and will not
affect the measurements , thus providing a sturdy and simple instrument
Torque Ring ter

Strain Gage for
Reading Torque
5° Graduations
Rotation Indicator
8 "-inch Casing with Side Fins
for Anchoring Torque Assembly
Torque Rod
A - rod for Applying Torque to

Vane . Made up in 5 - ft. Lengths
BX -Casing for Housing Torque
Rod and A -Rod
Vane Rod
BX - Casing - Point Containing
Bearing and Water Seals for Vent
Vane Rod
Opened Rubber "O" Ring Seal
while
Greasing Grease Chamber
Vane Varying Sizes
for
Fitting Greasing
by by
inch Diam
in in in
4 3 2
8 6 4
- - -
. .
. . .
inch Diam by Rubber Ring Seal

0
"
"
inch Diam
.
O
3
?
!
OF
SCALE OF INCHES SCALE INCHES
Reclamation Vane Test Apparatus

of
Fig VII The Bureau

-1
–
.
.
COHESIVE SOILS 145
suitable for field applications . The device is designed for measuring torsion
al forces from The torque applicator shown in Fig .
0 to 200 - foot pounds .
VII - 1, contains a sturdy ring 5 inches outside diameter , 1- inch high and 0.57 -
inch thick , which is attached to a 10 - inch gear driven by the other gears . The
ring has a section cut from it to permit deformation as torque is applied . The
deformation is indicated by a dial gage. A calibration chart is provided to
convert dial readings to foot - pounds of torque . The 10 - inch gear wheel hold
ing the ring is marked in degrees and has a pointer for registering rotation .
A piece of 8 - inch pipe , 2 feet in length with side fins for embed
standard
ding the surface of the ground , is used to anchor the torque appli -
1 foot into
cator . Pipe extensions can be used to vary the level of the instrument for
minor depth adjustments or to raise the instrument above water level when
necessary .
The vane rod stem consists of ( 1) an upper square section passing through
the torque applicator , (2) a center section made up of Standard A drill rod in
5 - foot lengths , ( 3) a lower , round stainless steel section on which the vanes
are attached , and ( 4) Standard BX casing pipe in 5 - foot lengths which encloses
the vane rod and contains a watertight bearing at the lower end . This bearing
is fitted with 0 - ring seals and has a grease chamber . The flush coupling con
nections on the A - rod must be adequately tightened to prevent slipping during
torque application . The 5 - foot lengths of BX - pipe and A - rod permit conveni
ent testing at intervals of 5 - foot depth . Special short pieces can be used to
vary the testing depth .
The vanes which are attached to the end of the vane rod are of three sizes ,
4 - inch dia meter by 8 - inches high , 3 - by 6 - inches , and 2 - by 4 - inches . The
vane rod passes through the lower bearing and a short length of the rod is ex
posed to soil when the vane is extended into undisturbed soil . The friction of
the bearing and rod must be taken into account , although it may be small and
relatively constant for the various degrees of force because there is no side
thrust in the lower bearing . A vane rigidly attached to the rod requires sepa
rate determinations of friction using a blank stem or assumed friction cor
rections in order to interpret shear test results . In order to determine this
friction value reliably , a modified vane was developed in which a definite
amount of rotation is permitted during which a friction correction is measured
for each test prior to applying force to the vane . A drawing of this modified
vane is shown in Fig . VII - 2a . It consists of a swivel stem containing a re
cessed portion in which contact lugs on the coupling are able to move through
80° of rotation before contact is made and force applied to the vanes . This
permits a series of readings to be made on the torque instrument to evaluate
this friction .
A drill rig and accessory equipment is generally used for placing the vane
stem at the desired testing position and pulling it after the test . In stiff or
firm soils predrilling is often necessary . In 1959, a feature was added to the
original equipment to permit using the vane rod stem as a drilling tool during
placement . This apparatus is shown in Fig . VII - 2b . A hardened face cutting
bit is attached to the lower bearing . Ports are made through the bearing so
that drilling fluid can be circulated through the BX pipe and out the bottom .
During drilling the vane is held up and inside the cutting bit by a top coupling
device which also serves as a connection to standard fittings for the drilling
fluid hose and the hoisting plug . The entire vane stem is then used as a rotary
drilling tool to advance the hole to the desired placement depth .
- - Boll support .
- - - Swivel stem .
D - - - - Vone rod .
Washer and nut for , - Vane coupling with

holding vone to swivel and contact
coupling . com lugs .
80° Free move -

----- - - - - 0 - ring water
--
ment before seal
contact is
mode with -of
vones -
( 0 ) MODIFIED VANE FOR FRICTION DETERMINATION
Fig . VII - 2. - Accessory Fe:

COHESIVE SOILS 147
N - rod coupling for

drilling - fluid line
ond clamp for holding
vone rod up during
drilling .
PX -0 -16235
Vone stem and bit ,

showing mud circulation
Vone retracted
inside rotory
cutting bit
PX - D- 14465
After drilling is
stopped , vone is
pressed into
undisturbed soil
P X- D -14466
Drilling in operation
(b ) ROTARY DRILLING TOOL FOR INSTALLING VANE STEM

the
Vane Test Apparatus

The instrument is calibrated for foot - pounds of torque by loading a lever

system shown in Fig . VII - 3. Several dial gage readings are obtained at inter
vals of 25 - foot - pounds of torque up to the maximum of 200 - foot - pounds . The
final calibration curve is the average result of several sets of readings with
the torque ring at different angular positions . Each instrument requires a
separate calibration curve. A typical example of a calibration curve is shown
in Fig . VII - 4. The calibration curve is of a type which permits readings of
small and large values of force of equal accuracy .
The equations for computation are as follows :
Total torque = T = T1 + 2T2
where T1 = torque resistance on the vertical cylindrical surface

T2 = torque resistance on the horizontal top and bottom , assuming
constant unit shear resistance
Consu
T1 = ( 2nrh )rs = 27 shr2

T, = 89 sr , when the height, h, is twice the diameter
xdx
dT2 = s(26
)x
x2dx
275
205
T2
=
*
/
=
sr3
87
1513
$
$
=
=
5 r
T
3T
,
refore
Therefore
=
28TE3
straight line
of
The relation shear strength torque can be shown by

to
-
graphs for the various vane sizes This shown Fig VII which also
is
in
-5
.
indicates the maximum shearing strength that can be measured under the 200
foot pounds limitation of the instrument for the various vane sizes
-
Test Procedure
.
C
.
briefly illustrated Fig

of
The installation the instrument VII The

is
in
6
.
.
-
sing with side fins about foot the soil Soil

ca
installed
is
inch anchor into

1
8
.
-
inside the casing removed

is
on
The vane rod stem assembled contain the vane the lower end the
is
to
,
-
BX
lower watertight bearing the round rod and the first sections pipe
on
of
,
and rod All vane rod and rod couplings must be tightened with forces
A
A
prevent slip
-
.
-
-
excess 200 foot pounds the capacity the instrument

to
of
of
in
,
-
ping during the test then installed

is
of
The stem the center the inch

in
8
-
.
casing soft soil the vane stem can be pressed into the ground without pre
In
.
boring Where overlying strata are relatively firm predrilled hole made
is
a
,
.
or the special drilling tool shown Fig VII

2b
used When the vane stem

is
in
.
-
with the vanes the up position located at the desired depth the upper
is
in
square rod the top and the vane rod ready for pressing into
is
attached
to
is
undisturbed soil
.
At this time the modified vane preparation for the free rotational
is
set
in
,
easily done applying

by
movement and friction determination This

is
a
.
wrench the square rod and turning the rod backward until slight touching
to
the vane lugs then pressed downward place the

of
felt The vane rod

to
is
is
.
undisturbed soil and location planned for testing

at
vane The minimum

in
is a
pressing distance that

18
used inches and the normal recommended

is
COHESIVE SOILS 149
PX - D- 20391
E - 1791 - 2
Fig . VII - 3. - Apparatus for Calibrating Dial Readings Against Foot - pounds
of Torque
.
.
TORQUE FT LO
-
CONVERSIONCURVE STRAIN GAGE
-
TO
IN
MOVEMENT TORQUE FOOT POUNDS

be
Each apparatusmust separatelycalibrated

(
.
an
This only example calibration

is
)
00
400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200 3400
D
STRAIN GAGEMOVEMENT FROMINITIAL READING 0001 INCHES

-
.
for
Fig VII Typical Calibration Curve Vane Shear Test Apparatus

-4
a
-
.
.
"2
.
VANE SIA
- "4
150
BV HEIGHT
.
-
--
30
LILIT OF TESTIMO
-:-
-
APPARATU
8612
.-
-
-
25
PSI
-
X
-0-
--3
20393 20396
STRESS
a
)(b
()
-8
Install inch anchor Install upper square vane
30
casing and pressvane rod and press vanes
to
stem position inches into undisturbed soil
SHEAR
SHEAR STRENGTH
--
.
"3
VANE DIAM
- 6"
BY HEIGHT
10
CONFERENCE
--4
-
.
VANE DIAM
20397
"4 "
BYE HCIGHT 20398
c
()
)(d
to
-0 is
50 Install torque assembly Test ready begin

100
200
200 250
250
150
--
.-
.
TORQUE FT LBS
of
of
-.6
.
-
Fig VII Method Installation the Vane
of
to
.
Testing
-.5-
Fig VII Conversion Torque Shear Stress Apparatus
COHESIVE SOILS 151
distance is 30 inches . The torque applicator assembly is then placed over the
square rod and rested on the BX casing pipe , and the BX clamp is tightened .
The square vane rod must be free in the square hole for the zero reading
which can be checked by rattling the rod . Finally , the base plate of the torque
applicator is securely clamped to the 8 - inch anchor casing so that it is suf
ficiently tight to resist the maximum torsional force of the instrument . The
test is then ready to begin .
The test is made in three steps to obtain the test results shown in Fig .
VI - 7 . Initial readings
of the dial indicator gage and the rotation protractor
are made after checking the rod for looseness . During all parts of the test ,
the crank is turned at a uniform rate of 12 turns per minute which is conveni -
ently paced at 1 revolution for each 5 seconds on a stopwatch . This corre
sponds to a rotational movement of the vane rod of 0. 1° per second . The read
ings are recorded at intervals of 5° of rotation . The first part of the test is
the friction correction determination . Six observations are made during the
first 30° rotation while using the standard rate . Following this , a more rapid
turning without observation can be used to complete the remaining portion of
the free movement. The operator must use care in observing , by the dial
gage , when the contact with the vane is reached so that the standard controlled
speed is again used during the application of force to the vanes . The second
part of the test is the determination of the undisturbed strength . While turning
the crank at the standard rate , strain gage readings are made at rotation
intervals of 5º and at the angle when the maximum gage reading is observed .
The test is then continued for 5 additional readings at 5° intervals . The third
part of the test is the determination of the remolded strength after failure .
The vane is rotated without observations through an angle of 90° so as to com
pletely shear the cylindrical surface . Testing at the controlled rate is again
resumed . Five additional readings are made at 5° intervals . Upon com
pletion , the crank handle is turned back until the square rod is free and a no
load reading of the gage is made to check the initial zero .
Fig . VII - 7 shows these test results in terms of dial gage readings and de
grees of rotation . Since the calibration curve is nearly linear , it is unneces
sary to evaluate foot - pounds for each observation made . Therefore , the foot
pound evaluation of the torque is only determined for the points of interest
which are those for undisturbed strength and remolded strength . The dial
gage reading determined during the friction observation is subtracted from the
dial gage readings for these two sets of observations and the resultant differ
ences are used to determine the foot - pounds . Using the equation shown previ
ously or the chart, the computation of shear strength in pounds - per - square
inch is obtained . The results of shear strength are then usually plotted
against depth for a particular test hole .
D. Examples of Test Results .
An example of an extensive series of vane tests are those made for Willard
Dam in Utah . This structure is a low dam extending for a considerable
distance on soft lake sediments of Great Salt Lake . Fig . VII - 8 is a plan view
of the dam which will be used to impound water for irrigation distribution near
Ogden , Utah . On the axis of the dam , a series of vane test holes were es
tablished at intervals of 500 - to 1, 000 - feet . These vane test holes were gener
ally made to depths of 50 feet , and three of the holes , shown by points on the
axis of the dam , were made to depths of 120 feet. An example of the vane test
results obtained from one of these deep drill holes is shown inset in this
figure . These results show a gradual increase in the undisturbed shear
VANE SHEAR TEST PLOT SHEET
HOLE NO . DH-202VT DEPTH _ 60 _ ft. VANE 3 in .
FRICTION UNDISTURBED REMOLDED
_ 1373 _ - 3164_ 2225 _

1216 _ 1373 1373
- 157 _ diff . _1791 _ diff. - 852 _ diff.
_ 16 _ _ ft - 10 _ 156 _ ft - 16 _ 77 _ ft - 10
| | 18. 9 _ psi .
9 3_ psi ,
3500
3164
3000
READING
2685
2500
GAGE
22 25
2000
H
1500
13731
1216
1000
80 105 145 155 180 290 310
1
1
OF
DEGREES ROTATION
Fig VII Vane Test Results

7
-
-
.
.
COHESIVE SOILS 153
resistance of the clay sediments from 5 psi near

the surface to about 10 psi
at a depth of 100 feet except for of firm materials of relatively
a few strata
high shear resistance . The test results show the range of strength of the soft
clay sediments and the sensitivity of these clays are registered by comparing
the remolded shear resistance to the undisturbed shear resistance . The
second inset graph in the Fig . VII - 8 is given to show detailed test results at
the one depth of 70 feet .
Another example of a series of vane shear tests was that made at Lovewell
Dam in Kansas . In this case , construction of the dam was in progress . A
clay layer in the foundation of the right abutment , which was at the location of
a meandering creek , was found to be very soft and saturated . It was decided
that the shear resistance of this clay layer should be further investigated
HOLE 201
WILLARD BAY Willard uton ,
- EAST DRAINA
DREM - WILLARD
DAM <- - LEVEE
HOLE 202 - -
seale
SCALE OF MILES
OF MILES
k - - SOUTHDRAIN
HOLE 203 - - -
SOIL DESCRIPTION AND VANE TEST RESULTS - HOLE 2021
0 - 4 Leon clay
6-15 silt
6 32 Leon clay
FEET
32- 35' Fine sand

35- 40' Silt -
- - EXAMPLE TEST AT 70 FEET
DEPTH
40- 45' clay

( sity , Orgonic) HOLE No. 202 DEPTH 70feet VANE3in.
Friction Undisturbed Remolded
1285 2020 1542
45- 70' Fot clay - Undisturbed 1206 1285 1285
- 120' Leon clay 100 79 diff . 735 diff . 257 diff .
eft . Ib. 67 ft . Ib. 25 ft. ib.
- 15
10
1205 20
SHEAR STRESS PSI
Remolded
-
Remolded
-
GAGE READING
Undisturbed
Friction
-<
-
lost ad
at
da
k
lo
a
ten mit
-
w
>
1000 1
150 175 200 225 250 350 375
OF
ANGLE ROTATION
Fig VII Example Vane Test Studies Willard

at
of
Dam Utah
,
8
-
-
.
.
before embankment placing continued , since
its
softness could affect the rate
of
embankment placing and the design dimensions Therefore
of
the dam
,
a
.
this soft clay strata These tests provided
of
series vane tests were made
in
.
quick information shear resistance permitting necessary adjustments
on
to
,
design and construction procedure with minimum delay
be
made con
. in
in
,
a
struction
Theoretical Discussions Vane Shear Tests
of
E
.
–
is .
the average
on
by
The shear strength obtained the vane test dependent
effective normal stresses acting the surfaces sheared by the vane
no on
In
a
.
natural deposit excess hydrostatic pressure this effective
is
which there
in
,
stress related the overburden pressures effective stresses are in
is
to
If
.
creased the vane shear strength should also increase according some shear
a to
resistance relationship normal stress This illustrated by series
is
of
to
.
vane tests which were performed the foundation of test embankment at
in
a
Willard Dam shown Fig VII
in
,
9
-
.
Before the embankment was constructed . tests were performed Drill
at
,
the proposed at Drill
of
Hole 126 below the center embankment and Hole 125
,
The results are shown by Lines the graphs
on
below low level berm
A
a
.
After completion the embankment the foundation gradually consolidated
of
,
thus causing effective pressures the foundation increase as pore
to
in
pressure dissipated Vane tests which were again performed adjacent
to
.
these locations later time intervals showed increases shear strength

at
in
.
At first as shown by Lines the graphs the significant increase
on
in
,
,
B
strength was primarily greater increase under the

at
the upper levels with

,
Tests made still later as

of
center the embankment than under the berm
,
.
shown by Line on the graph under the center of the embankment showed
C
at
considerably greater increase strength not only the upper level but also
in
,
at
lower depths
.
compare the results laboratory triaxial shear tests

of
vane tests
To
to
a
,
of
series undisturbed samples were obtained from separate sampling hole

a
at depths corresponding the initial vane tests made

at
Hole DH 126 Fig

to
in
.
-
VII These were tested for triaxial shear as described Part for undis
in
V
9
.
-
turbed saturated soil

.
The first example laboratory test results Fig VII

of
10
shown for
is
in
-
at
Sample No 162 167 which was obtained from Hole DH 126 elevation 4174
,
.
adjacent shear strength psi The

of
the vane test which showed

in
to
4
to .5
a
place shear resistance from the laboratory tests appeared be close that
25 in to
Fig VII
by
Specimen Although initial

10
shown there were variations

in
,
5
, .
.
-
the specimens average

an
these void ratios near The

of
1 of
void ratio
is
to a 1
.
.
void ratio for Specimen at failure was 169 Specimen showed slight
5 5
1
.
expansion and Specimen slight consolidation which was similar that

a
usually observed for the reloading portion of consolidation curve Speci

a
mens tested higher pressures consolidated appreciably greater amount

at
an
The estimated shear resistance obtained from the Mohr plot for each speci
,
s
'
men at the point tangency the envelope plotted on the log graph
of
is
to
p
e
,
for convenient comparison psi shear re

of
The determination
to
5
5
.5
.
sistance for Specimen compared reasonably well with the results the
of
5
vane test which was psi

4
.5
The second example laboratory test results Fig VII

of
11
for
is
shown
in
Sample No 162 168 which was obtained from Hole DH 126

at
elevation 4169
,
,
to .
of -
adjacent vane test which showed shear strength psi This sample
4
.2
a
.
AXIS OF TEST EMBANKMENT
SCALE OF FEET
EL
CREST 4240
.
30
LOW
.
%
2
- --
EL 4234
.
150
.
.- 6
'
EL
4211
ASSUMEDORIGINALGROUND
--
-
.
SURFACEEL 4205
VT
--
-
-
DH 125
4200
4200
-
-
-
DH 126 VT
4190 4190
EXPLANATION
,8
to
. March April 1955
30
Dataobtained .
FT
for
FT
®
- line -
June
30
1955
Constructionbegun
COHESIVE SOILS
, ,
-- --
Sept24 1955
4170 Constructionfinished
ELEVATION
5
,
10
10
Oct Oct
ELEVATION
1955
.-
Datoobtained
for
©
line
,
March 1959
13
--
Dataobtained
for
©
4160 line
4160
*
FailurewoSnotreached
of
of
becouse thelimits
4150
themachine .
41505
10
15
O
10
-
SHEAR RESISTANCE PSI
-
SHEAR RESISTANCE PSI
.
-.9-
in
Fig Vane
of
VII
a
Test
at
Studies the Foundation

155
Test Embankment Willard Dam

was of relatively low initial density with most of the specimens averaging
about 1. 6 initial void ratio , except Specimens 1 and 2 which had void ratios of
about 1. 45 . All shear specimens consolidated varying amounts , except Speci
men 6 which expanded slightly under low effective lateral pressure . The in
place shear resistance estimated on the e- log p graph , in Fig . VII - 11 , appears
O
PSI
STRESS
2
SHEAR
40
5
30
0
20
-
o
11111 TIITTITUTI
- -0
-- Wal Spec
.
-
Pan
YIN
20
fo
1
on Spec
.
da
5
.
Spec
. 4
RATIO
Spec
A
A
3
VOID
Spec
A
.2
uullul III
-
-
Spec
03
6
+
,
.
C
-
L
100 1000
10
LOAD PSI
-
Laboratory Triaxial Shear Test Laboratory Sample

16
Fig
10
VII 167
Z
-
-
-
--
.
.
Hole DH 126 Elevation 4174

-
--
COHESIVE SOILS 157
-
PSI
STRESS
SMEAR
20
30 40 50
-
la
Spec
Spec
or
AA
.
--
6
To
ca
--
-
NL
Od
AbiSpec
1
.
lolo Spec
I
3
. .
asl Spec
RATIO
2
VOID
Specimens 19 Spec
.8
ond
5
4
7
,
,
(
were unsuitable and

03
,
--
-
not completed
0
--
-
)
8
0
.
ktoo
Spec www
,
9
.
10
100 1000
LOAD PSI
-
11
Fig VII Laboratory Triaxial Shear Test Laboratory Sample 162 168
.
-
.

-
--
-
PSI
STRESS
SHEAR
1
5
-
20 30 40
-
1
2
.
NWA
RATIO
A4
Spec
.!
la
Spec
VOID
2
.
Til Spec
To
Fines
.3
,
03
0
-
-
ūruto
104
ab
Spec
LITT
. 5
Spec
4
.
100 1000
-
LOAD PSI
-
Fig Laboratory Triaxial Shear Test Laboratory Sample

12
VII 162 166

–
-
--
--
-
.

-
-
COHESIVE SOILS
tobe between that shown by Specimens 6 and 1. Specimen 1 consolidated :
small amount and is considered to show a shear strength near to that in - -
place, but unfortunately , it was one of the specimens initially denser than the
average. Considering that the in - place strength is probably near to that
shownby Specimen 1 but between that shown by Specimens 6 and 1, the esti -
mated in - place strengthwas 4 . 0 psi . This value compares reasonably well to
the
psi
vane test results of of
4
2
a .
.
The third example laboratory test shown Fig VII one involv
12
in
is
is ,
,
, .
-
ing
higher stress range The sample No 162 166 which was obtained
a
-
from Hole DH 126 at elevation 4179 adjacent vane test which showed
to
,
a
-
psi Although the characteristics

of
of
shear strength this sample were
9
0
,
.
rather irregular and good plot of

shear resistance was difficult de
to
a
a
termine the higher strength shown the vane test can be seen by
the labo
in
,
ratory test results by the larger stress circles The average initial void
.
Specimen consolidated only slightly
at
ratio was near void ratio
to
.1
1
a
.
log plot Fig VII Specimen
12
of
failure From the

045 shows
in
p
,
1
1
e
a
.
.
-
-
shear strength psi and Specimen
of
about which consolidated somewhat

of ,
8
.5
shear strength psi These results are reasonably close

10
more shows
a
comparisons the vane test result which was psi The envelope the
to
of
9
.0
shear stress circles difficult interpret because the irregularity .

is
of
of
to
Specimen Also the sample was insufficient obtain test at lower

to
5
a
.
stress condition than that shown by Specimen

of 1
.
The foregoing examples represent many comparisons
of
few vane tests
a
laboratory triaxial shear tests undisturbed samples from one project

on
to
.
very soft clay soil which most applicable
All
on
were made vane testing

is
to
,
butdifficult test the laboratory believed that stiffer clay deposit

is
to
in
It
a
.
which would show better laboratory results would be better for research com
These examples showed obvious relationship larger stress
an
parisons
of
.
sample where the vane test value was psi when compared
on
the
to
circles
9
a
smaller stress circles for the two samples where the vane test values were
and psi With this fact and the theories demonstrated mind less de
in
4
4
.5
,
.
tailed comparisons have been made for tests from other projects Judgment
.
select the shear resistance value for comparison

of
that the
to
to
was needed
vanetest but the results were significant particularly for broad variations
,
place strength
in
in
-
important viewpoints should be kept

Several mind when making compari
in
laboratory tests The shear resistance

of
vane test results those

to
of
sons
.
some average effective stress which diffi

of
the vane test

is
the effect
is
of
When sample
to
cult determine extracted foundation these

is
from
a
,
.
stresses are released and the sample expands An undrained unconfined

at ,
.
of
compression test which does not restrict further expansion the time
failure logically can show shear resistance less than that place On the
in
a
.
-
triaxial shear specimen which tested at effective pressure

an
other hand
is
,
a
high enough cause considerable shear resistance

to
consolidation can show

a
considerably greater than that place Therefore logical that place

is
in
in
by it
,
.
-
shear resistance would be most nearly represented specimen which

in
a
the
pressures are reapplied but are less than that which overcomes the in
,
herent strength and less than that which causes appreciable consolidation
.
any judgment given analysis for the

an
Allowance must be made

to
in
such
disturbance resulting from sampling unloading and
of
well known effects

re
,
,
-
loading characteristics and some drainage which cannot be avoided during

,
specimen preparation and testing Also should be understood that the

it
,
.
laminations of a deposit would certainly affect the strength for different orien
tations of either the vane or triaxial tests , but these effects would be most
critical for tightly packed deposits rather than the soft , loosely packed de
posits in which vane testing is most applicable .
APPENDIX . -- BIBLIOGRAPHY
1. Bennett , G . B . and Mecham , J . G . , Use of the Vane Borer on Foundation

Investigation of Fill , Proceedings 32nd Annual Meeting of the Highway
Research Board .
2. Bureau of Reclamation , Earth Manual (1958 ) .
3. Cadling , L . and Odenstad , Sten , The Vane Borer , Proceedings No . 2 ,

Royal Swedish Geotechnical Institute , Stockholm , 1950 .
4. Casagrande , A . , The Structure of Clay and Its Importance in Foundation

Engineering , Journal of the Boston Society of Civil Engineering , April,
1932 . (Reprinted in Contributions to Soil Mechanics , 1925 to 1940 , Boston
Society of Civil Engineers ) .
Casagrande , A ., The Structure of Clay and Its Importance in Foundation

Engineering , Journal of the Boston Society of Civil Engineers , April 1932 ,
Reprinted in Contributions to Soil Mechanics , 1925 to 1940 , Boston Socie
ty of Civil Engineers .
Casagrande , A . , Seepage Through Dams , Journal of the New England

Water Works Association , June 1937 , Vol. LI, No . 2, Reprinted in Contri
butions to Soil Mechanics , 1925 to 1940 , Boston Society of Civil Engi
neers .
7. Cedergren , H. R. , Utility of the Flow -net in Stability Analysis , Proceed
ings of the Purdue Conference on Soil Mechanics , September 2 - 6 , 1940 ,
Purdue University .
8. Dorsey , N. Ernest , ( Comp . ) , Properties of Ordinary Water Substance
(New York : Reinhold Publishing Corporation , 1940), p. 180 .
9. Ellis , Willard and Holtz , W . G ., A Method for Adjusting Strain Rates to

Obtain Pore - Pressure Measurements in Triaxial Shear Tests , ASTM
Symposium on Time Rates of Loading in Soil Testing , 1959 .
10 . Gibbs , H . J. , An Apparatus and Method of Vane Shear Testing of Soils ,

Symposium on Vane Shear Testing of Soils , Special Technical Publication
No . 193 , ASTM , 1956 .
11 . Gibbs , H . J. and Hilf , J . W . , Triaxial Shear Tests Holding Effective

Lateral Stress Constant , Proceedings of the Fourth International Con
ference on Soil Mechanics and Foundation Engineering , London , 1957 .
12 . Gilbert, O. H ., The Influence of Negative Pore Water Pressures on the

Strength of Compacted Clays , S. M . Thesis , Massachusetts Institute of
Technology , 1959 .
13. Glover , R . E . and Cornwell , F . E . , Stability of Granular Materials , Paper

No . 2172 , Transactions , ASCE , Vol . 108 , 1943 , p . 47 .
COHESIVE SOILS 161
14. Gould , James P ., Construction Pore Pressures Observed in Rolled Earth

Dams, Technical Memorandum 650 , U . S. Department of the Interior ,
Bureau of Reclamation , Denver , Colorado , April 1959 .
15. Hamilton , L . W . , The Effects of Internal Hydrostatic Pressure on the

Shearing Strength of Soils , Proceedings ASTM , Vol . 39 , p . 1100 , 1939 .
16. Henkel, D . J. and Gilbert , G . D . , The Effect of the Rubber Membrane on

Measured Triaxial Compression Strength of Clay Samples , Geo
the
technique, p . 20 , 1952 .
17. High Dam , Proceedings of First International Congress , Stockholm ,
Sweden , 1933 .
18. Hilf , J. W . , Construction Pore Pressures and their Effect on the Stability
of Rolled Earth Dams, M . S. Thesis , University of Colorado , Boulder ,
Colorado , 1947 .
19. Hilf , J . W . , Estimating Construction Pore Pressure in Rolled Earth
Dams , Proceedings of the Second International Conference on Soil Me
chanics and Foundation Engineering , Rotterdam , 1948 , Vol.
III
234
.p
,
.
An Investigation Pore Water Pressure Compacted Co
20
Hilf
, of
W
in
,
.,
.
J.
hesive Soils Doctoral Thesis University Colorado Department

of
S
),
(
.
.
Interior Bureau Reclamation Technical Memorandum 654
of
of
the
,
,
Denver Colorado October 1956
,
,
.
of
21
Hilf Use Materials from Structural Excavations Journal

W
of
,
.,
,
.J
.
Soil Mechanics and Foundations Division Vol No SM2 Paper No

83
,
,
.
.
1207 1957
,
Expansive Clays Properties Quarterly

22
of
Holtz Problems
W
and
G
,
,
.
.
.
No
54
89
the Colorado School Mines Vol

of
1959
p
,
,
4
.
Consolidation and Related Properties

of
23
Holtz and Gibbs

W
,
,
H
.
.
.
.
J.
126
No
Loessial Soils Symposium

on
Consolidation ASTM STP 1951

,
,
.
Engineering Properties Expansive

24
and Gibbs
. of
Holtz
W
, ,
,
H
.
.
.
.
J.
Clays ASCE Trans Paper 2814 Vol 121 641 1956

.p
,
,
.
Shear Strength Pervious Gravelly Soils

25
and Gibbs
of
Holtz
W
,
. ,
,
, H
.
.
.
.
J.
No
Proceedings Vol Paper

18
the ASCE
of
867 1956
,
,
of .
Lowe John and Karafiath Leslie Stability Earth Dams upon Draw
26
, ,
,
.
down Proceedings First Pan American Conference

on
Soil Mechanics
-
and Foundation Engineering Mexico City Mexico September 1959

,
Fundamental Principles Soil Compaction Engineering

27
Proctor
of
,
.,
,
R
, R
.
Vol III August through September

31
28
News Record 1933

,
,
.
.
-
Puri Their Physics and Chemistry

28
Soils York Reinhold

—
New
,
.,
N
A
:
.
Publishing Corporation 1949 393

p
,
,
.
.
)
Capillary Potential
of
29
Richards Lorenzo The Usefulness Soil

to
,
.,
A
.
Moisture and Plant Investigators Journal Agricultural Research July

C ,
December 1928 Washington Government Printing Office

,
.,
,
D
S
.
.
.
pp
1929 719 742

,
Skempton the Alluvial Plain

30
Vane Tests River Forth

of
W
the
in
,
.,
A
.
near Grangemouth Geotechnique Vol No 111 London

.p
,
1,
,
2
.
.
31 . Skempton , A. W . , The Pore - Pressure Coefficients A and B , Geotechnique ,
December 1954 .
32 . Taylor , D . W ., Tenth Progress Report to the U . S. Engineer Department ,
Massachusetts Institute of Technology , Soil Mechanics Laboratory , May
1944 .
33 . Terzaghi , Chas . , Principles of Soil Mechanics , Engineering News

Record , Vol. 95 , November 5 through December 31 , 1925 .
34 . Terzaghi , Karl, Simple Tests Determine Hydrostatic Uplift , Engineering
News - Record , June 18 , 1936 , pp . 872 - 875 .
35 . Walker , Fred C . , The Use of Residual Soils in Earth and Rockfill Dams ,
First Pan - American Conference on Soil Mechanics and Foundation Engi
neering , Mexico City , Mexico , 1959 .
Walker , Fred C . and Daehn , W . W ., Ten Years of Pore - Pressure

.36
Measurements , Proceedings of the Second International Conference on

Soil Mechanics and Foundation Engineering, Vol .
III
245 Rotterdam
p
,
,
.
1948
.
37 Waterways Experiment Station Soil Mechanics Facts Finding Survey

,
.
:
Progress Report Triaxial Shear Research and Pressure Distribution
, ,
Soils April 1947

on
Studies
.
PRESTRESS EFFECTS ON THE STRENGTH OF CLAYS
By Reginald A . Barron , 1 F . ASCE
SYNOPSIS
The mathematical relationships that exist between the shear strength enve
lopes and curves of principal effective stress and prestress ratio are de
veloped . Prestress data of an organic silty clay are correlated by means of
the developed mathematical relationships to the strength curves normally
used .
INTRODUCTION
The influence of prestressing by means of overconsolidation upon the shear

strength of clays has been known since the late twenties . In 1953 Casagrande
and Wilson2 presented data which indicated that prestressing also occurs in
the R shear test ( consolidated -undrained ) on normally consolidated clay, and
that prestressing results from a build - up of pore pressure during the test
which reduces the effective normal stresses . Hirschfeld ) , 4 investigated the
effects of prestress and presented detailed data in his doctoral thesis in 1958 .
Hirschfeld ' s studies revealed that prestressing occurred in all R tests .
Information on shear strengths is normally presented in the form oi
strength - stress envelopes . Data on prestressing may be conveniently pre
sented in a plot of principal effective stress ratio , 01 / 03 , versus the prestress
ratio , Rp , as shown in Fig . 1. The mathematical relationships that exist be
tween these two methods of presenting strength data are developed for the pur
pose of permitting ready correlation of the data . The prestress ratio , Rp , was
defined by Hirschfeld as " the ratio , Olon , of the maximum effective normal
stress which has existed on a plane in the past to the effective normal stress
on that plane at a later moment . ” While the prestress ratio may be determined
for any plane , it is generally associated with the failure plane . The prestress
ratio and the principal effective stress ratio are also those which occur at a
selected percent strain.
1. Engr ., Soils Branch , Civ . Works , Office , Chf. of Engrs . , Dept . of the
Army, Washington , D . C .
2. Casagrande , A . , and Wilson , S . D . , “ Prestress induced in consolidated
quick triaxial tests . " Proceedings of the Third International Conference
on Soil Mechanics and Foundation Engineering ( 1953 ) .
3. Hirschfeld , R . C . , The Factors Influencing the Constant Volume Strength
of Clays . Doctorial Thesis , Harvard University , Cambridge , Mass . , May ,
1958 .
4. " The relation between shear strength and effective stress . " Primer
Congreso Panamericano de Mecanicia de Suelos y Cimentaciones
( 1959 ) .
163
STRESS RATIO
UPPER LIMITING CURVE (2)
f=
TTI7TTIA
A
,7
RPP
EFFECTIVE
17g
LOWER LIMITING CURVE
1
(
)
PRINCIPAL
PRESTRESS RATIO
-
Fig Prestress Ratio Principal Effective Stress Ratio Chart
,
1
.
.
STRESS
SHEAR
-T
Ön
,
7

Ō
-
Fig Shear Strength Envelopes

2
.
.
MATHEMATICAL RELATIONSHIPS
The curves of the principal stress ratio and prestress ratio are related
to
the shear strength envelope Let the principal effective stress ratio be
a
.
the prestress ratio

of
function
]'
=
(f R
)
well
PRESTRESS EFFECTS 165
. ,
)Rp
the
is
at Rp
of
then = 1 0 value and
(1
F
f
(
)
sind
,
1
1
(
)
-
f
the slope angle
of
drained shear envelope
of
where the consolidated
is
Ø
-
(
)
', )Rp
of
normally consolidated clays The value for point Fig associ
is
,
1
,
on P
.
.
' f(
as an
of
ated with infinite number points etc the shear strength
p
,
',
,
P
'
, .
Hirschfeld the relationship be
by
As
envelope shown Fig defined
in
2
, .
.
tween the prestress ratio Rp
on
the normal effective stress the failure plane
,
,
oc
and the preconsolidation pressure
ön
,
, olloc
(3
)
pm
Now Fig
)3
(
sin
sin
73
8
)
57
(1
+
53
sind
)
ø
2 +
03
-
(1
°1
57
)
-
(4
in
<
+
.
5
=
)Rp
Eq
be
using left hand

by
Using may replaced
of
ģ1
the value Özf and
;
2 ),
( 1
. .(
-
Eq replacement for sin simplification
of
as
part there develops after

Ø
,
a
)
+
(R
(1
P
P
)
)
(5
+
1
1
P
)
(
.Eq
we
.Eq
Öv
Substituting
of
the value obtained from into obtain

),
(3
(5
)
+
(1
1
(f
]
91
(6
)
1
(
)
f
stress the failure plane obtained using Fig

on
The shear
is
,
,
3
?
.
.
ream
02
*-
[
-1
[
]
.
..
-*
*
.Eq
by
as
Eq
of
Substituting the value expressed into we obtain after

),
(6
7
õ
.
)
simple operations
R
(
)
.
PR
+
)
(8
1
P
|
(
)
|
obtaining
an
on
expression for the ratio the failure plane

on
. of
thus and
in
T
the function of the prestress ratio

of
terms
The possible range shear strengths of prestressed clay shown by the
of
is
a
cross hatched area of Fig bounded by curves and These limiting curves
4
. 2
-
1 .
can be transformed into the prestress ratio plot of Fig Fig

. of
For curve
1
.
is .
the principal effective stress ratio

so
Fig
of
constant that curve

is
,
a 4
, 1
1
a
horizontal line The shear strength constant tr for the

is
assumed be
to
,
.
upper limiting curve of Fig Further assumed that the maximum

is
it
2
,
is .
.
.Eq
pressure
oc
consolidation also constant Using

,
,
),
(8
a
.
[P(1 ) - 1] [ P(R ) -
PDIVI (
R
tan Ø = 1 1)
* * * 2V 7(1)
PIR
)
+
that for the limit curve Fig
of
so
1
.
[P ]
) )
(1 -
) [E(
3 )
)] 1
(1 +(1
(R 2
R
10
10
(
(
) )
-1
]
[T
+
-
STRESS
SHEAR
-T VECTOR
CURVE
og
,
-
ON
og

-
Fig
Mohr Stress Circle

3
.
.
STRESS
ΟΤΑΝ
=
τ
T2
TAN
SHEAR
0
-
-T CURVE
(2
)
CURVE
1
(
)

J
-
Fig Bounding Limits Prestressed Shear Strength

of
4
.
.
PRESTRESS EFFECTS 167
COMMENTS
The data , Fig . 5 , for an organic silty clay fromNew Haven , Connecticut ,
presented by Hirschfeld are used to demonstratethe mathematical relation
ship developed . The data are those for 15 percent strain of the test specimens
unless failure occurred at a lower strain . Curves relating the preconsolidation
ratio with the principal effective stress ratio have been drawn in Fig . 5 . Al
though the data for the S tests on undisturbed clay are limited in number , a
straight line appears to be a good representation available data . The of the
datafor the remolded R tests have a fair degree of scatter , and a moderately
curved line appears to fit the data best . It may be that the results of the S and
the
tests might be represented
by
remolded single curve The data for the
R
, .
undisturbed clay exhibit considerable scatter and
on
at
tests best the curve
R
,
Fig merely approximation
an
is
shown
in
5
.
Fig Fig .
through
The three curves shown have been transferred
to
in
6
5
.
.
use
for preconsolidation pressure equal All three curves

.Eq
unity
to
of
(8
.
)
above the shear strength envelope for normally consolidated clays on

lie
test
S
by
shown the long dashed

line
.
Fig are actually relative stress

be
noted that the curves

of
should
It
6
.
that strength curves for various preconsolidation pressures such

so
curves
,
,
Fig would all coincide Fig This follows from the

as
in
in
indicated
6
. 2
,
.
.
.
Fig may be that this coincidence of strength curves does

of
relationships
It
.1
not
actually exist because of the influence strength values of

on
decreasing a
voidratio with increasing consolidation
pressures
.
are shown for the nor

by
The lines as defined Casagrande

and Wilson
,
.1 2,
F
prestress ratio
of
mally consolidated cases For the line coincides

0
a
F
.
with the straight line strength curve The line for the remolded tests
R
S
F
.
-
clays Shear strength data

of
lies below that for the undisturbed the left
to
.
these lines for test must be obtained from preconsolidated specimens

R
F
.0 .
Shear strength data between these lines and the line for Rp equal are
to
1
F
F
of
necessity controlled drainage during the

of
obtain because the

to
difficult
tests This might performing
by
accomplished
be
an
test selected strain

to
S
a
.
andthen completing as
an
test
R
CONCLUSIONS
Considerable shear strength exists between the strength curves

difference
in
the
the
consolidated undrained tests and preconsolidated drained tests

for of
S
-
undistrubed clay Only minor difference strength exists between the

in
a
.
the preconsolidated undisturbed and the preconsoli

of
shear strength curves

S
datedremolded curves when plotted against effective normal stresses

R
It
.
wouldappear that the excess strength

of
the undisturbed test caused both

is
R
by
structure bond and soil particle arrangement and by prestress The close
,
strength between the undisturbed

of
relationship preconsolidated and the

,
remolded tests indicates that considerable remolding occurs during the

R
type test may be that further investigations will show that these curves
of
It
.
actually coincide hoped that the mathematical relationship developed

is
in
It
.
this brief paper may be prestress

of
further study
of
of
value the influence

in
uponshear strengths of clays

.
THEORETICAL LIMIT
RINCIPALEFFECTIVE STRESSRATIO
At UNDISTURBED-R
REMOLDED-
LL
SYMBOL BORING TEST
169 UND
R -S R R S
H H H H H
- - -
- - - - -
169 UND
169 REM
UNDISTURBED-
S
193
85 S
UND
45
193 UND-
THEORETICAL
LIMIT PRECONSOLIDATEDSPECIMEN
*
PRESTRESSRATIO
-
Fig Prestress Ratio Principal Effective Stress Ratio Data

,
5
.
.
7
.
LINE UNDISTURBED NORMALLY CONSOLIDATED
=
R
2
F
,
.
RELATIVE SHEARSTRENGTH
FLINE REMOLDED NORMALLY CONSOLIDATED

=
R
2
.0
,
.
FLINE UNDISTURBED NORMALLY

R
1
.0
,
-
.
CONSOLIDATED
8.0
UNDISTURBED
S R R
-
( ( (
)
- - -
id REMOLDED)
/T
06
UNDISTURBED)
.0
2
0
4
0
.8
0
.6
.
övlar RELATIVE NORMALSTRESS

-
Fig Relative Shear Strength Envelopes

6
-
.
.
PHYSICAL COMPONENTS OF THE SHEAR STRENGTH
OF SATURATED CLAYS
By M . Juul Hvorslev , F . ASCE
INTRODUCTION
During the period 1933 - 1936 the writer was engaged in research on the
physical properties of clays at the Technical University of Vienna , Austria ,
under the general direction of Professor K . Terzaghi . The results of this re
search were published as a thesis in 1937 , and abstracts thereof were pub
lished as conference papers in 1936 and 1938 . The writer has had but little
opportunity personally to perform additional experimental investigations of the
shear strength of clays , and this paper is to a large extent a re - evaluation and
restatement of some of the results of the above - mentioned research . The
paper deals primarily with the physical components of the shear strength of
remolded , saturated clays and with the various factors which influence these
components . The results of subsequent research by others are taken into con
sideration and summarized when appropriate , but the paper is not a complete
review of the very extensive and important research on shear strength of satu
rated clays performed during the last twenty years . The sources of error in
the
tests performed placed

by
and emphasis
on
the writer are discussed
is
,
the assumptions and limitations relating

of
clarification the conclusions and

to
formerly proposed criteria for the shear strength saturated clays New data
of
and illustrations are included but many figures are copies simplifications
or
,
the original figures the 1937 publication

of
in
Most of the data presented this paper concern the properties of remolded
in
by
of
clays As demonstrated CASAGRANDE the results

on
1932 tests
A
of ,
2
.
remolded clays cannot be used directly for solution practical problems in

volving undisturbed clays Nevertheless many relations concerning physical
,
clays were first determined means of tests

by
remolded clays
on
of
properties
,
and they apply with minor modifications and limitations undisturbed clays
to
but the coefficients entering these relations may be quite different for re
molded and undisturbed clays
of
The use remolded soils basic research

in
.
has important advantages especially regard uniformity the test speci

of
in
to
,
their stress history and separation

of
of
mens control
of
the influence the

,
of
many variables which govern the deformation and strength characteristics

However essential that research
on
soils remolded soils be supple

is
it
,
.
mented by research on undisturbed soils and that the results of the latter be
,
Sta
Army Engr Waterways Experiment

to
Cons the Soils Div

,
.,
.,
U
1
S
.
.
.
Vicksburg Miss
,
Items indicated thus CASAGRANDE refer the corresponding

to
1932
2
,
(
)
.
entries arranged alphabetically the Appendix References

in
,
II
.
.
-
169
verified and amended by observations and analyses of the behavior of proto

type structures .
Notation . - The letter symbols adopted for use in this paper are defined and
arranged alphabetically , for convenience of reference , in Appendix I.
1. INTRINSIC FORCES AND STRUCTURE OF CLAYS
Intrinsic Forces
Clarification of the physicochemical constitution of clays and of the intrin
sic forces acting in the soil - water system is of utmost importance to a better
understanding of the deformation and strength characteristics of these soils .
Earlier investigations of this problem by GOLDSCHMIDT (1926 ) , TERZAGHI
( 1925 , 1931 , 1941 ) , ENDELL (1936 ) , FREUNDLISH ( 1935 ) , and others have in
recent years been supplemented by extensive research in the fields of colloid
chemistry , mineralogy , soil physics , and soil mechanics . Significant results
have been obtained , and among the many papers dealing in part or whole with
the subject , reference may be made to those by MASON and WARD ( 1954 ) ,
ROSENQUIST ( 1955 ) , TAN ( 1957 ) , LAMBE ( 1958 ) , and to those in a symposium
on physiocochemical properties of soils published in the Journal of the Soil
Mechanics and Foundations Division , ASCE , April , 1959 . It is difficult to sum
marize the results of this research because of the remaining uncertainties and
differences in interpretation of the observed phenomena .
The conception that the usually flaky particles of the clay minerals are sur
rounded by double layers of bound and partially bound water has been strength
ened and further detailed . It is agreed that the clay particles are covered with
a thin film of strongly bound or adsorbed water which in turn is surrounded by
a thicker film of partially bound water . The structure of the water in these
layers is different from that of free liquid water ; the adsorbed water has some
properties similar to those of ice , but opinions vary in regard to its actual
structure , density , and viscosity . However , it has been suggested that the ad
sorbed water is so strongly bound to the clay particles that it is moved by dif
fusion rather than viscous flow , and that it prevents actual contact between the
clay particles at pressures normally encountered in soil deposits ; BOLT
( 1956 ) , ROSENQUIST ( 1959 ) . The water in the adsorbed and partially bound
layers primarily serves to transmit and modify the electrochemical forces
between the clay crystals , and the bipolar character of the water molecules
and the ions in the water are of primary importance to this function . The
principal intrinsic forces are the van der Waals forces and the Coulombic
forces . The van der Waals forces normally cause attraction but decrease ex
tremely rapidly with distance , and they predominate at points of very small
spacing of the particles . The Coulombic forces are ( 1) electrostatic attraction
between the positive charges at the edges of one particle and the negative
charges at the flats or surfaces of other particles , and ( 2) electrostatic repul
sion between two surfaces or between two edges of adjacent particles . Addi -
tional cohesive forces may in some cases be caused by hydrogen or potassium
bonds or may result from cementation by organic or inorganic compounds .
The magnitude of the resultant intrinsic forces depends on the type of clay
minerals , the size and corresponding specific surface of the particles , the type
of ions adsorbed on the surfaces of the clay crystals , the type and concentra
tion of the ions in the water , and on temperature . Since the Coulombic forces
decrease with the square of the distance and the van der Waals forces with
PHYSICAL COMPONENTS 171
about the sixth power of the distance from the particles , the intrinsic forces
are governed by the points of shortest distance between the particles rather
than by their average spacing ; that is , the intrinsic forces depend not only on
the void ratio but also on the geometric arrangement of the particles or the
soil structure ; see LAMBE (1953 , 1958 ) , ROSENQUIST ( 1959 ) . External and
gravity forces influence the intrinsic forces when the external forces cause
a change in the effective stresses in the soil -water system and a correspond
ing change in the spacing of the clay particles .
Considering the deformation characteristics of clays , it has been proposed
by GOLDSTEIN ( 1957 ) and others that the intrinsic forces be divided into two
groups , one producing elastic bonds and the other forming viscous bonds , but
it has not yet been possible to identify the basic forces and conditions which
cause formation of the two types of bonds . NASCIMENTO ( 1953 ) and BORO
WICKA ( 1959 ) have suggested that changes in free energy and corresponding
forces at the interface of water and soil particles can produce tension in the
bound water and a corresponding cohesion at very close spacings of the parti
cles . However , it has not yet been verified experimentally that such a tension
in the bound water exists , and the hypothesis has not been correlated with
other observed phenomena or with the concepts summarized in the foregoing
paragraphs .
The intrinsic forces in the soil -water system may in some cases undergo
thixotropic changes , defined by FREUNDLICH ( 1935 ) as isothermal , reversible
sol - gel transformations ; that is , the strength or coherence of a clay may be
decreased , without any change in water content or temperature , by large and
rapid deformations , but the strength is gradually regained when the deforma
tions cease or the rate of deformation decreases . The phenomenon may be
visualized as a disturbance and subsequent re - establishment not only of the
arrangement of the clay particles but also of the structure of the bound water
with consequent changes in the transmission of electrochemical forces be
tween the particles . The thixotropic properties of clays depend primarily on
the type and concentration of ions in the pore water and adsorbed on the sur
face of the clay minerals . Recent data and concepts relating to thixotropic
changes in strength of soils are presented in the papers by SEED and CHAN
( 1959 - A ) and MITCHELL ( 1960 ) . The latter has demonstrated that the pore
water pressure increases during a thixotropic disturbance of a cohesive soil ,
and that this pressure decreases during a subsequent thixotropic hardening of
the soil , which in a physical sense explains the changes in strength . Similar
observations have also been made by BISHOP , ALPAN , BLIGHT, and DONALD
( 1960 ) .
Clay Structure
TERZAGHI (1925 ) suggested various single - grain and honeycomb struc
tures as an explanation of the great variations in void ratio and behavior of
soils . Goldschmidt maintained in published lectures that the flaky clay parti
cles in sensitive clays lean upon each other and form an open and unstable
structure ; see ROSENQUIST ( 1959 ) . CASAGRANDE ( 1932 ) proposed a com
pound honeycomb structure of silt grains and flocculated clay particles for
sensitive undisturbed clays .
& The term " structure ” refers here to the geometric structure or arrange
ment of the soil particles , which also is called “ fabric ” in some recent publi
cations , where " structures includes both the geometric structure and the force
structure ; see LAMBE ( 1958 ) .
The structure of a clay is disturbed by remolding , and it may be assumed

that the clay particles have no preferred orientation after a thorough remold
ing , considering the entire remolded mass . However , a definite orientation
may be imparted to the clay particles during subsequent consolidation and
strength tests on the remolded clay . This possibility was investigated by the
writer by means of slaking tests . A silty clay was remolded at a water con
tent close to that of the liquid limit and then reconsolidated by confined com
pression . A thin slice was cut from the test specimen in a direction parallel
to that of the principal consolidating stress , and another slice was cut per
pendicular thereto . These slices were partially air - dried and then laid in
water and allowed to slake . As seen in Figs . 1 - A and 1 - B , the expansion oc
curs in a direction parallel to the principal consolidating stress , and the
fis
sures are perpendicular thereto the re
of
another part
of
ball was formed
A
.
molded clay and air dried this case the consolidation
In stress capil
,
e
i.
.
.
-
lary pressure of equal magnitude all directions and slaking test on

is
in
a
,
,
preferred orientation
of
in no
thin slice cut from the ball shows deformations
and fissures Fig As shown Fig similar results were also ob
C
1
2
,
,
.
.
-
slaking tests partially dried slices undisturbed clays

on
of
tained
in
.
The conclusions drawn from these experiments were that the clay particles
the remolded and uniaxially reconsolidated clay have preferred orienta
in
a
tion perpendicular the principal consolidating
of
stress and
to
the direction
,
that the consolidated test specimen may have anisotropic strength deforma
,
tion and permeability characteristics These conclusions were substantiated
,
unconfined compression tests described

by
of
the results Section
in
6
.
probable that shear strains promote orientation the clay particles
of
is
It
direction parallel the principal strain provided the strains are

of
that
to
in
,
a
active over sufficiently long period of time the strains also are very
If
a
large they may cause formation
as
of
slickensided failure surfaces found
in
,
the USAE Waterways Experiment Station Fig HVOR ,

at
torsion shear tests
3
;
,
SLEV and KAUFMAN 1952 However slickensided failure surfaces were not .
,
(
).
on
kg
formed when the normal stress the surface was less than cm2
0
.5
.
/
Recent research has yielded very interesting hypotheses and data

on
the
and compacted clays and
on
of
of
structure both undisturbed the influence the
structure on the deformation and strength characteristics of clays LAMBE
.
1953 1958 has suggested that clays which were flocculated during sedimenta
or ,
(
at
tion compacted water contents less than optimum water content have
a
the clay particles or type of structure

of
random orientation cardhouse

a
”
"
Conversely clays which were dispersed during sedimentation or compacted

at
,
water content greater than the optimum water content have parallel orien
a
the particles Possible variations from these general rules are dis
in of
tation
.
cussed paper by SEED and CHAN 1959 who also found that although
B
a
,
(
)
-
on
the structure may have pronounced influence the deformation character

a
compacted clays has relatively little influence on the maximum

of
istics
it
,
strength because reorientation of the particles the failure zone may take
in
a
place during shear or triaxial test

a
ROSENQUIST 1959 has developed ingenious techniques for obtaining stereo

(
photographs of the structure

by
means of
of
undisturbed clays
an
electron
microscope He found that the structure of marine clays general ran
in
is
a
.
dom orientation of the particles and resembles the cardhouse structure sug
by
gested GOLDSCHMIDT LAMBE 1958 and TAN 1957 However the

,
),
,
).
(
photographs also show that groups of particles may form packets with par
a
“
"
allel alignment the flaky clay particles

of
BUESSEM and NAGY 1953 SALAS

),
.
(
-
PRESSURE PERPENDICULAR PRESSURE HORIZONTAL PRESSURE UNIFORM
TO PLANE OF PICTURE IN PLANE OF PICTURE IN ALL DIRECTIONS
Fig . 1 Slaking Tests on Remolded and Reconsolidated Vienna Clay
А
SILTY CALCAREOUS CLAY VIENNA SILTY CLAY CRUMBLY CALCAREOUS CLAY
HORIZONTAL SLICES IN LEFT HALF AND VERTICAL SLICES IN RIGHT HALF OF PHOTOS
Fig . 2 Slaking Tests on Partially Dried Slices of Undisturbed Clays
LAKE PROVIDENCE CLAY ATLANTIC MUCK , PANAMA CANAL

VERTICALLOAD 2. 0 KG / CM2 VERTICAL LOAD 5.0 KG / CM2
Fig . 3 Formation of Slickensided Failure Surfaces in Torsion Shear Tests

and SERRA TOSA ( 1953 ) , and MITCHELL (1956 ) have used X - ray diffraction or
polarized light and the birefringent properties of clay crystals to investigate
the relative orientation of clay particles . They found that the particles in re
molded and uniaxially reconsolidated clays are oriented perpendicularly to the
direction of the principal consolidating stress . MITCHELL ( 1956 ) also investi
gated the structure of many undisturbed clays . He found that some clays have
a completely random orientation of the particles and that others have various
degrees of a preferred orientation of the particles . The degree of parallel
orientation of the particles is uniform in some clays ; in others , this degree of
orientation varies from spot to spot or zone to zone .
2. DE FINITIONS AND CONCEPTS RELATING TO SHEAR STRENGTH
Shear Strength and Effective Stresses

The failure conditions for a soil may be expressed in terms of a limiting
shear stress , called shear strength , or as a function of the principal stresses .
The two forms of the failure conditions are often but not always interchangea
ble . The shear strength of a soil , Tf, may be defined as the shear stress in
the plane of failure at the time of failure . In case of direct or torsion shear
tests , the plane between the stationary and moving parts of the equipment is
considered as the enforced plane of failure . The stress - displacement curve
for this plane usually has a well - defined peak point , Fig . 4 , and the corres -
ponding maximum value of the shear stress is considered to be the shear
strength for the purpose of the investigations discussed in this paper . It
should be noted that the shape of the stress - strain curve depends on the type
of loading , rate of deformation , and drainage conditions . The oldest and still
widely used expression for the shear strength is the Coulomb failure criterion ,
tp = c + Op tan ( 1)
where c is the cohesion , of the normal stress on the failure surface , and the
angle of internal friction . This equation is relatively simple , but the value of
c and depend on many factors and may vary between wide limits ; also the
field of application of Eq . 1 is limited to conditions duplicating those existing
during the test in which the values of the coefficients were determined .
A failure criterion of greater general applicability is obtained by expres -
sing the shear strength as a function of the effective normal stress , o's, in ac
cordance with Terzaghi 's fundamental concept that the strength and deforma
tion characteristics of soils are governed by the effective stresses rather than
the total stresses . This paper deals solely with saturated soils for which
Os = op - u (2)
and
Tp = c ' + of tan $ (3 )
where u is the pore -water pressure , which is assumed to be zero in the case
of fully drained tests . The pore - water pressures thus defined or measured
probably represent the pressures in the free or lightly bound pore water ; the
pressures in the more strongly bound water at points of minimum distance
SHEAR STRESS
PEAK STRENGTH
RESIDUAL STRENGTH
DEFORMATION OR STRAIN
Fig Shear Stress Deformation Diagram

4
.
TA
STRENGTH
SHEAR
by
LIT
1
.
|
Led TAN
O
'
T
=
T
!
-
-
Ezen
TAN
O
a
16
NORMAL STRESS
O
Fig Coulomb Shear Strength Diagram

5
.
TA
SIMPLE OVERCONSOLIDATION
REBOUND CURVE
(
)
SHEAR STRENGTH
NORMAL
CONSOLIDATION
SZ CYCLIC OVERCONSOLIDATION
-
RELOADING CURVE
(
ps
NORMAL STRESS
O
p
'.
Fig Shear Strength Hysteresis Loop

6
.
between the soil particles are not yet definitely known . The effective stress
concept may be said to be semiempirical , but its practical validity has been
demonstrated in many laboratory tests and field observations .
Influence of the State of Consolidation
Agraphical representation of Eq . 3 is shown in Fig . 5. The line OAC is the
shear strength line for a normally consolidated clay , for which Ø ' = 0 ; and
c ' = 0. The line BA is the shear strength line for the same clay preconsoli
dated at the pressure on and having Ø = Øn . If the clay is
preconsolidated
at other pressures , shear strength lines parallel to BA are obtained , as in
dicated by the dashed lines in the figure . This means that c ' is proportional
to op or
'=
tan
c
(4
)
and Eq may
be
then written
3
.
tan
jó
op
ostan
%
=
A
(5
$
:
-
)
For normally consolidated clay and
=
',
a
d
o
'
Tp
tan
tan
of
%
=
B
(5
.
)
-
the right side Eq may
as of
be
on
The first term the cohesion
as
considered
B
5
-
.
component and the second term the friction component and the anomaly of
,
having two angles for the same clay and
no
internal friction
of
and
Ó
,
',
',
cohesion for normally consolidated clay thereby eliminated As shown

is
a
.
later the shear strength lines BA Fig are simplifications of actual test
in
5
,
results and they represent changes both the friction and cohesion compon
in
,
are not the actual cohesion or angle

of
ents Therefore the values and

,
c
.
'
internal friction but merely mathematical components and

of
now called
is
;
c
'
the cohesion intercept and the angles of inclination of the shear strength lines
are called the angles The form and general concept
of
shear strength Eq
of
.
.
were first proposed by Tiedemann and were based extensions of in
on
A
5
-
vestigations by Krey see SEIFERT 1933 41 and BRENNECKE LOHMEYER

p
;
,
(
)
.
1938
).
(
The shear strength lines shown Fig and Eq

on
are based the results

in
5
.
of fully drained tests OHDE 1955 emphasizes that Tiedemann performed

.
fully drained tests normally consolidated test specimens but started tests
on
overconsolidated test specimens immediately after reducing the total nor

on
Op
mal stress and attempted constant water content

Of
maintain
to
to
from
a
equal that existíng the preconsolidation pressure correct but

at
This
to
if
is
;
.
the water content actually were held constant the values of

off
Tf
and would
,
also be practically constant and equal those existing at the preconsolidation

to
pressure that the shear strength lines BA would be nearly horizontal as

is
;
found by the writer and others Tiedemann obtained appreciable slopes of the
.
lines BA which indicate that partial consolidation swelling and equalization

or
,
the pore water pressures had taken place during the test The writer there
of
.
-
fore took the liberty calling Eq the Krey Tiedemann failure criterion
of
5
.
The shear strength relations shown Fig and Eqs and are simple
in
in
5
5
.
and practical but they furnish only approximate values of the shear strength
,
of overconsolidated soils As first shown by TERZAGHI 1929 the shear

,
(
)
.
strength diagram for a series of test specimens overconsolidated to corre

spond to various points on the rebound and reloading branches of the pressure
void ratio diagram , Fig . 10 , forms a hysteresis loop similar to that shown in
Fig . 6 . This diagram shows that different values of Te can be obtained for
identical values of op and on, and that the actual shear strength lines may
deviate appreciably from the straight lines shown in Fig . 5 when of is small
compared to on . The shear strength lines of the hysteresis loop have much
greater curvature when the shear strengths are determined by undrained tests
and plotted against total normal stresses , of; see Fig . 31 .
Shear strength lines with a slight double curvature are often obtained in
tests on undisturbed test specimens , which may be explained in the following
manner . Assume that the clay is preconsolidated at the vertical pressure op
and that the current overburden pressure for the test specimens is o'd, Fig .
6 . The shear strength line DB on the rebound branch is obtained when tests
are performed for values of of smaller than o' s, but the shear strength line
will pass from D to E on the reloading branch on the hysteresis loop when
tests are performed for values of of greater than od
Triaxial Stresses and Strengths
The foregoing discussion applies primarily to the results of direct box or
torsion shear tests . The results of the now commonly used triaxial tests are
usually evaluated by means of the Mohr diagram , as shown in Fig . 7 for an
overconsolidated soil and simplified conditions corresponding to those in Fig .
5 . If the angle abetween the direction of the major principal stress and the
planes of failure is measured or assumed , the shear stress and corresponding
normal stress on a failure plane can then be determined graphically as shown
in Fig . 7 or by
sin
lo - o )
za
p
os
cos
01
40
03
za
)
(
)
-
-
=
+
o
by
be
shear strength can then determined diagram similar that

to
The
at in
a
Fig but Tf and

of
noted that the maximum values 02 occur

of
is
it
4
,
.
the same time

or
constant throughout the test

is
strain when case the

In
a
soil has isotropic strength properties which conform the Coulomb failure
to
Eq shown mathematically that the optimum value

be
of
condition can
it
,
3
,
a
.
given by
is
45
or
=
90
20
a
=
'
(
)8
These equations can theoretically used for computing

be
or
when known
is
,
Ø
a
'
vice versa However Eq not correct when the soil has anisotropic
is
8
or ,
.
strength properties proper

or
when nonuniformities stress distribution

in
ties exist the test specimen see Section

in
6
.
The plane corresponding the point of tangency

of
the Mohr envelope has

to
am
of
of
the angle
45
the inclination 0m where inclination the

m
=
is
Ø
),
-
(
'
Mohr envelope but this does not signify that the Mohr envelope the shear
is
,
strength line and that the angle internal friction unless the inclina
of
'm
is
am
tion of the failure planes also and the soil has isotropic strength proper
is
ties As seen Fig the shear strength line intersects the Mohr circles
in
,
6
.
.
MOHR ENVELOPE
iva
Fig . 7 Mohr Diagram for Triaxial Tests
SHEAR STRENGTH LINE
Ø =ASSUMED
ANGLE OF INT .
FRICTION
90 - 07
a =45 - Ž
PK ,
pelin a
Fig . 8 Mohr Diagram for Direct Shear Tests
II STRONG OVERCONSOLIDATION LIGHT OVERCONSOLIDATION

SWELLING SETTLEMENT
SHEAR STRENGTH
TpaTd Tot ce
CV
Cu
To
=
ENERGY COMPONENT LT
.
EGATIVE
Td
POSITIVE
FRICTION COMPONENT
W
To
TANDE
=
0
f*
COHESION COMPONENT call

=
CV
RHEOLOGICAL COMPONENT
=
Fig Components of Shear Strength Constant Void Ratio

at
9
.
when either smaller or larger than am . Some investigators use the shear
a is
strength angle for normal consolidation , Øs, and others use Ø'r or Ø'm and
corresponding values of a for evaluation of triaxial test results . According to
the hypothesis proposed by the writer and described later , the effective angle
of internal friction , Øe , is smaller than Øi and Øm ; and it has also been
found that the measured values of a for remolded clays usually are larger than
am , although they are subject to considerable variation . The difference be
tween assumed and actual values of a may result in appreciable variations in
the computed values of Tf and of , but it should be noted that these variations
cause only a slight shift in the position of the shear strength line . Although
the actual shear strength line may be located slightly below the Mohr envelope ,
the latter is usually assumed to be the shear strength line in evaluation of tri
axial test results for practical purposes .
Difficulties connected with determination of correct values of Tf and of
from results of triaxial tests can be avoided by expressing the failure condi
tions in terms of the principal stresses . When the Mohr envelope is a straight
line with the inclination d'm and the cohesion intercept c'm, the failure condi
tion in terms of principal stresses is
pon
(0 - 03) ( 01 - 03) =
line
cos
sin
A
(9
= 2c
+
-
(
)
0
The actual Mohr envelopes are often slightly curved and their replacement
,
with straight lines equivalent assuming that the soil conforms the
is
to
to
Coulomb failure criterion Therefore Eq called the Mohr Coulomb
is
,
A
9
.
.
-
failure condition This equation may also be written the following form -
in
.
tan
201
03
66
Bo
);
01
B
(9
-
)
)
-
=
+
(?
0
.Fig
oj
By
as
plotting straight line

34
versus
of of
shown
,
,
in
+
g
0
a
-
(
')
')
s
Bs
.cg
obtained which has the angle inclination and the intercept The
is
parameters om and Eqs

be
by
comparison
of
determined
m
can then
A
9
c
-
.
'
and
B
9
.
-
sin
on
Bg
tan and cos

=
=
C
c
/
C
9
-
(
)
of
diagram much easier series Mohr circles and

to
is
Such draw than

,
it
a
facilitates averaging of scattered test results and determination

of
the mean
values of the parameters Examples application of this method for evalua
of
.
triaxial strength tests are described by KYVELLOS

of
of
tion the results

who indicates that the method originally was suggested Caquot
by
1956
),
), A
A (
similar method was developed

by
SKEMPTON and BISHOP 1954 who trans

(
formed Eq into
A
9
.
-
che
sin
cos
sin
D
(9
-
)
;
-
=
)(
-
0
+
)
plot versus yields straight line from the inclination and in

',
,
A
of 0
a
-
(
as
tercept which values ofØ Fig

be
and can determined illustrated

m
in
,
c
.
be '
'
.Eq
Finally may
27
written
-A
9
.
2cmn
sin
cos
sin
01
of
=
+
-
(
1
+
)
(9
E
(
1
-
)
)
o
and plots of oſ versus o'z form straight lines as shown in the Rendulic dia
gram , Fig . 38 . Further details of evaluation of the results of triaxial strength
tests are discussed in Section 9.
Conversely , the results of direct and torsion shear tests cannot be expres
sed in terms of principal stresses without knowing or assuming the angle of
inclination of the failure planes and / or the effective angle of internal friction .
For the conditions and assumptions shown in Fig . 8, where is the angle of
inclination of the shear strength line and Ø , is the effective angle of internal
friction , the failure criterion in terms of principal stresses is
sing
coca cos
or
(0 - 0 )
cos
cos
89
(
01
)89
10
= 2c
+
(
)
(
(
0
(
)
, de
Changes and may cause appreciable changes and the

in
in
+
Ø
),
0
o
.
(
Mohr circles but only relatively small changes the position and inclination
in
of
the Mohr envelope

.
Volume Changes and Surface Energy

Volume changes Changes pore water pressure and or volume of the soil
in
/
-
.
during shear tests constitute important factors
of
evaluation the test results
in
.
These changes are summarized below qualitative sense order facili
to
in
in
a
tate further definitions and explanations

of
observed phenomena
.
The pore water pressure increases and or the void ratio decreases when
-
, /
the average external pressure increased and vice versa

is
.
The pore water pressure increases and or the void ratio decreases when
-
normally consolidated or slightly overconsolidated clays are subjected
an
to
increase shearing stresses
in
The pore water pressure decreases and or the void ratio increases when
/
-
strongly overconsolidated clays are subjected shearing

an
increase
to
in
stresses
.
stress reversal shearing stresses may produce changes

or
decrease
in
A
pore water pressures and void ratios which are of the same sign but of
in
much smaller magnitude than those caused by comparable increase shear

in
a
ing stresses
.
The pore water pressure increases with increasing time when normally
-
slightly overconsolidated clays are subjected

or
consolidated constant shear

to
ing stresses and drainage prevented

is
the pore water pressure increases or decreases

not yet known whether
is
It
with increasing time when strongly overconsolidated clays are subjected

to
constant shearing stresses and drainage prevented

is
test specimen signifies addi

of
Surface energy change the volume

in
of A
-
.
tion or expenditure surface energy TAYLOR 1948 demonstrated that con

)
(
.
volume changes and corresponding changes surface energy

of
sideration
in
during shear tests sands can explain the difference shear strength of
on
in
dense and loose sand This concept was later formulated by BISHOP 1950
,
(
)
.
or
who introduced strength component called the dilatation component the

,
a
surface energy correction which denoted by Td this paper Assume that

in
is
Objections have been raised

dilatation because soils may
to
the term
”
b
undergo either volume during strength tests

an
decrease increase The

or
to in
a
term surface energy correction refers correction of the friction and

a
”
"
cohesion components but component and not correction of the total

it
is
a
a
,
Misunderstandings may be avoided

by
shear strength using the term volume

"
.
change component
."
the change in thickness of a direct shear test specimen is dy during a time in

terval đt at failure , and that the lateral displacement in shear is dx during the
same time interval , then the surface energy is ( A • Of • dy ) and the internal
energy corresponding to Ta is ( A To : dx ) , where A is the area of the test
specimen . If it is assumed that the two energy quantities are equal, then td is
defined by
Ta = of a ( 11)
BISHOP and ELDIN ( 1953 ) , and BISHOP (1954 ) have shown that with similar
assumptions the surface energy correction for triaxial tests can be expressed
by
Ta = op ( 12)
where dy is the rate of volume strain and de the rate of axial strain at failure .
Surface energy corrections are also made at stresses and strains before and
after failure in order to obtain stress - strain curves for the individual strength
components . The component , td , is positive for a volume increase , negative
for a volume decrease , and zero for constant volume or undrained tests .
Introduction of the surface energy component in the evaluation of shear
tests on sands furnishes an explanation of the difference in strength of dense
and loose sands . The resulting angle of internal friction corresponds to the
ultimate or residual shear strength of dense sands , Tr in Fig . 4 ; see also
BISHOP ( 1950 ) . The peak shear strength and the residual shear strength are
identical in the case of loose sands . It is to be noted that both friction and
surface energy components of sand are linear functions of the effective stres
ses , and it is a matter of definition or choice whether in practical applications
the strengths should be expressed by a friction component based on the peak
strength or on a surface energy component plus a friction component based on
the residual strength .
GIBSON (1953 ) and others have extended the concept of a dilatation compon
ent or surface energy correction to the evaluation of the results of drained
shear tests on clays . The writer has previously suggested , HVORSLEV ( 1953 ) ,
that a part of the volume change of clays during shear is an indirect conse
quence of the shearing strains , and that a corresponding part of the surface
energy is absorbed or contributed by internal forces as in consolidation tests
and should be disregarded in computing the surface energy component . It is
possible that a part of the volume change of clays during shear is caused by
interference or interlocking , as in sands , but it is also probable that the shear
ing strains cause a partial disturbance and thixotropic weakening of the bonds
or a decrease in the effective stresses between the clay particles . It is known
from consolidation tests that such a disturbance promotes additional consoli
dation of normally consolidated or slightly overconsolidated soils and addi -
tional swelling of strongly overconsolidated clays .
The external energy supplied during a volume decrease of the test speci
men may , at least in part , be absorbed internally by consolidation and produce
an increase in cohesion . The external energy expended during a volume in
crease may in part be supplied by internal swelling pressures and produce a
decrease of the cohesion . On the other hand , introduction of the dilatation
component or surface energy correction defined by Eqs . 11 and 12 primarily
changes the angle of shear strength ; see Fig . 9 .
In the evaluation of slow , drained shear tests it is usually assumed that the
excess pore -water pressure at failure is zero . However , measurable rates of
volume change at failure signify the existence of remanent , positive or nega
time , excess pore -water pressures . These pore -water pressures are a func
tion of the rates of volume change and the properties of the soil, and theoreti
cally they can be computed , but it is difficult to do so when the volume changes
are not uniformly distributed throughout the test specimen . Consideration of
the remanent excess pore -water pressures would increase the angle of incli
nation of the shear strength line , and it is probable that a part of the surface
energy correction actually is a correction for remanent excess pore - water
pressures .
Some of the problems discussed in the foregoing paragraphs may be said
to be of academic rather than practical interest , because the full surface en
ergy corrections for shear tests on clays usually are relatively small at the
moment of failure , although they may attain appreciable magnitudes before
failure .
The writer has not yet reached definite conclusions concerning the proper
accounting for the surface energy corresponding to volume changes during
shear tests on clays . Further investigations are needed concerning the magni
tude of remanent pore -water pressures in the zone of failure during slow
drained tests , and also for clarification of the mechanics and intrinsic forces
involved in the consolidation and swelling of clays , caused directly or indirect
ly by shear stresses and strains .
Components of the Shear Strength
As indicated in the foregoing discussion and as shown in Fig . 6, the meas
ured shear strength cannot be expressed as a unique function of the effective
normal stress and the preconsolidation pressure . The writer ' s investigations
show that the shear strength of remolded , saturated , and normally consolidated
clays can be expressed either as a function of the effective normal stress , o's,
or as a function of the void ratio at failure , ef. Furthermore , the shear
strength of these clays in states of both normal consolidation and simple cyclic
overconsolidation can be expressed by a combination of the two functions . It
has been suggested that the components of the shear strength corresponding to
the two functions be called the “ stress component ” and the " void ratio compo
nent , " and these terms are quite logical and attractive . Nevertheless , the
writer prefers in this paper to retain the established terms " effective friction
component ” and “ effective cohesion component ” in connection with extensions
of the Coulomb condition of failure , and the two components are designated by
To and ce . The terms " stress component " and " void ratio component ” may be
preferable when the failure conditions are expressed in terms of the principal
stresses or functions thereof. As discussed in the foregoing subsection , con
sideration of a " dilatation component " or " surface energy correction " has been
suggested in recent papers . This component is designated by Td , and the meas
ured shear strength , Tf, may then be expressed by
If = Tot ce tła (13 - A )
Most cohesive soils possess an apparent structural viscosity and their de

formations are of a visco - elastic character . The corresponding strength com
ponent may be called the " viscous component , " but factors other than viscosity
seem to be involved , and themore inclusive term " rheological component ” and
will be assumed that forms part
Cy
the
Cy
of
are proposed
notation the
It
a
, .
effective cohesion component ce because the effective friction component
,
,
by
of
be affected
Tg
remolded clay does not seem the increased rate
to
of
a
,
deformation after failure provided the soil structure not changed see Sec
is
;
,
However this assumption further experimental corrobo
of
need
is
tion
in
,
8
.
zero with increasing time

Cy
ration The value of converges decreasing
on
or
of .
deformation whereas Ce at the same time approaches
By an
rate ultimate value
,
,
which may be called the ultimate cohesion component
Cu
definition the
,
."
"
test duration or rate
of
exists any given
at
following relation deformation
13
Cv
Cu
Ce
B
)
=
-
(
rearranged
.Eq
be
may then the following

13
written form
in
A
-
14
Cy
Cu
Ce
Td
Te
Tø
Tø
)
(
=
+
-
the
the left side represent the external energy and those

on
on
where terms the
right side the internal energy
,
of
definition and experimental determination
of
For the purpose the individ

ual components the basic assumption made that the cohesion and rheologi
is
,
satu
of
cal components are constant when
or
the void ratio water content
1
(
)
rated clays constant the rate of deformation or test duration constant
is
is
2
,
,
(
)
and there no significant difference the geometric structure of the clays

is
in
3
(
)
during given test series see Section The probability that the latter condi
1
;
a
tion fulfilled enhanced by the observation that changes geometric struc

in
is
is
ture during strength tests tend eliminate initial differences and produce
to
to
similar structures the moment of failure SEED and CHAN 1959 shear
at
A
;
)
.
strength diagram for constant void ratio shown Fig and illustrates sim
is
in
9
.
plified ralations and definitions the above mentioned strength components

of
suggested by TERZAGHI this diagram .

As
obtained by selecting points

is
1938
,
(
with the same void ratio

or
on
failure the normal consolida

at
water content
and reloading branches diagram similar
of
that shown
to
tion rebound
in
,
Fig Further details of the method are shown Fig

23
by in
6
.
.
11 .
energy component Eqs

12
td
The surface defined and and

is
is
in
,
troduced here order its influence on the other components

to
demonstrate
is in
,
but
reference made
the comments thereon the foregoing section The
to
in
dilatation component
Td
subtracted from the shear strengths

Tf
obtained
is
,
,
the tests and represented The cor

by
the shear strength line AB Fig

in
in
. 9
.
.
rected shear strength line corresponds Eq

of
14
the left side

to
The
A
B
,
',
'
general effect of consideration the surface energy component slight

of
is
a
small decrease of
of
the inclination the shear strength line and

of
increase
a
by
zero intercept According experiments GIBSON

of the of its
1953 introduction
to
),
of (
.
energy correction causes degrees of

an
the full surface increase

to
1
inclination of the shear strength but the influence may greater

be
case
in
,
undisturbed clays td decreases with increasing test duration

of
The value
.
and probably zero for tests very long duration

of
also zero for all

un
is
is
It
.
or
drained constant volume tests

.
The effective friction component

of
function the effective stress

td
is
,
defined by the shear strength line obtained when

of
on and varied while

is
is
the cohesion and rheological components remain constant and corrections are
for
made variations of the surface energy component any The shear

if
,
strength line thus obtained usually straight and its angle

of
is
inclination de
,
,
be
called the effective angle

of
may internal friction The friction component

.
then expressed by
is
To = of tan = ( or - u ) tan : (15 )
where the pore -water pressure , u , appears as a part of the friction component
and may be a time - dependent variable . The angle de depends on the compo
sition of the clay and possibly also on the orientation of the flaky mineral par
ticles or zones of stratification with respect to the surface of failure , but !
has been found to be practically independent of the void ratio . Until further
test data become available , it is assumed that We is independent of time or
the rate of deformation .
The effective cohesion component , ce , represents the strength caused by the
intrinsic forces , and for the purpose of this paper it is defined as the zero in
tercept of the above - mentioned corrected shear strength line , A ' B' . The ef
fective cohesion can be expressed as a function of the void ratio , or the water
content in case of fully saturated clays ; but the coefficients in this function
vary with the soil constituents , including the ions in the pore water or adsorbed
on the surface of the clay particles , with the geometric structure of the clay ,
and possibly also with temperature , although the latter may primarily influence
the rheological component . It should be noted that the void ratio and geometric
structure referred to above are those existing at the moment of failure .
The ultimate cohesion component , Cu , is the intransient part of the effective
cohesion component or the value which ce approaches with increasing time or
decreasing rate of deformation . Methods for estimating Cu are discussed in
Section 7 . This component can be expressed as a function of the void ratio ,
and the coefficients in this function vary with the constituents and geometric
structure of the clay .
The rheological component , cy , is the transient part of the effective cohe
sion component or the intercept OB ' in Fig . 9, and it decreases to zero with
increasing time . Methods for estimating Cy are discussed in Section 7 . It is
assumed that the rheological component is a function of the void ratio and the
test duration or the rate of deformation , and the coefficients in this function
vary with the constituents and geometric structure of the clay and with temper
ature . It should be noted that the decrease in strength of an undrained test
specimen in part is caused by an increase in pore -water pressure with time ;
BJERRUM , SIMONS , and TORBLAA (1958 ) . Such an increase in pore - water
pressure causes a decrease of the effective friction component , Eq . 15 , and the
rheological component is in this case defined as the difference between the
total decrease in strength and the decrease of the friction component .
This concept or definition of the rheological component is admittedly a prac
tical expedient until results of further investigations of the rheological proper
ties of clays become available . Modifications of the definitions of several com
ponents will be required in case the results of future research should show that
the transient part of the shear strength affects not only the intercept but also
the inclination of the shear strength line in Fig . 9 .
Theresidual shear strength . Most clays decrease in strength after failure .
The ultimate value of the strength , called the residual shear strength and des
ignated by Tr in Fig . 4 , depends not only on the type , structure , and state of
consolidation of the clay but also on the type of loading , rate of deformation ,
drainage conditions , and the elapsed time after failure ; see Section 8. The re
sidual strength of some clays is attained only after very large deformations ,
and its dertermination then requires use of special equipment , such as the tor
sion shear apparatus .
the remolded clays shear strength after failure
of
is
case decrease
in
In
,
a
primarily pore water pressure and
by
caused transient increase thixo
in in
a
a
-
tropic loss strength which regained time upon cessation of the defor
in
is
,
mations However permanent decrease or increase strength may be
in
a
,
.
caused by void ratio after failure Most undis

an
increase or decrease
in
.
turbed clays are subject transient and permanent decrease
in
to
both
a
a
strength after failure The permanent part of this decrease strength
in
is
.
primarily caused
by
of
alteration the soil structure
.
Significance of the components The components of the shear strength de
.
the foregoing paragraphs are primarily parts mathematical ex
of
fined
in
pressions shear tests and they may be called physical

of
of
of
the results
,
phenomenological components which have not yet been definitely identified
with specific intrinsic forces Some investigators have suggested that there
.
essential difference between the physicochemical forces which cause

no
is
friction and cohesion remolded clays which case the cohesion may
in
in
”
”
"
,
as
be
of
viewed the result secondary and residual changes the spacing and
in
arrangement the clay particles possible that the result of further
of
re
It
is
.
search into the physicochemical and rheological properties clays may sug
of
gest modifications of the definitions and
of
other components
or
introduction
/
.
Consequently the writer prefers retain the terms effective angle
of
to
in
,
"
ternal friction and effective cohesion rather than using the terms true
"
"
"
"
angle of internal friction and true cohesion
of
Further delimitations the
”
.”
definitions are discussed the sections dealing with the individual strength
in
components
.
TESTING PROCEDURES AND SOURCES OF ERROR

3
.
Properties and Preparation of Soils Tested

The writer research dealt primarily with the shear strength calcare
of
's
ous silty clay from the vicinity of Vienna Austria

of
minor series tests

,
A
,
was also performed fat Tertiary clay from Rögle Klint

on
on
of
the shore
,
a
Little Belt Denmark The average index properties and coefficients

of
these
,
clays are summarized Table Both clays are subject appreciable thixo
in
to
.1
tropic changes strength which were investigated by means of the liquid

in
limit device and the Swedish cone test water contents slightly greater and
at
less than the liquid limit

.
Properties
of
Table Index Soils Used Tests

in
1
-
.
Vienna Little Belt

Property Clay Clay
77 _
cm
Specific gravity soil particles

76
of
2
,
.
.
/
Liquid limit 127

22
Piastic limit
Plasticity index 91
, mm
77
23
Clay size fraction 002

<
),
0
(
.
compression index for plot

, /cm In
2
.6
8
=
.
e
4
B
')
(
kg
84
10 07
void ratio for

=
2
.0
1
=
.
o
.
'
e
cm
of
94
10
coefficient
7
-
consolidation sec 145

x
-6
-
x
?
/
Complete consolidation and permeability tests were performed with con

solidometers , using a molding water content slightly greater than the liquid
limit . The virgin branches of the pressure - void ratio diagrams were also de
termined by means of the shear boxes , duplicating the conditions under which
the specimens for shear tests are consolidated . That is , molding water con
tents slightly lower than the liquid limit were used ; test specimens were con
solidated under various normal pressures , and the water contents in the cen
ter of the test specimens were determined upon completion of the consolida
tion . The results obtained for Vienna clay are shown in Fig . 10. The com
pression index B refers to the void ratio versus natural logarithm of pressure ;
see Fig . 18 . Similar results were obtained for Little Belt clay , and it is to be
noted that the virgin branches of the e - log o' diagrams are straight in all
cases , and that the diagrams obtained by tests in the shear boxes lie below
those obtained by means of the consolidometers because the molding water
content was greater in the latter case .
The soils were thoroughly remolded at a water content slightly lower than
the liquid limit , and each batch of remolded soil was stored for three to four
weeks before being used in order to obtain a more uniform distribution of
water content . Portions of a batch removed for preparation of test specimens
were again remolded to remove the thixotropic regain of strength during stor
age . Nine batches of Vienna clay were prepared during the research , and there
were slight differences in shear strength characteristics of the individual
tan
batches ; the minimum value of was 485 and the maximum value 514
,
0
on 0
's
.
which must be taken into consideration when comparing results of tests
material from different batches The available amount of Little Belt clay was
of .
small and batch consisted clay which was remolded again after being
III
,
used for the primary series tests and then utilized for supplementary
of
tests
.
Shear Testing Equipment
The principal strength tests were performed means of the Krey direct
by
shear apparatus and the Terzaghi shear boxes accommodating cm square

10
a
-
test specimen Investigations of the plastic deformations before failure and
in .
the changes shear strength after failure were made with torsion shear ap
a
paratus and ring haped test specimens having

of
11
95
an
outside diameter
.
-s
Fig 11 This equipment

95
cm and inside diameter of described

is
cm
in
in
5
,
.
detail earlier papers HVORSLEV 1937 Both types of equipment

to
in
1939
of ;
)
.
were designed for use incremental stress loading torsion shear appara
A
.
tus with both controlled stress and controlled strain types loading were later
of
built by the USAE Waterways Experiment Station HVORSLEV and KAUFMAN

;
The torsion shear apparatus Fig

11
1952 can be used either with divided

,
,
(
)
.
confining rings or with solid confining rings the first case failure occurs
In
,
.
midheight of the test specimen as the usual box shear test With solid
at
in
,
rings failure occurs short distance below the upper loading plate and this
a
,
arrangement was used for most tests of long duration since eliminates the
it
stress concentration and possibility leakage through the joint between the
of
at
upper and lower rings the arrangement with divided confining rings
in
Test Specimens and Failure Deformations

Investigation the distribution of water contents normally consolidated
of
in
test specimens showed that the water content the center than near
is
lower
in
the top and bottom surfaces This distribution may be caused either by ab
.
water from the porous stones during dismantling the equipment

of
of
sorption
test and or by
of
completion
at
of
concentration stresses and strains the

in
a
a
/
Special consolidation tests

of
center the test specimen the shear boxes

in
.
COEFFICIENTOF PERMABILITYIN CM/MIN

3 4 5 6 7 8 9 10 ' 2 3 4 5 6 7 8910 6 2 3 4 5 6 7 8910
e
VOIDRATIO
VOIDRATIODIAGRAMS
ANDPRESSURE
PERMEABILITY
52
44
W
%
STARTOF TEST:
=
1
e
,
.
IN
TEST CONSOLIDOMETER
II COEFFICIENTOF PERMEABILITY
=
B
7
.7
/
AT 10°C
11
. 0
IN
TEST SHEARBOXES
=Y
8
.4
B
06
TRIAXIALTESTSBY RENDULIC
28
W
STARTOF TEST
.4
-
LT
3 OS
LOGO 05
'
10
10
8901
3
4
5
6
7
8
9
3
4
' 5
6
/ 78
9
4
5
6
7
IN
EFFECTIVEPRESSURE KG CM2
O
No
and Permeability
10
Fig Vienna Clay

on
Confined Consolidation Tests

V
.
VERTICAL LOAD VERTICAL LOAD

TORQUEWHEEL
ANNULAR PISTON SOIL

DENTATED DENTATED
PLATES SOIL PLATES
>
CONFINING RINGS
BASE
UNDIVIDED RINGS DIVIDED RINGS
Fig Torsion Ring Shear Apparatus

11
.
indicated that the water content near the top and bottom surfaces of a normally
consolidated test specimen may be increased by an amount corresponding to
0. 5 to 1. 5 per cent of the dry weight of the soil during the dismantling of the
apparatus . The thickness required to eliminate post - test changes in water
content at the center of the test specimen depends on the consolidation charac
teristics of the soil . Specimen thicknesses chosen on the basis of these re
sults were 2 . 2 to 2 . 4 cm for Vienna clay and 1 . 8 to 2. 0 cm for Little Belt clay .
Data obtained in tests with specimens of various thicknesses show that the
horizontal and vertical deformations during a shear test do not increase
linearly with the thickness of the test specimen . It is estimated that in box
shear tests the effective thickness of a 2 - cm - thick test specimen is only about
1 . 5 cm . The shear deformations of a thin test specimen are more uniformly
distributed than those of a thick specimen , but the water content in the failure
zone of a thin test specimen cannot always be determined with adequate ac
curacy because of absorption of water during dismantling of the apparatus .
The shearing forces were transmitted to the test specimen through dentated
porous stones , with teeth 2 . 6 mm high and spaced 4 .6 mm apart . The internal
deformations in a test specimen were delineated by coloring thin vertical zones
during consolidation and slicing and drying the test specimen after completion
of the shear test . Fig . 13 - A shows the pattern of internal deformations for a
shear displacement of 4 mm . The displacement at failure was 3. 5 mm , and it
was estimated that the zone of nonuniform deformations or progressive failure
then extended over a 1. 5 -cm - wide zone at the ends of the test specimen . Pro
gressive failure also occurs in a zone along the sides of the test specimen ,
and colored sections near the sides indicated that this zone was about 0 . 75 cm
wide . Fig . 13 - B shows the deformations of a similar test specimen after a
shear displacement of 8. 5 mm , and it may be noted that inclined planes of fail
ure now have been formed , probably because the lateral confining stress at
the upper and lower rear ends of the test specimen had been decreased or
eliminated during the large deformations after failure . Such inclined zones of
failure were much less pronounced or absent for strongly overconsolidated
test specimens .
Variations in water contents of the soil near the teeth were investigated in
special tests with porous stones having very large teeth . Corrections were
made for water absorbed during dismantling , and it was found that the soil
around the points of the teeth was consolidated to considerably lower water
contents than the average for the test specimen . This zone of excess consoli
dation extended to a height approximately equal to half the distance between
the teeth .
Tests were also made with the induced shear zone slightly above the teeth
of the lower porous stone . The shear strengths obtained were nearly identical
with those for the normal test arrangement when the test specimens were nor
mally consolidated , but they were 8 to 10 per cent lower than those obtained
with the standard arrangement when the specimens were strongly over
consolidated . It was also found that failure occurs outside the above -mentioned
zone of excess consolidation near the teeth . These results were confirmed by
tests with the torsion shear apparatus . This difference in strengths obtained
with the induced failure zone in the center of the test specimen and near the
teeth of the dentated stones must be taken into consideration when comparing
test data obtained in box shear tests with those obtained in torsion shear tests
with solid confining rings .
SAMPLES FOR WATER CONTENT
X =DISPLACEMENT
A - DISPLACEMENT X =4. 0MM
4 5
L = 10CM B - DISPLACEMENT X =8.5 MM
EFFECTIVE AREA
= L (L - X)
NORMAL STRESS O =2.0 KG / CM2
0 = PA .EP T = , SAS DURATION TS =88MIN

FAILURE DISPLACEMENT 3.5MM
Fig . 12 Stresses and Water Fig . 13 Internal Deformations

Contents - Direct Shear Tests in Direct Shear Tests
IC - - - - -- -- - WATC CASE I - PARABOLIC VARIATION
AVERAGE RD =1 - 1 - )
STRESS
CASE II - LINEAR VARIATION
=1 - 11
TUITSIINID 1.00
0.75 CM
1. 5 CM CASE I
CM EFFECTIVE AREA
-L
10
A = 96. 5 CM2
0.90
DISTURBED AREA CASE II
Ab = 40.0 CM2
IVILIZATIINIMIZ 0.80
- X = 9.65 CM 0. 5 T / T 1. 0
Fig . 14 Influence of Progressive Failure in Direct Shear Tests

Finally , tests were made with nondentated porous stones having either a
medium coarse - or a medium fine - grained structure . The results obtained
with coarse - grained stones were nearly identical with those obtained with
dentated stones , but the shear strengths obtained with fine - grained stones
were slightly lower , and there were indications of slippage at the surface of
the fine - grained stones . These tests were made with normally consolidated
test specimens , and a greater amount of slippage can be expected in tests
with strongly overconsolidated test specimens, for which the normal stress
is small compared to the shear strength . Finely dentated stones will prevent
such slippage and are preferable to stones with relatively large teeth .
Calibration and Sources of Error
The calibration of the box and torsion shear test equipment covered not
only lever ratios and mechanical friction in the equipment and dials , sidewall
friction for various stress conditions , but also the normal stresses and fric
tion acting on the displaced ends of the box shear test specimen . The result
ing corrected normal force and shear force were applied to the effective
cross - sectional area at failure , Ae in Fig . 12 , and the errors in the average
values of o' f and Tf thereby obtained are believed to be smaller than 0 .01 kg /
cm ? or 1. 0 per cent of the stress , whichever is larger .
The influence of progressive failure on the strength obtained by box shear
tests may be estimated as shown in Fig . 14 . As mentioned above , the dis
turbed zone at the ends of the test specimen has a width of approximately 1. 5
cm , and that along the sides a width of about 0. 75 cm . The area of the dis
turbed zone is then 40 cm2, and the total effective area 96 . 5 cm2 for a dis
placement of 3. 5 mm at failure . The minimum value of the shear strength is
designated by Tb , the maximum value by Tc , and the average value by Ta. The
variation of strength within the disturbed zone may follow a parabolic curve ,
Case I, or be linear , Case II . The relation between TalTc and 76 / Tc may then
be determined as shown in Fig . 14 - C . For Vienna clay , Th is approximately
equal to the residual strength orth / T , = 0. 80 ( see Fig . 35 - A) for which Tat .
is 0. 96 to 0. 97 . The residual strength of Little Belt clay is not reached before
very large displacements have taken place ( see Fig . 35 - B ) and Tb / Tc is esti
mated to be about 0 .60 for which T / T, is 0. 92 to 0 . 94 . Corrections were not
made for the influence of progressive failure in the box shear tests , and the
values of Tf, determined as shown in Fig . 12, represent the average values ,
Ta , in Fig . 14 .
A theory for determination of the stress distribution in the ring - shaped test
specimen of a torsion shear test is presented in previous publications , HVOR
SLEV ( 1937 , 1939 ) . It was found that the maximum shear strength at failure is
0 to 3 per cent higher than the average maximum shear strength , depending
upon the shape of the stress - strain curve . This theory does not consider the
influence of stress concentration at the joint between the upper and lower rings
or at the teeth of the porous stones . As mentioned earlier , values of the shear
strength obtained by torsion shear tests agree well with those obtained by box
shear tests in case of normal consolidation but are 8 to 10 per cent lower for
strongly overconsolidated test specimens .
The above -mentioned influence on nonuniform stress distribution and pro
gressive failure is to a large extent eliminated in a shear box developed by
ROSCOE (1953 ) . Test specimens in most direct box and torsion shear equip
ment so far developed are confined laterally and are subject to the influence of
sidewall friction . Corrections can be made for the influence of the sidewall
friction , butthe magnitude of the lateral stresses cannot be controlled . It was

found in box shear tests with normally consolidated Vienna clay that the lat
eral pressure corresponds to a coefficient of earth pressure at rest Ko =
,
0 .67 . However , ,
the value of K increases with the degree of overconsolidation ,
because of residual lateral pressures , and the total lateral pressure may be
greater than the vertical normal pressure for strongly overconsolidated test
specimens . These large lateral pressures have some influence on the void
ratios , the failure conditions , and the width of the shear strength hysteresis
loop , Fig . 6 .
Several investigators have proposed that the confining rings in torsion
shear test equipment be eliminated and that the test specimen be covered with
a thin rubber or plastic membrane or with grease and subjected to known lat
eral pressures as in a triaxial compression test . It may be possible to find
membranes or coatings which are satisfactory for this purpose , but the test
specimen will still be subjected to lateral restraint by the end plates . On the
other hand , at short distances from the end plates, the test specimen may
undergo changes in outside and inside diameters during a test. These changes
are difficult to measure and have a significant influence on the torsional shear
resistance of the specimen .
Testing Procedures
Material from the main supply or batch of remolded clay was again re
molded and placed in the shear boxes to a thickness which would produce the
desired thickness of the test specimen after consolidation . The normal loads
were increased in increments similar to those used in standard consolidation
tests . About one week was allowed for normal consolidation , two weeks for
simple overconsolidation , and three weeks for cyclic overconsolidation . The
degree of consolidation thereby attained was well beyond that corresponding
to the end of the primary consolidation . The normal loads were removed for
a few minutes while transferring the shear boxes from the consolidation load
ing bench to the shear testing apparatus . This transfer was made six to twelve
hours before starting the actual shear test .
The shear loads were increased in increments which varied from 5 per
cent of the estimated failure load at the start of a test to 1 or 2 per cent of the
estimated failure load near the end of the test. The time interval between load
increases for slow tests was varied from about 10 minutes at the start to 30
minutes or several hours near the end of a test. An example of a time - loading
curve is shown in Fig . 15 , where Tf is the actual duration of the test . The
equivalent duration , Ts , for a constant rate of load increase was computed by
the geometric condition that the area OBTs is equal to the area OATf, but it is
believed that the actual equivalent test duration , producing the same degree of
consolidation , is greater than the values of Ts thus computed .
The equipment was dismantled as quickly as possible upon completion of a
test . A strip of soil 2 cm wide and 3 mm thick was cut from the center of the
test specimen . This strip was divided into five sections , and the water content
was determined for each section , Fig . 12 . The water content in the failure
zone was computed as the average of the water contents of sections 2, 3 , and 4 ,
since the water content of sections 1 and 5 often was changed by absorption of
water during dismantling of the equipment .
Influence of the Duration of Tests
influence of the test duration on the shear strength of Vienna clay was
The
investigated for various consolidation conditions and thicknesses of the test
specimen . Examples of the results obtained are shown in Fig . 16 . Shearing
stresses cause development of positive pore -water pressures in a normally

consolidated soil . These pressures and the void ratio decrease whereas the
friction and cohesion components increase , and the rheological component de
creases with increasing duration of the test . It may be noted that the shear
strength decreases slightly when T , is increased from about 1000 minutes to
2320 minutes . Another series of tests , performed with the torsion shear
appa
,
ratus shows a decrease in strength of only 1 per cent when Ts is increased
from 1300 minutes to 36 days . The results of tests on strongly overconsoli
dated Vienna clay are shown in the upper part of Fig . 16. In this case the
shearing stresses caused development of negative pore -water pressures ,
which were equalized with increasing duration of the tests . Consequently , the
friction , cohesion , and rheological components all decreased with increasing
duration of the test . The maximum value of Tg in this test series was only
515 minutes , and a further slight decrease in strength would undoubtedly have
been found for larger values of Ts .
Based on the test results shown in Fig . 16, an equivalent test duration Ts =
600 to 800 minutes was used for slow , drained tests on Vienna clay . The
actual degree of consolidation obtained cannot be determined directly from the
test results because of the influence of the rheological component combined
with slight variations or irregularities in water contents at failure . However ,
GIBSON and HENKEL ( 1954 ) have developed a very useful theory for determi
nation of the required duration of drained tests or , conversely , the degree of
consolidation obtained for a given test duration . For Ts = 700 minutes , H
( thickness of the test specimen ) = 2. 2 cm , and cy (the coefficient of consoli
dation ) = 87 X10 - 4 cm2 /min for normal consolidation , the theory furnishes the
following values of the degree of consolidation :
Average for entire test specimen , U . = 1 - ( H /2) 4 / 30TS

T = 0 944 .
901
At center of test specimen , U = 1 - (H / 2 )2 /20 To = 0 .
S
by
the theory
on
As mentioned the authors furnishes results which are the
,
the assumption that the shear stresses and

on
safe side because based

is
it
deformations are uniformly distributed over the entire height the test speci
of
men The writer found that the deformations and changes water content are
in
.
greatest fairly thick test

of
the center and that the effective thickness

in
specimen considerably smaller than the actual thickness and believed

is
is
it
,
that the degree Vienna clay

of
on
consolidation attained slow drained tests

in
per cent normal consolidation and greater for over

of
95
was least case

at
in
consolidation because the coefficient of consolidation much greater

is
then
,
than for normal consolidation

.
The supply Little Belt clay was very limited and only
of
few tests for de

a
,
the influence of the test duration were performed The results

of
termination
.
Fig
of of
of
some these tests are shown increase

17
an
Other tests show that

in
.
.
Ts from 970 minutes 2600 minutes causes less than per cent increase
to
shear strength On the basis of these results

Ts
was decided
to in
use 800
to
it
=
.
1600 minutes for slow Little Belt clay

on
drained tests However consid

,
,
.
ering the values

of
the consolidation coefficient Table and the thickness of

,
1
,
cm
the test specimen

of
on
slow tests Little Belt

to
the duration
=
H
1
8
,
2
0
,
.
clay should be about twelve times that used for Vienna clay order obtain
to
in
the same degree of equilization excess pore water pressures failure

of
at
-
This apparent contradiction between test results and theory was not discovered
MAX SHEAR LOAD
ACTUAL LOADING AREA OA TA =AREA OBT

CURVE
ACTUAL DURATION
EQUIVALENT CURVE
Ts
EQUIVALENT DURATION
=
UNIFORM RATE
OF
Tf
TS
TIME
Fig Procedure and Equivalent Test Duration

15
Loading
.
26
.0
CYCLIC
TMIN
DIRECT SHEAR OVERCONSOLIDATION
KG KG
CM2
=
O
5
.0
OF
42
/ /
%
100 KG CM2
=
WEIN
4 >
%
4
0
T
(
/
.
CM2
=
p
.5
0
o
'
26
CM
TFIN
=
H
.
2
.5
25.
0
26
.0
26.
25
1
Ts
== 23200
25
.7
TMAX
25
25
.7
84
26
.
01
.
26
26
We
0
..1
.
NORMAL CONSOLIDATION TORSION SHEAR
TE
OF DIRECT SHEAR TIME

252528 W
TA
KG
% CM2
=
'y
0001 63
2
.0
o
MIN
/
4
.5
. .0
4
28
.4 .4 .
WIN CM
=
8
H
.5
2
.4
MIN 1300
TFIN
. 1
36
22
100
DAYS
1
KG
(Tp
97
CM2
%
=
0
/
)
.
500 1000 1500MIN

.Fig
16
of
Influence Test Duration for Vienna Clay
WIN
%
TMAX
1850MIN
Ts
50
50
.1
.2
OF 49.
6
%
TRIN
70 KG
522 NORMAL CONSOLIDATION 01 CM2

% =
) 5
= .0
/
CM
75
KG CM2
=
TF 100
H
1
.
/
.
(
500 1000 1500MIN
Fig
17
of
Influence Test Duration for Little Belt Clay

.
before all tests had been completed , and a definite explanation thereof requires
additional tests . The influences of the rheological component and the remain
ing excess pore - water pressures at failure undoubtedly compensate each other
to some extent , but it is not certain that the influence of the rheological com
ponent alone can explain the test results shown in Fig . 17 . Reference is made
to similar results recently obtained by BJERRUM , SIMONS , and TORBLAA
( 1958 ) .
Summary : Reliability of Test Results
It is believed that errors in measured shear and normal stresses , adjusted
in conformity with the calibration of the equipment , is less than 0 .01 kg / cm2
or 1. 0 per cent of the stress . However , corrections were not made for the in
fluence of the progressive failure ; see Fig . 14 .
The dilatation or surface energy component was not considered because the
concept of this component had not been developed at the time of the tests . Suf
ficient data for computation of the component are not available , but the com
ponent determined by Eq . 11 would be very small because of the loading pro
cedure and relatively high rates of shear deformation at the time of failure .
The shear strengths determined include the rheological component , which
is relatively small in the case of slow , drained tests on Vienna clay . Estima
tion of this component is discussed in Section 7.
is believed that the degree of consolidation attained in slow , drained tests
It
on Vienna clay was at least 95 per cent . The influence of the remaining excess
pore -water pressures is in part compensated by neglecting the influence of
progressive failure in evaluation of the test results .
In spite of the test data shown in Fig . 17 , it is probable that the duration of
the slow , drained tests on Little Belt clay was too short and that appreciable
excess pore -water pressures existed at the moment of failure and affected the
computed values of the strength parameters . However , the errors may be rel
atively small because of the compensating factors , and the values obtained are
probably significant in a qualitative sense .
4. CONSOLIDATION CHARACTERISTICS AND SHEAR STRENGTH
The Equivalent Consolidation Pressure

A change in void ratio or water content causes a change in the shear
strength of a clay, and the consolidation characteristics of a clay form a part
of a complete expression of the shear strength . The virgin branch of a semi
logarithmic plot of the consolidation diagram is usually straight ( Fig . 18 ) and
can be expressed by the equation
e = e - C ( 16 )
where Cc the compression index and oo is the consolidation pressure cor

is
responding Expressed in terms of natural logarithms and using the
to eo
compression index B , originally defined by Terzaghi , the equation is
e = € - } in ) ( 17)
Eq . 17 was used in former papers and will also be used in the following devel
opments since it simplifies the mathematical expressions .
The equivalent consolidation pressure , de , corresponding to the void ratio

e is defined as the pressure o for a point on the virgin branch of the consoli
dation diagram with the ordinate e, or by rearranging Eq . 17
Ble
exp [
18
od = o
)
-
)
]
(
oe
The graphical determination Fig for the points
18
shown and
of
in
is
F
.
Gon the rebound and reloading branches of the diagram but the method and
,
apply any value of Gs may be used for
18
Eq The substitution
to
=
w
e
e
/
.
.
fully saturated clays where Gs the specific gravity the soil solids In
of
is
,
.
represented
by
quantity with the same
of
de
or
troduction
is
means that
or a
e
dimensions as Tf and which facilitates graphical analytical determina

of
,
the failure conditions for clays Representation byde

of
of
tion be
is
to
e
.
considered as mathematical expedient and not necessarily dependent
on
the
a
actual pressure void ratio diagram will be shown later that any line paral
it
;
-
lel the virgin branch the consolidation diagram Fig

of
18
be
used for
to
can
of ,
,
.
determination of The graphical determination Fig
18
de
de
shown was
B in
.
.
by
first used TERZAGHI and JANICZEK see TERZAGHI 1931 their
in
,
,
-
(
)
investigation of the unconfined compressive strength of clays
.
The clay lies below the virgin
of
state overconsolidation when

is
in
e
,
a
'o)
diagram or when
of
smaller than
oe
branch the consolidation and the
is
o
,
'
by
degree of overconsolidation defined

is
/0
19
ne = 0
(
)
The largest effective consolidation pressure
or
which soil sample test
is to
specimen has been subjected Fig preconsolidation

18
on
called the
in
,
,
.
or
pressure according CASAGRANDE and WILSON the prestress

to
1953
,
),
(
and
mg
20
=
/0
(
)
0
the prestress ratio

is
Graphical Presentation
of
Shear Test Results

The principal results obtained Vienna clay and
on
slow direct shear tests

in
Little Belt clay are shown Figs

of
20 21
22
and Some the data for Vienna

in
clay are also shown Figs

19
and where water contents are plotted versus

in
or .
the logarithms presenting the results

of
Tf
of
shear
of
of
This form
,
,
'os
tests was originated during the Cooperative Triaxial Shear Research Program
the Corps Engineers RUTLEDGE 1947 and
of
by
, of
was later used

it
,
,
(
HENKEL 1958 1959 1960 presenting results strength tests

of
on
both nor
in
,
(
mally consolidated and overconsolidated clays

.
The writer used the equivalent consolidation pressure express the

de
to
,
relation between void ratio and shear strength and equivalent pressure curves
,
for conditions the start and the end of the shear test are shown Figs
at
at
in
should be noted that the equivalent pressure curves for void

22
21 and
It
.
ratios at the end of the consolidation start the shear test proper are based
or
, of
on
data obtained standard consolidation tests since consolidation tests

in
in
Fig
11
the shear boxes were confined normal consolidation As shown

to
in
,
.
there some difference the position and slope of the virgin branches of the
is
in
consolidation diagram obtained consolidometers and shear boxes primarily

in
the molding water content

of
because differences These differences are

to
in
large extent eliminated when the consolidation diagrams are transformed into
a
equivalent pressure diagrams

.
VOIDRATIO
In ( !)- .- Geloo ) IN
W
WATERCONTENT
70
04
=
/;
=
1
0
.
n
9
-
(TF,
EXP
6 ace
30
'=
taw
[
.
o
%
CONSTANT)
=
2
(
.
6 ) W
WA
6
SLOW
(o'f, .)
QUICK
lo
TESTS
TESTS
)
START
VIRGIN WA
AT
OF
FAILURE
24
SHEAR
SLOW
W T T
88
( (
)
RELOAD
=
0
-
.
QUICK)
+
e
TEST
CONSTANT
-
REBOUND
|
-- 21
REMOLDEDVIENNACLAY
-- Clio 1B
15
/ 10
ob
0
0
IN 7
o
3
4
5
o
2
0
.7
1
0
.3
0
.5
0
.2
'
'
KG
PRESSURE
LOG CONSOLIDATION LOG STRESS CM2
Fig Equivalent Consolidation Fig Strength and Water

19
18
Shear
.
.
Pressure Content Normal Consolidation

-
10
15
KG
IN
LOG CM2
5
7
2
3
1
.0
0
.5
0
.7
O
0
.2
0
.3
0
.1
,
%
%
IN
IN
W NORMALSTRESS W
NORMALCONSOLIDATION
WATERCONTENT
WATER CONTENT
WATERCONTENT
)Wp
04
W
(0
)
't
w
.
.
(
L2 RELOAD ENDCONSOLIDATION
--
OF
START SHEARTEST
RELOAD REBOUND
REBOUND
SHEARSTRENGTH
WATERCONTENT
LULUI
KG
lle
91
CM2
4
0
'
/
.
.
-
20
osos
KG
10
IN
LOG CM2
.22
T
5
,
2
3
.7
1
.0
0
.5
0
0
0
.3
/
0
.1
Fig Shear Strengths Vienna Clay

20
and Water Contents

of
.
%
IN
WATER CONTENT START OF TEST /
ATER CONTENT
KG
N
30
Aw CONSTANT
%
=
.2
PRESSURE
WATER CONTENT
W AT FAILURE
EQUIVALENT
CMS
,T
KG
IN EQUIVALENT PRESSURE
AT FAILURE
SHEAR STRENGTH
EQUIVALENT PRESSURE
START OF TEST
CONSOLIDOMETER
(
SHEAR STRENGTH
-,C.
0
24
He 050
= =
. 0
TIEDEMANN
.
Mr 434
0
COHESION LINE
MM
SWELLING
IN
MOVEMENT OF SAMPLE SURFACE
CM
HEIGHT OF SAMPLE
=
H
2
4
.
SETTLEMENT
CONSTANT
+
1
e
T23 A4 A5
KG
IN
-
NORMAL STRESS KG CM2

-
;
IN
NORMAL STRESS
/ /
Fig Vienna Clay

of
21
on
Results Slow Direct Shear Tests

V
.
The shear strength line for normal consolidation is straight in case of

Vienna clay and slightly curved for Little Belt clay . In discussions of the
writer ' s original papers it has been suggested that the curvature of the shear
strength line for Little Belt clay is caused by insufficient duration of the tests .
As indicated in Section 3, it is probable that the duration of tests on Little
Belt clay was too short and that excess pore -water pressures existed at the
moment of failure . However , the writer does not believe that this could cause
the curvature of the shear strength line , since the coefficient of consolidation
is practically constant for the entire range of pressures used in the tests .
Furthermore , the shear strength lines for normal consolidation are straight
for both slow and quick tests in case of Vienna clay but curved for Little Belt
clay , Fig . 31 , which indicates that the test duration does not influence the
shape of these lines . In contrast thereto the results of consolidation tests on
Little Belt clay are similar in form to those obtained for Vienna clay ; that is ,
the virgin branches of the semilogarithmic consolidation diagrams are straight
for both clays . The results of tests on Little Belt clay exhibit other anomalies
which will be discussed later . It is difficult to remold a very fat clay com
pletely , and the influence , if any , of inadequate remolding on the anomalies
should be investigated .
Volume Changes During Shear Tests
As first shown by CASAGRANDE and ALBERT ( 1932 ), the void ratio or
water content of normally consolidated clays decreases during a shear test .
The movements of the surface of the test specimen and the relative position
of equivalent pressure lines , shown in Figs . 21 and 22 , indicate that both nor
mally consolidated and slightly overconsolidated clays are subject to a de
crease in volume during a shear test . Consequently , the shear stresses cause
a temporary increase in pore -water pressures and the shear strength in
creases with increasing duration of the test when drainage is possible ; see
Figs . 16 and 17 .
Other tests , HVORSLEV (1937 , 1938 ) , show that interruption of a very slow ,
drained shear test on normally consolidated Vienna clay and decrease of the
shear stresses cause a slight additional decrease in void ratio . RENDU LIC
(1937 ) found that interruption of an undrained triaxial compression test on nor
mally consolidated Vienna clay and decrease of the axial stress caused a slight
increase instead of the expected decrease in pore -water pressure . That is , a
decrease in shear stresses may cause a change in void ratio or pore - water
pressure which is of the same sign but smaller magnitude than that caused by
an increase in shear stresses . RENDU LIC ( 1936 , 1937 ) also observed that the
initial volume decrease of a normally consolidated clay during a triaxial com
pression test was reduced by a slight volume increase near failure . This re
versal of the volume change near failure may have been caused by the use of
a molding water content close to the plastic limit . The above -mentioned ob
servations are discussed in greater detail in Section 9.
The test data in Figs . 21 and 22 show that the volume of strongly over
consolidated clays increases during a shear test . Consequently , the shear
stress causes a temporary decrease in pore -water pressure and the shear
strength of such clays decreases with increasing time , Fig . 16 . BUISSON
( 1936 ) also found that the volume of certain dense clays increases during a
shear test . It appears , however , that the sign of the volume change during
shear depends on the stress history and prestress ratio rather than on the den
sity or void ratio of the clay . At a certain prestress ratio , the shearing stress
does not cause any change in volume or pore -water pressures , but this critical
ū
/
CM2
%
IN
KG
WATER CONTENT
WATER CONTENT END CONSOLIDATION IN

O
PRESSURE
WATER CONTENT
. AT FAILURE
EQUIVALENT
/
KG CM2
,T 8
IN
o
SHEAR STRENGTH
EQUIVALENT PRESSURE
ã AT FAILURE
5
5 EQUIVALENT PRESSURE
3
|
START OF TEST
5 CONSOLIDOMETER
)
AI
to
o
f SHEAR STRENGTH
8
Oto
MM
IN
SWELLING
MOVEMENT OF SAMPLE SURFACE
HEIGHT OF SAMPLE CM
=
H
1
9
.
SETTLEMENT
KG
IN
NORMAL STRESS CM3

/
Fig
of
22
Results Slow Direct Shear Tests Little Belt Clay

on
.
prestress ratio is quite different for simple overconsolidation and cyclic over
consolidation , and all cyclic overconsolidated test specimens of Little Belt
clay were subject to a volume decrease during the shear tests .
The volume changes observed in direct shear tests are probably influenced
to some extent by residual lateral stresses , but the general character of the
results has been verified by other investigators and by triaxial tests on both
remolded and undisturbed clays , and these results are important for estimat
ing the short - and long - term shear strengths of clays .
The results of shear tests on normally consolidated test specimens of
Vienna clay , Fig . 21 , also show that the decrease in water content , Aw , or void
ratio , A e , during a shear test is constant and independent of the normal stress
o't . This means that the pressure - void ratio line for failure conditions is par
allel to the virgin branch of the consolidation diagram , Fig . 19. These results
were also obtained by RUTLEDGE ( 1947 ) and HAEFELI ( 1951) and have been
verified by others . The degree of overconsolidation at failure , ne , of original
ly normally consolidated clays can then be determined as follows . The void
ratio at the end of the consolidation or start of the shear test is er , and that
at failure is ef = e , - de . The normal stress remains constant during a drain
ed shear test and can , according to Eq . 17, be expressed by
21
'
Ble
exp [
(
)
Os = o = a
e
-
)]
.Eq
pressure
by
The equivalent
18
failure
at
as
consolidation defined
is
,
,
ep
de
exp Ble
22
=
)
]
-
[
.
o
(
)
and Eqs yield
19
22
23
and then
,
,
.
exp
)]ep
se
23
Blec
=
)
·
(
-
[
(
)
The following values were obtained for Vienna clay batch 4e
=
=
,
,
B
V
8
.4
nc
70
0635 and
=
,
0
1
.
.
.
The numerical value of the change void ratio during

or
water content
in
normally consolidated test specimens of Little Belt clay de

on
shear tests
creases slightly with increasing normal stress and the above mentioned rela
,
tions for Vienna clay not apply Little Belt clay HENKEL 1958 1959
do
to
,
on (
.
obtained similar results drained triaxial compression tests normally

in
consolidated test specimens of Weald clay and London clay which are com
,
parable Vienna clay and Little Belt clay was found that the change
to
in
It
.
void ratio 4e was constant for Weald clay but decreased slightly with in
,
,
creasing consolidation pressure for London clay

.
The Shear Strength Normally Consolidated Clays

of
the writer original papers the shear strength of normally consolidated

In
s
'
clays was treated as special case the equations for the shear strength
of
of
a
both normally consolidated and overconsolidated clays However the relation

,
.
between shear strength and water content void ratio for normally consoli
or
dated clays can developed without special assumptions

be
as
as
long the shear

by
be
strength can expressed

tan
24
of
Ms
of
If
=
)
*
.Eq
the
by
virgin branch diagram the change void

of
17
the consolidation and

in
,
ratio during the shear test 4e

ef
of
independent the normal effective

ec
=
is
,
,
-
stress , of . These conditions are not fulfilled by and the following derivations
do not apply to Little Belt clay . According to Eq . 24
which inserted in Eq . 17 yields
- ec ) = Ble, - es - de ) ( 25 )
and
t
(exp
(-exp
(-exp
T Be
,
Be
Be
=
)
_)
=
26
C
(
)
or
(exp
exp
Bee
27
es
To
Ae
exp
=
Bee
,
T
=
|
C
[
(
-
)
]
)
-
)
-
(
)
where C1 and C2 are constants for These equations also apply given soil
to
a
The equa .
4e
the results of slow undrained tests and for which ef
=
=
0
e
,
,
.
tions show that the shear strength normally consolidated clay can be ex
of
a
of
pressed as the test
at
or
at
the start failure

of
function the water content
,
a
and that plot of the logarithm of the strength versus water content forms
a
a
straight line Fig parallel the virgin branch

19
CC the consolidation
to
of
in
,
,
.
diagram These results have also been obtained independently by RUTLEDGE

.
1947 HAEFELI 1951 and others HENKEL 1958 1959 performed slow
,
,
(
)
.
triaxial tests both drained and undrained test specimens of remolded and
on
normally consolidated clays

of
was found that the results both drained and

It
.
by
presented single straight line corresponding

be
slow undrained tests can

a
Fig These results and other comparisons made

by
19
Henkel show
to
CC
in
that the prestress developed remolded and normally con

on
undrained tests
in
solidated clays has no effect on the failure conditions when the duration of the
undrained tests equal that used slow drained tests However pre
to
is
in
,
.
stressing may have considerable influence on the deformation characteristics

of clay
a
The shear strength normally consolidated clay can also be expressed

of
a
pressure Eqs
of
as linear function the equivalent

19
24
consolidation and
a
yield
and
28
os
Of
and tan
=
=
p
n
/n
'/
.
.
(
(
)
By
.Eq
.Eq
for
inserting the expression

Øs
tan
de
18
To
26
and
in
is
,
0
-
transformed into Eq 27
.
Quick direct shear tests Vienna clay yielded straight shear strength
on
line for normally consolidated test specimens Fig 31 Therefore equa

A
,
,
.
.
-
tions similar those representing the results of slow tests can also
be
to
devel
oped for the results of quick tests That the shear strength obtained
in
is
,
.
as
be
of
quick tests can expressed linear function the total normal stress
,
a
or the equivalent consolidation pressure exponential function

of
an
as
oe
or
,
of the water content plot water content versus the logarithm

of
of
shear
A
.
straight line DD which lies slightly the right

of
strength Fig
19
to
the
is
in
,
,
a
line CC That the quick tests yield shear strengths which are slightly
is
,
.
greater than those obtained slow tests for the same water content at failure
in
The difference represents the influence the rheological component

of
and this
,
the equivalent consolidation

be
as
difference can also expressed

of
function
a
pressure or the water content or void ratio at failure . The influence of the
rheological component is discussed in greater detail in Section 7.
5. THE EFFECTIVE FRICTION AND COHESION COMPONENTS
Theoretical Derivations
It has been shown above that the shear strength of a normally consolidated
Vienna clay can be expressed as a linear function of the effective normal
stress , of , or the equivalent consolidation pressure , c'e, or as an exponential
function of the void ratio , ef , or the water content , wf, at failure . Conse
quently , the shear strength can also be expressed as a compound function of
of and o' or ef. An inspection of Figs . 20 and 21 shows that the shear
strength of overconsolidated test specimens of Vienna clay cannot be expressed
as a unique function of either of, o , or ef. However , the similarity of the
shear strength and equivalent pressure diagrams in Figs . 21 and 22 suggests
that the shear strength of both normally consolidated and overconsolidated
clays may be expressed as a compound linear function of of and e. Hence
it is postulated
Ip = Me + Korea ( 29 )
be
the
where tan The equation may also written

=
Ø
.
30
tan
(
)
9
e
*
*
*
+
This equation has the simple form

Eq
of
the Coulomb criterion and the
if
,
3
,
.
by
expressed mathematically
be
shear strength can this equation the points

straight line with the angle ,
of
of
should lie on inclination

a
'0
T
o
,
!
./
)
/
(
'
ordinate intercept Fig shows that points representing data

24
an
and
A
K
.
Vienna clay batch straight line with very little

on
obtained tests form

in
a
,
scatter even when certain tests have been duplicated because minor irregu
of
,
the first run these tests Similar results but slightly different
of
larities
of in
values de and were obtained tests with Vienna clay batch

in
,
K
I.
tests with Little Belt clay are shown

24
The results obtained Fig

in
in
,
B
.
-
and the points also lie close straight line but the scatter greater than
is
to
a
Vienna clay and line connecting the various points still forms
of
the case
in
a
,
faint hysteresis loop and separate line with flatter slope for the points
a
representing the state

of
normal consolidation As mentioned before

is
it
,
.
probable that appreciable excess pore water pressures existed the time of
at
-
failure Little Belt clay and the numerical results are

on
drained tests
in
in
,
partially compensating errors and cannot be considered fully reli

by
fluenced
the results indicate that the curvature
of
able However the shear strength

,
.
line for normally consolidated Little Belt clay part can be explained by the
in
observation that the change void ratio during the shear test decreases with
in
increasing values
of
o
s
'.
Eq
Eq
29
By
de
of
introducing the expression last member

22
for the
in
can be written as follows

,
31
)Bey
[exp
(-exp
KO
Ble
ko
(
)
,
@
p
=
=
'
'
c
-
))
where
(exp
32
Be
cz = ko
(
)
coefficient which represents the value
or
co
; of
the function the effective
is
a
see Fig merely
26
ef
hesion component for zero void ratio This
is
=
C
0
,
.
Eq possible that the above
of
32
mathematical interpretation
is
and
it
,
a
Eq
mentioned relations change when the void ratio approaches zero
29
can
.
then be written the alternate form
in
exp
of 33
Bec
To
tan
,
%
(
)
)
C
-
=
(
This equation shows that the shear strength can o expressed as
in be
function
a
the effective normal stress on and the void ratio the plane of failure
at
the
moment of failure and that this function the stress history
of
independent
is
,
of the clay
.
WA
WATER CONTENT
CONSTANT
CONSTANT
-
-
TA
SHEAR STRENGTH
SHEAR STRENGTH LINE FOR CONSTANT

WATER CONTENT
cew
TAN
=
T
ở
-
NORMAL STRESS OF
Fig Separate Components

of
23
Determination Friction and Cohesion

.
The symbol former papers here changed commonly

cz
because
in
is
to
y
C
used as symbol for the Poisson ratio

is
.
RATIO STRENGTH SHEAR
' . , .
10°o°
17°30
.0 2
07
834 ,c 34
9
KDO
+
116
= = = 0
$ = = = € e c
NORMAL STRESS RATIO

. .2 .0 .0 .
TAN
40
.
60
0
175
.8 0
145
8
P46
100
315
0
Parameters
= = =
40
= =
= 0
= B M K
Ty B K
He
.
TO
7
0
and Cohesion
O
10
V
.60
20
CLAY
Friction
CLAY
BELT
Ha
of
60
VIENNA
LITTLE
012
Determination
' :
Combined
'
17°30
=
Ø
.0 .0
.
24
100
145
Fig
= =
K K
RATIO STRENGTH SHEAR

should be noted that Eq . 33 is a mathematical expression of the shear

It
strength which is valid without special assumptions when the points ( 0 / 0 ,
Tilo ) straight line
lie
Fig The first member
as
24
of on
on
shown the
to in
,
a
.
.
the
right side equation equivalent the effective friction component
is
,
the effective cohesion component
Ce
and the second member defined
in Tø
to
,
,
,
Fig As previously mentioned correction were not made for the influ
,
9
.
.
ence of the surface energy component and assumed that the rheological
is
, it
,
component case of slow tests negligible or included the effective
in
is
is
in
cohesion component Special assumptions discussed connection with Figs
in
,
.
.
are required when the component parameters are determined sepa

23
and
9
rately and also explain the physical meaning of the components

of to
.
Introduction the equivalent consolidation pressure mathemati
is
of o
a
,
,
cal expedient and any line parallel the virgin branch the consolidation
to
,
may be used for this purpose The change from one line another merely
to
, be .
by
multiplied
of
of
constant the slope

de
causes the values the lines

to
;
a
will not be changed but the value the intercept will be divided
of
Fig
24
in
,
,
k
.
by the same constant hence the product Koe remains constant However
,
;
.
lines which are not parallel the virgin branch diagram
of
the consolidation
to
de since this would cause the straight

of
should not be used for determination

,
to or
Fig change the apparent slope
of
24
lines become curved the lines

be to
In
in
.
by
general may
at of
oe
most expedient determine the values means
of
it
,
the line which represents the water contents void ratios failure of nor
or
mally consolidated test specimens Fig

19
line BB
in
,
.
.
Eq 29 may be written two alternate forms when the clay normally
is
in
.
.Eq
according
23
od
consolidated which case

to
and
in
=
,
,
0
n
's
.
34
Tp
he
of
(
)
=
(K
n
(
OT
35
Tp
-A
on
)
Helme
=
+
(x
)
31
Eq
of
and by use
.
35
Men exp Bec

Te
,
+
)
-
=
B
c
)
(x
-
(
comparison Eqs yields

24
34
of
and
A
het nk
As
36
=
)
(
These equations show that the shear strength normally consolidated clay
of
a
can be expressed as an explicit function of either the effective normal stress

,
or the equivalent consolidation pressure or the water failure

or
at
void ratio
,
as also shown Eqs may be used for comparison of shear

35
Section
in
4
.
strengths normally consolidated clays when

of
on
drained and undrained tests

the pore water pressures and effective stresses are unknown see Section
7
;
-
and Fig may be mentioned that oe for normally consolidated clays

32
It
A
.
.
-
nearly identical with the values of the effective principal stress

oj
at
is
failure direct shear tests and triaxial compression tests BJERRUM 1954
in in
,
(
and Fig
of
38
but not the case triaxial extension tests see Section

9
;
Vienna clay are summarized

on
The results obtained

of
tests two batches

in
Table
in
2
.
206
Table 2. -- Strength Parameters for Vienna Clay

U
Batch
Batch
B
B o no "s Is Me le к
504
122
I 8. 3 0 .85 1 . 72 0. 26°45 105
0
.
' '
116
100
315
8. 4
70
84
V 485 25°50 17°30
0
0
'
1
.
0
.
.
.
.
The Effective Friction Component
The foregoing theoretical derivations and resulting failure conditions may
primarily apply clays which have been remolded at water contents close
to
to
the liquid limit and then reconsolidated water contents between the liquid
to
and plastic limits The simple expressions for the effective cohesion com
.
ponent may not apply clays remolded relatively low water contents or
to
at
to
undisturbed clays which some of the bonds between the particles have
in
a
character which different from that remolded clays Under such condi
is
in
.
tions method proposed by TERZAGHI 1938 for separate determination of
),
,
a
(
the effective friction and cohesion parameters may be used advantage and
to
,
this method also excellently illustrates the physical meaning of the components
.
Fig The basic assump
23
The above mentioned method illustrated
is
in
.
-
tions are that the effective cohesion component constant when the water con
is
tent constant and that the rheological component also constant when the
is
is
,
deformation or test duration are constant Fur

of
water content and the rate
.
significant difference
no
thermore must also be assumed that there
in
in is
it
,
the geometric structure of the test specimens given series at the time of
a
The points on the stress water content diagram Fig
23
failure
D in
D
B
E
,
,
F
.
, .
-
on -
have the same water content failure The corresponding points
at
E
,
F
.
the stress strength diagram are determined as shown Fig 23 These

in
A
.
.
-
-
points should and usually straight line the angle of inclination of
do
lie on
a
this line de defined as the effective angle of internal friction The ordi
is
,
-
the intercept
of
Ce
is
nate the effective cohesion component for the water

, G
,
,
content wf points This operation repeated for other water con

of
is
D
E
,
F
.
tents whereby data are obtained for establishing the relation between the
ef
,
fective ochesion component and water content The graphical determination of

.
We may practice be replaced by numerical computations using the coordin

in
points trength diagram

of
ates the stress

in
D
,
,
E
.
-s
BJERRUM 1954 GIBSON 1953 and others have made many tests on re
),
,
(
by
molded clays and determined the effective angles

of
internal friction means

of the Terzaghi method general was found that the angles obtained
do
In
it
,
.
not vary with the water content beyond normal scatter of test results This
.
part the reliability

of
of
establishes the method but for full confirmation

in
the physical interpretation the results necessary compare the angles

of
to
is
it
,
with angles of inclination of failure planes compression

on
tests the
in in
d
same clays Such comparisons are discussed Section and shown that
is
it
6
,
.
fairly satisfactory statistical confirmation obtained for remolded clays

is
Theoretically also possible determine the angle measuring the

by
to
is
it
é
\
angles the failure planes and using the theoretical relation

in - of
of
inclination
,
, a,
de
90
not reliable because of considerable

2a
however the method

is
=
and also because of the influence un

of
on
scatter the individual values

a
known anisotropic properties

of
the test specimen

.
The Effective Cohesion Component

After having obtained various corresponding values
of
the effective cohesion

component and water content by the Terzaghi method one may first investigate
,
if this component can be expressed as a linear function of the equivalent con

solidation pressure or as an exponential function of the void ratio or water
content , as indicated in Eqs . 29 and 33 . An example of such evaluations is
shown in Figs . 25 and 26 , using data obtained in tests on Vienna clay .
If the values of we are determined on basis of line AA in Fig . 25 - A , cor
responding to the virgin branch of the consolidation diagram , the relation be
tween ce and de is represented by the straight line AA in Fig . 25 - B. If the
values of de are determined by line BB , which is parallel to AA and repre
sents the water contents at failure of normally consolidated test specimens ,
the line BB in Fig . 25 - B is obtained . The tangents of the slopes of the two
lines are equal to the coefficients K . The values of Kand of de are different
for the two lines , but the product ce =koe is constant for a given water con
tent . On the other hand , when values of ol are determined by means of the
line CC , which has a flatter slope than AA , the dashed line CC in Fig . 25 - B is
obtained . Unless test values of ce corresponding to low values of de are
available , the line CC gives the impression of being fairly straight and inter
cepting the ordinate axis . Similarly , the use of line DD , which has a steeper
slope than AA , for determination of de results in the dashed line DD , which
gives the impression of intercepting the abscissa axis .
In case there is doubt about the proper slope of the reference line for de
termination of de , it is better to evaluate the test results by plotting log ce
versus wf as shown in Fig . 26 . When the points straight line the ef on
lie
,
fective cohesion component can be expressed by a
Bep
(exp
37
Ce
ez
A
=
)
-
-
(
)
the intercept with the abscissa axis and
be
cz
where can computed from

of is
, cy
the slope the line As previously explained the coefficient represents
,
.
of
the theoretical value zero void ratio and

at
the effective cohesion
is
it
a
more significant soil parameter than the value of which depends the on
K
,
of
choice reference line for determination

of
of
oe
values
a
Extensive investigations of the shear strength

of
remolded cohesive soils

by
have been made BJERRUM 1954 who verified that the cohesion component
,
(
clays remolded high water contents proportional

of
but found that

at
is
to
'de
this component for clays remolded expressed by

at
low water contents

is
37
Ce
KOC
-
(
)
=
+
6
quite possible that this relation correct since the shear strength char
It
is
is
acteristics of clays remolded water contents close the plastic limit may
at
to
clays remolded
be
of
different from those high water contents

is
at
also
It
.
possible that the reference lines used

by
Bjerrum for determination of the

values did not have the proper slope and produced de diagrams
Ce
oe
to of
,
(
similar Fig This possibility enhanced by the use of

25
CC and DD
in
is
B
.
.
-
clays remolded at different water contents single test series and Bjerrum
it in
a
realized that this testing procedure makes difficult obtain proper refer
to
a
om
ence line for determination Bjerrum also used the method shown
of
in
Fig slight curvature

26
for evaluation of some of the test results and found

a
.
ap
wf log
- ce
at
of
the line low water contents which imposes limit

on
the
,
a
(
A )
plicability of Eq
37
of .
Determination Strength Parameters by Triaxial Tests

The proposed expressions for the shear strength Eqs
33
are
29
and
in
or
on
based results obtained direct boxtorsion shear tests SKEMPTON

in
and BISHOP 1954 have developed two methods for determination of the
(
)
208
AD
CB
ce
% IN CONTENT
COHESION
AT
OF
W
END
.-
CONSOLIDATION
WATER
в
\ C
D
SHEAR STRENGTH
ce WF
%
KG
CM2
0
30
WAT FAILURE 107

28
0
17 34
25
0
51 23
CONFERENCE
22
21
00
21
.0 .0 . .2 . . .0
/ . .0.0 . .0.0 .1
20
68 85 07
с
ов AN
20
0
6
0
8
1
2
3
4
6
|8
L4 0
. 0
.
4
5
6
7
9
.
10
15
2
3
8
LOGO EQUIVALENT PRESSURE
-A
-
EQUIVALENT PRESSURE LINES COHESION VERSUS EQUIVALENT PRESSURES

of
a
as
. 25
Fig Cohesion Function Equivalent Consolidation Pressures
209
PHYSICAL COMPONENTS
effective friction and cohesion parameters by means of triaxial compression
tests ; these methods are also described by BISHOP and HENKEL (1957 ) . In
the first or direct method a diagram similar to that in Fig . 23 is used , and
Mohr circles are drawn for different stress conditions but at constant water
content . The envelope for such a set of circles is identical with the shear
strength line for the particular water content or void ratio ; hence , Om = 0
= The intercept of the envelope , cr , represents the effective cohesion
component for the particular void ratio . Envelopes for various void ratios
yield the same values of 0 , but different values of ch. The second and very
elegant method is illustrated in Fig . 27 . The basic triaxial failure condition ,
Eq . 9 - A , is transformed into that shown in Eq . 9 - D. Division by 0e , the
equivalent consolidation pressure obtained from the virgin branch of a triaxial
consolidation test under all - round pressure , gives this equation the dimension
less form
loj - ole co cos % on sind .

2 Ol ? -
sing
11 - sin . + (11 -- sin
sing ( 38-A)
which is similar to that of Eq . 30 . By plotting ( o; - 09 / 20 , versus og / 0 ,
a straight line is obtained with the angle of inclination B3 , the intercept cz ,
and the equation
( 38 - B )
which also may be written
( 01 - /2
cz0
tan
38
B3
;
03
-C
=
'
)
+
o
by
cr
Values equating the coefficients Eqs

38
and are obtained and

as O
in
's
-A
.
27
38
or
shown Fig
in
,
B
,
-
sin
tan B2
cz
%
- )
1
-
(
.
38
sin
kt
od
and
.Bz
)
=
=
O
-
C
(
=
(r
'
tan
+
1
cos
The parameters Ør and correspond We and the equivalent method

to
in
of +
K
suggested for evaluation the results of direct shear tests Eq

30
The
,
method has the advantage that facilitates evaluation

of
the test results and

it
the proper average values of the parameters from scattered

of
determination
test data but should be used only when significant data for determination of
it
,
as de
are available and when the effective cohesion component can be expressed
oe
constant times
a
be
Similar equations can also

by
plotting
oi
obtained versus
0
c
-
)
/
oz
gloj which forms straight line with the equation

+
'e
a
)
Bg
38
tan
al
01
03
E
-
(
450€
=
+
-
)
)
the
Øy and
be
. of
by
cr
and values can then determined equations which are

equivalent Eq
of to
C
9
-
.
Or and triaxial tests the University of London

Kr
Values
at
obtained
in
,
,
by
HENKEL 1958 may

be
are summarized
of
noted that values quoted

Kt
It
(
.
)
30
210
c
,c
=
- Bet
20/
exp
% IN WE CONTENT
, 0 - ,0(
WATER OR
SHEAR
RATIO
0
6.
1
2
3
0
4
0
5
do .
Lozol
STRENGTH
VOID
. 0
8
;.
SIN
0
-)
.0 0
-( cosec
SIN
-1.
.0 1
.0 0$
-
206 SINO
)0
=
CONFERENCE
(Wf
C
,
60
,)503
=
67
. 00
+
TANPA
10
100 1000
(
)
AFTER BISHOP AND HENKEL 1957
KG
/
CE
IN
LOG CM2
27
of
of
a
.
. 26
as
Fig Cohesion Function Water Fig Determination Friction and Cohesion Parameters
of
Content or Void Ratio From Results Triaxial Tests

by Henkel are considerably smaller than corresponding values of K obtained

in direct shear tests , Fig . 24 , GIBSON ( 1953 ) , BJERRUM (1954 ) . This differ
ence in results of direct shear tests and triaxial tests may in part be caused
by various sources of error in the two types of tests , but it may also in part
be due to the fact that the intermediate principal stress at failure in direct
shear tests has a value between the major and minor principal stresses , which
may cause the strength to be slightly higher than that obtained in triaxial com
pression or extension tests ; see Section 9 and Fig . 40 .
Comments on Validity and Applications
When the results of strength tests are plotted as shown in Figs . 24 and 27
and form a straight line , Eqs . 29 , 33 , or 38 constitute mathematically correct
expressions for the shear strength in the failure zones irrespective of the
value of the strength components or parameters in other directions . This
statement applies also to test results expressed by the Coulomb condition of
failure . The problem of possible variations of the strength components or
parameters in other directions than that of the planes of failure is discussed
in Section 6 .
The effective cohesion component for a given clay is expressed as a func
tion of the void ratio , Eqs . 33 and 37 - A , but it is undoubtedly also a function
of the geometric structure of the clay . However , the influence of minor initial
variations in structure on the failure conditions will probably be relatively
small when the deformations tend to produce similar structures at failure .
These expressions for the effective cohesion component are based on the re
sults of tests on remolded clays with a normal range of water contents or void
ratios , and it is possible that the parameters or the form of the relation may
change when the void ratios become very small or very large . Modification of
the equations may also be required for undisturbed clays in which stronger
chemical bonds or cementation between the clay particles have been formed in
course of time .
The strengths obtained by standard tests , and expressed by Eqs . 29 , 33, or
38 , are larger than the actual long - term strengths because the test data and
the equations include a time - dependent or rheological component , but so do
the same test data expressed by the Coulomb or Mohr - Coulomb failure con
ditions . Methods for determination of the rheological component and available
data on the time - dependent decrease in strengths are discussed in Section 7.
The parameters in the Coulomb and Mohr - Coulomb failure conditions vary
with the stress history of a clay, and a graphical representation of these fail
ure conditions consists of a family of shear strength lines and hysteresis loops .
The same test data may be expressed by Eqs . 29 , 33 , and 38 , in which the
parameters are independent of the stress history of the clay . Furthermore ,
a graphical representation is reduced to a single line as shown in Figs . 24 and
27. This constitutes a considerable simplification from a theoretical point of
view , and it furnishes a simple and consistent explanation of the variations in
shear strength . However , application of the failure condition expressed by
these equations and diagrams requires knowledge not only of the pore -water
pressures and effective stresses but also of the void ratios at the moment of
failure . Advances have been made in developing methods for estimating
changes in pore -water pressures and void ratios under laboratory testing con
ditions , but these methods have not yet been verified for stress conditions en
countered in many practical problems . Therefore , the proposed methods for
determination of the effective friction and cohesion components may currently
and primarily be of interest in research and for estimating changes in the
parameters and establishing proper limits for the use of the Coulomb and
Mohr - Coulomb failure criteria in the analysis of practical problems .
6. PLANES OF FAILURE AND INFLUENCE OF ANISOTROPY
Tests on Vienna Clay and Little Belt Clay

In the previously discussed direct shear tests the orientation of the flaky
clay particles was horizontal or parallel to the induced planes of failure . Sev
eral series of unconfined compression tests were performed in order to obtain
an indication of the influence of the direction of orientation of the clay parti
cles on the compressive strength , q, and the angle of inclination of the planes
of failure , a , and also to determine whether the corresponding values of 0 = a
90 - 2a agree with pe or Øs .
Remolded Vienna and Little Belt clays were reconsolidated in the shear
boxes under a normal pressure of 5 . 0 kg / cm2 . Three types of test specimens
were cut from the reconsolidated clays in such a manner that the angle be
tween the axis of the test specimen and the planes of orientation of the clay
particles was 90° for type I, 45° for type II , and 0° for type III specimens ; see
Fig . 28 . The test specimens had a length of 4 cm and a square cross section
with a side length of 2 cm . The square cross section was chosen in order to
facilitate preparation of the specimens and measurement of the inclination of
the failure planes . Some drying was allowed to take place during the prepara
tion of the specimens , and the final water content corresponded to an equiva
lent consolidation pressure of about 6 kg / cm2. The specimens were stored in
a closed vessel for several days in order to increase the uniformity of dis
tribution of the water content . The compression tests were performed with
load increments of about 2. 5 per cent of the estimated failure load and a time
interval of 2 minutes between
load increments . The inclination of the failure
planes was measured upon completion of the test, whereupon the specimen was
weighed and dried and the water content computed .
Four specimens of each type were prepared and tested, and good agreement
of the test results was generally obtained , but there were occasional devi
ations . It was assumed that major deviations primarily were caused by irre
gularities within the test specimens or in preparation of the end surfaces ;
therefore , each figure in the summary of the test results , shown in Fig . 28 ,
represents the average result of the three tests in best agreement out of four
tests . The photographs in Fig . 28 show test specimens which have been dis
torted by drying , since the shrinkage was greatest in a direction perpendicu
lar to the planes of orientation of the clay particles .
Deformations . Appreciable deformations after failure were in some cases
permitted in order to obtain full development of the planes of failure . Type I
test specimens were subjected to considerable bulging and lateral deforma
tions after failure , which caused a slight rotation of the planes of failure . On
the other hand , deformations after failure of types I and
III
specimens con
sisted primarily sliding along the planes
of
of
Type
of
failure specimens
II
.
Little Belt clay were subjected plastic flow direction parallel the
to
to
in
a
planes of orientation of the clay particles addition sliding along the

to
in
,
planes Type III specimens of Vienna clay had tendency vertical

, of
failure
to
a
in .
splitting which some cases was more pronounced than shown Fig
28
in
A
.
the clay particles

of
of
reorientation the zones failure was not observed but

in
may occur especially slow tests

in
,
.
II
II
III
VIENNA CLAY - BATCH IX
II
LITTLE BELT CLAY BATCH

II
–
MI
VIENNA CLAY LITTLE BELT CLAY

IX
!I
11
ulin
INIVO
110
1
mini
UN
,
Mu
11
WOW
m
H
10
48
42
41
44
46
38
q
0
0
0
.
0
/
.
.
RELATIVE
109
95
93
100 100 112

STRENGTH
|
|
399
439
36°
36°30 40°15 429

'
'
(
)
(
)
(
)
149
129
18°
17°
)69
Pa
90
20
9°30
(
=
'
|
|
)
(
-
pe BATCH 17°30 BATCH 109

V
'
(
I)
(
)
Fig Compression Tests

of
28
Influence Particle Orientation

in
.
Compressive strengths . The compressive strengths of the three types of
test specimens cannot be compared directly because of slight variations in
water content and strength of the individual specimens . However , TERZAGHI
and JANICZEK ( 1931 ) have shown that the unconfined compressive strength of
a normally consolidated clay can be expressed as a linear function of the equi
valent consolidation pressure , and this was verified in the experiments by the
writer , HVORSLEV ( 1937 , 1938 ) . Therefore , the ratio q / oe can be used for
comparison of the strengths of the three types of specimens .
As shown in Fig . 28 , type II specimens have compressive strengths between
specimens of Vienna clay have
14
those obtained for types I and Type
III
1
.
of
III
per cent greater strength type but type specimens Little
III
of
than those
,
Belt clay have per cent greater strength than type specimens and this dif
20
,
I
behavior prevents formulation general conclusions
of
ference The results
in
.
by
obtained for Vienna clay batch are verified previous incomplete test
IX
,
,
material prevented performance
of
series
of
on
other batches but lack addi

,
-
compression
in on
tional unconfined tests Little Belt clay The variations
in
.
compressive strength may be caused part by differences cohesion and
in
friction parameters and part by differences pore water pressures devel
in
in
-
few compression tests
of
oped during tests the three types specimens
on
A
.
were also made on type specimens of Vienna clay with circular cross sec
I
tion and was found that they had per cent greater strength than those with
it
8
,
square cross section

a
Inclination of failure planes Average measured angles
of
inclination
a
,
,
.
and the corresponding values of 20 which are not considered fully

90
in =
,
-
l
parentheses Fig For type speci

of
reliable are shown the table 28

in
I
.
by .
were probably increased slightly
of
mens the values deformations after

a
of
failure and there was considerable scatter the measured values for
in
a
,
type III specimens Little Belt clay The test results show excellent agree
of
ment between values of lo corresponding the primary planes of failure

to
in
,
type specimens and the angles effective internal friction Øe obtained

of
II
of ,
the direct shear tests As previously mentioned the duration the direct
in
,
.
Little Belt clay was probably too short and

on
shear tests not known

is
to
it
,
what extent the errors affecting 10° compensate each other

=
e
at I .
Many unconfined compression tests were performed speci

III
types
on
and
Vienna clay
of
of
of
mens other batches The axial deformations the end

.
these tests were smaller than for the tests mentioned above and the measured
,
angles of inclination of the failure planes may be more reliable The following
.
average values
of
were obtained these tests

in
0
Vienna clay III

VI
batch
,
V
18°
19°
19°
Specimens type
,
I,
= =
6
III
5°
type
19
Specimens 19° 19°

,
,
S
The corresponding values

of
of
of
Øe were 17°30 18º and those the angle

to
'
Øg
of
shear strength were 25°50 seen that the values for

to
26°45
is
Ø
,
s,
It
'.
'
'
specimens are slightly larger

III
types and than but much smaller than

e
I
O
s
.
'
Clays
on
Tests Other
compression tests various remolded English clays per
on
The results
of
the University
of
by
London are summarized

at
formed GIBSON 1953

It
,
of .
(
was found that the failure planes rotate and that the value of the angle incli
increases with increasing strain corresponding
Øa
to
of
nation Values
,
,
,
a
.
the initial values of a , were in excellent agreement with values of de which

had been corrected for the influence of surface energy or dilatation but were
1° to 2° larger than uncorrected values of Øe.
BJERRUM ( 1953 ) reports that values of land
Øé, obtained in tests on
five silty to very plastic Swiss clays , were in good agreement , the deviations
being of the order of + 1.5º. BJERRUM ( 1954 ) also performed two extensive
series of compression tests on remolded talus clay from Zurich . In one
series the compression tests were performed immediately after remolding
and in the other series after hydrostatic reconsolidation . The average values
of a obtained in the two series were in good agreement in spite of differences
in water content and strength , but the deviations of individual measurements
were rather large , the standard deviation of a being gº to gº . It was con
cluded that can be determined by tests on remolded clays with or without
reconsolidation , but single measurements are not reliable , and the number of
tests performed must be large enough to determine a statistical average value .
CASAGRANDE and WILSON ( 1953 - B, p . 36) state that measured angles of
inclination of failure planes and the corresponding values of % agree well
with the angles of shear strength , Øs, obtained in triaxial tests at Harvard
University on silty to highly plastic clays having values of Ø s ranging from
20° to 34º. Detailed test data in support of the statement are not presented in
this report, but CASAGRANDE and RIVARD ( 1959 ) state that values of ø for
an undisturbed plastic clay varied from 12° to 30° , and that the average was
21°, or 23° when maximum and minimum values were excluded . The corres
ponding values of Ø's ranged from 21° to 24º . On the other hand , HABIB ( 1953 )
found a = 45° or go = 0 for three clays and refers to similar results obtained
by TSCHE BOTARIOFF and BAY LISS ( 1948 ) .
Theoretical Considerations
CASAGRANDE and CARRILLO ( 1944 ) have developed analytical and graphi
cal methods for determination of the angle of inclination of and the strength
in the failure planes of anisotropic , cohesionless or purely cohesive soils .
HANK and MCCARTHY ( 1948 ) , using modified Mohr stress circles , have ex
tended these methods to soils possessing both cohesion and internal friction .
These theories show that anisotropy may have a marked influence on the op
timum angle of inclination of the failure planes . The simplest case is that of
a purely cohesivesoil in which the directions of the maximum and minimum
values the cohesion , c , and c9 , coincide with those of the principal stresses .
of
Assuming that the variation of the cohesion is represented by an ellipse , the
optimum angle of inclination , Qo , between the direction of the major principal
stress and the failure planes , and the shear strength , Tf, in these planes are
expressed by
tan a
12 = 2c1°2 /( ( y + cy ) (39 )
Reference is made to the above -mentioned papers for the more complicated
cases of soil possessing internal friction .
An indication of the influence of anisotropy or other irregularities can also
be obtained by determining the strength - stress ratio , T / T , for various planes
in an isotropic test specimen having constant values of c and in the Coulomb
failure criterion . As shown in Fig . 29 , the strength - stress ratio is expressed
1 - sin cos 20 ( 40 )
cos y sin 20
for
The ratio is equal to unity for a = 20 = 45 - 50 . Values of Ti / t devia
A2from the optimum inclination have been computed
or
tions
+
A
a
,
a
a
,
,
,
,
30° and are shown graphically Fig
29
for standard devia
in
gº 0°
and
is in to A
=
of =
a . , Ø
Ø
.
.
do by corresponds and
as
4º
found BJERRUM AQ
=
+
tion 1 954
in
,
)
(
TE 01 This means that the failure may occur plane with the
= to
015
=
1 a
1
1
T
.
.
/
when the shear strength this plane per
to
a 4°
in
inclination
1
5
.0
+
.
a
cent smaller than plane with the inclination an Such small difference
in
a
:
strength might be caused by anisotropy irregularities within the test speci
in
,
distribution of the external stresses the end surfaces
on
or
men nonuniform
,
the test specimen

of
.
Difference friction and cohesion parameters for the test specimens
28 in
of
by
Fig definitely determined means
be
shown cannot the above
in
anisotropy on changes pore water
of
mentioned theories since the influence
in
,
-
pressures not considered these theories and was not determined during
is
in
the tests
.
Summary and Comments

Theoretical considerations show that anisotropy and minor irregularities
strength properties and stress distribution may cause considerable varia
in
tions of the angle of inclination of the failure planes and measured values
a
,
,
,
this angle are subject appreciable amount
of
scatter However aver
of
an
a to
,
by .
for remolded clays obtained the writer Gib
of
age values
90
=
,
0
-
-
(
)
.
son and Bjerrum agree well with or are slightly larger than but are
e
,
,
-
-
do
much smaller than Øs These results not furnish definite answer the
to
a
.
hypothetical question of whether the effective cohesion component ce the
is
,
,
all directions for isotropic clay but they indicate that the pro
an
do
same
in
posed failure criterion and corresponding values Ve are better agree

of
in
ment with the test data than are values of obtained by standard methods
0
's
of evaluation and the Coulomb failure criterion the incli
in on
Published data
.
nation of failure planes undisturbed clays are limited extent and

on
tests
in
contradictory character
of
Fig the influence of particle orientation

28
on
on
Data the unconfined

in
compressive strength are preliminary character and not permit formu

, of
do
particle orientation giv

of
definite conclusions since the directions

of
lation
ing maximum and minimum strengths are different for Vienna clay and Little
Belt clay However the difference between maximum and minimum strengths
,
.
appreciable per cent for both clays and too large be neg
in 14
20
to
to
is
is
,
the influ
at
research Much more detailed investigations

of
lected least
,
ence of orientation clay particles and anisotropy general are needed

of
of
in
Triaxial tests should be performed specimens with horizontal inclined

on
vertical and random orientation the clay particles The deformation char
of
,
failure planes pore water pressures strengths and

of
acteristics inclination
,
,
-
the effective cohesion and friction parameters should be determined for each
the clay particles
of
type orientation
of
THE RHEOLOGICAL COMPONENT AND PROPERTIES

7
.
Early Investigations
of
Plastic Deformations
TERZAGHI the concepts and discussed the physical me
1931 reviewed
A
-
(
plastic
clays He pointed out that the plastic deformations
of
chanics
of
flow
.
may start shear stress considerably below the shear strength determined
at
a
by
laboratory tests and may continue for long periods

of
time and he empha

,
sized the importance of these phenomena the design of foundation structures

in
.
--
--
--
--
--
120
ola
Cх сот Ф
(0 ; - o ; )
SIN
ac T = a
2
COS
;
,
,
QC
=
0
(
2
OM +
0
-)
0
(
{
0
-0
)
2
Tq
=
TAN
+
O
c
SIN
Tp
COS
Q
1
-
2
a
SIN
COS
Y
/T 8
og
TR
=
.+a
RATIO
450
Q
. =
-
STRESS
30°
=
30°
P
=
STRENGTH
=
0
P
0°
=
00
1
.
10
10
8
+
4
+
-
+
-
+
2
-
. +
4
-
AA DEVIATION
=
FROM
a
Fig
29
Strength Stress Ratio Versus Inclination

.
-
These considerations were primarily based on the results of undrained com

pression tests on plastic clays .
Prompted by the above - mentioned paper by Terzaghi , the writer investi
gated plastic deformations of Little Belt clay under fully drained conditions.
The test was performed with the torsion shear apparatus , Fig . 11 , and the test
specimen was strongly overconsolidated in order to avoid a gradual decrease
in void ratio and corresponding increase in cohesion with increasing shear
stresses , which might impede the full development of plastic deformations .
The shear load increment of 0 . 1 kg / cm2 , and the time interval between load
increments was about ten days . Total horizontal and vertical displacements
for each load increment are shown in Fig . 30 - A . Failure occurred 10 to 15
minutes after the last load increase , and the actual shear strength is undoubt
edly slightly less than that corresponding to the final shear load , 0. 50 kg / cm2 .
The relation between the shear strain , y, and the twist , e, of the torsion shear
apparatus is
Y = 0 = 4 . 64 radians (41 )
Ion?
where n = R1 /R2, Ry and R2 are the inside and outside radii , and He is the ef
fective height of the test specimen . The actual height was 1. 45 cm , but the
shear strains are not uniformly distributed , and other tests indicated that the
equivalent height of a test specimen with uniform shear strains is about 1. 0
cm ; HVORSLEV ( 1937 ) .
Only the total deformations were observed . However , elastic deformations
took place within a few hours after load application , and subsequent deforma
tions were plastic or permanent. Plastic deformations for each load incre
ment and in the period from 100 to 240 hours after each load application are
shown in Figs . 30 - B and - C . The inclination of tangents to the curves in Fig .
30 - B represents the velocity of plastic deformations . These velocities were
determined for 120 and 240 hours after each load application , and are shown
in Fig . 30 - D. It is seen that these velocities were zero at 120 hours for shear
stresses less than 0 . 09 kg / cm2, and at 240 hours for shear stresses less than
0 . 15 kg / cm2 or about 30 per cent of the shear strength . At greater shear
stresses the velocities at 120 and 240 hours after load application increased
approximately linearly with the increase in shear stress until failure was ap
proached . Similar relations would undoubtedly have been obtained for still
longer periods after load application , had the time intervals been increased .
The diagrams in Fig . 30 - C show that the plastic deformations for a consid
erable period increased linearly with the logarithm of time. However , the de
formations for the stress Ta = 0 . 1 kg / cm² ceased after a period of 220 hours .
At the same time after load application the diagrams for T2 = 0 . 2 and 0 . 3 kg /
cm2 exhibit a slight curvature , and it is possible that the deformations for
these stresses would also cease in the course of time . The diagram for Ta =
0 . 4 kg / cm2 is straight for the entire time interval of 10 days , but this period
was too short for estimating whether the deformations at this stress would
continue at a decreasing rate and ultimately cease or cause failure of the test
specimen .
As shown in Fig . 30 - A , swelling occurred after each load increment and
also after failure . This means that an increase in shearing stresses caused a
decrease in pore -water pressure , but this increase was equalized in the course
of time with a corresponding increase in void ratio and decrease of the friction
VERTICAL SWELLING AH IN MM
0 0.1 0.2 0. 3
?
/
CM2
CM/ AFTER FAILURE
CG
,T
KG
IN
240HOURS AFTER
AVERAGE SHEAR STRESS
AVERAGE SHEAR STRESS

VERT . SWELLING LOAD APPLICATION
64
=
y
0
.
120HOURS AFTER
TORSION SHEAR LOAD APPLICATION
SIMPLE OVERCONSOLIDATION
CM
PRECONSOLIDATION KG TEST SPECIMEN

5
?
/
CM
. 11 45
CM
=
LTESTING LOAD KG
H
= = 1
1
.
.1 /
95.95
STRESS INCREMENT KG CM2 OD CM
0
L I
/
10
TIME INTERVAL DAYS CD CM
5
0
8
OF
TOTAL TWIST 102RADIANS RATE TWIST AO 106RADIANS PER HOUR

.
OF
STRESS AND DEFORMATIONS STRESS AND RATE DEFORMATION

A
D
-
-
to
=
.2
0
RADIANS
TANGENT AT
. 0.3
*-
240HOURS
=
=
,
10
=
04
IN
To
-
Io
=
00
(
4
0
.
=
0
.4
0
=
A0
=
3
0
T
.
HOURS
=
.3
0
100
TWISTAFTER
Ta
=
In
0
.2
=
0
2
.
To
To
=
=
0
1
0
1
.
.
100 120 140 160 180 200 220 240 100 120 140 160 180 200 220 240
TIME AFTER LOAD APPLICATION HOURS LOG TIME AFTER LOAD APPLICATION HOURS
=
T
-
.
DEFORMATIONS VERSUS TIME

of
Fig Belt Clay Before Failure

30
Plastic Deformations Little

.
and cohesion components of strength . This primary decrease in strength with
time contributed to the elastic and plastic deformations . However , for the
test described , the primary swelling is theoretically 99 per cent complete in
less than 100 hours , and plastic deformations occurring more than 100 hours
after load application may represent the influence of the viscous or rheologi
cal strength component , secondary swelling , and thixotropic and structural
changes in strength .
Recent Rheological Investigations
Over many years HAEFELI ( 1939 , 1953 ) made extensive field and labora
tory investigations of continuing plastic deformations of snow and ice , extended
these investigations to soil , and called attention to the great importance of such
deformations in the design of foundation structures . HAE FELI and SCHAERER
( 1946 ) also suggested methods for determination of shear - strain relations for
soils by means of triaxial tests .
GEUZE ( 1948 , 1960 ) investigated plastic deformations of cohesive soils in
the field and also in the laboratory by means of drained direct shear tests .
He found a linear relation between the shear strains and the logarithm of time
in both field and laboratory investigations . This relation can be expressed by
log
't ï
42
= At (Ya +
)
(
t
)
yd
is at
the strain one day after placing

74
where the total strain the time

is
is
in t,
the stress increment AT the time days after placing the stress incre
,
t
corresponding
an
the increase strain for
to
of
ment and increase
in
is
,
= a
i'
t
full logarithmic cycle or tenfold increase of Both yd and refer AT
to
a
t.
by
of dif
on
and depend also the total stress The rate strain obtained
is
,
1
T
.
of
ferentiation Eq
42
,
.
at
at
43
=
2
3
:
ï
(
)
The elastic strains attain their maximum value less than one day and sub
in
sequent increases strain are plastic strains Therefore Eq represents ,

43
in
,
.
the rate of plastic strain and shows that this rate varies inversely with time
.
Eqs 42 and 43 apply also with minor modifications and limitations the test
to
,
,
.
Fig Geuze also suggested that the volumetric strains should

30
data shown
in
be considered computing the pure shear strains and he called attention

to
in
the similarity between the plastic shear strains and the volumetric strains
by
caused secondary consolidation

.
GEUZE and TAN 1953 and GEUZE 1960 investigated the shear deforma
(
clays
by
torsion shear and triaxial

of
of
tions constant water content means

at
tests The tests were performed and evaluated accordance with the princi
in
.
ples and theories of rheology and plasticity and the authors deserve credit for
of ,
introducing these principles into the field soil mechanics The mean com
.
pressive stress was held constant order produce states of pure shear
to
in
and the elastic and plastic deformations were separated by unloading the test
specimen before each load increase and observing the rebound was found
It
.
it , up
that the shear strains were reversible elastic character

or
to
fo
in
,
T
called the yield limit For stresses greater than the yield limit part
of
the
.
strains were permanent plastic For the soils investigated appears that
or
these plastic strains approach constant velocity during the rather limited
a
time intervals between the load increments and that this terminal velocity
,
increases linearly with the stress increase

by
expressed
as
,
apa
44
(T - f ) =n
(
)
where the coefficient structural viscosity also called the Bingham vis
is
of of
,
îl
by
cosity since the behavior such clays can be
represented Bingham rheo
a
logical model However being disputed by some in
the Bingham concept
is
,
.
vestigators and the yield limit theoretical rather than definite physi
fo
is
,
a
pure elastic deforma
of
cal limit since there transition between the state
is
a
tions and that of plastic deformations with velocities given Eq
by
44
When
.
.
the shear stress exceeds another limiting value called the flow limit and des
,
plastic strains becomes greater than that indicated
of
ignated by the rate

ff
,
.Eq
and the authors suggest that this greater rate

44
of
by
flow ultimately
,
may increase with time and cause failure of the soil support of this sug
In
.
experiments
by
gestion reference made CASAGRANDE and WILSON
to
is
,
GEUZE 1960 proposes that the design earth and foundation struc
of
1951
(
)
.
the above mentioned yield and flow limits rather than

on
tures should be based

-
the shear strengths obtained currently used laboratory and field tests
on
in
.
Determination of the elastic and plastic components the total strains by
of
complete unloading the test specimen before each load increase may cause
of
changes the stress strain properties of the soil which many cases may
in
in
,
-
objectionable unloading may be avoided by utilizing the sugges

be
However
,
.
by
Wilson that the elastic strains are approximately equal the instan
to
tion
taneous strains which can be determined as the zero time intercepts of the
,
stress strain curves for individual load increments see CASAGRANDE and
;
-
WILSON 1949 and WILSON and DIETRICH 1960

(
)
.
Egs 43 and represent two basic types of plastic deformations of clays

44
.
.
may many cases require very long time intervals between load increments
in
It
these types of flow actually occurs as read

of
order determine which is

to
ily in
seen by examination of Figs GOLDSTEIN

30
30
and 1957 1958 has

C
B
)
(
.
.
-
developed more complete equations for the plastic deformations which take
,
the effect of previous load increments into consideration The above

.
mentioned relatively simple relations may be altered when the clays undergo
thixotropic when migration pore water
or
or
structural changes strength

of
in
occurs during the plastic deformations clays may

of
The deformations
in
.
many cases be represented by rheological models for which mathematical ex

pressions of the relations between stresses strains and time have been de
,
veloped and are described various papers and books rheology see for
on
in
example REINER 1954 Applications of visco elastic stress strain theories

(
)
.
and rheological models specific soil problems and for various assumed re
to
lations are explained the excellent paper by SCHIFFMAN 1959

in
)
.
The Rheological Component of Strength

comparison of the results of slow and quick direct shear tests
on
Vienna
A
clay and Little Belt clay Fig 31 shows that quick tests yield lower strengths
,
,
.
than slow drained tests for normally consolidated and slightly overconsolidated
clays but higher strengths for strongly overconsolidated clays The primary
,
causes of these differences strength are the development of positive or

in
1
(
)
negative pore water pressures during quick tests and corresponding decreases
in -
increases void ratio during slow drained tests and changes the vis
or
in
,
(2
)
cous or rheological component strength with time or the rate

of
of
deformation
.
The part of the difference shear strengths obtained

by
quick and slow

in
direct shear tests attributable the rheological component can be determined

to
,
for normally consolidated clays by means Eq

35
and for overconsolidated

of
.
222
SLOWSHEARTESTS
Y
VIENNA CLAY
CM2/ KG INTH STRENGTHSHEAR
CM2/ KG TRIN STRENGTHSHEAR

SLOWSHEAR TESTS
II
LITTLE BELT CLAY
QUICK SHEARTESTS
I
SHEAR STRENGTH
VIENNACLAY
QUICKSHEARTESTS
I
LITTLE BELT CLAY

CONFERENCE
56
1
4
.
3
.
2
.
/
IN
OF
KG
CM2
OF
NORMALSTRESS
KG
NORMALSTRESS
5IN
/ CM2
-B
LITTLE BELT CLAY
-A
VIENNA CLAY of
31
.
Fig Results Quick Direct Shear Tests
clays by direct comparison of strengths at the critical degree of overcon

a
solidation , when there are no changes in pore - water pressure or void ratio
during a shear test. The data shown in Fig . 31 cannot be used for the above
mentioned purpose because the quick and slow tests were made with different
batches of clays , but comparable data were obtained in tests on batch I of
Vienna clay .
The strength parameters for Vienna clay I are shown in Fig . 32 - A , and
the strength obtained in slow drained tests on normally consolidated samples
is
.
Tp = 0 324 0 .
+ 0 105 o ' (45 - A )
or by use of Eq . 35 and nc = 1 72 .
294
105
189
.
oC
45
Tp = 0
B
+
.
*
0
-
the 0
.
(
0
)
To
where
de
of
324 friction component and
ce
189
=
is
=
0
105
0
=
.
d
'é
.
the effective cohesion component
is
/
CM2
.10 .0
QUICK TESTS MIN
=
O T
,
3
TFKG
05 -
O
=
o
1
:
.
144
105
Ty
333
=
350
o
=
0
0
1
.
0
.
*
.
THE 324
= = =
. 0. .0
105
*
TA
72
x
.
1
I
COMPONENI
0
189
600
FRICTION SLOW TESTS MIN

=
>
,
! 0 T
-
504
If
)10504
7207
=
. =
To
0
1
.
.
'
294
10
189
=
=
0
0
o
0
.
I
.
I
KG
10
0
CM2
1
.O
..
6
.
NORMAL CONSOLIDATION
A
-
ZERO SLOW TESTS QUICK TESTS RATIOS

VOLUME CHANGE 600 MIN
>
MIN SLOW QUICK

=
T
,
3
T
,
Ir
of
324 oil
. Il
5
0
0
-
.0
-
42 32 32
81 6868
39 3636
83
AH
72
= =
83
.1
0 0
.0
0
50
82
. 0 0.
0
. 0.
.0
0 0.
00.
. 0
.
.
SH
.
.
|
0
.1
|
10 01
91
49
89
80
0
0
1
3
0
-
.
0
.
.0
.
0
.
OVERCONSOLIDATION
B
-
Rheological Component Vienna Clay

for
Fig
32
of
Influence
I
.
During quick shear tests on normally consolidated samples the water con
,
tent in the shear zone decreased 0 . 4 per cent , corresponding Ae = 0.011 which
inserted in Eq . 23 yields nc = 1. 10 . However , the time required for disman
tling the apparatus and cutting samples from the shear zone was about equal
to the duration of a quick shear test, and it is
probable that a part of the
change in water content in the shear is caused by internal migration of zone
water after failure . Therefore , it is assumed that the actual change in water
content at the moment of failure only was 0. 2 per cent , corresponding to nc =
1. 05 or 0e = 1. 05 O., where of is the total normal stress at failure and also
the effective stress at the start of the test . The shear strength obtained in the
quick test on normally consolidated samples can then be expressed by
350
333
Op
.
46
Ep = 0
A
=
)
-
0
o
'
.
by
by
assuming that the friction component
or
not affected time
is
46
189 144
B
=
+
g
)
-
0
0
.
.
T
ce
quick
oe
where the effective cohesion component obtained
in
of is
144
=
0
.
Eqs yields the following strength time

45
tests comparison and 46

A
-
.
.
ratios
,
Tp
slow
105
slow
] )
7 0
(
(
)
.
88
73
0 0
..
and
e
73
-
00
quick
..
=
533
0
.
144
=
=
=
:
Couick
)
Ś
(
Data obtained by tests Vienna clay are summarized
on
overconsolidated
I
cyclic
of
at
Fig The critical degree overconsolidation
32
which there
is in
,
B
.
-
.
change pore water pressure or void ratio during shear test can be
no
in
,
a
-
by
determined approximately Fig

21
The surface movement

of
examination
,
.
.
and the diagrams representing equivalent consolida

of
AH zero for
is
=
1
0
,
,
at
tion pressures
Tf of
at
the start and end the shear tests intersect each other
of
Corresponding
of
were then determined assuming

ce
values and
=
1
3
,
.
.
that the friction component

of
for both slow and quick tests

To
324 The
=
0
.
average values of the strength time ratios are .

Ce
86
76
for and for

Tf
0
0
.
.
-
various approximations :
on
of
The evaluation these test results based

is
,
possible but minor influence thixo
of
and does not consider differences

in
a
tropic changes strength remanent pore water pressures

or
differences
in
in
Therefore the test results furnish only rough indication the influence of
of
,
the rheological component strength The time strength ratios cited above
of
for Tf and
86
are average
88
of
75
an
for These ratios indicate

ce
to
0
0
.
the rheological component with time reduces by

byof
12
that the variations

to
Tf
per cent and

ce
per cent when the test duration

14
25
increased from
is
minutes 600 minutes The rheological decrease strength

or
200 fold
to
is
in
.
-
not complete 600 minutes but its ultimate value cannot be

of
test duration
at
a
or,
without further assumptions

of
by
determined means available test data

.
Investigations discussed the following subsection show that the compres

in
sive strength of saturated clays decreases linearly with the logarithm of time
,
Fig and Eq probable that this relation also applies

33
48
the decrease
is
to
Ce It
.
Tf .
the coefficient pz
of
with time
of
and Values Eq 48 are obtained by

in
:
dividing the reduction percentages given the preceding paragraph by log

in
,
log log 200 which yields

05
Tf
06
Ce
for
11
600 and for

to
=
=
3
2
.3
,
0
-
ce )
:
by
at
be
by
Values the time designated

of
Cet can then determined the

t,
equation
Cet
108
47
11
cez11
=
(
)
)
.
0
where the time in minutes , and cez is the value of Ce for t = 3 minutes .
t is
Numerical values of Cet / cez obtained by this equation are shown in Table 3.
It is not yet known whether Eqs . 47 and 48 are valid for very high values of t .
It is probable that ce with time approaches a limiting value , Cu, which is the
actual cohesion component . The rheological component , cy , at the time t is
then determined by Cy = Cet - Cu . For the purpose of illustration it may be
assumed that cu is equal to the 10 - year value of Cet : Corresponding values of
cu / cet and cy /cer are also shown in Table 3, and it is seen that cu is 43 per
cent and Cy is 57 per cent of ce in a common slow test with t = 10 hours .
Table 3 - Influence of the Rheological Component for Vienna Clay

Time t
3 min 10 hr 10 Days 1 Year 10 Years
1 . 00 0 . 75 0 . 60 0 .43 0 . 32
0 . 32 0 .43 0 . 54 0 . 75 1.00
0 . 68 0 . 57 0 .46 0 . 25 0. 00
The data in the preceding paragraph

and Table 3 are speculative and are
presented in order the influence of the rheological component
to demonstrate
and a method when adequate test data are available . In the example cited , cu
forms a substantial part of ce determined by slow tests , but cu may be close
to zero for other clays with the same time - dependent decrease in strength
but having smaller values of the parameters for the effective cohesion com
ponent .
The Long - Term Compressive Strength
CASAGRANDE and WILSON ( 1949 , 1950 ) have made extensive and very
thorough investigations of the long - term compressive strength of clays and
clay - shales at constant water content and under sustained loads . It was found
that the strength decreases linearly with the logarithm of time when the clays
are saturated . This approximate relation , which has been verified by other
investigators , is illustrated in Fig . 33 and can be expressed analytically by
pa
log
48
9 + = 9a - ä log
)
=
(
t
(
t/
9
.)
(1
-
t(
/t
)
]
by
or
as
suggested GOLDSTEIN 1957

(
log
49
+
=
q
9
(t
/t
)
(
)
where the time corresponding zero strength and

to
the decrease
is
to
is
49 in
q
,
strength for logarithmic cycle of time or the strength for Eq

to
=
a
0
1
t
advantageous for further mathematical developments whereas Eq

48
an is
offers
,
.
by
opportunity characterize the decrease strength single parameter

to
in
which may be called the coefficient

qa
rheological decrease
of
Pa
=
q
in
,
/
strength
qa
This coefficient varies with the reference strength and time

ta
.
to ,
value Pb corresponding
its
qb
and another reference strength and time

to
,
by
easily
Pab
is
determined
or
the relation paqa

q
=
,
pq
50
Pp
pa
96
11
P2192 log
=
/tty
(
)
=
)
/
/
-
)
]
Casagrande and Wilson discuss both transient and long term strengths and use
,
-
of
minute strength reference value for determination

as
the one the coeffi

a
-
of
cent decrease strength which the following will be designated

by
21
in
in
,
Several other investigators are primarily concerned with the long term
-
226
.
LOG TIME
,S IN
DISPLACEMENT 2
1
CM 2
.4
1
.
.
1
8
.
1
5
.60
0
.
3
0
0
0
,
95
1
1
2=
1T
OI
)-
. ,,
. -(=
OVERCONSOLIDATION 483MIN
elet
0L06
-,
-1,2-
0
( BLOGS
Ž .
%
81
)
-4 100
AT
STEIN 1957
OR
1951
47 OR
*. 06
-4= 0=,9
160LOSTEIN
%
72
%
,T, =
=
.02
/
,
04
CM2 1300MIN
73 KG
KR
STRENGTH
STRENGTH
min
.10
.
LOG SCALE
S
40
5
35
25
30
20
TO
15
10
.6
TWISTOF SHEARINGAPPARATUS 100RADIANS

-A
VIENNA CLAY
33
.
Fig Undrained Strength Versus Test Duration
IN
CM
SHEAR STRENGTH
DISPLACEMENT
,8 .
4
2
. 2
1
3
.
1
8
0
.
0
6
5
.
2
9
.0
)
(
-
-
=
FROMBJERRUM SIMONSTORBLAA 1958 16

=
.1 ,
.1 ,T1-15-
O
',
CM2
MIN
720
/ / OVERCONSOLIDATION
1
5 .
KG HOURS
12
-
TO
=
0
=
300 NORMAL CONSOLIDATION 01 1160MIN

IN
= =
100
T OC
''
0 =0 .
)) 045
=
,T, 8
=
.5 8.
%
=
- , 1.0
0
te
30
AT 31
CM
-
CONFERENCE
0
({ 460HOURS
TO5°
Tg C
..
29250
'' == =
045
=
3
=
.
.
0
8
A
%
=
ta
40
AT 52
CM
B
=
,
-
SIN
TAN AVERAGE RATIO
/'
c
ØC=
COSO
40
35
25
30
20
15
5
10
0
.0
TWISTOF SHEARINGAPPARATUS 100RADIANS
+
KG
/; IN
240
)0
CM2
-B
LITTLE BELT CLAY
and After
-
Before
35
.
Fig Stress Strain Diagrams for Shear
34
Fig
and Time
.
Strength Versus Effective Stress
Failure
strength , in which case it is more significant to use a reference strength cor

responding to that obtained in normal slow tests . The writer suggests that the
1000 -minute strength be used as the reference strength , and that the corres
ponding value of the coefficient of decrease in strength be designated by Pm :
According to Eq . 50 , the relation between p1 and Pm is
Pm = P _ /( 1 - ,
30 (51 )
The data presented by CASAGRANDE and WILSON (1951) yield the following
coefficients of decrease of the undrained strength of clays and clay - shales
with time
General range P = 0 . 04 to 0 . 09 PM = 0 .05 to 0 . 13
Cucaracha clay - shale P = 0 . 07 to 0 . 19 p = 0 . 09 to 0 . 46
BJERRUM -SIMONS - TORBLAA ( 1958 ) investigated strength - time relations

for a normally consolidated marine clay in
its
undisturbed state using
,
of
of
controlled strain type loading The results consolidated undrained tests
.
-
agree with Eq 48 provided the test duration greater than one hour and is
to ,
,
.
pm
yield The authors also found that the pore water pressure
06
07
=
0
-
.
.
.
at failure increases with the test duration for normally consolidated test
specimens and that this increase pore water pressure and corresponding
in
,
part but not all
of
decrease effective stresses can explain the decrease
in
in
strength with increasing test duration The test data were evaluated by the
.
by
method represented Eq Fig
34
and the results are summarized
in
,
B
9
-
.
.
.
that the angle of shear strength decreases with increasing test
is
seen
It
d
,
',
duration The lines OA and OB correspond the shear strength lines and
re
to
.
present changes both effective stresses and void ratios As mentioned by

in
the authors the data obtained were not sufficient for computation of the effec
,
tive friction and cohesion parameters and determination whether the decrease
strength with time affects not only the cohesion but also the friction para
in
meter
.
BJERRUM SIMONS TORBLAA 1958 also conducted consolidated drained

(
)
-
triaxial tests the marine clay and found that the strength independent of
on
is
,
the test duration when the latter greater than one day The authors suggest
is
the rheological component

by
compensated
of
that the decrease this case

by in
is
component secondary
an
of
increase the cohesion caused consolidation and

,
decrease void ratio

in
BISHOP and HENKEL 1957 investigated the long term strength of Weald
(
clay and London clay and the data obtained yield the following values pm
of
,
clay test
06 04 04 06
London normal consolidation

, ,
drained
,
=
,
e
P
0
.
Weald clay normal consolidation drained test

a
,
=
,
0
P
Weald clay overconsolidation drained test a

= =
,
0 0
.
em
Weald clay normal consolidation undrained test

,
surprising that the same values

of
om were obtained for normally con

is
It
solidated and overconsolidated Weald clay which may indicate that the influ
,
or
secondary volume changes was negligible that compensating factors

of
ence
were active The higher value
of
pm obtained the undrained tests reflects

in
.
the influence of increasing pore water pressures

-
GOLDSTEIN 1957 presents time strength data for unspecified clay

an
= (
09 )
which yield pm
.13
Goldstein defines the failure strain and time

to
,
0
0
.
caused by given stress as that which the rate of strain starts increase
at
to
a
,
and he found that this strain is independent of the test duration for a given
soil , state of consolidation , and testing procedure . Therefore , the time of
failure for a given stress can be determined by extrapolating the time - strain
curve for this stress to the failure strain . In a subsequent discussion GOLD
STEIN (1958 ) also proposed amended analytical expressions for the stress
strain - time relations and derived corresponding expressions for the long
term strength .
Some of the data presented by CASAGRANDE and WILSON ( 1949 , 1951)
show but little influence of the test duration on the failure strain , whereas
other data indicate an increase in failure strain with increasing test duration .
BJERRUM - SIMONS - TORBLAA ( 1958 ) found a definite decrease in failure
strain with increasing test duration . WHITMAN (1960 ) has conducted extensive
investigations of the influence of high rates of loading on strength and strain
and found that the failure strain of plastic soils is nearly independent of the
test duration , whereas that of brittle soils decreases with increasing test
duration .
Some of the previously mentioned tests for determination of the time
dependent decrease in strength and the failure strain were performed with
controlled stress loading and others with a controlled rate of strain . It is
possible that results obtained with the two types of loading may not be directly
comparable because of differences in the rate of strain at failure for a given
test duration . There may also be differences in the definition of the failure
strain , and attention is called to the previously mentioned definition by GOLD
STEIN ( 1957) .
Tests on clays at constant water content yield one rate for the decrease in
strength with time, and another rate is obtained by drained tests . It is possi
ble that the decrease in strength of natural clay deposits may have a magnitude
which is between the two rates , because a slight migration of pore water may
cause a considerable change of the pore - water pressure in the failure zone .
A migration of pore water takes place in undrained compression tests but it
may not be of the same magnitude as that which occurs in natural deposits
over periods of time .
long
discussed in this section refer to saturated clays . As shown by
The data
CASAGRANDE and WILSON ( 1951) , the strength of partially saturated , com
pacted cohesive soils , held at constant water content , may initially decrease
slightly and thereafter increase with time .
Summary and Comments
The principal causes of continuing shear deformations and time - dependent
changes in strength of cohesive soils may be summarized as follows :
a . Primary changes in pore - water pressure after an increase in shear
stress . An increase in shear stress causes a decrease in pore - water
pressure in strongly overconsolidated clays , but this decrease may
be equalized in the course of time by migration of pore water , which
in turn produces a gradual increase in void ratio and decrease in ef
fective stresses , and a corresponding decrease in strength and in
crease in strain .
b. Secondary changes in pore - water pressure caused by a sustained in
crease in shear stress . It has been demonstrated that the pore -water
pressure of a normally consolidated clay increases with time when
triaxial test specimens are held at constant water content and sub
jected to a sustained axial load . The behavior of strongly overcon
solidated clays under similar conditions is not yet known .
c. The above -mentioned effect of an increase in shear stresses may

initially be increased or decreased by a concurrent change in the ef
fective , mean normal stress (oj + 02 + og ) / 3 .
d. The existence of a viscous or rheological component of the shear
strength and its variation with time and the rate of deformation .
e. Thixotropic changes in strength with the rate of deformation , which
often may be difficult to separate from concurrent changes in the
rheological component of strength .
f. Structural disturbance of the clay in the zone of current or former
failures and corresponding transient and permanent changes in shear
strength .
The deisgn of many foundation structures is governed by the deformations
rather than the limiting strength of the foundation soils , but adequate experi
mental rheological data are generally lacking , and the restrictions imposed
by the deformations are often taken into consideration by using allowable
strengths which correspond to small but arbitrarily selected strains on the
stress - strain diagrams obtained in standard laboratory strength tests . This
procedure is a practical expedient which should be amplified and delimited by
detailed and systematic research , utilizing the methods and theories of plas
ticity and rheology . However , modifications may be required , since these
theories have been developed for materials with volume change characteristics
which are different from those of soils , especially in regard to the volume
changes caused by pure shear stresses . Investigations of the long - term rheo
logical properties of clays are by nature time - consuming , but basic research
into the problem , and correlations with index properties and the state of con
solidation of clays , will undoubtedly lead to short - term tests or methods for
making reliable estimates .
8. CHANGES IN SHEAR STRENGTH AFTER FAILURE
Knowledge of the changes in shear strength and of the shape of the stress
strain curves for soils after failure is of practical importance for estimating
the effects of progressive failure or previous disturbance of the soil in sta
bility and bearing capacity problems and for various special purposes . Deter
mination of the stress - strain curve after failure may be difficult when large
deformations are involved , which generally is the case . The accuracy of
standard box shear tests , unconfined and triaxial compression tests decreases
with increasing deformations because of changes in the effective cross section
of the test specimen . The ROSCOE ( 1953 ) direct shear apparatus maintains a
constant cross section of the test specimen and can be used for relatively large
but nevertheless limited deformations . Deformations of any magnitude can be
obtained with the torsion ring shear apparatus, and the cross section of the
test specimen remains constant , but the deformations after failure are confined
to a relatively thin zone . Limited data on the stress - strain curve after failure
can also be obtained by means of vane shear tests .
The following data on changes in shear strength after failure of Vienna clay
and Little Belt clay were obtained by means of the torsion ring shear appa
ratus shown schematically in Fig . 11. The tests were performed by incre
mental changes in torque or shear stress . Points on the stress - strain curves
after failure were obtained by rapidly reducing the shear load until the de
formations ceased and immediately thereafter increasing the shear load until
failure again occurred . The tests were in fact conducted as slow drained tests
before failure , and as fairly rapid tests with short interruptions after failure .
However , rest periods between subsequent repetitive loadings permitted equal
ization of static pore -water pressures . The increase in deformations with in
creasing distance from the center of the test specimens was taken into con
sideration by the method of evaluation of the test results described in previous
publications, HVORSLEV ( 1937 , 1939 ) .
Tests on Vienna Clay
The residual strength . Typical stress - twist or stress - displacement curves
are shown in Fig . 35 - A . The curve for normally consolidated Vienna clay has
a transient minimum of 73 per cent of the maximum strength , which undoubt
edly is caused by development of excess pore -water pressures after failure ,
and the subsequent increase in strength to 81 per cent is due to partial equal
ization of these pore -water pressures . An increase in shear stresses tends
to cause a decrease in void ratio of normally consolidated clays , and an ad
ditional decrease in void ratio or increase in pore - water pressures occurs
after failure . In contrast thereto , strongly overconsolidated clays undergo an
increase in void ratio or a decrease in pore - water pressure before and after
failure . Therefore , the stress - twist curve for strongly overconsolidated
Vienna clay does not exhibit a transient minimum value of the shear strength
during continuous and fairly uniform deformations after failure , and the rela
tive residual shear strength , 72 per cent , is smaller than that for the normally
consolidated clay .
Regain of shear strength . Upon reaching an apparent residual value of the
shear strength , the above -mentioned shear tests were interrupted by reducing
the shear load , and the tests were repeated after rest periods of increasing
length . The results obtained for the normally consolidated Vienna clay are
shown in Fig . 36 . It may be noted that the original shear strength of this clay
was 1 . 00 kg / cm2 , shown as 100 per cent in Fig . 35 - A . The residual strength
increased from the transient minimum of 0. 73 kg / cm2 to 0 . 83 kg / cm2 in less
than 82 minutes and remained constant thereafter . After a rest period of 10
minutes the shear strength had increased from the residual strength to 99 per
cent of the original strength ; thereafter the strength increased linearly with
the logarithm of time and reached 1. 11 kg / cm2 after a rest period of 238
hours . It is estimated that the decrease in water content after failure was
about 1. 0 per cent , which corresponds to an increase of 0. 09 kg /cm2 in the ef
fective cohesion component .
A similar increase in strength upon cessation of de formations was also ob
served for strongly overconsolidated Vienna clay , but the rest periods were
not of sufficient length to determine definitely whether the original strength
ultimately would be regained . The overconsolidated clay underwent a small
increase in void ratio after failure , and it is probable that the original shear
strength would not be fully regained by a longer period of rest .
The results of these tests indicate that the decrease in shear strength after
failure primarily is a thixotropic phenomenon for remolded Vienna clay , and
that the structure of the remolded clay has not been altered appreciably by the
failure .
Residual strength parameters . The final residual strengths obtained for
normally consolidated and overconsolidated test specimens of Vienna clay
form a shear strength diagram similar to that shown in Fig . 21 . The friction
and cohesion parameters were determined as shown in Fig . 37 , where de is
the equivalent consolidation pressure corresponding to the final water content
LOG T =REST PERIOD IN HOURS

0. 1 2 4 6 81. 0 2 4 6 8 10 2 4 6 8 100 2 4 6 8 1000
/
1. 12
KG CM2
,T
1. 10
IN
1.08
TO
SHEAR STRENGTH
06
02
1
.
00
1
.
98
0
.
96
0
.
94
0
.
VIENNA CLAY
V
REST THIN KG CM2

INO
/
,
92
NORMAL CONSOLIDATION PERIOD MAX

0
MIN
.
KG
CM2
=
2
.0
o
/
'
00
73
90
. . . 0
. . . . . . .1
.
0
.
WATER CONTENTS 10 99 81
1 0
0 0
25
'
%
AT FIRST FAILURE
+
% 0
02
.
82 83
88
NO
'
FINAL VALUE 24 01
0
.
323 055 83
1 1 1 1
0
'
KG
AW
%
AOI
08 07
=
CM2
1
.0
,
KG 0
/ .9
12H 83
-
86
0 0 0
. . .
0
.
7 601W
09
Ace CM2 35H 83

0
.
-
122H 10 83
84
KG
83
FINAL MINIMUM 11
0
CM2 238H 11 83
.
0 0
.
. .
. .
MAX 157 83
1
82
.0
80
L
0
.
Fig
Regain Shear Strength with Time for Vienna Clay

36
of
.
11
VIENNA CLAY
SHEAR STRENGTH
MeR 315 Ke 045 LITTLE BELT CLAY

0
=
.
0
-
.
090
030
=
K
HeR
=
0
.
.
17°30
=
Ø
-
'
NON
BIO
=
5°10
'
045
=
030
=
0
K
0
.
1
.
*
0
3
0
O 7
0
8
0
.6
.
0
5
.
1
0
.2
.
.
10
NORMAL STRESS RATIO

.
Fig
37
Friction and Cohesion Parameters for Residual Strength

.
in the failure zone , and of is assumed to be equal to the total normal stress ,
os. A comparison of Figs . 24 - A and 37 shows that failure does not affect the
effective angle of internal friction but reduces the effective cohesion para
meter . It was originally believed that the residual strength parameters thus
determined were reliable , because the void ratios apparently had reached
stable values and the soil structure had not been changed appreciably . MIT
CHELL ( 1960 ) and others have recently shown that changes in the pore - water
pressure take place during thixotropic changes in strength , in which case the
values of of used in Fig . 37 are not the true effective stresses . However ,
Mitchell also demonstrated that thixotropic changes do not affect the strength
expressed in terms of effective stresses . Therefore , the strength parameters
shown in Fig . 24 - A also apply to the residual strength , and a comparison with
Fig . 37 indicates that the transient pore - water pressure caused by a thixo
tropic decrease in strength is a constant for a given void ratio , but this ap
plies only to the final residual strength and not to the transient minimum
strength exhibited by remolded and normally consolidated Vienna clay . As
seen in Fig . 37 , available data for determination of the residual strength para
meters are very meager , and additional investigations are needed .
Tests on Little Belt Clay
The stress - twist diagrams in Fig . 35 - B show a much greater decrease
in strength after failure than that obtained for Vienna clay , and it was found
that the regain in strength upon cessation of the deformations was incomplete .
The strength of the overconsolidated test specimen increased from the resid
ual value of 30 per cent to 37 per cent during a rest period of 4 days and to 55
per cent in 60 days . The determination of the residual strength parameters in
Fig . 37 should be corrected for pore - water pressures corresponding to the
thixotropic change in strength , which in this case accounts for only a part of
the total decrease in strength . It is evident that failure cause a decrease of
both the friction and cohesion parameters and also a permanent change in the
soil structure . The failure surfaces were distinct and smooth but not glossy .
As previously mentioned , the curvature of the shear strength line for nor
mally consolidated Little Belt clay , Fig . 22 , was not caused by too short a du
ration of the tests . It is possible that a part of the above -mentioned structural
disturbance of Little Belt clay occurs before failure , and that this disturbance
increases with increasing normal pressure and may account for the curvature
of the shear strength line . The fact that structural disturbance occurs under
the above -mentioned circumstances may also indicate that it is not possible
to obtain complete dispersion of a highly plastic clay by mechanical remolding .
Other Investigations
TIEDMANN (1937 ) developed a torsion ring shear apparatus similar in
principle to that shown in Fig . 11 but with a different design of the details .
Tests were performed on four remolded and undisturbed clays , ranging from
silty clays to highly plastic clays . The test specimens were normally con
solidated , and the testing procedure was similar to that described for Vienna
and Little Belt clays . The residual strengths varied from 20 per cent to 60
per cent of the maximum strength and were attained after displacements of 10
to 14 cm .
As previously mentioned , a torsion ring shear apparatus with both incre
mental stress and controlled strain types of loading has been developed by the
USAE Waterways Experiment Station . A series of tests on Atlantic Muck , a
highly organic clay , was performed with this apparatus . The test results are
presented in a detailed report by the WATERWAYS EXPERIMENT STATION
(1951 ) and are summarized by HVORSLEV and KAUFMAN ( 1951) . The test
specimens were undisturbed and normally consolidated or overconsolidated
under various pressures . Controlled strain type of loading was used , and the
starting rate produced failure in 3 to 6 minutes ; thereafter the rate of dis
placement was increased five times , but the total duration of the tests was
nevertheless sufficient to permit appreciable changes in pore -water pressures
and void ratio , which made the evaluation of the test results rather difficult .
Transient minimum strengths after failure were observed for normally con
solidated and lightly overconsolidated test specimens . The residual shear
strengths ranged from 30 per cent to 90 per cent of the maximum strengths
and were attained after displacements of 3 to 10 cm , according to the state of
consolidation of the clay . Slickensided failure surfaces were produced after
large displacements , Fig . 3 - B , but only when the normal stress was equal to
or greater than 1. 0 kg / cm2.
The form of stress - displacement curves for soils , before and after failure ,
is important for estimation of the tractive power of vehicles . BEKKER (1956 ,
pp . 263 - 273) has proposed mathematical expressions for complete stress
displacement curves , originally obtained by tests resembling direct box shear
tests . Improved equipment for in - situ determination of the stress
displacement curves has recently been developed by the Land Locomotion Re
search Branch , Ordnance Tank - Automotive Command , Department of the
Army; see PAVLICS (1958 ) and VINCENT ( 1959 ) . The apparatus , called a
bevameter , consists of an annular plate with ribs or grousers which is pressed
into the soil , subjected to a desired normal load , and rotated .
Indications of the form of stress - displacement curves before and after fail
ure can also be obtained by vane tests . These tests are primarily used for de
termination of maximum and residual shear strengths of clays in situ , and rel
atively few stress - displacement diagrams have been published . MARSAL
(1957 ) presents stress - displacement diagrams obtained by vane tests in the vol
canic clays of Mexico City . The residual strengths ranged from practically
zero to 30 per cent of the maximum strengths and were attained after displace
ments of 8 to 12 cm . Neither the effective nor the total normal stresses ex
isting during vane tests are definitely known .
The large deformations required to attain the residual shear strength of
plastic clays cannot be produced in triaxial or unconfined compression tests .
However , these tests may be used for determination of complete stress - strain
curves of brittle clays and clay shales. Examples of such curves are shown in
a paper by CASAGRANDE and SHANNON ( 1949 ) .
Influence of the Rate of Deformation
Published data on the influence of the rate of deformation on the residual
shear strength and the shape of the stress - strain curve after failure are very
limited . A single test for this purpose has been made at the WATERWAYS EX
PERIMENT STATION ( 1951) and fragmentary data may be obtained from some
triaxial tests and vane shear tests . An increase in the rate of deformation in
creases the rheological or viscous component of strength , but it also increases
the thixotropic decrease in strength . Proceeding from very low to very high
rates of deformation , the combined effect is at first a decrease in residual
strength ; thereafter the residual strength appears to remain fairly constant
over a rather wide range of rates of deformation ; finally the residual strength
increases at very high rates of deformation . Migration of water from or to the
shear zone increases the residual strength of normally consolidated clays and
decreases that of strongly overconsolidated clays ; therefore , the slope of the
stress - deformation curve after failure is also affected by time and the rate of
deformation .
When controlled or incremental stress type of loading is used , the rate of
deformation increases with increasing shear stress , Fig . 30 ; under the last
load increment , this rate may remain fairly constant for a considerable
period , but a progressive increase in the rate of deformation occurs at and
after failure with corresponding thixotropic changes in strength and changes
in the pore -water pressure . This type of loading tends to produce a decrease
in shear strength after failure of normally consolidated and lightly overcon
solidated clays , but such a decrease in strength may not always occur in un
drained tests on strongly overconsolidated clays .
Loading by a very low and constant rate of strain before and after failure
may reduce the thixotropic decrease in strength after failure to an insignifi
cant amount , and it facilitates equalization of pore -water pressures and
changes in void ratio . As a consequence , this type and rate of loading may not
produce a significant decrease in strength after failure in undrained tests on
overconsolidated clays or in drained tests on normally consolidated and lightly
overconsolidated clays , unless failure causes a permanent or actual disturb
ance of the soil structure . However , this type and rate of loading tends to pro
duce a decrease in strength after failure in drained tests on strongly overcon
solidated clays , WROTH ( 1958 ) .
There is need of detailed investigations of the influence of the rate of de
formation on the shear strength after failure , and there is an especially great
need of comparisons of the results obtained in very slow tests with controlled
stress and controlled strain types of loading .
9. TRIAXIAL CONSOLIDATION AND FAILURE CONDITIONS
The data and concepts presented in the foregoing sections are primarily
based on the results of confined consolidation tests and direct box or torsion
shear tests . Some comments on consolidation and failure relations obtained by
triaxial tests are made in this section , but these comments should not be in
terpreted as a complete review or summary of the subject.
Triaxial Stress Notation
In evaluation of the results of triaxial tests , it is generally assumed that the
stress distribution is uniform . Referring to Fig . 38 - A , O2 = 0'a is the effec
tive axial stress and ox = oy = or are the effective lateral stresses , and
these are assumed to be the principal stresses . In recent years the test re
sults are often expressed in terms of the first , second , and third stress in
variants
J1 = + 02 + 03 (52 - A )
J2 = 01°2 + 0203 + 0301 (52 - B )
33 = 1 °2 93 (52 - C )
or the equivalent octahedral stresses
Poct
53
v2
oz
97
oz
A
}
}
(
)
.
-
=
)
Vig
ogle
Poet
352
oz
53
Vals
Oz
03
=
-B
-
)
(
-
(
-
(
)
}
)2
)2
The octahedral normal stress also called the mean normal stress and desig
is
nated by For effective stresses triaxial tests these formulas take the
in
p
.
form
54
A
Oct
)
-
20
p
'
3
(0
)
-
=
and
Foct
54
(0
0
)
-
)B
(
-
for compression
o1
for extension tests
or
where tests and
=
=
o
.
Sources of Error Triaxial Tests
in
Changes external forces caused by piston friction and by the rubber mem
in
brane have been investigated considerable detail and are now fairly well
in
Some difficulties are still encountered accurate measurements
of
known
in
.
changes diameter and volume of the test specimen during test

in
or
The most disturbing
of
of
source error the influence the end restraint
is
the loading plates and the test specimen which

of
friction between the ends
,
may cause nonuniform distribution of stresses strains volume changes and
a
,
pore water pressures the triaxial test specimen brief review
of
theoret
in
A
-
ical and experimental investigations the influence of end restraint and other
of
by
the writer HVORSLEV

of
sources error contained discussion 1957

in
is
,
,
a
)
papers
by
and details are presented

of
this conference WHITMAN 1960

in
,
(
)
BISHOP ALPAN BLIGHT DONALD 1960 and SHOCKLEY AHLVIN 1960
),
.
(
)
-
Nonuniformity exists not only axial but also radial directions and the
in
in
pattern of the nonuniformities varies with the external stress conditions and
the state of consolidation of the test specimen therefore this pattern changes
;
during test Direct determination pore water pressures water contents

of
a
,
.
volume changes various parts of test specimen show that internal mi

or
in
gration of pore water takes place during undrained tests and also that volume
a
one part drained test specimen may occur simultaneously with

of
decrease
in
volume decrease another part of the specimen The experimentally deter

in
a
mined degree of nonuniformity at large strains much larger than that ob

of is
tained by theoretical analysis of the influence end restraint but the causes
a
of this additional nonuniformity have not yet been definitely determined

.
error may not too seri

of
of
be
The influence the above mentioned sources

-
ous when the strains are small but the difficulties encountered proper eval
in
,
uation of the test results increase rapidly with increasing strains Further in
.
vestigations of the magnitude of the nonuniformities under various stress

,
strain and consolidation conditions are needed order develop appropriate

to
in
,
methods of correction or establish practical limits for the reliability of re

to
sults of triaxial consolidation and strength tests

.
Definitions Triaxial Failure and Strength

of
by
The shear strength determined direct shear tests defined as the peak
is
strength Tf the stress strain diagram Fig similar definition may

in in
A
,
4
,
,
.
.
-
considering the stress conditions

by
triaxial tests
on
also be used certain

assumed failure or critical planes Investigation
of
on
the stress conditions

.
such planes is very useful , as in form of the vector curves by CASAGRANDE

and WILSON (1953 ) , but it does involve assumptions in regard to the inclina
tion of the planes . Such assumptions are avoided by defining the state of fail
ure in triaxial tests as the peak value of (01 - 03 ) = ( 0 - 0 ) , and this defi
nition is used in this paper unless otherwise noted .
The peak values of Tand ( o - on) occur for the same stress condition and
strain when the assumed inclination of the critical plane is the same through
out a test. However , these peak values may occur at different strains when
the inclination of the critical plane is assumed to be that furnishing the maxi
mum obliquity of the resultant stress and varies with the principal stress
ratio ; TAYLOR ( 1948 , 1955 ) .
The peak value of the effective principal stress ratio , oj / 03 , has been and
is still being used to define the state of failure in triaxial tests , although this
definition may lead to inconsistencies when the soil also is assumed to con
form to the Mohr - Coulomb failure criterion and when the Mohr envelope does
not pass through the origin ; TAYLOR ( 1950 ) . Eq . 9 - A may be written in the
following form
'
cos
op ( of - 04 ) - 2c
.
sin
(9
F
-
)
Ø
os
01
occur simul
03
03
of
which shows that the maximum values and
-
is (
'
)
/
taneously or the same strain when held constant during the test see
at
oh ;
o
'
HILF and GIBBS 1957 possible maintain constant value

to
of
in
It
is
a
.
(
drained tests and also some undrained tests on partially saturated soils
in
,
fully saturated clays
on
but cannot be held constant undrained tests

in
'oz
.
og
saturated and normally consolidated clays
on
During undrained tests de
,
creases with increasing strain and the maximum value of may then 09
o
On (
)
/
.
01
of
occur after the maximum value

attained the other hand
is
,
0
s
09 on -
.
(
'
)
increases during undrained saturated and strongly overconsoli

03
tests
dated clays and may then occur before the peak value
of
peak value
, oʻ
,
)
/
by
general definition
of
reached failure
is
of
of
the state the

to In
;
0
o
.
(
produce greater values

03
of
peak valuě
04
tends and smaller values

Ø
',
(
)
/
by
of
of
than those obtained when failure defined the maximum value

is
(o
c
) '
It 03
.
-
09
increase whereas
is
often observed that continues

to
,
;
0
o
(o
-
(
)
/
remains fairly constant during the last part of undrained triaxial tests
on
strongly overconsolidated clays may not be

of
maximum value
A
0
.
-
(
strains ordinarily used triaxial tests but maxi

or
at
reached attainable
in
value does exist according

of
the critical void ratio concept ROSCOE

to
mum
SCHOFIELD WROTH 1958 which discussed later this section
in
The
is
),
-
changing stress conditions during the last part test are very close
of
to
such
a
or
03
of
the failure envelope and the peak value any value

04
of
of
those
,
)
/
os
during the last part

of
the test combined with the corresponding

0
,
-
(
of )
may used for expressing the failure conditions ef

be
of
value terms
oſ
in
,
fective stresses
.
Two definitions of comparative strengths are used evaluating the influence

in
of various triaxial stress conditions the strength of clays The first may
on
.
be
of
termed the constant volume strength and the peak value for
09
is
(o
-
)
the same void ratio but varying stress conditions The second type of compar
.
ative strengths may called the effective stress strength and

be
the peak value

is
of ( 0 - 04 ) for equal values of os , or , considering stresses on assumed fail

ure planeš , value of Tf for equal values of os . Comparative effective stress
strengths indicate the relative position of the failure envelopes or shear
strength lines for various stress conditions .
These comments on the state of failure are generalizations and simplifi
cations of actual conditions , and there is need of further and more detailed
investigations of the definitions of the state of failure . These investigations
should take into consideration the influence of the type and rate of loading ,
the development of surface cracks , increasing nonuniformities in the test
specimen with increasing strains , the yield and flow limits proposed by
GEUZE (1960 ) and the possible existence of a limiting failure strain as sug
gested by GOLDSTEIN (1957 ) .
The Rendulic Diagram
Basic experimental relations between triaxial stress conditions , void
ratios , and pore -water pressures for normally consolidated clays were first
established by RENDULIC (1936 , 1937 ) . The importance of this classic work
has remained practically disregarded until recently verified and amplified
through a series of very valuable investigations at the University of London ,
HENKEL (1958 , 1959 , 1960 ) . The tests by Rendulic were performed in Vienna
clay , and the most complete series of tests reported by Henkel were made on
Weald clay . The two clays are quite similar , and their index properties and
the molding water contents used are given in the following table .
Table 4 . -- Index and Molding Water Contents
Vienna Clay Weald Clay

Liquid limit 47. 6 43
Plastic limit 22 . 8
com
Molding water content

.
o.
27 5
Liquidity index 0 . 19 0 64
Rendulic tests Both drained and undrained compression and extension

,
.
tests were performed on Vienna clay The test specimens were drained by
a
.
sand and mica mixture and the pore water pressures meas
of
central core
,
a
ured were those existing this core RENDU LIC 1937 states that correc
in
)
.
changes
of
of
tions were not made for the influence cross section the test
in
specimen with increasing strains Therefore data obtained for large strains
,
,
.
including failure conditions are not reliable The low initial or molding water
,
content should be noted since undoubtedly affected the test results REN
it
,
DULIC states that the molding water content corresponds pressure

to
1936
a
(
kg starting the liquid limit This state

of
cm2
at
confined consolidation
in
6
,
is /
ment probably misprint and not repeated the 1937 and more com
in
is
it
a
plete version the paper equivalent

of
The initial water content void

is
to
a
.
ratio of Fig corresponds equivalent con

76
10
an
and that
it
to
of is
seen
in
it
0
,
.
solidation pressure
kg
about cm2
2
.
/
very comprehensive and carefully executed series tri

byof
Henkel tests
A
.
at
Weald clay were performed the University

on
of
axial tests London Hen

kel and graduate students over period of five years This series included
a
drained and undrained tests compression and extension tests various types
of
,
both normally consolidated and overconsolidated

on
drained tests and tests

,
test specimens tests were performed London clay but

of
on
similar series
A
.
were confined to compression - type tests . It should be noted that the molding
water content and the corresponding value of the liquidity index were much
higher than those used by Rendulic . The test specimens were drained at the
end , and the pore - water pressures were measured there , but drainage was
aided by strips of filter paper on the surface of the test specimen . There
were minor differences in consolidation characteristics of the batches of
the
clay used at the beginning of the test series .
and end
The Rendulic diagram . The pressure - void ratio relations determined by
drained tests may be plotted in the conventional semilogarithmic diagram , and
an example of data obtained by Rendulic for all - round or hydrostatic pressure
is shown in Fig . 10. The points straight line but because
is lie
of
on
the low
,
a
molding water content this line considerably below and has flatter slope
a
,
than the virgin consolidation diagram obtained by confined consolidation and

an
initial water content close the liquid limit

to
.
ingenious method for comprehensive graphical repre
an
Rendulic devised
sentation of relations between void ratios and stress conditions triaxial
in
tests Points representing these stress conditions lie the plane of sym
in
.
metry BOC Fig 38 where the ordinates are and the abscissae are
ol
in
A
,
,
-
.
plane Rendulic plotted contours of equal water content
or
this void
In
O
V
2
.
of
ratio as shown Fig Weald clay
38
for states normal consolidation
of
in
,
B
-
.
.
The line DD and lines parallel thereto represent states
of
of
constant values
the mean effective stress or the invariant and these lines are tangent
p
,
',
,
J
their intersection with the line on which
at
the constant void ratio contours

to
,
represents states of all round stress or Rendulic found fairly good
oa
=
o
'r
-
agreement between the results of drained and undrained tests and he conclu
,
ded that any point the diagram represents unique relation between void
in
independent the stress path provided

of
ratio and stress condition which

is
,
this path does not cause temporary decrease void ratio

in
a
general
by
The above mentioned results and conclusions Rendulic have

in
-
been verified by Henkel The shapes

of
the contours for drained and undrained

.
tests are nearly identical but there are some cases minor differences
in
in
,
the stresses the position As yet

or
of
the contours uncertain whether

is
it
.
these differences are systematic or merely represent unavoidable differences

the properties of the test specimens but the possibility exists that further
in
investigations may show that the stress paths some cases may affect the
in
position of the contours

at .
Weald clay defined

, of
of
on
failure the maximum values

as
Values coa
,
,
)
)0g
and form two straight lines OE and

oj
Fig
38
are also shown

of
in
,
B
-
-
.
, (
OF which are envelopes the constant void ratio contours and pass through
to
the point origin when the test specimens are normally consolidated
of
These
.
lines correspond angle of incli

an
the Mohr Coulomb failure criterion and

to
by
be
and they were found Henkel independent

to
of
nation
m
22°10 the
, =
',
type test drained or undrained

of
of
and various combinations increases and

,
of
decreases
or
ca
and
The unified Rendulic diagram The constant water content contours for
a
.
saturated and normally consolidated clay appear have nearly identical

to
shapes for water contents appreciably smaller than the molding water content
.
replaced with
be
of
Therefore family such contours can single curve which

,
a
the in
of
the ratios oaoc and where the stress

at
function
is
is is
,
0
; 0
a
o o
, c
'
/
tersection of the contour and the line OA that the equivalent con
is
,
all
solidation pressure for hydrostatic pressure Such unified Ren

or
round
-
dulic diagrams for normally consolidated Vienna clay and Weald clay are
E DE
(o : + 20 =
W OR e CONSTANT
CONSTANT
900
SPACE
DIAGONAL
D
=
0
0 V2
2.0 4. 0 0 10 6.0
0,
Woonkbnog29 INTERSECTION OF C, A ,
AND OE , AT O ', =2.21 0
=
oC
'
RENDULIC (1937)
VIENNA CLAY
=
0
O
oc 117
.0
47
171.1
RENDULIC ENVELOPE
L -
TO - - - .
= OF
INTERSECTION 40
21 A
. B,
0 ,
To
OF
AT
CONSTANT AND
O
R
,
/
OF
CONTENT
sos RATIO
0
.8
SLOPE
..
1
:2
CURVES
0 VOID
WATER
OR
7
0
=
HENKEL 1958
|)0
(
WEALD CLAY
0
.6
LINE
+ OF
. 0
84
CONSTANT
00
+0 (=);
HENKEL
38
0
FAILURE ENVELOPES
.
090889TOSTI
0
.
+
Dz
RENDULIC ENVELOPE
est
to
0
.4
0
.6
0
.2
1
.0
1
.2
ORIC
Fig Diagram
38
Unified Rendulic Normal Consolidation

-
.
shown in Fig . 38 - C and are for convenience of operation plotted in the XZ

plane instead of the plane of symmetry . In the XZ plane the line OA has a
slope of 1 : 1 and the line DD a slope of 1: 2. The contours appear in a slightly
distorted form , but they are still tangent to the line DD in point A . There are
some irregular variations in the shape of individual constant water content
contours obtained by Rendulic , and the unified contour shown in Fig . 38 - C re
presents the average shape of several contours obtained for all - round consoli
dation pressures close to 4. 0 kg / cm2 . The unified diagram cannot fully re
place the family of contours , but it is convenient for a general discussion of
relations and for comparison of characteristics of different clays .
Failure envelopes for Weald clay are also shown in Fig . 38 - C but as func
tions of ca and or instead of the above -mentioned ratios . The corresponding
dashed lines for Vienna clay are envelopes for the contours obtained by Ren
dulic , but they are not the actual failure envelopes since the test data were not
corrected for the influence of changes in cross section of the test specimen
during the tests . Corrections for changes in cross section of the test speci
men , estimated by the writer , indicate that the failure envelopes for Vienna
clay are fairly close to those for Weald clay , but they do not pass through the
point of origin .
Influence of the molding water content . The unified diagram for Vienna
clay , Fig . 38 - C , does not represent the contours corresponding to low pres
sures or water contents approaching the molding water content , which resem
ble those obtained by Hendel for overconsolidated Weald clay . Rendulic also
found that a slight swelling occurred at and after failure in drained tests on
normally consolidated Vienna clay and , as indicated above , the corrected fail
ure envelopes do not pass through the point of origin . Data available to the
writer do not show any case in which a clay normally consolidated from a
water content close to the liquid limit exhibits a volume increase at and after
failure . The anomalous behavior of a normally consolidated Vienna clay is
probably caused by the low molding water content used by Rendulic . The
slight difference in unified water content contours of Vienna clay and Weald
clay before failure may indicate actual differences in soil characteristics , but
it may also in part be the result of differences in molding water content or in
the locations where the pore - water pressures were measured .
It is often convenient to use a low molding water content in the preparation
of triaxial test specimens of remolded saturated clays , but this practice may
cause abnormalities in the test results obtained . The influence of the molding
water content on volume change and strength characteristics of remolded satu
rated clays should be investigated , and proper limits should be established for
this water content or the liquidity index with respect to the consolidation pres
sures used in the tests .
Stresses and Void Ratios
The stress conditions at failure of normally consolidated Weald clay , ex
pressed in terms of the equivalent consolidation pressure o' and obtained
from the unified diagram in Fig . 38 - C , are summarized in the table on the
following page .
The unified diagrams in Fig . 38 - C and the numerical data in Table 5 rep
resent generalizations of several contours or tests , and individual test results
may deviate therefrom to a minor extent . However , it is believed that the data
are sufficiently reliable for a discussion of qualitative relations and trends.
As previously mentioned , Henkel found that the angle of inclination of the Mohr
envelope is practically the same for compression and extension tests ; that is ,
the effective stress strengths are nearly identical for the two test conditions .
On the other hand , the data in Table 5 show that the constant volume strength
in extension tests is only 81 per cent of that obtained in compression tests .
This reduction in strength is primarily caused by the development of greater
pore - water pressures in extension tests .
Table 5 . -- Stress Conditions at Failure of Weald Clay
Stresses Compression Ratio

Extension
. 0 38. of
ó.47 .
1 04 o
o
0 84
ó.
0 0
o
2011
0 . 69 1 . 04
ó
0 66 of 0
o0 .46
p
=
81
51 57
o?
. 0
o
.
0
=
q
(
) -
0
(0 )
= 0
17 77 00 81
62 67 21 22
o7
.0 0. 2.rio .1
0
.
0
. . ini1.
03 +
+
(
1
.
ó
/ 01
pi
/
53 87
0
0
.
9
1
O
.
0
Triaxial consolidation produces virgin consolidation diagrams which
in
,
a
semilogarithmic plot are parallel those obtained by confined consolidation
to
,
,
Fig provided the initial water contents are equal

10
often stated that

is
It
,
.
.
the void ratio primarily the major principal stress

of
or
function that
is
,
a
as
diagrams plotted
of
are nearly independent

of
the consolidation function

a
the principal stress ratio

of
which case the constant water content contours

in
,
should be straight and parallel the lines AB and AC Fig

38
seen is
to
in
It
.
from this figure and Table that the above mentioned statement nearly cor
is
5
rect for some stress conditions but oversimplification for other situa
an
at is
tions major principal stress failure compression tests

oi
The
in
=
is
0
04 .
close agreement with the statement but the major prin

oc
which
is
in
=
,
1
.
cipal stress For consoli

of
84
0 an
on
failure extension test

at
= in
to =
is
0
.
which corresponds
17 on
dation with the ratio confined consolidation

7
0
',
.
oc
clays for Weald clay and

of
12
for Vienna
oc
is of
=
17 =
=
to g
,
1
.
clay that the major principal stress per cent greater than that
12
is
;
required for all round pressure

in -
the equivalent consolidation pressure during

or or
The change void ratio

a
drained compression test obtained by extending the

is
constant
=
=
,
,
0
line C1A1 intersection with the failure envelope OE1 which yields of
to
=
,
oe also applies hence

04
but the relation

oc
21
af
=
2
1
d
,
.
.
'
off
21
or
04
ol
12
=
12
=
55
2
=
.
o
and
1
'
.
o
=
2
.
o
'
2
.
A
-
(
where the equivalent consolidation pressure

di
failure and ng
at
is
the
is
overconsolidation ratio The following relations for drained extension tests

.
are obtained similar manner

in
a
off
and
21
84
63
or
=
55
no
2
.
63
o
'
0
.
=
o
' 0
=
.
'o
2
.
(
- B
)
which shows that the volume change during a drained extension test is appreci
ably greater than that in a drained compression test on a normally consoli
dated clay .
The diagrams in Fig . 38 show that the void ratio is not a unique function of
the mean effective stress , p ' , or the first stress invariant , Jy , since the con
tours then should be parallel to the lines DD and D , D1 . Changes in void ratio
are a function of changes in both normal stresses and shearing stresses and
may possibly be expressed as a function of changes in the mean stress , p ' ,
and the octahedral shearing stress , Toot Development of mathematical ex
pressions for the complete Rendulic constant water content contours, or for
the change in void ratio caused by an arbitrary change in triaxial stress con
ditions , would aid the understanding of the failure criteria for clays and would
permit more reliable estimates of the settlement of structures founded on
massive clay deposits .
HENKEL (1958 , 1959 , 1960 ) presents a series of interesting correlations
between water contents and stress conditions at failure . Many of the diagrams
are similar to those used by RUTLEDGE ( 1947 ) and shown in Figs . 19 and 20 ;
but in other diagrams the relations are shown in arithmetical coordinates and
can be used for predicting pore - water pressures at failure in undrained tests
from the results of drained tests , or void ratio changes in drained tests from
the results of undrained tests .
Henkel correlates the water content with the effective mean stress , p ' . It
was mentioned in the foregoing paragraph that the void ratio is not a unique
function of p ' ; however , at failure there is a definite relation between oj , ok ,
and p ', and correlations between w or e and p ' at failure can then be estab
lished . Although the relations between oi , oi , q, and p ' at failure are quite
different for compression tests and extension tests , there is but little differ
ence between correlations of w and p ' for the two types of tests , because the
value of p ' at failure in extension tests is only slightly larger than that ob
tained in compression tests ; see Table 5 and line D2D2 in Fig . 38 - C . How
ever , it should be noted that whereas the value of p ' at the start of an un
drained test is equal to o'c, at failure p ' is only 0. 66 o'c in compression tests
and 0 .69 o'c in extension tests . When the effective mean stress , p ', is kept
constant during a test , the test specimen is subjected to pure shear , but a nor
mally consolidated clay undergoes a volume decrease as seen by inspection
of the families of constant water content contours for Vienna clay , Weald clay ,
and London clay . This result is contrary to that obtained by the conventional
theory of elasticity , but it has been corroborated by compression and extension
tests in which p ' actually was kept constant during the tests ; see Fig . 18 in
HENKEL (1958 ) . Strongly overconsolidated clays undergo a volume increase
during similar tests .
Stresses and Pore -Water Pressures
The pore -water pressure , u, created by an increase in external pressure
or total stresses during undrained tests on a normally consolidated clay can
be determined by plotting the point ( oa / oc , Or / oC ) in the Rendulic diagram ,
Fig . 38 - C , and drawing a line through this point parallel to OA until it inter
sects the unified contour . The horizontal or vertical distance between the two
points is equal to u / oc . The pore -water pressures at failure in compression
and extension tests are shown in Table 5 . A similar graphical determination
of the pore -water pressure can also be used for overconsolidated clays pro
vided the contour for the particular water content and state of overconsolida
tion is available . As previously mentioned , other diagrams proposed by
HENKEL ( 1958 , 1959 ) can also be used for prediction of pore - water pres
sures .
SKEMPTON ( 1954 ) proposed that the change in pore - water pressures at
failure , Au , can be determined by the coefficients Af and Bf and the equation
(56 - A)
su = Bpl6oz + Aplon - 40z )]
The coefficient Bf is close to unity for saturated clays , and the equation is
generally used in the simplified form
soz
407
103
(Ap
56
Au =
B
+
-
)
)
Skempton also suggested that the equation be given form which indicates the
a
of changes
the mean principal stress changes
of
influence
oz
instead
in
by in
p
,
,
.
HENKEL 1958 1960 generalized the above mentioned equations in
,
B
-
-
(
or
troducing stress invariants octahedral stresses and suggested the follow
,
ing expression
902
103
ac
Viso
Ap
10
10
32
sole
,
,
Au
12
57
(
(
)
-
-
=
-A
be
which also may written the shorter form

in
4p
su
57
3a
=
Atoct
+
B
-
(
)
There considerable difference values of Af for compression tests and
in
is
extension testswhereas corresponding differences values of af are much

in
,
smaller and some cases negligible Therefore Eqs 57 have greater gen
in
,
.
eral validity than Eqs and they represent an advance

56
available means
in
,
.
for predicting pore water pressures

-
The coefficients Af and

af
vary with the consolidation characteristics and

the stress history the clay
of
These coefficients decrease with increasing

.
overconsolidation are zero for the critical degree of overconsolidation and

,
become negative for strongly overconsolidated clays The pore water pres
.
sures shown Table refer test specimens which have been normally
to
in
,
,
5
,
u
consolidated under all round pressure and then stressed failure Test data
to
-
anisotropically consolidated specimens Weald clay are not available

of
for
to
the writer but assuming that the unified Rendulic contours also are valid for
,
this case estimated that the values of are appreciably smaller and
is
it
,
those of af are considerably larger than the corresponding values for isotrop
ically consolidated test specimens RENDULIC 1937 found that stress re
a
.
versal or
of
decrease A0 Aon does not produce

of
decrease the pore

a
in a
(
water pressure but this pressure case the test speci

of
small increase
a
normally consolidated Likewise the writer found HVORSLEV 1937

is
men
,
,
.
stress reversal during Vienna clay pro

on
that drained direct shear test

a
slight additional decrease

of
duces the void ratio These findings are

in
a
agreement with Eqs 57 since the sign

of
stress change does not affect the

,
, a
.
the change pore water pressure but

of
sign should be noted that the value

af in
it
-
of
the coefficient quite small for stress reversal

is
view of the above mentioned large variations the pore

of
the values
In
in
-
pressure coefficients with the stress conditions and stress history

of
the soil
, ,
not possible provide reliable numerical data for practical applications

is
to
it
but the pore pressure coefficients must in each case be determined by tests
which simulate the stress history and conditions of the prototype soil as
closely as possible . Much more research on the problem of pore -water pres
sures is needed , and development of more comprehensive mathematical ex
pressions for the triaxial consolidation characteristics of clays, mentioned
in the foregoing subsection , will also assist in solving this problem .
The Critical Void Ratio Concept
ROSCOE - SCHOFIELD -WROTH ( 1958 ) have advanced a very important con
cept concerning the existence of a yield surface and a critical void ratio line
for remolded saturated clays which are subjected to loading by means of a
constant and low rate of increasing strain . With this type of loading the clay
can pass through a yield or failure point without collapse, and continue to de
form while the path of stresses and void ratios follow a yield surface until a
critical void ratio is reached . Thereafter additional deformations take place
without further changes in void ratio , pore - water pressure , and stress condi
tions , provided the rate of strain is not changed . Very large strains may take
place , particularly for overconsolidated test specimens , before the critical
void ratio line or ultimate state is attained .
WROTH (1958 ) states that the peak shear strength or failure condition co
incides with the ultimate state on the critical void ratio line for normally con
solidated clays, but the peak shear strength observed in drained tests on
strongly overconsolidated clays occurs well before the ultimate state is
reached . That is , a part of the decrease in void ratio and corresponding in
crease in strength caused by overconsolidation is active at the moment of fail
ure , and the final increase in void ratio and elimination of the effects of over
consolidation occurs after failure . The peak shear strengths referred to are
those actually observed plus a surface energy or dilatation correction , which
can be quite large before and still be significant at the peak shear strength ,
but this correction is zero for the ultimate state where changes in void ratio
or pore - water pressure cease .
The authors express the results of triaxial tests by means of the variables
p ', q , and e ; the latter may be replaced with the water content , w , for saturated
clays . With these triaxial coordinates the yield surface for overconsolidated
clays is represented by the equation
,
exp
ūp '
Be
58
q = + ē
)
(-
)
(
which identical Eq 33 and the coefficients have the same relative

to
is
form
in
to ,
.
meaning but are shown with bar indicate that they refer triaxial consoli
to
a
dation and stress conditions However the coefficient Te not coefficient

is
a
it ,
.
does not indicate the ratio of shear

of
effective internal friction because

failure surface The member Tep may
be
stresses normal stress

to
in
a
'
called the normal stress component and the member exp Be the effective
C
at ,
)
-
void ratio component The yield surface terminates the critical void ratio
The pro
.
qw
.ey
line where the variables attain their ultimate values piv and
,
is ,
,
jection of the critical void ratio line

by
plane
on
the represented the

p
(e
,
)
equation
Pů
ew
exp õlen
Pń
59
=
A
-
[
)
-
by
and the projection

of
the plane
(p
q
,
)
qu = ūg Pu (59 - B)
where (en , ph ) is a known point on the critical void ratio line . Some of the
symbols in these equations are different from those used by Roscoe - Schofield
Wroth . The change was made to obtain agreement or avoid conflict with the
symbols used in this paper . When the projection on the ( e , p ) plane is plotted
in semilogarithmic coordinates , the critical void ratio line appears as a
straight line which is parallel to the virgin consolidation diagram obtained by
all - round triaxial pressure , Fig . 39 - A . These lines correspond to AA and BB
in Figs . 19 and 20 , but the critical void ratio line is common to normally con
solidated and overconsolidated test specimens . Eqs . 59 - A and 59 - B correspond
to Eqs . 21 and 24 for states of normal consolidation .
Published data on triaxial tests on clays , performed with controlled - strain
type of loading , often do not conform to the critical void ratio concept , but
Roscoe - Schofield -Wroth point out that the tests in such cases were terminated
before the ultimate state was attained , and PARRY (1958 ) demonstrated that the
changes in void ratio or pore - water pressure , at the time the tests were termi
nated , were in direction of the critical void ratio line . This line has not yet
been reached in compression tests on strongly overconsolidated clays , but to
do so requires very large strains with consequent uncertainty in proper evalu
ation of triaxial test results . Furthermore , strongly overconsolidated clays
undergo swelling during drained tests , and Roscoe - Schofield -Wroth suggest
that the volume changes in this case may tend to become concentrated in a rel
atively narrow zone , and that the void ratio in this zone is greater than the
average void ratio of the entire test specimen . This possibility constitutes
another probable source of error in triaxial test , which should be investigated .
The data obtained by the writer in tests on Vienna clay and with controlled
stress type of loading , Figs . 19 and 20 , diverge appreciably from the critical
void ratio concept , especially for strongly overconsolidated test specimens .
The probable cause of this divergence is that the rate of strain increases with
increasing stress and creates increasing thixotropic reduction in strength so
that complete failure occurs before the ultimate void ratios are attained . It
was observed that the void ratio of normally consolidated test specimens de
creased after failure , even with test durations of several weeks , and that the
void ratio of strongly overconsolidated test specimens increased after failure .
Reference is made to Fig . 36 , which shows that the void ratio decreased from
25 per cent at to 24 per cent after failure with a corresponding increase in
strength after cessation of deformations and the thixotropic disturbance .
The concept of the existence of a critical void ratio line produces a great
simplification in the determination of ultimate void ratios and strengths in
cases where the rate of strain is limited by natural conditions until the ultimate
state is attained , and such cases undoubtedly occur in the field . However , there
are other cases where actual failures have a catastrophic character , and where
the void ratio and stress conditions at and after failure may be more nearly
simulated by those obtained in tests with controlled stress type of loading . To
the writer ' s knowledge , published test data do not permit a reliable comparison
of failure and ultimate stresses , strains , and void ratios obtained by very slow
tests performed with controlled stress and controlled strain types of loading ,
and this problem merits careful investigation .
Roscoe - Schofield - Wroth expressly limit their investigations and concepts to
triaxial compression tests , o > o , = 0', . It is hoped that these investigations
VOIDRATIO
.4
ULTIMATE STRENGTH
ALL ROUNDPRESSURE
-
0 CVR LINE
(eu,
Pi
qu
Fs
=
:
pues
se
CVR LINE CONSTANTI
FROMPAPERBY
FIELD WROTM(1958
ROSCOESCHOE
)
-
-
VO ULTIMATE MEAN STRESS
PA
P
LOG MEAN STRESS
..
LOG
P
Fig Critical Line for Compression Tests

39
The Void Ratio

.
.
'
RIGHT COMPRESSION TEST
A
ID
SECTION MOHR
1
72
EXTENDED MOHR
); 1:
VONMISES COULOMB
SLOPE
- =
0
D
0
,
(0,';
EXTENDED
=
DIAGONAL
TRESCA
0
SPACE
=
•
Da
EXTENSION TEST
MOHR
_
0
0=;;
-
-
-
OR -
-
-
MISES
>
TRESCA
K
os
D2
V2
v2
V2
03
=
;
=
0
0
FAILURE ENVELOPES RIGHT SECTION

D
D
A
B
-
-
-
MOHR COULOMB HAYTHORNTHWAITE

-
KIRKPATRICK
POSSIBLE
FAILURE
-
OR SURFACE MOHR
=
Ø
O
TRESCA COULOMB
UU
11
EQUATIONS
38
33
VONMISES AND
)
(
)
(
-
COHESIVE SOILS COHESIONLESS SOILS

C
-
0
-
Fig Principal Stress Space

40
Failure Surfaces the

in
.
will be extended to extension tests , oi = 0', > o's , and to tests with inter
mediate values of ob . If this is done , it will undoubtedly be found that the
concept of a critical void ratio line is valid for all stress conditions . How
ever , it is probable that some of the coefficients or parameters in Eqs . 58
and 59 will vary with the relative value of 0%, since this stress has the same
influence on the variables p ' and pi, as oy and o , but appears to have rela
tively little influence on q and qu , as discussed in greater detail in the follow
ing two subsections .
Triaxial Failure Conditions
The basic Coulomb failure condition , Eq . 3, may be combined with the Mohr
failure condition and expressed in terms of the principal stresses by the vari
ous forms of Eqs. 9, known as the Mohr - Coulomb failure criterion . The
amended Coulomb failure condition with the void ratio as an additional vari
able , Eqs . 29 and 33 , is expressed in terms of principal stresses by Eq . 38 in
accordance with investigations by Skempton , Bishop , and Henkel. Finally ,
Roscoe , Schofield , and Wroth have proposed Eqs . 59 as the ultimate failure
condition at large deformations and a constant , slow rate of strain . These
equations are not mathematically complete , and more comprehensive expres
sions for the failure criteria and their representation by lines and surfaces in
the principal stress space are discussed in this subsection . The general ob
jective is to examine the limits of validity of the criteria , or to obtain expres
sions which are valid for all stress conditions without change in form of the
equations and the numerical value of the coefficients or parameters .
Failure Surfaces . Comprehensive triaxial failure or yield criteria were
first developed for metals , crystalline rocks , concrete , and other materials
and are described in various textbooks on strength of materials and theories
of plasticity ; for example , NADAI ( 1950 ) , PRAGER and HODGE (1951 ) , and
TIMOSHENKO (1956 ) . Application of these theories to soils and determination
of the corresponding failure surfaces in the principal stress space are de
scribed in the excellent papers by RENDU LIC ( 1938 ) and KIRKPATRICK
( 1957 ) . These theories and their graphical representation are summarized in
the following paragraphs in order to provide a basis for the subsequent dis
cussion .
Fig . 40 - A shows a section of symmetry in the principal stress space , simi
lar to that in Figs . 38 - A and 38 - B. The axes and principal stresses can in
turn become the major , intermediate , and minor principal axes and stresses .
It is assumed that the failure lines O B and OC are straight and represent
the Mohr - Coulomb failure condition , Eqs . 9, and that the parameters 0 ' and
c ' are constant or independent of the value of the intermediate principal stress .
The right section DD is shown in Fig . 40 - B , where D1 , D2 , and Dz are the
points of penetration of the three axes ; B , represents failure in compression
tests with o > 0 = 0',, and C, represents failure in extension tests with
o'y = o'z > 0 . It may be noted that the change from B , to C , involves a
change in direction of the major principal stress . The points B2 , C , and B , ,
Cz are obtained in a similar manner ; for example , the point B3 represents
failure for the condition of = 0! > on . It can be shown mathematically that
the straight lines connecting these points represent failure conditions for inter
mediate values of the second principal stress . The failure surface correspond
ing to theMohr - Coulomb condition of failure is then a pyramid with the irregu
lar hexagon B , C , BzC , B , C , as base or right section . This surface is repre
sented by six equations similar to Eq . 9 - A or variations thereof , which can be
combined into the following complete equation
[ ;- ) -(2c'
cj2
gºje
oğlan ein
cos
603
og
)
' ] ]
( [ [
c
-; -
"
6
( (0
60
opje
gove
con
03
-(
'
e
"
4
co
+
]guja
oin
cos
03
o;
60
2e
"
-(
(
)
-o
The subscript formerly used for the Mohr equations has been omit
in
m
,
Ø
,
,
'
point iso
at
the all round stress
at
ted for convenience the stresses
If
A
-
.
point are
B
RV
RV
and
,
and
R
RX2
,
-
.:
=
!
+
0
12
/3
.'
=
-!
0
7
/6
.
(0=
R
-
)0
/ 3
point
at
and
C
.Eq RV
and
RV1
,
0,
12
=
!
-
=
/
0
+
.
2
3
'
/
0
=
.
6
(
-
)
1
0
R
/3
which inserted yield the following relation between the radii
in
=
R
A
9
-
AB for compression tests and Re AC for extension tests

=
sin
R
)61
1
-
'
(
sing
writ
+
1
R
when this parameter

be
of
noted that this ratio independent

is
should
It
',
c
unity with decreasing values

of
constant and that approaches as
is
it
0
a
shown Table
'
in
6
.
Ratios Rc and Re for Mohr Failure Surface

of
Table Coulomb
6
.-
-
-
gi
10°
15°
20°
25°
30°
35°
40°
0°
5°
796
R
000 943 889 842 753 715 678 647

/
0
.
000 061 123 188 258 328 400 473 546

R
1
.
.
/
For the Mohr Coulomb failure criterion reduces

to
=
Ø
-
'
62
=
O
'c
(
)
o
-
and five corresponding equations which can be combined general single

in
a
equation similar Eq 60 This the Tresca or Guest failure condition and

is
to
,
.
.
the corresponding failure surface prism for which the right section
is
is
,
a
regular hexagon and Re Rc The following failure condition

=
63
ūp
ūklo
(
)
=
(
+
-
0
)
0
+
'
+
c
+
0
)
0
'c
'
and the five corresponding equations have a failure surface which is a

pyra
mid with a regular hexagonal right section . Eq . 63 is called the extended
Tresca failure condition , but it is credited to Sandels by JOHANSEN ( 1958 ,
p . 32 ) .
Theoretical and experimental investigations by von Mises , Hencky , Huber ,
Lode , Nadai , Schleicher , and others lead to the following failure condition
(0 - 0 )2 + (0 - 0 )2 + (0 - 0 )2 = constant
= constant (64 - A)
which generally is called the von Mises yield or failure condition and applies
to many ductile metals or plastic materials . The failure condition is also
given in the form
(0 - 0 ) - (0 - 0 )2 -
103
501
-B64
)];
-
0
2)
=
' +
[c
0
+
+
0
(
)
as
of
known the extended von Mises condition The members these equations
.
can easily be replaced with stress invariants or octahedral stresses and the
,
failure condition may then be expressed the more general form
in
64
flooct
-C
Foct
)
=
which generally credited Schleicher These equations are symmetrical

to
is
respect principal stresses the corresponding failure surfaces are

to
with the
;
as
surfaces revolution with the space diagonal axis and the right section
of
is
,
circle which circumscribes the Tresca hexagon Some of the above

a
mentioned equations were originally developed as yield criteria and later also
used as failure conditions
.
Cohesionless materials Comprehensive investigations of triaxial failure

.
conditions for soils have primarily been made of interest

on
sands and
is
it
,
review the results obtained KJE LIMAN 1936 performed tests sand by
on
to
(
.
means of equipment which permitted independent control of the three principal

stresses cubical test specimen The results of compression tests
on
>
a
o
,
.
0 . ate og
corresponded
of
35º whereas the results tests for intermedi

to
=
=
Ø
,,
0
'
the second principal stress corresponded

of
values
oi
= to
>
>
=
o Ø
,
),
0
o
'
43° may be noted that Kjellman did not perform extension tests
It
o
,
s
'
>
BISHOP and ELDIN 1953 found good agreement between values ob

, of
Ø
(
drained compression and extension tests

on
tained sand However BISHOP

in
the possibility that may slightly larger for inter

be
1954 calls attention

to
Ø
(
'
mediate values of the second principal stress and support thereof he refers
in
by ,
theoretical investigations the problem the re

of
CAQUOT 1934 and

to
to
(
sults of preliminary investigations the University of London

at
Further
ex
.
periments with the recently developed apparatus for conducting strength tests
plane strain may yield data which will clarify the problem
of
under conditions
.
KIRKPATRICK 1957 developed equipment for triaxial tests

on
thick walled
(
cylinders sand which permits control the second principal stress

of
of ' of
He
,
found close agreement between values

of
for compression and extension

Ø
tests but obtained slightly larger values for intermediate stress condi
Ø
,
'
The corresponding right section

of
tions the failure surface circumscribes

.
the irregular hexagon of the Mohr - Coulomb surface and is shown by the
dashed lines in Fig . 40 - C . These results are in part corroborated by the
above - mentioned limited test data by Kjellman . In contrast thereto , HABIB
(1953 ) , PELTIER (1957 ) , and HAYTHORNTHWAITE (1960 ) found values of Ø !
in extension tests which are considerably smaller than those obtained in com
pression tests . Haythornthwaite concludes that the Mohr - Coulomb failure con
dition in some cases may furnish unsafe values and suggests that it be re
placed with a criterion based on the triangular surface B1 B , B2 in Fig . 40 - C .
These very substantial differences in results obtained by various investigators
emphasize the difficulties encountered in performing triaxial extension tests
and the need of further investigations of sources of error in triaxial testing
equipment and procedures .
Cohesive materials . Compression tests on marble by Kármán and exten
sion tests on the same material by Böker , as reported by RENDU LIC ( 1938 )
and JOHANSEN (1959 ) , show that the effective stress strengths in extension
tests are 7 to 9 per cent greater than those obtained in compression tests .
JOHANSEN ( 1959 ) also presents an excellent review of failure conditions for
concrete and arrives at the conclusion that these conditions in general agree
with the Mohr hypothesis and that there is so much scatter in the test data that
it cannot be stated with certainty whether the intermediate principal stress
has an appreciable influence on the strength .
RENDULIC (1938 ) reviews the major yield and failure hypotheses and cor
responding surfaces in the principal stress space , but states that his com
pression and extension tests on Vienna clay do not furnish reliable data on the
failure conditions, since the changes in cross section of the test specimens
were not determined . On the other hand , he estimates the yield conditions on
basis of the shape of the stress paths or contours and suggests that they are
represented by a hyperbolic surface of revolution or a variant of the von Mises
yield condition , Eqs . 64 .
TAYLOR ( 1948 ) reiterates a general belief that intermediate principal
stress has a minor influence on the strength , and that values of Ø ' obtained in
extension tests are about 10 per cent greater than those for compression tests .
TAYLOR and CLOUGH ( 1951 ) summarize the results of compression and ex
tension tests on undisturbed samples of Cambridge clay . It was found that the
undrained or constant volume strength in extension tests is as much as 20 per
cent smaller than that obtained in comparable compression tests . This differ
ence in strengths is correctly attributed to differences in pore -water pres
sures at failure , see Table 5 , and it is also suggested that these pressures are
a function of the change in the mean principal stress . However , the shear
strength lines for critical planes , Fig . 14 in the report , show that the effective
stress strengths for extension tests are equal to or slightly larger than those
obtained in comparable compression tests . In a later review of these test re
sults , TAYLOR (1955 ) , it is stated that the batch of clay used for the extension
tests had physical characteristics which under comparable testing conditions
gave about 10 per cent lower strengths than the batch of clay used for the com
pression tests . Therefore , had all the tests been performed on clay with iden
tical strength characteristics , the effective strengths for extension tests would
probably be perceptibly greater than those obtained in compression tests .
HIRSCHFELD (1958 ) performed undrained triaxial compression and exten
sion tests on undisturbed samples of three different clays . The results ob
tained for normally consolidated test specimens are expressed by the effective
principal stress ratio at failure , o / on , and show considerable scatter
because of variations in the physical properties of the individual test speci

mens . Values of of / 0 , for extension tests were in some cases greater and
in other cases smaller than those for compression tests , and definite conclu
sions in regard to the influence of the intermediate principal stress could not
be formulated .
The extensive series of triaxial tests on remolded Weald clay , HENKEL
( 1958 , 1959, 1960 - A , 1960 - B ) , include both drained and undrained compres
sion and extension tests of various types . The effective stress envelopes for
normally consolidated test specimens were practically identical for all types
of tests . For overconsolidated test specimens the value of 0 ' for extension
tests is about 0. 5° smaller than that for compression tests ; however , the small
cohesion intercept , c ' , for extension tests is about 50 per cent greater than
that for compression tests .
Referring to Fig . 38 - C , Table 5, and Eqs . 55 , the relative values of os or
the overconsolidation ratios at failure in extension tests are greater than
those in compression tests . According to Eqs . 33 and 38 the effective shear
strength lines for extension tests should then lie slightly above those for com
pression tests . This problem is discussed in greater detail in the following
subsection .
Definite conclusions cannot be formulated on the basis of the test data sum
marized above , but it seems probable that the Mohr - Coulomb hexagon repre
sents the inner limit of the failure surface for clays , Fig . 40 - D. Some of the
test data and consideration of the void ratios at failure indicate that points re
presenting extension tests may lay slightly outside the Mohr - Coulomb hexa
gon . Strength data on clay for intermediate values of the second principal
stress are not yet available , but considering the results obtained by KIRK
PATRICK ( 1957 ) for sand , Fig . 40 - C , it seems possible that the actual failure
surface for clays may have a curved form as shown by the dash - line diagram
in Fig . 40 - D . With decreasing values of \ ' , the Mohr -Coulonb hexagon ap
proaches the regular Tresca hexagon , and it is possible that the strength of
plastic clays with very small values of Ø may be represented by the von Mises
circle and extended failure conditions , Eqs . 64 .
Comments on the Variables
Stresses on principal or critical planes . The results of triaxial tests are
obtained in terms of principal stresses , and the required number of assump
tions is generally reduced when these stresses are used in the evaluation and
application of the test results . However , in many investigations and practical
applications it is desirable or necessary to express the test results in terms
of stresses on theoretical or assumed critical planes or planes of failure . The
angle of inclination of such planes , a, is generally determined by Eq . 8, but the
assumed values of 0 ' vary . As previously mentioned , minor variations in the
values of a have very little influence on the position of the computed shear
strength line , but such variations may cause appreciable differences in the in
dividual values of of and af and in correlations between these stresses and
consolidation pressures , void ratios , or water contents .
Stress invariants . The use of stress invariants or the therefrom - derived
octahedral stresses often appreciably simplifies the expression of relations
between stresses , strains , and volume changes , and also simplifies certain
yield and failure criteria , such as those by Tresca and von Mises , Eqs . 63 and
64 . The complete Mohr - Coulomb failure condition , Eq . 60 , which is symmet
rical with respect to oj, 0 , and o's , can also be expressed in terms of the
three stress invariants , but the resultant equation becomes very complicated ,
as emphasized by JOHANSEN (1958 ) . However , COLEMAN ( 1960 ) has recently
suggested a relatively simple expression in terms of stress invariants for a
curved failure surface passing through the six points of the Mohr - Coulomb
hexagon , Fig . 40 . The curved and continuously convex failure surface shown
in Fig . 40 - D may possibly be represented by an equation similar to the one
proposed by Coleman . Stress invariants or octahedral stresses should not be
used in the partial Mohr - Coulomb failure condition , Eqs . 9, which do not con
tain the intermediate principal stress .
The consequences of using the first stress invariant , , or the mean Ji
stress , p', in expressing the results of individual series of strength tests on
soils which conform to the Mohr - Coulomb failure criterion may be domonstra
ted as follows . The intermediate principal stress may be expressed by
(
03 = 2 01 + 0 ,) + 0 - 03) (65 - A )
where n is a coefficient which varies between - 1 and +1 . The mean principal

stress is then
p ' = ( +
,)
0 + (0 - 0 ) (65 - B )
and
(0; + o ) = 2p -
5401
65
03
-
?
C
-
(
)
which introduced the partial Mohr Coulomb failure condition Eq
in
,
A
9
.
-
-
yields
sing
cos
'
66
p
2
'
".
. +
sing
+2
A
c
singl
)=
-0
-
(
)
'(
i0
3
+
1
this equation the coefficients for and are not constant but vary with the
In
p
c
'
'
the intermediate principal stress

or
or
For compression tests
of
value
,
n
.
Eq
66
and becomes
=
1,
=
A
',
n
0
-
.
o
-
z
'
Ec to
66
Colo sin
Tºj
cos
B
2
_
2
'
-
(
)
1 _
singine
'
°3!)
sing
=
-
' c
.
1
1
.
or
and for extension tests and

=
=
+
,
,
;
0
1
n
si
2
c
cos
+ +
0
sing
-
',
(66
at
p
-C
)
IH
'
2
+
.
2
c
-0
)
(0
=
by
for equal values

of
The strengths obtained these equations and have the

p
c
'
'
following ratio
sin
"
}
67
.1
+
-0
'e )
(0
)
(
sinf
01
1
-
(
which is identical with Eq . 61, as it should be since Eq . 66 - A has the same

form as the extended Tresca failure criterion , Eq . 64 - B , for which R , = R . ,
whereas R , > R , for the Mohr - Coulomb criterion . If the coefficients for c
and p ' in Eq . 66 - A are determined by compression tests and used without
change for estimation of the strength in extension tests , the result will be con
siderably greater the actual strength ; for example , for Ø ' = 25° the over
than
estimation amounts nearly 33 per cent ; see Table 6 . Likewise , plotting
to
strengths or principal stress ratios at failure versus corresponding values of
Ji or p ' instead of ( o + 0' ) or o' may cause a considerable increase in the
spacing of the diagrams representing compression and extension tests , with
consequent difficulty in utilizing the correlations for intermediate or unknown
values of os .
· Weighted intermediate stress . In cases where the actual failure conditions
are between those represented by the Mohr and von Mises failure criteria ,
unified expressions or diagrams cannot be obtained by use of principal or
octahedral stresses or stress invariants . JOHANSEN (1958 ) proposes , in such
cases , that the stresses in a plane for which the direction cosines of the nor
mal are ( 1, Q, 1) / 12 + a2 be used as variables . These stresses are
(68 - A )
and
8
;-
cję
-;
og
;
;
a°
Velo
J2
]?
o
+
–0
(0+
68
B
-
(
)
a²
+
2
and the proposed failure condition then

is
Floa 69
-A
ta
,
)
K
)
which may simplified observing that Tais approxi

be
by
for small values

of
,
,
a
Tm
mately equal
or
to
0
(
0
-
)
69
Im
Floa
,
B
K
)
)
-
(
Eqs 69 are similar the extended Tresca and von Mises criteria but with
to
.
weighted principal stress As previously men

of
values
the intermediate
.
tioned von Mises and Böker found that the Mohr envelope for extension tests
,
marble lies above the one for compression tests JOHANSEN 1958 plotted
on
)
Tm
as
using
of
these test data functions and and found that the two
=
,
,
0
4
o
envelopes then merged into single curve The method proposed by Johansen
a
may be used advantage case further experiments

ac
to
should show that the

in
tual failure surface for clays lies between those corresponding the Mohr and
to
von Mises failure criteria

.
Eqs
by
as
As suggested
29
33
Cohesion function the void ratio and

of
,
,
a
.
38
the Coulomb failure criterion may be expressed

as
the cohesion func

in
,
tion of the void ratio order take the stress history of the soil and the cur
to
in
vature of the shear strength lines into consideration Fig 38 shows that the
C
-
.
.
of
failure are nearly identical for compression and

of
03
values and
at
0
s
'
/
extension tests on normally consolidated test specimens of remolded Weald

clay . However , in undrained extension tests the normal stresses and shear
stresses at failure are 81 per cent of those for undrained compression tests ,
see Table 5, but the void ratios and equivalent consolidation pressures are
identical for the two test conditions . This means that the void ratio in exten
sion tests is smaller and the equivalent consolidation pressure is larger than
in compression tests with the same effective stresses . In the case of drained
tests , the effective stresses at failure are nearly identical , but the equivalent
consolidation pressures at failure in extension tests are considerably greater
than those in compression tests , Eqs . 55 . Therefore the cohesion component
and the effective stress strength in extension tests should be greater than in
compression tests , or the values of the strength parameters , Ø . and k, must
be different for compression tests and extension tests . The following values
of the strength parameters for Weald clay were obtained by PARRY ( 1956 ) and
are quoted by HENKEL (1958 ) .
Table 7 . - Effective Friction and Cohesion Parameters for Weald Clay
Not Corrected for Corrected for

Surface Energy Surface Energy
Type of Test e к
K
Compression tests
Drained 180 0 . 05 - 20° 0 . 02
Undrained 200 0 .02 - 200 0 . 02
Extension tests
Drained 17 .50 0 . 04 .
19 50 0 .03
Undrained 166 0 . 06 160 0 . 06
Values of difor extension tests are slightly smaller and those of k tend to be
slightly larger than corresponding values for compression tests . The differ
ence between the results of drained and undrained tests decreases when cor
rections are made for the surface energy corresponding to the rate of volume
changes at failure but, as previously mentioned , it is also possible that this
difference in test results in part may be due to remanent pore -water pres
sures in drained tests and slight errors in measuring the significant pore
water pressures during undrained tests .
The influence of constant values of the parameters We and Kon the effec
tive stress strengths in compression and extension tests may be estimated as
follows for normally consolidated Weald clay . It is assumed that We = 189 ,
K = 0. 05 , and that the inclination of the failure planes is a = 45 - 60e . The
shear strength is then , according to Eq . 33 ,
325
70
tan
05
Tp = of
%
(
)
+
+
0
0
o
'
.
.
.
k
Assuming
21
for compression tests

oz
of
Table the normal effective

is =
,
2
5
/
stress failure
at
sin
71
42
poster
(
)
-
-
+
=
o
o
.
.
(
)
(
)
1
For drained compression and according Eq

on
55
tests
to
=
0
.,
A
.
-
0 = 2 12 o. ' 2. o'z
= 12
of by use of Eq. 71
100
72
ore
.12
42
49
(
)
=
=
0
2
1
/1.
.
shear strength drained compression tests then
is
in
The
OP
73
49
Tfc 325 005 400
=
=
0
0
x
.
(
)
.Eq
Eq
tentatively assumed
71
that also valid for extension tests and
- is
is
,
55 It
.
then yields
B
63
42
85
=
O
0
2
0
/
.
.
and
Tp
74
05
85
325 418
=
)
(
0
0
x
.
.
or
This equation corresponds
26
ok
22°41 and
to
418
=
tan
=
=
2
o
0
ø
.
.
s
'
/
s
'
.
'
Recomputation with this value principal stress ratio yields

of
the
and
43
26
oC
84
69
88
on
=
=
1
1
o
'
.
.
.
.
:
shear strength drained extension tests

is
The then
in
75
Tre
05
88
325
=
+
0
=
.
419
.0
.1
x
)
0
0
.
The strengths expressed

73
is by
Eqs are comparable

75
and effective stress

.
strengths and their ratio

,
Speltte
76
419 400 048

=
=
0
/0
1
.
(
)
That the effective stress strength extension tests per cent higher
is
in
is
4
8
,
that compression tests undrained tests for both compres

In
=
in
than
o
o
.
Tfc
varies also
is
sion and extension tests and but the ratio Tfe 048
,
,
1
o
'r
.
/
when comparisons
os
Similar computations
of
are made for the same value

.
ge
02
.02
for 20° and yield Tie Tfc

=
=
= in in K
1
.
.
/
As mentioned much greater values

of
Section are obtained direct

in
,
on K
5
shear tests than triaxial tests Direct shear tests Vienna clay yielded
.
Assuming for the purpose

of
10
de 17°30 and illustration that these

=
0
.
.
'
values of the parameters are valid for triaxial tests clay with character
on
a
istics similar those of the Weald clay and taking the shape of the Rendulic
to
curves for Vienna clay into consideration Fig 38 estimated that the
is
C
in it
,
,
.
Tfc 09
which constitutes perceptible difference strengths
=
1
a
,
.
appears beyond reasonable doubt that the void ratio failure extension
at
in
It
tests smaller than that comparable compression tests According the

to
in
is
hypothesis formulated by Eq 33 such difference void ratios should cause

in
a
,
.
extension tests yield slightly higher effective stress strengths than compres
to
of
sion tests As shown above the difference between theoretical values The
,
.
be
, as
magnitude the difference
Tfr
of
very small and the same order
is
of
and
difficult
of
the test results shown Table Therefore
to
tween some
in
in is
it
7
.
verify or disprove the correctness of the theoretical difference strengths by
possible Parry that the strength parameters
by
as
experiments shown
It
is
,
.
vary slightly with the stress conditions However considering the sources of
,
.
error and difficulties encountered triaxial testing also possible that
is
in
it
,
future tests with improved equipment and procedures may verify that triaxial
extension tests yield effective stress strengths which are slightly higher than
those obtained compression tests
in
.
10
. CONCLUSION AND SUMMARY
The principal topic of this paper the shear strength of saturated re
is
,
molded and reconsolidated clays but major part deals with related proper
a
ties which influence the shear strength The paper ,
part
of
restatement
is
in in
a
.
the results of earlier investigations by the writer and part review of re
a
cent research by others conformity with the objectives of this conference
In
,
.
emphasis placed on discussion of fundamental relations rather than on

is
prac
of
methods for determination and application shear strength clays
of
in
tical problems
.
The Shear Strength

clays
of
Mohr Coulomb failure criteria The shear strength commonly
is
-
determined by the Coulomb or Mohr failure criteria When these criteria are .
total stresses the parameters are subject
of
expressed terms such wide
to
in
variations that they must be determined by tests which closely simulate stres
ses stress history drainage conditions and time the prototype structure
in
,
.
Nevertheless failure criteria expressed terms of total stresses are often
in
,
practical problems but their limitations

of
convenient use the solution

to
in
should be realized
.
The same failure criteria expressed terms of effective stresses ac

in
cordance with the Terzaghi concept Eqs and have much greater general in
3
9
,
,
.
validity Even then the parameters vary with the stress history of the clay
or ,
.
and practical applications require estimation the pore water pressure

of
graphical representation cri

of
effective stresses failure

at
the failure
A
.
teria consists of shear strength line for normally consolidated clays plus
a
a
-
family hysteresis curves for overconsolidated clays Fig

of
6
.
.
by
Stress void ratio failure criteria research extending investigations

In
.
-
Terzaghi and Janiczek the writer found that the strength of remolded and re
,
consolidated clays can be expressed as effective friction component which

an
the effective normal stress plus effective cohesion compo

of
an
function
is
a
nent which function of the void ratio or the equivalent consolidation pres
is
a
,
sure at failure Eqs 29 and 33 This concept has been extended triaxial
to
,
.
by
Skempton Bishop and Bjerrum Eqs

38
tests and stress conditions

,
.
.
The parameters of the stress void ratio failure criterion are independent
-
of the stress history of the clay and graphical representation

of
the crite
a
,
single straight line Figs This represents

24
27
rion reduced and

to
is
a
,
, .
considerable simplification from theoretical viewpoint but application of the

a
practical problems requires estimation pore water pressures

of
criterion
to
-
or
many cases
or
void ratios failure which

at
and
in
not convenient even

is
,
/
possible considering current knowledge the volume change characteristics

of
of clays However the stress void ratio criterion offers consistent expla
,
a
of .
the commonly used and simpler failure criteria

of
nation the deficiencies

,
and it facilitates estimation of corrections to and the limits of application of

these criteria .
The effective friction and cohesion components may be considered as phy
sical or phenomenological components , which possibly are related to primary ,
secondary , and remanent changes in spacing of the clay particles , but probably
not to separate types of intrinsic forces . The analytical expression for the
effective cohesion component , Eq . 37 - A , is probably a simplification of the
actual relations , and its validity may be limited to the normal ranges of void
ratios or water contents of a given clay .
Strength and void ratio . The shear strength of normally consolidated clays ,
excepting some very fat and active clays , can be expressed as an explicit func
tion of the void ratio at the start of a test or the void ratio at failure , Eqs . 26 -
27 and Fig . 19 , but the parameters of these functions, or the strengths at con
stant void ratio , vary with the relative value of the intermediate principal
stress at failure , Fig . 38 and Table 5 . Similar expressions for the strength of
overconsolidated clays are much more complicated , or have a limited range
of validity , because this strength also is a function of the stress history of the
clay .
The above -mentioned simple relations between strength and void ratio are
valuable from a theoretical standpoint and when dealing with uniform clays ,
but they may be difficult to apply to many undisturbed clay deposits in which
there are relatively great variations in water content and composition with
short distances even though the variations in strength are small .
Inclination of planes of failure . Possible variations in the shear strength
parameters or the effective cohesion component with direction have been in
vestigated by measurement of the angles of inclination of failure planes in un
confined and triaxial compression tests , Fig . 28 , but definite conclusions can
not be formulated . Anisotropy and minor irregularities may have an appreci
able influence on the inclination of these planes , Fig . 29 , and there is consider
able scatter in measured values . However , the average values of the angles
of inclination for remolded clays are in much better agreement with the
stress - void ratio criterion than with the Mohr - Coulomb failure criterion .
Published data on angles of inclination of failure planes in undisturbed clays
are limited in extent and vary much more than those for remolded clays .
Volume - change energy . It has been proposed that the effective friction and
cohesion components be modified by introduction of a surface energy correc
tion or component , Fig . 9, which is a function of the rate of volume change at
failure , Eqs . 11 - 12 . However , the origin or dissipation of the volume - change
energy and its relation to electrochemical and external forces need clarifica
tion . The rate of volume change at failure may indicate the existence of cor
responding , remanent excess pore -water pressures in slow drained tests .
Time - dependent decrease in strength . An increase in shear stresses
causes a decrease of the pore - water pressures and an initial increase in
strength of strongly overconsolidated clays . A subsequent equalization of the
pore -water pressures causes an increase in void ratio , a decrease in effective
stress , and a decrease in strength . Rapid or undrained tests on strongly over
consolidated clays yield greater strengths than slow drained tests on the same
clays , Fig . 31 . These phenomena are explained by the volume - change charac
teristics of clays and the stress - void ratio failure criterion .
A secondary or rheological decrease in strength occurs irrespective of the
state of consolidation which , according to investigations by Casagrande , Wil
son , and others , is proportional to the logarithm of time, Fig . 33 and Eq . 48 .
This decrease in strength is in part caused by a secondary increase in pore

water pressures and a corresponding decrease of the effective stresses and
the effective friction component . Another part is caused by the existence of
a " rheological component ” which decreases to zero with
increasing time .
This component is in part a viscous phenomenon , but may be the result of
various factors which have not yet been clarified .
It is assumed in this paper that the rheological component is a part of the
effective cohesion component , and that the remaining part of this component
constitutes the ultimate cohesion component , Fig . 9. Available experimental
data are inadequate for verification of this assumption and for determination
of the individual parts of the total decrease in strength . It is possible that the
ultimate cohesion component may be negligible for some clays , and may con
stitute an appreciable part of the effective cohesion component for other clays .
Principal stress space . Comprehensive failure conditions and correspond
ing failure surfaces in the principal stress space are reviewed in the preced
ing section of the paper . The relative value of the intermediate principal
stress affects the pore - water pressures and the constant volume strength to
a considerable degree , Fig . 38 and Table 5, but appears to have only little in
fluence on the failure conditions expressed in terms of effective stresses . The
Mohr - Coulomb failure surface probably represents the lower limits of the
shear strength , and it is possible that the actual failure surface for clays is a
curved surface which circumscribes the Mohr - Coulomb hexagon , Fig . 40 .
However , definite conclusions cannot yet be formulated because of scatter in
some of the limited published data available .
Influence of deformations . Geuze has suggested that shear stresses ex
ceeding a certain " flow limit " cause continuing deformations and ultimate fail
ure of a clay . Goldstein has suggested that failure occurs when the shear
strains exceed a certain limiting value . The design of many foundation struc
tures is governed by the deformations rather than by the limiting strength of
the soil , which often is taken into consideration by using allowable strengths
which correspond to small but arbitrarily selected strains in stress - strain

diagrams obtained by standard laboratory tests . These suggestions and meth
ods need to be amplified by further detailed and systematic research utilizing
the methods and theories of plasticity and rheology .
Shear strength after failure . A permanent decrease in strength after fail
ure may be caused by an increase in void ratio or a change in structure of the
clay . However , a major part of the decrease in strength after failure of re
molded clays is caused by a transient increase in pore -water pressure and / or
thixotropic changes in strength , in which case the strength is regained in time
after the deformations cease . The magnitude of this reversible decrease in
strength depends to a large extent on the rate of deformation at and after fail
ure .
Properties Influencing the Shear Strength
Structure of clay . Recent investigations by Lambe , Rosenquist , Tan , and
others constitute a significant advance in clarification of the nature of the in
trinsic forces acting in the clay - water system and of the many factors which
govern the magnitude of these forces . The intensity of the forces decreases
rapidly with distance from a clay particle , and the minimum distance between
the particles is of greater importance than the average distance . The result
ant intrinsic force between particles of a given clay is a function of both the
void ratio of the clay and the structural arrangement of the particles . Initial
variations in structure may have considerable influence on the deformation
characteristics and the pore -water pressures , but investigations by Seed ,

Mitchell , and Chan indicate that such variation may have relatively little in
fluence on failure conditions expressed in terms of effective stresses , proba
bly because the deformations tend to produce similar structures at failure .
Changes in void ratio . Volume - change characteristics of clays during
drained tests are fairly well known from a qualitative standpoint . An increase
of the mean , effective normal stress produces a decrease in void ratio and
vice versa . An increase in shear stresses causes a decrease in void ratio of
normally consolidated and lightly overconsolidated clays but an increase in
void ratio of strongly overconsolidated clays . A shear stress reversal or de
crease in shear stresses may produce void ratio changes of the same sign but
much smaller magnitude as an increase in shear stresses . The volume de
crease caused by a change in shear stresses contradicts the results obtained
by the mathematical theory of elasticity , and this fact should be taken into con
sideration when the theory and related failure criteria are applied to clays .
Roscoe , Schofield , and Wroth have demonstrated that void ratios and effec
tive stress conditions of normally consolidated and overconsolidated clays ap
proach the same " critical void ratio " line at very large strains . This ultimate
state may be reached after failure or may coincide with the peak shear
strength of normally consolidated clays , depending upon the testing procedure .
However , in drained tests on strongly overconsolidated clays the critical void
ratio is attained after failure for both controlled strain and controlled stress
types of loading .
Quantitative data on volume changes of clays have been obtained by confined
consolidation tests and by triaxial compression and extension tests but not yet
for stress conditions corresponding to plane strain or intermediate values of
the second principal stress . The common statement that the void ratio pri
marily is a function of the major principal stress is approximately correct for
some stress conditions but is an oversimplification for other stress condi
tions , Fig . 38 . The change in void ratio during a drained shear test on a nor
mally consolidated clay is a constant for a given clay and testing procedure ,
Fig . 21 , but this change in void ratio varies with the stress conditions at the
start of the test and with the relative value of the intermediate principal stress
at failure , Fig . 38 .
Analytical expressions have been developed for changes in void ratio during
confined consolidation and hydrostatic consolidation tests . Changes in void
ratio for other stress conditions corresponding to those in triaxial compres
sion and extension tests can be determined graphically by means of the Ren
dilic constant void ratio contours and other methods suggested by Henkel . The
change in void ratio of clays is a function of changes in both the effective nor
mal stresses and the shear stresses , and efforts should be made to develop a
comprehensive theory of volume changes which will permit estimation of the
change in void ratio caused by an arbitrary change of any principal stress .
Changes in pore - water pressure . Changes in void ratio under drained con
ditions and in pore -water pressure under undrained conditions are inter
dependent . An increase of the mean normal stress tends to produce an in
crease in pore -water pressure and vice versa . A change in shear stresses
tends to produce an increase in pore -water pressures of normally consolidated
and slightly overconsolidated clays and a decrease in the pore - water pressure
of strongly overconsolidated clays . Upon completion of the initial or primary
change in pore - water pressure , a sustained change in stresses produces an ad
ditional or secondary change in pore - water pressure which increases with
time, as demonstrated by Bjerrum and associates . This phenomenon is rela

ted to the rheological properties of clays, but has not yet been investigated in
adequate detail.
Previously mentioned methods and diagrams for graphical determination of
changes in void ratio can also be used for estimation of changes in pore - water
pressure under undrained conditions , Fig . 38 . In recently improved analytical
methods by Skempton and Henkel , the change in pore -water pressure at fail
ure is expressed as a linear function of the changes in the mean normal stress
and in the octahedral shear stress , Eq . 57 , but the coefficients in this equation
vary with the stress conditions at the start of the test , with the relative value
of the intermediate principal stress at failure , and with the stress history of
the clay . Therefore , those coefficients must be determined by tests which
simulate the stress conditions and state of consolidation in the prototype
structure .
Development of more comprehensive methods and theories for determina
tion of changes in void ratio and pore -water pressures is needed to enable a
more general application of the principle of effective stresses to the solution
of problems dealing with deformations, bearing capacity , and slope stability
of clays . Such developments will also improve the accuracy of estimates of
settlements of foundations on massive clay deposits .
The rheological properties . Plastic deformations or creep of clays over
extended periods of time often occur in natural deposits and have been ob
served and recorded . The results of very limited laboratory investigations
indicate that the plastic deformations , depending on the stress conditions and
type of clay , may ( 1) ultimately cease , ( 2) continue for an indefinite period of
time at a decreasing rate , and ( 3) attain a fairly constant velocity which ulti
mately may increase and cause failure of the clay , Fig . 30 and Eqs . 42 - 44 .
However , available experimental data are inadequate for estimating the type
of deformation which may take place under field conditions .
Theories of plasticity and rheology have been developed and verified ex
perimentally for other materials . These theories may also be applied to clays ,
provided the special volume - change characteristics and thixotropic properties
of clays are taken into consideration . Adaptations of these theories and the
corresponding rheological models to clays have been proposed by Geuze , Tan ,
and Schiffman , but additional systematic experiments of long duration are re
quired in order to verify the hypotheses and to separate the influence of many
factors which affect the plastic deformations of clays .
Fundamental research into the rheological properties of clays should also
yield data on the nature and extent of the time - dependent decrease in strength
on the existence and values of yield limits , and for improvement of currently
used practical expedients for estimating stress conditions which will not cause
excessive ultimate deformations .
Sources of Systematic Variations and Errors
The following testing procedures , definitions , and other factors may cause
systematic variations or errors in the test results and are in need of additional
investigations .
1. Nonuniform distribution changes in stresses , strains , void ratios , and
pore -water pressures in a triaxial test specimen .
2 . Review and critical comparison of various definitions of the state of fail
ure and corresponding strengths and strains .
Influence of remolding water content in relation to reconsolidation pres
3. In
sure on stresses , void ratios , and pore -water pressures at failure .
Influence of the direction of anisotropy or a preferred orientation of

4.
clay particles on strengths , deformations , and pore -water pressures .
5 . Remanent pore -water pressures in slow drained tests and their rela
tion to the rate of volume change at failure .
6 . Influence of the type of load application in very slow tests - - controlled
stress increments versus controlled rate of strain - - on strengths ,
strains , void ratios , and pore -water pressures at and after failure .
When comparing the results of various test series and observations in both
laboratory and field , the possible existence and influence of compensating fac
tors or errors should be carefully investigated and taken into consideration .
Conclusion
The papers and reports referred to in this review constitute only a part of
published data on the shear strength of saturated clays . Much additional and
pertinent data can undoubtedly be obtained from the detailed test records which
have served as a basis for published summaries of the test results . Neverthe
less , full understanding of the strength characteristics of clays , and the appli
cation of this knowledge to practical problems require extensive additional re
search , not only into the shear strength proper , but even more into related
properties of clays .
The two general problems , concerning properties which influence the shear
strength , in greatest need of additional experimental and / or mathematical in
vestigations are ( 1) relations between changes in stresses , void ratios , and
pore - water pressures , and ( 2) the rheological properties of clays , including
the time - dependent decrease in strength .
The above -mentioned research will require very accurate experiments and
relatively large quantities of uniform clay preparation
of
for
test specimens
.
will undoubtedly advantageous and even necessary perform
be
Therefore
to
it
,
some of the experiments on remolded and reconsolidated clays but these ex

,
periments should be supplemented by tests undisturbed clays and the final

on
results should be verified and amended as necessary by comparisons of the

clays under field conditions
of
estimated and actual behavior

.
possible and even probable that the more comprehensive failure cri
is
It
teria and theories for deformation and changes void ratio and pore water
in
pressures clays may be relatively complicated and difficult prac

of
apply
to
to
tical problems single easily applicable failure criterion for all possible
A
,
.
field conditions probably does not exist but should be possible develop
or to
it
,
simple failure criteria for several limited ranges use cur

of
conditions
to
,
rent practical methods and criteria with appropriate adjustments and delimita
The more comprehensive
of
tions failure criteria and theories should be

.
great assistance explaining and correcting the deficiencies currently used

in
in
practical methods other practical methods

of
the development estimat

in
in in
,
ing adjustments required by differences between stress conditions test

a
stipulating the proper limits

of
specimen and the prototype structure and

in
,
by
application of data obtained various types of tests and field observations

.
APPENDIX NOTATIONS
-
I.
establishing the following list ob

of
notations efforts have been made

to
In
by
tain agreement with the letter symbols suggested the ASCE ASTM Com
-
Glossary
of
on
mittee Terms and Definitions Soil Mechanics and also with

in
the list of symbols recommended for use papers for the Fifth International
in
Conference on Soil Mechanics and Foundation Engineering . The prime marks

used for o' , c ' , and Ø ' indicate effective stresses or values of c and obtained
for relations expressed in terms of effective stresses . The prime mark has
been omitted in cases where it is believed that the omission will cause no mis
understandings .
area
Aь
partially disturbed area in box shear tests
effective area in box shear tests

pore pressure coefficient , Skempton
pore pressure coefficient , Henkel
Terzaghi compression index equation natural logarithms
in
(
)
Terzaghi compression index for all round triaxial pressure
-
pore pressure coefficient Skempton
compression index ,
equation common logarithms
in
(
)
cohesion total stresses
,
cohesion effective stresses

,
or
effective cohesion cohesion component

component time
at
effective cohesion
t
effective cohesion for water content

w
Cew
intercept
of
Mohr envelope effective stresses

,
cohesion intercept total stresses

,
cohesion intercept effective stresses

5
,
ultimate cohesion component
5 coefficient
of
consolidation
3 rheological strength component
at
coefficient theoretical effective cohesion zero void ratio

=
of
coefficient effective void ratio component triaxial

,
Nico
void ratio
o
or
of
of
void ratio end consolidation start shear test

at
void ratio failure

at
con
co
kg
void ratio corresponding cm2

to
0
1
(
)
/
utlimate void ratio after failure

å
om
change void ratio during shear test
in
up exponential function
flow limit
of of
T
yield limit
T
263
PHYSICAL COMPONENTS
specific gravity of soil solids

specific gravity of water
thickness of test specimen

effective thickness of test specimen
J2 and Jz = stress invariants

coefficient of earth pressure
coefficient of earth pressure at rest
side length of test specimen in box shear tests
liquidity index
natural logarithm , loge
common logarithm , log10
porosity
coefficient of relative value of o' z

overconsolidation ratio
prestress ratio
mean total normal stress = ct = ( 01 + 02 + 03 ) /3

mean effective normal stress = poct = (oj + 0 + 04 ) 3/
ultimate mean effective stress at CVR line
compressive strength = (o - 01 ), subscripts as noted in text
compressive strength at time t
ultimate compressive strength at CVR line
log
compressive strength for cycle

of
decrease in time
failure surface radius for compression tests
failure surface radius for extension tests

actual duration shear test
of
equivalent duration
of
test for uniform loading

of
rate
as
time subscripts noted text

in
,
time corresponding zero strength

to
pore water pressure

-
water content
of
of
or
at
water content end consolidation start shear test

failure
at
water content
kg
water content corresponding cm2

to
0
1
(
)
/
change water content during shear test

in
lateral displacement during shear test

or
deformation
vertical settlement or swelling during shear test

angle between plane of failure and major principal stress
for
direction cosine or weighting coefficient
o
' s
angle corresponding
of
inclination
to
of
angle inclination corresponding
to
plane failure
of
of
angle inclination
of
optimum
angle inclination corresponding
to
of
'or øs
clay particles
of
angle stratifications
of
orientation
oj
oj
oz
of
angle inclination
, , versus
-
ž
(
)
versus
;
angle
(01
versus
of
inclination
ž
2
0
0
}
(
)
-
+
shear strain
shear strain one day after load application
shear strain time

at
for log cycle

of
increase shear strain time
in
compressive tensile strain

or
coefficient structural viscosity

of
twist radians during torsion shear tests

in
of of of
coefficient effective cohesion

coefficient void ratio component triaxial tests
in
for
coefficient effective cohesion residual strength

coefficient friction
of
effective tan
=
tan
of of
coefficient shear strength for overconsolidation

=
Ø
zloga
for
coefficient shear strength normal consolidation tans

=
coefficient effective stress component triaxial tests

of
in
of
coefficient total strength CVR line triaxial

at
tests
in
Liga
Poisson ratio
e rheological decrease subscripts
of
coefficient strength text

é
in
in
,
value for
of
minute
o
=
p
1
t
value for minutes

o
of
1000
=
p
as
subscripts
o
total normal stress for
of
;
effective normal stress

o
o effective axial stress triaxial tests
in
effective consolidation pressures

com
equivalent consolidation pressure

at
plane failure failure
on
effective normal stress
of
effective consolidation pressure corresponding kg cm2
to
,
(1
e
)
/
olgan
effective prestress preconsolidation pressure
or
effective radial stress triaxial tests
in
effective octahedral normal stress
=
p
'
iga
on
y and effective principal stresses

',
=
o
shear stress
ay average shear strength plane failure
of
in
ay shear strength plane failure
of
minimum
in
at
of
maximum shear strength plane failure
in
at
of
or
dilatation surface energy component shear strength

met shear strength
het shear strength compression tests

in
met shear strength extension tests

in
or shear strength corresponding

to
o
of octahedral shear stress

moite residual shear strength
a
of
friction component shear strength

of of
angle shear strength total stresses

,
angle shear strength effective stresses

,
angle
of of
effective internal friction

D20 effective angle internal friction for residual strength
of of of
angle inclination Mohr envelope

of
angle shear strength for overconsolidation

angle shear strength for normal consolidation
equivalent angle
90
internal friction
of
2
a
-
(
APPENDIX REFERENCES
II
.
-
University
pp
Theory
of
BEKKER 1956 land locomotion 520 The

M
,
G
(
.”
.
.
).
.
"
Michigan
of
Press Ann Arbor

,
BISHOP measurement shear strength

of - of
W
Discussion
of
of
1950
,
A
.
.
(
).
"
.pp
soils Geotechnique Vol 113 116

,
2
,
.”
of
BISHOP 1954 Discussion shear characteristics saturated

W
,
A
pp a
(
“
.
.
).
silt measured triaxial compression Geotechnique Vol

45
43
in
,
4
,
,
"
.
.
-
the slip circle stability analysis

of
of
BISHOP 1955 The use

W
the
in
,
A
«
(
.
.
).
Geotechnique Vol
pp
slopes
17
,
,
5
7
"
.
.
-
BISHOP , A . W . , ALPAN , I, BLIGHT , G . E ., and DONALD , I. B. (1960 ). “ Fac

tors controlling the strength of partly saturated cohesive soils . " Paper pre
sented at ASCE Research Conference on Shear Strength of Cohesive Soils ,
University of Colorado .
BISHOP , A . W . and ELDIN , A . K . G . ( 1953 ) . " The effect of stress history on

the relation between 0 and porosity in sand . ” Proc . Third Int . Conf . Soil Mech .
Found . Eng ., Zurich , Vol. 1, pp . 100 - 105 .
BISHOP , A . W . and HENKEL , D . J . ( 1957 ) . “ The measurement of soil proper

ties in the triaxial test . ” 190 pages . Edward Arnold Ltd . , London .
BJERRUM L . (1953 ) . Discussion on shear strength of clays . Proc . Third Int .

Conf. Soil Mech . and Found . Eng . , Zurich , Vol . 3, pp . 112 - 113 .
BJERRUM L . (1954 ) . “ Theoretical and experimental investigations on the
shear strength of soils . ” Norwegian Geotechnical Institute , Oslo , Bulletin No .
5 , 112 p .
BJERRUM , L ., SIMONS, N . , and TORBLAA ( 1958 ). " The effect of time on the
shear strength of a soft marine clay ." Proc . Brussels Conference 58 on Earth
Pressure Problems , Vol . 1 , pp . 148 - 158 .
BOLT , H. G . (1956 ). " Physico - chemical analysis of the compressibility of
pure clays . " Geotechnique , Vol. 5, No . 2, pp . 86 - 93 .
BOROWICKA , H . ( 1959 ). « Ueber eine neue Theorie der Scherfestigkeit
bindiger Böden . ” Mitteilungen des Institutes für Grundbau und Bodenmechanik
and der Technischen Hochschule Wien . Heft 2 , pp . 9 - 35 .
BRENNECKE , L . and LOHMEYER , E . (1938 ). “ Der Grundbau . " (Fifth Edition

by E . Lohmeyer , Vol. 1, Part 1) 248 pages . Wilhelm Ernst & Sohn , Berlin .
BUESSEM , W . R . , and NAGY,
B . ( 1953 ). " The mechanism of deformation of
clay . " Proc Nat . Conf. on Clays and Clay Minerals , University of
. Second
Missouri , pp . 480 -491 . ( Publication No. 327 , National Academy of Sciences ,
Washington , 1954 .)
BUISSON , M . ( 1936 ) . " Shearing tests , compressive tests on cylinders . Com

parisons of results in view of determining the angle of friction and the settle
ments in soils . " Proc . First Int . Conf . Soil Mech . Found . Eng . , Cambridge ,
Vol 2, pp . 121 - 125 .
BURGERS , J . M . and SCOTT BLAIR , G . W . ( 1948 ) . " Report on the principles
of rheological nomenclature . " Proc . First Int . Congress on Rheology ,
Scheveningen .
CAQUOT , A . ( 1934 ) . “ Equilibre des massifs a frottement interne . Stabilite
des terres pulverulents et cohérentes . ” 100 pp . Gauthier - Villars , Paris .
CASAGRANDE , A. (1932 ) . structure of clay and importance founda
its
" The
in
pp
engineering Civ Eng

19
tion Journal Boston Soc Vol 168 209

,
.,
,
. ”
-
.
ALBERT
re
CASAGRANDE and
on
1932 Research the shearing

,
,
G
A
“
.
.
(
).
Report Massachusetts Institute

of
of
sistance soils Technology Cambridge

,
,
."
anisotropic
of
CASAGRANDE and CARILLO 1944 Shear failure

,
.,
,
N
A
. .(
., )
“
.
pp
materials Journal Boston Soc Eng

31
Vol
74
Civ 1944 87
,
,
."
.
-
CASAGRANDE , A ., CORSO , J . M . , and WILSON , S. D. ( 1950 ) . " Investigation

of effect of long - time loading on the strength of clays and shales at constant
water content." Report to Waterways Experiment Station , 34 pp . Harvard
University .
CASAGRANDE , A . and RIVARD , P . J . ( 1959) . " Strength of highly plastic

clays . ” Norwegian Geotechnical Institute , Publication No . 31, pp 1 - 23 . Re
printed as Harvard Soil Mechanics Series No . 60 .
CASAGRANDE , A . and SHANNON , W . L . ( 1949 ) . " Strength of soils under dy
namic loads . " Trans . Am . Soc . Civ . Eng ., Vol . 114 , pp . 755 - 772 .
CASAGRANDE , A . and WILSON , S . D . ( 1949 ) . " Investigation of the effect of

long - time loading on the strength of clays and shales at constant water con
tent. " Report to U. S. Waterways Experiment Station , 77 pp . Harvard
University .
CASAGRANDE , A . and WILSON , S . D . ( 1951 ). " Effect of rate of loading on the

strength of clays and shales at constant water content . " Geotechnique , Vol.
2 , pp . 251 - 263.
CASAGRANDE , A . and WILSON , S . D . (1953 - A ). “ Prestress induced in

consolidated - quick triaxial tests. " Proc . Third Int . Conf . Soil Mech . Found .
Eng . , Zurich , Vol. 1, pp . 106 - 110 .
CASAGRANDE , A . and WILSON , S . D. (1953 - B ) . “ Effects of stress history on

the strength of clays . " Harvard Soil Mechanics Series No . 43 , Cambridge ,
Mass . 37 pp ., 79 figs .
COLEMAN , J . D . (1960 ) . " Suction and the yield and failure surface for soil in
theprincipal effective stress space . ” Geotechnique , Vol. 10 , pp . 181 - 183 .
CRAWFORD , C. B . (1959). " The influence of rate of strain on effective stres

ses in sensitive clay . " Papers on Soils 1959 Meetings , Am . Soc . Test . Mat . ,
Special Publication No . 254 , pp . 36 -48 .
ENDELL , K . and HOFFMAN , U . (1936 ) . " The chemical nature of clays . "
Proc. First Int. Conf . Soil Mech . and Found . Eng . , Cambridge , Vol. 1, pp 51 -
54 .
FREUNDLICH , H . ( 1935 ) . " Thixotropy .” Hermann & Co ., Paris .

GEUZE , E . C . W . A. ( 1948 ) . " Horizontal earth pressure against a row of
piles . ” Proc . Second Int. Conf. Soil Mech . Found . Eng . , Rotterdam , Vol. 4 , pp .
135 - 140 .
GEUZE , E . C . W . A . ( 1960 ) . “ The effect of time on the shear strength of

Civ
clays . ” Paper presented at the Am . Soc . Eng Convention New Orleans

in
,
.
March 1960
.
GEUZE shearing properties

of
and TAN
W
1950 The
,
, ,
C
K
E
.
.
.
.
. T.
.
).
"
- (
pp
soils Geotechnique Vol 141 161

,
2
.
.
."
GEUZE
of
and TAN The mechanical behaviour

W
1953
,
,
E
K
T
.
. .
.
.
"
.
.
, .
(
.pp
on
clays Proc Second Int Conf Rheology Oxford 247 259

,
-
.
.
."
Experimental determination
of
GIBSON 1953 the true cohesion and

,
R
“
.
.
).
(
true angle internal friction clays Proc Third Int Conf Soil Mech
of
. in
.
."
pp
Found Eng Zurich Vol 126 130

1
,
,
.,
-
.
.
GIBSON , R . E . and HENKEL , D . J. (1954 ) . " Influence of duration of tests at

constant rate of strain on measured drained strength . " Geotechnique , Vol. 4,
pp . 6 - 15 .
GOLDSCHMIDT , V. M . (1926 ) . “ Undersökelser over lersedimenter ” ( Investi

gations of sedimentary clays ) . Nordisk Jordbrugsforskning , Proc . Third
Conf ., Oslo , pp . 434 -445 .
GOLDSTEIN , M . ( 1957 ) . " The long - term strength of clays . " Proc . Fifth Int.
Conf. Soil Mech . Found . Eng . , London , Vol. 2, pp . 311 - 313 .
GOLDSTEIN , M . ( 1958 ) . “ Discussion of failure deformation of clays as a func
tion of time. ” Proc . Brussels Conference 58 on Earth Pressure Problems ,
Vol . 3, pp . 91 - 92 .
HABIB , P. ( 1953 ) . “ Etude de l'orientation du plan de rupture et de l'angle de
frottement interne de certaines argiles . "
Int
Proc . Third Conf Soil Mech
.
Found Eng
28
Zurich Vol pp 31
1
,
-
.
.
HABIB the intermediate principal
of
of
1953 Influence variation
on M
the
,
"
(
.
.
).
shearing strength
of
Int
stress the soils Proc Third Conf Soil Mech
.
."
.pp
Found Eng Zurich Vol 131 136

.,
1
,
-
.
HAE FELI 1938 Mechanische Eigenschaften von Lockergesteinen

,
R
“
.
).
(
.'
.pp
Schweizerische Bauzeitung Vol 111 299 303 321 325

,
,
.
.
-
-
HAE FELI 1939 Schneemechanik mit Hinweisen auf die Erdbaumechanik
R
,
"
"
.
.
(
Snow mechanics with reference soil mechanics Geologie der Schweiz

to
,
).
(
.pp
65
Geotechnische Serie Part 241 Bern

,
,
3
HAE FELI Shearing strength and water content complement
to
1948
,
R
A
«
(
- . )
.
" .
the shearing theory Proc Second Int Conf Soil Mech Found Eng Rotter
.,
.
.
pp
33
44
dam Vol
,
3
,
.
HAE FELI Investigation and measurements the shear strengths

of
1951
,
R
.
).
"
(
pp
of saturated cohesive soils Geotechnique Vol 186 208

2
,
,
."
.
-
HAE FELI 1953 Creep problems soils and ice Proc Third
, in
snow
R
,
,
"
.
).
(
.'
.
pp
Int Conf Soil Mech Found Eng Zurich Vol 238 251
.,
3
,
.
.
-
HAE FELI and SCHAERER 1946 Der Triaxialapparat The triaxial

,
,
C
R
”
.
, ).
(
67 (
.pp
51
65
81
apparatus Schweizerische Bauzeitung

53
64
Vol 128
,
,
-
-
).
HANK and McCARTHY 1948 Shear failure anisotropic

in
,
,
R
L
.
J.
.
of E.
).
"
(
materials possessing any values

of
cohesion and angle internal friction

."
.pp
Proc Highway Research Board Vol

28
449 456
,
-
.
the triaxial test for soils

of
HAYTHORNTHWAITE 1960 Mechanics

M
,
(
., R.
.
., ).
"
."
61
pp
Proc Am Soc Civ Eng SMF Div Vol

86
35
No SM5 Also issued

,
-
.
of .
.
(
April 1959 as Technical Report

by
Division Engineering Brown Univer

in
sity
.)
HENKEL 1958 The correlation between deformation pore water

D
,
,
"
.
J.
).
(
pressure and strength characteristics of saturated clays Thesis University

,
."
April
of
London 1958
,
HENKEL 1959 The relationships between the strength pore water

,
D
,
“
.
J.
(
).
pressure and volume saturated clays

of
change characteristics Geotech

.
'
-
nique Vol pp 119 135

9
,
-
.
.
PHYSICAL COMPONENTS - 269
HENKEL , D. J . ( 1960 - A ) . “ The relationship between the effective stresses and

water content in saturated clays . " Geotechnique , Vol. 10 , pp . 41 - 54 .
HENKEL , D. J . (1960 - B) . “ The shear strength of saturated remoulded clays . "
Paper presented at the ASCE Research Conference on Shear Strength of Co
hesive Soils . University of Colorado .
HILF , J . W . and GIBBS , H . J. (1957) . “ Triaxial shear tests holding effective
lateral stress constant . " Proc . Fourth Int. Conf. Soil Mech . Found . Eng . ,
London , Vol. 1, pp . 156 - 159.
HIRSCHFELD , R. C . (1958). “ Factors influencing the constant volume strength

of clays . " Thesis , Harvard University , Cambridge , Mass ., 398 pp .
HIRSCHFELD , R . C . (1959). " The relation between shear strength and effec
tive stress . ” Proc . First Pan American Conf. Soil Mech . Found . Eng . , Mexico
City .
HVORSLEV , M . J . (1936 ). " Conditions of failure for remolded cohesive soils . "
Proc . First Int . Conf. Soil Mech . Found . Eng ., Cambridge , Vol. 3, pp . 51 - 53 .
HVORSLEV , M . J . (1937). « Ueber
die
Festigkeitseigenschaften gestörter
bindiger Boden On the strength properties of remolded cohesive soils
”
(
)
, .
Thesis 159 pages published by Danmarks Naturvidenskabelige Samfund
,
,
Nr
Ingeniorvidenskabelige Skrifter Series Copenhagen

45
,
,
A
.
.
.
HVORSLEV The shearing resistance
of
1938 remolded cohesive

M
,
“
.
J.
).
(
soils Proc Soils and Foundations Conf Engineer Department Boston

.,
,
U
S
.
.
.
."
pages
30
Section
E
,
HVORSLEV 1939 Torsion shear tests and their place the deter
M
in
of ,
"
(
)
.
J.
the shearing resistance Proc Am Soc Testing Mate

of
mination soils
.”
.
pp
rials Vol
39
999 1020
,
,
.
HVORSLEV soil properties

on
1953 Discussion Proc Third Int

M
,
.
.J
).
(
.
.pp
Conf Soil Mech Found Eng Zurich Vol 122 124

.,
3
,
-
.
irregularities
of
HVORSLEV distribution stres

of
957 Discussion
in
M
,
.
J.
).
(
ses strains and volume changes triaxial test specimens Proc Fourth Int
in
,
.
pp
Conf Soil Mech Found Eng London Vol 105 107

.,
3
,
-
.
HVORSLEV and KAUFMAN 1952 Torsion shear apparatus and

M
,
,
38 R
“
.
J.
, .
I.
).
(
testing procedures Bulletin No USAE Waterways Experiment Station

,
.”
Vicksburg Miss
,
og
JOHANSEN Brudbetingelser for sten beton Failure condi

W
1958
,
K
”
.
.
).
(
.pp
Bygningsstatiske Meddelelser Vol

19
tions for rocks and concrete 25

,
,
).
Teknisk Forlag Copenhagen

44
,
.
JURGENSON The shearing resistance Journal

of
1934 soils Boston

,
L
,
.pp "
(
21 )
."
.
, .
Soc Civ Eng Vol 242 275

.,
-
.
, of
KIRKPATRICK The condition failure for sands Proc

W
1957
M
,
”
.
.
).
- .
.
"
(
pp
Fourth Int Conf Soil Mech Found Eng London Vol 172 178
.,
1
,
.
KJE LIMAN Report on apparatus for consumate investigation

an
W
1936
,
"
(
.
).
the mechanical properties Proc First Int Conf Soil Mech

2 of
soils
of
.
.
.
.
Cambridge Vol
pp
Found Eng
16
20
.,
,
.
.
-
KOHLER , K . ( 1959 ) . «Vorgänge in der Scherbüchse und deren bodenmechan

ische Deutung . " Die Bautechnik , Vol . 36 , pp . 25 - 27 .
KYVELLOS , M . G . (1956 ). “ Etude de la curbe intrinseque des sols compactés
et non saturés . " Annales de l' Institut Technique du Batiment et des Travaux
Publics , Vol. 9 , pp . 386 - 411 .
LAMBE , T. W. ( 1953 ) . " The structure of inorganic soil .” Proc . Am . Soc . Civ .
Eng ., Separate No . 315 .
LAMBE , T . W . ( 1958 - A ) . « The structure of compacted clay ." Proc . Am . Soc .
Civ . Eng ., SMF Div ., Vol . 84 , No. SM2, Paper No . 1654 , 34 pages .
LAMBE , T . W . (1958 - B). « The engineering behavior of compacted clay . ”
Proc . Am . Soc . Civ . Eng ., SMF Div . , Vol. 84 , No . SM2 , Paper No. 1655 , 35
pages .
MARSAL , R. J. ( 1957 ) . " Unconfined compression and vane shear . " Confer
- for
ence on Soils Engineering Purposes Mexico ASTM Special Tech Publ
.
.pp
No
232 229 241

,
.
MASON and WARD 1954 Chapter Physical and chemical as

,
,
A
2
pp P
“
M . .
.
.
.
(
elasticity and plasticity

63
pects Building materials

38
their
in
”,
,
“
-
."
North Holland Publishing
by
Co
Edited Reiner Amsterdam
,
.,
.
.
natural clays and its relation en
of
MITCHELL fabric
to
1956 The
,
K
J.
.
).
"
(
gineering properties
pp
Highway Research Board Vol
35
Proc 693 713
-
.
.
."
MITCHELL aspects thixotropy soils Proc

of
1960 Fundamental
in
,
K
.”
.
. . .
).
.
"
, (
. J
. 52
pp
Eng
86
19
Am Soc Civ Soil Mech and Found Div Vol No SM3
.,
No -
.
.
NASCIMENTO 1953 Capillarity and soil cohesion Publication
,
U
"
."
).
(
de
Engenharia Civil Lisbon

37
13
Laboratorio Nacional
p
,
,
.
pp
of of
NADAI Theory hapters

of
1950 flow and fracture solids 572

,
16 A
on (
.
).
.
"
.) ."
strength and conditions (C

of
15
and theories failure McGraw - Hilli

Co
Book New York

.,
OHDE 1955 Ueber den Gleitwiderstand der Erdstoffe Veröffent

,
."
“
(
).
J.
lichungen der Forschungsanstalt für Schiffahrt Wasser und Grundbau No

6
,
,
.
-
Berlin
.
PARRY Strength clay Univer

of
1956 and deformation Thesis

,
,
H
R
)
.
.
.
.
"
."
sity of London
.
by
PARRY
, of
of
1958 On the yielding soils Discussion paper

, ,
G
H
R
.
.
.
(
).
"
, ”
8 .
pp
Roscoe Schofield and Wroth Geotechnique Vol 183 185

-
.
PAVLICS OTAC instruments for the measurement physical

of
1958
,
F
.
"
).
(
the Army Or
of
No
Research Report
.pp
Department
14
31
soil values
,
,
5
.”
-
.
dinance Tank Automotive Command Research and Development Division Land

,
,
-
Locomotion Research Branch Centerline Michigan

,
PELTIER 1957 Recherches expérimentale sur courbe intrinseque

M
,
la
R
“
.
.
(
).
rupture des sols pulverulents Proc Fourth Int Conf Soil Mech Found
de
. .”
.
.pp
Eng London Vol 179 182

.,
1,
-
.
perfectly plastic solids

of
PRAGER and HODGE 1951 Theory

. W,
, G
P
“
.
.
.
).
."
(
264 pages John Wiley Sons Inc New York

&
.,
.
REINER , M . (1954 ). « Theoretical rheology . " Chapter 1, pp . 3 - 37 , in “ Building

materials , their elasticity and plasticity . " Edited by M . Reiner , North Hol
land Publishing Co . , Amsterdam .
RENDU LIC , L . ( 1936 - A ). “Relation between void ratio and effective principal
stresses for a remolded silty clay . ' Proc . First Int . Conf. Soil Mech . Found .
Eng . , Cambridge, Vol . 3, pp . 48 - 51 .
RENDULIC , L . (1936 - B ) . “ Porenziffer und Porenwasserdruck in Tonen ” ( Void

ratios and pore water pressures in clays ) . Der Bauingenieur , Vol. 17 , pp . 559
564 .
RENDULIC , L . (1937 ) . " Ein Grundgesetz der Tonmechanik und sein experi
menteller Beweiss . " Der Bauingenieur , Vol . 18 , pp . 459 -467 .
RENDULIC , L . (1938 ). " Eine Betrachtung zur Frage der plastischen

Grenzzustände ” ( A consideration of the problem of the limiting conditions of
plasticity ) . Der Bauingenieur , Vol. 19, pp . 159 - 164 .
ROSCOE , K . H . ( 1953 ) . " An apparatus for the application of simple shear to
soil samples . ' Proc . Third Int. Conf. Soil Mech . Found . Eng ., Zurich , Vol. 1,
pp . 186 - 191.
ROSCOE , K. H . , SCHOFIELD , A . N. , and WROTH , C . P . (1958 ). " On yielding

of soils . " Geotechnique , Vol. 8 , pp . 22 -53, Discussions , Vol. 8, pp . 134 - 136 ,
183 - 185 , 186 - 187 , Vol. 9 , pp . 71 -83 .
ROSENQUIST , I. Th . (1955 ). “ Investigations in the clay - electrolyte -water

system . ' Norwegian Geotechnical Institute , Oslo , Publication No . 9, 125 pages .
ROSENQUIST , I. Th . ( 1959 ) . " Physico - chemical properties of soils : Soil

watersystems . " Proc . Am . Soc . Civ . Eng . , Journal SMF Division , Vol. 85 ,
No . SM2, pp . 31 - 53 .
RUTLEDGE , P . C . (1947 ). “ Review of cooperative triaxial research program
of the Corps of Engineers . " Progress Report on Soil Mechanics Fact Finding
Survey , pp . 1 - 178 . Waterways Experiment Station , Vicksburg , Mississippi .
SALAS , J . A . J. and SERRATOSA , J . M . (1953 ). " Compressibility of clays . "

Proc . Third Int . Conf. Soil Mech . Found . Eng . , Zurich , Vol. 1 , pp . 192 - 198 .
SCHIFFMAN , R. L . (1959). " The use of visco - elastic stress - strain laws in
soil testing . " Papers on Soils 1959 Meetings , Am . Soc . Test . Mat. , Special
Publication No . 254 , pp . 131 - 155 .
SEED , H . B. and CHAN , C . K . ( 1959 - A ) . “ Thixotropic characteristics of com

pacted clays . " Trans . Am . Soc . Civ . Eng . , Vol . 124 , pp . 894 -925 .
SEED , H . B. , and CHAN , C . K . ( 1959 - B) . " Structure and strength characteris

tics of compacted clays . " Proc . Am . Soc . Civ . Eng ., Journal SMF Division ,
Vol. 85 , No. SM5 , pp . 87 - 128 .
SEIFERT , R . ( 1933 ) . " Untersuchungsmethoden
um festzustellen ob sich ein
gegebenes Baumaterial für den Bau eines Erddammes eignet . " Proc . First
Congress on Large Dams, Stockholm , Vol. 3, pp . 5 -67 .
SHOCKLEY , W . G . and AHLVIN , R . G . (1960). “ Nonuniform conditions in tri

axial test specimens . " Proc . ASCE Research Conference on Shear Strength of
Cohesive Soils , University of Colorado .
SKEMPTON , A . W . (1954 ) . “ The pore -pressure coefficients A and B. ' Geo

technique , Vol. 4, pp . 143 - 147 .
SKEMPTON , A . W . and BISHOP , A. W . (1954 ). Chapter 10, “ Soils , ” pp . 417
482 in “ Building materials , their elasticity and plasticity . ” Edited by M .
Reiner , North Holland Publishing Co . , Amsterdam .
SOEDERBLOM , R . ( 1959 ) . " Aspects on some problems of geotechnical chem
istry . ” Geologiska Föreningens Förhandlingar , Vol . 81 , No . 4 , pp . 727 - 732 .
TAN , T - K . (1957 ) . Discussion on structure and rheological properties of
soils . Proc . Fourth Int. Conf . Soil Mech . Found . Eng . , London , Vol. 3, pp .
87 - 89 .
TAN , T - K . (1959) . " Structure mechanics of clays . ” Scientia Sinica , Vol . 8 ,

pp . 83 - 97 .
TAYLOR , D . W . (1948 ). " Fundamentals of soil mechanics . " 700 pp . John

Wiley & Sons , New York .
TAYLOR , D . W . (1950 ) . Discussion of triaxial testing of soils . American
Society for Testing Materials , Special Technical Publication No . 106 , pp . 224
225 .
TAYLOR , D . W . (1955 ). “ Review of research on shearing resistance of clay . ”

Report submitted to the USAE Waterways Experiment Station by Massachusetts
Institute of Technology .
TAYLOR , D . W . and CLOUGH , R . H. (1951) . " Research on shearing resistance

of clay . ' Report submitted to the USAE Waterways Experiment Station by
Massachusetts Institute of Technology .
TERZAGHI, K . ( 1925 ) . “ Erdbaumechanik auf bodenphysikalischer Grundlage . "
Deuticke , Wien , 399 pp .
TERZAGHI , K . (1929 ) . " The mechanics of shear failures on clay slopes and
the creep of retaining walls . " Public Roads , Vol . 10 , pp . 177 - 192 .
TERZAGHI , K . ( 1931 - A ) . “ The static rigidity of plastic clays . " Journal of
, Vol. 2 , pp . 253 - 262.
Rheology
TERZAGHI , K . ( 1931 - B) . “ Festigkeitseigenschaften der sedimente , Schüttungen

und Gele . " Chapter in Auerbach und Hort , " Handbuch der physikalischen und
technischen Mechanik , " Vol . 4, Part 2, pp . 513 - 578 , Barth , Leipzig , 1931 .
TERZAGHI , K . ( 1938 ) . “ Die Coulombsche Gleinchung für den Scherwiderstand

bindiger Böden .” Bautechnik , Vol. 16 , pp . 343 - 346 .
TERZAGHI , K . (1941) . “ Undisturbed clay samples and undisturbed clays . "

Journal Boston Society of Civil Engineers , Vol . 28 , pp . 211 - 231 .
TIEDEMANN ,
die
B . (1937 ). “ Ueber Schubfestigkeit bindiger Böden On the

”
(
pp
shear strength soils Die Bautechnik Vol

of
15
cohesive 400 403 433 435

,
,
).
.
of
TIMOSHENKO 1956 Strength materials Part Third Edition

,
S
,
.
(
.
Co "
)
II
.”
.
pp
572 van Nostrand York

D
New
.,
.
.
TROLLOPE and CHAN Soil structure and the

,
1 ' D
1960
H
step
,
strain
C
.
.
.
. K.
“
(
).
. -
phenomenon Proc Soc Civ Eng Soil Mech and Found Div Vol
86
Am
.,
.pp.
.,
.
,
.
.
39
No SM2
,
.
.
-
TSCHE BOTARIOFF , BAY LISS , J. R . ( 1948 ) . “ The determination of

G . P . and
the shearing of varved clays and of their sensitivity to remolding . "
strength
Proc . Second Int. Conf . Soil Mech . Found . Eng ., Rotterdam , Vol . 1, pp . 203 -
207.
VINCENT , E . T . ( 1959 ) . “ Theory of land locomotion . " 62 pages . Land Loco

motion Mechanics , Summer Session , College of Engineering , University of
Michigan .
WATERWAYS EXPERIMENT STATION (1951 ) . " Torsion shear tests on Atlan
Muck The Panama Canal Technical Memorandum No 328 USAE
tic
,
3
,
.
-
.
, '
Waterways Experiment Station Vicksburg Mississippi
.
WHITMAN The behaviour soils under transient loadings
of
1957
R
,
V
.
.
"
(
.)
."
pp
Proc Fourth Int Conf Soil Mech Found Eng London Vol 207 209
.,
1
,
-
.
.
WHITMAN Some considerations and data regarding the shear
1960
R
"
.
.
)
.
strength of clays
Proc ASCE Research Conference Shear Strength of
on
."
Cohesive Soils University

of
Colorado
,
WILSON and DIETRICH 1960 Effect pressure

of
consolidation
,
,
R
S
.
.
.
, of J.
" .
"
(
. )
elastic and strength properties clay

on on
Proc ASCE Research Conference

of .
Shear Strength Cohesive Soils University

of
Colorado
of
WROTH 1958 Soil behaviour during shear Existence critical
,
C
“
.' .
.
(
.
)
Engineering London Sept

pp
26
void ratios 1958 Vol 186 409 413

,
,
(
-
.
.
STRENGTH AND DEFORMATION CHARACTERISTICS OF VARIOUS
GLACIAL TILLS IN NEW ENGLAND
By Kenneth A. Linell , 1 F . ASCE , and H. F . Shea 2
SYNOPSIS
General characteristics of New England glacial tills are described herein ,

and data on gradation , compaction , shear strength , stress - strain , Atterberg
limits , and clay mineral properties are presented for five typical tills . The
field performance of these five tills is compared to the extent to which such
data are available . Performance of Otter Brook Dam , the upstream face of
which bulged about 3 feet during construction , is described in detail . It is
concluded that research is needed to develop methods of embankment design
in which allowable deformation is the controlling factor rather than available
strength . Glacial tills are shown to vary widely in performance character
istics , and it is concluded that caution must be exercised when tills which
show exceptional stress - strain characteristics are proposed for use in outer
sections of embankment or in homogeneous embankments . Based on experi
ence at Otter Brook Dam , thin internal drainage sections are effective in re
ducing deformations of cohesive , homogeneous type embankments during
construction .
INTRODUCTION
Continental glaciation is believed to have covered New England several

times. While clear - cut evidence of successive advances and recessions of
ice is present in many areas over most of New England , there appears to be
no proof that these represent other than stages within the classical Wisconsin
glaciation . The general directions of movement of this glaciation are shown
in Fig . 1. The seaward limit of glaciation runs from Long Island eastward to
Nantucket Island and then is believed to extend northeastward , off the present
coast ( 1) . 3 The predominant glaciation is the classical Wisconsin .
Deposits of glacial till are perhaps the most widespread products of this
glaciation in New England . Even where alluvial materials cover the surface ,
till may commonly be found under these deposits overlying the bedrock . Till
has been defined by Terzaghi and Peck as " an unstratified glacial deposit of
clay , silt, sand , gravel , and boulders " ( 2) . Lawrence Goldthwaite has de
scribed till as " ground up rock debris which was carried by the glacier and
deposited into a compact , unstratified mass of angular fragments of all sizes :
clay , silt, sand , stones and boulders ” ( 3) . Flint ( 4) has stated “ . . . its
1. Chf . , Foundation
and Materials Branch , U . S. Army Engr . Div ., New
England , Waltham , Mass .
2. Asst. Chf . , Foundation and Materials Branch , U . S. Army Eng . Div . , New
England , Waltham , Mass .
3. Numerals in parentheses refer to corresponding references in the
Appendix .
275
IN
IMA
H
N
i
.
.
uran
SOIL CONSTRUCTION LOCATION OF

NO NAME PROJECT TILL DEPOSIT
.
LIMESTONETILL LORINGAFB LIMESTONE ME

1
.
BANGORTILL DUW AFB BANGORME
,
D FRANKLIN FALLS FRANKLINFALLS WEBSTER
H
N
,
.
.
DAM TILL DAM
16
BLACKWATERDAM BLACKWATER
1
DAM WEBSTER
.H
N
,
.
TILL
SURRY MT. TILL SURRYMT DAM SURRY NH
,
.
N.
OTTER BROOK OTTER BROOK ROXBURY
H
,
.
.
TILL DAM
.
R
PETERBORO EDWARD PETERBOROUGH

WONN
,
.
V
TILL
CONNU MCDOWELLDAM
H
N
.
.
.
HOPKINTON TILL HOPKINTONDAM HOPKINTON
, . M
. N
N ,
H .
.
VICTORY TILL VICTORY DAM VICTORY
,
UNION VILLAGE UNION VILLAGE UNION VILLAGE
TILL DAM VT
.
11
HARTLANDTILY HARTLANDDAMN HARTLAND VT

N
,
.
.
. .
. 12
Approximate southerly BALL MT TILL BALL MT DAM JAMAICA VT
,
.
.
, , , .
. 13
TOWNSHEND TILL TOWNSANO DAM TOWNSHENDVT

limit glaciation
.
of
. 14
TULLY TILL TULLY DAM ROYALSTON MASS.

.
16 15
BIRCHHILL TILL BIRCHHILL DAM ROYALSTONMASS

S
.
.
BARREFALLS TILL BARREFALLSDAM BARRE MASS
.
.
WORCESTERTILL WORCESTER AUBURN MASS
,
.
DIVERSONPROJECT
W
HOLYOKE TILL HOLYOKE DAM HOLYOKE AND

.
SPRINGFIELDMASS
,
BRIMFIELDTILL BRIMFIELDDAM BRIMFIELD MASS

E
, ,
.
20 BUFFUMVILLE BUFFUMVILLE CHARLTONMASS.

.
Arrows indicate approximate TILL DAM

glacial movement 21 WESTVILLE TILL WESTVILLE DAM SOUTHBRIDGE
of
MASS
1
directions
.
22 WEST HILL TILL WEST HILL DAM UXBRIDGEMASS

across New England
, ,
.
. .
.
MANSFIELD MANSFIELD WINDHAM CONN

HOLLOW TILL HOLLOW DAM
24 THOMASTONTILL THOMASTON DAM PLYMOUTH CONN
.
,
|
Fig Location Map

1
.
.
-
may consist clay

99
of
outstanding characteristic
is
that nonsorted
%
is
It
it
particles or large boulders or any combination

99
these and intermediate

of
%
sizes Flint further classifies till into lodgement till which deposited
is
."
from drift transport glacier and tends be highly com

of
the base
to
in
in
pact and ablation till which deposited from drift transport within
. or
is
in
,
upon the terminal area shrinking glacier and tends noncompact

of
be
to
In
a
GLACIAL TILLS 277
New England , till deposits containing substantial clay sizes frequently have a
surface layer 3 to 7 or more feet thick which may be ablation till or may be
lodgement till which has been altered by frost action and chemical weather
ing . This surface layer is characteristically brown or yellow in color , in
contrast to the underlying grayish or bluish unweathered till , and has lower
density , higher water content , and higher permeability . Excavations frequent
ly encounter copious seepage from the surface till layer , with very little from
the underlying compact , more impervious till . These differences in charac
teristics are very important from the construction standpoint , since the
surface till is considerably weaker than the unweathered till and requires
flatter slopes for equal stability . When this material must be excavated to
reach the underlying more desirable till , it is generally used in the interior
of embankments , for seepage control blankets , or in other applications where
shear strength is less critical . In till deposits which have very little clay
size material , the surface zone of modified material may be almost un
distinguishable .
remainder of this paper all data and discussions will , unless speci
In the
fied otherwise , refer to the main bodies of the till deposits , below any sub
stantially wetter and looser surface zones .
In engineering literature , glacial tills are usually described as well - graded
materials , virtually ideal for earth dams , with very low permeabilities , high
angles of internal friction , and high dry unit weights . Many glacial tills do
indeed have these ideal characteristics . However , as indicated in the quo
tation from Flint given above , tills may have a very wide range of charac
teristics , and these are dependent on such factors as the nature of the rock
or soil over which the glacier passed prior to the point of deposition and the
manner of deposition . Thus , tills may in fact vary from soils in which the
clay fraction is so predominant that they have the characteristics of clay to
soils composed so predominantly of large angular particles as to have the
characteristics of a compact rock fill . Also , tills with similar gradations
may show widely different shear characteristics
Geologi sts have concluded from world -wide glacial studies that although
small amounts of hard and durable stones may be carried great distances by
ice (up to many hundreds of miles ) , the greatest proportion of rock debris is
deposited within a relatively short distance of its origin , as within 10 to 20
miles . The characteristics of materials deposited by glaciers thus tend to
be closely related to the local bedrock , in the approximate direction from
which the flow came . These observations are confirmed with respect to the
mineral composition of till deposits in New England . Where the bedrocks are
types such as gneisses, schists , phyllites , and slates which weather into sub
stantial proportions of clay minerals , the till deposits are found to contain
substantial percentages of clay - sizes , are highly impervious , and may be
quite plastic . Where the bedrocks are granitic , however , the percentage of
clay sizes is low and the tills tend to be more sandy , much less impervious
and nonpla stic . Thus , knowledge of the local bedrock geology is an invaluable
aid in searching for borrow materials of specific desired characteristics .
Experience shows, however , that the above - mentioned general and very
useful relationship is not sufficiently accurate to distinguish differences in
the characteristics of tills which are superficially quite similar but differ
subtly in mineral composition . This is true for several reasons . First , most
bedrock in New England is hidden by a mantle of overburden , and usual rock
formations of limited extent which may contribute special characteristics to
glacial drift may not be anywhere exposed to view . Second , it is possible that
coarse and fine products of glacial erosion may be transported varying
distances and thus cause a complex intermixing of materials of various ori
gins . Third , in well - graded materials such as tills it is possible that very
small amounts of particular materials or minerals may have a decisive in
fluence on the strength and other characteristics .
DEFINITIONS OF TERMS AND SYMBOLS
Definitions of terms and symbols used in this paper correspond to those

listed in " Glossary of Terms and Definitions in Soil Mechanics , ” Report of
the Committee on Glossary of Terms and Definitions in Soil Mechanics of the
Soil Mechanics and Foundation Division ( Proc . Paper 1826 ), vol 84 , No . SM4,
October 1958 , part 1, Journal of the Soil Mechanics and Foundations Division ,
Proceedings of the American Society of Civil Engineers , with the following
additions and exceptions :
ACTIVITY
The ratio of plasticity index to percentage of 0 .002 - mm size for the
portion of a soil passing the No . 40 U . S. standard sieve .
ATTERBERG LIMITS
The liquid limit and the plastic limit of the portion of a soil passing the
No . 40 U . S. standard sieve .
BOULDER
A rock fragment, usually rounded or semirounded by weathering or
abrasion , having an average dimension of 10 in . or more .
COBBLE
A rock fragment , usually rounded or semirounded , with an average di
mension between 3 and 10 in .
PROCTOR COMPACTION (STANDARD PROCTOR )

A laboratory soil compaction test in which the portion of soil passing
the No . 4 U . S. standard sieve is compacted in three equal layers in a 1 / 30
cu - ft mold (4 - in . diameter by approximately 4 . 6 in . high ) by 25 blows per
layer of a 5 . 5 - lb hammer having a 2 - in . - diameter striking face and a free
fall of 12 in . (Same as ASTM D 698 - 42T and AASHO T - 99 - 57 , Method A. )
Wp Optimum water content as determined by the Proctor compaction
test .
Dry unit weight at optimum water content as determined by the
Proctor compaction test .
GENERAL PHYSICAL CHARACTERISTICS OF NEW ENGLAND TILLS
The U . S. Army Engineer Division , New England , has used glacial tills ex
tensively in earth dam and airfield construction in New England over more
than 20 years . Locations of some of the larger projects involving these ma
terials are shown in Fig . 1.
Gradation curves and data for typical samples from each of the 24 sources
shown in Fig . 1 are presented in Fig . 2. Unified Soil Classification System
GLACIAL TILLS 279
symbols are also shown . A nominal maximum size of 3 in . has been used for
the gradation curves and data . Actual embankment materials would contain
varying amounts of coarser materials . When screened to a maximum size of
3 in . , as shown in Fig . 2 , the gradation curves of these geographically widely
distributed materials fall within a surprisingly consistent range . However ,
from the construction standpoint there are substantial differences in proper
ties of the tills according to the shapes and positions of the curves within the
band . For example , the Mansfield Hollow till, which is curve number 23 at
the lower boundary of the range , is a very bony till of only moderately low
permeability , whereas the soils which show higher than average fines toward
the lower ends of the curves have plastic characteristics and are highly im
pervious . Gradations of five typical tills for which shear strength and other
data are presented in some detail in this paper are shown separately in Fig . 3
for easy comparison . Gradations of minus 200 -mesh sieve fractions of these
five soils are shown in Fig . 4 .
General physical data on the five tills selected for shear data presentation
are summarized in Table 1. It should be carefully noted that the test results
shown therein were obtained on the minus No . 4 sieve fractions of the tills .
Test results using complete gradations would show lower moisture contents ,
higher dry unit weights , and lower permeabilities . The results represent
tests on many specimens from each source , including specimens varying con
siderably in gradation from the curves shown in Figs . 3 and 4 . Thus , the
average test values do not necessarily equate with the gradations of Figs . 3
and 4 , which represent individual selected specimens . The test data show
average natural moisture contents and dry unit weights both above and below
the values determined by Standard Proctor Compaction . Permeabilities range
from 6 x 10 - 9 to 0. 5 x 10 - 4 cm / sec .
Available Atterberg limits data on the minus 40 - mesh fractions of the tills
are shown in Table 2. Atterberg limits tests were not performed on some of
the older projects . However , it may be possible to fill in some of these gaps
before final publication of this paper .
LABORATORY TESTING APPARATUS
Special mention is made of the liquid limit and shear testing apparatus
used for the five tills which are described in some detail .
Records are not available on exact details of liquid limit apparatus used
in the older tests . However , liquid limit tests for Otter Brook , North Hart
land , Ball Mountain , and Thomaston tills were performed with the Casagrande
apparatus (5 ) . Principal differences between this and other types are the use
of a “ Micarta ” base and a grooving tool designed to accurately control the
height of groove . For the Hopkinton till , a liquid limit device with a hard
rubber base was used with the Casagrande grooving tool . As shown in
Table 2, many of the soils are nonplastic . The highest average plasticity in
dex , 12 . 4, was obtained on the Thomaston till . Fig . 5 shows envelopes of the
Atterberg limits test results for the five typical soils tested in shear . The
Thomaston and Otter Brook tills are seen to be distinctly more plastic than
the Hopkinton , North Hartland , and Ball Mountain tills . It should be noted that
till formations in New England may show moderate variations in plasticity
within nominally homogeneous deposits . This is illustrated in Fig . 5 by the
elongated envelopes for the Thoma ston , Hopkinton , and Otter Brook tills . The
U . S. STANDARD SIEVE SIZE

3IN . IN . NO. 4 NO. 10 NO:40 NO. 200
WEIGHT
_Grobatioh Ronge
BY Curves llthrough24
FINER
PERCENT
10
01
100 001
0
0
.1
.
IN 1
.0
.
GRAIN SIZE MILLIMETERS
GRAVEL SAND
OR
BOULDERSCOB SILT CLAY
'S
'L
Coorse Medium Fine

Coarse Fine
UNIFIED SOIL PERCENT PERCENT PERCENT

FINER THAN FINER THAN
OF
NAME SOIL CLASSIFICATION PASSING
I
NO 002 mm
.
SYMBOL NO 200 SIEVE 02 mm
0
0
.
.
.
3632
SM SC 42 18
LimestoneTill
-
10
Bongor Till SC 47
12 109 8 7 6 5 4 3 2
FranklinFallsDomTill SC 50 31 13
SM 37 23
123 12 9
BlackwaterDomTill
3116
SurryMI Till SC
.
Otter Brook Till SC 45

SM 32 17
W
PeterboroTill
.
HopkintonTill SM SC 50 33
-
32
Victory ML 29
a
7 1819
Union VillogeTill SM 39
Hortland Till SM 40
N
.
I
Ball Mi Till SM 30
.
Fig Typica
2
-
.
.
GLACIAL TILLS 281
U . S. STANDARD SIEVE SIZE

3IN . IN . NO.4 NO.IO NO .40 NO.200
Gradolign Range
Curves through 24
--
- 6
II
15
22
1000
1.0
001
100 10 0.1 0.01
.
0.001
0
IN
GRAVEL SAND
BOULDERS
COB OR
SILT CLAY
S
'L
'
Coorse Fine coorse Medium Fine
UNIFIED SOIL PERCENT PERCENT PERCENT

NO
NAME OF SOIL CLASSIFICATION PASSING FINER THAN FINER THAN

.
SYMBOL NO 200 SIEVE 02 mm 002 mm

0
0
.
.
.
24 22212019181716151413
TownshendTill SM 40 21
Tully Till SM 20
18
Birch Hill Till SM

BarreFalls Till SM 31 21
WorcesterTill SC 40 28
1526
Holyoke Till
56303241
SC
Brimfield Till SC
E
.
BuffumvilleTill SM 12
WestvilleTill CL
WestHill Till SM 24
27 7 9
23 MansfieldHollowTill SM 18
ThomastonTill SC 40 10
NOTE
on
All gradationsdetermined
minus inch material
3
Glacial Tills
of
radations
.
Table .
1 – New England Tills - General Characteri
Soil Paront Rock in Probablo Description and Classificat

I dontir. Sotl Nane SourceArea of Tiu of Till
No.
Onaise, schist andquartsito, Clapey SAND( SC) odth fov

Ottor Brook T111 cacoous ; Littleton formation. cobbles andboulders oxcop
on murtace.
med
Gnaiss, schist and quartsite,
SA
Gravelly alw clayey
SM
SC
,
,
Hopkinton Tin cacoous Littleton formation Only occadonal
(
-
).
plus granitic phami boulders.
SM
Phyllito and sericitio schist
10
Gravelly SAND
#
,
Hartland til
(
)
.
with calcareousolenonto. Coarn fragmentsconposad
N
,
angular schiet
SM
Mun
Gravell dlty SAND
,
(
)
Bell Till Gneiss and schiet with numerouscobblor and
.
boulders.
SC
McA schist with ignoou, clayoy SAND( very
th
v
)
Thomas
ton 1111 intrusions utulo gravel and only occ
.
cobbles and boulders
.
Noto
where single values are given without ranges, they are individual typical data
Wharesingle values are given with ranges below they are averagevalues. .
,
on
However since the moisture contant anddensity values used are not always
,
correspondingspecimens average values for these two properties are not

,
alwaysstrictly comparable
.
StandardProctor for minus entorial only

(2 (1
) )
Permeability values for Otter Brook till

de
Hartland and Thomaston

.N
,
tonned fromconsolidation tost results Permeability of Ball Mountain

till deterdned falling hoad method
by
.
.
material the upper several feet the North Hartland till deposit non
is
of
in
plastic whereas the underlying material slightly plastic should be noted

is
it
;
,
on
that the shear tests reported this paper were run the upper nonplastic
in
till rather than the plastic underlying material Some the other tills
of
in
classified as nonplastic are known minority slightly

of
Table contain
to
in
plastic zones
.
Activity ratios defined as the ratio plasticity index percentage finer

of
to
,
weight are also shown

by
tills including the

16
than 002 mm Table for

in
0
,
.
five selected for presentation this paper Ratios range from

of
shear data
in
zero for nonplastic materials such as Ball Mountain till high

91
for
to
of
0
a
Westville till
.
GLACIAL TILLS 283
i Five Typical Soils Tested in Shear
In-Placo Characteristics
Characteristics , at Optimum
Spec
. Grarity ma il fraction in LaboratoryCompaction
( 1) Ccofficient of
Permeability oa
Pino Dry Unit Dry Unit ninus #lawt' i ,
Iraction Moister Moistun en sec(2)
Content, Weight Contant, Weight
por pes
100
0. 00006x
%
- 11
77
120 est 125. 127pcs

1
3
.3
- 7
•
582
)
.
&. .
10
12
.
11
16
83
115 126last. 124 126)
2
.5
.8
,
(2
2
•
10
.
(
)
-
)
.
ntar spring 02
-4
x
0
(
.
117per
129
- .72
Not tostad.
.9 2
10
75
136 167)
(h
(7
.6
-
.
)
.
• .74
01
- 9 .5
x10
156
132 est
. 9
127. Under
722
0
.
· 5
.
75
.)
.11
121 134) aftar standard

(7
.0
1
(2
8
3
0
.
(
)
.
)
.
conpaction
.
x10
- .76
3 10
02
- 10
125. 122.
to
742
. 7
8
3
.0
0
5
"
. .
11
78
.
.13
105 135) 110 132) ufuar standard

(8
.6
(2
(5
)
)
.
)
.
compaction
,
• .744
. 10
. 13
to
124. 129. 0008 09

722
x 0
0
5
.2
4
4
.
15
76
10
122 129) 13. 122 136) aftor

(8
.7
.0
(2
7
.0
,
)
(
)
)
.
(
.
-'
standardcom
paction
,
Typical laboratory compaction curves for seven tills including the five
,
selected for shear data presentation this paper are shown Fig The
in
in
4 6
. .
.
tests were performed by the Standard Proctor Method

on
minus No sieve
Surry Mountain till which was made
on
material except for the test

on
minus
,
material Compaction tests materials with larger maximum size

on
in
1
.
-
than the No sieve would show higher maximum densities and lower opti
4
.
mum moisture contents as suggested by the position the curve for Surry
of
,
Mountain till For some tills compacted dry unit weights

of
complete
,
.
gradation embankment material not infrequently run as high as about 145

lb
per
cu
ft
.
284
-.2
Table New England Tills Atterberg Limits
Liquid Activity
.
(
Limit Plasticity Minus No 40
)
Remarks
.
No Soil Name Index Sieve Fraction
6
)(
)(
1
)(4
)5(
)(
2
22
(
(
)
63.-
0
20
18
24
TIL
Limestone
28 ( 0
). .
)
12
6 8 9 18
67
(
)%( )3( )- )-
17
32
Bangor Till
0
.
8
- .
8
0
.
36
20
SHEAR STRENGTH
Franklin Falls Dam Till

'
Blackwater Dam Till Described byothers as silty

"
""
with some clay sizes

to
Similar Franklin Falls Till
Mt
.
Surry Till
11
26
.0
)5.
.(5
. -4
72
(14
.
.-8 728
).4
3 13
Otter Brook Till

CONFERENCE
5 6 7
-
West Peterborough Till Non Plastic SeeRemarks Generally non plastic plastic
4
;
In -
%
,
25
phases liquid limit

18
.;
.
3
4
.9. 8.-
15
plasticity index
+
40
.0
activity
21
(
0
7
),0
-
.0 39
-5.
). 2
8
14
21
Hopkinton Till
'
"
Variously described as silty
9
Victory Till
"
"
and cla yey
10
Union Village Till
.0
). 0
24
-0( 11
.(
1
.-
14
17 0
4 24
).0
North Hartland Till

L
-
TICITYwith 29 and PI
-19
-7,
0
10
,8
14
Tully Till
to
Plasticity zero very low
15
Birch Hill Till Non Plastic
- -
Barre Falls Till - Non Plastic
1
53
8
6
.
Worcester Till 24
6
6
,0 .0
55
25
Holyoke Till
25
,0
)4.
. . , -3
0
)7
.(
(
. 4.7.-
East Brimfield Till
19
1 3 27
10
21
-
Buffumville Till Non Plastic
26
11
(
)
-
-0(
)
Westville Till
.0
91
18
32
16
-
West Hill Till Non Plastic
-
Mansfield Hollow Till Non Plastic
4
GLACIAL TILLS
23
12
. 0
.-,0
(
,3)
Thomaston Till
(
87
,-
.
25
6 33
10
4
18
NOTES
on
40
Atterberg Limits tests performed material passingNo sieve
:12
..
is
of
to
of
(
,
)
Activity defined as ratio Plasticity Index percentage 002mm clay size
"(
"
'
.A
,I
.
)2. . ,
.
See Colloidal Activity of Clays by Skempton Vol 3rd Int Conference on

).
Soil Mechanics and Foundation Engineering 1953 The values givenare for theminus
.
No
40
sieve fraction of the sample
in
3
.
' ..
. .. ,
.
no
Rangesare shown parentheses Where rangesare shown results represent singletests

285
U. S. STANDARD SIEVE SIZE

3
3IN FIN. NO.4 NO. 10 NO. 40 NO. 200
SENA
-- --
-
WEIGHT
---
BY
FINER
.c SC
OTTERBROOK TILL BTT B5 ClayeySAND
2412118 6
, ,
, 1,
10,
(
)
-
FD
HOPKINTON TILL 147LB Gravellysilty layey
SANDISC
SI
,
,
-
-
NORTHHARTLAND TILL BT 153 Silty SAND SM
- ,
(
)
-
SM
, B3
BT
BALL MTN TILL 102 G ravellys ilty SAND
PERCENT
, ,
, ,
(
)
.
SC
THOMASTON TILL BTT B1 ClayeySAND
-5
(
)
NI 24
10
1000 100 01 001
0
0
.1
0
.0
IN 1
.
GRAVEL SAND
OR
BOULDERSCOB Coorso Medium Coarsel Medium Fine SILT CLAY
S
L
''
Fig Gradations Five Typical Soils Tested

of
Shear
in
.3
-
.
.
IIIIIIIIII
SC
OTTERBROOKTILL BTT- 85 Clayey
SAND
2412118 6
, , FD,
, 1,
,
SH
, (
)
HOPKINTONTILL 147 LBIO Grovellysilty layey
SANDSC
SM
,
, ,
BT -
-
( .c
NORTHHARTLANDTILL BT 153 SiltySAND
- ,
)
, B3-
BALLMTN. TILL 102 Gravelly siltySAND(SM)

,
,
Clayey SC
BI
THOMASTONTILL BTT- SAND(

PERCENT FINER BY WEIGHT
5,
)
200MeshSieve
001
05
02 01 005 002
0
0074
0
0
0
IN0
.
1
0
.
.
.
.
.
Fig Gradations Minus 200 Mesh Fractions for Five Typical

of
4
-
.
.
Soils Tested Shear

in
.
GLACIAL TILLS 287
PLASTICITY INDEX
60
OS
Area
Covered Line
-
ByLarger
A
Thomaston
Till
Piot
PI
20
40
60
80
INDEX
100
LIQUID LIMIT Otter Brook Till
PLASTICITY CHART
After Casagrande
A
(
)
.
id
PLASTICITY
)
-20
HopkintonTill Line
LL
.-
(
73
=A
0
PI
North HortlandTill
BallMountainTill
15
20 25 30 35
LIQUID LIMIT LL
Plasticity Index Liquid Limit Minus No

vs
Fig 40
Sieve Fractions
5
-
.
.
.
1401
. SITE SAMPLE
. , , , , , , 112
Surry Mountain Dom
BH BT
B1
I
,
- -
Zero
FT
71
North Hartland Dom 830
7 6 5 4 3 2
. . . . . .
BallMountain Dom BTT B9 6

- -
CU Air Franklin Falls Dom ADT No
1
2
.
:'2.
voids
CF FD
Hopkinton Dam 147 LBIO

PER
- -
.
- 14
Otter Brook Dom VB2

Gs
=
81
2 ThomastonDom BTT
5
70
POUNDS
89
IN
1
DENSITY
DRY
IAL
10
11
12
13
14
15
6
MOISTURE CONTENT
IN
PERCENT
NOTE
on
Tests performed Minus No sieve material by Standard

4
- .
Proctor Method ASTM D698 42T exceptfor Surry Mountain

(
on
Dom BT 117 B2 whichwas performed Minus inch

,
1
,
,
,-
material
Fig
Typical Compaction Curves

6
.
-
.
.
.
288
.. .ft.
.ft. 10ft
/T-
TestDuration
TestDuration
sq
Minutes
Minutes
/ 8 /
.1
5 sq 8
sq
810
T
-
APPLIED PRINCIPAL STRESSES

Tyga
T
- .ft .0.5
50
.. ..
f.
6
.
ft 0 2
/. .T/t .
TRESS
STRESS
STRESS
INITIAL
INITIAL
115
LbsIcu
115
cu
TEST
115
115
Ya YA
50
.
Lbs
sq
11
.2.4 .0=
4
22 50
.0 .. 55
DEVIATOR
1313 +
C$
DEVIATOR
2 10
/T .0 .1
50 ft
.
SHEAR
sq O
0
..
0
.ft. /T f.
10ft
./t8
10
TestDuration
TestDuration
8-
sq
/T-
Minutes
Minutes
sq sq 1o
-
T
1816

TRIAXIAL
f.
. ..
\.
./t 9..7 .9 .f.t 6 6
. /.
11 44
STRESS
.
STRESS
INITIAL
32
INITIAL
129
126
125
123
Yocu Yocu
30
.2
4 32
.C:
Lbs
T
Lbs
sq
4 4
|
16
..30.
0
140
1 1216
.
08
=$
DEVIATOR
.f.t
121212
DEVIATOR
3864
08
2 1 2 2
/ 1..2 .4 UNDRAINED
08
sq 1632
T
0 0 0
.ft
12 12
.f.t .f.t
12
/T-
TestDuration
TestDuration
. ..
sq
..
Minutes
10 10
04
Minutes
10
556
/ /T
sq
sq
24
27
-8
27
T
22
.
23
-8 8
. .. 4
UNCONSOLIDATED
32
ft
/.
ft
.
32
STRESS
STRESS
1184
1224
INITIAL
I.2
1277
1223
INITIAL
120
yacu cu
14
Lbs
Lbs
16
. .
2
.C
16
0 0
Tsaft
4
120
%
9 3old 12
DEVIATOR
DEVIATOR
12
JY
=1:7
0 .11 ./ .
pel
/T .12.4
ft 11
00
$
32
sq 0816 2 sg 1 16
T
o
/Tsq
IN
PERCENT STRAIN AXIAL

IN
PERCENT STRAINAXIAL STRESS SHEAR

ft
-
.
.
/Tsg
5
f1
- TAG 084 TAGIT 005 TAG
.c.
0
c0
.0. .
23
.
-C 0
14
38°
6
8
0
2
6
4
6
0
2
10
10
10
12
STRESS SHEAR
sq
sq
sq
T
/T
T
/
/
.f.t
..ft
.ft
-
8-
8-
APPLIED PRINCIPAL STRESSES APPLIED PRINCIPAL STRESSES APPLIED PRINCIPAL STRESSES
CONSOLIDATED UNDRAINED TRIAXIAL SHEAR TEST
T
-
.
f.t
ft
T/
..
T
-
/
..
-
SHEAR STRESS sq
sq
SHEAR STRESS
sq
- SHEAR STRESS
4
6
6
/8
ft 8
10
12
02
.0
2012 140
InchesDISPLACEMENT
INITIAL INITIAL INITIAL
Normal Normal Normal
HORIZONTAL
of
of
of
Stress Rate Strain Stress Rate Strain Stress Rote Strain

.
ft
T
.
ft
ft
T
T
.
In
In
./.
.
Min
sq
sq
Min Min
cu
Lbs
cu
Lbs Lbs
Yocu
sq 0
122 0010 126 00032 115 00078
9
122 0010 125 00024 138 1156 00072
.f.t .8 .6 .7
/ . ..
0 00
/ .. .
12 4
00
. 0. 0.
2
.. 8. . 9.
/ ...
.8 .9.2
In 0 0 0
/. .0.0.0
124 128 00024 115 00085
121212
101111
0010
/ .2.2.4 .4
.f.t 000 0
ft 5 2 6
.8.
./ . . .
ft 115 000092
1413
.
GLACIAL TILLS
- PeakValues PeakValues
=
.T/ ft
0
sq
10
C$
=
.50
.
C0
.:
34
in
at
33° Values Displacement Average
PeakValues
0
of
=
5 in
=
0
Values Displacement
C
:
C0
C0
:
33°
32
=0
32
2
4
6
0
2
0: 4
0
2
4
6
10
12
10
10
TasqSTRESS SHEAR
ft
8T-
/
sq
sq
T
ft
.ft
8.
.
8.
sq
NORMAL STRESS NORMAL STRESS NORMAL STRESS
CONSOLIDATED DRAINED DIRECT SHEAR TEST
-,
N
-,
,
,
-5
SC
SC
THOMASTON BTT B1 ClayeySAND OTTERBROOKBTT Ciayey

SAND HARTLAND BT 153 SiltySAND
SM
166
%
cu
/- , 3.
11
)..ft
(/
Yp
)(.f.t
B5Yp
.0 6 ,,1 =
cu
%
%
11
2
)(.f.t
/
,3
12
/T– 0.
..
,,..
3
cu
.. /
.- 0
Wp 123 Lbs Wp Lbs Wps Yp121 Lbs
126
NOTE
on
.
.
4
No
Sheor tests performed minus moteriol
.
-7.
.I-
Fig Typical Shear Test Curves

289
290
-
/
.f.l
T
-
/
..ft
DEVIATOR STRESS
sq
DEVIATOR STRESS
18
20
0
22
24
2
10
12
14
T 16
18
16
10
12
sq14
18
1
.
IN 08
216
PERCENT STRAIN
INITIAL INITIAL
Tast
AXIAL
Duration Test
Duration
YO
.
.
Au Minutos Lbs Minutes
yocu
bs
131 32 124
08
ft 1 3
2
5
131 123
16
.. ...5
. 8...8
l .1..4
131 123
32
SHEAR STRENGTH
.
it
C
2
..1C
.
tisoit That
.0.3..
sa 16 39
1538
/T
-
-
1
STRESS
CONFERENCE
SHEAR
4
8
2
6
8
14
20
24
10
14
18
20
10
12
16
18
22
12
T
-
/.
.ft
.
-
APPLIED
16sq
PRINCIPAL APPLIED PRINCIPAL STRESS

11
STRESS The
UNCONSOLIDATED UNDRAINED TRIAXIAL SHEAR TEST
..ft
f.t
R
/
-
-
STRESS
sq
DEVIATOR STRESS DEVIATOR STRESS

6
**
12
T14
16
18
20
24
/16
10
12
14
.18
sq16
10
20
32
34
2J6
/
.
1
IN 08
PERCENTSTRAIN
INITIAL INITIAL
Tost
Duration Test
Duration
AXIAL
14
YA
T
!
/
/T
Lbeleu Minutos Losluft Minutes
9
%4
11
.
IO 08 1129 12
.
o59.12
.?2
30 229
10
STRESS
SHEAR
4
6
0
10
12
14
16
18
20
0
22
24
2
6
8
APPLIED
T
10
-
./. ft
12
PRINCIPAL
16
STRESS sq 18
20
-
".
CONSOLIDATED UNDRAINED TRIAXIAL
APPLIED
SHEAR
PRINCIPAL
TEST
STRESS
.T/ 89
T
-
/
.
SHEAR STRESS
sq
T
-
/
SHEAR .
1
.
8
STRESS
.
sq
-
10
12
6
14
ft 16
8
2
10
Inches
12
tl14
16
18
25
DISPLACEMENT
Hot
p01
INITIAL INITIAL
Normal Normal
Stress
of
Role Stroin Stress
%
of
f.t
Los Role Stroin
1
Min
ft
./.
Icu
Lbeku Min
ft 5
130
.
00018
8
122
HORIZONTAL
130 sq 0
. 2 0013
ve 100
00018
/1 50.1..
3
sq 500
123 0012
... In 000
.586.
. ..9.5
, 130 00018
0 0
/ ...
in000
.11.5.0.1
11 122 0011
/sq
T
- PeakValues
.C.4
*
Tsh
36
388
C6
.0..
3840
GLACIAL TILLS
PeakVolues
. in
Values Displacement
0
C
STRESS
of22
.
0
4
36
.:3.6
.
6
8
10
12
0
14
2
4
18
.50 1 6
20
22 24
8
10
12
/
16
18
ft
SHEAR
NORMAL
20
STRESS
16sq
T
NORMAL STRESS
f1
14sq
CONSOLIDATED DRAINED DIRECT SHEAR TEST
,-
,
HOPKINTON 147
LB
FD
clayey
10
SAND
,-
BALLMOUNTAIN
BT 102B3 Silty
SAND
9
%
2
.
ft
Wp
SCcu
129 Lbs
%
..0
,.
- , 9.
).
Yp
../ (./
Wp
10
122 Lbs
Ml
SMcu
Yo
NOTES
in
. )( ../ .
on
Till
Sheartests Hopkinton performedminus moterial
.
.
-T , ..2 ,4
Sheor
onon
tests BallMin Till performedminus No moterial
on
Initialwatercontents
to
of
anddensitiesare prior application
or
anyconsolidating
load pressure
..8-
II
Fig
-.
Typical Shear Test Curves

291
292
SHEAR STRENGTH CHARACTERISTICS OF FIVE TILLS
The following five of the soils listed in Fig . 1 have been selected for
presentation of results of detailed shear strength studies in this paper :
Identification No . Till
Otter Brook
coco
Hopkinton
North Hartland
BallMountain
Thomaston
The Otter Brook till was selected because the unusual bulging which oc
of
curred Otter Brook Dam during its construction which described later
is
in
,
this paper The other four soils were selected because extra study was
in
which they were used as result

of
of
made the dams
or
these soils the

in
a
Otter Brook Dam experience
.
by
on
Shear tests the Hopkinton till were performed the Army Engi
S
.
.
neer Waterways Experiment Station Vicksburg Mississippi Tests on the
,
.
North Hartland till were performed by the Division Laboratory Army
U
,
S
.
.
Engineer Division North Pacific Portland Oregon The remainder
of
the
by ,
, ,
.
tests were made the Division Laboratory Army Engineer Division
,
U
S
.
.
New England Waltham Mass
,
All shear tests were run sieve fractions except Hopkinton

on
minus No
4
,
.
till for which minus material was used Figs and show stress
in
3
/8
8
.
-
, .
strain curves for the five selected tills all cases unconsolidated
In
.
undrained and consolidated undrained tests were run triaxial compression

in
-
.
by
of
Consolidated drained tests were run the direct shear method Results
-
.
selected moisture content conditions rela
at
shear strength tests conducted

optimum are shown together with data
on
tive Table strain conditions

in
to
3
,
at
which strength values were selected and sample dimensions For the
.
Hopkinton till adjustment was made for the slight difference optimum
in
,
moisture contents between minus No and minus materials mold

in
on in
4
3
8
.
.
-
/
ing specimens for shear tests molding water content

of
The effect the

.
shear strength Otter Brook till Fig summary plot

of
indicated and
is
in
a
,
.
shear strength curves for all five soils various molding water contents
of
at
Fig
or
dry optimum Proctor optimum moisture

10
wet shown
is
of
in
.
.
contents for the minus No fractions the soils used the shear tests
of
in
4
.
were as follows
:
Till Moisture Content

%
,
11
Otter Brook
3
.
Hopkinton
9
9 0 2 2
. . . .
North Hartland 11
10 10
Ball Mountain
Thomaston
Direct shear test apparatus were essentially one type with only minor
of
Specimens ranged from

at 38
variations the different machines used

to
in
2
.
square and height Porous stones were provided top

to
in
in
in
4
3
4
1
.
.
/
specimen permit drainage Shearing was accomplished by

of
and bottom
to
separating the top and bottom halves the shear box slightly and shearing
of
-
UNCONSOLIDATED UNDRAINED TRIAXIAL COMPRESSION TESTS
sq
/T ft
08
.2
16
32
/T
..
/T
sq
sq
.f.t
.f.t
.1
.4
=
=
Loterol Pressure Lateral Pressure Lateral Pressure
. N
FT
.
SQ
/
-
T CONSOLIDATED UNDRAINED TRIAXIAL COMPRESSION TESTS
32
sq
ft
08
sq
.4
/T
/T
sq
=
.f.t
..
ft
T=
.1
16
.2
=
/T
..
Lateral Pressure Lateral Pressure Lateral Pressure
–
STRENGTH
-
CONSOLIDATED DRAINED DIRECT SHEAR TESTS
GLACIAL TILLS
sq
.f.t
/T0.
4
sq
=
:
sq
.f./t
.1
0
T
0
"
T
.2-
Normal Pressure Normal Pressure
.f/t Normal Pressure
SHEAR
NOTE
as
Optimum water content
by
determined Stondord Proctor
is
11
.3 %
Compaction Test
+
-2
+
+
+
0
2
4
6
+
+
+
+
+
+
-
-
-
2
0
2
4
6
1
2
0
2
4
6
4
MOLDING WATER CONTENT MINUS OPTIMUM WATER CONTENT

vs
.
-9.
Fig Shear Strength Molding Water Content
293
SM
,
,
,1-
(
.)
Otter Brook Till BTT B5 Clayey Sand

-- - -
Table 3. -- Summary of Selected S
A, Compi
Method Stressor Strin AzountWet
Identiti ConditionUsed Dry of Op1
Cation Soil Shoar to Dotard
.no Motstan a
No. Nano TypeTest Test ShearStrength
, UU
to
Unconsolidated
undrained triadul 20%strain - 3.6
Ottar , CU
Consolidatedundrainod triaxdal
Brook
till
those
11dainddrained, CD
Conso direct 0. 5 in. dplace
you
Unconsolidated , U
undrained triadal m . str . - 2. 3
NAX. stress
20%strain 12. 1
Hopkinton Consolidatnd , cu
undrained triadal MAX. stress
205stran 1
20% stran 12. 8
drinod , CD
Consolidated direct nar, stres, - 2. 1
. 05in. dipl . -0. 2
.07 in. dipl . 1. 7
Unconsolidated , U
undrained triadal max
. stress 12. 0
1
North
Hartland Consolidated , cu
undrained triedal
Tu
tad
CD
Consolida drained direct

,
one
§
that
Unconsolidatod
undrained triadal max stress
co w
,
.
Ball
Mountala Consolidated
undrained tranda
CD ,
Mu
Consolidated
drained direct
38
,
dipl
in
0
5
.
.
UU
Unconsolidated
undrained triedal Rex. stresa
,
yoq
CU
do
ThomastonConsolidatod
undrinod tradal
,
TIU qop
CD
Consolida
tad drained direct
,
qos
ho
y
All tost specimens

compactedb Standard
on
Proctor procedures Testoperforend

til
.
ndnusNoli natorial exceptHopkinton

,
on
whichwereperfomad rdnus natorial

in
3
/8
-
.
the sample horizontally under

of
constant normal vertical load The rate

a
.
(
strain was maintained essentially constant and low enough insure consoli
to
dation during shearing

.
Both strain control and stress control triaxial shear test apparatus were
the five tills were as follows
of
used The specific types used for each

:
.
piston
by
Otter Brook Till Stress control type axial load applied dead
to
;
-
.
weight sleeve type piston with brass sleeve and ball bearings top and bottom
.
;
Hopkinton Till hy
by
Strain control type axial load applied piston

to
;
; .
-
draulic pressure polished steel piston running between two Thompson linear
motion ball bearings piston sealed by two rings
in
O
;
"
"
.
GLACIAL TILLS 295
rength Test Results on Five Typical Soils
Sin of Test
degreestonsAtt . Spacio SaturationNotas Torting Agency
New
dia
0, 9 Not satinind prior England

Divison Laboratory
to
| 1. 6 in. tost
in
3
*
.
.
. 1
752
.
0
dia
10
84 Satirated prior
to
tant.
.0
in
1
3
.
.
75
aquan
in
in
3
.0
x
.
O thiclonose
Notstirind prior Waterwaye

Expo
to
din. test rimontStation
1
. . .6
in
in
2
.8
7
1
/4
1515
*
-
.
.
11
Satıntad prior
to
.0
tost
.0 0
.
. 1
75
0
38
1n
quer 786
in
2
0
x
.
thickness
dia
73
Saturatedprior NorthPacific Diridon Laboratory

to
tast
.1
26
in
in
.4 . 0
.
.
.
0 0
Satinted prior
to
tost
00
. .
05
yuto Not stanted prior

to
tasto
, in
in
13
4
x
x
.
thick
0
1
in
.
dia dia
prior NewEnglandDiridon Laboratory

to
Notstarted tast
in
in
0
2
6 6
x x
.
8,
.
.
. .
Saturatedprior
to
in
in
tast
2
.8
equam
00
in
in
Saturatedprior
.0 .
to
3
0
.7
do
do
tasto
:
.
New
dia Motsatinind prior EnglandDivision Laboratory

in
in
to
tant
.4
3
.5
.9 0
x
.
.
.6 0
0
Satunted prior
to
dia
in
in
tanto
1
.4
3
.5
x
.
.
75
quan Satuntadprior
to
in
tasto
.0
in
0
x
3
,
.
.
by
North Hartland Till Strain control type axial load applied piston
to
;
–
.
motor driven gear system through external proying ring sleeve type piston
;
-
with aluminum sleeves and seals

.
Ball Mountain Till Strain control type axial load applied

by
piston
to
;
.-
-
motor driven gear system through external proving ring sleeve type piston
;
with brass sleeves and ball bearings top and bottom

.
Thoma ston Till Strain control type axial load applied piston by motor
to
;
.
-
driven gear system through proving ring proving ring and piston located
;
by
inside triaxial chamber load applied piston

on
screw rod
;
.
BALL MOUNTAIN TILL , BT - 102 , B3 , Gravelly , silly SAND (SM)
HOPKINTON TILL , FD - 147 , LB 10 , Gravelly , silty SAND (SM - SC )
.
.!
+182 ci +O
.
+0.1
T-22
-023V +07
FT
/T
SQ
- THOMASTON TILL , BIT - 5 , B1, Cloyey SAND (SC)
+
2
ro
STRENGTH
SM
NORTH HARTLAND TILL BT 153 and 166 Silty SAND
,
,
-
)
SHEAR
3
.5
prior
Soturored Notspturored
+ +
.54 3.1
C
To test prior test
to
e
BIT
85
SC
OTTER BROOK TILL clayey SAND
,
,
-1
(
)
17
-
23
+ +
42
12
4
NORMAL STRESS TI SQ FT
-
.
.
UU TRIAXIAL CU TRIAXIAL CD DIRECT

-
-
-
NOTES
Figures beside curves are molding woter contentminus optimumwater content
as
UU trioxial specimens not saturated prior testing except noted

to
CD
CU
os
specimens saturatedprior consolidation except noted

on to
and
All tests exceptHopkinton performed minus No material Hopkinton tests
4
,
performed on minus material

in
3
/8
-
.
Fig Shear Strength

10
Curves
--
.
.
.
GLACIAL TILLS 297
The shear tests were run by various personnel, on equipment of different

sizes and design , in three widely - separated laboratories , over a period of
more than four years . Although the personnel were experienced soil tech
nicians , the tests were planned and performed as routine design tests , not as
research tests . Thus , the degree of accuracy of the test results is considered
" average . ” Also , as shown in Figs . 7 and 8 and as indicated in the fifth
column of Table 3, some of the triaxial tests did not reach strength peaks at
20 % strain which was used as a limit for selection of test values . Thus the
strength values at this maximum strain , which were used as the actual shear
strengths , are , to varying degrees , approximations of actual maximum shear
strengths . However , it is believed that the conclusions drawn in this paper
are commensurate with the degree of precision of the test results .
Inspection of Table 3 and Fig . 10 shows the following ranges of total shear
strengths for the five tills , at or near optimum moisture content (within - 0 .6
to +0 . 2 %), selected at normal stresses of about 5 T / sq ft .
UU CU CD
Lowest d = 17° 0 = 23° 0 = 32°

values C = 0 . 9 T/ sq ft C = 0 . 5 T / sq ft C = 0 . 0 T / sq ft
Thomaston till Thomaston till Otter Brook till
Highest = 38° Ø = 38° $ = 38°
values C = 0 . 2 T / sq ft C = 0 . 4 T / sq ft C = 0 . 4 T / sq ft
Ball Mountain till North Hartland till Ball Mountain till
Substantial reductions in unconsolidated - undrained shear strength are shown

for specimens compacted wet of optimum with the Otter Brook material , as
illustrated in Fig . 9, showing one of the sharpest reductions.
Other tests have indicated progressive increase in shear strength with in
crease in maximum size of the soil fraction used in the shear tests. There
fore , shear strength of the actual embankment materials containing sizes up
to and including cobbles and boulders may generally be assumed to exceed the
values given herein .
BULGING OF OTTER BROOK DAM
During the late summer and fall of 1957 bulging of the embankment of
Otter Brook Dam located near Keene , New Hampshire , occurred during con
struction in sufficient amount to cause considerable concern .
Otter Brook Dam is of rolled fill construction . The height is about 133 ft .
As shown in Fig . 11 , it is of homogeneous , impervious cross section , except
for a pervious fill drainage blanket and chimney , and rock - fill and gravel bed
ding slope protection . Both upstream and downstream slopes are 1 on 2 - 1 / 2 .
On the right abutment the foundation is rock , consisting of mica schist with
some granite and gneiss. The left abutment consists of a deep deposit of
glacial till . In the valley bottom , glacial till is covered by a thin layer of
silty gravelly sand outwash with cobbles and boulders , and by other minor
deposits . A cutoff with 12 - ft base width was extended to depths of 5 ft on the
left abutment and 10 ft in the valley bottom . Impervious fill was placed di
rectly on bedrock on the right abutment .
The glacial - till embankment material was obtained from a borrow area
opened in the left abutment above the elevation of top of dam . It consisted of
bow
Dom
¢
Dom DOWNSTREAM
UPSTREAM
2
25' .
.,
L
.
.
Bridge Pier
El
Top 802
of
Dam
!
.
Spillway Crest 781
El
is
.
'
.
70
na
Fill
Rock Fill ETTO Fill
k
6
2
Rock
.5 Fin
10
32
'
Gravel Fill
3A 6
.
'
18
'3
28
38
1A
Gravel
2A
,
,
-
,
8
Fill lmpervious
683 ervious
Impervious Fille
Fill
1
'
.
EEEEEEE
Po
CAVA Pervious Fill
m
m
Silty Grovelly SAND and Silty SAND and
Bedrock
,
.
111 Clayey sondy GRAVEL Sondy CLAY
Bedrock
,
with Boulders TILL
)
Silty SAND and Silty Sondy ,
GRÁVEL with Boulders
,
OTTER BROOK DAM
Dam
- DOWNSTREAM
Dam
517
. Top
UPSTREAM
El
]
of
'[1
.
El
Spillway Crest 424

'
KPI 15
Fill
lo
Rock
\'
Random
\\
lear
A
Fill
l
Sond andGrovel
TULIWA Impervious Fill Fine Rock FM Bedrock
)
Bedrock Select Pervious Fill . Sond and Grovel

:
:
.
.
THOMASTON DAM
UPSTREAM Dam DOWNSTREAM

A
Dam
572
EDTop
E1
of
Fill
El
Spillway Crest 546

in
.5
tham Pervious
'
ti
l'
una
dopo Gravel Bedding Rock Fill

2
'
l'
799
87 GravelBedding
EL
480
X
I' l'
i
(
Rock Pervious Fill

Fill
EL
450
.4
13.
4
NU
El
ImperHous 422
.
'
Fill
6,
.
5
.
,
6
trong Is Bedrock Pervious

Per Fire Gravel Bedding
Bedroch
report
Bock
Fill
NORTH HARTLAND DAM

piezometer
3
IN
SCALE FEET
Fig
Typical
11
Dam Sections
-
-
.
.
I.
GLACIAL TILLS 299
a well - graded gravelly clayey sand (SC ) having from 10 - 20 % gravel , cobble ,
and boulder - size material . Gradation curve is shown in Figs . 2 and 3 . Gra
dation of minus 200 - mesh fraction is shown in Fig . 4 . Stones larger than 6 in .
were removed . The fill was placed in 8 - in . layers and compacted with six
passes of a 35 , 000 - lb sheepsfoot roller . Additional compaction was provided
in an irregular pattern by the hauling equipment traveling over the surface of
the compacted fill . This additional compaction is reported to have been suf
ficient to have caused some weaying of the surface of the fill under the equip
ment . Explorations during design of the dam had indicated that the borrow
material would be 2 to 3% wet of optimum , as indicated in Fig . 13 . However ,
these explorations were made in the winter and spring when the borrow area
would have been at its wettest . Actually , the summer of 1957 proved to be
one of the driest on record , and substantial drying of the material occurred
both in the borrow pit and on the embankment . As a result , the material as
actually placed averaged slightly dry of optimum moisture content and the
average dry unit weight exceeded that at laboratory optimum , as indicated in
Fig . 14.
The footing the service bridge pier ( see Fig . 11) was poured on 12
of
August 1957 at
el 739. This is about 53 % of the full height of the dam . Sever
al days later it was noticed that the bridge pier had apparently moved in an
outward and slightly downward direction . Reference stakes were therefore
installed at various elevations and stations on both upstream and downstream
slopes of the dam and periodic observations were initiated . Additional stakes
were added as the dam was increased in height . The rate of embankment con
struction and the measured movements of the pier are shown in Fig . 15 . It
was found that movements were closely related to the placement of fill in that
they halted within a day or two after placement was stopped and began again
soon after placement was resumed . Construction was halted for seven days
in September 1957 after the rate of movement appeared to become increasing
ly rapid as the narrowing of the top of fill caused the embankment to rise
more rapidly . During this halt the properties of the materials and the analy
sis of design were thoroughly reconsidered .
Stability analyses of the upstream slope using consolidated - drained shear
strength and the pore pressures obtained from the piezometers indicated an
unstable condition . However , analyses using unconsolidated -undrained shear
strength indicated adequate strength . During these studies , unconsolidated
undrained shear tests performed on 2. 65 - to 3.64 - in . - diameter undisturbed
samples carved from 1 - cu - ft specimens removed from the fill gave shear
strength values of 6 = 9 to 32 degrees , C = 0. 7 to 1. 5 T / sq ft , average , 0 = 19
degrees , C = 1. 3 T / sq ft . These results indicated strength somewhat greater
than design strength . The specimens contained from 6 to 10 % of plus No . 4
sizes , whereas the design tests were run on minus No. 4 material. It was
concluded that the embankment was stable in spite of the first of these analy
ses and in spite of the somewhat alarming rates of movement , and construction
was resumed . However , the rate of raising of the fill was restricted there
after as a precaution . When the embankment was completed in October 1957 ,
the total movement of the bridge pier had reached 3. 0 ft horizontally and 0 . 5
ft downward . Although the bridge pier remained plumb , this required some
modification of the pier in order to accept the already fabricated two - span
steel service bridge to the gate tower . It is likely that there had been some
bulging of the embankment before placement of the bridge pier . Also , the em
bankment was raised about 5 ft above the base of the bridge pier footing before
300
UPSTREAM DOWNSTREAM
Dom
Fill
Rock
of
Top
Dam 437
ih
'' 42
Spillway 416 Gravel Filk
ile
El o
.'
10
Spillway Crest 416
E1
El
.
. ',
_
385 425 384
59
125425
2515
Impervious Fill Compacted Pervious Fill
'
to
7
+
)
(
,
Grovelly Silty SAND TILL 100
HOPKINTON DAM
Dam
DOWNSTREAM
UPSTREAM 210
'
EI
& 8
Top
TOP Dam IO52
ofEl
."
.
1049
'.
El
Spillway Crest
1017
SHEAR STRENGTH
165
-
Gravel Fill
B
Processed
Fine Rock Fill
A
I
Impervious Fill
PCH
.6
Fill Rock Fill
'
El
.
845
.2
25
CONFERENCE
=
.
-
Boulders
-
a
with
-
Sondy Graver
,
Gravel Filt
(
Grovelly Silly SAND
V
TILL
m
~
)-
~
~
Bedrock
-
•
Tip
Piezometer
BALL MOUNTAIN DAM
IN
SCALE FEET
.I-I
12
.
-.
Fig Typical Dam Sections
GLACIAL TILLS 301
PERCENT
IN
CONTENT
WATER
NATURAL
+
4
6
-
NATURAL WATER CONTENT MINUS OPTIMUM WATER CONTENT
Fig Natural versus Optimum Water Content Otter Brook Till

13
.
.
.
-
45
LEGEND NOTE
SAMPLES
on
TotolSample Hystograms
based
40F
-
-
-
totol 185samples
of
MinusNo. Fraction
o
4
AVERAGEMOISTURECONTENTS AVERAGE DENSITIES

»
OF TotalSample TotalSample 129 pcf

%%
. 9
.7
.5 .8
- -
- -
10
30
MinusNo MinusNo. 126 pct

TOTAL NUMBER
.4
OF
Averageoptimum
PERCENT
Dryunitweight
, at
wotercontent averageoptimum
-
- ,
11
%
125 pcf
#
#
4
.3
.7
-
-
118
10
IN 11
12
13
14
120 122 124 126 128 130 132 134 136 138
9
IN
MOISTURE CONTENT PERCENT DRY UNIT WEIGHT POUNDS PER CU FT

.
.
Fig Density Summary Glacial Till Impervious Fill

14
Field Moisture
-
-
.
.
Otter Brook Dam

.
DTTTT
FEET
825
ment EI, 802.0
Topof Embonk Topof DamEl.802O
800
800
IN
FEET
7756
Sta
12
00
ELEVATION TOP OF FILL
+
.
750 IN
Sta
00
13
+
,
so
.
Bridge
725 Pier
ELEVATION
Upstream
Slope
Slope
-
7004 Downstream
of *
00 00
12 13
III
Base Dam Sta
LIIIIIIIIIII
+ +
.
Fill 671
of
675 Min. ElBase Sto.
-
.
650 650
625
2
MEASURED MAXIMUMHORIZONTAL
MOVEMENTOF OUTER SLOPES
IN
FEET
Vertical Movement
FEET
IN
|
T
OF PIER
IIIIIIIIIIIIIIIIIII
T1
MOVEMENT
HorizontalMovement
PERCENT OF OVERBURDENPRESENT
PZ 28
PRESSURE
-
38
PZ
-
PZ 28
-
38
PZ
3A
-
PIEZOMETRIC
2A
CPZ Стттт
-
Fldefective
)
‫لللللللللللللللللللل‬
TTTTTTTTTTTTTTTTTT
40
30PZ 3A
ABOVE TOP OF FILL
PIEZOMETRIC LEVEL
PZ
28
-PZ
38
28
-
PZ
-
3A
PZ
-
3B
PZ
-
2A
PZ
NOTE
-
Only onthlyreadings
bi
ALLIILLILIIIIIIIIIIIIII
-m
shown after 1957

detective
JA ODF JA
)
DIF DIF
O
A
- A
O
A
N
A
1958 1960
J
1957 1959
-
and
Fig Pore Pressures Slope Movements Otter Brook

15
Dam
.
-
.
.-
-
GLACIAL TILLS 303
the latter was poured . Hence the maximum bulge of the embankment at the
pier was probably slightly in excess of 3 ft . The maximum horizontal move
ments measured at the bridge pier and at representative stakes are plotted
against elevation on the dam in the diagram at the upper right of Fig . 15. It
will be seen that the elevation of greatest horizontal movement must have
been very close to that of the base of the bridge pier . Except for the point at
el 773 , the points for which horizontal movements are shown in Fig . 15 were
not established until some time after completion of the embankment to the
elevations indicated ; therefore , the indicated maximum horizontal movements
of these points are less than actual total movement. On the downstream
slope , as shown in the diagram at upper right of Fig . 15 , movements were
significantly smaller than on the upstream slope , the maximum movement
from the time observations were started being 0. 5 ft . The effect of the move
ments on the finished embankment is today visible only as a barely noticeable
bulging of the upstream rock - fill slope . Thus , if the bridge pier had not been
located as it was on the upstream embankment of this dam , or if its con
struction had been delayed until after completion of the embankment , the bulg
ing might actually have passed unnoticed . This suggests that bulging of simi
lar relative magnitude may have occurred in other earth dams in the past
without being observed .
In April 1958 , approximately six months after construction was completed
and with piezometers still indicating relatively high pore pressures in the
embankment , a reservoir pool of 65 . 9 ft was stored behind the dam and was
drawn down very rapidly with no slope movement occurring , thus helping to
confirm the stability of the upstream slope .
Six closed - type piezometers were planned in the original design and were
installed in the impervious fill during construction . After the slope move
ments began to cause concern , an additional five open - type piezometers were
installed . Pore pressures indicated by the closed - type piezometers ranged
from about 10 ft below to 35 ft above the top of embankment after adjustments
following installation were essentially complete . At completion of the em
bankment , pore pressures ranged from 30 to 65 % and 38 to 55 % of the
pressure of the overlying fill as indicated by the closed - type and the open
type piezometers , respectively . Since completion of the dam , all piezometers
have shown decreasing pore pressures .
The closed - type piezometers within the interior of the embankment have
dropped at a rate of about 1 to 2 % of overburden pressure per month . The
open - type piezometers near the outer slopes have dropped at a slower rate ,
ranging from 0 . 1 to about 1% per month . When the reservoir was filled to
65 . 9 ft in April 1958 , the closed - type piezometers located in the interior of
the embankment showed negligible reaction . The open - type piezometers near
the upstream slope showed a slight temporary increase . Fig . 15 shows
records with time of piezometric pressure as per cent of overburden pressure
and piezometric level with respect to top of fill for the following four typical
closed - type piezometers :
Piezometer Tip
No . Station Offset Elevation
PZ - 2A 12+40 705 . 8
PZ - 2B 12 +
40 75 ft upstream 711 . 3
PZ - 3A 13 +90 708 . 4
P2 - 3B 13+90 75 ft upstream 711 . 6

DEFORMATION AND PIEZOMETER MEASUREMENTS AT OTHER DAMS
Since completion of Otter Brook Dam , programs of slope movement and

piezometer observations have been established for three subsequent dams
currently under construction . Initial results are reported below . It is antici
pated that more conclusive indications will be available before final publi
cation of this paper .
North Hartland Dam , North Hartland , Vermont . - This dam will be 185 ft
high . Its cross section is shown in Fig . 11. The embankment is of homoge
neous , impervious cross section , except for a central pervious chimney and
downstream pervious drainage blankets , located at the base of the embank
ment and at about the third points of the height, and rock - fill and gravel bed
ding slope protection . The upstream slope is 1 on 2 on the upper part and 1
on 2 - 1/ 2 below . The downstream slope is 1 on 2. Properties of the North
Hartland till are compared with the Otter Brook till in some detail elsewhere
in this paper . The design data indicate the till should exhibit less plastic
properties than the Otter Brook till , at its natural borrow pit moisture con
dition . The dam is 42 % higher than the Otter Brook Dam and the downstream
slope is steeper (1 on 2 as compared to 1 on 2 - 1/ 2); however , the upstream
rock - fill section is much larger and internal drainage is provided at inter
mediate elevations for relief of pore pressures in the downstream portion .
Both slope movement and piezometer observations were started in the
initial sea son of fill placement , in 1959 , and will be continued until completion
of the dam , scheduled for December 1960 . As of January 1960 , with the em
bankment at 111 ft, 60 % of full height, upstream slope movements range from
1 to 3 in . horizontally and 1 to 4 in . vertically . These movements are con
sidered as having occurred mostly in the heavy rock - fill section , as down
stream slope movements do not exceed 2 . 5 in . horizontally and 1 in . verti
cally . After initial stabilization adjustments , closed - type piezometers located
some 72 to 78 ft below present top of dam have indicated steadily decreasing
pore pressures , and in January 1960 pressures ranged from 7 to 14 % of over
burden pressure . For the set installed at 20 to 28 ft below present top of
dam , the indicated pore pressures in January 1960 were 11 to 44 % of over
burden pressure . However , these piezometers probably have not been in for
sufficient time to become fully stabilized , as readings continue to be some
what erratic , and it is estimated that actual pore pressures are in the range
of 25 to 45 % of overburden pressure .
Observations on this dam to date do not suggest that difficulties of the
magnitude encountered on Otter Brook Dam will be encountered during its
construction .
Ball Mountain Dam , Jamaica , Vermont . - This dam has a maximum height
of 265 ft . Its cross section is shown in Fig . 12 . It has a central core of
glacial till , with rock - fill shells on both sides . Because of the heavy shells
and the nonplastic nature of the till , no bulging is anticipated . Open - type
piezometers installed in the fill , first during the initial stage of construction
and later when the embankment had reached a height of 123 ft (46 % of full
height ) , have indicated zero or negligible pore pressures .
Thomaston Dam , Thomaston , Connecticut . - This dam will be 142 ft high .
As shown in Fig . 11 , the dam consists of an inclined impervious fill blanket ,
a central random fill section , and heavy rock - fill shells upstream and down
stream , with gravel bedding and filter layers between the rock - fill and earth
fill sections . The upstream slope is 1 on 2. The downstream slope has a
GLACIAL TILLS 305
26 . 5 - ft -wide berm which carries a relocated railroad . Above and below the
berm the downstream slope is 1 on 1. 5 . Because of the heavy confining rock
sections , no bulging of the glacial - till impervious section toward the shoulders
has anticipated . Mea surements on points established on the slopes have
been
indicateda maximum horizontal movement of 1- 1 / 4 in . and negligible verti
cal movement. As of January 1960 , the embankment has been essentially
completed except for the closure section .
As of January 1960 , closed - type piezometers indicate pore pressures in
the impervious fill of 11 to 17 % and zero to 5 % of overburden pressure at
depths of 85 and 55 ft , respectively , in areas with fill essentially at final
height . Undoubtedly the use of a relatively thin impervious section in this
case is very effective in preventing development of large pore pressures .
Pore pressure in the more pervious random fill (also in areas essentially at
final height ) , with tips 75 ft below top of fill , shows pressures from 1 to 2 .6 %
of overburden pressure , corresponding to piezometric levels of 2 to 5 ft .
Hopkinton Dam , Hopkinton , New Hampshire . - During design of Hopkinton
Dam , West Hopkinton , New Hampshire , consideration was given to a program
of slope movement and piezometer observations in that structure , a cross
section of which is shown in Fig . 12 . However , as result of study of the
properties of the Hopkinton till , essentials of which are reported in this
paper , and because the dam is only 76 ft high , it was concluded that , with de
sign adopted which , as shown in Fig . 12 , includes a central pervious drainage
section as a precautionary measure , no unacceptable deformations would oc
cur and such measurements would not be needed .
CLAY MINERAL AND PH ANALYSES
To investigate the possibility that some peculiarity of the mineral or

chemical nature of the Otter Brook till might be responsible for the unusual
behavior of the embankment , clay mineral and pH analyses were obtained on
the material. Since Hopkinton , North Hartland , and Ball Mountain Dams were
then under design , analyses were also obtained of the glacial till materials
planned for these structures to aid in evaluating the possibility of similar
deformation tendencies in these soils . Four specimens of till from the
principal borrow or excavation source for each of the four dams were sent to
Dr . Ralph E . Grim , Urbana , Illinois , for analysis . Specimens of the Otter
Brook and North Hartland tills were also sent to the Waterways Experiment
Station , Vicksburg , Mississippi , for analysis . The nonclay fractions of the
Ball Mountain and Hopkinton tills were examined by the Division Laboratory ,
U. S . Army Engineer Division , New England . The analyses were made by X
ray diffraction analysis , differential thermal analysis , and microscopy . Re
sults presented in the reports submitted by Dr . Grim and the Waterways Ex
periment Station (6 ) , (7) , (8 ) , ( 9) and as obtained by the New England Division
Laboratory are summarized below .
a . Material retained on 200 - mesh sieve . The Otter Brook samples ap
peared to have been derived principally from granitic gneiss and other gneiss
high in mica ; the material retained on the 200 -mesh sieve consisted princi
pally of such rocks and of minerals derived from their breakdown , namely ,
quartz , feldspar , and mica (miscovite and biotite ) .
The fraction of the North Hartland till retained on the 200 -mesh sieve
consisted mainly of fine - grained micaceous phyllites with smaller amounts
of chlorite and muscovite schist , granite , and quartz , and in the finer sizes of
minerals derived from the breakdown of such rocks .
The Ball Mountain till appeared to have been derived from granitic or
gnei ssic rocks and schists high in mica . In the No . 30 to No. 200 sieve range
the most abundant constituent was quartz , with lesser amounts of feldspar ,
mica , and various volcanics .
The Hopkinton till appeared to have been derived mainly from granites ,
gneisses, and schists which comprised about 69 % of the larger sizes . Other
rock types and minerals were quartz and quartzites 17 %, gabbros and diabas
es 10 %, and feldspar , garnet , conglomerate , and miscellaneous 4 %. The ma
terial showed a greater percentage of quartz and less weathering than did the
Ball Mountain material at comparable size ranges .
b . Material immediately below 200 -mesh size . In all four materials the
principal minerals in the size range immediately below the 200 -mesh sieve
were quartz , feldspar , and mica . Quartz and feldspar were the predominant
constituents in the Hopkinton , North Hartland , and Ball Mountain tills , but
mica was predominant in the Otter Brook material and its percentage in
creased with decreasing grain size .
c . Minus 2 micron material . As shown in Table 4, illite was by far the
predominant material in the minus 2 micron size range for all four soils . In
the Otter Brook and Ball Mountain tills it comprised 75 % or more of the ma
terial . However , as shown in Figs . 3 and 4, the minus 2 micron material
comprises a variable fraction of the four soils .
If the percentages by weight of minus 2 micron material, based on com
plete typical gradations to 3 in . , are multiplied by the percentages of illite
shown in Table 4 , the following relative values are obtained :
Relative Percentage
Soil of Illite
Otter Brook till 9 - 9.6
North Hartland till 0. 8 - 1. 1
Ball Mountain till 0. 8 +
Hopkinton till 7. 2 - 8. 4
The second and third most abundant clay minerals in the four soils , as
shown in Table 4, were kaolinite and chlorite , with amounts varying from
0 to 40 %. No significant amount of swelling clay of the montmorillonite group
was reported .
Dr. Grim noted that the clay minerals in the Otter Brook specimens ap
peared somewhat less well ordered and / or of smaller particle size than in
the North Hartland samples , and that this higher degree of dispersion into
smaller particles showed up in the oriented aggregates that were prepared
for the X - ray diffraction analyses . He also reported that the illite in the
Hopkinton and Ball Mountain tills showed only a small amount of such degra
dation effect . He further noted that both this characteristic and the larger
amount of illite in the Otter Brook samples would be in accord with the higher
plastic properties of the Otter Brook till and with the deformation character
istics ; it would also be consistent with leaching and weathering .
Dr. Grim also suggested that differences in deformation characteristics of
the materials might be caused by slight differences in sodium content and
might correlate with differences in pH . He noted that in the glacial tills of
Illinois there are occasionally small areas with relatively higher pH values
-.4
of
.
on
Table New
.
England Tills Results Clay Mineral Analyses Four Soils
Tablol
.
Hahand tu
of
.
Ronaldo ChaMpon Analyse lourSoil
on
moralCar onto Sall HallMountain
T111
1?
.
TIL
Orlar
Inotilo nerdan Merons
Tu
North
Hartland
nu lopetation
Brook
,
,
min Mostabundant
constitueat Mootabandant
consultant Mostabandant
conetocat
%
to
.-
80
to
or
703
75
non
80
60
.
:" , .
of
slighe degne allwount degnding
.ded
Second
mostabundant
to
or
to
20
%
%
10
Nono less
10
25
, %
constituent
20
Second
%
most
abundant
or
,o
Chlorina loss
10
Irace toss
,
156
to
la
constituent Los Suhtly decided
20
.
10 pecsma
2015
-
to
Mod loor Med Layers chloritsand Tnce quiteabundant
la
monteonllonit sample
to
Chloritoandmontorillond
un
ola
.)
MixedLayers and poate udth posedly
t
,. ,.
(.
borullon another
surple
55 Ss
of of1n
vorderlite
?
u
in
in
3
foldeper Inco wpleo Horad promat som
.
specima
la
. .
1
in
1
pobolo
.
Trece uplo Trace pecina
,
Tricednen Trucodinera
GLACIAL TILLS
..
not reported not reported
u
Caleita Motod prosent one
an
.
specimens
.
Quarte
to
Hono Mono wsylittle
%
Vendculito
in
Mone
.
25
specimen
one
of
,
Lwolo
,
theOttorBrook wenselected for claydiner analysisonefromtheborrow
Tu
to
pit depth foot
3
17
22
.
and fromthecompactadembantomnt
of
,,
portion each of the salplog tooted thewatonyakoperirent Suatlon
va
by
.,
.
,
,
Vicksburg
of
.
Missised
ppt andanother portion each RalphGris UrbanaIllinois
moles
1
theNorth Hartland
Tiu vor loctodfromborrow
of
ani dru bolosaplosand

to
thosewasdivided make
of8
1•
Dr 4
of
total specimasSamples representad
by of
to
depths footbelowtheoriginalground
86
marface vor stod
by
is
thewatonayo kuperintSuation and Oria
Dr
ar
,I
Data WallMountain Tiu
for
I. .85
0
malto for specimens

to
fromborrowAnos roprosenting depths footbolo
,. .
.
originalgroundsurface Analysesperformed Oria
Dr
biby
,1
.
Datafor HopkintoaTul resultsfor pocinonsfrom Canal
en
of
of to
representing
depths footbolovoriginal
No
10
.
1 88
. .
..
rounderface Analymas performed Orie
by
,2
MM
. .
deronfructions
wererepented watsodinauation of
by 1 Drby
without
we chordealflocculento
de
)1( ) )(3 )1( ()(5 )6(

to
of
,
Dagradingon theremoval bondingcationsbetween
to
the licato layers tanding make

307
Erind andbonplastic thedooralfine

which have very different physical properties than the surrounding till and
are called “ slick spots . " A good reason for the relatively higher content of
sodium which exists at these spots has never been produced . The New
England Division , therefore , arranged for pH tests on the four tills . Results
are shown in Table 5. As will be seen by comparison with Table 2 , the soils
in Table 5 are arranged in order of decreasing average plasticity from top to
bottom . The most acid soil , the Otter Brook till , is actually the one with
highest plasticity . Thus , it is evident that the plasticity characteristics of
the soils are the opposite of what might be expected from the Illinois experi
ence .
The relatively high pH of the unweathered North Hartland till is believed
attributable to the presence of carbonates in that till . There are limestone
belts in Vermont to the north and west of the North Hartland location from
which such material could be derived . A little calcite was noted in the North
Hartland soil in the size ranges below 200 mesh . None was found in the
sample obtained from the shallow depth of 5 - 6 ft which had a pH of 5 . 0. This
material may have experienced substantial weathering and the calcite may
have been removed from this zone by leaching . The possibility of slight
carbonate influence on the pH values of the Ball Mountain and Hopkinton tills
may be suspected . One Hopkinton till specimen did in fact show dolomite as
a nonclay mineral size component . Differences between the weathered and
unweathered North Hartland till materials are graphically illustrated by dif
fractometer traces for minus 2 micron fractions in Fig . 16 . This figure also
shows a comparable trace for Otter Brook till .
The consistent strongly acid pH of the Otter Brook till may reflect (a )
its
basically different mineral composition the possibility that may have
it
of ,
. at b
(
)
weathering
or
been subject depth
22
least the borrow area

in
to
to
ft
,
a
The possibility of deep weathering

of
of
combination both these factors

a
c
(
)
correlate with the previously mentioned observation

by
Dr Grim
of
would
a
higher degree dispersion .
the Otter Brook till does not
of
However
in
it
,
.
agree with the fact that the material the borrow pit does not show
at
depth
in
downward weathering such as discoloration

of
the usual visual evidence
A
.
fourth possibility concerning which that the ma
no
evidence available
is
is
,
by
terial may have experienced weathering prior its deposition glacial

to
action its present location

in
The possibility that differences cation exchange capacity or type

in
of
in
exchangeable ion could cause important differences the behavior of these

in
tills was considered investigate this point were made

no
However tests
to
,
.
The Waterways Experiment Station expressed the following opinion

on
this
aspect does not appear that the type exchangeable ion likely
of
is
to
It
"
:
produce important effects on the behavior the Otter Brook and North
of
Hartland tills saturating illites with different cations on

of
the effects
]
.
.
.
the Atterberg limits are considerably the changes

of
smaller than the effects

particle sizes
in
."
summary the above analyses thus indicate the following three ways
In
in
,
which the Otter Brook till different from the other three soils investigated
is
all size ranges below the 200 mesh micas and illites are present
In
in
a
,
.
greater amount the Otter Brook till than the Hopkinton North
in
in
Hartland and Ball Mountain tills

,
The clay minerals the Otter Brook material are less well ordered
in
.
b
and or are of smaller particle size than the other soils

in
/
.
GLACIAL TILLS 309
Table 5 . - New England Tills - pH Determinations on Four Soils
SOJL SAMPLE SO!RCE PH

NO.
Otter Brook Mil CF 16 , UBI Dam embankment

CF 14, UBU
CF 10 , UB5
BTT - 1, B5 17 to 22 depth in borrow pit . 5. 0
de
Hopkinton mill FD - 59 , LB- 6 | Excavation Area , 10- 17 ft . depth
FD- 34, LR- 8 32 -40 B
FD- 34, LB- 9 56- 65 *
FD- 50 , LB - 7 76 - 88 m
N. Hartland Till , B1
BT - 1113A 5 to 6 ft . depth in borrow pit
BH- 71, B15 21 .5 to 25 ft . depth in borrow pit 7. 8
BH- 71 , B21 | 30 to 35 ft . depth in borrow pit
BH- 71, B24 38 to 40 # # #
Ball Mountain Till BT - 22 , Bl 1 .5 - 5 . 0 ft . depth in borrow pit

BT - 100, B3 1. 6 - 3.4mm
BTT - 1 , B? 8. 0 - 8. 5 #
BTT - 2 , B6 8. 0 - 8. 5
310
1000
North Hortland Till
Sample No
1
.
15
Depth 20 Feet
-
500 Unweathered 500
)
_
(
ec
/1
300
LIIIII
200
Illite
: = = =
100
-
Chlorite
F Q C
Quartz
Feldspar
Cal Calcite
40
." .=
Cai
40
10
20
15
25
30
40
35
o
5
500 500
-
NorthHartland Till
Sample No
2
300
.
300 Depths Feet

5
6
-
SECOND
/
Weathered
Q
Put
(
I/
LIIIII
200
VI Illite
100 100
:
ML Mixed Layer Clay
= ==
'
-
Chlorite
,
Q C
COUNTS
Quartz
to
Feldspar
40
40
40
35
30
25
20
15
10
2000 0 2000
–
-
Otter Brook Till
Sample No
4
.
Embankment Specimen
INTENSITY
1000 1000
-
500 500
300 300
200 200
100
100
100
Illite
-
Illite
= : =
Į
K C I
t
amination
I" =
Chlorite
Kaolin
30
15
10
25 20
DIFFRACTION ANGLE 20
Fig
16
Diffractometer Traces Minus Micron Fractions

2
,
.
.
-
Waterways Experiment Station

(
)
GLACIAL TILLS 311
c . The Otter Brook till shows a strong acid reaction , whereas the other
tills generally vary from neutral to alkaline .
DISCUSSION
As shown in the diagram at the upper right of Fig . 15 , there was no move
ment at the base of Otter Brook Dam , the foundation being very strong . The
bulging varied over the height of the dam , the maximum bulge occurring in the
vicinity of el 730 - 750 . There was no abrupt differential movement , as would
occur at a shear plane . Thus , the entire mass of the embankment material
was involved . The bulging represented the accumulated movement required
to develop sufficient shear stress to support the embankment under the satu
rated condition which developed as the height and weight of overlying material
increased . The importance of the pore water pressures in the bulging of the
upstream face is indicated by the fact that on the downstream side of the em
bankment, where the internal drainage section permitted greater drainage of
the till , substantially lower deformation was experienced . However , it is be
lieved that the inherent property of the Otter Brook till of requiring relatively
large strains to develop given levels of shear resistance is probably the cause
of the large magnitude which the movements reached .
The bulging of Otter Brook Dam bears some similarity to the partial
failure of the 60 - ft high Clendenning Dam during its construction in 1937 (10 ) .
The upstream face of Clendenning Dam bulged as much as 9 ft horizontally .
The core was constructed of a fat clay . As at Otter Brook Dam , there was no
movement at the toes . However , unlike Otter Brook Dam , rupture in shear
did occur . The top of the dam sagged approximately 6 ft , cracks appeared in
the crown of the dam , and a cave zone developed downstream of the crown .
Thus , a failure condition was actually reached in Clendenning Dam , whereas
in Otter Brook Dam the stresses remained within the available strength of the
material. As at Otter Brook Dam , the embankment was reported to have been
overcompacted .
To aid in ascertaining what correlations may exist between the per
formance of the Otter Brook , Thomaston , Hopkinton , North Hartland , and Ball
Mountain tills and the various physical and strength characteristics of these
soils , various summary data have been brought together in Table 6 . The
following will be observed from inspection of this table :
a. The strain characteristics of the five soils as shown by the triaxial
tests correlate approximately with percentage of minus 0 .002 -mm material ,
plasticity index , and activity . However , the latter factors provide no clue
which would lead one to suspect the exceptional strain characteristics of the
Otter Brook till .
b . The North Hartland and Ball Mountain tills have not shown detrimental
pore pressures or deformation characteristics during construction of North
Hartland and Ball Mountain Dams to date , and none are anticipated . These
soils show strains of 7 % or less at peaks in triaxial tests on specimens com
pacted at optimum , have under 4 % material finer than 0 .002 mm , have plas
ticity indexes of less than 5, and have negligible activity values . (Allowing
for the higher plasticity index of North Hartland till in the embankment , as
compared with the laboratory specimens . ) Thus , tills within this range of
characteristics are likely to perform satisfactorily during construction in
homogeneous section dams up to the height of North Hartland Dam (which is
. –6
.
of
in
on
Table New England Tills Summary Key Information Five Soils Tested Shear
312
GradationData ShearStrength
.)0 at ApproxOptiu
.51(-
.
%
%
at
or
11
Strin Peak Max Por Deloruation Plasticity
in
of
Strun Stress Triaxial Tosta Prossures Inder Activity Relative Deendation UU CD
1
,
Soil
(
to CU
|.,|
Group ForSoil CompactedNear During Dan Mms Mmg 1111ta of
&
(.
.
.
7
)8
Opt ma Migr Construction No Mms Clay
10
40
(.No
s
During Minus Minus
)
(|
|
-
0
.
of Dan Construction natorial material Gradation Minerals
m
200Mash 002
ot
/
/1
oq
degrooo degreest
mot degree
960
Lo
.0
0
9
9
6
.3
.969-.
-
ho
Verybigla Ottar Brook Substantial Noticeable
11
72
21
95
19
9.5
.1
2
2
. / 0
1111 Noticeable
Thomaston Narrowtill sectionnot
to
16
20
to
12
40
87
17
23
33
.
0
0
9
0
|
1
Till susceptible significant tostad testad
11
Not
Mot
11
or
pressuredevelopment
.
deformation
High
Hopkinton Measurenents
not
13
20
.
.
to
50
Tin planned Lovdan
7
30
.0
39
27
8
4 .
|5 |0 38
|
/ |
.| 22-
. 1.
|.0 o
121 Small
762
758 17
to
NorthHartland Moderato Small SZ Lossthan
.
.
)( 2
27
33
38
37
to
3
7
0
0
1
1
0
data
to
8
|.0
0
data
-
1111
.-
34
|
.
.
plastie OtterBrook 11
Non
T111
Noneanticipe
.0
of
Ball Mountain
/-
2
1
2
Negligible tod because Non plastic 84 Slight
W
to
38
|
|.0
2
36
0.
|2 38
|
.
to
ools
Till dato heavyrock
.
shells
is
NOTES Percentage left diagonal figurefor minus inch material
toto
ofof
is
Parcentage right diagonal figurefor dnus testeve fraction ..
-
.
ALI test values shown
on
for NorthHartlandTill vereobtained non plastic material
-3 .
Kabankomentnitorial actuallyplacedaveragesslightly plastic
GLACIAL TILLS 313
185 ft high ) when compacted at field moisture contents within about 1% of opti
mum .
c. Otter Brook till has shown high pore pressures and detrimental slope
bulging during construction . This soil does not reach strength peaks in tri
axial tests below 20 % strain when compacted near optimum , has 12 % minus
0 . 002 - mm material , plasticity index of 11. 3, activity of 0. 72 , and a high per
centage of mica or illite at all sizes below 200 mesh . Any till exhibiting
similar characteristics should be studied very carefully before acceptance
for a homogeneous fill design or similar application , and special design
measures such as internal drains or increased shell section thicknesses may
be required . It is believed that the amount of strain shown by the stress
strain curve in triaxial tests is the most direct indicator of the probable field
performance .
d. The Thomaston and Hopkinton tills are intermediate between the above
mentioned till types . The Thomaston till was used in a narrow impervious
section buried deeply within the Thoma ston Dam , and unacceptable pore
pressure development or deformations could not occur . The Hopkinton Dam
will be only 76 ft high ; measurements of pore pressures and deformations are
not planned because of this low height and because a central pervious drainage
section has been included as a precautionary measure . However , careful
study would be in order if material with the strain characteristics of these
soils were used in a massive dam section of considerable height with rela
tively thin slope covering , particularly if moisture contents a few percentage
points above optimum were anticipated .
It was probably fortunate that the summer of 1957 was one of the driest on
record at Otter Brook Dam , as the bulging would undoubtedly have been even
greater at a few percentage points higher moisture content .
CONCLUSIONS
The following general conclusions may be drawn from this study :

a. Methods of slope stability design which are based on available strength
alone are inadequate when applied to situations such as outlined herein for
Otter Brook Dam . Research is needed to develop methods of slope design
which may be used in situations where deformation rather than strength is the
controlling factor .
b . Glacial tills in New England may vary substantially in characteristics ,
from very strong , stable , nonplastic materials which produce peak strengths
at very low strains , to soils in which the properties of the fine fractions are
predominant in controlling stress and strain characteristics and which may
show the performance characteristics of clays .
c. When a glacial till produces stress - strain curves in unconsolidated
undrained or consolidated - undrained triaxial shear tests which show peaks at
very high strains or which continue to show increasing stress at 20 % strain ,
special care in analysis should be exercised if the till is to be used in the
outer section of an embankment or in a homogeneous embankment where sub
stantial movement would be detrimental . Tills containing substantial per
centages of mica and illite may be more likely to show these stress - strain
characteristics .
d . Thin internal drainage sections, such as used on Otter Brook Dam , are
effective in reducing deformations of cohesive , homogeneous type em
bankments during construction .
ACKNOWLEDGMENTS
During the bulging of Otter Brook Dam , Dr . Arthur Casagrande of Harvard

University , and Mr. George Bertram of the Office , Chief of Engineers , De
partment of the Army , gave invaluable suggestions, advice , and assistance to
the U . S . Army Engineer Division , New England . The valuable contributions
of Mr. Thaddeus C . Johnson , formerly of the U . S . Army Engineer Division ,
New England , who was in immediate charge of the various soil tests , analy
ses , and studies on the Otter Brook Project , are also acknowledged .
APPENDIX . -- REFERENCES
1. Fenneman , N . M . , Physiography of Eastern United States . McGraw - Hill ,

New York , N . Y. , 1938 , pp . 381 , 385 - 386 .
2. Terzaghi , K ., and Peck , R . B ., Soil Mechanics in Engineering Practice .

John Wiley & Sons , Inc ., New York , N . Y . , 1948 , p . 7 .
3. Goldthwaite , L . , Glacial Till in New Hampshire . New Hampshire State

Planning and Development Commission , 1948 .
4. Flint, R . F ., Glacial and Pleistocene Geology . John Wiley & Sons , Inc .,
New York , N . Y . , 1957 , pp . 109 , 111 - 112 .
5. Casagrande , A., “ Notes on design of the liquid limit device ." Geotech
nique , vol. VIII , No . 2 (June 1958 ).
6. U. S . Army Engineer Waterways Experiment Station , Composition of
Samples of Glacial Till from Otter Brook Dam , New Hampshire, and
North Hartland Dam Site , Vermont . Miscellaneous Paper No . 6 - 262 ,
Vicksburg , Miss . , March 1958 .
7 . Grim , R . E ., Report Mineral Analysis of Samples Submitted by

of Clay
the Corps of Engineers , U . S . Army , New England Division . Contract
report to U . S. Army Engineer Division , New England , Waltham , Mass . ,
12 February 1958 .
Report of Analyses of Samples from Ball Mountain Dam Site ,
Vermont . Contract report to U . S . Army Engineer Division , New England ,
Waltham , Mass . , 26 July 1958 .
9. , Report of Analyses of Samples from West Hopkinton Dam
Site , New Hampshire . Contract report to U . S. Army Engineer Division ,
New England , Waltham , Mass . , 26 July 1958 .
. Creager , W . P. , Justin , J. D . Hinds , J., Engineering

for
10 and Dams John

.
.pp
Wiley Sons Inc New York 1945 772 775

&
.,
.,
,
N
-
.
.
STRENGTH PROPERTIES OF CLAYS DERIVED FROM VOLCANIC ROCKS
By Robert J. Pope and M . W. Anderson ,2 F . ASCE
SYNOPSIS
The soils in the Willamette River Basin are derived from volcanic rocks
and are extremely heterogeneous . This paper contains a description of the
formation and constituents of these soils , followed by a discussion of labora
tory and field methods for estimating shear strengths to be used in the sta
bility analysis of slopes in undisturbed soils , of compacted earth dams , and
of compacted cores in rock - fill dams . Research is needed in order to develop
more reliable methods of testing and more economical designs of structures
in and of such soils .
INTRODUCTION
Soils of the Pacific Northwest Cascade Mountain Range are almost entirely
derived from volcanic rock and are usually extremely heterogeneous . Prob
lems encountered in estimating the shear strength properties of these ex
tremely heterogeneous soils are discussed in this paper . A reasonably accu
rate determination of the in - situ shear strength properties cannot always be
obtained by laboratory tests . In these instances a more reliable value of shear
strength is obtained by analysis of unstable slopes. The test data and analyses
presented in this paper have been selected from problems encountered during
design and construction of several earth dams and highway and railroad re
locations for the Willamette River Basin Reservoir System , shown in Fig . 1,
and are therefore not of the completeness or accuracy of research work .
ORIGIN AND GEOLOGICAL ASPECTS OF SOILS

OF THE CASCADE RANGE
The rocks composing a major section of the Cascade Range , and all that
part lying in Oregon , are volcanic in origin . Geologically they are very young ,
only five to thirty million years old . Coming from within the earth and orig
inating in an extremely hot and almost moistureless envoironment , they have
since been exposed on the earth ' s surface to moist and relatively cool condi
tions that are incompatible with their mineral composition . The lavas , mostly
andesites and basalts , were abundant in minerals that are low on the weather
ing stability list . They were either filled with bubbles or emerged as com
plete froth , or were crisscrossed with cooling joints . More important , the
1. Chf . , Soils Engrg . Sect., U . S. Army Engr . Dist . , Portland , Oreg .

2. Soils Engr . , U . S. Army Engr . Div . , North Pacific , Portland , Oreg .
3. S. S . Goldich , “ A study in rock weathering , ” Jour . of Geol . , vol 46 ( 1938 ) ,
pp 17 - 58 .
315
316
.
.
CZ00 12230 OSOWAR
)12
30
z
.
10
Pe
Рал
RIVER
WILLAMETTE
.
OC 450 views
REGIONAL
MAP
SCALEMILES
IN
.
121
30
.
122
00
RIVE
LINCOLN
5.
.
00
00
SAUSLAW
NAL
FOREST
11
DIAN WAARON
ESERVATION
WASCO
ETTERSON
CO
CO
BIG CLIFF
.
.
REREGRES
MARION
DETROIT LINNA
RES
ONVOGE
RUS GREENPETER
.
PAM RES
AMAN
-
OKNOTER
TT
HOLLEY
.
DEN IRES
.4.
LANE
-
LUME
BLUERIVER
.
RESMA
FERN RIVEN
RIDGE COUGAR
RES RES
Sicu
EUGENE .
4400
NATION
FALECREEK
NATIONA
DEXTER RES C
.
REREGRESI
10окрит o
,
POINTRES
A
.
s
•
Co
Scroll
.
123 DORENA
con
cauere
30
RES s
CONV17M
4
DESCHUTES
COTTAGE
QUA
GROKE
VOLCANIC
m
RES
.
121
WILLAMETTE
30
arice
sre
.
DE
SCHUTES
CO
Shell
Mosby
LEGEND
ROCKS
.
4330 EXISTING
JALOUS
430 RESCAVOIRS
FORES
>
UNDERCONSTRUCTION
12
Too
DOUGLAS AUTHORIZED
RESERVOIRS
CO
MO
OU
TAINS
SALE MILCS
IN
.U
S
,
.
ARMYENGINEER
DISTRICT
PORTLAND
*
123
00
122 CORPS ENGINEERS
30
OF
..1
Fig Wilamette River Basin Reservoir System
317
mass of the range is composed of about 60 percent consolidated beds of vol

canic fragments and ash (tuffs ) , most of which originally were high in volcanic
glass of intermediate to basic composition , which is relatively unstable . Vol
canic rocks of these types are easily attacked by chemical reations de pending
upon moisture and warmth , and the fossil record shows that during their ex
istence both moisture and temperature were favorable for these reactions .
Because of their physical character they readily absorb all of the local pre
cipitation of their times , and from the beginning suffered considerable weath
ering and decomposition . Only those volcanic rocks high in silica , or those
with dense , even grain , were able to withstand these attacks .
Since their origin these rocks have been subjected to at least one period of
intense strain and faulting with the creation of numerous large open joints , at
least two periods of weathering , and one period of hydrothermal alteration ,
all of which have greatly weakened them . During the periods of alteration and
weathering a large percentage of the original rock material was converted to
clay; in the lavas this deterioration progressed inward from the joints . Sub
sequent rapid stream downcutting upset the equilibrium of these weakened
rocks , and during the excessively wet glacial ages they rave led and slid in
great volume onto the terraces and valley floor . In some cases as much as
a hundred thousand cubic yards of weathered lava and tuff slid down en masse
and came to rest virtually intact upon old “ Pleasant Hill" type intense ly
weathered terrace grave ls . Much of the clay came from the breakdown of
more recent volcanic ash , particularly in the case of the older gravels . Tests
show that the clay fraction of these gravels is composed mainly of the clay
mineral halloysite which has a tubular structure . Some of this clay , a long
with large quantities of local residual clay , washed down into the open joints
previously created .
The result of these geologic activities in the Cascade Range has been the
creation of a set of physical conditions almost unique in their adverse effects
on construction work . Not only are large areas of the original rock material
in a weak condition , but the sequence of erosional events has also left behind
great sloping deposits of very heterogeneous debris , much of it still moving
as lands lides or in a condition to do so upon the slightest provocation .
Construction of the Willamette River Basin Reservoir System required
evaluation of the shear strength properties of these complex soils . Railroad
and highway relocations crossed miles of old slide areas. Numerous cut slope
failures occurred during construction , and two major slides have deve loped
following operation of the Lookout Point Reservoir . Construction experience
established that the slip planes on which the majority of the slides moved were
ancient in origin ; other slope failures in residual soils have been due to the
influence of highly plastic clay which was deposited in fractures as the rock
weathered and now forms a network of thin seams at random inclinations and
angles .
LABORATORY EVALUATION OF UNDISTURBED SHEAR STRENGTH
Shear strength evaluation of soils derived from volcanic rocks for the un
usual case where weathering processes and reworking during geological his
tory have produced a reasonably homogeneous , saturated soil does not impose
unique problems in laboratory technique or test data interpretation . Consis
tent test results are also obtained from samples of residual soils which do
VOLCANIC ROCKS 319
not contain rock fragments , old faults seams, or slickensided surfaces . Typi
cal triaxial test data for such a soil are shown in Fig . 2 . These materials
generally have a relatively high angle of internal friction as determined by
drained , direct shear tests . Since undisturbed samples of these materials are
seldom saturated , an angle of internal friction is also frequently indicated by
unconsolidated undrained shear tests . The degree of saturation is usually
within a range of 85 to 95 percent . Although drainage is not permitted during
unconsolidated undrained tests , volume changes occur due to compression of
air and gas on application of the chamber pressure and during shear strain .
The shear strength is therefore not independent of the applied pressure , and
and indicated angle of internal friction is obtained
Shear strength evaluation of residual soils which are extremely hetero
geneous and other soils containing various percentages of rock fragments ,
hard remnants of residual matter , slickensides or thin clay seams of random
orientation is difficult by normal laboratory testing procedures . Typical scat
tering of shear test data for a heterogeneous soil is illustrated in Fig . 3. In
many cases it is impossible to duplicate with reasonable agreement , test data
obtained on specimens cut from a one -cubic - foot sample . Triaxial shear ap
paratus of 6 - inch - and 12 - inch - diameter size were constructed in an effort
to improve results obtainable on soils containing rock fragments and slicken
sided surfaces . Larger samples permitted inclusion of 2 - to 4 - inch rock
fragments without excessive influence or disturbance during trimming opera
tions . Test results were more uniform , but the cost and difficulty of obtain
ing large samples from test pits and handling samples of this size are a dis
advantage .
Three - stage triaxial shear tests , suggested by Professor Donald W . Taylor ,
during design studies on Lookout Point Dam were used in an effort to improve
the reliability of shear data on soils in which inconsistent results are obtain
ed due to material differences between test specimens . These tests were per
formed on both 2 . 8 - inch - and 6 - inch - diameter samples . A strength envelope
is obtained by this method with one test specimen indicating the gain in
strength due to compression of air or gas . The first - stage procedure is the
same as a standard unconsolidated undrained test performed at chamber pres
sures usually approximating the overburden load . However , the strain is car
ried no further during the first stage than necessary to obtain the maximum
deviator stress . The chamber pressure is then increased to the value desired
for the second stage and strain resumed until a second peak in deviator stress
is obtained . The third stage is a check on the first stage by reducing the
chamber pressure to that first used and applying additional strain . Typical
test results are shown in Fig . 4 . Testing experience is necessary in order to
change stages at the proper strain conditions. Three - stage tests are not re
commended when samples suitable for conventional tests can be obtained , as
the results are largely dependent on the skill of the laboratory technician and
the influence of the first - stage stress conditions on the second stage is ques
tionable .
EVALUATION OF THE MASS SHEAR STRENGTH
Reliable laboratory shear test results are often insufficient for evaluation
mass shear strength of residual soils derived from volcanic rock . The
of the
mass shear strength can be evaluated with reasonable accuracy from analysis
ft
/
sq ft
, /
MAJORPRINCIPALSTRESStors
sq
SHEARSTRESS tons
1q
APPLIEDPRINCIPALSTRESSESIons
ft
,
/
TEST NO
/ , .
Watercontentwo 423
ft %
INITIAL
,
Dry density Ibs
cu
78 180
8
.
.
Voidratio
e
193 115
, ,
1
_
.
.
%
SaturationSo.
27
TEST
W
%
ofter saturation
ws.
C
,
.
.
BEFORE
Saturation %
S
,
,
.,
/ / , % sq
Consol pressuretons
ft
,
% , sol /
W
afterconsolwc.
STRAIN
.C
.,
.
Voidratioaftercon e63 11
FAILURE
186 103
11 1420 1 1
.
. .3 . .55 .
.
.
Majorprinstres tonssqft
3 5060
prin
VOLUMETRIC
0 0
, ,
%, s
.
sq
50
Minor stresstons
2
ft
.
. . . 401
.
Watercontent wi
er ,
,
AT Voidratio
100093 .
,
180 098
4 1 1
Specimen diameterinches
,
Initialheightinches
1298
,
10
30
.
20 Testtime failure min

to
.
%
AXIALSTRAIN
,
INCHESTNO
OFMESHPERINCH STOGRAIN
N
SIZEI MILLIMETERSSTYPE
OF TEST
.V
S
,
(
TYPE OF SPECIMEN
.
Constant 63 Undisturbed
-
ton
+
Strain controlled 024

C0 =
-
BY WEIGHT
sq
PERCENTFINERBY WEIGHT
Un consolidated
Un drained tons/
+
ft
,
5
|
REMARKS Description Gray still clay

,
,
:
weathered tuff
.
COARSER
PERCENT
PROJECT LOOKOUTPOINT RAILROAD

RELOCATION
SAMPLE DATAMinnow Station Slide
822 008 E1
N
409 791
_
- BY
SUBMITTED Portland District

EXPL NO 224 SAMPLENO.
D
17
( 1 /O GR
GRAVEL SAND FINES

32
33
ELEV 944 DEPTH

- 2
cl
.0
5
.
S .
SILT
W
Classification CH MH NO. 55 305

(
-
.
-
74
of 12
TEST 56
LL 61
JC
Drown
G
2
.
.
-
-
26
PI Do checked
GHD Date Roport ChiolSoilMechanics
Branch
L
.
AXIAL
NPD FORM 335 TRI AXIALCOMRESSION TEST REPORT

For
TRI
35
-
o
:
.
REPORT
-
3
APRIL 50 COMPRESSION
CORPSOF ENGINEERS
NORTHPACIFICDIVISIONTESTINGLABORATORY
Fig Test Data Homogeneous Soil

,
2
.
.
VOLCANIC ROCKS 321
15
/ ft
, /
MAJOR PRINCIPALSTRESS tors
sq
SHEARSTRESS tons
, 7
ft 9
sq
213
APPLIEDPRINCIPALSTRESSEStons/
TEST NO
20
21
Wotercontentwo
19
%
98 24
INITIAL
,
,
.3 8
98 . .6
1
86 . 6
.
92 6 9
.
% Ibs
.
.
Dry density
. .
cuff
51
106 106
., e ,
105 /
8
|
Voidratio
, ,
615 607 628 742
_ 1
.
__ .
Saturation
.
S
93
_
TEST
_
_
_
W
after saturation ws. %

C
.
,
.
FAILURE BEFORE
Saturation
%
,
S
,
, ,
.
/s / XX % sq
Consolpressure fons/
ft
%
W
ofterconsolwc
STRAIN
C
.
.,
.
ft 63
Voidratio aftergourel 583 5991594 714

,
.
3 5 .
3 8 .
sq
Majorprinstresstons
VOLUMETRIC
8 0055
00 . 0042
00 .6 5099
prin
36
3 o
, ,
1 6
1 8
20. .
21. .
.
. . 19. .
24. .
Minor stresstons qft
00
% ,
Watercontent wr
61
et ,
.
9
.
.
avoid ratio 6031
,
5971 601 715

.
4 2 .
Specimen
4 2 .
diameterinches
00
00
,
6 . 2
2
.
. .
Initialheightinches
. .
01
10
,
041 071
10
3 4
20
.
.
to
Testtime failure
16
,
/1
2
6
%
AXIALSTRAIN
,
OF
MESHPERINCH STOGRAIN
O
SIZEI MILLIMETERSTYPE OF TEST

V
.S
,
.
TYPE OF SPECIMEN
Constant63 Undisturbed
Strain controlled tan
PEACENTFINERBY WEIGHT
-
BY WEIGHT
Un consolidated
Un drained tons/sq
ft
,
REMARKS
Tan moist firm weathered tuff
,
,
COARSER
on on
23 22
Tests Sample
& &
3 1
4 2
Tests Sample
PERCENT
PROJECT HILLS CREEK DAM
SAMPLE DATA
54 DHBY

22
EXPL NO
. 23
238 SAMPLENO
&
-
.
/ GR
.0 .
GRAVEL SAND FINES

51
ELEK 1517 DEPTH

- 7
55
.
5 .
.
5
NO
Classification
W
ML CL
O
57
(
.
-
LL
75
14
4501 Drown
GHD
G
May 1954
2
.
PL 15 Dio" Checked Date Report

of
ChiefSert
Mechanics
Branch
(
NPD FORM TRI AXIAL COMRESSION TEST REPORT

335
-
APRIL 50 CORPSOFENGINEERS
Fig Test Data Heterogeneous Soil

,
.3
.
ft 3.0 1
/
sq ft
, /
MAJOR PRINCIPALSTRESS tors
sq
SHEARSTRESS fons
APPLIEDPRINCIPALSTRESSEStons 1911
/
hase KKST. NO.
2
1
Watercontentwo 17
ft %
INITIAL
4 3
,
. .
cu
Dry density Ibs 110
/
Voidratio 554
0
., e
, ,
.
.
%
Saturation
S
TEST
W
%
ofter saturationws.
.C
,
.
.
, .%
SaturationSs
FAILURE BEFORE
% sq
, Consol pressuretons/
ft
%
.
wc
W
afterconsol CON
.C
STRAIN
,
,
.
10
Voidratioofter SOCIA60 543 534 5471
0 171 5 0
. . . . .
3 10 .
. . .
Final
00 53
80
Majorprinstresstons qft 43
prin
1 5
)
(
VOLUMETRIC
,
.5, o,
. .
/ /s
00
00
Minor stresstonssqft
%,
Watercontent we
er ,
3
AT Voidratio
10
01544 531 564 Final
,
)
|
.
1215
1190
Specimen diometerinches
,
. .
Initialheightinches
,
118
16
10
to
Testtime faites. Min
|
.
%
AXIALSTRAIN
,
INCHED
NOOFMESHPERINCH
OF
TQGRAIN
SIZEI MILLIMETERS
TYPE OF TEST
U
S
TYPE SPECIMEN
.
.
.
.
100 Carstedt Phase Indisturbed

- 3
-
ton
Strain controlled
sq23
43
=
1 0
, .
Un
15
PERCENTFINERBY WEIGHT
consolidated
Un drained tons/
WEIGH
ft
,
REMARKS Description
:
Light brown speckled tan moist Pirm

COARSER $
,
| ||
BY gritty weathered tuff

.
PERCENT
PROJECT HILLS CREEK DAM
SAMPLE DATA Left Abutment Foundation

97
98
20tota
200 590
N
E
;
,
,
-
SUBMITTEDBY Portland District

EXPL NO. 41 SAMPLENO.
D
23
-
/O GR
GRAVEL SAND FINES

23
18
19
ELEV 1598 DEPTH

- 1
.6
.
5 .
-
.
52
CL
Sa
Classification CLAY NO 329

.
-
(
)
.
75
LL Drown
JC
19
G
2
.
PI 24 Do Date Reporty SoilMechanics

Branch
of
Checked
cun
L
NPD FORM 335 TRI AXIALCOMRESSION TEST REPORT

,
APRIL 50 CORPSOFENGINEERS
Fig Test Data Phase Triaxial Test

,
4
3
.
.
VOLCANIC ROCKS 323
of active slides by determining the average shear stress of failure along the
slide plane . The average shear strength has been estimated both by analysis
of slides inadvertently caused by cutting overly steep slopes and by analysis
of test slopes de liberately cut to create slide conditions . Analysis of natural
slopes, presently stable but which show evidence of slide activity within re
cent years indicating they are barely stable , has also provided useful data .
There are two approximations involved in these analyses . Analysis by slices
is approximate due to neglecting forces acting between slices ; and A . W .
Skempton4 has shown that the $ = 0 analysis applied to an actual measured
shear surface will at least theoretically lead to incorrect estimates of the
forces and stability . The first approximation leads to a high estimate of the
shear strength , and the magnitude of the second cannot often be determined .
Even so , estimating the shear strength by analysis of slides has been found
more dependable for extremely heterogeneous soils than reliance entirely on
laboratory tests . The following cases are examples where the shear strength
could not always be estimated with reasonable accuracy by laboratory shear
tests .
Minnow Creek Cut . - A cut slope failure east of Minnow Creek on the high
way relocation for Lookout Point Reservoir was typical of failures which
develop in extremely heterogeneous materials . The natural ground slope , as
shown in profile in Fig . 5 , was about 10 degrees and construction plans re
quired a cut of about 40 - foot depth at highway center line . The overburden
soils , derived from tuffs , generally classified as ML and MH , but the plasti
city varied over wide limits as shown on the plasticity chart in Fig . 16 . The
water content was variable but often as high as the liquid limit , and the sensi
tivity was in the order of 4 to 8 . Slickensides were noted in a test pit at
depths of 30 to 40 feet . Consolidated drained laboratory tests indicated an
angle of internal friction of about 27 degrees with a cohesion of 0 . 2 ton per
square foot. Unconsolidated undrained tests gave a strength ranging general
ly from 0 . 5 to 1.5 tons per square foot , although several tests indicated
strengths as low as 0.25 ton per square foot as shown in Fig . 6. The cut was
designed for a l - on - 2 slope in expectation that the average mass strength
would be appreciably higher than indicated by the few low test va lues .
Failure cracks deve loped beyond the top of the cut slope before excavation
was completed . A system of horizontal drains was installed at the toe of the
cut ; and although considerable seepage was intercepted , slide movement con
tinued until a slope of approximate ly 1 on 4 was attained . Examination of the
cut face revealed an irregular pattern of old fault seams in the residual soil
which were filled with a soft CH clay . Additional tests performed on samples
obtained from the weaker materials indicated the angle of internal friction
obtained by drained shear to be as low as 12 degrees .
The wide scattering of shear strength values under conditions encountered
in this cut makes evaluation of the mass shear strength from any reasonable
amount of exploration and laboratory testing very doubtful. Since the mass
strength is determined by orientation of old fault planes and uneven weather
ing of the residual soil , the evaluation of shear data is difficult even if a large
amount of reliable test data is available . Averaging of strength data is use
less and design on the basis of the minimum strength in all cases is uneco
4 A . W . Skempton , “ The 0 = 0 analysis of stability and its theoretical basis ,

Proceedings of the Second International Conference on Soil Mechanics and
Foundation Engineering , vol 1 (1948 ), p 72 .
Original ground line
and
Ground line after ly clou - - - - Cloy Talus
I on 2.5 slope failure Turn Breccia
5
Revised slopen land
& :
Basolt Bosolt
Hwy
-
-
-
-
-
- R
-
- .
- R.
Ton
Sirt - -
4
Tolu
--
-
t
--
-
-
--
and Tuff Breccia
onlis Qy Seoms
with Clay
Se Clay
-
-
-
-
-
-
-
-
7
IN
SCALE FELT
10100
-
Fig
Profile Minnow Creek Cut

5
.
.
Range
/1-
Cong Test
D
S
.
.
.
642
SHEAR STRESS
Range
UU Triax Tests
.
Design Strength
-
APPLIED STRESS f42

T
/
-
Fig Range Shear Strength Creek Cut

of
Minnow
,
6
.
.
Hwy
.
slide orc 670 RR

of
Rodius
3,
Gr
'
1000L Orig Line

.
.
Proboble Slide Surface

Old Hwy Normal Pool 925 tu
El
.
L
.
900 Swee. Groves

stiff clay
14 naciona
So. Gr Boulders
Cloy
. and
Sa
-
cu Stirr Intact Cloy Weo. and Gr.

Gr
Sa
- eo
Te
V
-
.
.
-
.
.
-
w
.
.
.-
Very stiff Intact Bedrock Decomp Andesite

-
300
100
800
700 600 500 400 300 200 100 200
70
STA 15 00
+
PROFILE AT
R
R
.
.
.
piezometer pipes and

Slide surface determined by depth
of
in
kinks
exposure slip plane hand dug test pit
in
of
Fig Voss Slide

of
Profile
7
.
.
VOLCANIC ROCKS 325
nomical . In this case the cut slope was revised as shown in Fig . 5 . Fortu
nately , higher strength materials were encountered approximately 300 feet
from the centerline of the highway , which controlled the extent of sliding .
Voss slide . - The difficulties experienced with cut slope failures along the
Lookout Point Reservoir were corrected during construction without inter
ference with the operating railroad or highway . However , two slides occurred
on the reservoir slopes after completion of the work , due primarily to draw
down of the reservoir . The re located alignments crossed these slide areas ,
and daily maintenance was required during drawdown seasons until the slides
were corrected . A profile of the Voss slide area is shown in Fig . 7. The
slide mass was about 400 feet in width at the railroad grade and widened to
600 feet at its maximum width near the toe . The toe of the slide mass was
approximately 75 feet be low full pool elevation . The re located railroad and
highway were completed and placed in use in 1953 . The reservoir was filled
for the first time in the spring of 1955 and drawn down to elevation 825 that
fall . Slight abrupt movement of the railroad grade was observed during the
early stages of drawdown during 1955 and 1956 . During the 1957 drawdown ,
slide movement continued for the full drawdown cycle and amounted to about
2. 5 feet . The slide accelerated rapidly during the 1958 drawdown until it was
moving at a rate of about 3 inches a day . Total movement of the railroad
roadbed amounted to about 24 feet horizontally and 4 feet vertically during the
1958 drawdown season . The head of the slide extended into the highway fill ,
requiring a short detour around it . Railroad traffic was maintained under
slow orders but almost daily realignment of the tracks was required .
As indicated on the profile , the average natural ground slope poolward of
the railroad grade was about 10 percent. The slip plane was located in bor
ings and test pits at a depth of about 45 feet at the railroad and s loped toward
the reservoir with an overthrust at the toe without evidence of an upward
component . The overburden above the plane is a stiff clay with the relic
structure of weathered gravel . Two other inactive fault zones were found in
a test pit below the active zone of movement. Charcoal and slickensided zones
were also found in borings below the active slide plane . Thin strata of com
plete ly weathered sand and gravel and very stiff intact clays are found be
neath the slide plane to a depth of about 90 feet where partly weathered sandy ,
clayey gravel is encountered . A rising rock line and stable terrace gravels
prevented extension of the slide uphill beyond the limits first noted by crack
ing of the highway pavement .
The slide plane , which was quite accurately located by explorations , ap
proximated an arc with a 3700 - foot radius . An analysis for determination of
the average shear stress along this arc indicated 0 . 15 ton per square foot with
the pool full during which sliding did not occur and 0. 2 ton per square foot
with the pool at minimum flood control elevation 825 during which continuous
sliding occurred . The shear stress at the time movement started must have
been approximately within these limits , and the shear stress under which
movement continued is considered close to the lower value of 0. 15 ton per
square foot . Unconfined compression test data shown in Fig . 9 would indi
cate the shear strength at the depth of sliding to be about 0. 85 ton per square
foot. Considerably lower shear strength exists along the slide plane than was
obtained in any of the samples tested in the laboratory . The lower strength
along the slide plane is believed due mainly to the existence of an old slide
plane remaining from past slide activity in this area . This belief was sup
ported by the observation of both the active slide plane and two additional in
active faults at lower elevations in a test pit . The active slide plane was fill
ed with soft fat clay gouge and the thin inactive faults showed slickensided
surfaces . The most economical solution to this slide problem was to relocate
the railroad and highway on stable ground .
Minnow Station Slide . - This slide area was also located in the reservoir
area of the Lookout Point Project and involved the roadbeds of both the re
located highway and railroad . A general profile of the slide is shown in Fig .
10 . The slide width was approximately 500 feet along the railroad alignment
and average ground slope in the reservoir area was 6. 5 percent . Slide move
ment occurred during the first pool drawdown . Movement accelerated when
the pool was lowered below elevation 870 . Control survey hubs were estab
lished on several lines across the slide mass to determine the rate and di
rection of movement . Measurements were continued as the pool was raised
and lowered for flood control purposes . Explorations were accomplished with
churn drills equipped to obtain 6- inch - diameter drive samples . Continuous
sampling procedure was followed whenever possible . During inspection of
INDEX
A -LINE
PLASTICITY
OH
MH
M7 MLIOL
-
100
20
40 60 80
LIQUID LIMIT
Fig Plasticity Chart Voss Slide

,
8
.
.
/
ft2
7
on
samples from Drill Holes

of
UC tests plotted
depths os
.
115
lbs
STRENGTH
Unit weight fmo

=
.
/
. 276
o
.
COMP
0273 °272 273

togethe1
0272 276 027
il
.
l
272273
0
.8
4273 273
_
0272 0272
UNCONF
0272 Pressure of assumed

96
depth
of
slide
!
3
OVERBURDEN PRESSURE
64
OVERBURDEŇ PRESSURE
/
1
Fig Unconfined Compression Test Data Voss Slide

,
.9
.
VOLCANIC ROCKS 327
samples at the laboratory , the least disturbed specimens of the various ma

terial groups were selected for shear strength tests . The remaining samples
were inspected to classify and describe the material and to check for slide
planes and slickensides . Overburden material is highly plastic , medium stiff
to stiff clay containing variable amounts of rock fragments . Atterberg limits
test results are shown in the plasticity chart , Fig . 11 .
The rock fragments varied from fresh to completely decomposed , and the
degree of weathering was not consistent with depth . Explorations encountered
organic material, indicating the area to be an old slide mass . Inspection of
samples did not reveal a definite slide plane but zones of slickensided mate
rial were observed . Deflection of the casing in exploration holes indicated
movement at a depth of approximately 60 feet near the railroad roadbed . Re
sults of laboratory tests performed on undisturbed samples are shown in Fig .
12 . Since materials were relatively homogeneous in the movement zone and
satisfactory samples were obtained , test results did not give scattered data
as had been obtained at other slide locations . For comparative purposes ,
analyses were made of the total slide mass to determine the average shear
strength required along the failure plane for a safety factor equal to one , con
sidering several pool elevations . The section analyzed and a sample compu
tation are shown in Fig . 10 . This analysis indicated that the average shear
strength required along the slide plane ranged from 0. 26 to 0. 41 ton per
square foot , depending upon the pool elevation . The plot of shear strength
versus depth shown in Fig . 12 indicates the shear strength at the 60 - foot
depth of movement should be about 0. 7 ton per square foot . This difference
between computed and test values suggested the possibility of a progressive
failure caused by loss of toe support . A sample analysis assuming loss of toe
support is also shown in Fig . 10 , and indicated strength requirements for this
assumption are in close agreement with laboratory test data . Development of
tension cracks in the slide mass supported the explanation of progressive
failure . Slide corrective measures consisted of reducing the driving force by
excavation and providing a stable highway roadbed .
Hills Creek Dam Left Abutment Foundation .The left abutment foundation
for Hills Creek Dam is predominantly a deeply weathered series of lapilli
tuffs . The overburden material derived from the lapilli tuffs is a reddish
brown sandy clay which is relatively compact and still has much of the struc
ture of the parent rock . This material has medium to high plasticity with
liquid limits ranging from 36 to 73 and plastic limits ranging from 20 to 39 .
Depth of residual soil varied from 20 to 60 feet but soft weathered tuff , only
slightly stronger than the overlying material, extended an additional 50 to 100
feet in depth . Core borings indicated that both the weathered and fresh foun
dation rock have been extensively sheared and fractured to a considerable
depth . The rock structure is relatively open as a result of this fracturing ,
and soft secondary clay which has been deposited by percolating ground water
is found as a partial filling in the fractures of both weathered and unweathered
rock at all depths . Except for a shallow surface layer , the residual soils have
retained remnants of the original rock joints and fractures .
The permeability of the upper 20 to 60 feet of residual soil in its undisturb
ed state varied from 10 - 3 to 10 - 7 feet per minute , but when remolded the
permeability decreases to a value in the order of 10 - 7 to 10 - 8 feet per minute .
Permeability of the weathered tuff and sound bedrock is controlled by the ex
tent to which the joints and fractures in the rock have been filled by deposi
tion of secondary clay . The ground -water table is 100 to 200 feet below
400 800 1200 1600 2000
328
O
'
.
Hwy
.R.
R
238
.-
R
IZ
1000 GA
El
Normal Pool . 925
--
-
El
.
.
Min
-
)€
pool 825
.i
Sin 1125
800
Assumed Failure Plone
.A
.B
Average Shear Strength Determination Average Shear Strength Determination Assuming
Pa
of
Computation Loss Toe Support
ft
no
241
El
Arc Length Assume Minimum Pool 825 and shear resistance
40
sq
.f.t
.0
=
C
Assumed Strength downslope from Point
No
T
ft
PA
ET Pool 266 width 186 width
of
En
.1.2
. /T .f/Tt.f/Tt
ET Pool 925 250 Weight
W failure wedge
El
/ /T
::
ft
Assume Pool 825 2920

rt
./
):
MO Tangential Component 2920 1125 328

ft
T.
.():+
(::
T /
:(
40
()
218
()
)
.(-
266
238
241
1860 328 514
238
‫مان‬ Total Driving Force
PAPA
)..0(-):
: == ).ft
- ) ()T
/
186 Average Strength Required
of
Computation Driving Force

sq
/T
.f.t
. . . .0: 60
(( () .
.
Wt of Failure Wedge 4230 Aver

Pt
=
./
T
f/t.
Tangential Component 4230 1125 476
:)T
.i/t
::
Total Driving Force 186 476 662
.1 .2 .3 .4
Average Required Shear Strength Along Slide
00
.7: .
T) 1
( :
Plane for Safety Factor
Tottal Driving Force
of
Aver Length Slide Plane
a- :
662
2
.
41
r/t
T
.0=
. .
Aver
10
.
.
Fig Average Strength Determination
VOLCANIC ROCKS 329
CH
- LINE
INDEX
PLASTICITY
OH
MH
MI7
cZ
MLOL
-
[
ТО
100
60
80
40
LIQUID LIMIT
Fig Plasticity Chart Minnow

11
Station Slide
,
.
.
SAMPLES FROM DRILL HOLE 224

D
-
UNCONFINED
JU
UC TYPE TESTS
&
COMPRESSION
TESTS
.
-
FT
STRESS
40
!
DEPTH
I
SHEAR
801
0
.5
1
0
2
- 5
5
2
.0
.
0
.5
.
1
.0
1
.5
APPLIED STRESS DEVIATOR STRESS

T
/
Fig Shear Strength Data Station Slide

12
Minnow
-
.
.
or
of
ground surface except near the toe near river

is
the abutment where

at
it
of
level The degree generally between

of
saturation the overburden soils

is
.
percent but ranges from percent

88
92
75
98
and
to
Shear strength values obtained from unconsolidated undrained tests

on
the
residual soil varied over wide range due variations the basic soil
to
in
a
properties compressibility
of
of
voids ratio water content degree saturation

,
plasticity residual rock structure The wide variation shear strength

in
and
,
.
was basically due to differing degrees of decomposition of the parent rock .
Typical shear test data are shown in Figs . 3, 4 , and 13 , and the scattering of
test values obtained from one test pit is illustrated in Fig . 14 .
Due to wide scattering of test results , the mass shear strength of the abut
ment was studied by analysis of the steeper portions of the natural abutment
slopes . Slopes appeared to be stable under existing ground - water conditions ,
but soil accumulations and several large slide scars supporting second
growth fir trees indicated that shallow slope adjustments have occurred within
recent years . The average stress was determined for different depths of
circular arcs and by infinite slope method of analysis . A comparison of the
strength of the residual soil, assuming the slope was on the verge of failure ,
determined by the above method with the results of consolidated drained and
unconsolidated undrained shear tests on samples of these materials is shown
in Fig . 15 . The shear strength indicated by slope analysis is lower than any
value obtained by consolidated drained direct shear tests but compares favor
ably with the lower values obtained by unconsolidated undrained triaxial tests .
It was therefore considered a conservative value for design purposes . A
similar analysis was made for the strength of the lower stratum of weathered
tuff , using the strength previously determined for the portions of the arc
which pass through the residual soil.
COMPACTED SHEAR STRENGTH AND CONSTRUCTION EXPERIENCE
ML - MH Silts . - Plastic silts of ML - Mh classification derived from volcan

ic rocks are frequently encountered on projects constructed in the Willamette
River Basin . These materials are generally found in terrace deposits locat
ed high above the valley floor except where massive slides have transported
them to lower elevations . The terrace deposits are frequently decomposed
to a homogeneous mass with no remaining visual evidence of the parent rock
structure ; but on the steeply inclined slopes of the valleys , the homogeneous
soils have been eroded , exposing less weathered material with a clearly evi
dent remnant rock structure . The less weathered materials usually contain
a high proportion of hard rock fragments . Since the water content of residual
MH soils often approaches the liquid limit , a pronounced loss of shear
strength at natural water content occurs upon destruction of the remnant rock
structure by remolding . However , the angle of internal friction obtained by
consolidated drained direct shear tests may be as high as 35 degrees . Ex
perience with the use of these soils on several projects is cited in the follow
ing paragraphs .
Lookout Point Relocations . - Construction of railroad and highway reloca
tions for the Lookout Point Project involved a three -mile section of this type
of clayey silt . Typical Atterberg limits are shown on the plasticity chart
of Fig . 16 . Standard Proctor compaction tests indicated a maximum dry den
sity of 70 to 90 pounds per cubic foot and optimum water content of 30 to 50
percent . The natural water content was generally between 42 and 64 percent .
Typical shear strength data for the compacted material are shown in Fig . 17 .
Although these materials would have been excellent for construction of em
bankments if they could have been compacted at a water content reasonably
close to optimum , only a small percentage could be so used due to high natural
water content . A reduction of the water content of this type of material suf
ficient for proper compaction is not practical during highway and railroad
construction . Due to loss of shear strength on remolding , scrapers and trucks
VOLCANIC ROCKS 331
/
MAJOR PRINCIPALSTRESS tons qft
ft
, /
sq
SHEARSTRESS tons
34
sq
APPLIEDPRINCIPALSTRESSEStons/
ft
,
TEST NO
26
Watercontentwo.
ft %
27
INITIAL
25
.4
,
.4 .0
7 0
cu
Dry density Ibs
86. . .
95 91 , 94
/
5
.
Voidratio 796 789 870 864
, ,
., e,
6
0
0
.
.
.
92
%
Saturation 91 85
S
ws
TEST
W
aftersaturation
%
.C
,
,
.
AT FAILURE BEFORE
, %
Soturation
S
,
,
,
% sq
Consolpressuretons/
ft
%
W
afterconsolwc
STRAIN
.C
,
.
voidratioafterconsola 757 10 776 60

3 .
. .
64
92
74
74
Majorprin stresstonss41
VOLUMETRIC
0
o
0
. ,
,
/ /
.
20 Minor stresstons
sq
p rin
ft
Unconfined
2
. 0
% ,
4
. 0
.
|
. . . 26.
. . . 26
Watercontent we
er ,
voidratio 7340 754 70 5

,
5 2 0
Specimen diameterinches
98
98 98
9870
,
5 2
3 1
4 1
. .
. .
Initialheightinches 98 01
,
10
20 30 Testtime failureMin
to
AXIALSTRAIN
%
,
INCHESNOOP
MESHPERINCHVS. STOGRAINSIZEI MILLIMETERSTYPE OF TEST
N
,
TYPE OF SPECIMEN
III Constant 63 Undisturbed
ton
Strain controlled
5015
-
' 0
sq9
-
ACENTTINERSY WEIGHT
BYWEIGHT
0 .
Un drained
un consolidated tons/
,
ft
c
REMARKS Description Light brown

,
:
for
Totol Sample 2
Grddation _ wet medium clay and weathered
,
COARSER
angular rock
.
2
_
2
PERCENT
PROJECT HILLS CREEK DAM SITE
SAMPLE DATA Tunnel

97536 98899
N
8
-
-
.
.
TABY

NO
11
EXPL SAMPLENO.
/ GR
GRAVEL SAND
26
27
FINES ELEV 1326 DEPTH

.5
3
-
.
'
Classification Gravelly
| W
Sandy CLAY CH NO.

( 10 O
52 322
5
-
-
(
).
74
LL
56
JC
Drown Dec 1952

G
DOLLOFF
2
PI
.
JL
, .
.
Do Checked
cum Date Report ChielSoilMechanics
of
Bronch
"
NPD FORM TRI AXIAL COMRESSION TEST REPORT

OF
-
APRIL 50 CORPS ENGINEERS

1
Fig
13
Unconsolidated Undrained Test Hills Creek

,
Dam
.
.
332
Sample Depth LL PIW
%
Sat
4 ft
- -
.0 .
.
43 20 23 72
3
.0
5 8 20
.5 24
43 89
.2 3
0
4
2 .0
- -
.
16
1
.7
4
1
8
20 21
82
.5 .0 .5 6
. 6
4
1
.5 .6 . 153
3
15
- - - -
31.
.
*=0.
.
. 32
1
40 14 94
. 38 5
37 42 17 16
ton
9
24
42 43 49 21
7
-2.7 ton 49 18 13 77
. 12
2
0 .
= 2
*4-C
C=
.
45
15
.
20
*5-
-
A
#
.
$
ton
T
7
=
2
.
ton
-
C
-
6
*
Ź
39
,
*- .)
0
$=
STRESS
(C.D
tan
63
08
\
.
-0.
0
=
/C
•
23
ton
.0
=
SHEAR
•
.05
ton
21
.0
0
=
-C
-C=
3
*
13
*2
.
34 ton
0
.
:
0
C
.
-
#
7
-
W Trioxiol tests except Direct sheor
D
U
.U
C
2
*
-
.
.
.
3
Ft
APPLIED STRESS
T
-
.
/
Fig
Fa
Pit
Hills Creek
14
Test Data Test Dam

-
-
.
.
Direct Shear Tests

x
-
Arcs from Triaxial Tests

.U
U
.
./
2
ft
T
-
STRESS
17
1
SHEAR
.
0
.4
=
Design Strength Envelope

2
STRESS
Ft
APPLIED
.2
T
-
/
Fig Design Strength Hills

15
Shear Test Data Creek Dam

&
.
.
-
-A
-A
Line CH Line
INDEX
INDEX PLASTICITY
PLASTICITY
чон
.
.
&
мн Он
&
CL
./
ML OL
&
OL
SEML
CS2
ML
OL
MEMLE
oo
20
20 40 60 80 100 20
80
40 60 100
LIQUID LIMIT LIQUID LIMIT
-
.. 18
.
- .
16
Fig Plasticity Chart Minnow Creek Cut Fig Plasticity
A Chart Borrow
B
&
Areas Cougar Dam
,
60
2
.
f4
w
.
/
.C.
Proctor Opt
%%
T
VOLCANIC ROCKS
-A
CH
WW
Line
INDEX
-
%
35 32 32
!
-.C.-.C. -.C.
tw
-.C
W
.
34
CL
STRESS
%
34
ZWW
.
.C- -
UU Triax Shear
PLASTICITY
CD Direct Sheor MHE OH

TE
9
MHEOH
W
C
.
..
45
SHEAR
w
%
Nat
I
CI
- M
.
MLE OLT
C
%
W
52
2
20 40 60
:C.. -.. 4
100
3
80
APPLIED STRESS LIQUID LIMIT
..
..
17
Fig Shear Strength Data Remolded Fig Plasticity Chart Borrow
333
19D
&
.
--
Samples Minnow Creek Cut Area Rt Abutment Cougar Dam

could operate in some cut sections only during favorable sunny weather . Em
bankments were designed to utilize wet material in uncompacted random fill
zones where the shear strength and settlement properties were not important .
Cougar Dam . - The rockfill dam for the Cougar Reservoir Project is pres
ently in the initial construction stage . This dam will have a maximum height
of 425 feet , and 800 , 000 cubic yards of impervious material will be required
for the steeply sloped upstream core . Adequate imprevious borrow sources
were scarce in the steeply incised canyon of the McKenzie River . Several
high elevation benches of deep residualMH soils were investigated but later
rejected for use in the imprevious zone of a high dam by consultants for the
project . Typical Atterberg limits for these plastic sandy silts are shown in
Fig . 18 . Standard Proctor compaction tests indicate a dry density ranging
from 75 to 92 pounds per cubic foot and optimum water content from 27 to 42
percent . The natural water content averaged about 10 percent higher than
optimum . Typical shear strength data are shown in Fig . 21 .
Further investigations indicated that stripping of relatively shallow over
burden and weathered rock from the lower valley slopes would provide a mix
ture of rock fragments and plastic fines with essentially the properties of a
clay binder . Stripping of the slopes by dozer operation blends the heterogene
ous materials and improves the gradation and provides more uniform plasti
city properties . Typical Atterberg limits for these materials are shown in
Fig . 19 . This material was considered less susceptible to cracking due to
differential settlements and consequent piping and therefore more suitable for
the imprevious core . Typical gradation curves are shown in Fig . 20 . The
rock content has an appreciable influence on the shear strength properties as
indicated by the difference between the strength curves on the minus No . 4
sieve fraction and minus 1 - 1 / 2 - inch fraction as shown in Fig . 22 . The test on
samples containing rock fragments to 1 - 1 / 2 - inch size were performed on
6 - inch - diameter triaxial shear specimens . Rock fragments larger than 6
inches will be removed before placement . Standard Proctor compaction tests
on the fraction of material passing a No . 4 sieve indicate a dry density rang
ing from 96 . 3 to 104 .2 pounds per cubic foot and optimum water content from
20 . 2 to 23. 9 percent . The natural water content on the same size fraction ay
erages about 7 percent over optimum . The importance of placement water
content relative to the unconsolidated undrained shear strength is indicated by
the shear strength curves of Fig . 22 . The unconsolidated undrained shear
strength at optimum water content is indicated by c = 0. 2 ton per square foot
and an apparent angle of internal friction of 26° . This is reduced to c = 0 . 8
ton per square foot and an apparent angle of internal friction of 0° when the
water content of the fraction passing a No . 4 sieve is 8 percent above optimum .
Drying the soil to optimum water content is considered impracticable from an
operational standpoint , but a reduction in water content to within approximate
ly 5 percent of optimum will be required to obtain sufficient shear strength
for an adequate safety factor at the end of construction . Although the imper
vious core is relatively thin , an appreciable gain in shear strength during con
struction due to the dissipation of pore pressures cannot be expected . Piezo
meters will beinstalled to obtain pore pressure data as a check on shear
strength during construction .
CL - CH Clays . - Clay soils of volcanic origin have been used for the imper
vious core for two earth - fill dams constructed in the Willamette River Basin .
Dorena Dam was completed in 1949 and Lookout Point in 1954 . Both dams are
basically thin central core designs with gravels as the supporting shell ma
terials . Seasonal rainfall limited core placement to approximately 100 days
SIE VE ANALYSIS
HYDROMETER ANALYSIS
IN
SIZE OF OPENING INCHES
.,
NUMBEROF MESH PER INCH US STANDARD
IN
GRAIN SIZE MM
100
100
III
111
WEIGHT
BY
VOLCANIC
WEIGHT BY FINER CENT PER

COARSER CENT PER
ROCKS
34
IN
200
IN
GRAIN SIZE MILLIMETERS CORPS ENGINEERS
OF
UNIFORMSOILCLASSIFICATION
Coarse Fine Coarse Medium
COBBLES Fine
GRAVEL SAND FINES
20
..
,
,
Fig Gradation Curves

335
Cougar Dam Impervious Core

per year on these projects , and a reduction in the high natural water content
of plastic imprevious soils was assisted by the use of large borrow areas .
Construction work was concentrated on placement of core materials during
the dry summer season and the gravel shells were filled out during the wetter
seasons. Imprevious borrow materials and construction operations are de
scribed in the following paragraphs .
Dorena Dam . - Dorena Dam has a height of 140 feet and a total volume of
3 million cubic yards of which 680 , 000 cubic yards of material are impervious .
The impervious borrow was obtained from a large flood plain deposit of plas
tic clay 1 to 15 feet in depth overlying clayey gravel . A wide range in the
liquid Limit is illustrated by the plasticity chart in Fig . 23 . Because of the
natural water content of the borrow material it was necessary to design the
embankment for placement of impervious material at water contents higher
than optimum . For the design studies it was assumed that core material
would be placed 3 percent wetter than optimum and dissipation of pore pres
sure would not occur during construction . The shear strength of the clay core
used for design is shown in Fig . 24 . A broken strength curve was used for
the analyses . For normal loads under 3 . 3 tons per square foot values of tan
= 0. 60 and c = 0 applied , and for larger normal loads values of tan 0 = 0 . 12
and c = 1. 57 tons per square foot were used . The former coefficients were
the values selected from drained direct shear test data as representative of
the strength of clay in the shallower portions of the core that are subject to
saturation and may lose the " cohesion " temporarily acquired by compaction .
The coefficients tan 0 - 0. 12 and c = 1.57 are the values selected from un
consolidated undrained triaxial shear test data as representative of the
strength in the deeper portions of the core where pore pressures would result
in a reduction in shear strength .
The natural water content averaged about 45 percent and Standard Proctor
optimum water content was about 32 percent . The major construction problem
was to reduce the water content to about 3 percent above optimum to permit
satisfactory placement of material in the embankment . Drainage ditches were
excavated in the borrow area to lower the normally high water table . Disking
and roto - tilling in 3 - to 6 - inch depth increments accelerated drying of the
moderately stiff clay . Two to four days were required to process each lift
excavated from a borrow area . The soil, after tilling , would surface - dry into
hard cloddy lumps unsuitable for compaction . However , when dried to well
below optimum moisture content and then brought back up to optimum by the
addition of moisture , the hard surface would s lake , producing a material of
more uniform water content and improving compaction conditions .
Lookout Point Dam . - Lookout Point Dam has a maximum height of about
275 feet and a total volume of 8 million cubic yards of which 1, 157 , 000 cubic
yards of material is impervious . The material for the impervious core was
generally a reddish brown to brown plastic clay and silt resulting from de
composition of quaternary terrace gravel and tertiary pyroclastic rocks . The
plot in Fig . 25 of Atterberg limits obtained on embankment control samples
again illustrates the variable characteristics of volcanic soils . The stability
of a dam with a narrow central impervious core supported by free - draining
gravel shells depends only slightly on the shear strength of the core material.
For this reason , the main effort in design and construction control on the
core at Lookout Point was toward obtaining a permanently impervious core .
Test fills constructed prior to contract award indicated that 95 percent of
Standard Proctor density could be obtained with material placed in 6 - to
95
at
%
2 Specimens placed Proctor
W
.
on
..
Density corresponding
ft
ofof
Cos
,
/ wey side optimum except
.
T noted
-A
LINE
INDEX
-
Direct Shear
CO
--
UU Triaxiol Shear
CL
STRESS
PLASTICTY
он
¢
мн
SHEAR
CLEMA MLOL
0 %
Proctor Density
2
20
o 40 60
80
100
LIQUID LIMIT
f4
2
APPLIED STRESS
21
..
..
,
23
Fig Shear Data Borrow Fig Plasticity Chart
&
,
A
B
Areas Cougar Dam Dorena Dam Borrow
VOLCANIC
Opt
,
CD Direct Shear 2 CUU Triax Shear Tests

.. C
%
.
3
=
W
UU Triaxial Shear . obore opt

f4
-- .
,
)"*
- -1(&
)*
(
Mox size mot 12 /
ROCKS
/ )"-)4*-(.
opt
+5
%%
.
- Opt opt -
zby Design Strength Curves
< 61
0
CD
STRESS From Testen ton

2
.
14
Opt
./
12 T
STRESS
= 57
.
01
=
c
2
=
Opt
60
SHEAR
% %%
./T . 4+. 3-8.+
(. ."* )4( )"( 14
Opt
SHEAR
.0-0
ton
2
8
6
8
10
T
66 -
-
APPLIED STRESS APPLIED STRESS TA12
22
-
24
,,
.
.
Fig Shear Data Borrow Fig Shear Strength Data

337
Data
Strength
,
..D
Area Cougar Dam Dorena Dam Borrow
CH
A - LINE
INDEX
PLASTICITY
CL
OH
MH
MI
ICE
MLOL
:
-
20 40 60
LIQUIT LIMIT
Fig Plasticity Chart Lookout Point Dam

25
Borrow
,
,
.
.
IMPERVIOUS CORE MATERIAL
/-
ft2
Design lay
T
=)
30
, (C
24. 2°
=
C
.0
SHEAR STRESS
Range
of
Results
Average
79
6°
M2
C
1
T
8
,
-
/
-
.
APPLIED
APPLIED STRESS T4
TA
2
STRESS
--
?
Fig
26
Unconsol Undrained Shear Strength Data

.
.
Record Samples Lookout Point Dam

,
/T
IMPERVIOUSCORE MATERIAL
ft2
- Avecs
11
0287
C
-
/
SHEAR STRESS
nge
of
Results
Design clogo
12
30
;.
C(
242 142
//
-0
7
-
.
472
APPLIED STRESS
-7
/
27
Fig Consol Drained Shear Strength Data

,
.
.
Record Samples Lookout Point Dam

.
VOLCANIC ROCKS 339
12 - inch looselifts and compacted with 4 passes of a 4 - wheel pneumatic - tired

roller with 85 , 000 poundsgross load and 85 pounds per square inch tire pres
sure . Standard Proctor densities ranged from 73 to 92 pounds per cubic foot
and optimum water content from 28 to 44 percent . Since the natural water
content in the borrow areas ranged from 20 to 60 percent, it was necessary
to open relatively large borrow areas and provide sufficient time for aera
tion . However , tilling or reworking of material in the pit was not necessary
and unusually long dry weather seasons during construction premitted ex
tended operations . Placement water contents were controlled primarily by
action of the pneumatic roller . The upper limit of water content was that
which permitted placement and compaction with the specified roller without
excessive deformation of the material. The lower limit was the minimum
water content with which satisfactory compaction could be obtained . Con
struction control data indicated the average placement water content was 3. 3
percent over the average optimum of 35 . 4 percent and the average density
was 96 percent of Standard Proctor . A comparison of design strength values
and record sample test results is shown in Figs . 26 and 27 . Since the natural
water content of impervious borrow sources located in the Willamette River
Basin is generally high , it is necessary to determine workable placement
procedures and adapt the embankment design to the available materials .
High pore pressures are to be expected under these conditions .
CONCLUSIONS
Construction of earth dams and railroad and highway re locations for the
Willamette Basin Project required the use of plastic silts and clays often
containing rock fragments or gravel and usually having in - situ water contents
over optimum . Evaluation of the undisturbed shear strength properties of
these extremely heterogeneous residual soils derived from volcanic rock pre
sents a complex problem . Improved laboratory equipment and procedures and
a better understanding of the shear strength properties of partially saturated
soils of this type are urgently required . However , experience has indicated
that laboratory shear tests often cannot be relied upon for determination of
stability of these soils against sliding . Old slide planes , clay filled fault
cracks , and other geological aspects greatly influence the mass strength . An
understanding of the geological history and stratigraphy is essential for the
location and study of potential slide areas . Closely spaced borings , often sup
plemented by test pits , are required for detailed studies of old slide areas ,
and experience has shown that even after considerable test data are available ,
a true evaluation of the effect of a change in stress due to construction work
on the stability of such slopes is not always obtained . Analysis of slopes
which have failed has been a useful aid in evaluation of the mass shear
strength under these conditions .
Construction of earth dams on the western slopes of the Cascade Range
usually requires the use of highly plastic clays and silts at water contents
higher than optimum for the impervious zone . Additional studies of the com
pacted shear strength properties of these soils , particularly those containing
a highproportion of rock fragments or gravel, are required for improved

economy and safety in design of future dams .
ACKNOWLEDGMENTS
Design and construction of the Willamette River Basin Reservoir System

dams and relocations have been under the direction of the Portland District ,
Corps of Engineers , Colonel Walter L . Winegar , District Engineer . The writ
ers are grateful to Dr . Paul W . Howell for his contributions to this paper on
the geological aspects of soils of the Cascade Range .
NONUNIFORM CONDITIONS IN TRIAXIAL TEST SPECIMENS
By Woodland G . Shockley , 1 F . ASCE , and Richard G . Ahlvin ,? F. ASCE
SYNOPSIS
Small triaxial specimens of fine sand were prepared at relative densities

ranging from dense to very loose , subjected to different degrees of axial
strain , and volume changes were determined for increments of specimen
height. The tests showed that yolume expansion occurs in the failure zone of
the specimens, and volume decrease occurs at the ends . Very large triaxial
specimens incorporating pressure cells were also built . These showed the
vertical stress at midheight to be greater in the center than toward the edge ,
and stress near the base to be greater toward the edge than in the center . A
specimen incorporating strain gages showed a similar pattern for strains .
INTRODUCTION
When triaxial shear test results on practically any type of soil are evalu
ated , it is commonly assumed that the stress and strain relations are uni
form throughout the specimens, or at least they are uniform in the plane or
zone of shear . It is recognized that different conditions may exist near the
ends of the specimens because of the restraining effects of the specimen cap
and base ; however , if the failure plane or zone is outside the end zones of the
specimen , the end effects are usually neglected . Computations of stress ,
strain , and volume change ( if any ) in triaxial specimens customarily are
made assuming average values for the entire specimen . If nonuniform con
ditions exist in a triaxial specimen , then stresses , strains , and volume chang
es computed on the basis of average specimen conditions may not be at all
indicative of the changes that are occurring in the zone of shear . This could
be especially important in cases where failure occurs as a bulging of the
specimen rather than as a well - defined shear plane.
In connection with investigations conducted by the Waterways Experiment
Station for the Mississippi River Commission from 1947 to 1952 on the nature
and behavior of the banks of the Lower Mississippi River , extensive tests
were conducted on remolded samples of Mississippi River sands to determine
their strength and stress - strain characteristics . One phase of this study was
directed toward finding the so - called “ critical void ratio ” of the sands by tri
axial shear test methods. It became apparent as work progressed that in
order to evaluate the test results with confidence , careful techniques must be
developed to insure uniform density throughout the “as - built " test specimens,
and that data were needed on density changes in different parts of the test
1. Asst. Chf., Soils Div . , U . S. Army Engr . Waterways Experiment Sta .,

Vicksburg , Miss .
2 . Chf ., Special Projects Sect. , Flexible Pavement Branch , Soils Div . , U. S .
Army Engr . Waterways Experiment Sta ., Vicksburg , Miss .
341
specimens as the specimens were strained in shear . Tests performed to ob
tain the latter data are described in this paper . Also presented are data from
other investigational programs on volume change in triaxial specimens of clay
and the distribution of stresses and strains in 35 - in . - diameter triaxial speci
mens of sand .
TEST PROGRAM AND APPARATUS
The test program on Mississippi River sands was conducted in two gener
al series : In the first series , triaxial specimens of dry sand placed at differ
ent relative densities were strained under axial loading under constant lateral
pressure (partial vacuum ) ; in the second , saturated triaxial specimens of the
same sand were tested at constant volume, one group being tested with
constant pore pressure and another group with constant lateral pressure . In
both series , density changes in different parts of the test specimen were
mea sured at specific degrees of axial strain .
The sand used in all of the tests was a clean , fine , subrounded sand from
a deposit on the Mississippi River . It has a D10 size of about 0 . 22 mm and a
uniformity coefficient of about 1. 4. The laboratory maximum density is about
106 lb per cu ft and the minimum density is about 90 lb per cu ft ( 1). 3 A
grain - size distribution curve of the sand is shown in Fig . 1.
The triaxial test specimens were 2 . 8 in . in diameter and 6 . 5 in . high .
Specimens were prepared in loose , medium , and dense conditions in split
molds in which the rubber membrane had already been placed . After numer
ous trials to determine placement procedures that would give the most uni
form initial densities , the following procedures were adopted . All specimens
were placed in the dry condition . Loose and very loose specimens were pre
pared by carefully placing the sand in the mold using a spoon designed by
L . S. Chen (2) ; for the loose specimens, oven - dried sand was used , and for the
very loose specimens, sand that had been stored in a humid room for 24 hours
was used . Medium and dense specimens were prepared by placing the sand
in 10 layers of equal thickness and striking each layer with blows of a 220 - g
hammer having a 2 - in . -diameter striking face . Medium - dense specimens
were prepared using a 1- in . hammer drop and 3 - 3 - 4 - 4 - 5 - 5 - 6 - 6 - 7 - 7 blows
per layer, respectively , for the 10 layers starting at the bottom , or a 2 -in .
hammer drop and 0 - 0 - 0 - 1- 2 - 3 - 4 - 5 - 6 - 7 blows per layer . Dense specimens
were prepared using a 6 - in . hammer drop and 5 - 7 -7 - 9 - 9 - 11 - 11 - 13 - 13 - 15
blows per successive layer from bottom to top . These procedures were found
to give initial specimen densities that varied less than +1 lb per cu ft through
out the specimen .
The first series of tests was performed on dry specimens at different
initial relative densities and in which the lateral pressure ( 12 . 28 psi ) was
provided by applying a partial vacuum to the interior of the specimen . This
procedure permitted testing the specimens in the open without confinement in
the triaxial chamber . In order to measure density changes during axial
strain , the rubber membrane had a series of equally spaced horizontal lines
drawn on the surface. Full - scale photographs of the specimen were taken at
0 , 2 . 5 , 5 . 0 , 7 .5 , 10, and sometimes 15 per cent of axial strain . A typical test
3. Numerals in parentheses refer to corresponding references in the

Appendix .
IN
STD SIEVEOPENINGS INCHES US STD SIEVENUMBERS
US
HYDROMETER
100 140 200
14
16
20
30
40
50
70
10
100 MUTT
LARGE TRIAXIAL SPECIMENS SECTIONED VOLUME CHANGE

WEIGHT
WEIGHT
LABORATORY SPECIMENS
.
AND COMPARATIVE LABORA BY
BY
TORY SPECIMENS
NONUNIFORM
FINER CENT PER

COARSERCENT PER
CONDITIONS
So Jioo
0
.
.
0
5
100
.
0
QI 05 001 0005 001
IN
GRAVEL SAND
OR
SILT CLAY
D
COARSE FINE COARSE FINE
I
MEDIUM
..- 1
in
.
Fig Sand Used Triaxial Tests

343
setup isshown in Fig . 2. The average diameters and vertical distances be
tween grid lines were measured on the photographs and , when corrected for
photographic distortion , were used to compute the apparent volume changes .
Sufficient external measurements were made with vernier calipers on select
ed specimens to check the validity of this process . A further check on the
validity of the volumechange calculations was made by saturating selected
specimens at specific degrees of strain , allowing them to drain , quick
freezing them , cutting the specimens into increments transverse to the axis
( usually four ) , and determining the density of each increment by volumetric
methods.
The second series of tests was performed on saturated specimens of sand
at medium -dense and loose , relative densities . The sand specimen was pre
pared dry , then saturated by passing ammonia gas through it , after which
water was introduced and the sample allowed to saturate from bottom to top .
By this process , degrees of saturation of 99 per cent or greater could be
achieved . A schematic diagram of the triaxial apparatus is shown in Fig . 3.
T27
Fig . 2 . - Triaxial Test Specimen
Dry Sand Tests
PISTON
TO
COMPRESSED
AIR
LINE
PROVING RING
-
PRESSURE GAGE
PRESSURE
CLUETTE
REGULATOR
PLEXIGLASS
CHAMBER
-
DOUBLE STAGE GRADUATED
- TITTITILITZ
PROVING FRAME STANDPIPE
ZITUZ
NONUNIFORM
NIN
-
PRESSURE NVACUUM PRESSURE
RESERVOIR GAGE
ZZZZ
WATER LEVEL MEMBRANE
SOIL
CONDITIONS
SAMPLE
CUZZLUCILLEZ
LITETIT
99
TO AMMONIA
TTTIIIIIIII SUPPLY
.
of
. -.3
Fig Schematic Diagram Triaxial Shear Apparatus
345
The test specimens were strained at constant volume under axial loading us
ing two different test procedures : In one the lateral pressure (03 ) was held
constant and the pore pressure ( u) was controlled to maintain a constant
volume ; in the other the pore pressure was held constant and the lateral
pressure was controlled . After the specimens were strained to approximate
ly 10 per cent axial strain , they were drained , quick - frozen , and cut into four
axially transverse segments , and the density of each segment was determined
by volumetric methods . In order to investigate possible “ crust action ” (that
is , the formation of a denser and therefore stronger crust around the peri
phery of triaxial specimens as a result of the confining effects of the mem
brane and the applied lateral pressure ), additional specimens were prepared
as above , except that 1. 5 - in . - diameter cylindrical plugs were taken from each
of the axially transverse segments and the density was measured . As a check
on the validity of the procedure, unstrained saturated test specimens were
treated in the manner described above .
Nontilting specimen caps were used on the triaxial apparatus in all tests
so as to insure uniform vertical deformations during axial strain . The rubber
membranes used in the tests had a thickness of 0. 05 cm . Tests were carried
out on the membrane itself and on sand specimens confined in one , two , three ,
and four membranes to determine the effect of membrane restraint . These
tests , together with theoretical computations , indicated that deviator stresses
resulting from membrane restraint were less than 1 psi at axial strains less
than 20 per cent . In view of this finding it was concluded that membrane re
straint had little effect on the test results reported herein .
TEST RESULTS AND DISCUSSION
Results from the first series of tests on typical samples are shown in
Fig . 4 . Here are plotted volume change versus axial strain for four dry sand
specimens ranging in density from dense to very loose . Although measure
ments were made for 13 individual vertical segments of the specimens , for
purposes of illustration they have been averaged into three approximately
equal- size groups , designated as top , middle , and bottom zones , since the
volume changes in the individual segments of each zone were reasonably con
sistent . In all cases the middle zone encompassed the visual failure zone of
the specimen ; this was usually not at the geometric center of the specimen ,
but slightly below that region .
It may be seen in Fig . 4 that the volume changes in the middle or failure
zone of the specimens are significantly different from those at the ends . For
all specimens , except the very loose specimens at strains less than about 3
per cent , there is a volume increase in the failure zone , whereas at the ends
of the specimens there is a consistent volume decrease . This is a somewhat
different picture than is obtained from total volume change as usually
mea sured in triaxial tests and as shown by the da shed lines on the figure .
Total volume changes tend to follow the conventionally thought - of pattern of
volume increases for dense specimens and volume decreases for loose speci
mens . However , it is readily seen that the total volume changes are not in
dicative of volume changes that are occurring in the zone of shear . There
fore , interpretations of test data based on total volume changes may be
completely at variance with those determined on the basis of changes in the
failure zone alone.
NONUNIFORM CONDITIONS 347
It is not difficult to visualize the reason for the pattern of volume change
in the failure zone of very loose specimens . The initial small strains will
tend to cause the more delicately balanced particles to readjust and to fall
into adjacent open pockets , thereby causing a slight decrease in volume.
Further strain forcing the particles to move produces a tendency for them to
roll over one another , as the previously available open spaces have been
filled . Thus , increased strain causes a volume increase in the zone of shear .
Results of tests on typical specimens of saturated fine sand are presented
in Fig . 5 . Densities of the four quarters of the specimen (open circles ) are
shown for medium - dense and loose placement conditions after the specimens
had been strained to 10 per cent axial strain in constant volume shear . Data
are shown for constant volume tests in which the lateral pressure (03 ) was
held constant and the pore pressure (u ) was varied , and for tests in which the
pore pressure was held constant and the lateral pressure was varied .
SHEAR PLANE DEVELOPED

BEFORE 15 % STRAIN BULGE FAILURE
INCREASE
VOLUMECHANGE PER CENT
DECREASE
IN
15
IN
DENSE 105 LB PER CU FT MEDIUM DENSE 100 LB PER CU FT

.2
.1
-
-
INCREASE
FAILURE BULGE FAILURE

DECREASE
150
-
10
11
AVI
JA
IN

"
'
LB PER CUFT LB FT
93
CU
96
LOOSE VERY LOOSE PER

0
.4
.
-
-
LEGEND
IN
IN
oTOP NOTE DIAM HIGH SPECIMENS.

X
6
.5
2
.8
-
-
-
-
:
TOTAL SAMPLE CONSTANT LATERAL PRESSURE

12
•
28 PSI VACUUM
=
-
O A --
MIDDLE FAILURE ZONE)

(
.
).
D A
BOTTOM
Fig Volume Changes Triaxial Specimens

in
.4
-
.
Dry Fine Sand

As procedures , the left -hand plots in Fig . 5 show re
a check on placement
sults of tests on specimens which were saturated and consolidated under the
initial lateral pressure , but before axial strain was applied . The long -dashed
lines represent the average density of the entire specimen after consolidation
as determined by external measurements ; the open circles show densities of
the individual quarters of the specimens as determined by the freezing and
TOP
SPECIMEN
OF
QUARTERS
воттом
98
90
92
94
96
IN98
94
96
98
100 102 104 100 102 100 102
DRY UNIT WEIGHT LB PER CU FT
CONSOLIDATED CONSTANT SPECIMEN CONSTANT SPECIMEN
U
10
10
PER CENT AXIAL STRAIN PER CENT AXIAL STRAIN
MEDIUM DENSE 101 LB PER CU FT

-
QUARTERS OF SPECIMEN
92
94
96
98
88
90
92
94
96
98
90
92
94
96
98
100 100
"
IN
DRY UNIT WEIGHT LB PER CU FT

CONSOLIDATED CONSTANT SPECIMEN CONSTANT SPECIMEN
U
10
10
PER CENT AXIAL STRAIN PER CENT AXIAL STRAIN
LB FT
95
LOOSE PER CU
–
LEGEND
OF
IN
NOTE
IN
ENTIREQUARTER SPECIMEN DIAM HIGH SPECIMENS

2
.8
X
6
5
-
:
.-
,
-
-
.
IN
DIA CYLINDRICAL PLUG CONSTANTVOLUME TESTS

1
.5
-
FROM CENTER OF QUARTER INITIAL 25 PSI

=
AVERAGE DENSITYOF INDIVIDUAL QUARTERS

AVERAGEINITIAL DENSITYAFTER CONSOLIDATION
Fig Volume Changes Triaxial Specimens

in
5
–
.
.
Saturated Fine Sand

cutting process ; and the short - dashed lines show the average density of the
entire specimen as obtained by averaging the individual densities of each
quarter . It may be seen that the densities of the individual quarters of the
consolidated specimens were within 2 lb per cu ft of the average initial den
sity after consolidation , and the density determined by averaging the four
quarters was within 1 lb per cu ft of the initial density after consolidation .
These variations are considered to be minor and are the result of slight yari
ations in the sample placement procedures, changes resulting from the
process of consolidating the specimens , and inherent inaccuracies in sampling
and weighing techniques .
The densities of the specimen quarters after axial strain show the same
tendency for variation as noted for tests on the dry sand specimens described
previously ; that is , there was a greater decrease in density in the middle of
the specimens than at the ends . It will be noted that in some cases there was
a density decrease at the ends of the specimens as well as at the center in
contrast to the first series of tests on dry sands where there was a consistent
density increase (volume decrease ) at the ends . Sufficient tests were not con
ducted to explain this discrepancy in behavior , but it may be surmised that
the saturation of the specimen in the second series of tests had a definite in
fluence on the specimen behavior during axial strain .
Fig . 5 also shows that density changes tended to be greater for the constant
lateral pressure tests than for the constant pore pressure tests ; the reasons
for this difference are not known . Further , there was a tendency for greater
volume change with the medium - dense specimens than with the loose speci
mens , as would be expected .
The solid circles shown in Fig . 5 represent densities of 1. 5 - in . - diameter
cylindrical plugs taken from the center of the specimen quarters . These data
indicate lower densities for the cylindrical plug than for the entire quarter ,
thus implying a higher -density annular " crust ” for the specimens . However ,
the cylindrical plug densities for the unstrained specimen also show a trend
for lower density , as well as wide variations in density from one quarter of
the specimen to another . It is believed that the techniques of trimming and
measuring the axial segments for density determinations were not sufficiently
precise to produce reliable results . Therefore , any conclusions based on
densities of the cylindrical plugs are open to question .
In summary , it may be stated that even for constant volume tests on satu
rated sands , there are nonuniform density or volume changes throughout the
specimens as they are subjected to axial compression between rigid end
plates .
TESTS ON CLAY
The test results described above raised the question as to whether non
uniform changes also occur during undrained compression of saturated clays .
Consequently , the Waterways Experiment Station requested the Massachusetts
Institute of Technology to make some preliminary studies under an existing
contract . The results of these investigations have been reported by Taylor
and Clough (3 ) and are summarized below .
Four pairs of specimens of undisturbed , Boston blue clay were tested .
The specimens were 2 . 8 in . in diameter and ranged in height from 5 . 5 to 11 . 0
in . They were initially consolidated in a triaxial apparatus under a chamber
pressure of about 109 psi, after which one specimen of each pair was subject
ed to axial compression with no drainage permitted . Each specimen was cut
into a number of horizontal segments and the water content of each segment
was determined . It was postulated that changes in water contents between
similar segments of the pairs of specimens which had and had not been sub
jected to axial compression would be indicative of corresponding volume
changes , since the specimens were saturated .
Results of these tests showed that the average over - all moisture contents
of the respective pairs of specimens checked closely , the maximum discrepan
cy being 0 . 3 per cent ( initial moisture contents of all specimens ranged from
28 to 32 per cent ) . The specimens subjected to axial compression decreased
in moisture content at the center third by an average of 0 .73 per cent and in
crea sed at the ends by 0. 43 per cent . The data showed considerable scatter
and gave only a rough idea of the relative magnitude of the changes in water
content involved . They are , therefore , of use mostly from a qualitative stand
point. Nonetheless , the data do show definite moisture - content changes within
saturated clay specimens in undrained compression , thus substantiating the
contention that nonuniform volume changes occur in triaxial specimens sub
jected to axial compression .
Geuze and Tan (4 ) also show a comparison of water contents in identical
specimens of a pottery clay ( LL = 69 . 0 , PI = 39 .6 ) before and after the un
drained triaxial test . The central part of the specimen along the vertical axis
and the bulged portion of the specimen in the lower third show significant in
creases in water content ( 1 to 3 per cent ) at the cost of the outside portion in
the upper half of the specimen which did not bulge . These tests show a
pattern of water - content change similar to that obtained for the tests on sands
reported herein . The difference between the tests by Geuze and Tan and those
by Taylor and Clough may be explained in part by the fact that the latter took
horizontal slices across the sample , thereby averaging out any changes in
water content between the periphery and the center of the specimen ; further ,
the samples of Geuze and Tan bulged , whereas those tested by Taylor and
Clough developed distinct shear planes .
OTHER TESTS PROVIDING CONFIRMATION
Nonuniform conditions within triaxial specimens were further demonstrat

ed by a series of tests conducted by the Waterways Experiment Station from
1953 to 1956 . These tests were part of a study of pressure cell action and
involved triaxial specimens 70 in . high and 35 . 7 in . in diameter ( 1000 - sq - in .
cross section ) .
Specimens were constructed in a segmented confining jacket and were con
tained by a 1 / 32 - in . - thick rubber membrane . The soil material used was a
uniform fine sand placed air - dry . A grain - size distribution curve for the
sand used is included in Fig . 1. As with the small tests described previously ,
reasonably uniform density was obtained throughout the specimens by an in
crease of the compaction effort per layer on successive layers from bottom
to top . About 85 per cent relative density was attained in the specimens perti
nent to this presentation .
Confinement was accomplished by evacuation of specimens through a
porous concrete ba se plate , and vacuum was measured at both top and bottom
of the specimen to insure uniformity . A vacuum of 27 in . of mercury (13 . 26
psi) was used in all of the tests reported here . Load was applied to the
specimen by hydraulic jacks working against a large load - reaction truss .

The load applied was measured by a Baldwin load cell . None of the speci
mens were tested to a large percentage of strain (normally less than 2 per
cent ). Fig . 6 is a photograph of a large triaxial specimen and associated
equipment .
The large specimen size was adopted to permit placement of 6 - in .
diameter pressure cells at various locations within the mass . Pressure -cell
connecting cables were introduced through a buried chamber and box conduit
and through an aperture in the base plate . These cables were carefully
spiraled within the specimen to minimize adverse effects . Fig . 7 is a sche
matic diagram of the entire test equipment .
Of interest to this study are vertical stress measurements registered by
pressure cells within the specimens . Readings were made at a number of
radial positions at midspecimen height and at points just above the base . The
pressure cells used were of the WES type (5 ) , 6 in . in diameter , and were
Fig . 6 . - Large Triaxial Apparatus

352
A
IS
HANDCRANE MOUNTED
-
ON THEREACTION LOADTRUSS
FOR HANDLINGTHE FORMING REACTION
JACKET SEGMENTSANDHEAD LOAD
.
PLATE TRUSS
SAFETYHOOKFORHEADPLATE
-
LOADCELL HYDRAULICJACK
FLEXIBLE TUBE
EXTENSOMETERDIAL
SAFETY HOOK
HEAD PLATE
-
EXTENSOMETER
DIAL MANOMETER
SOIL
AIRTIGHT CONNECTIONS
ARE RUBBERMEMBRANE
MADEWHERETHEMEASURING
INSTRUMENTCABLESENTER THE SPECIMEN
CHAMBER USEOF PIERCED
BY
RUBBERSTOPPERS GACE SUPPORT
-
GAGESUPPORT
VACUUMPUMP RUBBERMEMBRANE
SURGE TANK CABLE OUTLET
CHAMBER
-
BASE PLATE
VV
.1
PANINIWANINI
-
FLEXIBLE COUPLING VACUUMLINEUSED PULL

TO
MEMBRANEAGAINSTFORMING
JAGKETDURINGCONSTRUCTION
OF SPECIMEN
-
CABLE ENTRANCECHAMBER
of
.
Fig
-.7 Schematic Diagram Large Triaxial Apparatus

placed with a minimum center - to - center spacing of 12 . 5 in . (four cell radii ) .

Since specimens were a nominal 18 in . in diameter , this necessitated use of
several test specimens to attain data for a given elevation within the speci
men at the desired number of offsets laterally . In consequence , a question
has been introduced regarding similarity between duplicate specimens which
may account for some scattering of measurements ; however , not to the extent
to prejudice the data . Cells were placed at midheight and just below the head
plate in addition to the lateral spacing mentioned . This is a vertical spacing
of over 10 cell radii and is more than adequate to avoid cell interaction .
Fig . vertical stresses measured at the various offset positions
8 shows the
for a 30 - kip specimenload (nominal 30 - psi average stress ) . In the upper
half of the figure is a plot of stress versus offset for points at midheight of
specimen , while in the lower half is a similar plot of points just below the
EDGEOF SPECIMEN
30
LINEOF SPECIMEN
CENTER
VERTICAL STRESS AT MIDHEIGHT
OF SPECIMEN
PSI
10 12
IN OFFSETIN INCHES
STRESS
OF
EDGE SPECIMENS
-
301
20
VERTICAL STRESS NEAR TOP

ENTERLINE OF SPECIMEN
OF
-
SPECIMEN
10
12
6
IN
OFFSET INCHES
Fig Distribution Stress Triaxial Specimen

of
in
8
.
.
-
head plate . In both cases significant variation in stress magnitude occurs
between the center and the edge of the specimen . At midheight , stresses in
the center exceed those at the edge ; whereas near the base , edge stresses are
larger . There is also considerable variation of stress vertically within the
specimen . This is in accord , at least quantitatively , with theoretical treat
ments by Filon (7) , D ’ Appolonia and Newmark (8) , and Pickett (9) . In both
plots , extensions of curves to the specimen edge , where data were not avail
able , have been made with these theoretical results in mind .
Some reservation should be stated here in regard to the quantitative validi
ty of these stress measurements . Overregistration of pressure cells , par
ticularly in granular media , might be expected to yield stress measurements
somewhat too large , and this may be a factor in the results being reported .
However , whether some pressure cell overregistration is involved or not , the
results indicate a distinct nonuniformity of stress throughout the specimen .
One large triaxial specimen was constructed in conjunction with the tests
of pressure cells which incorporated a number of 6 - in . strain gages placed
to measure vertical strains . These strain gages were of a differential
transformer type devised by personnel at the U . S . Army Engineer Ohio River
Division Laboratories (6) . Since the gages were a maximum of 2 in . in di
ameter , they could be placed much closer together without interference than
could the pressure cells . Thus , all comparative measurements could be de
termined on a single specimen .
The specimen containing strain gages was constructed , confined , and load
ed similarly to those incorporating pressure cells . Strains were measured
at a number of vertical positions and at various offsets within the specimen .
In all, 12 strain gages were used .
The strains measured by these 12 gages when the specimen had been load
ed to produce a vertical deflection of 1/ 2 in . at the top are shown in Fig . 9.
Strain is plotted versus offset position , and curves have been drawn through
the points for each elevation . The strong resulting pattern indicates that
vertical strains are small near the top of the cylinder and it may be presumed
they would similarly be small near the bottom . Measurements were larger
toward the middle , with the largest strains occurring just below midheight
(26 in . above base ) . This is about the same vertical position at which the
tests discussed in the earlier part of this paper showed maximum volume in
crease , and at which the " bulge ” typical of many triaxial test specimens
occurs .
The measured strains plotted in Fig . 9 at 26 - , 35 - , and 44 - in . elevations
also show lower values toward the edge than closer to the central axis . The
bottom two curves have been drawn through the single points for which re
sults are available , following a pattern indicated by theoretical treatments
( 7) , (8) , (9) . It is entirely possible , and perhaps even likely , that the lowest
curve in Fig . 9 should bend upward at the larger off sets . The gross pattern ,
however , clearly indicates a nonuniform distribution of vertical strains
throughout the specimen .
Comparative stress - strain curves for a large triaxial specimen (nominal
36 by
70 in . ) and a 2 .8 - by 6. 5 - in . laboratory specimen are shown in Fig . 10
as matter of interest . Material for both specimens was the dry sand for
a
which a size -distribution curve is shown on the left in Fig . 1. Confining
stress was 13. 26 psi. The large test used vacuum confinement , while the
small test involved the pressure confinement normal to most triaxial tests .
Several cycles of loading are plotted for each test . The patterns are strongly
similar except that the strain at equivalent stresses for the small specimen
is about twice that for the large specimen . This is probably an indication that
the large specimen is somewhat stronger , and the smaller specimen , at the
30 - psi maximum stress applied , is somewhat closer to failure than the large
specimen at the same loading .
CONCLUSIONS
The tests reported herein on dry sands show that in triaxial compression
there is a volumeincrease in the middle third , or failure zone , of specimens
over a density range from dense to very loose , and that a volume decrease
occurs at the ends of the test specimens . A similar change occurs in saturat
ed sand specimens subjected to axial strain under " constant volume” con
ditions . Qualitative tests on clay specimens in undrained shear show moisture
changes indicative of volume decrease in the zone of shear , and volume in
crease at the ends of specimens . Tests on large triaxial specimens of dry
26 IN. ABOVE BASE EDGE OF

SPECIMEN
MIDHEIGHT (35 IN.

PER CENT
IN
POINT
44 IN . ABOVE BASE
A
STRAIN AT
VERTICAL
53
IN
ABOVE BASE
|
CENTER LINE
0
.5
OF SPECIMEN
IN
ABOVE BASE
7
10
15
it
CO
IN
OFFSET INCHES
Fig Distribution Triaxial Specimen

of
Strain
in
.9
-
.
sand show higher than average values of both vertical stress and vertical
strain in the portion of the mass near the vertical axis just below midheight ,
where the maximum bulge is normally obtained in a triaxial specimen , with
lesser values both toward the edges and toward the ends .
These tests clearly indicate that measurements and interpretation of over
all volume changes , or lack thereof, of specimens in triaxial shear may not
be indicative of the conditions existing in the failure zone of the specimens .
By implication , if nonuniform volume changes are occurring , then nonuniform
stress and strain conditions also exist . This is directly established for both
stress and strain by the results of pressure - cell and strain - gage measure
ments large specimens . Thus , it is believed that the triaxial test in its
on the
present cannot be used as a rational basis for interpretation
form of soil be
havior in the field with respect to volume changes and stress conditions , al
though the measured strengths in the triaxial test may approximate field con
ditions . At best , the triaxial test can only be considered as an empirical tool
until such time as more can be learned concerning the test itself and the field
condition it attempts to duplicate .
30
20
PSI
IN 2.8-IN.-DIAM. X6.5-IN -HIGHSPECIMEN

AVERAGEAPPLIED STRESS
PRESSURECHAMBERCONFINEMENT
0.2 0.8 1.0

0
.6
IN
AVERAGESTRAIN PER CENT

IN 35
IN
DIAM.
.- 7
-
-
.
70
HIGHSPECIMEN
.0
x
VACUUMCONFINEMENT
0
.8
0
.2
0
.4
0
.6
IN
AVERAGESTRAIN PER CENT
Fig Comparison Large and Small

10
Stress Strain Curves

of
from
-
-
.
.
Triaxial Specimens
APPENDIX . - REFERENCES
1. U . S . Army Engineer Waterways Experiment Station , Summary Report of

Soils Studies . Potamology Report 12 - 2, October 1952 .
2. Chen , L . S., Stress Deformation and Strength Characteristics of Cohesion
less Soils. D . Sc . Thesis , Harvard University , December 1944 .
3 . Taylor , D . W. , and Clough , R . H . , Report on Research on Shearing Charac
teristics of Clay . Massachusetts Institute of Technology , June 1951.
4. Geuze , E . C . W . A . , and Tan , T. K ., « The shearing properties of soils . "
Geotechnique , vol. II, No . 2 (December 1950 ) .
5. Army Engineer Waterways Experiment Station , Pressure Cells for

U . S.
Field Use . Bulletin No . 40 , January 1955 .
6. U. S. Army Engineer Ohio River Division Laboratories , Final Report on
Development of Earth Strain Measuring Device . June 1953 .
7. Filon , L . N . G ., " The elasticequilibrium of circular cylinders under

certain practical systems of load . " Phil. Trans . , Royal Society , London ,
Series A , vol . 198 ( 1902 ) , pp . 147 - 233 .
8. D ’ Appolonia , E . , and
Newmark , N . M . , “ A method for solution of the re
strained cylinder axial compression . ” Proceedings , First U . S.
under
National Conference of Applied Mechanics , Am . Soc . Mech . Eng . (1951) ,
pp . 217 - 226 .
9. Pickett , G ., " Application of the

Fourier method to the solution of certain
boundary problems in the theory of elasticity . " Journal of Applied Me
chanics , Am . Soc . Mech . Eng . , vol . 11 ( 1944 ), pp . A - 176 - 189 .
STRESS -DEFORMATION AND STRENGTH CHARACTERISTICS OF A
CLAY COMPACTED TO A CONSTANT DRY UNIT WEIGHT
By A . Casagrande , 1 F . ASCE , and R . C. Hirschfeld , 2 A . M . ASCE
I. INTRODUCTION
This paper describes an investigation of the stress - deformation and

strength characteristics of a clay compacted in the laboratory by kneading
compaction . It is essentially an abridgement of Reference 1, but omitting the
following :
( 1) Tables giving the as -molded dry unit weight , water content , and degree
of saturation for each of the PH , PC , and PN tests described below .
( 2) Tables giving the as -molded dry unit weight , water content and degree
of saturation , time of loading , modulus of deformation , and stresses at
failure for each of the ő , Ř , and S tests described below .
( 3) Details of test procedure and computation of results .
(4) All stress - strain curves , except for some typical curves from each
test series .
(5 ) Appendix discussing pore - pressure measurements in triaxial tests .
(6 ) Appendix discussing use of vector curves for describing pore -pressure
build - up .
The tests were all performed on one type of clay which comes from the
Corps Engineers ' Canyon Dam Project in Texas . All triaxial tests were
of
made on specimens compacted in the Harvard miniature compaction device
to a dry unit weight of approximately 105 lb / cu ft , corresponding to 95 % of
the maximum dry unit weight for Standard Proctor compaction , and at water
contents ranging from 13 % to 19 % which is from about 3 % dry to 3 % wet of
Standard Proctor optimum water content .
This investigation consisted principally of complete series of triaxial Q ,
Ř , and S tests . A few direct shear tests of the S type were made for com
parison with the triaxial S tests .
Three types of special tests were made in which the pore - pressure build
up was investigated for various stress conditions , but in which no failure con
dition was produced . These tests were performed as follows :
( 1) Specimens subjected to increasing hydrostatic confining pressure , at

constant water content - PH tests .
( 2) Specimens loaded in a consolidation apparatus ( i . e . , with a lateral
strain of zero ) at constant water content - PC tests .
( 3 ) Specimens loaded in the triaxial apparatus with constant 01 / 03 ratio , at
constant water content - PN tests .
1. Prof . of Soil Mech . and Foundation Engrg . , Harvard Univ ., Cambridge ,

Mass .
2 . Asst . Prof. of Soil Mech . , Harvard Univ . , Cambridge, Mass .
359
II . NOMENCLATURE
Notation . -- The letter symbols adopted for use in this paper are defined in
Appendix I.
CLAY
III
CANYON DAM DESCRIPTION AND CLASSIFICATION TESTS
--
.
The clay was obtained borrow area for the Canyon Dam project
in
in
a
between depths
of
Texas was excavated from the horizon and
ft
It
9
A
.
-
.
The average As excavated this clay was
16
situ water content was
in
,
.
-
chunky and friable Before shipment Harvard had been dried about
it
to
to
.
water content and then mixed without removing any coarse particles
%
5
.5
.
The color this material the dry state very light tan and when
of
is
in
,
a
The liquid limit and the plastic limit
15
34
mixed with water darker tan
is
a
.
.
These values place the material on the plasticity chart high above the line
,
-
indicating that the colloidal fraction consists largely montmorillonite type
of
-
clay minerals Its toughness the plastic limit and its high dry strength
at
.
agree with these results

.
Fig are plotted grain size curves this clay which were determined
of
In
1
.
at the Waterways Experiment Station and Harvard For practical purposes
80 .
the entire sample passes No the material passes
of
40
70
mesh and
to
%
20 .
No 200 mesh Only about falls into the clay fraction finer than 002
%
0
.
.
particles larger than No
40
of
mm The small percentage mesh consists
in
.
.
Al
shells The specific gravity
of
part
of
71
the solids was found be
to
2
.
.
.
the grain size curves this material could be described

of
though the basis

on
clayey silt its physical characteristics are such that should be classi
as
it
, ,
a
fied as tough inorganic clay low plasticity CL

of
a
.
(
GENERAL INFORMATION CHARACTERISTICS

IV
ON COMPACTION
.
OF CANYON DAM CLAY

by
Full scale compaction both

Compaction means sheepsfoot
of
Field
,
1
-
.
.
sq
rollers 480 eight passes layers and rubber tired rollers

lb
in
in in
, .,
, ,
12 8
-
-
. .
- (
), )
/
wheel load four passes layers gave the same field

25
000
lb
,
(
compaction curve which shown Fig These results which were fur
. 8 of is
in
,
2
of .
.
nished through the Office the Chief Engineers were obtained from field
,
density determinations diameter by high undisturbed samples

10
in on
in
in
-
-
.
in .
high samples which were taken

on
Note Field tests diameter

in
2
2
,
:
(
.
-
thin wall tubes gave results which scattered excessively

,
. -
.)
Laboratory Compaction with Standard Proctor Tests The results of

2
.
--
two series standard Proctor tests performed the Waterways Experiment

at
of
Station are plotted Fig They yielded maximum dry density

of
111
in
,
. a
.3
.
.
and an optimum
of
15
1b
. cu
content water
%
ft
7
/
Laboratory Compaction
with Harvard Miniature Compaction Device
In
3
-
.
Fig
of
are plotted the results compaction tests carried out

of
nine series
3
.
with the Harvard miniature compaction device at different compaction efforts

.
The curve obtained using thirty tamps layers practi

20
on
of
be 10
each
is
lb
-
cally identical with the field compaction curve as can seen Fig
in
,
2
.
.
not possible obtain by means com

of
the Harvard device

is
However
to
it
a
,
paction curve which similar shape the Proctor curve This confirms
to
in
is
.
STRESS - DEFORMATION 361
100
Project 5810
80
Smallar
Determined at WES
Percont
Determined at Harvard
20
01 001
0
1
0
0
.
.
.
Dom
Grain Size
-
Fig Grain Size Curves

1
-
-
.
.
115
Project 5810 field Compaction
Standard Proctor
ft Campaction
/
–
cu
-
lb
Weight
Aarvard Miniature
Unit
Compaction
Thirty
20
tamps
10 lb
Dry
-
(
of layers
on
each
20
18
16
14
12
Molding Water Content percent

-
.Fig
and
of
Comparison Field Laboratory Compaction Curves

2
-
.
- - - - - - -
Moto: Compactionefforts
for individual curves are
Project 5810 indicated as follows :
"8- 20- 10' means"eight
20- poundtamps on aach
of 10 Layarı .
f-t
15-40- 10 6
/-
cu
1b
105
8-40- 10
• 10013
95%
Dry Unit Weight
1010
2020
-3060
4+
de
- -
-
85%
20
15
1064
-
-
20- 10
8
-
40 10- 100
-
15
10
10
-
-
10
10
8
-
-
85
12
14
16
18
20
22
Molding Mater Content percent
-
Fig Compaction Curves Harvard Miniature Compaction

3
-
.
.
Project 5810 Note All specimenscon

:
pacted Harvard
in
miniaturo compaction
device using com
a
paction et fort of
20
fifteen pound
-
10
on
tamps
of
ouch
Curing time day Layers
1
.
ft
/
cu
-
1b
Unit Weight
15
Curing time days

•
Curing time days

8
-
-
Dry
2022
12
16
18
14
-
Curing
on
Fig Effect Time Compaction Curve

of
of
4
-
.
.
STRESS -DEFORMATION 363
the findings by other investigators that kneading compaction and dynamic

compaction result in compaction curves having different shapes .
V. TEST PROCEDURES AND EQUIPMENT
1. Compaction . - The Harvard miniature compaction device which was

used this investigation is described in detail in Reference 2 . The diameter
in
of the tamper is 0 . 5 inch , and the dimensions of the cylindrical test speci
mens formed by means of this device are h = 2.82 in . and d = 1. 31 in .
In the course of a preliminary investigation it was found that the time
which elapses between mixing the air dry material with water and the com
paction of the specimens , i . e . , the " time of curing " , has a slight effect upon
the compaction curve, as can be seen in Fig . 4. For this reason a constant
time of curing of about 24 hours was used for all specimens . The curing was
carried out by letting the mixture stand in sealed jars .
Each triaxial test was started within a period of between 0 . 5 and 1. 0 hour
after compaction of the specimen .
A special series of tests was made for the purpose of determining the ef
fect of the cumulative compaction effort on the lower layers , due to using a
constant number of tamps on each layer throughout the height of the speci
men , and also due to the force required to extrude the specimen from the
mold . The results of these tests are summarized below :
( 1) For specimens on the wet side of optimum , no significant variations in

unit weight throughout the height of the specimen were found .
(2) For specimens compacted dry of optimum , extrusion from either end
produced a unit weight 2 to 4 % greater at the end where the extruding
force was applied than at the opposite end .
(3) When eliminating the extrusion process by using a split mold , no sig
nificant variations in unit weight were measured that might be due to
the cumulative compaction effort .
In view
2.
of these results , all test specimens were compacted in a split mold .
General Information on Triaxial Apparatus and Technique of Tests
The triaxial apparatus used in this investigation is of the same type as de
.
scribed in detail in Reference 2, and modified for pore - pressure measure
ments as described in Reference 3.
Commercial grade rubber membranes were used with wall thickness
averaging 0. 0025 in . The membranes were first soaked in silicone oil for a
period of at least one week in order to reduce leakage . Each membrane was
used only once . For the higher confining pressures two membranes were
used . The strength of all specimens was high enough to permit neglecting the
membrane correction .
The following details of technique of testing are common to all triaxial
tests made in this investigation :
( 1) The entire chamber and several feet of 1 / 8 in . O . D . Saran tubing con

necting to the air pressure reservoir were filled with tap water , as
discussed in Reference 4.
( 2) Stress - controlled load application was used as in preceding investi
gations , except for one series of tests in which strain control was used .
For tests in which pore - pressures were measured , each increment of
axial load was made about 10 % of the estimated failure load , and was
allowed to remain constant for a period of five minutes . Near failure ,

the size of the axial load increments was usually reduced to one - half ,
but each of these smaller increments was still allowed to act for a
period of five minutes .
(3 ) The procedure for measuring pore - pressures is identical with that de
scribed in Reference 3. The basic principle of this procedure consists
in finding a level in the capillary tube at which the meniscus neither
rises nor falls . At the end of consolidation in an R test , the water level
in the burette is lowered until the meniscus is in the capillary tube . An
increment of deviator load is applied to the specimen . Assuming that
this increment of deviator load induces an increase in pore - pressure ,
water will start to flow out of the specimen , and the meniscus will rise
in the capillary . In order to prevent the water from flowing out of the
specimen , the air pressure acting on top of the meniscus is increased
in one sudden increment . As a result of this air pressure increase
there will be an instantaneous drop in the level of the meniscus . This
instantaneous drop is due to the elastic expansion of the drainage con
nections between the porous stones and the capillary , and also due to
the compression of any gas bubbles trapped in the drainage line or be
tween the porous stones and the ends of the specimen . After the sudden
increment of air pressure is applied , the equilibrium level of the
meniscus should be lower than it was originally because of the ex
pansion of the lines and the compression of the bubbles . When the air
pressure acting on the meniscus is now held constant , one can observe
whether the meniscus is rising or falling . If it rises again , this means
that water is tending to flow out of the specimen -- the air pressure act
ing on the top of the meniscus is still smaller than the porewater
pressure . Then , additional increments of air pressure must be applied
until the meniscus no longer tends to rise under a constant air
pressure . If the meniscus continues to fall under constant air pressure
( after an increment of air pressure has been added ), it means that
water is flowing from the drainage system into the specimen - the air
pressure on the top of the meniscus is larger than the pore - pressure in
the specimen and must be reduced . For increasing pore - pressure the
equilibrium level of the meniscus in the capillary will continuously
decrease .
VI. SPECIAL TESTS FOR INVESTIGATION OF PORE - PRESSURE

BUILD - UP
Pore - pressure build - up in triaxial tests is influenced by many variables .

In order to investigate the mechanism of this build -up , it is necessary to re
sort to special tests in which the influence of these variables is separated .
In this chapter are described the results of three series of tests for
explor
ing the effects on pore - pressure build - up of certain simple types of loading .
The pore - pressure build - up in these tests was measured using the same pro
cedure as described above .
1 . Specimens Subjected to Increasing Hydrostatic Confining Pressure , at
Constant Water Content (PH Tests ). These tests were performed by subject
ing triaxial test specimens to an incremental increase of hydrostatic confin
ing pressure without allowing any change in water content . These tests are
designated PH tests ( P standing for pore - pressure , and H for hydrostatic con
fining pressure , 01 = 02 = 03 = 0c ) . The chamber pressure was increased in
approximately ten equal increments , and each increment was allowed to re
main constant for five minutes .
Eight PH tests were performed on specimens compacted as close as possi
ble to a dry unit weight rd of 105 lb / cu ft, with the actual yd values ranging
from 102 . 4 to 108 . 6 lb / cu ft . In these tests the water content was the variable
and ranged from 13. 7 to 18. 8 %, with a corresponding range of degree of satu
ration from 60 . 8 to 87 . 2 %. In Fig . 5 is plotted the pore - pressure build - up and
axial strain in these tests as a function of the hydrostatic confining pressure
oc
If we assume for a hypothetical specimen ( 1) 100 % saturation , (2) water
and solids incompressible , (3 ) for the time of such a test (approximately one
hour ) , that the pore - pressure changes due to secondary time effects are
negligible , and (4) zero pore - pressure at the start of the test , then a change
in hydrostatic confining pressure should immediately be reflected in an equal
change in pore - pressure . If Fig . 5 ( a ) this change would plot as the 45 - degree
line through the origin . If the test is started with a definite tension in the
pore water , e. g . as would be the case in 100 % saturated , undisturbed clay
specimens , we would still obtain a 45 - degree line , but with a negative inter
cept on the pore - pressure axis .
In Fig . 5 ( a ) it can be seen that with decreasing degree of saturation (at
approximately constant dry unit weight ) the deviation of the curves from the
45 - degree line increases . This can be readily understood from the following
consideration . As the hydrostatic confining pressure increases , this pressure
is at least partially transmitted to the trapped air which is reduced in volume
in accordance with Boyle ' s law , and which is partially dissolved in the pore
water in accordance with Henry ' s law . For a detailed description of this
phenomenon see e. g. J . Hilf (5) . 3 As a result , the degree of saturation rapid
ly increases . In tests PH6 , PH7 , and PH8 (the three upper curves in Fig . 5
(a ) , the degree of saturation of the specimens at the start of the test averages
87 %. Since the corresponding volume of air in the specimen is small , it is to
be expected that practically all of this air would be driven into solution under
a relatively small hydrostatic confining pressure , and that the curve of pore
pressure versus hydrostatic confining pressure at higher pressures would be
approximately a straight line with a slope of 45° . Fig . 5 (a ) indicates that the
curves do attain 45° slopes at a hydrostatic confining pressure of the order
of 5 kg / sq cm . On the other hand , for the curves in Fig . 5 (b), for which the
degree of saturation ranges between 60 and 70 % at the start of the test, it can
be seen that even at confining pressures between 10 and 14 kg / sq cm the slope
deviates substantially from 45 degrees , indicating that the degree of satu
ration is still significantly smaller than 100 %.
Theoretically , all test curves plotted in Fig . 5 (a ) should have a negative
intercept on the vertical axis , because negative pore - pressures exist in the
specimens as compacted . However , since the pore - pressure was measured
through coarse porous stones , it was impossible to measure negative
pressures .
From the plot of axial strain vs . confining pressure in Fig . 5 (b ) the follow
ing conclusions can be drawn :
3. Numerals in parentheses thus , J . Hilf (5 ) , refer to corresponding refer

ences in Appendix II .
Project 5810
PHB 18.8 86. 5
PH7 18. 3 86. 2
PES 17.9 87. 2
PH TESTS
POREPRESSURE
/
VERSUS
be HYDROSTATIC
CONFININGPRESSURE OPHS 76. 3
b 16. 9
-(
>
)
n
ION einssard
14 16 16
PBA 72
.0
.5
PR3 72
1
.7 .
.
PH2 60
.8
13
Po 62
.7
.5
Toot Water Dogno
BO Content of
Saturation
10
0 12
ng 14
16
6
/kg
Bydrostatic Confining pressure
a
(
)
-
kg
Dydrostatic Confining Pressure oga

(0
/
12 )
-
10
14
16
8
Project 5810
percent
PBA
-
HO
P5
PHO
PHD-
O
Axial Strin
PHLS
PH3
PH2
PR2
A
B
(
)
PH
TESTS
AXIAL STRAIN VERSUS
HYDROSTATIC
CONTINING
PRESSURE
Fig PH Tests
5
-
.
.
( 1) For low confining pressures , the compression of specimens compacted

on the wet side is greater than the compression of specimens com
pacted on the dry side , despite the fact that the wet specimens contain
much less air . This is obviously because the structure of the wet
specimens is weak whereas that of the dry specimens is strong .
(2) At high confining pressures the wet specimens are fully saturated and ,
therefore , since these are constant water - content tests , there can be
no further volume change . The total change in volume from the start
of such a test is approximately equal to the volume of air at the start
of the test . Since the dry specimens still contain air under the high
confining pressures , they still continue to compress . Since the total
air content of the dry specimens at the start of the tests was much
larger than for the wet specimens , the total compression of the dry
specimens is larger under the high pressures .
2 . Specimens Subjected to Increasing Non - Hydrostatic Confining Pressure ,
but Maintaining a Constant 01 / 03 Ratio , and at Constant Water Content . (PN
Tests ). – These tests were performed by subjecting triaxial test specimens to
an incremental increase of both 01 and 03 such that the ratio 01 / 03 was main
tained constant throughout the test, and without allowing any change in water
content . These tests are designated PN tests (P standing for pore - pressure
and N for non - hydrostatic confining pressure ) .
The chamber pressure was increased in approximately ten equal incre
ments , and simultaneously with each increment the load on the hanger was
increased such as to maintain the desired 01 / 03 ratio . Each increment was
allowed to remain constant for five minutes . The test results are plotted in
Fig . 6 showing pore - pressure as a function of the total major principal stress
01 .
In Fig . 6 (a ) are plotted the results of six PN tests on specimens compacted
as close as possible to a dry unit weight of 105 lb / cu ft , with the actual yd
ranging from 104 . 2 to 107 . 0 lb / cu ft . The water content was also intended to
be constant , with the actual range from 12. 8 to 13. 5 %. Six different 01 / 03
ratios , ranging between 1. 25 and 2. 50 , were used in this series of tests . All
observed points for these six tests fall into the narrow , shaded band shown in
Fig . 6 ( a ), indicating that the 01 / 03 ratio has a negligible effect on the relation
between pore - pressure and total major principal stress 01 for specimens
compacted at this water content .
In Fig . 6 ( b ) are plotted the results of seven PN tests on specimens com
pacted as close as possible to a dry unit weight of 105 lb / cu ft , with the actu
al yd ranging from 104 . 3 to 108 . 0 lb / cu ft . The water content was also intend
ed to be constant , with the actual range from 15 . 9 to 16. 5 %. Three different
01 / 03 ratios , ranging from 1. 25 to 1. 75 , were used in this series of tests .
When examining these curves in terms of average curves for each 01 / 03
ratio , we are again forced to the conclusion that this ratio has a negligible
effect on the results , at least up to 01 = 10 kg / sq cm . For higher 01 values
tests No . PN 12 and 13 begin to deviate from the general trend because in
these tests the combination of water content and 01 / 02 ratio was such as to
produce failure at the end points of these two curves .
In Fig . 6 (c ) are plotted the results of seven PN tests on specimens com
pacted as close as possible to a dry unit weight of 105 lb / cu ft, with the actu
al ya ranging from 105 . 7 to 106 . 8 lb / cu ft. The water content was intended to
be a constant , with the actual range from 18 . 6 to 19 . 1 %. Three different
Test No. 07/ 03

ama
/
ng
SERIESPN 1
PNI
PN2
1. 25
1.50
-)
kg WaterContent - 12. 8 to 13.5%
Dry Unit Weight - 104. 2 to 107. 0 lb/cu ft
PN3
PNA
1.75
2. 00
2. 25
(
u Degreeof saturation - 57.0 to 63. 3% PNS
PNG 2.50
Por Pressur
All observedpoints for
PN
PN
(A
)
to
Tosts 11.
PN
6
TESTS within the shadedband
POREPRESSUREVERSUS
TOTALMAJORPRINCIPALSTRESS
20
16
18
10
- 12
a 14
8
ng
/kg
Total Major Principal stress To
)
- 15
II . to16
3 to
Water Content
0 .5
.9
-
.2 Lb
cu
Dry Unit Weight 104. 108. ^

ft
0
%/
79
to
Degreeof Saturation 69
PN -
SERIES
/
aan
ng
-)u
75
02
PN13
50
PALO 762
1
0,
PNI2
.
-
1
kg
-
/
.
(
(B
)
PN
PNE
TESTS
25
03
1
.
-
POREPRESSURE
//
Pore Pressuro
VERSUS
PN9
03
50
1
.
--
-
-
//
16
10
) 12
cm 14
18
20
1
/kg
sq
Total Major Principal stress

0,
(
-
- 18
.8 to19
7 to
Mater Contant
to %
.6
1
-
.8 1b
cu
Dry Unit Weight 105. 106.

ft
8
84
%/
88
Degreeof Saturation
PN-
SERIES III
a
/
sq
24
.125
pano
-)
kg
,
/0
-
vas
(u
PN
TESTS
VERSUS
Por Presouro
POREPRESSURE
50
03
PNS
.
-
9
/
PN
20
PN19
75
03
PN18 Measuredpore pressure

1
4
S
-
.
/
.
not reliable
16
10
cm12
14
8
uq
/kg
va
Total jor Principal stress

(o
,
)
-
Fig
PN
Tests
6
.-
.
01 / 03 ratios were used in this series of tests , ranging from 1. 25 to 1. 75 .

Because the applied stress increments were excessively large, the specimen
in test No . PN 18 failed prematurely after application of the third increment ,
and the measured pore - pressures are not considered reliable . Again it may
be concluded that the 01 / 03 ratio has a negligible effect on the results . The
end points of the curves for tests No . PN 16 , 17 , 18 , 19 and 20 represent
failure conditions .
The PH tests , Fig . 5, may be considered as also belonging to the PN series
in the sense that they represent a ratio 01 / 03 = 1. 0 . In fact , when comparing
the curves in Fig . 6 with those in Fig . 5( a ) for corresponding water contents ,
one can see that the relationships between pore - pressure and total major
principal stress are similar .
3. Specimens Loaded in Consolidation Apparatus , at Constant Water
Content (PC Tests ) . - These tests were carried out in a specially built con
solidation apparatus in which drainage could be prevented and pore - pressures
could be measured . The vertical load (01) was increased in about ten incre
ments without allowing any change in water content . Each increment was al
lowed to remain constant for five minutes . These tests are designated PC
tests ( P standing for pore - pressure and C for consolidation apparatus ) .
Ten PC tests were performed on specimens compacted as close as possi
ble to a dry unit weight of 105 lb / cu ft , with the actual yd ranging from 101 . 8
to 108 . 8 lb / cu ft . The reason for this large range is that it is difficult to con
trol the compaction for such specimens with a thickness of only 0 . 5 in . and a
diameter of 4. 25 in . In these tests the water content was the variable , rang
ing from 12 . 9 to 18. 6 %. Pore - pressure and axial strain are plotted as a
function of the total major principal stress in Fig . 7 .
The pore - pressures in tests No . PC 7 and 8, plotted with solid circles in
Fig . 7 (a ) , were not measured by the standard procedure which was otherwise
used in this investigation . In these two tests the water level was kept at a
constant position in the capillary , and this was done deliberately in order to
observe the difference in the pore - pressures measured using this procedure
as compared to the procedure used in all other tests . Since the water
contents of the specimens in tests No . PC 7 and 8 were about 18 %, these two
curves should lie in the range of the upper three curves , namely for tests
No . PC 5 , 9 and 10. Instead there is an enormous difference which indicates
that the method of measuring pore -pressures by maintaining the water level
in the capillary at a constant position , is unsatisfactory when the time allowed
for each load increment is short .
When comparing the development of the strains in Fig . 7 (b ) with corre
sponding pore - pressure curves in Fig . 7(a ) one can see that for the three
uppermost curves (PC 5 , 9 and 10) the slopes of the curves indicate that for
pressures higher than about 6 kg / sq cm , the specimens must have been , for
practical purposes , fully saturated ( 45 - degree slopes in the pore - pressure
plot and horizontal slopes in the strain plot) . However , there is a discrepancy
for tests PC 4 and 6 , which indicate a state of practically full saturation ac
cording to the strain curves whereas the pore - pressure curves are still devi
ating from the 45 - degree slopes in the range of high pressures . Since errors
in pore - pressure measurements are usually such that the measured pore
pressures are somewhat too low , it is believed that the deviations of pore
pressure curves PC 4 and 6 , in Fig . 7(a ), from the 45 - degree slope in the
high pressure range are due to this error .
PC10 18. 6 79. 1

18.6 78.4
SUB ( A) O PCS 16. 5 80. 3
PC TESTS
POREPRESSUREVERSUS
/
be TOTALMAJORPRINCIPALSTRESS 18. 3
17. 3
78.8
79. 3
-(
BX
) 16. 5 76. 8
n PC7and PCB Measured
pore
-
pressures not reliable

.
Brod anslard
16
75
.8
.2
Test Water Degree
No. Content of
Saturation
18
73
. . .7
9 .6 0
66
.
13
27
12. 60
16
18 20
cm14
- 12
10
a
6
eq
kg
Total Hajor Principal stress
(0
,
)
14 can
ng
/kg
04
Total major principal stress

12 -
)
(
16
10
8
PC
TESTS
AXIAL STRAINVERSUS
percent
- PC3
Axial Strain
PC2757
selo tortas PC9
нарса
Somesoil squeezedout
[
aroundsides of loading
Po plate
Fig PC Tests
.7
.
-
The curves of pore pressure versus total major principal stress plotted
in Fig . 7 (a ) are similar to those plotted in Figs . 5 (a ) (PH tests ) and 6 ( PN
tests ), for specimens tested at comparable water contents .
4 . Conclusion . -- The investigations described in this chapter indicate that
the pore -pressure build - up in specimens compacted to the same dry unit
weight is principally a function of the water content (which in turn is related
to the degree of saturation ) and of the total major principal stress . The total
principal stress ratio has practically no effect on the pore - pressure build - up
when the latter is expressed as a function of the total major principal stress .
VII . O TESTS
1. Summary of Test Results . - A total of 33 @ tests were carried out in

four series , such that in each series the water content was approximately
constant. It was attempted to compact all specimens as close as possible to
the same dry unit weight of about 105 lb / cu ft . The average water content ,
dry unit weight, and degree of saturation for each series are listed in the
table below .
Series Number Average Water Average Dry Average Degree

Content Unit Weight of Saturation
pcf
QI 13 . 6 105 . 3 61 . 0
ÕII 16 .2 106 . 3 74 .8
ÕIII 18 .2 107 . 3 86 .8
ÕIV 19 . 0 106 .9 88 .8
2.Stress - Strain Curves for ő Tests . - Typical stress - strain curves ( for a
consolidation pressure of about 6 kg / sq cm ) are shown for each of the four
series in Figs . 8 - 11 . The maximum deviator stress is defined as the com
pressive strength of the specimen . The plotted pore - pressure is the pore
pressure (u - 40 ) induced by the deviator stress , i . e. , the total pore - pressure
u minus the pore - pressure uo produced by the hydrostatic confining pressure
oc
When comparing all stress - strain curves within each series of tests (which
are plotted in Reference 1) the following relationships are noted :
( a ) Series QI (ave w = 13 .6 %). - The strength increases rapidly with in

creasing confining pressure Oc . This is even more clearly seen from the Q
envelope in Fig . 13 (a ) .
The induced pore - pressures (u - 40 ) are negligible for ő tests performed
at small confining pressures Oc. For confining pressures of the order of
14 kg / sq cm , the induced pore - pressures at failure are of the order of magni -
tude of one - half of the compressive strength .
(b ) Series QII (ave w = 16 . 2%) . — The rate of increase of strength with
confining pressure is less than in Series I , as can also be seen from the Q
envelope in Fig . 13 ( b ) .
The induced pore - pressures at failure are approximately 20 % of the com
pressive strength for small confining pressures , and for large oc values they
-
sq
Stress kg
-
cm
Stress kg sq
372
0
1
2
3
/4
5
6
10
-
12
6
/8
come
14
Project 5810 Project 5810
-
kg
240
/ sq
cm
Mgo
+ Deviator Stress
-
03
.)
-
Induced
102
150
Pore Pressure
kg
/
|
u
-(.
sq
cm
percent
+
Deviator Stress -
percent
-
-
- Induced Od
101
Pore Pressure
Strain
Strain
SERIES
II
7
SERIES
kg
eq
00
./ com
To
33
-.*.6-ti
min
/
kg
og
en
00
.6- -
11
t
min
90
)
(
-.8
Q12
04
.
–.9
-
.
( Test
)
Fig Stress Strain Curves Test Fig Stress Strain Curves
/
-
sq
kg
sq
/
-
cm
kg
Stress stress
cm
3
Project 5810 Project 5810
-
|-
Mso
kg
sq
/.
180
com
-
Deviator Stress
03
01
)
-(
Mgo
-
Induced
25
Pore Pressure
kg
/
)u
-u(
sq
cm
percent
percent
+
- Deviator Stress
-
07
03
-(
)
Induced
Strain
Pore Pressure
Strain
)u
-u(
STRESS - DEFORMATION
5
7
IV
SERIES III SERIES
/
.6
sq
0
kg
05
/
.6
cm
sq
kg
08
cm
- -ty
100min
-=65
min
373
)
(
11
-
Q31
-.
Test
.
Curves
)
Strain
(
Fig Stress
-
10
-.
.
Fig Stress Strain Curves Test Q22
600
Project 5810
500
II .2 / %
16
=
v
7
/
6%
come
. 13
Average
--
sq
-
SERIES
=
--
w
kg 300
Ö
--
Average
--
M50
SERIES
200
Average
w
=
18
2.
SERIES
%
III
b
100
.0%IV 19
•
w
Average
Ö
SERIES
16
12
14
10
come
6
kg
sq
Hydrostatic Confining Pressure
(0
/
)
-
Fig Deformation versus Confining

12
Modulus Pressure Tests

of
Q
-
.
.
)
increase the order of magnitude the compressive strength For large
of
to
confining pressures the shapes the stress strain curves and the pore
of
pressure curves are quite similar

on
those obtained tests undisturbed

to
in
Q
samples normally consolidated medium sensitive clays The stress

of
strain curves this series have characteristic abrupt bend

at
about
to
in
2
a
strain sort of yield point particularly at the higher confining

%
3
a
,
pressures whereas most tests very

at
other water contents there

at is
in
,
a
)
gradual change
up
of
curvature the maximum deviator stress strains

to
. in
the order of 20
%
respectively
18
19
Series QIII and QIV ave and these

%
In
=
w
,
.2
.0
c
-
).
(
)
two series the strength increases only very slightly with

oc
as
be
can seen
,
clearly from the envelopes Fig

13
and
in
Q
(d
(c
.
).
)
The induced pore pressure failure increases for the larger oc values
at
-
one half the compressive strength but then drops off slight
of
the order
to
it
;
-
by
ly
for large strains This drop may be explained the tendency for com
.
pacted material small effective stresses expand similar manner

at
in
to
to
a
undisturbed clay specimens tested under small effective stresses Since

is .
swelling not permitted tests this tendency for expansion reflected

is
in
Q
by drop the pore pressure

in
a
.
-
Fig plotted the secant

12
Modulus Deformation M50

is
of
In
d
-
).
.
(
)
deformation M50 corresponding the compressive

50
modulus
of
of
to
%
(
of
strength as This modulus increases with decreasing water

Oc
function
a
)
increases with increasing

Oc
The total range

of
content and M50 these

in
it
,
.
STRESS - DE FORMATION 375
De
jater
Project 5810
13
(A)
AverageWaterContent
%
. -
.6
averago content
/
com
sq söz
-1)
kg SERIES
5
( CIRCLES
O
Shearstros
10
/ 12
14
16
18
20
B
cm
kg
sq
Total NormalStress
(0
)
-
cm Project 5810
16
/
%
AverageWaterContent
.2
(8
-
)
sq
-
kg SERIES 11
0
() CIRCLES
O
T
ShearStress
đi
ở
|
*
*
/ 12
20
24
16
18
2
cm
Total NormalStress kg sq
(a
)
-
cm Project 5810
/
1B
sq AverageWaterContent
%
2
-
kg
-)
(C
)
( SERIES 111
7
T
CIRCLES
O
by
ShearStross
Specimentailed buckling
az
azz
õio
en
lazo
y
14 1
il
r
2 I
10
/ 12
10 20
cm
sq
Total Normalstress kg
(0
cm
)
-
/
sq
Project 5810
kg
-)
19
Averagewater Content
%
Tv D
.0
(
-
)
(
1 SERIES
7
CIRCLES
Shearstress
120
220
oso
14 an
an
X
n
cm
sq
Total NormalStress kg
(3
)
-
/
Fig
13
.
tests was between about 20 and 550 kg / sq cm . Fig . 12 does not show a con
sistent pattern , although a general trend is established . It is believed that a
major reason for the irregularities in these results may be the dependence of
shape of the initial portion of the stress - strain curves on the time elapsed
between the compaction of a specimen and the start of the triaxial test , which
ranged in these tests between about 30 and 60 minutes . Since the M50 is de
fined by the first portion of the curve which in turn is determined by the first
3 to 5 points , the total elapsed times to the points defining the M50 values
range between about 45 and 85 minutes . It is known from other investigations
on the stress - deformation characteristics of remolded clays that such a range
of elapsed periods from the moment of actual remolding has a decided effect
on the shape of the stress - strain curves .
3 . Q Envelopes . - In Fig . 13 are plotted the strength circles in terms of
the total stresses ( Q circles ) and the Q envelopes . Fig . 14 is a summary
plot showing all of the Q envelopes . Comparing the envelopes for the test
series with the highest and lowest water contents , one can see that for the
high water content (Series IV ) the envelope is horizontal for confining
pressures above about 3 kg / sq cm , whereas for the low water content (Series
I ) the envelope is still rising slightly at the highest confining pressures used
in this investigation . From this comparison it may be concluded that speci
mens compacted at the high water content were for practical purposes fully
saturated under confining pressures above 3 kg / sq cm , and that the volumes
of these specimens did not decrease with further increase of confining
pressure ; whereas the specimens compacted at low water contents could not
have been fully saturated even under the highest confining pressures , and that
their volumes were decreasing with increasining confining pressures , with
corresponding increases in effective stress and strength .
4 . Vector Curves for Q Tests . -- In Fig . 15 are plotted the vector curves
for all Q tests . They were computed assuming that the failure plane is in
clined at 600 to the major principal plane, assuming a friction angle of 30
degrees . From the results of S tests , presented later in this paper , it was
determined that the S line is inclined at approximately 30 degrees . The end
points of the vector curves in Fig . 15 lie well above the 30 - degree line .
5 . Analysis and Discussion of Results of Q Tests . – The variations in
strengths of these compacted specimens as a function of water content and
confining pressure are analyzed with the help of several special plots which
are described below .
In Fig . 16 are plotted curves , derived from the Q envelopes , representing
the shear strength as a function of total normal stress , i . e. , the shear stress
and total normal stress on the 60 - degree planes (the assumed failure planes )
for circles that are tangent to the Q envelopes . In addition , similar curves
are interpolated for a sequence of water contents . When examining Fig . 16 ,
it should be recalled that the optimum water content for standard Proctor
compaction is approximately 16 %, and that these Q tests were made on speci
mens which all had approximately the same dry unit weight , viz. about 105
lb / cu ft , corresponding to 95 % of the maximum dry unit weight for standard
Proctor compaction . Therefore , the interpolated curves in Fig . 16 apply only
to specimens compacted to this dry unit weight . The most significant result
which Fig . 16 illustrates is that for low water contents the strength of the
material increases rapidly with increasing confining pressure , whereas for
high water contents there is only a small increase in strength for the low
10
Project 5810
a
/
vq
-)T
1
kg Series
(
ShearStress
Series 11
7
Sarios 111
7 7
IV
Series
18
20
10
kg 12
14
16
8
com
nq
Total NormalStress
(0
)
-
Fig /
Envelopes Summary Plot
14
Q
-
.
.
/-a
13
vq Project 5810 AverageMaterContent

%
.6
-
kg
(
Plano Tool
A
(
)
Envelope
7
SERIES
0
1
VECTORCURVES
60°
on
shear stress
18
o 12
a 14
16
S
ng
kg
on
Bffective momalstress 60° Plane

(
)
-
/
a
/-
16
ng Project 5810 AverageMater Content

-
)
kg
30
(
(B
)
plane TO
Envelope
Ö
SERIES
7
II
VECTORCURVES
600
on
Hour Stress
2013
fou
doua los 1016

o
2
a 7
8
ng
kg
on
Effective Mormalstress 60°

-
/
Fig
15
and
(b
(a
.
)
/-
an
ag
kg
Project 5810 Average Water Content - 18. 2%
(C)
III
30°
Plane 1760
SERIES
7
VECTOR CURVES
727
econocoa
60° 1826
on 020 1823
Shear stress
eza
725
Ha
zz
4
.0
ng3
a 5
3
) .0
2
.5
2
.0
1
5
.0
1
R
.
0
.5
/kg
on
Effective Normal Stress 60° Plane löco
-
19
Average water content
%
=
Project 5810
.0
'
/-
can
D
(
)
sq
IV
)
ky SERIES
Ö
-
Xoooooo
VECTOR CURVES
Dobra
(
TO
60° Plane
on
Shear stress
031
o29 730 33
1
4
1
6
1
.0
1
/kg .2
.6
0
8
.4
0
0
2
.
.
.
cm
sq
on
Bffective Normal Stress 60° Plane loco

-
)
Fig
15
and
d
)
)
(
.
(c
.. . .. . 2. 0.
=
% 0
6 % %
5 %2
0 0
% % %%
13
= -
14 14
- - ==
16 16 17
=
18 19
18
W w W
5810
Project
16
60° Plane
on
/
com
14
sq
Stress
kg
Normal
12
Plane
Total
60°
10 on
versus
Stress
Plane
8
Normal
60°
on
6
Total
Strength
4
..
Shear
—
16
2 Fig
ana
kg
sq
on
Plane 60° Strength Shear

/
-
380
low confining pressures and no increase in the range of higher confining
pressures .
In Fig . 17 are plotted the shear strengths corresponding to the horizontal
portions of the curves in Fig . 16, which for practical purposes are the ordi
nates along the right margin in Fig . 16. In addition , the actual range of water
contents for each series of Q tests is shown in Fig . 17 .
In Fig . 18 is plotted the pore - pressure as a function of total major princi
pal stress during the axial load increase for each Q test . The starting point
of each curve is the pore - pressure that was induced by the hydrostatic con
fining pressure . The failure point is shown by a solid circle . It is well to
compare this plot with Fig . 5 (a ) in which the pore - pressure is plotted as a
function of hydrostatic confining pressure . If one were to draw curves
through the starting points of the curves for tests at the same water content
in Fig . 18 , one would obtain curves of pore - pressure versus hydrostatic con
fining pressure , i. e. curves the same as those plotted in Fig . 5 (a ) . A detailed
comparison of Fig . 18 with Fig . 5( a ) shows , in fact , good agreement. On the
other hand , when connecting the solid circles in Fig . 18, for tests performed
at the same water content , one obtains a set of curves which represents the
relationship between the pore - pressure and the total major principal stress
at failure . These curves are drawn in Fig . 18 for the average water content
of each of the four series of tests . Finally , the failure points ( solid circles )
como
/
sq
-
kg
MaximumShear Strength
Range
of
water content
and average water content
13
14
15
15
16
17
18
19

-
Fig Shear Strength versus Molding Water Content

17
Maximum
–
.
.
STRESS -DE FORMATION 381
5810
.6
%
13 28
=
w
Project
26
Average
.
2
%
16 24
=
.
w
2
%
Average
22
.
18
0=
%
Stress
w 7
19 20
=
/-
Average
com
w
Principal
sq
18 kg
),0(
Average
Major
16
stress
Total
14
Principal
versus
14
=
-7
°c 12
Major
Pressure
10
Total
-.
Pore
8
6
.
18
Fig
4
2
= 2
0
cm
sq
kg Pressure Pore
u
/
-
)
(
Fig
all
tests performed the same confining
18
were connected for
at
in
.
pressure oc but giving
of
various water contents
at
series inclined
a
straight lines
.
Fig
of
18
In the intersection curve for constant water content with
a
.
straight line for given oc value defines the total major principal stress and
a
the pore pressure
or
of
failure for this combination
at
water content and
.
-
From these values one can also compute the compressive strength and the
effective principal stresses at failure lines Fig
18
These two sets indi
in
of
.
.
orderly variation the pore pressure and total major principal
an
cate
of
-
stress failure as function water content and on By interpolation one
of
at
,
.
can plot the two sets
of
lines shown Fig This diagram
19
probably the
is
in
.
best summary presentation
of
of
the results
of
the four series tests Again
,
.
emphasized that the relationships shown this figure apply only
it
is
in
to
specimens compacted dry unit weight
of
this were
cu
to
about 105
of lb
If
ft
a
.
the minimum unit weight required
an
the construction earth dam then
in
,
Fig would permit the estimation
19
of
the strength
of
the material and the
.
pore pressure water content and total major

at
failure for any combination
of
-
principal stress which for the period during construction usually taken as
is
(
the weight overlying material

of
)
.
VIII TESTS
Ř
.
Stress Control Tests with Back Pressure Saturation

R
A
-
.
of
Specimens
kg
back pressure
sq
cm
ap
Saturation was
of
about
1
A
.
-
.
6
-
/
plied the porewater during the consolidation phase tests for the of
to
these
Ř
purpose increasing the degree of saturation
of
Ř to
100
.
Summary tests with stress control

of
of
21
Test Results total

A
2
.
.
-
-
loading and back pressure saturation were performed three series such
in
,
-
the specimens as compacted was
ap
of
that each series the water content

in
proximately constant was attempted compact all specimens as close as

to
It
.
possible the same dry unit weight The average as

, of
cu
about 105
to
ft
lb
of /
molded water content dry unit weight and degree saturation are listed
,
in
the following table
.

pcf
%
ŘI
13
106 63
8
0
.5
.
.
RII
16
87 76
107
3
.6
.
RIII 19 106
0
.0
.
Stress Strain Curves for Stress Control Tests Typical stress

R
3
-
.
effective consolidation pressure of about

cm
strain curves for

an
kg
sq
8
(
are shown for each Figs

of
the three series

20
22
in
The maximum deviator

.
.
-
stress defined as the compressive strength From comparison of all the

is
a
.
stress strain curves which are plotted Reference the following obser
in
1
-
vations can be made

:
For small consolidation pressures only negligible pore

(a
’,
o
)
pressures develop during the axial load increase regardless the water
of
,
28
26
)
Failure
24
22
(
at
Stress
20
/
cm
Principal
sq
,)(o-
.
5
0. 18 kg
.
x17
5 18X
Major
|
0. 18
19 16
Total
Stress
versus
14
Principal
Pressure
12
Major
10
Pore
Total
of
a
/kg
sg 8
Interpolations
-
3 6
-.
4 .
19
Fig
cm sq / kg - Failure at Pressure Pore

-
sq
kg
sq
Stress
kg
-
cm
cm
Stress
384
/4
3
6
5
/4
0
3
2
/
1000kg
sq
cm
sq
/ kg
-
cm
M50 1500
Project 5810
–
Project 5810 Deviator Stress
)0
-(
02
Deviator Stress
oz
!
,-
10
Induced
Pore Pressure
)0
:0(
-
Induced
Pore Pressure
U
u
.-
(
percent
percent
- -
Strain
Strain
R
SERIES II
SERIES RI
sq
kg
/
70
cm
sq
kg
/
=
26
.8- -
com
01 ty
min
55
.7 -, 70
0t
min
)
.
(
Ř7
R13
-
No
Test No
)
-.
21
.
Curves
.
Strain
(
Stress
-
Fig
-.
Test
. 20
Fig Stress Strain Curves
Stress - kg / sq cm
6 8 10 12 14
Project 5810
*so • 1100 kg / g ở
Deviator Stress
10, - og
percent
Induced
-
Pore Pressure
(u - )
Strain
SERIES III
R
crne
/kg
sq
50
O
7
70 .
'
-
ty
min
-
No
Fig
22
Stress Strain Curves Test R20

(
)
-
-
.
.
which the specimen was compacted The consolidation pressure

at
content
(
.
equalthe chamber pressure oc minus the back pressure

uo
os For
is
to
).
-
high confining pressures the pore pressure at failure all tests was about
in
,
the compressive strength

of
75
%
For the tests with low and medium water contents Series and
II
,
(b
)
(
)
the stress strain curves are almost straight lines strains approaching
to
-
failure The failure strains are exceptionally small ranging between one and
,
.
two per cent , reflecting the brittleness of these specimens . For tests with
high water content (Series the stress strain curves are gradually curved
III
,
,
)
). -
reaching failure at high strains
15
20
to
%
(
deformation M50 corresponding the com
of
50
The secant modulus
of
to
%
pressive strength plotted Fig
23
of
as function
is
the consolidation
in
a
.
pressure on can be seen that this relationship has some similarity the
to
It
.
envelopes that can be drawn circles Fig For consolidation
24
the
sq to
in
R
.
.
pressures larger than kg cm the modulus increases approximately
in
8
/
direct proportion with or However the ratio M50 0s increases with lower
/
.
consolidation pressures The water content at which the specimens were
.
compacted does not seem have an important influence
on
the M50 values
to
obtained these tests although the failure strains differ greatly
in
, .
contrast with the corresponding plot for the Fig
12
tests the results
In
of ,
.
Fig
of
23
are much more consistent the discussion the modulus
In
in
of
deformation for the tests was pointed out that variation the time which
it
Q
of
specimen and the start axial load increase
of
elapsed between compaction
a
may have been responsible for the erratic results However the results
in
,
.
plotted Fig 23 the relative variations the total time which elapsed from
of in
in
of ,
.
the compaction specimen the start axial load increase was very
to
a
small Also the stress strain curve for much more sensitive
is
test
to
Q
a
,
-
.
slight variation the molding water content than the test which the
in
in
in
R
specimen hydrostatic confining pressure before the
is
consolidated under
a
application
of
deviator stress
.
I
3000
Project 5810
13
%
Average Molding Water Content
5
-
.
Series
Ř
2500
16
II
Average Molding
%
Mater Content
3
.
-
Sorie
Ř
:
2000
a
/
vq
-
kg 1500
M30
1000
f
500
Series III
+
19
Average Holding
%
mater Content
.0
-
10
12
6
ng
kg
Effective Consolidation Pressur

a
-
Fig
of
Modulus Deformation versus Effective Consolidation

23
-
.
.
Pressure Tests
(R
)
4. Circles for Stress - Control Ã Tests . -- In Fig . 24 are plotted the

R
strength circles in terms of total stresses .
The R lines were drawn tangent to the circles for the highest confining
pressures which are believed to be sufficiently high so as not to be affected
to any significant degree by compaction prestress . For the lower confining
pressures , the circles rise above the R lines , indicating a prestress effect
due to compaction .
The inclination of the R lines for high confining pressures is greater for
the high water content tests . A comparison of the slope angles and of the
water content is listed below .
Series Number Average Water Content of Slope Angle

Specimens as Compacted of R Line
O
(degrees )
RI 13 . 5 16 . 7
RII 16 . 3 17 .2
RIII 19 .0 19 .6
5 . Vector Curves for Stress - Control R Tests . -- In Fig . 25 are plotted the
for
vector curves the tests with stress control loading and back pressure
R
-
They were computed the assumption that the failure plane
on
saturation
is
at .
the major principal plane assuming friction angle
of
inclined 60°
to
a
,
.,
i.
30°
Fig
were drawn through the origin tangent
25
lines the circles

to
Ř
Ř
,
,
.
corresponding the highest confining pressures which are believed
be
to
to
(
sufficiently high as not any significant degree by com

so
be affected
to
to
paction prestress The slope angles lines and the average water
of
the
R
)
.
contents are compared below for each series

of
tests
.
Series Number Average Water Content of Slope Angle

Specimens as Compacted of Line
Ř
degrees
%
)
RI
31
13
5
2
.
RII
31
16
.8
3
.
RIII
36
19
5
0
.
.
large difference between the slopes

, ais
of
There
on
the and series the

II
I
lines see pre

on
series
III
one hand and the the other hand both for the
R
,
ceding section and the lines Related this difference appears be the
to
to
Ř
.
)
shape are basically different

of
the stress strain curves which for series

III
-
than for Series and

II
.
of I
The shape the vector curves for specimens compacted

at
low water
content Series RI quite different from the shape for specimens compacted
, is
,
high water content Series KIII This will the following

be
discussed
at
in
.
section
.
Analysis and Discussion

of
Results for Stress Control Tests The

Ř
Ř
6
.
-
-
.
at
tests show that the water content

at
which the specimens are compacted

,
equal dry unit weight has important effect

an
on
the stress deformation and

,
-
Project 5810
a SERIESR1
/
ng
R CIRCLES
-)
kg
(
T
13
AverageMoldingWaterContent-
%
.5
ShearStress
16
.
Lafail
_
10
/ 12
14
16
18
20
6
ng
Total Monal stru kg
(0
a
)
-
Project 5810
(B
)
/-
11
SERIES
R
ng CIRCLES
R
)(1
kg
16
%
.3
shearstress
Rial
iis
14 L
16
18
20
10
/ 12
8
ng
Total Hol stres kg

(a
a
)
-
10
Project 5810
c(
)
a SERIES III
Å
/
sq CIRCLES
R
-) a
kg
19
*(
%
.0
ShearStress
024 Val
iro
Bus R06 R18 R19 R21 R20
20
. 10
/ 12
14
16
1
ng
Total formalstres kg
a
(o
)
-
Fig
24
.
Project 5810
6
-
(A )
kg SERIES Ã IL K Line
)
VECTOR
CURVES
31.
7092 AverageMoldingWaterContent
13
%
.5
|
-
60° Plane
aceto
on
ShearStross
10
12
sq14
16
18
B
20
on
cm
Ettective lonal stress 60° Plan
g
/
Project 5810
s
/s Line
Ř
(
)B
SERIES
IL
|
VECTORCURVES
IT AverageMolding
16
Content-
%
.3
60° plano
on
i
shearstru
Suz
13
12
a 14
16
18
20
ng
kg
on
Effective Mommal
stress 60° plano foco)
-
/
Project 5810
koga
SERIES III
I
VECTOR
CURVES
19
AverageMoldingMaterContent-
%
.0
60° Plan
on
shearstress
16 Ruiz 19 R19 R20 R21

10
ng14
12
16
18
2
on
/kg
Betective tonal struc 60° Plane

a
.
Fig
25
.
strength characteristics . The most surprising result is the large difference
in the slopes of the R and R lines when comparing the tests for low and high
water contents . For the high water contents the slope angles are substantially
greater than for the low and medium water contents . The possibility that ex
perimental errors are responsible for this difference is discussed below .
( 1) Failure Strains . - At high failure strains the membrane and end re
straints tend to increase the strength of specimens above the true value .
Since high failure strains developed in the tests with high water content , this
might be a possible cause of error . Although most of the medium water
content tests show failure strains of the same high order of magnitude (15 to
20 %) as the tests with the high water content , the slope angles for these tests
are closer to the results obtained for the low water contents for which the
failure strains were chiefly in the range of one to two per cent .
(2) Technique of Measuring Pore - Pressures . - Identical technique for
mea suring pore - pressures was used in all tests .
( 3) Differences in Pore Pressures Between Middle and Ends of Speci -
men . -- In the low and medium water content series the measured pore
pressure increased rapidly near failure , whereas in the high water content
series it stayed practically constant over a substantial range of strain .
Therefore , at least for the high water content series , the measured pore
pressures are believed to be representative for the pore - pressures in the
middle of the specimen , and the slope of the R line should be reasonably
correct . On the other hand , for the low water content tests it is conceivable
that the measured pore - pressures are smaller than the actual pore - pressures
in the middle of the specimen . This would tend to make the slope angle of the
Ř line too small . If this factor were significant, its effect could be substanti
ally reduced by allowing much more time for the last strain increments prior
to failure . This was actually done in strain - control tests which are discussed
under a subsequent heading . A comparison of the failure points of these
vector curves with those in Fig . 25 shows excellent agreement, proving that
the substantial additional time allowed in the strain - control tests not in did
fluence measurably the envelopes all water contents
at
R
Membrane Leakage Any significant membrane leakage would nor

(4
)
.
--
mally show during the two days during which the specimen
up
allowed to
is
consolidate Because the subsequent axial load increase has duration of

a
.
only about two hours leakage that was small that was not de
so
of
rate
it
a
,
tected during consolidation would not be significant However significant

if
a
,
.
leak should develop during the consolidation phase then depending the lo
on
of ,
the specimen the pore

of
the leak near one end the middle

or
cation
in
),
is (
pressure which the end of the specimen may be either larger

at
measured
or smaller than the pore pressure the middle of the specimen Such
in
-
errors would cause results which deviate erratically from the general trend
.
believed that the shape the stress strain curves not subject
is
of
to
It
is
any important errors Therefore the fact that basically different shape re
to a
,
.
sults for the high water contents as compared the low and medium water
contents may be considered as indirect supporting evidence that the differ
,
the slopes lines reflect actual differences and are not

of
ences the and

in
caused by errors
.
Strain Control Tests with Back Pressure Saturation

B
Ã
-
-
.
Purpose and Technique

of
Tests the stress control tests the

In
1
,
.
.
--
specimens compacted low water content failed abruptly

at
Therefore
,
.
strain - control tests were performed in order to investigate the shape of the
stress - strain curves in the range of failure , and particularly after failure .
The rate of strain was about 0. 15 % per minute up to and past the maximum
deviator stress ; and then the rate was increased to about 0. 30 % per minute.
In these strain - control tests the deviator stress built up so rapidly at the
start of axial loading that the pore - pressure could not be measured with suf
ficient accuracy at the small strains . However , for large strains the deviator
stress remained essentially constant for a sufficient length of time so as to
ensure that the pore - pressure measurements near the end of each test are
reliable .
2 . Summary of Results of Strain - Control R Tests . - A total of six Ř tests
with strain - control and back - pressure saturation were performed in three
series with different water contents . In each series two tests were per
formed , with consolidation pressures of approximately 2 kg / sq cm and 8 kg /
sq cm . The average as -molded water content , dry unit weight and degree of
saturation for each series are shown in the table below .

pcf
RIV 13 .0 105 .8 58 .7
RV 16 .2 105 .2 72 . 2
RVI 18 .8 106 .7 86 . 7
3 . Stress - Strain Curves for Strain - Control R Tests . – The deviator stress
and induced pore -pressure are plotted as a function of strain in Fig . 26 for
the two tests in Series RIV , in Fig . 27 for Series RV , and in Fig . 28 for
Series RVI. As explained in the preceding section on “ Technique of Test ” ,
the measured pore - pressures for low strains (less than about 5 %) are con
sidered inaccurate . The stress - strain curves for the low and medium water
contents (Series IV and V) are practically straight lines to strains approach
ing failure ( of the order of 0 . 5 to 1. 0%), similar to the results obtained by
means of stress - control. For the high water contents ( Series VI) the stress
strain curves are also nearly straight lines , but only up to deviator stresses
between 60 and 80 % of the strength . Then the strains increase , first rather
abruptly , and then with a further gradual increase in deviator stress until
failure is reached at strains of the order of 20 %.
A characteristic of all these stress - strain curves appears to be a " double
failure ” , namely a " yield strength " at a very low strain which develops im
mediately following the straight - line portion of the stress - strain curve , and
an " ultimate strength ” that develops at very high strains . In some tests , e. g. ,
in Fig . 26 (a ) , particularly for low water content and low confining pressure ,
the yield strength is greater than the ultimate strength . For the same water
content , but with higher confining pressure , Fig . 26 (b ) , both strengths are
about equal . It is conceivable that for still higher confining pressures , but
with the same water content , the ultimate strength would be greater than the
yield strength . For the intermediate water content , Fig . 27 (a ) and (b) , the
ultimate strengths were slightly greater than the yield strengths for both the
low and high confining pressures . For the high water content tests it can be
seen in Fig . 28 (a ) and (b ) that the yield strengths are between 60 and 80 % of
the ultimate strengths .
-
sq
kg
Stress
sq
kg Stress
cm
cm
392
/2
0
1
2
3
/4
5
6
sq
=
kg
/
cm
=
M50 1500
/
1600kg
sq
cm
Woooopas
Project 5810
Induced Project 5810

Pore Pressure
-u
)u
(
Deviator Stress
-(04
03
Induced
Pore Pressure
(u -.u
percent
percent
- -
SHEAR STRENGTH
Strain
Strain
-
Deviator Stress
02
03
-(
)
R
IV
SERIES
IV
R
SERIES
CONFERENCE
A )
La
)B(
STRESSSTRAINCURVES
-
STRESSSTRAIN CURVES
.
.
No
TEST R22 TEST R23
No
81
( / sq
kg
- 1.-
!
cm
kg
0
08
.7-
/ sq
cm
=
ot
min
75
min
85
Fig
26
.
/
-
kg
og
Stress
sq
/
-
cm
kg
Stress
cm
/ kg
ng
-
1200
-
450
-
kg
/
ng
ant
800
Mso
Project 5810
Project 5810
21
Induced
Pore Pressure
u
-u(
)
Hottbot
Deviator Stress
-(
04
07
10
Induced
Pore Pressure
| (
u
-.
percent
percent
- -
in
Strain
Stra
Deviator Stress
-
!
Og
101
v
Å
R
v
SERIES SERIES
)A(
STRESSSTRAINCURVES STRESSSTRAINCURVES
TEST TEST
R25
R24
eq
kg
No 02
() . /
an
ng
No 01
. /. kg
com
- .8- -
04
70
75
min
0t
- 2--
min
393
Fig
27
.
The type of stress - strain curves which is seen in Fig . 26 , 27 and 28 , has
also been obtained in triaxial tests on cemented sands . The yield strength
corresponds to the failure of the cementation which occurs at very small
strains . Then , at very much greater strains the ultimate strength is mobi
lized which corresponds to the strength of the noncemented sand . Depending
on the confining pressure , the ultimate strength of the cemented sand is
smaller , equal or larger than the yield strength .
This characteristic of a double failure is not as clearly defined in the
stress - strain curves of the stress - control tests because in these tests it is
difficult to measure satisfactorily the shape of the stress - strain curves be
yond the yield strength , with the exception of those tests where the deviator
stress continues to increase well beyond the yield strength . Nevertheless ,
the pattern of the double failure , as described above , can still be observed .
The similarity of the double failure to that observed for cemented sands ,
suggests that there may be a bond at the points of contact that acts similar
to cementation , possibly due to thixotropy . The failure of this bond would ac
count for the very low strain corresponding to the yield strength , whereas the
high strains at the ultimate strength would reflect the mobilization of internal
friction as occurs in loose sand at high strains .
4 . Strength Characteristics . – The R circles and the modulii of deformation
(M50 ) from these tests are consistent with those obtained in the Ř tests with
stress - control loading and are not included in this paper . The vector curves
are substantially different from those obtained in the stress - control tests ,
probably because of errors in pore - pressure measurement at the start of the
test when the load built up very rapidly under strain - control loading .
C . Stress - Control Ā Tests with Seepage Saturation

1. Purpose of Tests . -- The method of saturating specimens with back
pressure is not generally used . Therefore , it appeared desirable to investi
gate whether the method of saturation has a significant effect upon the results
of the strength tests .
2 . Special Tests to Investigate Degree of Saturation . It is extremely diffi
cult to measure the degree of saturation of specimens subjected to seepage
saturation . Therefore , special tests were made in order to obtain a relative
measure of the degree of saturation achieved . The results of these tests are
plotted in Fig . 29 .
In Test A the specimen was first consolidated under a hydrostatic confin
ing pressure with a back - pressure of about 6 kg / sq cm on the pore water .
Then the drainage valves were closed and the hydrostatic confining pressure
was increased in increments , measuring the pore - pressure and the change in
height of specimen for each increment . The height of this specimen changed
very little during the test , as can be seen in the top of Fig . 29 . Curve A in
the bottom of Fig . 29 shows that this specimen must have been almost fully
saturated .
In Test B the specimen was first consolidated under a hydrostatic confin
ing pressure , but with no back - pressure on the pore water. Then water was
forced through the specimen under a differential head of about one meter for
a period of about two days , in the same manner as for the R tests with seep
age saturation . Then the drainage valves were closed and the hydrostatic
confining pressure was increased in increments , measuring the pore
pressure and the change in height of specimen for each increment . Under an
kg
nq
-kg
nq
stress Stress
com
.P
/- 8
cm 10
12
14
/ kg
ng
-
/ -
M50 1000
-
sq
kg
/
M50 2000
_cam
cm
Project 5810
Project 5810
Induced
Pore Pressure
-u
(
10
Induced
Pore Pressure
–Deviator Stress
(|u
.-
10
og
-
!,
porcent
percent
- -
Deviator Stress
og
10
!,
-
++
Strain
Strain
VI
&
SERIBS
VI
SERIBS
)A(
)B(
STRESSSTRAINCURVES
-
STRESSSTRAINCURVES
13ST
TEST
kg
ng
a
HO 16
./
- .
3
27 vq
kg
a
No 19
. /.
-
70
0t
--
ala
0&
8 55
•
min
22
395
28
.
Fig
A
-
--
Back pressure saturation
396
63
=
G
3
cu
13
8
lb
-w
107
percent
-A -B
Seepage uturation
.
.
-
R
Gw
1b
16
79
w
cu
=
Ya
107
Seepage satuntion
na
lb
=
=
13
,3 ,7 8
63
ft ft f't
-C w
/ / / cu
. . .
-7 = -,
. 9. 1.
,, =, G,
. 2. 0.
106
Strain Axial
u
-.
.0
and chamber pressure
and pore pressure
at end of consoli
dation and satura
tion process
SHEAR STRENGTH
com
.
0c
chamber pressure
eq
. .
==
/ uc total pore pressure

kg
-
."
-
CONFERENCE
uc
20
15
16
14
12
6
10 ng
kg
/
aan
- %
B-
°C
in
.
-.
29
Fig Axial Strain and Pore Pressure versus Change Hydrostatic Confining Pressure
increase in hydrostatic confining pressure of about 10 kg / sq cm the height of

the specimen decreased by about 1. 3 %, as can be seen in the top of Fig . 29 .
Such a decrease in height can only develop if the specimen contains a sub
stantial air content . This air content is also reflected in the shape of the
curve B at the bottom of Fig . 29 which shows the pore - pressure build - up .
However , under high confining pressures , curve B in the top reaches practi -
cally a horizontal slope , while curve B in the bottom reaches a 45 - degree
slope , both indicating that under the high confining pressures a state of
practically 100 % saturation was reached .
Test C was carried out in the same manner as Test B, except that the
specimen was compacted at a lower water content . The curves in Fig . 29 for
Test C indicate that this specimen contained even more air than specimen B .
Under a confining pressure of 13 kg / sq cm the height of specimen C de
creased by 2 . 8 %.
These special tests prove conclusively that seepage saturation is not ef
fective in achieving 100 % saturation with a clay such as Canyon Dam clay
(coefficient of permeability about 10 - 7 to 10 - 8 cm / sec ) . In contrast , satu
ration by means of back - pressure is effective , provided one uses a sufficient
ly high back - pressure .
3. Technique of Tests . The technique of performing these tests was simi
lar to that used in the stress - control tests with back -pressure saturation ex
cept that no back - pressure was used , and water was forced through the speci
men from the bottom to the top under a differential head of about one meter ,
for a period of about two days . The actual amount of water that flowed
through the specimen during this period was very small , usually not exceed
ing about three cubic centimeters . During the consolidation phase drainage
was permitted from the top porous stone only . During axial load increase
pore - pressures were measured through the bottom porous stone only .
4 . Summary of Test Results . - A total of 12 Ř tests with stress - control
and seepage saturation were performed in two series . In each series the
water content of the specimens as compacted was approximately constant . It
was attempted to compact all specimens as close as possible to the same dry
unit weight of about 105 lb / cu ft . The average as -molded water content , dry
unit weight, and degree of saturation for each series are shown in the table
below .
_ _ _ _
pcf
RVII 13 .4 106 . 3 61 . 7
RVIII 16 .3 107 . 3 77 . 0
5 . Stress - Strain Curves for Ř Tests with Seepage Saturation . – Typical

stress - strain curves (for an effective consolidation pressure of about 8 kg / sq
cm ) for each series are shown in Figs . 30 and 31 . When comparing all the
stress - strain curves (plotted in Reference 1) with the stress - strain curves
of the tests with back - pressure saturation , it is found that these tests with
seepage saturation have resulted in much more erratic stress - strain curves .
The reason for this is not known . It seems possible that the variations in
degree of saturation resulting from mere seepage saturation could be very
1500
Average Molding Water Content - 16. 3%
Series Ā VIII
/
cm
sq
1000
-
kg
Project 5810
Mgo
500
Series VII
Ř
-
13
%
.4
-
10
cm 12
14
16
4
sq
Effective Consolidation Pressure kg
/
Fig
30
Modulus Deformation versus Effective Consolidation

of
Pressure
.
.
-
much greater and more erratic within specimen than when using back
a
pressure saturation
.
general the pore ressure build these tests was substantially up

In
in
,
-
-p
specimens saturated by means of back pressure

on
smaller than the tests

in
.
probably due the greater air content the specimens saturated by
of
This
to
is
seepage flow
.
the compressive
of
50
The secant modulus deformation M50 for
of
the consolidation pressure

as
strength plotted Fig

30
is
of
function
in
o
c
.
.
'
When comparing these results with the corresponding results plotted
in
Fig specimens saturated with back pressure one can see
23
. on
for the tests
,
.
Fig
30
that the moduli are substantially smaller This result probably
is
in
caused by the greater air content which made these specimens more
spongy
”
.
"
31
Circles for Tests with Seepage Saturation Fig are plotted
In
R
in Ř
6
-
.
is .
the strength circles terms total stresses instructive compare

of
to
It
.
the following table the slope angles

of
the lines obtained these tests

in
in
R
with those from the tests with back pressure saturation

-
SLOPES OF LINES TOTAL STRESS CIRCLES

R
Stress Control Strain Control

-
-
with Back Pressure with seepage

-
Saturation Saturation
0° 30
6° 60 7°
20 23
16
Water Content
. .
Low
. . .
19 17
Medium Water Content

High Water Content
As already pointed out before the tests with back ressure saturation gave
,
-p
slope angles which increased with the water content at which the specimens
were compacted was shown that these specimens were almost 100 satu
%
, It
.
rated contrast the tests with seepage saturation gave much larger slope
In
.
angles The difference largest for the low water content and con
6°
is
(6
.
)
.
siderably smaller 4° for the medium water content No tests were per
2
(
)
.
.
on
with seepage specimens with high water content

is
formed saturation
It
.
likely that in such tests there would be little difference in the slope angles
since the degree of saturation would have been high even if no attempt at satu
ration had been made. Thus one is forced to the conclusion that the high slope
angles for the R lines derived from tests with seepage saturation are related
to incomplete saturation and the resulting small pore - pressure build - up .
IX . S TESTS
A. Triaxial S Tests
1. Summary of Test Results . - A total of 10 triaxial S tests were per
formed in two series , such that in each series the water content of the speci
mens as compacted was approximately constant . It was attempted to compact
all specimens as close as possible to the same dry unit weight of about 105
lb / cu ft . The average as -molded water content , dry unit weight, and degree
of saturation are shown in the following table .

pcf
SI 13 . 3 107 . 5 62 . 5
SII 19 .0 106 . 3 87 .0
2 . Stress - Strain Curves for Triaxial S Tests . - Typical stress - strain

curves ( for a chamber pressure of about 4 kg / sq cm ) for each series are
shown in Figs . 32 and 33 . All stress - strain curves have erratic steps which
increase in magnitude approximately proportional to the deviator stress . The
largest steps are very abrupt and are of the order of magnitude of 10 to 15 %
of the deviator stress . In appearance , and because they are roughly pro
portional to the deviator stress , these steps would strongly indicate that they
have been caused by piston friction . On the other hand , detailed tests which
were conducted in the past on these pistons with ball bushings ( see Fig . 4 in
Ref . 6 ) , proved that even with large eccentricity and for periods of rest up to
one week , the friction remained well within tolerable limits . However ,
further tests are planned to investigate the cause of these steps , including ad
ditional check tests on piston friction . Since temperature variations during
these tests remained within a range of +2° C, and since these variations were
gradual , this could not be the cause of the observed steps .
In addition , much more gradual , and smaller steps occurred ( e. g . in Fig . 33
corresponding to a deviator stress of 3. 7 kg / sq cm ) whenever the size of the
load increments was reduced . The smaller the load increments are , the
smaller are the pore - pressures which are induced at the moment of the ap
plication of an increment. Although these increments and corresponding pore
pressures are small , it seems that the build - up of the load by means of larger
increments , i. e. , with larger temporary pore - pressures , causes greater
strains .
For low water content and low confining pressures the failure strains are
very small . For all other tests failure was approached gradually at high
strains . Failure of all specimens occurred usually only by bulging . In some
tests shear planes also formed , but only at very high strains .
400
Project 5810
)A -
(Ř
SERIES VII
R
CIRCLES
com
sq
/
13
%
-
. 4
kg
-
)7
(
Stress Shear
Z
R28
R30
!
igal las
ok
0
2
4
6
8
10
12
14
16
18
20
eq
-)o
(
kg
/
can
Total Mormal stress

Project 5810
)B(
Ř
SERIES VIII
R
CIRCLES
can
sq
16
%
3
.
-
/ Average Molding Water Content
kg
-
)
1
(
0°
.
20
I
Stress Shear
R34
R36 R37 R39 R38
18
16
14
10
6
20
4
cm
sq
12 kg
/
( 0
-)
Total Normal Stress

401
. 31
Fig
)
14
Project 5810
a
/
ng .
S9
II
8
days
No
.3- - (
kg
SERIBS
12 73
Test
22
a % ty
/
10
sg
Curve
-8
kg
-
Dovlator Stroo
Strain
-
Stress
.
33
Fig
Percent Strain
-
)
.
Project 5810
14 S4
a
/ No
(
ng
12
SERIBS SI
days
.3- -
kg
89 Test
a 10
/
19
ng
-8
Curve
kg
-
viator Stross
Strain
-.
Stress
.
32
Fig
percent Stnia
-
As in Figs . 32 and 33 , usually two stress - strain curves were drawn : a

dashed curve which goes through all computed points (including the peak
points of the steps ) , and a full - drawn curve which is an estimate of the true
stress - strain curve assuming that the steps were caused by some testing
error . Strength circles for both stress - strain curves are plotted in Fig . 34
whenever the dashed and the full - drawn curves give different compressive
strengths .
The secant modulus of deformation M50 for 50 % of the strength is plotted
in Fig . 35 as a function of the confining pressure . For the tests on specimens
compacted at low water content the modulus appears to be affected very little
by the confining pressure . However , the specimens compacted at a water
content well above optimum show approximately a linear increase of M50
with confining pressures . It should be emphasized that the M50 values de
rived from S tests have not the same physical meaning as those derived from
Q or R tests , or unconfined compression tests , in which no change in water
content takes place during axial loading . In the case of S tests on specimens
compacted at high water contents , a substantial decrease in water content was
taking place (the amount increasing with increasing confining pressure ) which
is reflected in the axial strain . On the other hand , for the specimens com
pacted on the dry side of optimum the specimens were actually swelling dur
ing axial load increase , the swelling being greatest for the smallest confining
pressures .
While it is difficult to measure accurately the volume changes of a triaxial
specimen during the test , one obtains a reasonably accurate measure of the
total volume change by measuring the water content at the beginning and at
the end of each test . The actual changes in weight AW were measured in
these tests . All specimens compacted on the dry side ( Series SI) increased
in weight . Under the smallest confining pressure (0 . 44 kg / sq cm ) the weight
increase was about six grams , and under the highest confining pressure (7 . 88
kg / sq cm ) the weight increase was about one gram . The specimens compact
ed on the wet side (Series SII ) decreased about 5 grams in weight under the
largest confining pressure of 7 . 63 kg / sq cm , while under the smallest con
fining pressure of 0. 51 kg / sq cm the specimen actually increased in weight
by about 0. 2 grams . These measured changes in weight are affected by the
amount of water which the specimen has absorbed from the porous stones in
the few minutes between the removal of the chamber pressure and the weigh
ing of the specimen . It is estimated that this error did not exceed one gram ,
and in most tests was much smaller .
3. Strength Circles for Triaxial S Tests . - In Fig . 34 are plotted the
strength circles for all triaxial S tests , and the S lines drawn through the
origin tangent to the largest circles . As explained in the preceding section ,
two circles were plotted for those tests for which the stress - strain curves
did not yield one clearly defined strength value because of the steps in the
stress - strain curves.
As can be seen in Fig . 34 , the strength circles for the smaller confining
pressures extend only slightly above the S lines . The slope angles for the S
lines is slightly larger for the specimens compacted wet of optimum , as can
be seen from the following comparison .
404
Project 5810
)A(
cm
SHEAR
Series SI
sq
13
. 3
%
-
/ Average Molding Water Content
kg
-
STRENGTH
Stress Shear
CONFERENCE
Note See text for meaning

.
of double circles
S3 S4
20
0
10
2
4
12
14
6
16
18
/
o
(
-)
cm
sq
Normal Stress kg
Project 5810
()B
Series SII
19
Average
=
0
Molding
. %
Water Content
cm
sq
/
kg
-
Stress Shear
STRESS -DEFORMATION
Note See text for meaning

.
of double circles
S6 S7 S8
Hslo
!
|
|
2
4
6
3
20
14
16
15
20
kg
cm
)o-
s 12/sq
Normal Stress
Fig
34
. .
( S-
Strength Circles
405
Test Series Water Content slope Angle

Number When compacted of S Line
2 (deg ) .
SI 13 . 3 29 .6
SII 19 .0 31 .5
B. Direct Shear Tests on Large -Diameter , Thin Specimens

Apparatus and Technique of Tests . - Triaxial S tests are very time
1.
consuming to perform . On the other hand , direct shear tests on thin remolded
specimens can be carried out in a relatively short time . Therefore , it was
considered of interest to determine whether direct shear tests could be sub
stituted for triaxial tests for determination of the slope of the S line .
The apparatus used in this investigation , illustrated in Fig . 36 , was built
in 1954 for the purpose of investigating the effect of time of loading in long
duration direct shear tests on the slope of the S line . These tests were per
formed on Boston Blue Clay remolded to the liquid limit . The results of this
investigation are summarized in an unpublished laboratory report (7) . The
principal feature of this apparatus , as compared to standard types of direct
shear boxes , is the use of very thin specimens of large diameter. The
300
| sories SI
Avenge Holding Mater Content - 13. 3%
/
cm
sq
)-
kg
(
Mgo
Deformation
Sorias SII
19
%
Average Holding Mater Contant

.0
-
of
Modulus
6
5
3
4
2
anu
ng
Consolidation Pressure kg
-
Fig Deformation versus Confining Pressure Tests

35
Modulus
of
(s
)
-
.
.
Extensometer -
Steel Plate - Steel Ball
Rough Surfaces , - Lucite Ring

WOWY
Micarta
FREE , Porous Carbon
Sample
Micarta
1 6. 3" Diam. (200 cm Area )
Shear Load
Loading Frame
Normal Load
Load
SECTION
CH
PLAN
Fig . 36. -- Apparatus for Direct Shear Tests on Thin Specimens

diameter of the porous plates in this apparatus is 6 . 3 in . and the thickness of
the specimens at the start of the test generally ranged between 1 / 8 and 3 / 16
in . Because the specimens are very thin in relation to the diameter , it is
possible to perform the tests without providing the specimens with lateral
confinement . The porous stones were made of porous carbon plates and they
were serrated perpendicular to the direction of movement . The serrations
were made by means of a hacksaw which was scraped laterally over the sur
face of the stone several times , thus creating grooves with a spacing corre
sponding to the spacing of the teeth .
The total time of loading to failure ranged in these tests between 195 and
260 min . Stress - control loading was used . The size of the load increments
was reduced during each test in several stages . The first stage consisted of
load increments approximately equal to 10 % of the estimated failure load . In
each subsequent stage the size of the load increment was made one -half of the
size used in the preceding stage . Each increment was allowed to remain in
place for a period of ten minutes , except that in some cases only five minutes
was allowed for each of the last few , very small increments .
2 . Test Results . - A total of six direct shear S tests were performed with
this apparatus , on specimens remolded and consolidated from the liquid limit ,
under normal pressures between 0 . 1 and 0 . 3 kg / sq cm .
A typical stress - displacement curve for these tests is shown in Fig . 37 .
Every time the size of the load increments was reduced , the stress
displacement curve developed a small discontinuity , similar to the gradual
steps observed in the triaxial S tests ( e. g. , corresponding to a shear stress
of 0. 075 kg / sq cm in Fig . 37 ) .
In Fig . 38 are plotted the shear strengths versus normal stresses ( Tests
DS1 to DS6 ). It can be seen that a straight line through the origin , sloping at
an angle of 31 . 9 degrees , passes almost exactly through the points represent
ing these six tests .
C. Direct Shear Tests in Conventional Shear Frame

1. Apparatus and Technique of Tests . - Because of the large area of the
shear apparatus described above , tests with this apparatus are limited to
normal stresses of not more than about 0. 3 kg / sq cm . Therefore , two tests
were made in a conventional shear frame with normal stresses of 0 . 5 and 1. 0
kg / sq cm . These tests were also started at the liquid limit , and the incre
mental loading was performed essentially in the same manner as described
above for the direct shear tests on large diameter , thin specimens . The size
of these specimens was as follows : Thickness at the start of tests was ap
proximately one cm ; the shear frame was 6 x 6 cm square . Coarse Norton
porous stones with rough surfaces but without dentations were used . The
total time of loading in these two tests was 140 and 160 minutes .
2. Test Results . -- In Fig . 38 are plotted the shear strengths versus normal
stresses ( Tests DS7 and DS8 ) . It can be seen that these results fall practi
cally on the same straight line through the origin , sloping 31. 9 degrees , which
was determined from the results of the six direct shear tests with large
diameter , thin specimens .
X . CONCLUSIONS
The investigation described in this report covers only the effect of vari
ations in molding water content on the stress - deformation and strength
all
characteristics of specimens of one compacted clay ,
the same dry unit
to
weight .
The following discussion and conclusions should be considered tentative
,
subject the investigation
of
modification as the results the later phases
of
to
become available
Compaction Characteristics . Confirming results other investigators
of
1
,
.
.
-
was found that the field compaction curve can be closely duplicated
by
it
means kneading type compaction whereas different shape curve

of
of
is
of
a
-
by
dynamic laboratory compaction Fig Close simi

of of
obtained means
2
.
).
(
larity laboratory and field compaction curves does not
it
the shape
of
in
laboratory and
of
self however ensure that the stress strain characteristics
,
,
-
field compacted specimens will also be similar each other because the two
to
-
of
compaction may still result different types
of
methods structure
in
.
kg
sq
Shoar Struss
a
-
08 /
02
04
06
10
12
14
0
0
.
.
project 5810
100 inch
/1-
Displacement
ng
/kg
20
=
0
a
260 .
min
DS
No
Fig Stress Displacement Curve

37
Test
3
-
-
.
.
.
(
)
410
Project 5810
DSB
an
/vq
kg
-
)
7
( in
Tests conventional shear frame
SHEAR STRENGTH
stru
Shear
D65
DS6
CONFERENCE
on
,
-
DS3 ests large diameter thin specimens
DSA
DS1
DS2
0
1
.0
2
0
3
.
.0
4
5
6
.07
.0
8
.0
9
.1
0
kg
.0 a/
.0 –)o(.
nq
Normal stro
38
Fig
—
. ..
Direct Shear Test Results
Furthermore , in such comparisons one must pay attention to another im

portant variable , namely the time which elapses between compaction and the
testing of the strength of a specimen . Clay compacted in an embankment
would normally be subject to a much greater lapse of time before strength
testing than would be the case for laboratory - compacted specimens which are
usually tested immediately after compaction . The fact is that very little is
known about the long - term changes in stress - deformation and pore - pressure
characteristics of compacted clay in an embankment . In the last analysis ,
the results of the laboratory investigations on compacted clays are of little
value, indeed , and may even be misleading , as long as the other question is
not thoroughly investigated .
Even from the standpoint of laboratory investigations , the time which
elapses between compaction and strength - testing may have a significant in
fluence and this time should be controlled within close limits . Also , the
" time of curing ” which elapsed between the time when the soil mass was
mixed to the desired water content , with the addition of water , and the time
when the specimens were compacted , was found to have an influence on the
shape and position of the compaction curve ( Fig . 4 ) .
2 . Special Pore - Pressure Tests. – The following three special types of
tests were performed for the purpose of investigating the effects of vari
ations in water content at constant dry unit weights ( i . e. , really the effect of
variations in degree of saturation ) , and of the 01 / 03 ratio upon the pore
pressure build - up :
( 1 ) Specimens subjected to increasing hydrostatic confining pressure at

constant water content (PH Tests ) .
( 2) Specimens subjected to increasing nonhydrostatic confining pressure ,
but maintaining a constant 01 / 02 ratio , and at constant water content
( PN tests ).
(3) Specimens loaded in consolidation apparatus at constant water content
(PC Tests ).
At constant water content , approximately the same relation between pore
pressure and total major principal stress 01 was obtained in all three types
of tests , regardless of the 01/ 03 ratio ( Figs . 5, 6 and 7 ).
When testing specimens compacted at high water contents ( i. e., small air
contents ), one observes that for low pressures the curve of pore - pressure vs.
applied stress starts out slightly flatter than 45 degrees ; but already under a
relatively small applied stress the small quantity of air contained in the
specimen is driven into solution and the remainder of the curve is a straight
line with a 45 - degree slope . In contrast , when testing specimens compacted
at low water contents (i. e., with high air contents ), this curve is very flat
(i. e ., very small pore - pressure build - up ), and approaches a 45 - degree slope
only under high pressures .
3 . @ Tests . The induced pore - pressure at failure for specimens com
pacted several per cent on the dry or several per cent on the wet side of opti
mum , are very small for small confining pressure and are of the order of
50 % of the strength (deviator stress at failure ) for high confining pressures .
For a specimen compacted approximately at optimum water content , the in
duced pore - pressure at failure is approximately equal to 20 % of its strength
for low confining pressure , whereas for high confining pressure it is ap
proximately equal to the full compressive strength of the specimen .
The stress - strain curves and the pore - pressure build - up for specimens
compacted near optimum water content are similar in shape to those obtained
in Q tests on undisturbed samples of normally consolidated , medium sensitive
clays . In contrast , for low and high water contents , the stress - strain curves
have more the appearance that one expects for remolded clay .
There is a poorly defined trend for the M50 modulus to increase with de
creasing water content and increasing confining pressure . At the higher con
fining pressures one would expect the modulus to approach a constant value
which , however , is not discernible in Fig . 12 .
The Q envelope ( i . e . , envelope to the total stress circles ) for high water
low
contents is practically horizontal whereas for the water contents the en
velope rises very substantially
sq
for confining pressures
cm
up
kg
. 14
to
. /
this investigation as can be seen Fig
13
the maximum used For
in
in
),
(
water contents approximately equal optimum the envelope attains practi
to
,
cally horizontal slope for confining pressures above
cm
kg
sq
im
The
a
.
portance the shear strength com
on
minor variations
of
of
water content
in
pacted specimens brought out by the summary plot Fig
16
which shows
is
in
.
the shear strength on the failure plane as the total normal stress
of
function
a
the failure plane for various water contents
on
The best summary presen
.
Fig where the pore
of
the results tests
19
tation the shown
is
of
in
Q
,
.
pressure at failure plotted vs the total major principal stress at failure
is
for various confining pressures Or and water contents This figure permits
.
the strength and pore pressure failure for any combination
of
at
estimation
,
-
water content and total major principal stress which for
an
of
embankment
(
during construction usually taken as the weight of overlying material
is
)
.
Tests The principal tests for this investigation were carried out
R
Ř
4
.
.
--
practically fully saturated specimens Full saturation was achieved by

on
applying substantial back pressure the pore water during consolidation

to to
a
.
by
tests was attempted achieve saturation the conventional

In
few
it
of R
a
procedure allowing water seep through the specimen under small head
to
.
was found that this procedure was quite ineffective achieving saturation
in
in It
impervious material degree

an
of
such Because the very small change

in
.
this procedure the results these seepage

by
of
of
saturation achieved
,
"
saturated tests were found be useful demonstrating the importance of

to
in
”
the degree of saturation on the stress strain and strength characteristics

-
from tests
R
with back pressure saturation

of
The main series tests was

Ř
e
,
(i
.
.
carried out two groups one with stress control and the other with strain
in
,
-
control No significant difference was found between the results these two
of
of .
types tests except that the strain control tests define the stress strain
,
-
-
curves much better beyond the point failure was

of
On the other hand

it
,
.
found that the measured pore pressures were subject serious errors the
to
in
-
strain control tests during the first part the loading when the rate
of
of
load
-
increase was very high

.
The induced pore pressure failure for specimens compacted all water
at
at
-
contents negligible for small consolidation pressures whereas for high

is
pressures the compressive strength

of
75
consolidation about
is
%
it
The stress strain curves for specimens compacted at low and medium
-
water contents are practically straight lines strains approaching failure

. to
of the order as shown Figs For specimens

of
26
in
to
and 27
%
0
.5
.1
0
),
(
compacted high water contents the stress strain curves are straight lines
at
-
up
corresponding
of
deviator stresses between the strength

60
to
and 80
%
to
,
strains of the order of 0. 5 to 1.0 %, but failure is reached only at strains of

the order of 20 %, as shown in Fig . 28 .
For all water contents , the stress - strain curves indicate a " double
failure ” , consisting of a yield strength ” at a very low strain which develops
immediately following of the stress - strain curve ,
the straight - line portion
and an " ultimate strength ” that develops very high strains ( Figs . 26 , 27 ,
at
and 28) . For low molding water contents and low confining pressures the
ultimate strength is generally less than the yield strength , like the stress
strain curve for a sensitive , undisturbed clay . For low water contents and
high confining pressures the yield and ultimate strengths are about equal .
For the intermediate water content , at all confining pressures , the ultimate
strength was slightly greater than the yield strength . At the high water
content the yield strength lies between 60 and 80 % of the ultimate strength .
The similarity of this double failure to that observed for cemented sands ,
suggests that there may be a bond at the points of contact that acts like a ce
mentation , possibly due to thixotropy . The failure of this bond would account
for the very low strains corresponding to the yield strength , whereas the high
strains at the ultimate strength would reflect the mobilization of internal
friction as occurs in loose sands at high strains .
The modulus M50 at all water contents is approximately in direct pro
portion to the consolidation pressure , for consolidation pressures larger than
about 8 kg / sq cm ( Fig . 23 ). The ratio M50 / 0'c increases with lower consoli
dation pressures .
The slope angle of the R lines increases slightly with increasing water
content, from about 17 to 20 degrees ( Fig . 24 ) . The slope angle of the R
lines increases substantially with increasing water content , from about 31 to
37 degrees (Fig . 25 ) . At this stage of the investigation it is not yet clear
whether such high angles reflect a real property of the material , or whether
these results are influenced by experimental errors as discussed in Chapter
VIII .
The results of tests with seepage - saturation for low water contents are
substantially different from the results of the tests with back - pressure satu
ration , because there still remained a large volume of air in the specimens .
On the other hand , for high water contents the differences between the results
of tests using the two procedures become much smaller , because the degree
of saturation of the specimens as compacted is already high . For low water
contents , seepage - saturation resulted in ( 1) smaller pore - pressure build - up ,
( 2) lower M50 values , and ( 3) higher strengths , than back - pressure saturation .
It is emphasized that most of the R tests were carried out on fully saturat
ed specimens , whereas the Q and S tests were conducted on the specimens
with the air content as compacted . In order to arrive at a complete under
standing of the influence of the air content on the stress - strain and strength
characteristics and the pore - pressure build - up , it is planned to include in the
second phase of this investigation additional R tests in which no attempt is
made to increase the degree of saturation , and S tests in which full saturation
is achieved by applying sufficient back - pressure .
5 . Triaxial S Tests . There is a possibility that in the triaxial S tests in
which the rate of load application is very slow ( time of loading = 16 to 26
days ) , piston friction of an objectionable magnitude may have built up . This
question will require further investigation .
The slope angle of the S lines increases with increasing water content
from 29 . 6 to 31 . 5 degrees (Fig . 34 ) .
6 . Direct Shear S Tests . -- Since direct shear tests are frequently used for
obtaining the S envelope , it was desirable to compare the results of triaxial
and direct shear S tests . For the direct shear S tests a special apparatus
was used in which large diameter , very thin specimens between rough porous
stones , and without lateral confinement , were tested . This apparatus was de
signed specifically for low normal pressures , with a maximum of about 0 . 3
kg / sq cm . So far , only preliminary tests were made in which the material
was placed in the apparatus at the liquid limit . The S line obtained from
these tests slopes at an angle of 31. 9 degrees ( Fig . 38 ). In addition , two di
rect shear S tests were made in conventional shear frames , using loads up to
one kg / sq cm , and they gave essentially the same slope angle for the S line
( Fig . 38 ) .
XI. ACKNOWLEDGMENTS
This investigation was carried out under Contract No . DA - 22 - 079 - CIVENG

59 - 17 between the Waterways Experiment Station and Harvard University .
The authors are greatly indebted to Mr . George E . Bertram , Office of the
Chief of Engineers , not only for his initiative in promoting this project , but
also for many valuable suggestions . The stimulating interest by the staff of
the Waterways Experiment Station is gratefully acknowledged . The tests
were carried out by the junior author and the following members of the staff:
Mr . Walter Ferris , formerly Lecturer on Soil Mechanics ; Mr. Peter Edinger
and Mr . Bruce Maclver , graduate students .
APPENDIX I. - NOTATION
The nomenclature used in this paper is described below .

DS Test An S direct shear test .
Degree of saturation
Gw
Secant modulus of deformation corresponding to 50 % of the
M50
maximum deviator stress , i. e. , the ratio of 50 % of the maxi -
mum deviator stress to the corresponding strain .
PC Test A pore - pressure test in which the specimen is loaded in a
special type of consolidation apparatus ( lateral strain equal to
zero ) , at constant water content .
PH Test A pore - pressure test in which the specimen is subjected to in
creasing hydrostatic pressure , at constant water content .
PN Test A pore - pressure test in which the specimen is subjected to in
creasing stresses , at constant water content , with the ratio of
01 / 03 maintained constant throughout the test .
Q Test A triaxial compression test at constant water content in which
the specimen is first subjected to a hydrostatic confining
pressure and then the axial stress is increased to failure .
Q Test A triaxial compression test the same as the Q test except that
pore pressures are measured .
Q Circle The strength circle from a Q or ő test , plotted in terms of

total stresses .
ā Circle The strength circle from a ā test, plotted in terms of effective
stresses .
Q Envelope The envelope to a series of Q circles .

Envelope The envelope to a series of Q circles .
R Test A triaxial compression test in which the specimen is first
consolidated under an effective hydrostatic consolidation
pressure o'c, and then , without permitting any further change
inwater content , is subjected to axial load increase to failure .
Ã Test A triaxial compression test the same as the R test except that
pore pressures are measured .
R Circle The strength circle from an R or å test, plotted in terms of
net total principal stresses , of and oſ .
Ř Circle The strength circle from an test , plotted in terms of the ef
fective stresses .
R Envelope The envelope to a series of R circles .

Ā Envelope The envelope to a series of R circles.
R Line The R envelope for normally consolidated clay specimens ; it
is believed to be generally a straight line through the origin .
Ã Line The R envelope for normally consolidated clay specimens ; it
is believed to be generally a straight line through the origin .
Shear strength .
S Test A triaxial test in which the specimen is first
compression
consolidated under a hydrostatic confining pressure , and is
then subjected to axial load increase which is applied in small
increments , allowing full consolidation under each increment ,
until failure is reached ; also , a direct shear test in which the
specimen is first consolidated under a normal stress , and is
then subjected to shear stress which is applied in small in
crements , allowing full consolidation under each increment ,
until failure takes place .
S Circle The strength circle from an $ test .
S Envelope The envelope to a series of S circles .

S Line The S envelope for normally consolidated clay specimens ; it is
believed to be generally a straight line through the origin .
Time of loading ; the total elapsed time from the start of axial
load increase in a triaxial test , or shear stress increase in a
direct shear test, to failure .
Pore - pressure measured at the ends of a triaxial specimen .
Initial back - pressure , i . e. the pore - pressure applied at the
ends of the specimen and maintained constant during consoli -
dation under the hydrostatic confining pressure in an R or R
test .
Vector curve The locus of points representing the shear stress and effective
normal stress acting on the plane along which failure ultimate
ly takes place during a triaxial test throughout the application
of deviator stress .
Water content .
Dry unit weight.
The inclination of an R line .
The inclination of an Ē line .

Total normal stress .
Total major principal stress .
The net total major principal stress , i. e. , the total major
principal stress minus any initial back - pressure up . For
uo = 0, oi = 01 :
Total minor principal stress .

The net total minor principal stress , i. e. , the total minor
principal stress minus any initial back - pressure uo . For
uo = 0 , 03 = 03 .
01/ 03 Total principal stress ratio ; the ratio of the total major
principal stress to the total minor principal stress .
(01 - 03 ) Deviator stress ; the difference between the major principal
stress and the minor principal stress .
(01 - 03 )max The compressive strength ; the deviator stress at failure .
Hydrostatic confining pressure ; the chamber pressure in a tri
axial test .
The effective consolidation pressure in an R or S test , i. e. , the
hydrostatic confining pressure o minus the initial back
pressure uo .
Effective normal stress .
Effective major principal stress .

Effective minor principal stress .
Effective principal stress ratio ; the ratio of the effective
major principal stress to the effective minor principal stress .
The effective principal stress acting on a plane inclined at an

angle of 60° to the major principal plane .
Shear stress .
The shear stress acting on a plane inclined at an angle of 60°
to the major principal plane.
The inclination of the S line .
APPENDIX II . - REFERENCES
1. Casagrande , A ., and Hirschfeld , R . C . First Progress Report on Investi -

gation of Stress -Deformation and Strength Characteristics of Clays (Re
port to Waterways Experiment Station ). Soil Mechanics Series No . 61 ,
Harvard University , Cambridge , Mass ., 1960 .
2. Casagrande , A ., Corso , J . M ., and Wilson , S. D . Investigation of Effect of
Long - Time Loading on the Strength of Clays and Shales at Constant Water
Content (Report to Waterways Experiment Station ). Harvard University ,
Cambridge , Mass ., 1950 .
Casagrande , A . and Wilson , S. D . Report to Waterways Experiment Station
on Triaxial Research Performed During 1950 - 1951. Harvard University ,
Cambridge , Mass . , 1951 .
4. Casagrande , A . and Wilson , S. D. Investigation of Effect of Long - Time

Loading on the Strength of Clays and Shales at Constant Water Content
(Report to Waterways Experiment Station ) . Harvard University , Cam
bridge , Mass . , 1949 .
5. Hilf , J . An Investigation of Pore - Water Pressure in Compacted Clays
( Technical Memorandum 654 ) . U . S. Bureau of Reclamation , Denver , 1956 .
6. Casagrande , A . and Wilson , S. D . " Prestress Induced in Consolidated

quick Triaxial Tests , " Proceedings , Third International Conference on
Soil Mechanics and Foundation Engineering , Zurich , 1953 , vol. 1, pp . 106
110 .
7. Corso , J. M . "Notes on Direct Shear Tests on Thin Samples of Remolded

Clay ” ( unpublished report ) , Harvard University , Cambridge , Mass . , 1955 .
EFFECT OF CONSOLDATION PRESSURE ON ELASTIC AND
STRENGTH PROPERTIES OF CLAY
By Stanley D . Wilson ,1 F . ASCE , and Rudy J. Dietrich ,2 A . M . ASCE
INTRODUCTION
When axial load is applied to a specimen of soil, as for example in an un

confined compression test , the application of each load increment produces an
instantaneous deformation followed by creep deformation , the rate of which
decreases rapidly . In 1949 and 1950 Harvard University undertook for the
U . S. Waterways Experiment Station an investigation of the effect of long - time
loading on the strength of clays and sha les at constant water content , and as
part of this research the senior author deve loped techniques for separating the
sudden deformation from the creep deformation (1) " . The ratio of applied
stress to sudden strain was referred to as the " instantaneous modulus of de
formation " (Mi). For any one specimen Mi was found to be near ly constant for
all values of applied stress up to a failure condition and it was concluded that
Mi corresponded approximately to Young 's Modulus of Elasticity , E . Young ' s
Modulus was also measured by a vibration method using a special apparatus
developed for this purpose ( 2 ) .
It appeared from the earlier work at Harvard that there was a definite re
lationship between Mi and the compressive strength of various clays . More
recent inventigations using improved apparatus and techniques have shown that
for sands , the modulus of elasticity is dependent on the effective pressure be
tween the grains (3) . Since the compressive or shear strength of a cohesive
soil is governed primarily by the effective stress under which the clay has
been consolidated , it appears reasonable that there may be a correlation be
tween strength and elastic modulus for cohesive soils .
Skempton (4) has established a correlation with a modulus of deformation
for London clay .
This paper presents data which indicates that a correlation does exist for
the elastic modulus of clays . Included are detailed descriptions of the soils
tested and the techniques of testing . It is believed that the results will be of
particular interest to those engineers involved in the design of foundations for
radar towers , power plants , dams , missile - launching silos and other impor
tant structures subject to transient loads where the resulting elastic deforma
tion must be held within specified limits .
DEFINITION OF TERMS AND SYMBOLS
Unconfined Compression (U ) Test is a test in which an unconfined specimen

is subjected to axial loading until the specimen fails . All U tests referred to
1. Partner , Shannon and Wilson , Seattle , Wash .

2. Proj. Engr ., Shannon and Wilson , Seattle , Wash .
3. Numerals in parentheses refer to corresponding entries in the Appendix
References .
419
in this paper were of the controlled - stress type , with the axial load applied in
increments at regular time intervals .
Quick Triaxial Compression ( Q) Test is a strength test in which the speci
men is subjected to hydrostatic pressure and then the axial load is increased
to failure without allowing any drainage during the entire test .
Rapid Triaxial Compression (R ) Test is a strength test in which the speci
men is first consolidated under a given hydrostatic pressure and then , without
permitting further consolidation , subjected to axial load increase to failure .
Slow Triaxial ( S) Test is a strength test in which the specimen is first con
solidated under a given hydrostatic pressure and then subjected to an axial
load increase which is applied in small increments , allowing full consolidation
for each increment until failure is reached .
Compressive Stress is the axial load per unit of cross - sectional area of a
test specimen . In computing the compressive stress corresponding to an
axial load on the specimen , a corrected cross - sectional area is used which is
computed from the following formula :
original area
corrected area = 1 - strain
Compressive Strength (a) . In unconfined Compression Tests the compres

sive strength ( q) is the maximum stress reached during the test . In a triaxial
compression test , the compressive strength is the maximum deviator stress.
Instantaneous Strain is obtained by plotting the curve of deformation vs time
for each increment of load , and extrapolating this curve back to zero time .
The difference between this extrapolated strain and the total strain immedi
ately prior to applying the load increment , is defined as the instantaneous (or
sudden ) strain .
Instantaneous Modulus of Deformation ( Mi
) is the slope of the straight line
obtained when stress vs instantaneous strain is plotted .
Effective Preconsolidation Pressure (Pc ) is that all around pressure under
which a specimen of clay would have to be first consolidated in order that the
resulting compressive strength as subsequently determined by U, Q or R tests
would agree with the actual compressive strength .
Dynamic Modulus of Elasticity (Ed) is the modulus of deformation comput
ed from data obtained in a vibration test using the equipment described in this
report .
Static Modulus of Elasticity (Es ) is the modulus of deformation obtained at
the end of a repetitive loading period during which small on - off load incre
ments are used . This procedure eliminates creep effects .
TESTING EQUIPMENT AND PROCEDURES
Instantaneous Modulus of Deformation (Mi) is obtained from study of time

deformation readings taken at specified time intervals after application of
each load increment in U, Q or R tests , from which time curves are plotted .
Fig . 1 a shows typical time curves for an R test on Cucaracha clay shale
(Soil B) . These time curves represent " creep " of the specimen under con
stant load between increments . By extending any time curve back to the zero
time corresponding to application of the load increment , it is possible to ob
tain the total creep during the time interval . When the total creep is subtract
ed from the total deformation , the instantaneous deformation accompanying the
CONSOLIDATION EFFECT 421
Time - seconds Deviator Stress - tons per sq ft

20 40 0 10 20 30
Creep Under !
Stress=24.84 tons per sq ft wo . Mi =107,000
Elastic Deformation
-
% Elastic Stroin
Deformation
W Creep
Creep Under Deformation
Strain
sq
Stress 2760 tons per
ft
:
on
Time Deformation Curves Test Cucoracha Clay Shale
(a
R
-
(b
)
Fig Determination Modulus Mi

of
of
Instantaneous Deformation
1
.
.
(
)
Axial Stress
Strain
Axial
Fig Static Modulus Es

of
of
Determination Deformation
2
.
.
(
)
application was not found possible

of
obtained record the instan

to
load
It
is
ap
taneous deformation directly for two reasons first the load could not be
,
,
plied instantaneously without impact and second the movement the dial
of
,
;
by
was too rapid read eye and possibly lagged behind the specimen
to
deformation
The solid stepped lines in Fig . 1 b show the actual deformations of the spec
imen , separated into instantaneous movements and creep movements . The
curved line represents the total combined deformation , and is the stress
strain curve that would normally be plotted when time curves for each load
increment are not obtained .
The dashed line shown in Fig . 1 b represents the summation of instantane
ous strain plotted against total stress . The slope of this line is referred to as
the " instantaneous modulus of deformation (Mi) . " For most soils tested , all
curves of stress vs instantaneous strain have been found to be straight lines
passing through the origin in contrast with the curves of stress vs total strain
which have ever increasing curvature as failure is approached .
Static Modulus of Elasticity (Es) is determined in the following manner :
The specimen is placed in the triaxial apparatus under a given confining
pressure . A small axial stress is applied in increments and then reduced .
This cycle is repeated until creep deformation ceases to be significant and
the unit deformations between corresponding load increments becomes con
sistent, Fig . 2. The static modulus of elasticity is then computed using the
small load increments and corresponding deformations .
It should be noted that the reverse hysteresis phenomena which develops as
creep becomes insignificant , as shown , Fig . 2 , is believed to be caused by a
small mechanical lag in the extensometer . Apparatus compressibility was de
termined and deducted from extensometer readings .
An alternate procedure for determination of Es is to load the soil axially
up to some desired percentage of the failure load , then to maintain this load
constant for an extended period of time. Fig . 3 shows such a test conducted
as part of the 1949 WES research (1) , in which a load corresponding to about
two -third of the failure stress was held constant for 69 days . The specimen
was next partially unloaded and reloaded and the straight - line slope of the re
sulting stress - strain curve is also referred to as Eg.
Dynamic Modulus of Elasticity (Ed) is computed from theories of elasticity
based on laboratory determination of the resonant frequency of soil specimens
subjected to forced vibration . Vibration methods have been used extensively
in metallurgy , geology , and concrete technology , but limited application is
found in the field of soils .
Japanese investigators (5 ) describe theory , apparatus , methɔds and results
of tests to determine the elastic constants for cohesive soil specimens . In
general , the soil specimen was placed upright on a diaphragm which was vi
brated ( longitudinal and torsional ) with known frequencies . The frequency was
varied until the sample developed resonance which was determined by observ
ing a peak amplitude of vibration at the top of specimen . The resonant fre
quencies were then used to compute appropriate elastic constants , including
modulus of elasticity , modulus of rigidity and Poisson ' s Ratio . Ine lastic be
havior of short soil specimens at high frequencies was attributed to " solid
viscosity ” properties of the soil .
English investigators at the Road Research Laboratory (6 ) describe an ap
paratus and method for measuring the dynamic modules of elasticity for small
specimens of cohesive soil . Specimens were held centrally in a vertical posi
tion , and longitudinal vibrations were produced by the impact of a small steel
ball falling on the top of the specimen . The frequency of vibration was mea
sured at the base and used to evaluate the modulus of elasticity .
Stress - tons per sq ft

. 0. 2 0. 3 0. 4 0.5 0.6 0. 7 Os
Loaded in increments every 10 minutes
-
%
Elapsed time of 69 days
Strain
psi
Es
050
, 4
-
-
-
-
:
t
at
Unloaded and reloaded
one minute intervals
-
Fig Alternate Procedure for Determination

of
3
.
.
Static Modulus Elasticity ES

of
DrivingRod
-
Bross Stem
- -
Aluminum Clamp
CothodeRay
HOV Oscilloscope
Detail
Tie Rods
CrystalPhono
To
Pickup Oscilloscopo RubberSuspension

RubberSuspension Cop CrystalPhonoPickup
Specimen Encased Cap
-
Rubber Specimen
in
Encased
Membrone Standfor RubberMembrane
in
VacuumLine
Phono
To
VocuumPump Bross Plate DrivingRod

Pickup Brass
Aluminum То Plate Speaker
е
TIOV. Diaphragm Vacuum Driver

AC Amplifier SpeakerDriver Pump System
A
SeeDetail
te
•
Amplifier System
BrassStem Bose
LeadBose Plate
To
Amplifier
-
Signal
HOV
Generator
Longitudinal Vibration Apporotus Torsional Vibration Apparatus
Fig Diagram Vibration Apparatus

of
Schematic
4
.
.
As part vibration apparatus pat

at
the 1950 WES research Harvard

of
,
a
terned after that developed the Japanese Earthquake Institute was designed
at
Rather limited tests were performed

on
and constructed both sands and

(2
).
clays and the results are included herein

,
.
Fig . 5. Vibration Apparatus for Determination of

Dynamic Modulus of Elasticity (Ed)
A
-Line
TypicalGumboClays-
Plasticity Index
TypicalGlacial Clays
AT TypicalOrganic Silts and Clays

-
I
-_ -
LE art
*-_
SX
20
40
60
80
100 120 140

Liquid Limit
Fig Plasticity Chart

6
.
.
In 1958 the authors undertook similar investigations in connection with

missile - site investigations , and new equipment was designed and constructed
in our laboratory in Seattle (3) . Two types of vibration apparatus are now be
ing used , and these have been modified considerably from equipment develop
ed at Harvard . One type provides longitudinal vibration while the other unit
applies torsional vibration to the soil specimen . The apparatus is shown in
schematic form in Fig . 4 and in the photograph , Fig . 5 . The functioning of the
apparatus is described as follows :
An audio - frequency generator , whose output is amplified by a simple audio
amplifier , supplies a variable frequency voltage to a driver unit adapted from
a loudspeaker . For the longitudinal vibration apparatus , the moving diaphragm
of the driver is direct - connected by an aluminum rod to a clamped rim dia
phragm of aluminum having a natural frequency several times greater than
that of the soil specimen to be tested. For the torsional vibration apparatus ,
the moving diaphragm of the driver is directly connected by an aluminum rod
to an aluminum clamp to provide a torsional twist to the specimen . The soil
is encased in a membrane between a base which is attached to the oscillating
diaphragm or to the torsional brass stem and a lightweight cap . A standard
phonograph crystal, the needle of which rests against the cap , is suspended
by rubber thread from a tie rod or stand and the output of this pickup is re
corded on the screen of a cathode ray oscilloscope .
At the start of a test , the specimen is placed under a vacuum and the os
cillator is adjusted to a low frequency . Since the applied frequency is lower
than the natural frequency of the specimen , the amplitude of vibration of the
free end of the specimen is equal to or less than that of the vibrating disc .
Generally , this does not produce measurable output voltage from the pickup .
By gradually increasing the frequency , a state of resonance is attained at the
peak amplitude . The motion of the free end increases and reaches a maxi
mum and then diminishes with further increase in frequency .
Considerable trouble in earlier testing resulted from vibrations in the
frame . These vibrations produced false peaks in the output readings and were
often confused with vibrations of the specimen . Harmonic frequencies also
prove troublesome at times and an electronic switch connected to two sepa
rate pickups was helpful in studying phase relationships .
The restraint for the specimen being vibrated has been verified to corre
spond to the lower end clamped , the upper end free . For this condition , with
an applied frequency equal to the natural frequency of the soil (resonance ) ,
the following equation is valid for longitudinal vibration .
V = 4fl = ( 1 - 01 ua
2 . 2 . 2.
)
E,9
162 21 ]
Where
V = velocity of wave propagation ( in . / sec . )
+
f = resonance of frequency (cycles / sec )
l = length of specimen (in . )
M = Poisson ' s Ratio
a = radius of
specimen (in .)
Ed = dynamic modulus of elasticity (lbs / in . 2 )
y = unit weight (lbs / in . 3)
g = acceleration of gravity (in . /sec2 )
The units shwon in parentheses are those used in this report . However ,
any consistent system of units may be used with the above equation .
Poisson ' s Ratio may not always be known for the soil being tested . How
ever , for a sample whose length is twice the diameter and for any assumed
value of Poisson ' s Ratio the maximum error introduced by neglecting the
term in parenthesis will not exceed one or two percent , and this error is less
than the calibration errors in the oscillator . Therefore , this term was neg
lected and the following equation used :
Ed
16 f² er
9
The resonance frequency is obtained by observing which imposed frequency

of vibration gives the greatest output amplitude of the crystal pickup . Since
the input amplitude cannot be kept constant , and the output of the crystal is
affected by various experimental factors , the actual output readings by the
cathode ray oscilloscope have no significance .
For torsional vibrations , the same equations are used to compute the dyna
mic shear modulus (Gd ) . From the theory of elasticity the moduli of elasti
city and shear are related by the following expression .
Ga = 2176 ) Ed
Where
u = Poisson ' s Ratio

Computation of Elastic Constants from Seismic Velocity . The seismic
ve locity is related to the modulus of elasticity as follows :
2 ( l gu )
-
=
Ed
d (
(11+u
+p ) ( 1 - 2u )
= Erk
od
Where
Ve is the seismic or longitudinal wave velocity ( in . / sec )

In the above expression the constant K is a function of Poisson ' s Ratio u
which may vary within the limits of 0 to 0. 5 . As Happroaches the limiting
value of 0 . 5 for an incompressible material, k becomes very large . Thus ,
computation of E from the seismic ve locity is not reliable for those materi
als which have a Poisson ' s Ratio of about 0 . 5 . It is believed that fully
saturated clays , especially those that are normally consolidated , do have such
high values of u .
Included in this paper are rather limited data on the relationship between
seismic velocity and the elastic properties as determined from vibration tests .
PROPERTIES OF CLAYS TESTED
Elastic and strength data have been obtained for twelve soil types covering
a wide range of soil characteristics and physical properties . The pertinent
and the Plasticity Chart Fig
for
data these soils are summarized Table

in
.
I
.6
Velocity
to to
Seismic
1500
3400
fps
6500
7000
5400
2250
'
s
.0 .0 . 0. 0 .0 . . .
Poisson
54
50
54
53 61 46
Ratio
62 57
0 000
. 3. . 8. 0. . . .
1 10
10
91
86 82 90 82 8845
125
Elasticity
.5 .8 . .0 . 2. - .2 . . .
Ga Av
.4 .8 9 7
50
35 88 10
65
05 70 50 35
E 9 14 5
.0 .4 .5 3. . 2. . 6. - 2. . . . 0. . ..
11 11 13 18 25
1000 psi
of
. 8
40 91 73 9 5 00 0
70
TESTED
45 83 90 70
8 16 26 28 42 5
.0 .5 . .
Ea 9
40
12 10 15 16 23 30
.9 .5 .5 .0
Modulus
0 6 .7
50
77 30
105
46
68
89
_
M₂ 36 26
. . .. .. . .
OF SOILS
.. 9 0
Compressive Strength
. ... . . .
0 6
Tons per
ft
08 35 20 30 = 4 2 . .1 7. 1. .1 0. 1. 1. .2 = 6 6 6 .
72
72
52
00
86
42
sq 1
14 23 29 AV 0 15 14 0
07
10
36 39
Av
00 01
23 45 9080
3 5 5
R
ccccccccccccc &
PROPERTIES
Type
U R U
-.I . 2. 8. . 5. . . 2. 3. 222 0. .8 6. . 4. 2. - .8
4 .
Content
4 4 9 0 18
Water
+
20
21
400
60 22 16
43
19
-
35 19 20 18 49 45 26 27 15 52
Index
Plast
icity
18 63 11
106 24 17
11 17 18 1320
TABLE
Plastic
26 37
Limit
21 31 21 17 13 16 17 16 242225
4004
Liquid
34 24 35
Limit
132 61
39 94 32 33 3545
Ohio River Sandy Clay
WhidbeyIsland Clay
Mississippi Gumbo
Mexico City Clay
Cambridge Clay
Cucaracha Clay
Birch Bay Clay

DaheBentonite
BearpawShale
DuwamishSilt
Montana Clay
Name
Idaho Clay
.
Shale
2 c D G H J
K1
B1
No F1 11
Li
A . Cambridge Clay used in this investigation is a gray - green , medium
plastic , inorganic glacial clay, medium stiff and brittle in the undisturbed
state , soft and sticky when remolded . Preconsolidation pressure , believed
due to effects of surface drying , is in the range of 2 . 0 to 2 . 4 tons per square
foot, as determined by a consolidation test .
B . Cucaracha Clay Shale was obtained from the Panama Canal Zone (7) .
It may be described in general as a mottled gray - green and dark gray, soapy ,
slickensided clay - shale . The material in different cores and occasionally in
the same core varied in color , hardness and jointing . All samples contained
slickensided surfaces of varying dimensions , some apparently open and unce
mented , others closed and apparently recemented .
C. Mexico City Clay is an extremely compressible , bentonite clay of vol
canic origin , light gray -brown in color , and having a natural water content of
approximately 400 %. It is medium soft , yet brittle , in the undistrubed state ,
and extremely soft and sticky when thoroughly remolded . The liquid limit is
of the same order as the natural water content , and the clay plots well above
the A - line on the plasticity chart . The estimated preconsolidation load as
determined from a consolidation test is 1. 1 to 1. 2 per square foot .
D . Mississippi Gumbo is a highly plastic (CH ) clay from the alluvial val
ley of the Mississippi River . The upper strata have been oxidized and partial
ly dried , but the sample tested was from depth 30 feet and presumably had not
been dried . The preconsolidation load as determined from a consolidation
test was in the range of 1. 8 to 2 . 5 tons per square foot . Failure was often ac
companied by bulging and tilting of the specimen , the clay being quite insen
sitive and non - britt le .
E . Oahe Bentonite was obtained from a seam of bentonite in the vicinity
of Oahe Dam , Pierre , South Dakota . This bentonite has a soapy feel , is fairly
hard and very brittle . It is mott led , gray - green in color , and generally tend
ed to part rather easily along inclined planes which appeared to be existing
joints . When exposed to surface drying , fine hair cracks appeared on the sur
face and small fragments bounded by such hair cracks could readily be de
tached . When thoroughly remolded and mixed with water , this material re
verts to a highly plastic bentonite clay . Unfortunately , the Atterberg limits
of this particular sample were not obtained .
F . Ohio River Sandy Clay was obtained from the flood plain of the Ohio
River near Mariemont , Ohio . It is a fairly homogeneous , light brown sandy
clay of low plasticity (CL ), medium stiff and somewhat brittle in the undis
turbed state and soft when remolded .
G . Bearpaw Shale was obtained from the South Saskatchewan River Project
in Canada . It is a clay - shale of Upper Cretaceous Age and the sample was of
a highly plastic , slightly organic , fat clay , dark gray in color , of stiff consist
ency in the undisturbed state , high dry strength , and considerable toughness
at the plastic limit . Although there is litt le doubt that this clay - sha le has or
ganic admixtures , it is located on the plasticity chart well above the A - line .
It is possible that the influence of the organic colloids is overshadowed by the
large content of montmorillonite .
H . Duwamish Silt is a recent alluvial deposit of the Duwamish River in
Seattle , Washington . It is highly organic and was obtained below the ground
water table at depth 10 feet . It is normally consolidated . The specimen test
ed was first consolidated in the triaxial apparatus under an all -around stress
of 1. 0 tons / square foot .
I. a normally consolidated
Birch Bay Clay is inorganic glacial clay from
Birch Bay , Washington , containing small pieces of gravel dispersed through
out . It is quite similar in most respects to the Cambridge Clay (Soil A ) and
to the Whidbey Island Clay ( Soil K ) . The clay is remarkably uniform in soil
properties from just below the weathered surface zone to a depth of in ex
cess of 150 feet .
J.Montana Clay is from the vicinity of Great Falls , Montana , and is be
lieved to be a weathered glacial till . It is a brown plastic clay of very stiff
to hard consistency , containing small gravel pieces and occasional stones up
to several inches in diameter . It also contains pockets of gypsum , limonite
and calcite . No ground - water was encountered at depths up to 83 feet , and
from just below the surface zone down to the bottom of the hole the clay is
remarkably uniform in both strength and other characteristics . Based on con
solidation tests , the clay appears to have been preconsolidated under a stress
of about 10 tons per square foot , either by partial drying or by ice .
K . Whidbey Island Clay is from a soft deposit of blue glacial clay and the
sample tested was obtained within a few feet of the ground surface .
L . Idaho Clay is from lenses of silt and clay encountered in arkosic allu
vial outwash deposits of silty sands in the vicinity of Mountain Home , Idaho .
CONSOLDATION PRESSURE AND STRENGTH RELATIONSHIPS
Previous investigations on sands using both vibration and repetitive load

ing procedures indicated rather conclusive ly that the modulus of elasticity of
sands depended primarily on the relative density and on the normal pressure
between the grains . For studying the elastic properties of clays it was de
cided to first group the clays according to their plasticity characteristics and
for clays having similar plasticity characteristics to investigate the relation
ship between effective consolidation pressure and modu lus of elasticity .
Fig . 7 shows typical failure envelopes for Q, R and S tests on normally
consolidated , fully - saturated inorganic clays . For such clays , the Q envelope
will normally be a horizontal line , the R envelope will be inclined at an angle
OR of about 15° above the consolidation pressure , and the S envelope will be
inclined at about 30° . This simplified presentation neglects the effect of pre
stress induced in R tests (8) , (9) , and effects of rate of loading (7) , (10 ) , etc .
It may be seen that the compressive strengths of a clay determined from
unconfined compression (U ) tests , Q tests , or R tests at or below the consoli
dation pressure , all have about the same value , and that this strength is de
pendent upon the inclination of the R envelope and the magnitude of the effec
tive consolidation pressure (P ) . Therefore , it is reasonable to conclude that
there may be a relationship between compressive strength and elastic proper
ties , and that this same relationship may hold true for all clays having similar
R envelopes . In the inset of Fig . 7 is plotted a theoretical relationship be
tween the angle of inclination ºr of the R envelope , and the ratio of the ef
fective preconsolidation pressure to the compressive strength (Pc / o ). This
ratio diminishes quite rapidly as op increases .
The relationship between preconsolidation stress , compressive strength
and modulus of elasticity were investigated on a specimen of Whidbey Island
Clay ( Soil K ) . The tests were conducted in the following manner .
1. The specimen was consolidated in a triaxial chamber under a hydro

static stress of 0. 5 ton per square foot , which is known to be greater than
Envelope
S
10 15 20 25 30 35 40 45
PR - degrees
EffectiveConsolidation
Pressure
Shear Stress
(2) 1 R Envelope
11PR
Envelope
Q
-
UnconfinedStrength Rapid Triaxial Quick Triaxial
-
Strength Strength
Slow Triaxial Strength
Normal Stress
Fig Typical Failure Envelopes for and Tests

Q
,
R
7
S
.
.
20
000
,
Strength
15
000
,
Compressive
-
Pressure
psi
Consolidation
10
Es 000
,
and
Ed 000 Legend
5,
Ed
+•
Es
OO
4
sq sq
Compressive Strength tons per

ft ft
- -
Consolidation Stress tons per
Fig
on
of
Effect Consolidation Stress Elastic

8
.
.
Properties Whidbey Island Clay

of
the " in situ ” effective preconsolidation stress . After primary consolida
tion was complete , the static modulus of elasticity (Es ) was determined by
repetitive axial loading at small stress .
2. The specimen was carefully removed from the chamber and set up on
the vibration apparatus , with an applied vacuum equal to the original cham
ber pressure , and Ed and Gd determined .
3 . The specimen was again set up in the triaxial chamber and consolidated
under a higher pressure , following which Es , Ed and Gd were again deter
mined . This procedure was repeated several times at successively higher
pressure .
4 . Finally , the specimen was failed by axial loading in an R test , defining
an R envelope of 11°.
In Fig . 8 are plotted Ed and Eg vs both consolidation stress and compres

sive strength . The excellent agreement between Ed and Es (the results are
within the limits of accuracy for the procedure used to determine E . ) , and the
consistency of the data indicate that there is a very definite relationship be
tween the modulus of elasticity and the consolidation stress .
SUMMARY OF RESULTS FOR ALL CLAYS INVESTIGATED
In Fig 9 are plotted values of Ed, Mi and Es for all clays with compressive
.
strengths of less than 7. 0 tons per square foot . These data establish a rather
well - defined relationship which agree very closely with the data shown in Fig .
8. However , the Mexico City Clay ( C) and the Duwamish Silt (H ) , both of which
differ radically from the other clays in plasticity characteristics (See Fig . 6)
fall well below the average line . Two samples of Idaho Clay ( L ) fall well
above the average Line , but there is some question as to the reliability of these
particular data .
In Fig . 10 , the data of Fig . 9 have been extended to cover all clays tested .
the lower range has been taken from Fig fit
III
Curve and extended

to
in
9
,
.
three specimens with compressive strengths

up
of
Cucaracha Clay Shale

B
(
)
29 tons per square foot two specimens

of
on
Also shown are tests

to
Oahe
.
Bentonite and one on Bearpaw Shale These limited data appear de

to
(G
E
(
IV )
)
.
fine Curve which may apply the highly plastic bentonitic clays and shales
to
Note that the data from the very unusual Mexico City Clay fall even
be
low
curve
IV
.
The Duwamish Silt was the only highly organic silt tested and the data
,
(H
)
do
from the one test not warrant conclusions as probable behavior

of
other
to
organic materials
.
Also shown Fig Curve represents test data

10
are Curves and

on
in
II;
.
well graded gravelly sand up maximum size Curve represents data

to
a
II
"
specimens medium density Standard Ottawa sand The original data

of
on
3
(
)
.
the relationship between the major principal effective

as
were developed
stress and order show on comparative plot with clays the data
to
In
E
,
.
have been recomputed on the basis the compressive strength the sample
of
would have developed had been loaded failure unchanged minor princi
to
at
it
pal stress Thus the strengths correspond the envelopesshown Fig

to
in
,
7
S
.
.
.
PROBABLE ERRORS
lastic properties
of
The the soils tested were such that the frequency cor
e
responding maximum response could plus

of
be
determined within limits

or
to
432
50
000
,
Legend
- 40
, Ed
psi
000
омі
Es
+
Elasticity
30
000 Curve
,
of
20
,0
Modulus
10
000
,
7
6
5
4
3
2
sq
ft
-
Fig Relationship Elasticity

of
of
Modulus
to
9
.
.
Compressive Strength Clays

in
gravels
100000
,
and
sands
graded
graded
clays
,, --
1906 es
-
residual
80
000
,
,
psi
well
clays
well
Compact
moact
sands gumbo
,
I
Elasticity
clays
60
000 graded
,
and
Poorly
silts
silts
of 51
, I.
clays
Deoanic
II,
Inorganic
40
Modulus
000
,
bentonitic
ET
plastic
Highly
, V
000
HA
20
,
Dane od
9
12
16
20
24
28
3
2
sq

ft
-
Relationship Elasticity Compressive Strength

of
Fig
10
Modulus
of
to
.
.
minus one percent , but possible variations in the calibration of the oscillator
could have increased this to an error of two percent . In addition , small but
indeterminate errors are introduced by the weight of the cap , membrane ef
fects , effect of traction forces on the bottom of the specimen , effect of pick
up restraint , and by simplifying assumptions made in the derivation of the
theory . It is reasonable to conclude that not all these errors are cumulative ,
and that the probable error in any one frequency determination is of the order
of two percent . The length and the unit weight can be determined with an ac
curacy of at least one percent. Since the elastic moduli are proportional to
the frequency squared , the length squared and the unit weight , it is probable
that the maximum error in any one determination is of the order of 5 to 10 %
Poisson 's ratio was computed from the relationship
Ed -1
=
2Gd
Since many of the errors will be of the same percentage for both Ed and
Gd it is probable that the maximum error in the ratio does not exceed 5% ,
,
but even this small an error results in a 5 %, error in Poisson ' s ratio . As
may be seen from Table I , computed values for k for soft clays range from
0 .46 to 0.62 , whereas true values must be less than 0.50 . Therefore , techni
ques and equipment must be improved before the method can be used for the
reliable determination of Poisson ' s ratio .
SE ISMIC DETERMINATION OF YOUNG ' S MODULUS
The limited data on seismic velocities included in Table I have been ana
lyzed, but no consistent relationship between the seismic ve locity and Young ' s
Modulus has been found . Therefore it is concluded that Poisson ' s Ratio bears
no relationship to strength . In order to compute Young ' s Modulus from the
compressional wave ve locity one must determine either Poisson ' s Ratio or
the shear wave velocity . For large values of Poisson ' s Ratio there may be
considerable error in the computed value of E , un less his known precise ly ,
and this is seldom the case . A preferred method , according to Evison (11 ),
is to determine the shear ve locity instead of the compressional velocity , since
the shear velocity is not as sensitive to small changes in Has u approaches
the limiting value of one - half .
SIGNIFICANCE AND LIMITATIONS OF ELASTIC PROPERTIES
There are many applications in the field of soil mechanics and foundation
engineering in which knowledge of the elastic properties of the foundation
soils is required . Determination of the elastic properties by means of the
procedures outlined in this paper will eliminate the overconservative practice
of using data from load tests , plate bearing tests or triaxial tests , since all
of these tests include creep effects .
The test data presented herein are believed applicable only to transient
loads of low stress intensity . Limiting ranges of stress and duration have not
been established .
It is possible that in evaluating the lateral resistance of piling to dynamic
loads the modulus of horizontal subgrade reaction could be based upon appli
cation of elastic theory .
SUGGESTIONS FOR FURTHER RESEARCH
The test data presented

in this paper were obtained from tests on undis
turbed samples
of clays , silts and shales . These soils were either fully satu
rated , or in a state approaching complete saturation . It is suggested that fur
ther research be undertaken to study the elastic properties of compacted soils ,
and of soils which are only partially saturated .
It is pointed out that none of the data presented herein have been substanti -
ated by full - scale field tests .
Final Report to U . S . Waterways Experiment Station on Investigation of

Effect of Long - Time Loading on the Strength of Clays and Shales at Con
stant Water Content . By A . Casagrande and S. D. Wilson . (Unpublished ) .
Harvard University , Pierce Hall , Cambridge, Mass . , July , 1949 .
2. Report to U. S . Waterways Experiment Station on the 1949 - 1950 Program

of Investigation of Effect of Long - Time Loading on the Strength of Clays
and Shales at Constant Water Content . By A . Casagrande , J. M . Corso and
S. D . Wilson . (Unpublished ) . Harvard University , Pierce Hall , Cambridge ,
Mass ., July , 1950 .
3. “ Dynamic Triaxial Tests on Sands , ” by W . L . Shannon , George Hamane ,
and Rudy J. Dietrich . Proc . First Panamerican Conference on Soil Mech
chanics and Foundation Engineering, University of Mexico , September ,
1959 .
4. " Tests on London Clay from Deep Borings at Paddington , Victoria and the
South Bank ,” by A . W . Skempton and D. J . Henkel, Proc . Fourth Interna
tional Conference on Soil Mechanics and Foundation Engineering , London
1957 .
5. M . Ishi moto and K . Lida “Determination of Elastic Constants of Soils by
Means of Vibration Methods , ” Bull . Earthquake Research Institute , Vol.
15 ( 1937 ), p. 67.
6. " An Electrical Method of Measuring Young ' s Modulus for Small Samples of
Cohesive Soil , " Road Research Laboratory , Dept . of Scientific and Indus
trial Research , March , 1945 .
7. " Research on Stress - Deformation and Strength Characteristics of Soils
and Soft Rocks Under Transient Loading , ” by A . Casagrande and W . L .
Shannon , Soil Mechanics , Series No. 31 , Harvard University , June , 1948 .
8. " Prestress Induced Consolidated - Quick Tmaxial Tests , " by A .
in
Casagrande and Wilson , Proc . of the Third International Conference
S. D.
on Soil Mechanics and Foundation Engineering , Switzerland , 1953 , Vol. I ,
pgs . 106 - 110 .
9. " The Relation between Shear Strength and Effective Stress , " by R . C.
Hirschfeld . Proc . First Panamerican Conference on Soil Mechanics and
Foundation Engineering , University of Mexico , September , 1959 .
10 . " Effects of Stress History on the Strength of Clays , " by A . Casagrande

and S . D . Wilson , Soil Mechanics , Series No. 43 , Harvard University ,
June , 1953 .
11 . “ The Seismic Determination of Young ' s Modulus and Poisson 's Ratio for
Rocks in Situ ,” by F . F .
Evison , Geotechnique , Vol. VI, No . 3, September ,
1956 .
THE RELEVANCE OF THE TRIAXIAL TEST TO THE SOLUTION
OF STABILITY PROBLEMS
and
By Alan W . Bishop1 Laurits Bjerrum2
INTRODUCTION
1
.
The purpose the present paper show how the actual properties
of
is
to
of
cohesive soils measured the standard undrained consolidated undrained

in
-
and drained triaxial test are applied the more important
of
the solution
stability problem encountered by the practicing engineerto
classes
of
The
.
failure criteria chosen and the shear parameters
by
which they are expressed
found most convenient and most appropriate sta
of
are those the methods
to
bility analysis used the practical shear parameters
of
The relation the
to
.
more ba sic shear parameters proposed for example by Hvorslev 1937
is
3
,
,
)
outside the scope of the present paper and discussed elsewhere Skempton
is
(
and Bishop 1954 Bjerrum 1954
b
;
,
)
.
The practical shear parameters serve take full account the principal
of
to
differences between cohesive soils and other structural materials such as
,
the dependence strength on the state of stress and
on
the conditions of
of
drainage
.
Although the purpose the paper present the logical relationship

is
of
of
to
various standard tests the different classes stability problem at

of
to to
the
tention must also be drawn the apparatus ,
of
the various limitations which

usually performed in
the triaxial test These include non uniformity of stress
is
-
.
and strain particularly at large deformations and the inability of the appara
,
simulate the changes the principal stresses which occur -

of
tus direction
to
in
many practical problems To enable the quantitative importance of these

in
perspective emphasis the direct corre

on
limitations be seen laid

is
to
in
lation between laboratory tests and field observations stability instabili

or
of
wherever case records are available

ty
)
may seem the practicing engineer that many his problems are too
of
to
It
small scale or are soils too lacking homogeneity apply detailed

to
in
in
in
quantitative methods stability analysis even for the application

of
However
,
.
semi empirical rules important determine into which class the

is
it
to
of
stability problem falls

.
THE PRINCIPLE OF EFFECTIVE STRESS

2
.
One of the main reasons for the late development of Soil Mechanics as
a
Civil Engineering has been the difficulty recognising

of
systematic branch
in
Imperial College
of
Reader Soil Mech Science and Tech Univ

of
in
.,
.,
1
.
London England
,
Dir Norwegian Geotech Inst Oslo Norway

.,
.,
,
3 2
. .
Items indicated thus Hvorslev 1937 refer corresponding entries listed

to
,
),
(
alphabetically the Appendix Bibliography

in
437
that the difference between the shear characteristics of sand and clay lies
not so much in the difference between the frictional properties of the com
ponent particles as in the very wide difference - about one million times - in
permeability . The all - round component of a stress change applied to a satu
rated clay is thus not effective in producing any change in the frictional com
ponent of strength until a sufficient time has elapsed for water to leave (or
enter ) , so that the appropriate volume change can take place .
The clarification of this situation did not begin until the discovery of the
principle of effective stress by Terzaghi ( 1923 and 1932 ) and its experimental
investigation by Rendulic ( 1937 ) . An examination of current design methods
might suggest that the impact of Terzaghi ' s discovery had yet to be fully felt .
For soil having a single fluid , either water or air , in the pore space , the
principle of effective stress may be expressed in relation both to volume
change and to shear strength :
(a ) The change in volume of an element of soil depends , not on the
change
in total normal stress applied , but on the difference between the change in
total normal stress and the change in pore pressure . For an equal all - round
change in stress this is expressed quantitatively by the expression :
AV /V = - C.(40 - Au ) ( 1)
where AV / V denotes the change in volume per unit volume of soil ,

Ao denotes the change in total normal stress ,
Au denotes the change in pore pressure ,
and Cc denotes the compressibility of the soil skeleton for the particu
lar stress range considered .
It may be noted in passing that this equation shows that a decrease in pore
pressure at constant total stress is as effective in producing a volume change
as an increase in total stress at constant pore pressure , a fact which is con
firmed by field experience .
(b ) The maximum resistance to shear on any plane in the soil is a function ,
not of the total normal stress acting on the plane , but of the difference be
tween the total normal stress and the pore pressure . This may be expressed
quantitatively by the expression :
tan
'
Tq= c + (0 - u)
(2
)
the plane
on
where denotes the shear stress failure

te
at
denotes the apparent cohesion terms

of
effective
in
,
į |
c
'
Di denotes the angle shearing resistance

of
stress
,
the plane considered

on
denotes the total stress

o
a
,
and denotes the pore pressure

u
cases the effective normal stress thus the stress difference

is
In
both
by
usually denoted the symbol

,
0
o
-
'.
The validity the principle effective stress has been amply confirmed
of
of
,
by
for
the experimental work Taylor

( of
saturated soils Rendulic 1937

,
),
(
by
1944 Bishop and Eldin 1950 and Laughton 1955 and indirectly the
),
),
;
b
)b
(
Stresses and pressures are here considered as mea sured with respect
to
a
.
atmospheric pressure gauge pressures

as
zero The actual datum

e
).
.
.
(i
does not of course affect the value of the effective stress

.
given these papers

of
of
treatment the influence contact area

is
in
A
b
.
.
TRIAXIAL TEST 439
field records referred to in later sections of this paper . For partly saturated
soils , however , a more general form of expression must be used , since the
pore space contains both air and water which may be in equilibrium at widely
different pressures , due to surface tension . A tentative expression has been
suggested for the effective stress under these conditions ( Bishop , 1959b ;
1960 ) , of the form :
Ó = 0 - u, + x (ui - Up) (3)

where un denotes the pressure in the air in the pore space ,
u2 denotes the pressure in the water in the pore space ,
and X is a parameter closely related to the degree of saturation S and
varying from unity in saturated soils to zero in dry soils .
The parameter x and its values under various soil conditions are discussed
in more detail elsewhere ( Bishop , 1960 ; Bishop , Alpan , Blight , and Donald ,
1960 ) . It may be noted in passing that for a given soil condition , the value of
x measured in relation to shear strength may differ from its value measured
in relation to volume change . However , the large positive pore pressures
likely to lead to instability in rolled fills will in general only occur if the de
gree of saturation is high , where x may be equated to unity with little error .
The additional complication of observing or predicting pore air pressure may
therefore hardly be justified in such cases.
In most stability problems the magnitude of the body forces and of the ap
plied loads is known quite accurately . It is in the magnitude of the shear
strength that the main uncertainty lies and it is therefore useful to examine
the variables controlling the value of Tf in equation (2) .
The magnitude of the total normal stress o on a potential slip surface may
be estimated with reasonable accuracy from considerations of statics . The
shear parameters c ' and Ø are properties which depend primarily on the soil
type and to a limited extent on stress hi story ( see Table I in section 6 ) . Pro
vided representative samples are taken and tested in the appropriate stress
range , little error need arise in evaluating c ' and '. This aspect of any in
vestigation does , however , call for sound judgment and a knowledge of geology .
It is in the prediction of the value of the pore pressure u that in many
problems the greatest uncertainty lies . The development of cheap and relia
ble field devices for measuring pore pressure in soils of low permeability
has , however , transformed the situation as far as the practicing engineer is
concerned by enabling predictions to be checked and a control to be kept on
stability during construction work .
Much of the uncertainty about the pore pressure prediction has arisen
from a failure to distinguish between the two main classes of problema :
( a) Problems where pore pressure is an independent variable and is con

trolled either by ground water level or by the flow pattern of impound
ed or underground water , for example , and
(b) Problems in which the magnitude of the pore pressure depends on the
magnitude of the stresses tending to lead to instability , as in rapid
construction or excavation in soils of low permeability .
C. See , for example , Casagrande , 1949 ; U . S. B . R . , 1951 ; Penman , 1956 ;

Sevaldson , 1956 ; Kallstenius and Wallgren , 1956 ; Bishop , Kennard , and
Penman , 1960 .
d. This distinction is discussed in detail by Bishop ( 1952 ) .
problems which initially fall into class (b ) the pore pressure distribution
In
will change with time and at any point the pore pressure will either decrease
or increase to adjust itself to the ultimate condition of equilibrium with the
prevailing conditions of ground water level or seepage . The rate at which
this adjustment occurs depends on the permeability of the soil (as reflected
in its coefficient of consolidation ) and on the excess pore pressure gradients ,
which depend both on the stress gradients and on the distance to drainage
surfaces .
The least favourable distribution of pore pressure may occur either in the
initial stage or at the ultimate condition or , in special cases , at an intermedi
ate time depending , for example , on whether load is applied or removed , and
on other specific details of the problem , as discussed in section 6 .
3. PORE PRESSURE PARAMETERS
Inslope stability problems the influence of pore pressure on the factor of

safety is most conveniently expressed in terms of the ratio of the pore
pressure to the weight of material overlying the potential slip surface . This
ratio was used by Daehn and Hilf (1951 ) in the form of an overall ratio of the
sum of resolved components of the pore pressure and of the weight of soil , to
express the results of the stability analysis of four earth dams , based on the
field measurement of pore pressure .
Bishop ( 1952 and 1954 b ) showed that, for a slope in which the ratio of the
pore pressure u to the vertical heade of soil yh above the element considered
was a constant , the value of the factor of safety F decreased almost linearly
with increase in pore pressure ratio u / yh . Subsequent work by Bishop and
Morgenstern (in course of preparation for publication ) has shown that, both
for pore pressures obtained from flow patterns (class (a ) problems ) and for
those obtained as a function of stress ( class (b) problems ) , the average value
of the pore pressure ratio u / yh is the most convenient dimensionless parame
ter by which to express the influence of pore pressure stability ( Fig . 1) . The
ratio is denoted ry :
Where the pore pressure is independent of stress , its value is obtained di
rectly from the ground water conditions or flow net and expressed as the
all
averagef value of ru . In other cases class problems the ratio must

b
(
(
)
either be obtained from field measurements or predicted from the observed

relationships between pore pressure and stress change under undrained con
ditions and from the theory of consolidation This turn necessitates an
in
.
the stress distribution within the soil mass

of
estimate
.
For the inclusion the laboratory results the stability calculation

it
is
of
in
express them terms pore pressure parameters The de

of
convenient
of to
in
velopment these parameters and their application practical problems

to
detail elsewhere Skempton 1948 Bishop 1952 Skempton

is
described
in
b
;
;
,
,
(
1954 Bishop 1954 Bishop and Henkel 1957 Bishop and Morgenstern
,
,
;
;
a
1960
).
the average bulk density the soil and vertical

is
of
of
the distance the

y
h
e
.
surface above the element

.
by
Details the averaging method are given Bishop and Morgenstern

of
,
f.
1960
.
the possibility pore pressure changes clays

of
Attention was drawn

to
g
in
.
under the action of deviator stress before proved practicable

to
, it
a
measure them erzaghi 1925 Casagrande 1934

,
).
(T
TRIAXIAL TEST 441
3.00
SLOPE = 3 :1
DEPTH FACTOR - 1. 00
Ø - 30°
SH - 0.05
200
1.00
0. 2
.
0 2
04
0. 4 .
0 6
0. 8
.
0 8
loO
Fig . 1. - The linear relationship between factor of safety and pore

pressure ratio for a slope or cut in cohesive soil .
For a change in stress under undrained conditions the change in pore

pressure may be expressed as Au , where
Au = B (A03 + A (A01- 403)]
where 401 denotes the change in major principal stress ,

103 denotes the change in minor principal stress ,
(in both cases total stresses are considered )
and A and B denote the pore pressure parameters (Skempton , 1954 ) .
Triaxial compression tests show that for fully saturated soils B = 1 to

within practical limits of accuracy , and that the value of A depends on stress
history and on the proportion of the failure stress applied . This is illustrated
in Fig . 2 . where the A values for normally and overconsolidated clay are
given . The values of A at failure are seen to be very dependent on the over
consolidation ratio (defined as the ratio of the maximum consolidation
pressure to which the soil has been subjected to the consolidation pressure
immediately before the undrained test is performed ) .
NORMALLYCONSOUDATED OVER- CONSOUDATED

CONSOLIDATION
PRESSURE. P CONSOLIDATION
PRESSURE P
OVER- CONSOUDATION
RATIO. /
g - o
AXIAL STRAIN AXIAL STRAIN
AU AXIALSTRAIN
AXIAL STRAIN
los
STRAIN
AXIAL
AXIALSTRAIN AXIALSTRAIN
0
.5
10
(d
)
0
.3
.
..
16
OVER CONSOLIDATION
RATIO
-
Fig The dependence the pore pressure parameter

of
2
--
.
.
stress history
on
.
TRIAXIAL TEST 443
For partly saturated soils the value of B lies between 0 and 1 depending on
the degree of saturation and the compressibility of the soil skeleton . Typical
values of A and B are used in section 6h .
It should be noted that equation 4 takes no account of the change in the
intermediate principal stress 102 , or of possible changes in the directions of
the principal stresses . In the majority of stability problems the conditions
approximate to plane strain , in which the intermediate principal stress does
not equal the minor principal stress as in the standard cylindrical com
pression test . Theoretical studies ( Skempton , 1948 a and b ; Hansen and
Gibson , 1949 ; Bishop and Henkel , 1957 ) indicate that the form of the equation
remains the same , but that the triaxial test underestimates the value of A ,
and the limited amount of test data so far available (Wood , 1958 ; Cornforth ,
1960 ; Henkel , 1960 ) supports this view . Little is yet known about the influence
of the rotation of the principal stresses on the value of A . The importance of
these limitations in practice can at present only be assessed from the over
all check with observed pore pressures in the field .
It should also be noted that the principle of superposition can be applied to
pore pressure changes in soil only in a very restricted sense . Where the
purpose of the test is the accurate prediction of pore pressure at states of
stress other than failure , a more accurate result is obtained if the stress
increments occurring in practice are closely followed in the test by making
simultaneous changes in the values of both on and oz . The test result is then
conveniently expressed in terms of the relationship between pore pressure
and major principal stress , for the specified stress ratio , using the ex
pression :
Au = 3 . 20
(5 )
The influence of stress ratio on this parameter is illustrated in Fig . 3 for

a compacted earth fill .
It should be noted that the value of the parameter B only gives the change
in pore pressure due to stress change under undrained conditions . The actu
al pore pressure depends also on the initial value up before the stress change
is made ( Fig . 4) , and is given by the expression :
u = ug + Au (6)
i .e. u = 4, + B . 10 ,
(7)
In natural strata uo is determined from the initial ground water conditions ,
being positive below ground water level and negative above . In rolled fill the
initial value is usually negative , reaching quite high values in cohesive soils
placed at or below the optimum water content ( Hilf , 1956 ; Bishop , 1960 ;
Bishop , Alpan , Blight , and Donald , 1960 ) .
In cases where no dissipation of pore pressure is assumed to occur , the
pore pressure ratio ru used in the stability analysis is directly related to B :
1,1 = u / yh = l / yh . (uo + B.Jo, ) (8 )
h. The value of Bisgenerally given with respect to changes in pore water

pressure . A slightly different value relates the change in air pressure
to the change in total stress .
444
+
ver
150
ist
IN
/.
L
PORE
PRESSURE
LB AU
1=
4
0
RATIO
SHEAR STRENGTH
0
K
au
PRESSURE
PRESSURE PORE
JOIAL STRESS
=of
0F
=
S
1
5
PORE
IN
K
= oSf
00 1
0F
_ -.( -.(
=
.-,OK ,.',..) ,','.:.)
CONFERENCE
INCREASE
-
vel
50
100 150
IN
IN
?
LB
INCREASE TOTAL MAJOR
/.,
+
U
PRINCIPAL STRESS AO
=U
.B.
AG
–3
..
Fig The influence principal stress ratio
on
Fig
-.4
the pore pressure parameter

.
boulder Pore pressure change under undrained

+
%
of at
:B.1
.
clay compacted optimum test conditions

TRIAXIAL TEST 445
In the special case of the construction of earth fill embankments the aver
age value of 40 , along a potential slip surface approximates to yh (Bishop ,
1952 ) , and equation (8) becomes :
ry = B + up/ yh (9 )
For earth fills of low plasticity placed wet of the optimum the term up /vh
is small ,and a further approximation is sometimes used in preliminary de
sign :
r, = B ( 10 )
Some typical examples of the use of pore pressure parameters are given
in section 6.
4. STANDARD TYPES OF TRIAXIAL TEST
The type of triaxial test most commonly used in research work and in
routine testing is the cylindrical compression test . A diagrammatic layout of
the apparatus is given in Fig . 5.
The cylindrical specimen is sealed in a thin rubber membrane and subject
ed to fluid pressure . A load applied axially , through a ram acting on the top
cap , is used to control the deviator stress . In a compression test the axial
stress is thus the major principal stress 01 ; the intermediate and minor
principal stresses ( 02 and 03 respectively ) are both equal to the cell
pressure .
Connections to the ends of the sample permit either drainage of water or
air from the voids of the soil or , alternatively , the measurement of pore
pressure under conditions of no drainage .
In most standard tests the application of the allround pressure and of the
deviator stress form two separate stages of the test ; and tests are therefore
classified according to the conditions of drainage obtaining during each stage :
( 1) Undrained testsi . - No drainage , and hence no dissipation of pore
pressure , is permitted during the application of the all - round stress .
No drainage is permitted during the application of the deviator stress
(01 - 03 ).
( 2) Consolidated - undrained tests . -- Drainage is permitted after the appli
cation of the all - round stress , so that the sample is fully consolidated
under this stress. No drainage is permitted during the application of
the deviator stress .
(3 ) Drained tests . —Drainage is permitted throughout the test, so that full
consolidation occurs under the all - round stress and no excess pore
pressure is set up during the application of the deviator stress .
In order to illustrate the inter - relation between results of the different

types of test , saturated and partly saturated soils will be considered sepa
rately .
i. Alternative nomenclatures have been used both in Europe and the U . S . A .

The present terms are considered to be the most descriptive of the test
conditions .
AXIAL LOAD
io
AIR RELEASE
AIR
LOADING RAM
I b
PRESSURE GAUGE RELEASE
VALVE
WWW
RUBBER
—
TOP CAP
.
.
RING
-
POROUS DISC
-
FLEXIBLE TUBE
-
WATER
SAMPLE ENCLOSED
IN
RUBBER
A
MEMBRANE
RUBBER
RING POROUS DISC
-
VI SEALING RING
IM
OR
LTO CELL PRESSURE CONTROL CONNEXIONS FOR DRAINAGE
]
PORE PRESSURE MEASUREMENT
Diagrammatic layout
of
Fig the triaxial cell

5
.
.
-
.
EFFECTIVE STRESSES 283 TOTAL STRESSES

-,
1,
(
)
O
Fig soil stress circles

on
Undrained tests saturated total and effective

6
.
.
.
-
:
TRIAXIAL TEST 447
(a ) Undrained Tests on Saturated Cohesive Soils . -

These tests are carried out on undisturbed samples of clay , silt and peat
as a measure of the existing strength of natural strata , and on remoulded
samples when measuring sensitivity or carrying out model tests in the labo
ratory .
The compression strength (i . e . the deviator stress at failure ) is found to
be independent of the cell pressure , with the exception of fissured clays (dis
cussed in section 6 ) and compact silts at low cell pressures . The correspond
ing Mohr stress circles are shown in Fig . 6.
If the shear strength is expressed as a function of total normal stress by
Coulomb ' s empirical law :
T = Cu totan ,
(11)
where cu denotes apparent cohesion ,
in terms of total stress ,
u denotes angle of shearing resistance ;
it follows that , in this particular case ,
u = 0
(12 )
Cu = (01 - 03)
The shear strength of the soil, expressed as the apparent cohesion , is used
in a stability analysis carried out in terms of total stress , which , for this
type of soil , is known as the Du = 0 analysis (Skempton , 1948 a and b ). Since
the value of cy may be obtained directly from the unconfined compression
test (where 03 = 0) , and from the vane test in the field , it is a simple and eco
nomical test, but is often used without regard to the class of stability problem
under consideration .
For fully saturated soils the increase in cell pressure is reflected in an
equal increase in pore pressure and the effective stresses at failure remain
unchanged . If pore pressure measurements are made during the test only one
effective stress circle is obtained ( Fig . 6) , and tests at other water contents
must be carried out to obtain the failure envelope in terms of effective stress .
In Fig . 7 (a ) an example is given of the changes in pore pressure during
shear in an unconfined compression test and in Fig . 7(b ) the Mohr circles are
given in terms of total and effective stresses .
The A value measured in the undrained test on a sample of natural ground
is very different from the value in situ under a similar change in shear
stress . This results from the stress history given to the sample by changes
in pore pressure which occur during sampling and preparation due to the re
moval of the insitu stresses , quite apart from disturbance due to the sampler
itself . The release of the deviator stress existing in samples normally con
solidated with no lateral yield is a major factor contributing to this effect .
Tests on samples anisotropically consolidated in the laboratory ( Bishop
and Henkel , 1953 ) and on undisturbed samples ( Bishop , 1960 ) show that the
effective stress in the sample when under an all - round pressure or uncon
fined can be less than half the effective overburden pressure in situ . Yet
when the shear stress is increased to bring the sample to failure , the un
drained strength closely corresponds to the in situ strength deduced from
stability analysis or from vane tests . This is consistent with the experiment
al observation that , for a limited range of soil types and stress paths, strength
(07- 07 - 9.92 LB / INY
DE VIATOR 6
STRESS
(01 - 05)
LB / IN
REMOULDED LONDON CLAY
LL =66% .
PL - 274
w - 35.4%
AXIAL STRAIN E %
.
PORE
PRESSURE
LB / IN
Up - 4.5 LB / IN
- 10
Fig . 7a . - Pore pressure change during shear in an unconfined

compression test .
,
TOTAL STRESS EFFECTIVESTRESS
..
9 10
12
NORMAL STRESSop
NORMAL STRESS
- 4.5 P.S.T.
Fig . 7b . -- Total and effective stress circles for the unconfined

compression test .
TRIAXIAL TEST 449
and water content are uniquely related (Waterways Experiment Station , 1947 ;
Henkel , 1959 ).
If it is indeed this fact which provides the empirical justification for the
use of undrained compression tests in the Qu = 0 analysis , then to reconsoli
date the samples in the laboratory under the existing overburden pressure
will inevitably lead to an overestimate of the in situ strength of the soil , since
reconsolidation is almost always accompanied by a decrease water content .
.
in
(b) Consolidated - Undrained Tests on Saturated Soils

These tests are carried out on both undisturbed and remoulded samples of
cohesive soils , primarily to determine the values of c ' and ' , but also to de
termine the values of A and to study the effect of stress history .
In the standard test the sample is allowed to consolidate under a cell
pressure of known magnitude (p ), the three principal stresses thus being
equal. The sample is then sheared under undrained conditions by applying
an axial load . As in the case of the undrained test in the previous section , the
cell pressure at which the sample is sheared does not influence the strength
( except in dilatant silts at low pressures ) as illustrated in Fig . 8e . The test
result , in terms of total stresses , may thus be expressed by plotting the value
of cy against consolidation pressure p, Fig . 8b .
For normally consolidated soils the ratio Cup is found to be constant , its
value depending on soil type . However , strengths measured in undrained tri
axial tests and vane tests on strata existing in nature in a normally consoli
dated state , when plotted against the effective overburden pressure , lead to a
lower estimate of cu / p than is found with samples consolidated under equal
all -round stress in the laboratory . The difference increases as the plasticity
index decreases and appears to be due to two causes:
(i ) A naturally deposited sediment is consolidated under conditions of no

lateral displacement , and hence with a lateral effective stress con
siderably less than the vertical stress . The ratio of the effective
stresses , termed the coefficient of earth pressure at rest , is generally
found from laboratory tests to
lie
the range
35
the lower
in
,
0
.7
0
-
.
values occurring soils with low plasticity index Terzaghi 1925

in
;
,
a
Bishop 1958 Simons 1958 This cause alone can account soils
in
,
;
a
).
of low plasticity for difference for ex

50
cu
of
of
the value
%
in
p
a
/
(
ample Bishop and Henkel 1953 Bishop and Eldin 1953

;
,
)
.
Reconsolidation the laboratory after the stress release associated

in
ii
(
)
with even the most careful sampling technique leads lower void
to
a
ratio than would occur the natural stratum under the same stress
in
the pore pressure parameter particular sensitive

of
The value
is
in
A
the resulting modification soil structure and this turn leads

to
to
in
in
higher undrained strength

a
.
of
For these reasons the use the results consolidated undrained tests
of
,
-
expressed terms of total stress either by the parameter Cu or by the

in
p
/
appendix justified practical applications

be
of
value øcu see can few

in
1
),
.
(
the pore pressure

of
measured during the undrained stage

is
However the
if
,
the effective parameters

of
test the results can be expressed terms

in
,
c
'
and Experience has shown that these parameters can be applied wider
to
0
a
'.
range practical problems

of
The relationships between the total stress pore pressure and effective
,
typical series
of
stress characteristics obtained consolidated undrained

in
a
-
-
9
WATER
CONTENT
----
7-
------ CONSOLIDATION
PRESSURE
D
(b)
-
-
UNDRAINED
STRENGTH -
Gu _
9
_
_ CONSOLIDATION
PRESSURE
D
+
o
0.5
0. sti
CONSOLIDATION
PRESSURE
D
- OS
NORMALSTRESS
EFFECTIVE
TOTALNORMALSTRESS
OAT FAILURE
:
ALLCONSOLIDATED
SAMPLES UNDERPL
Fig . 8. - The relationships between the total stress , pore pressure and
effective stress characteristics for a series of consolidated
undrained triaxial tests on saturated cohesive soil .
TRIAXIAL TEST 451
triaxial tests are illustrated in Fig . 8 . The points , a , b , and c represent

normally consolidated samples ; the point d represents an over - consolidated
sample , the overconsolidation ratio being Pb / pd , Fig . 8a . For normally con
solidated samples the effective stress envelope is a straight line with c ' equal
to zero ( Fig . 8d ) , the value of 0 ' depending on soil type . Over - consolidation
results in an envelope lying a little above this straight line ; the section of this
envelope relevant to any particular practical problem can generally be repre
sented with sufficient accuracy by a slightly modified value of 0 ' and a co
hesion intercept c '.
The most marked effect of over - consolidation is , however , on the value of
A , which , with increasing over - consolidation ratio , drops from a value typi
cally about 1 at failure to values in the negative range ( Figs . 2 and 8c ) . These
low A - values are , in turn , largely responsible for the high undrained strength
values resulting from over - consolidation ( compare point d with point a in
Fig . 8b ).
Values of c ' and ' are usually based on the effective stress circles corre
sponding to maximum deviator stress . However , in some over - consolidated
clays in which large decreases in pore pressure during shear are associated
with very large failure strains , a slightly larger value of $ ' is obtained by
plotting the state of stress at a smaller strain approximating to the point at
which the ratio of the principal effective stresses 01 /03 ' reached its maxi
mum value . The difference in the value of 0 ' is generally not important from
a practical point of view , but in making comparisons between the values of Di
obtained from consolidated - undrained and drained tests it is necessary to
specify which definition of ' is being used ).
( c ) Drained Tests on Saturated Soils . -
Drained tests are carried out on both undisturbed and remoulded samples
of cohesive soils to obtain directly the shear strength parameters relevant to
the condition of long term stability , when the pore pressures have decreased
(or increased ) to their equilibrium values .
In the standard test the sample is allowed to consolidate under a cell
pressure of magnitude p and is then sheared by increasing the axial load at a
sufficiently slow rate to prevent any build - up of excess pore pressure . The
effective minor principal stress 03 ' at failure is thus equal to p , the consoli
dation pressure ; the major effective principal stress 01 ' is the axial stress .
The test results lead directly to the effective stress shear parameters c ' and
0 ' , which fc drained tests are often denoted cd and a
The drained tests also provides data on the volume changes which occur
during the application of the equal all - round stress and the deviator stress .
(d) Inter - Relationship between the three Types of Test on Saturated Soil . --
aspects of this inter - relationship are of practical interest to the engi
Two
neer concerned with stability problems : (1) The degree of reliability with
j . Whether this difference reflects an actual characteristic of frictional ma

terials or merely the increasing nonuniformity of stress in the cylindrical
compression tests at large strains is still open to question . It is , how
ever , clear from tests on sand published by the Waterways Experiment
Station ( 1950 ) and other unpublished tests at the Norwegian Geotechnical
Institute andat Imperial College that for very loose soil structures the
maximum deviator stress may occur at smaller strains than the maximum
stress ratio , and here the difference in 0 (of up to 15° or so ) undoubtedly
represents a physical property of the soil .
which the effective stress envelope defined by the parameters c ' and i can
be assumed to be the same for undrained , consolidated - undrained and drained
tests ; and (2 ) the extent to which volume changes in drained tests are an indi -
cation of the magnitude of pore pressure changes in consolidated - undrained
tests .
In Fig . 9d are compared the results of undrained , consolidated - undrained
,
and drained tests on clay from the foundation of the Chew Stoke Dam (de
scribed by Skempton and Bishop , 1955 ). The close agreement between the ef
fective stress failure envelopes may be noted . It is also of interest to note
from the low values of Af that an undisturbed sample reconsolidated in the
laboratory behaves as though it were 'over - consolidated ' even at cell
pressures greatly in excess of the in situ pre - consolidation pressure . The
intercept c ' will in general not be zero for this part of the failure envelope .
That there should be close agreement between the effective stress enve
lopes for consolidated - undrained and drained tests on normally consolidated
samples has been shown theoretically by Skempton and Bishop ( 1954 ) using
the concept of true cohesion and friction due to Hvorslev ( 1937 ) . Since Ø' is
to some extent time - dependent , it is necessary to use similar rates of testing
in making an experimental comparison , and to ensure adequate time for pore
pressure measurement in the consolidated - undrained test and for drainage in
the drained test . The predicted values of p ' from the consolidated - undrained
test are the higher , but only by 0 - 1° in typical casesk .
However , for heavily over - consolidated clays the position is generally re
versedl and the drained test is usually found to give the higher value , due to
the work done by the increase in volume during shear in the drained test, and
to the smaller strain at failure .
The volume changes in drained tests have for some time been known to
correlate qualitatively with the pore pressure changes in undrained tests .
Experimental data on two remoulded clays have recently been presented by
Henkel ( 1959 and 1960 ) who has described a simple graphical procedure from
which the quantitative relationship may be obtained .
(e ) Undrained Tests on Partly Saturated Cohesive Soils .

These tests are most commonly carried out on samples of earth -fill ma
terial compacted in the laboratory under specified conditions of water content
and density . They are also applied to undisturbed samples of strata which
are not fully saturated , and to samples cut from existing rolled fills and trial
sections .
The compression strength is found to increase with cell pressure ( Fig .
10a ) , as the compression of the air in the voids permits the effective stresses
to increase . However , the increase in strength becomes progressively small
er as the air is compressed and passes into solution , and ceases when the
stresses are large enough to cause full saturation , ºu the approximating to
zero . The failure envelope expressed in terms of total stress is thus non
linear , and values of Cu and u can be quoted only for specific ranges of
normal stress .
k . The sign and magnitude of this difference may change if the failure strains
are very dissimilar , as in long term tests reported by Bjerrum , Simons ,
and Torblaa ( 1958 ) .
1. If the failure envelopes corresponding to maximum deviator stress are
compared .
TRIAXIAL TEST 453
τΙ SILTY CLAY- CHEWSTOKE

LL =43%
50
PL - 25%
w - 397% DRAINEDTEST EFFECTIVE
STRESS
Ft
8022 "
30 CONSOLIDATED
UNDRAINED
TEST
EFFECTIVE
STRESS
20
A=0:341 A- 0.47
Ö 67
UNDRAINED
TEST
A 032 TOTALSTRESS
O 70 80 90 100
10 20 30 40 50 60
NORMALSTRESSO P.S.
Fig . 9. - Undrained , consolidated - undrained and drained tests on undisturbed

samples of Chew Stoke silty clay : maximum deviator stress .
ºu
RANGE OF On
-
TOTAL STRESSES
og
----
=- - - - K
EFFECTIVESTRESSES
-TOTALSTRESSES
-
-- -
--
--
91
ö o
Fig . 10 . - Undrained tests on partly saturated cohesive soil (a) in terms of

total stress , (b) in terms of effective stress .
If the pore pressure is measured during the test , as is usual where field
pore pressure measurements are to be used to check the stability during con
struction , then the failure envelope can be expressed in terms of effective
stress , Fig . 10b . The effective stress envelope is found to approximate very
closely to a straight line over a wide range of stress .
However , rather more difficulty arises in defining accurately the effective
stress envelope for a partly saturated soil than at first apparent. The first
difficulty lies in testing technique , to which attention was drawn by Hilf (1956 ) .
This problem is discussed in detail by Bishop (1960 ) and Bishop , Alpan ,
Blight , and Donald ( 1960 ) , where it is concluded that accurate pore water
pressure measurements can be made in the triaxial apparatus in partly satu
rated cohesive soils provided a porous element of very high air entry value
is used and provided a considerably reduced rate of testing is accepted .
The second difficulty lies in the form of the expression for effective
stress ( equation 3) , which includes a term for pore - air pressure as well as
pore -water pressure for values of the factor x other than unity . The use of
the simple expression for effective stress of total stress minus porewater
pressure leads to an over - estimate of effective stress of ( 1 - x ) (u1 - u2 )
where (u 1 - u2 ) is the difference between pore - air pressure and pore water
pressure . Since values of (u1 - 12 ) of up to 40 lb . per sq . inch have already
been measured on rolled fill in the triaxial test, and the value of x approxi
mates to the degree of saturation , significant errors in effective stress result
from the use of the simpler expression . This is particularly marked near the
origin of the Mohr diagram and may lead to the apparent anomaly of a nega
tive ' cohesion ' intercept ( Bishop , Alpan , Blight , and Donald , 1960 ). However ,
pore pressures set up under construction conditions are only critical if the
water content of the fill and the magnitude of the stresses lead to almost full
saturation , and in this case the error is small enough to be ignored in many
practical problems .
(f ) Consolidated - Undrained Tests on Partly Saturated Cohesive Soils .

These tests are carried out on samples of compacted earth - fill material
and on undisturbed samples . They may be necessary to determine c ' and '
when the degree of saturation of the samples is not low enough to result in a
sufficient range of strengths in the undrained tests to define a satisfactory
failure envelope .
Consolidated - undrained tests in which a backpressure is applied to the
pore space to ensure full saturation before shearing are carried out to ex
amine the effect on the values of c ' and ' of the submergence of fill or foun
dation strata . Back - pressures of up to 100 lb . per sq . inch are often required
to give full saturation on a short term basis .
( g ) Drained Tests on Partly Saturated Cohesive Soils . -

Drained tests are carried out on both compacted and undisturbed samples
to obtain directly the values of c ' and Øl for the condition of long term stabili
ty . Generally a backpressure is applied to ensure full saturation of the
sample before the application of the deviator stress , during which the back
pressure is held constant .
.
(h ) Inter - Relationship between the Three Types of Test on Partly Saturated
Soil
Here again two aspects of this inter - relationship are of practical interest
to the engineer concerned with stability problems : ( 1) The comparison of the
TRIAXIAL TEST 455
values of c ' and ' obtained from the different types of test ; and (2 ) the pre
diction of pore pressure changes from volume changes .
Tests carried out at Imperial College have generally shown that the differ
ence between the values of Ø ' measured in the different types of test are not
very significant from a practical point of view . The value of c ', however ,
tends to correlate with water - content at failure . Where all the samples de
fining a failure envelope show a marked increase in water content in the
consolidated - undrained or drained test with a back - pressure , c ' is generally
reduced . With the lower values of c ' obtained by using the improved pore
pressure technique described elsewhere ( Bishop , 1960 ; Bishop , Alpan , Blight ,
and Donald , 1960 ), the difference in c ' obtained in the different tests are less
marked , and in some soils are not of practical significancem ( Fig . 11 ) . The
range of soil types so far tested using this technique is , however , rather
limited .
It is generally easier to make accurate measurements of pore water
pressure under undrained conditions than to make the necessarily very accu
rate measurements of volume change and degree of saturation on which pore
pressure predictions depend . Studies at the Bureau of Reclamation by
Bruggeman et al . ( 1939 ) , Hamilton (1939 ) , Hilf ( 1948 and 1956 ) have shown
that the change in pore - air pressure can be related to observed volume chang
es by the use of Boyle ' s law and Henry ' s law . However , the magnitude of the
difference between pore - air and porewater pressure still has to be found ex
perimentally . For practical purposes , where the pore water pressure is the
more significant factor , it is therefore more convenient to measure it direct
ly , particularly if the effect of stress ratio on pore pressure is also to be
studiedn .
(i )Advantages and Limitations of the Triaxial Test .

The advantages and limitations of the triaxial test have been discussed in
some detail elsewhere (Bishop and Henkel , 1957 ) , and will be referred to only
briefly here .
The principal advantages of the triaxial test as performed on cylindrical
specimens are that it combines control of the drainage conditions and the
possibility of the measurement of pore pressure with relative simplicity in
operation .
The principal limitations are that the intermediate principal stress cannot
be varied to simulate plane strain conditions , that the directions of the princi
pal stresses cannot be progressively changed , and that end restraint may
modify the various relationships between stress , strain , volume change , and
pore pressure .
For most practical purposes the advantages outweigh the limitations , and
it will be apparent from section 6 that a very satisfactory correlation does in
fact exist between laboratory tests and field observations of stability in many
important engineering problems .
m. Some difference will in general arise from such factors as the different
strains at which 'failure ' is taken to occur , the different rates of volume
change at failure , and , in soils having true cohesion in the Hvorslev
sense , the different water contents of the samples defining the failure
envelope .
The effect of stress ratio is discussed by Bishop (1952 and 1954 a ) and
Fra ser ( 1957 ) .
- --
-
=
=
s
-as --
--
20 40 SO
60 80 100 120 140 160
NORMALSTRESSon pisci
EFFECTIVE .
STRESSCIRCLESFOR DRAINEDTESTSWITHFULL SATURATIONSHOWNBY BROKENLINE

STRESSCIRCLESFOR CONSOLIDATEDUNDRAINED ,
TESTSWITH FULL SATURATION
, SHOWNBY SOLIDLINE
PLOTTEDIN TERMSOF EFFECTIVESTRESSAT MAX. DEVIATORSTRESS
ENVELOPE(1) REPRESENTS
DRAINEDTESTS
UNDRAINED TESTSIN TERMSOF OU (CIRCLESNOT SHOWN )
X- VALUES
UNDRAINEDTESTSIN TERMSOF EON(3) WITH ASSUMED
STRAINRATE 0-38%
. PERHOUR IN ALL TESTS
Fig . 11.- Undrained , consolidated - undrained and drained tests on boulder

clay compacted at an initial water content 2% dry of optimum :
clay fraction 4 %.
--- home - ,
-
Piton
---
--
- (b)
Annot
1
w En- En. 1
ni WV
ton
tono
e
(a
)
Fig stability analysis

of
12
Forces the slices method

in
–
.
.
.
TRIAXIAL TEST 457
5. METHODS OF STABILITY ANALYSIS
The stability of soil masses against failure under their own weight , or
under the action of applied loads , can be examined either by methods based
on elastic theory or by methods based on the principle of limit design .
In the first case the stress distribution is calculated and the maximum
stresses are then compared with the strength of the soil. As a practical
method it is , however , open to several serious objections. Firstly , it is diffi
cult to assess the error resulting from the assumption that the soil mass is
a homogeneous elastic material having elastic constants which are inde
pendent of the magnitude of the stresses . Secondly , it has been shown that ,
even if these assumptions were true , local overstress would occur in a typi
cal earth dam section when its factor of safety (by a slip circle method ) lay
below a value of about 1. 8 ( Bishop , 1952 ) . The same applies in principle to
earth slopes and foundations .
In consequence elastic methods are not applicable to the calculation of the
factor of safety when studying observed failures or for design work on em
bankments and cuts where 1. 5 is accepted as a working value for factor of
safety . Elastic methods are , however , useful in giving an estimate of the
stress distribution studies and for pore pressure prediction .
In most practical stability problems , therefore , the engineer is concerned
with the factor of safety against complete failure , rather than against local
over stress . The most general definition of factor of safety against complete
failure , which can be applied irrespective of the shape of the failure surface ,
is expressed in terms of the proportion of the measured shear strength that
must be mobilized to just maintain limiting equilibrium . The shear strength
parameters to which the factor of safety is applied in setting up the equations
expressing the condition of limiting equilibrium depend on whether the analy
sis is carried out in terms of effective stress (c ' , $ ' analysis ) or total stress
du = 0 analysis ) . The two cases will be treated separately .
( a ) Effective Stress Analysis . –

In the effective stress analysis the proportion of the shear strength mobi -
lized for limiting equilibrium is expressed :
T = (c' / F) + ( 0 - u) (tan ' / F) ( 13 )
The value of the factor of safety F is obtained by assuming limiting equi

librium along a trial slip surface (usually the arc of a circle in cross
section ) , balancing the forces and solving for F . The value of o is determined
from the equilibrium of the soil mass above the failure surface by an appro
priate graphical or numerical method . The method of determining the value
of u will depend on the class of stability problem .
(1) In (a ) problems , where the pore pressure is an independent yari
class
able , the value of u will be obtained from ground water level if there is no
flow , or from a flow net if a state of steady seepage exists . The flow net can
either be calculated or based on field measurements of pore pressure .
( II ) in class (b ) problems, where the magnitude of the pore pressure de
pends on the stress changes tending to lead to instability , the most practical
method of approach is that adopted in earth dam design . Here a prediction is
made of the actual pore pressure likely to obtain in the stable dam , which
should thus check with the field pore pressure measurements usually made
during construction . This prediction is based on an approximate stress
distribution within the dam , the undrained pore pressure parameter B and a
calculated allowance for pore pressure dissipation , the value of B being re
adjusted if necessary to match the calculated factor of safety .
Where field measurements of pore pressure are available they are of
course substituted directly in the analysis .
While any method of stability analysis can be used which correctly repre
sents the statics of the problem , the more complex soil profiles or dam
sections involving a number of zones of c ' and ' and irregular distributions
of pore pressure can be handled most readily by a numerical form of the
method of slices (Bishop , 1954 b ) .
As applied to the slip circle analysis (Fig . 12 ), the method leads to an ex
pression for the factor of safety :
sec a
: ] ' . (W (1 – 1w)+ (X . - Xo +1)] }1 +
tan
al
tan
.gʻ
14
F = Wsina {c b + tano ”
(
)
This expression
takes full account both horizontal and vertical forces
between The vertical shear force term
the slices of
which cannot be eliminat
,
, .
mathematically
can however be put equal accu
ed
zero with little loss
in
to
racy The method agrees within about with the modified friction circle
to
%
1
.
method described by Taylor 1948 two cases which have been checked
in
(
.
The programming the digital electronic computer DEUCE
of
for the nu
'
'
additional and over
by
merical Little and Price 1958 has given
an
method
it
(
whelming advantage since any specified pattern slip circles can be analysed
of
,
seconds per circle using about

at
of
30
50
rate about slices This leaves
5
a
.
-
the engineer free investigate the effect varying his assumptions about
of
to
soil properties and pore pressure and modify his design without the heavy
to
,
computation previously involved

of
burden
.
of
The extension the slices method noncircular surfaces has been under
to
by
at
Kenney present being

is
taken Janbu 1954 and 1957 and 1956 and

it
(
. )
programmed for the computer
Total Stress Analysis

b
–
.
(
)
the total stress analysis the proportion the shear strength mobilized
of
In
expressed for the condition as

Pu
is
=
,
,
0
TEC
15
/,
F
(
)
Fig the expression safety using the

12
of
of
the notation for the factor

In
,
.
slip circle analysis becomes

:
sina
W
16
c,
F
-
l!
(
)
When ºu the inter slice forces enter into the calculation only non
=
if
0
a
-
circular slip surface used

is
cy
For saturated soils the apparent cohesion equal

of
one half
un
is
the
to
compression strength obtained from un

12
drained and its value qu

is
)
.
(e
or
undisturbed samples
of
on
drained tests from vane tests

cu
The value
.
usually varies with depth and appropriate values must be used around the
trial failure surface
.
TRIAXIAL TEST 459
It should be noted that the use of this method is correct only where the
field conditions correspond to the laboratory tests conditions, i . e . where the
shear stress tending to cause failure is applied under undrained conditions .
It cannot in general be applied using undisturbed samples from slopes, for
example , where the water content has had time to adjust itself to the stress
changes set up by the formation of the slope .
The validity of the ºu = 0 method is in fact restricted to saturated soilsP
and to problems in which insufficient time has elapsed after the stress change
considered for an increase or decrease in water content to occur . It is there
fore an ' end of construction method ' . Whether the factor of safety subsequent
to construction will have a lower value depends on the sign and magnitude of
the stress changes . The particular cases are discussed in Section 6 .
The use of total stress methods in which , is not zero , or in which the
angle of consolidated undrained shearing resistance ®cu is used , is , in the
opinion of the authors , to be avoided except in special cases , owing to the
difficulty of determining the physical significance of the factor of safety thus
obtained .
( c ) Relationship between Total and Effective Stress Methods of Stability

Analysis . -
Since the failure criterion and the associated method of stability are only
convenient means of linking the stability problem with the appropriate labo
ratory test , a soil mass in limiting equilibrium should be found to have a
factor of safety of 1 by whichever method the analysis is performed . As total
stress methods can only be applied under undrained conditions, it is conveni
ent to demonstrate this point by a simplified analysis of a vertical cut in satu
rated clay immediately after construction ( Fig . 13 ) .
To simplify the mathematics of the problem it is assumed that the un
drained strength cu does not vary with depth , and that the effective stress
failure envelope is represented by c ' = 0 and a constant value of 0 ' . The
failure surface is assumed to approximate to a plane without tension cracks .
The critical height H under these conditions is known to be equal to 4cu / y
where y is the density of the soil . The factor of safety of the soil adjacent to
a vertical cut of depth H can be calculated in terms of either total or effective
stress :
( I) Total Stress . - From equation 16 :
col) / (3
/ sin
F =({ W
. a)
sin
) YH
17
cu
cosec cot
= =
/) H
a /(1
sin2
a
(
)
?
.
.
( (
2cu cos
(y
H
a
.
Putting dF da giving the lowest value

of
of
obtain the value we

to
=
0
F
. /
obtain 450
=
a
Substituting gives
17
equ
in
yH
4c
18
=
F
)
/
Substituting 4cu for we obtain

=
/y
1
F
the fact that the principal stress directions

by
The error introduced

in
0
.
most practical problems differ from those the laboratory discussed

in
is
by Hansen and Gibson 1949

)
.
(
Stiff fissured clays under exception

an
normal stress are

of
reduction
.p
.
WAWWV
--
-- -
3
--
- --
ova
VIIVIT US
77
.
VAIV
145
No
45°
la
-
777
FAILURE PLANE FAILURE PLANE

Øv
METHOD METHOD
=
C
',
0
$
'
Fig Simplified analysis stability vertical

13
of
of
the cut saturated

in
a
-
.
.
cohesive soil immediately after excavation using both total and

,
effective stress methods

.
Effective Stress From the Mohr diagram Fig

of 13
we can obtain
in
II
(
)
.
-
the pore pressure element of soil at failure the major

an
terms
in
in
principal stress the triangle OPQ that

of
follows from the geometry

It
.
:
sin
01
03
01
03
19
1601
=
2
({
2
}
$
u
–
-
-
(
'
)/
)/
(
)
Putting and rearranging

01
03
we
cu
obtain
=
2
(
)/
:
-
20
sin
sin
Co
=
+
,
)
0
1
u
“
-
.
[(
’)
(
')]
/
TRIAXIAL TEST 461
For a plane slip surface , and with c ' = 0 the expression for F given in equ .
15 simplifies to the form :
cos
tan
21
sin
F = [ 1/ (
W
W
ul
$
a
a
–
.
)]
'(
)]
(
)
stress corresponds
of
For failure plane the state
on this case
to
the
in
a
,
Rankine active state and thus the major principal stress 01 equal yh
is
to
,
,
the vertical head
of
soil above the element
by
Substituting given and putting
of
21
20
cu
equ the value equ
in
=
u
,
.
.
we obtain the expression for
y
H
4
,
F
/
:
tan
22
cot
sin
sin
sin
2a
=
[1
0
1
a
(
)
F
'(
]
[(
)]
)
/
)/
-
-
·
Putting dF da we now find that the minimum value Fis given by the
of
=
+ 0
= /
Substituting this value equation and express
22
inclination 45°
in
0
a
/
of 2.
'
angles terms we find that the expression again reduces
to
ing the
in
2
,
$
/
=
1
F
This illustrates two important conclusions Firstly both total

comparison
,
.
stress methods stability analysis will agree giving factor
of
and effective
in
a
safety soil mass brought into limiting equilibrium by change
of
for
of
in
1
of a
stress under undrained conditions Secondly although the values factor of
,
. of .
the rupture surface
on
safety are the same the position depend
to
found
is
in ,
the analysis The closer this value approximates

of
the
to
the value used
of Ø
true angle internal friction the more realistic the position the failure
of
is
,
confirmed by the analysis

of
the Lodalen slide terms

is
surface and this
in
,
of
and Sevaldson 1956 and Section

Ø
c
6
,
(
'
).
short term stability problems
is
of of of
The choice method saturated soil

in
in
thus matter practical convenience and the method generally

=
is
0
u
a
its simplicity unless field measurements pore pressure

of
used because
It ,
be used as control should be noted however that for factors

of
are
to
1 a
,
.
safety other than the two methods will not general give numerically equal
in
the effective stress method the pore pressure predicted

, is
of
values
In
F
.
for the stresses the soil under the actual loading conditions and the value
in
expresses the proportion then necessary for equilibrium

of
and tan
of
Ø
F
.
'
'
implicitly pore
of
uses value
on
The total stress method the other hand

a
pressure related the pore pressure at failure the undrained test The
to
in
analysis
Pu
high factor safety shown for example slope

of
of
of
the
in
=
0
over consolidated clay which the pore pressure shows marked drop dur
in
a
-
ing the latter stages of shear will therefore not be reflected the effective
in
stress analysis such marked way

in
cannot be too strongly emphasized that comparison between effective

It
stress and total stress methods can only be logically made when the shear
stress tending cause instability has been applied under undrained con
to
ditions The use of the method under other conditions cannot be justi
=
0
P
.
fied theoretically and practice often leads very unrealistic results see
to
in
Table
V
).
This point Terzaghi

by
discussed more fully Skempton

is
1936 1948
.q
;
,
,
b
and Bishop 1952

,
.
6 . THE APPLICATION OF STABILITY ANALYSIS TO PRACTICAL

PROBLEMS
In this section the stability analysis of a number of typical engineering

problems will be examined . The purpose of the examination is in the first
place to obtain a clear qualitative picture of what happens to the variables
controlling stability during and after the construction operation or load change
under consideration . The second purpose of the examination is to indicate
the most dangerous stage from the stability point of view and to select the
appropriate shear parameters and method of stability analysis .
It is not possible to generalise about the solution of practical problems
the
without considering principal properties will
of
the soil each case
in
It
.
that the permeability the soil has an
im
is of
have been apparent from section
2
portant bearing
on
the way which the stability problem

in treated the
In
.
more permeable soils sands and gravels the pore pressure will be
in
g
e
)
(
.
.
fluenced by the magnitude the stresses tending instability only

of
lead
to
to
transient loading construction and long term
of
of
under conditions Both end .
problems will fall into class which pore pressure independent
an
is
in
(a
)
variable Only the less permeable soils the relative merits alterna
of
do
in
.
tive methods analysis have be considered most practical cases

of
to
of in
.
Table are listed representative values the shear strength parame
In
ters some typical soils arranged order decreasing permeability of

of
The
in
.
permeability values will be noted and will be apparent that
of
wide range
it
,
here that the largest quantitative difference between the soil types lies
is
it
.
Permeability and Shear Strength Parameters Typical Soils
of
Table
1
.
.
-
Signifies Undisturbed Samples

(*
).
Plasticity Permeability
cm
Material
PI
Index sec.
.sq
K
Degrees
Ib
|
ft
./
Approx
%
.)
45
fill
Rock tunnel spoil

x 5
:
..
.
...
..
.
10
Alluvial gravel Thames Valley 43

5
-2
:
.
.
.
..
.
.
.
.
.
..
.
.
.
- 33
Medium sand: Brasted

. .
..
..
..
.
.
-5 10
20
35
Fine sand
3 1
-4
x x
.
.
.
..
..
.
.
.
.
10
Silt Braehead 32
:
.
.
.
.
..
.
.
.
.
low
Normally consolidated clay plasticity

of
10
Chew Stoke
.1
5
-6
x
*
Normally consolidated clay high plasticity

of
10
Shellhaven
1
-8
x
*
Over consolidated clay low plasticity

of
-
10
Selsetboulder clay 321/

1
-8
x
:
*
Over consolidated clay high plasticity

of
-
- 20
1010
London clay 5x 250

*
-4 -
10
20
Quick clay
1
x
*
The more important problems are

:
Bearing Capacity Clay Foundation

of
(a
-
.
)
This problem may be illustrated most simply

of
terms the construction

in
soft clay stratum

of
low embankment on saturated horizontal with

a
a
14a
Fig diagrammatically the variation with time

is
of
surface the
In
shown
.
.
TRIAXIAL TEST 463
factors which govern stability , i . e . the average shear stress along a potential
sliding surface and the average pore pressure ratio .
The excess pore pressure set up in an element of clay beneath the em
bankment is given by the expression :
40g
403
40
Au = B [
,
A
(
)]
-
(4
)
general positive and have its
be
For points beneath the embankment will
Au
in
positive for
at
greatest value
of
the end construction since and
is
=
,
A
1
normally lightly overconsolidated clay slow or the
or
Unless construction
is
.
clay contains permeable layers little dissipation pore pressure will occur
of
,
during the construction period After construction completed the average
is
value of .
will decrease as redistribution and dissipation the excess pore
of
ru
pressures occur until finally the pore pressures correspond ground water
to
,
level
.
by
The factor safety given the effective stress analysis will thus show
of
a
will rise
or
at
minimum value near the end construction after which

of
it to
it
,
the long term equilibrium valuer For the long term stability calculation
.
is obviously appropriate take the values of and from drained tests

to
Q
c
'
'
.
construction case the same values may also be used for
of
For the end
,
though more logical take the value from undrained and consolidated
is
to
it
undrained tests expressed terms effective stress the error

of
on the
is
in
likely ,
is
conservative side and be small

to
.
of
The use
of
the effective stress method for the end construction case
means however that the pore pressures must be predicted or measured
in
,
Typical field measurements pore pressure under an oil storage

of
the field
.
tank are illustrated Fig 14b after Gibson and Marsland 1960 However
in
,
(
)
.
.
field measurements are usually limited the more important structures and
to
earth dams and the application

of
of
this method the end construction

to
to
,
case will depend estimated pore pressure

on
other instances have

to
in
values As this estimate involves an assumption about the stress distribution

.
by
how nearly limiting equilibrium approached

is
which influenced and

is
(
usually avoided by going directly

of
the determination the value

is
it
of
to
A
,
the Pu analysis which applicable

of
the end construction case with

to
=
is
0
zero drainage
.
The undrained shear strength be used analysis

is
to
the ♡u obtained
in
=
0
unconfined compression
on
from undrained triaxial tests

or
tests undisturbed
in (
samples or from vane tests the field the majority of problems involv
In
,
ing foundations soft clay where quite clear that the long term factor
on
is
it
,
higher than the value at the end

of
safety
of
construction there then no

is
is
need for the more elaborate testing and analysis required by the effective
appreciable dissipation pore pressure likely
of
of is
stress method However

if
,
.
occur during construction uneconomical not take advantage

is
to
to
in
it
it
,
calculating the factor of safety and the effective stress method then re
is
quired
.
at
Newport
by
The failure bauxite dump reported Skempton and

of
a
of (
may be taken as example analysis

of
an
Golder 1948 the use the ♡u

=
,
),
conditions Fig After relatively rapid tipping

15
of
for end construction

,
.
).
(
The position the most critical slip surface will change as the
of
course
of
r
.
pressure pattern alters

.
- - - - - - - - t - -
2 G.W.L.
HEIGHTOF FILL
AVERAGESHEARSTRESST ON A GIVEN SURFACETHROUGHP
- TIME
TIME
PORE PRESSUREAT P
DUE TO G.W
.L.
TIME
AVERAGEPORE PRESSURE
RATIO AROUNDSLIP SURFACE
AVERAGE
TIME
- O METHODAPPLICABLEHERE
FACTOROF SAFETYAGAINSTFOUNDATIONFAILURE (c', O' METHOD
)
FACTOR
OF
SAFETY
TIME
RAPID PORE PRESSURE

DISSIPATION POREPRESSURE
EQUILIBRIUM
CONSTRUCTION
Fig . 14a . – Variation with time of the shear stress , local and average pore
pressure , and factor of safety for the saturated clay foundation
beneath a fill .
TRIAXIAL TEST 465
- SOFT CLAYS -- - 5 -
1111111
1111111
1111101
Tiilillil
SETAMINATED CLAYEY SANDY _ SILTS
HEAD OF
IN
WATER
10
TANK FEET
(
--
M
D
N
A
F
J
J
1953 1954
5
10
IN
CHANGE
A2
PORE WATER
PRESSURE
5
OF
FEET WATER)
(
-5
14b
Fig Pore pressure changes soft clay foundation filling and

on
in
a
–
.
emptying storage tank after Gibson and Marsland 1960

,
a
).
(
analy
by
height
at
safety
25
failure occurred
of
of
feet the factor the

=
;
0
u
a
sis was subsequently found which can be accepted as agreement

08
be
to
to
1
,
.
experimental accuracy
of
within the limit

.
this case the fill was granular material and its contribution the
to
In
shearing resistance was small cases where the fill cohesive material
is
In
a
.
high undrained strength fullvalue of this strength

of
the use the

of
to in
the
,
analysis gives misleading results The explanation appears be that

=
0
.
up
shear deformations set the soft clay foundation under undrained loading
in
LAUXITEDUMP
• 36°
25FT
. LB/FTP
T . 102
- A
SOFT
- -
- - - - - S e AFTERFALRE
SURFACE
MADE CAOUND c . 0
sc · 1330 LOFT
MOTTLEDCLAY
OS LMTT
365 LOFT
SOFT BLUE CLAY
PL . 26
EMARL
FEET ..... 1oSCALE20 30 40 FEET
Fig . 15. – Failure of a bauxite dump at Newport (after Skempton and

Golder ,1948 ).
set up tensile stresses in horizontal direction in the more rigid fill above and
result in vertical cracks . A description of cracks wider at the bottom than
at the top and passing right through the fill is given by Toms ( 1953a ).
As the factor of safety rises with time , no long term failures can be quoted
in this category .
The calculation of the ultimate bearing capacity of a structural foundation
on a saturated clay is in principle the same as the problem treated above .
However , the bearing capacity is not calculated by assuming a circular slid
ing surface , but is computed from the theory of plasticity for both the total
and effective stress analyses , the results being expressed directly as bearing
capacity factors .
For large foundations on soft clay the ultimate bearing capacity will in
crease with time after loading . For small , shallow foundations on stiff clay
the ultimate bearing capacity will decrease with time, but in most cases
settlement considerations will govern the design .
An example of the u = 0 analysis of an end of construction foundation
failure has been given by Skempton ( 1942 ) . Here a footing 8 feet by 9 feet
founded on a soft clay with Cu = 350 lb . / sq . ft . failed at a nett foundation
pressure of 2500 lb . / sq . ft . ( Fig . 16) . Using a bearing capacity factor of 6 . 7
for this depth to breadth ratio , a factor of safety of 0. 95 is obtained .
Two series of loading tests on the stiff fissured London clay may also be
mentioned , where
the ºu = 0 analysis has led to factors of safety about 1. 02
( Skempton These tests are particularly interesting as showing that the
1959 ).
effect of fissures , which can lead to serious difficulties with end of con
struction problems in open excavations ( see section 6b ) does not prevent the
successful use of the Pu = 0 analysis in bearing capacity calculations under
end of construction conditions.
TRIAXIAL TEST 467
100 TONS
STANCHION
SHEAR STRENGTH
o 400 800 1200 LB / FT2
TOP SOIU
FIRM - BROWN - CLAY
MOTTLED CLAY = 8 -9 %
SOFT BLUE CLAY
PEATY CLAY
FIRM SANDY CLAY
W | LL | PL
40 70 29
40- 60 70 28
60 70 27
NETT FOUNDATION PRESSURE AT FAILURE = 2 500 LB /FT ?

IF 9 = 6 .70 FACTOR OF SAFETY = 0 . 05
Fig . 16 . - Failure of a foundation on soft clay at Kippen (after Skempton ,

1942 ) .
An example of the long term failure of a small heavily loaded foundation of

stiff clay is more difficult The long term failure of tunnel arch foot
to find .
ings described by Campion ( 1951 ) probably falls in this category .
It may however be concluded that with few exceptions the end of con
struction conditions is the most critical for the stability of foundations and
that for saturated clays this may be examined more simply by the Qu = 0
analysis . From the field tests and full scale failures tabulated in Table II it
is apparent that an accuracy of + 15% can be expected in the estimate of factor
of safety . One of the exceptions is dealt with in section 6 ( i ) . Where partial
dissipation of pore pressure occurs during construction an analysis in terms
of effective stress is used , and examples of the analysis are discussed in
section 6 (h ) .
Table II. - End of Construction Failures of Footings and Fills on a
Saturated Clay Foundation : Pu = 0 Analysis .
1. Footings, loading tests.
Data of clay
Safety factor
Locality W - PL Reference
=0
W LL PL
| PI analysis
Loading test, Marmorerá - 0.25 0.92 Haefeli , Bjerrum

Kensal Green 1.02 Skempton 1959
Silo , Transcona 0.95 1.09 Peck, Bryant 1958
Kippen 0.52 0.95 Skempton 1942
Screw pile, Lock Ryan 1.05 Morgan 1944, Skempton 1950
Screw pile, Newport 1.07 Wilson 1950
Oil tank, Fredrikstad 0.67 1.08 Bjerrum, Øverland 1957
Oil tank A , Shellhaven 0.73 1.03 Nixon 1949
Oil tank B. Shellhaven 1.05 Nixon (Skempton 1951 )
Silo , U.S. A. 0.98 Tschebotarioff 1951
Loading test, Moss 1.10 NGI
Loading test, Hagalund 1.37 0.93 Odenstad 1949
Loading test, Torp 1.99 0.96 Bjerrum 1954c
Loading test, Rygge 1.44 0.95 um 1954c
2. Fillings
Chingford 145 0.50 1.05 Skempton, Golder 1948

Gosport 80 0.48 0.93 Skempton 1948d
Panama 2 80 111 0.53 0. 93 Berger 1951
Panama 3 110 125 0.70 0.98 Berger 1951
Newport 50 0.71 1.08 Skempton, Golder 1948
Bromma II 100 1.00 1.03 Cadling, Odenstad 1950
Bocksjön 100 1.17 1.10 Cadling , Odenstad 1950
Huntington 400 0.98 Berger 1951
Failures Excavations ºu Analysis

III
Table Construction
of
in
End
=
0
.
:
.
-
Clay
of
of
Data Factor
safety
Location
0 :
Soil type Reference

W
PI PL
-
PL
,
0
PI analysis
0610 90
Huntspill Skempton Golder 1948

. .1 .4 0
05 6
1 1 . 0
,
.
Congress Street Ireland 1954

0
.7 . . 1
Skattsmanso Intact clay

I
09 03 Cadling Odenstad 1950

,
Skattmanso
,
. 1
II
Bradwell Stiff fissured 02 Imperial College 1959

0
1
-
clay
La
kempton Rochelle
,
)
(S
TRIAXIAL TEST 469
(b ) The Stability of Cuts and Free - Standing Excavations in Clay

The changes in pore pressure and factor of safety during and after the
.
excavation of a cut in clay are illustrated in Fig . 17.
The change in pore pressure can conveniently be expressed by putting
B = 1 and re - arranging equ . (4) in the form :
Au = [(20 + 40g)/ 2) + (A - 1/2) (40, - 403) (23 )
The reduction in mean principal stress will thus lead to a decrease in pore
pressure , and the shear stress term will also lead to a decrease in pore
pressure unless A is greater than 1/ 2 , if the unknown effect on pore pressure
of changing the directions of the principal stresses is neglected . An estimate
of the stress distribution can be made from elastic theory if the initial factor
of safety of the slope is high , or from the state of limiting equilibrium round
a potential slip surface if the factor of safety is close to 1.
- - - - - -
INITIAL P.W.P.
-- 7 - - - - - - - ...
NAL C.W.L.
LORICINALC WL
FINAL C.W.L.
FINAL P.W.P.
END OF EXCAVATIONPORE PRESSUREA=I
END OF EXCAVATIONPORE PRESSURE AO

EQUIPOTENTIAL
LINE
- - - - ORIGINAL C.W.L.
FINALGWL
PORE
PRESSURE
AT P
TIME
THOD APPLICABLEHERE
FACTOR - FACTOR OF SAFETY (C, Ó' METHOD

)
OF
SAFETY
TIME
RAPID PORE PRESSURE

REDISTRIBUTION PORE PRESSURE
EQUILIBRIUM
EXCAVATION
Fig . 17 . - The changes in pore pressure and factor of safety during and
after the excavation of a cut in clay .
In Fig . 17 the changes in pore pressure at a representative point are shown
for the values A = 1 and A = 0. The final equilibrium values of pore pressure
are taken from the flow pattern corresponding to steady seepages .
Using values of c ' and 0 ' from drained tests or consolidated - undrained
tests expressed in terms of effective stress the factor of safety can be calcu
lated at all stages from equation 14 . In the majority of cases , unless special
drainage measures are taken to lower the final ground water level , the factor
of safety reaches its minimum value under the long term equilibrium pore
pressure conditions .
An example of the investigation of a long term failure of a cut in terms of
effective stress has been given by Sevaldson ( 1956 ) . The slide took place in
1954 in a clay slope at Lodalen near Oslo , originally excavated about 30 years
earlier ( Fig . 18 ). Since the slide occurred without any apparent change in ex
ternal loading , it can be considered to be the result of a gradual reduction in
the stability of the slope . Extensive field investigations and laboratory
studies were carried out to determine the pore pressure in the slope at the
time of failure and the shear parameters of the clay .
Triaxial tests gave the values c' = 250 lb . / sq . ft. and $ ' = 32º. An effective
stress analysis using equation 14 gave a factor of safety of 1. 05 , and con
firms the validity of the approach to within acceptable limits of accuracy .
Where the final pore pressures are obtained from a flow net not based on
field measurements , allowance should be made for the fact that the perme
ability of a water laid sediment is generally greater in a horizontal direction
( Sevaldson 1956 ) . The highest wet sea son values obviously represent the
most critical conditions .
The excavation of cuts in stiff fissured and weathered clays presents some
special problems which have been discussed in detail by Terzaghi , 1936 a ;
Skempton , 1938 c ; Henkel and Skempton , 1955 ; Henkel , 1957 ; etc . The re
duction in stress enables the fissures to open up and they will then represent
weak zones which a sliding surface will tend to follow . The fissures will also
increase the bulk permeability of the clay (an increase of about 100 times is
reported by Skempton and Henkel , 1960 ) so that the pore pressure rise lead
ing to the long term equilibrium state will occur more rapidly .
The presence of fissures is reflected in the factors of safety obtained us
ing the effective stress analysis with field values of pore pressure and values
of c ' and ' mea sured in the laboratory on 1- 1/ 2" diameter samples . An
analysis of three long - term cutting failures in London clay by Henkel ( 1957 ) ,
using the laboratory values of c ' = 250 lb . / sq . ft . and Øl = 20° , gave factors of
safety of 1. 32 , 1. 35 , and 1. 18 . Putting c ' = 0 gave values of 0. 78 , 0 .81 , and
0 .82 respectively and obviously underestimated the factor of safety . If the
value of c ' required to give a factor of safety of 1 is plotted against time after
construction (Henkel , 1957 ; De Lory , 1957 ) , it is found that the value of c '
shows a definite correlation with time ( Fig . 19) and , as will be seen in section
6 (c ) , appears to approach zero in natural slopes on a geological time scale .
An effective stress analysis based on a value of c ' related empirically
with time for each clay type will obviously give a close approximation to the
correct factor of safety . For large scale work and for remedial measures
on active slips where the large strains tend to reduce c ' , it is prudent to en
sure a factor of safety of at least 1 with c ' = 0 .
s . This method has also been given since 1956 by Skempton in his lectures
at Imperial College .
TRIAXIAL TEST 471
F=1-01
ACTUALSLIP CIRCLE Fr 1.05
SAFETYFACTOR1-07
30 -
4 61216M
CRITICALSLIP CIRCLE
SAFETYFACTOR100 1
1
MEASURED
1 PORE
PRESSURE
10- Ft
CURVES
OF EQUALPOREPRESSURE
0 2 4 6 10
METRES
SECTION SAFETYFACTORS
No. (BISHOP1054
CO-ANALYSIS )
1.10
100
1-10
WEIGHTED
AVERAGE SAFETYFACTORFORTHE
WHOLESLIDE F . 1-05
Fig . 18 . – Long term failure in a cut at Lodalen (after Sevaldson , 1956 ) .
300
ODENOTES
ANALYSIS
BYCIRCLES
10 O DENOTES
ANALYSIS
BY PLANES
DENOTESTHEREIS REASONTO
BELIEVE
POINTSHOULD BE
HILL
WEMBLEY MOVEDIN DIRECTIONSHOWN
STABILITY SOFT
MILLLANE
- --
LB
150 EAST
PAPRVILLAGE
UMBRIDGE WOOD SEN
ooo G
7
tis
2100 NORTHOLT
-
-
-
-
UPPER
HOLLOWAY
so
SUDBURY
MI
S
.
n
!
!
10
20
30
40
50
00
70
80
Too
0
-
-
AT
AGE FAILURE
YEARS
Fig Long term stiff fissured London clay

19
failures Correlation
of
in
-
-
.
.
apparent cohesion required effective stress analysis with age cut

of
in
c
'
failure after De Lory 1957

at
)
.
(
The mechanism of the drop in c ' before a failure is initiated is not clearly
understood , but may be associated with stress concentrations due to the
presence of fissures , the progressive spread of an over stressed zone in a
soil which tends to dilate and absorb water on shear , and the effect of cyclical
fluctuations in effective stress due to seasonal water level changes . On the
limited evidence so far available from Lodalen and Selset ( see section 6 ( c ) ) it
does not appear to occur in any marked way in non - fissured clays .
In temporary work , where the end of construction condition is of primary
interest , the factor of safety may obviously be calculated by using the u = 0
analysis and the undrained shear strength . This method may also be used
with advantage where it is necessary to check that the initial factor of safety
is not lower than the long term value , as it avoids the necessity of explicitly
determining the stress distribution and pore pressure values at the end of
III
of
its
construction . Four examples use are given Table
in
.
Also included Table III an example the use
of
of
the Qu method for
is
in
=
to 0
construction conditions London clay which led
of
an
end cut in over
in
by a
,
estimate safety consequence
Whether this simply
of
of
the factor 70
is
a
.
pore
or
the opening fissures due stress release changes
of
in of
due
to
to
in
,
pressure even the short period the high bulk permea
of
to
excavation due
bility apparent
at
not yet clear The reduction strength low stresses
is
is
in
,
stiff fissured clay even undrained tests the triaxial apparatus Fig
in
20 in
in
.
(
but hardly appears adequate account for the
70
error conservative
to
%
),
A
.
safety must obviously be used ses and the rapid ad
of
factor similar
ca
in
,
justment the equilibrium pore pressure condition must be allowed for
to
in
prolonged construction operations
.
Natural Slopes
.
c
(
)
Natural slopes represent the ultimate long term equilibrium pro
of
state
a
file formed by geological processes The pore pressures are controlled by
.
the prevailing ground water conditions which correspond steady seepage

to
,
subject minor seasonal variations ground water level Natural slopes
to
in
therefore fall into class which the pore pressure

an
independent vari
is
in
(a
)
able
.
principle the analysis the long term equilibrium

as
the same
is
of
that
In
cut or excavation However the pore pressures will have already reached
of
a
their equilibrium pattern which can be ascertained from piezometer measure

ments the field and the natural processes will
of
softening leaching etc

in
.,
have already reached advanced stage An analysis based laboratory

an
on
.
this material would therefore be expected

of
tests lead close agreement

to
to
with observed slopes limiting equilibrium

in
Relatively few natural slopes limiting equilibrium have yet been ana
, in
lysed effective stress but some representative examples are col

IV of
terms
in in
lected Table The two cases involving intact clays Drammen and
,
.
Selset are being examined greater detail but the preliminary values
of
in
,
factor of safety 03 respectively show that the method can be

of
15
and
1
1
.
used with reasonable confidence Had tended zero the Selset slope would
to
c
. of .
'
of
have shown factor safety less than for example which outside
is
,
a
,
0
.7
the limit experimental error

of
stiff fissured clays special account has the pro

be
of
However
to
taken
in
,
in -
gressive reduction which appears eventually approach

of
the value
to
',
c
zero the failure zone since the shear strains and water content change
in
associated with failure are very localized and tests

do
of
on
the bulk the soil

not reveal the decrease
in
c
'.
TRIAXIAL TEST 473
30
SHEAR STRESS FISSURED LONDON CLAY : AVERAGEOF51TESTS
IN LB PER SQ.IN.
APPROXIMATE
AVERAGEOVER
BURDENPRESSURE
IN SITU
20
C. =16
- 3LB
PER SQ.IN.
10 20 30 40 50 60 70 80 90 100
TOTALNORMALSTRESS LB PER SQ.IN
Fig . 20 . - The strength of stiff - fissured London clay in undrained tests

(after Bishop and Henkel , 1957 ) .
. 300
: WATERLEVEL
ELEVATIONABOVESEA LEVEL FT
WATERLEVEL
AT SURFACE
B. H. ROADBEFORE SLIP
B.H. RAILWAY ROAD AFTER SLIP
200 B.H. RIVER BANK BEFORE SLIP
RIVERBANKAFTERSLIP
COALMEASURE
.M
B
.
CLAY SHALES
APPROX SLIP SURFACE
.
AVERAGEDEPTH TO
17
SLIP SURFACE FT
:
.
100
200
300
500
600
700
800
do
-
CHAINAGE FT
:
.
..
FT
/
SQ
: 21°
C
'
LB
SHEARSTRESS
DRAINEDSTRENGTH
500LB Sa FT.
/
.
UNDRAINEDSTRENGTH
) SQ
LB
FT
600
( .1
•
/
.
.
04
A
SHEARSTRENGTH
-
.
FORFAILURE
150 •400LBISQFT.
/ c
.'
LB SQ. FT.
L
000/
D
000
1
000 000 000

3
.
5
4
.
.
.
.
SQ
LB
EFFECTIVENORMALSTRESS FT
:
/
.
.
VE
POREPRESSURE(
)
-
IN
UNDRAINED
TEST
Fig Natural slope failure stiff fissured clay

21
Jackfield after
at
in
-
-
.
.
Henkel and Skempton 1955 Cross section and shear strength data
,
.
):
.sq
Shear strength for limiting equilibrium 400

.ft lb
ft
- =
./ ,
0 1
lb .4 .0
F
.
. ) ).
./
= A (
.sq
Undrained shear strength 150 21º

lb
=
,
0
c
(
'
'
01 sq
600 Drained shear strength 150

.sqlb
ft
=
1
4
.0
F
(c
.
.
.
= lb .
(
'
. ./
/
sq
.sq45
21° 580 Drained shear

ft
ft
' ),
=
lb =
1
(F
./
), .
.
.
).
strength
07
21° 430
@
ft
=
=
0
1
F
c
.
).
(
'
(
./
Table IV . - Failure of Natural Slopes : Effective Stress Analysis .
Data of Clay Factor of

safety:
Location Soil type W-PL Reference
c, ở
w | LL
PI analysis
Drammen Normally 1.09 1. 15 Bjerrum and Kjærnsli

Consolidated 1957
(Intact)
Selset Overconsolidated 18 1. 03 Imperial College
(Intact)
Jackfield Overconsolidated 20 1.45 Henkel and Skempton
(stiff -fissured) (1.07 with 1955
c' = 0)
Table V . - Long Term Failures in Cuts and Natural Slopes : Øu = 0 Analysis

(after Bjerrum and Kjaernsli , 1957 ).
1. Overconsolidated, fissured clays
Data of clay Safety

Type
factor,
Locality of W-PL Reference
R =0
slope | PI
PI analysis
Toddington Cutting - 0.34 Cassel, 1948

Hook Norton Cutting - 0.96 8 Cassel, 1948
Folkestone Nat . slope - 0.22 14 Toms, 1953b
Hullavington Cutting - 0. 18 21 Cassel, 1948
Salem, Virginia Cutting - 0.10 3. 2 Larew , 1952
Walthamstow Cutting 3. 8 Skempton , 1942
Sevenoaks Cutting Toms, 1948
Jackfield Nat. slope 0.00 Henkel /Skempton,
1955
Park Village Cutting 0.00 Skempton , 1948
Kensal Green Cutting 0.00 3.8 Skempton, 1948c
Mill Lane Cutting 3. 1 Skempton, 1948 c
Bearpaw, Canada Nat . slope 0.09 6.3 Peterson, 1952
English Indiana | Cutting 0.13 5.0 Larew , 1952
SH 62, Indiana | Cutting 0. 19 1. 9 Larew, 1952
2. Overconsolidated, intact clays
Tynemouth Nat . slope 1.6 Imperial College

Frankton, N. Z. Cutting 0.20 1.0 Murphy , 1951
Lodalen Cutting 0.72 1.01 N.G. I.
3. Normally consolidated clays
Munkedal Nat . slope

0.85 0. 85 Cadling/Odenstad,
1950
Save Nat . slope 0.80 Cadling /Odenstad,
1950
Eau Brink cut Cutting 1.02 1.02 Skempton, 1945
Drammen Nat . slope 1.09 0.60 N.G. I.
TRIAXIAL TEST 475
The landslide at Jacksfield provided a good example of the application of

the effective stress analysis to a natural slope in stiff - fissured clay (Henkel
and Skempton , 1955 ) and is illustrated in Fig . 21 . The slope of the hillside is
10 . 5°, and when the slip took place in the winter 1951 - 52 , a soil mass 600 feet
by 700 feet and 17 feet in thickness moved gradually downward about 100 feet .
The calculated average shear stress in the clay was about 400 lb . / sq . ft .
Drained tests on undisturbed samples gave c ' = 150 lb . / sq . ft . and Ø' = 219 ,
which with the observed pore pressures gave a shear strength of 580 lb . / sq .
ft . and a factor of safety of 1 .45 . Putting c ' = 0 gave a shear strength of 430
1b. / sq . ft . and a factor of safety of 1 .07 .
For natural slopes in stiff fissured clays it therefore appears necessary to
use c ' = 0 in the effective stress analysis . This is confirmed by observations
made by Skempton and De Lory (1957 ) on the maximum stable natural slope
found in London clay , and by Suklje ( 1953 a and b ), and Nonveiller and Suklje
( 1955 ) in other fissured materials . It is interesting to speculate on whether
the drop in c ' is due to the fissures , or whether both are due to some more
fundamental difference in the stress - strain - time relationships between the
fissured and intact clays .
A second class of soil which gives rise to special problems includes very
sensitive or quick clays . These clays show almost no strength in the re
moulded state , and they will therefore tend to flow as a liquid if a slide oc
curs . A small initial slip in a slope may therefore have catastrophic conse
quences as the liquified clay will flow away and will not form a support for the
exposed clay face , with the result that the whole of an otherwise stable slope
may fail in a series of retrogressive slips taking place under undrained con
ditions .
A factor which affects the quantitative analysis in the case of quick clays
is the influence of sample disturbance on the values of c ' and ' measured in
the laboratory . Soft clays of low plasticity are very sensitive to disturbance
and reconsolidation in the triaxial test is always accompanied by a reduction
in water content . Particularly where the initial water content is above the
liquid limit laboratory tests appear to overestimate the value of 0 ' . The in
vestigation of a recent slide in quick clay in Norway has given a value of 0 .
calculated from the statics of the sliding mass which is less than 50 % of the
value measured in the triaxial test .
The occurrence of quick clays is limited to certain well defined geological
conditions , and where they are encountered special precautions in sampling ,
testing and analysis are always necessary (Holmsen , 1953 ; Rosenqvist , 1953 ;
Bjerrum , 1954a and 1955c ) .
The application of laboratory tests to the stability of natural slopes raises
two general matters of principle . The time scales of the load application are
so different in the laboratory and in the field that it is perhaps surprising
that satisfactory agreement between the results can be obtained at all . Labo
ratory results quoted by Bishop and Henkel (1957 ), and Bjerrum , Simons and
Torblaa ( 1958 ) indicate that under certain conditions may have a lower
value at low rates of loading . This effect may be partly offset by the fact that
the worst ground water conditions which touch off the slip are only of season
al occurrence ; and by factors such as the effect of plane strain on the value
of Ø and the omission
of ' end effects in the stability analysis . It is also well
known that considerable creep movements occur in slopes still classed at
stable . However , with the exceptions noted , the overall correlation between
laboratory and field results is quite acceptable from a practical point of view .
Secondly , it may well be asked why the factor of safety cannot be calculated
with equal accuracy using the Pu = 0 analysis and the undrained strength of
samples from the slope where pore pressure and water content equilibrium
have been attained . A large number of case records of slides in both natural
slopes and cuttings are summarized in Table V and it is evident that as a
practical method it is most unreliable , giving values of factor of safety rang
ing between 0 . 6 in sensitive clays to 20 in heavily overconsolidated clays .
The fundamental reason for the difference is that in the undrained test the
pore pressure is a function of the stress applied during the test , and it is not
necessarily equal to the pore pressure in situ . To obtain a factor of safety
of 1 for a slope in limiting equilibrium using undrained tests would require
that the same pore pressure should be set up in the sample when the in situ
normal and shear stresses were replaced . This is in general prevented by
the irreversibility of the stress - strain characteristics of the soil and by the
changes in the principal stress directions . The latter occur even with in situ
tests .
The position is made worse by the fact that the water content changes both
in overconsolidated and in sensitive clays are very localized at failure and
samples which do not pick up these layers can have little bearing on the sta
bility analysis . A sample from the 2 inch thick slip zone at Jackfield in which
large strains had occurred was found to have a water content 10 % above the
adjacent clay and an undrained strength within 12 % of that required for a
factor of safety of 1 ( Henkel and Skempton , 1955 ) . This layer was difficult to
find and sample , and as clay outside the failure zone had an undrained strength
nearly four times as great the method has little predictive value .
( d ) Base Failure of Strutted Excavations in Clay .

During excavation in soft clay, base failure sometimes occurs accompanied
by settlement of the adjacent ground . Failures of this typet have occurred in
excavations for basements , in trenches for water and sewage pipes , and in the
shafts for deep foundations .
The construction of temporary excavations is generally carried out suf
ficiently rapidly for pore pressure changes to be ignored . The change in
stress thus occurs under undrained conditions and the stability can be calcu
lated using the Pu = 0 analysis and undrained tests .
The factor of safety F against base failure can be derived from the fa
miliar bearing capacity theory , considering the excavation as a negative load .
This leads to the expression (Bjerrum and Eide , 1956 ):
F = N. . cu/ yD + q) (24 )
where D denotes depth of excavation

y >> density of the clay
density of the
U the undrained strength of the clay beneath the bottom of the
excavation
q » the surface surcharge ( if any )
dimensionless bearing capacity factor depending on shape
and depth of excavation .
t. Bottom heave failures can also occur in clay if a pervious layer containing
water under sufficient head lies close beneath the excavation . For ex
ample see Garde -Hansen and Thernöe 1960 , and Coates and Slade 1958 .
TRIAXIAL TEST 477
The analysis of the failure of seven excavations is given in Table VI. The
results indicate that in practice an accuracy of within + 20 % can be expected .
It should be noted that this type of failure is not caused by inadequate
strutting , but the loads and distortion after its occurrence may initiate a
more general collapse .
Table VI. - Base Failure of Strutted Excavations in Saturated Clay : Øu = 0

Analysis (after Bjerrum and Eide , 1956 ) .
: :S F
: :D ./ : ./ ./
Average safety
p ym
theoretical
m
Shear strength
/B /
Safety factor
Sensitivity
m m
Dimensions
factor
L
Surcharge
B
Density
sq cu sq
Site
Depth D
tons
tons
tons
BXL
Pumping station
1
,
.
54 90 75
75
50
60
82 16 03
0
0X
0
0
.0 .10
.
Fornebu Oslo
5
5
.0
. . 1
.
. 1. 1.
.
.
,
50
60
100
0
.
0 1
1
5
1
2
20 5
Storehouse Drammen
X
– –
4
.8
.
,
3 2
. .
28
50
Pier shaft Göteborg
.
0
.90
1
.5
3
.10
0
.9
.
,
Sewage tank
,
4
64
69
.
20
93
6
.7
11
96
0
0
.10
Drammen
0
.
.
X
5
.5
8
.0
0
.
Test shaft
N
5
84
.
, , , (
)
85
140
0
9
®
7
4
1
.0
Ensjøveien Oslo
.1
.1
2
.7
0
.
1
.5
.
|
Excavation Grey
6
08
72
78
10
.
80
.1
.7
0
.0
5
0
.1
4
.pl
.0
1
1
X
5
.8
8
.1
Vedels Oslo
-
.
|
Kronibus shaft
7
61
"
"
40
.
80
19
.8
5
3
.7
0
.5
3
Tyholt Trondheim
.1
0
.0
7
.
X
4
.7
.4
2
|
,
Earth Retaining Structures

on
Pressures
.
Earth
e
(
)
an earth retaining structure

the displacement sufficient for the full
of
is
If
development of plastic zone the soil adjacent the earth pressure

to
it
in
a
will be function of the shear strength of the soil This condition apparent
is
a
much temporary work and many permanent structures

ly
satisfied The
in
in
pressure
of
of
distribution the deformation the structure and

is
function
of
a
the soil and can only be predicted after detailed consideration

of
the move
,
ments involved
.
For temporary excavations intact saturated clay generally suf

is
in
it
ficient calculate the total earth pressure using the bu analysis and the
to
=
0
undrained shear strength Justification for this procedure be found

is
to
in
.
by
the field measurements published Peck 1942 Skempton and Ward 1952
),
),
(
and Kjaernsli 1958 More recent measurements soft clay carried out by
in
.
(
the Norwegian Geotechnical Institute however indicate that the total earth
pressure may exceed the value determined analysis and that the
by
the
=
0
u
of
of
ratio the actual the calculated load increases with the number struts
to
used carry
to
it
.
The long term earth pressure logically computed using the effective
is
stress analysis with values taken from drained tests or

of
and
c
'
'
consolidated undrained triaxial tests with pore pressure measurements

to
-
gether with the least favourable position This will

of
the water table most

in
.
cases represent rise earth pressure Few examples are available

to
in
a
.
confirm this analysis other than of gravity retaining walls which are them
selves founded in the same clay stratum . The problem is then in effect one
of overall stability since the slip surface passes beneath the wall which is
then of little more consequence than one ' slice ' in the slices method of analy
sis ( Fig . 22 ) .
A number of failures of this type in stiff fissured London clay have been
analysed by Henkel (1957 ) , and here consistent active and passive pressures
on the walls have been obtained using the effective stress analysis and ob
served water levels , together with the reduced values of c ' shown in Fig . 19 .
It should be noted that where the excavation in front of the wall is deep , the
presence of the passive pressure is insufficient to prevent the occurrence of
progressive softening .
The behaviour of gravity retaining walls can of course throw little light on
the end of construction earth pressures in fissured clays . Measurements of
strut load have however been made by the Norwegian Geotechnical Institute in
a trench in the weathered stiff fissured crust overlying a soft clay stratum
( Di Biagio and Bjerrum , 1957 ; Bjerrum and Kirkedam , 1958 ) . Here the soft
ening appeared to proceed more rapidly than in cuts in London clay , for after
only a few months the strut loads corresponded to the value given by the ef
fective stress analysis with c ' = 0 .
The evidence from this cut and the indirect evidence from the Bradwell
slip ( section 6 (b )) indicates that the u = 0 analysis does not correctly repre
sent the behaviour of stiff fissured clays under decreasing stresses even
shortly after excavation . The rapid dissipation of negative pore pressures
due to the presence of open fissures is obviously an important factor in tem
perate climates . Under long term conditions the u = 0 method is also in
applicable for the reasons given in sections 6 (b ) and (c) .
(f) The Stability

It is impossible
Dams
of Earth .
deal adequately with all the stability problems arising
to
in earth dam construction in one short section . However , the most important
TENSION
CRACK
& DEEP
- - : LEVEL
WATER
MEASURED
LONDONCLAY
RAILWAY
CUTTING PROBABLE
FAILURE
SURFACE
LONDON
CLAY
Fig . 22 . - Retaining wall failure in stiff - fissured London clay

(after Henkel , 1957 ) .
principles may be illustrated by considering the stability of a water retaining

dam built mainly of rolled earth fill ( Fig . 23 ) .
TRIAXIAL TEST 479
The stability of the slopes and foundation of an earth dam against shear
failure will generally have to be considered under three conditions :
( 1) During and shortly after construction ,

(2) With the reservoir full ( steady seepage ) , and
( 3) On rapid drawdown of the impounded water .
Additional considerations arising from the possibility of failure in a clay

foundation stratum are outlined in section 6 (h) . In this section attention will
be limited to the fill .
The stability may be calculated for all three conditions in terms of ef
fective stresses . This involves the measurement of c ' and Øi in the laboratory
and an estimate of the pore pressure values at each stage . The use of ex
plicitly determined pore pressures in the analysis enables the field measure
ments of pore pressure which are made on all important structures to be used
as a direct check on stability during and after construction . It also enables
the design estimates to be checked against the wealth of pore pressure data
now becoming available from representative dams - for example the extensive
work of the U . S. B. R . recently summarized by Gould (1959 ) and special cases
such as the Usk dam (Sheppard and Aylen , 1957 ) and Selset dam ( Bishop ,
Kennard and Penman , 1960 ) .
To work directly with undrained test results expressed in terms of total
stress may be unsafe in low dams as it implies dependence on negative pore
pressures which will subsequently dissipate ; and uneconomical in high dams
in wet climates as no account is taken of the dissipation of excess pore
pressure during the long construction period .
For the effective stress analysis the values of c ' and ' are generally ob
tained from undrained triaxial tests with pore pressure measurement . In
earth fill compacted at water contents well above the optimum a series of
consolidated - undrained tests with pore pressure measurements may have to
be used instead , in order to obtain a sufficient range of effective stresses to
define a satisfactory failure envelope .
For the analysis of the condition of long term stability under steady seep
age and for the case of rapid drawdown it is necessary to consider the effect
of saturation on the values of c ' and ' . As mentioned in section 4 (h ), test
results show that, in general, the value of $ remains almost unchanged .
Where c ' has an appreciable value in the undrained tests this will decrease .
However , tests using the improved techniques described by Bishop ( 1960 ) and
Bishop , Alpan , Blight , and Donald ( 1960 ) have failed to reproduce the high co
hesion intercepts previously reported in undrained tests . Provided c ' has
been accurately measured in either type of test , the differences may only be
of significance in important works where the margin of safety is small .
Whether c ' is likely to become zero in rolled fill on a really long term basis
is discussed by Bishop (1958 b) and Terzaghi (1958 ) . Evidence so far does
not appear to point to such a reduction .
The principal factors controlling the pore pressure set up during con
struction are :
( i ) The placement moisture content and amount of compaction , and hence

the pore pressure parameters ;
(ii ) The state of stress in the zone of the fill considered , and
iii
dissipation pore pressure during

of
of
( The rate construction

)
.
TOP WATERLEVEL
ROCK FILL
TAARTmani
EQUIPOTENTIALTHROUGHP
I FQUIPOTENTI
BEFOREDRAWDOWN
IMPERVIOUSFILL
AVERAGESHEARSTRESSON A GIVEN SURFACETHROUGHP
CONSTRUCTION 1. RESERVOIRFULL _
RESERVOIREMPTY
IMPOUNDING RAPID DRAWDOWN
DISSIPATION
PORE PRESSURE
ASSUMINGZERO DISSIPATION
PORE
PRESSURE
AT P
TIME
FACTOR
OF
SAFETY
TIME
017
OF OF OF
) 18 u, 40 +
MAX VALUEOFu, (ZERODISSIPATION (nc oc +hp WHERE DENSITY WATER
X
,
-
)
( net
+ het
Xc
Ow
hw
DENSITY FILL
IMPERVIOUS
♡
-
)
-r
(
DENSITY ROCK FILL

hp
8
ne
FOR FULL SATURATION

~
n
4
)
]
“
–
-
-
-
-
l-
-
-
OF
SPECIFICPOROSITY ROCKFILL
n
-
Fig
23
shear stress pore pressure and factor

of
The changes safety

in
,
.
–
.
for the upstream slope

an
of
earth dam
.
TRIAXIAL TEST 481
In section 3 shown that the pore water pressure set up under un
it was
drained conditions can be expressed in the form :
u = u, + B . 10. ( 7)
In Fig . 24 the values of u , and B are plotted against water content for a
series of samples prepared with the compactive effort used in the standard
compaction test . This clearly shows the sensitivity of the value of the initial
pore pressure to the placement water content ; the importance of this effect ,
both in design and construction , cannot be overemphasised .
It is usually assumed that the value of 0 , is equal to the vertical head of
soil y . h above the point considered , although the direction in which 01 acts
is not necessarily vertical . This is a reasonably satisfactory assumption
when averaged around a complete slip surface , but tends to overestimate the
pore pressure in the centre of the dam and underestimate it near the toes
( Fig . 25 , after Bishop , 1952 ) . It enables the pore pressure ratio required for
the stability analysis to be expressed , under undrained conditions as
ru = B +ud/ yh (9)
or , at the higher water contents where B is large and u , small , more simply
as
r, = B
( 10)
However , in most earth fills a considerable reduction in the average pore

pressure results from dissipation even during the construction period . A
numerical method of solving the practical consolidation problem with a mov
ing boundary has been given by Gibson (1958 ). It should be noted that in many
almost saturated soils even a small amount of drainage has a marked effect
on the final pore pressure , since it not only reduces the pore pressure al
ready set up , but also reduces the value of B under the next increment of load .
The theoretical basis of this reduction is discussed by Bishop ( 1957) , and it
is confirmed by field results from the Usk dam ( Fig . 26a ) .
As an example of the distribution of pore pressure at the end of con
struction the contours from the Usk dam are illustrated in Fig . 26b . The ef
fectiveness of the drainage layers placed to reduce the average pore pressure
in the fill will be apparent .
The average pore pressure ratio along a potential slip surface at the end
of construction may be kept within safe limits either by restricting the size
of the impervious zone , by strict placement water content control or by
special drainage measures (as at the Usk and Selset dams). Which is the
more economical procedure will depend on the climatic conditions and the fill
materials available .
For reservoir full conditions the pore pressure distribution may be pre
dicted from the flow net corresponding to steady seepage . Accuracy is diffi
cult to obtain owing to the non - uniformity of the rolled fill and differences in
the ratio of horizontal to vertical permeability , so conservative assumptions
should be made . However , with properly placed drainage zones the average
value of ru for the downstream slope is generally less than during con
struction , except in low dams or with rather dry placement.
135
701, 801, 907. 1001. SATURATION
120
COMPACTION DIAGRAM
115
10
WATERCONTENT %
9 10 11 12
-S
To
INITIAL PORE WATER PRESSURE

U. IN P.S. I
-20
0.2
THE PORE PRESSUREPARAMETERB,

(0 - 30 PSI. STRESSRANGE)
10 11 12
WATERCONTENT %
Fig . 24 . -- Represents Values from Standard Compaction Test under Equal

All Round Pressure Increase : Compacted Boulder Clay , Clay Fraction 6 %
TRIAXIAL TEST 483
31 : OS
-
L - - 6
20 - 6
115 - 16
IIO IOS oo 95 90 1
Fig . 25 . - Major Principal Stress o , as a Percentage of y . h , the Vertical

Head of Soil Above the Point Considered (After Bishop , 1952 ).
A method of predicting the excess pore pressures resulting from rapid

drawdown has been proposed by Bishop ( 1952 and 1954a ) . In this method the
change in pore pressure on drawdown is assumed to take place under un
drained conditions and is deduced from the stress change and the pore
pressure parameters ( see Fig . 23 ) . For saturated fills the value of B is taken
as l ; the change in the value of 01 is due to the removal of the water load
from the face of the dam and the drainage of water from the voids of the rock
fill .
This method shows reasonable agreement with the results of the field
measurements on the Alcova dam (Glover , Gibbs and Daehn , 1948 ) . Two re
cent cases of very rapid drawdown soon after completion are not in such good
agreement, but both involve complicating factors (Bazett , 1958 ; Paton and
Semple , 1960 ) .
Fig . 23 shows diagrammatically the variation in pore pressure and factor
of safety for the various phases in the life of the upstream slope of the em
bankment calculated as described above . The lowest values of factor of safety
are usually reached at the end of construction and on rapid drawdown .
For the downstream slope , end of construction and steady seepage are the
two critical stages . However , during steady seepage the danger is generally
not so much from the pore pressures , which are easily controlled by drainage
measures , but from the possibility of piping and internal erosion in the foun
dation strata , and from crack formation in the fill .
For small earth dams built largely of saturated soft clay , the stability dur
ing the construction period can be calculated by the Du = 0 analysis using the
undrained strength (Cooling and Golder , 1942 ) . However , where the fill is
much stronger than the clay foundation strata satisfactory results are not ob
tained (for example , Golder and Palmer , 1955 ) for the reasons given in section
6 ( a ) . The long term stability can of course be determined only by the ef
fective stress analysis .
of Slopes in Sand and Gravel on Drawdown . -
( g ) Stability
relatively pervious soils of low compressibility the distribution of pore
In
pressure on drawdown is controlled by the rate of drainage of pore water
./.
IN
SA
:
LB " P171
POREPRESSURE
ENDOFCONSTRUCTION
CLAYCORE
DRAINAGE
BLANKET DRAINAGE
BLANKET
10
SHUT
DOWN 631
PO
17
.
..
..
ROLLEDCLAYP •P7.
FIL
50
3040 100150200FT.
10
0
20
50
SQ
OF
LB
IN
VERTICAL
HEAD SOIL
:
/
.
.
(a
..
)
IN
POREPRESSURELBISO
: SHUTDOWN
SHUTDOWN
ENGOF CONSTRUCTION
30
40
50
60
70
80
SO
OF
IN
HEAD
VERTICAL SOIL: LEI
.
.
DE
NOTESPOREWATERPRESSURE
+
MEASURINGCELL
IN -
ALL CELLSARE NOT THESAMEVERTICALPLANE
THEY ARE SHOWNTHUSONLY FOR CONVENIENCE
980BERM
IN
ACTUAL
LEVELREACHED OCT1953
.
292
2
948
:
1
20 948BERMANDDRAINAGE
BLANKET
BE
-
- 4040-
- - -
- - -
GAUGEHOUSE
1 No
- -
2
916
20
.
2020
916
-
-
-
-
-
-
-
-
-- --
DRAINAGE
MATTRESS
-
MATTRESS
DRAINAGE
-
890 ABOVE PUDDLECORE

.D
.B
L
STONETOE
.
STRIPPEDGROUND
LEVEL CONCRETE
CUT- OFF
1012
W
1006
T
3 L
.
.
:1 .
:
980 BERMANDDRAINAGEBLANKET
2
12
1:
948 BERM ANDDRAINAGEBLANKET

(c
)
GAUGEHOUSE
No.
2
916
a
DRAINAGEMATTRESS
890 ABOVE PUDDLE CORE

.D
.B
STONETOE
L
STRIPPEDGROUNDLEVEL CONCRETECUT- OFF
Fig pore pressure from the Usk dam

26
Field measurements
of
The
a
(
)
:
.
.
–
reduction pore pressure and the value due pore pressure

of to
of
in
(b in
dissipation after Bishop 1957 and contours pore pressure

;
,
c
(
of )
(
)
expressed as percentage the vertical head soil October

of
in
,
a
1953 and October 1954 after Sheppard and Aylen 1957

,
).
(
TRIAXIAL TEST 485
from the soil. This condition can be represented by a series of flownets with
a moving boundary as shown by Terzaghi (1943 ) and Reinius ( 1948 ) .
The flow pattern is a function of the ratio of drawdown rate to permeability
and the values of the pore pressures to be used in the stability analysis can
be taken from the appropriate flow - net . The influence of the greater perme
ability in the horizontal direction is considerable , but, in one case examined ,
tended to increase rather than reduce the factor of safety .
The values of c ' and ' are obtained from drained tests , c ' approaching
zero for free - draining materials .
An example of a drawdown failure in Thames gravel is shown in Fig . 27 .
The initial slope of the gravel was 33° and the permeability about 0. 05 cm /
sec . The value of Ø ' in the loose state was 36º . Failure occurred when the
pool was lowered at a rate of about 1 foot per day (Bishop , 1952 ) .
(h ) The Stability of a Clay Foundation of an Embankment where the Rate of

Construction Permits Partial Consolidation . -
It is not uncommon in earth dam construction to encounter geological con
ditions in which the foundation strata include a soft clay layer at or near the
surface , of sufficient extent to be likely to lead to failure in an embankment
having conventional side slopes ( for example Cooling and Golder , 1942 ;
McLellan , 1945 ; Bishop , 1948 ; Skempton and Bishop , 1955 ; Bishop, Kennard
and Penman , 1960 ) . It is then necessary to assess the economics and practi
cability of a number of alternative solutions . The soft layer may be excavat
ed , if its depth and ground water conditions permit ; or an embankment with
very flat slopes may be accepted , safety being calculating using
its
of
factor
the method which assumes zero drainage Alternatively account may
=
,
0
u
the dissipation pore pressure which occurs due

of
of
be taken natural
to
drainage for example Bishop 1948 due special measures such as

or
to
,
,
(
vertical sand drains designed accelerate consolidation for example

to
,
,
(
Skempton and Bishop 1955 Bishop Kennard and Penman 1960 this case
In
;
,
).
effective stress analysis

an
is
used
.
expression for the initial excess pore pressure

An
saturated soft clay

in
a
layer where has been obtained by Bishop 1952

=
B
)
:
25
Ap
.po
2A
Au
42
)
(
=
+
.p
V
K
K
1
2
)2
?
[(
]
(
[(
)/
)/
–
–
-
where Ap denotes change total vertical stress due the fill

to
in
shear stress along the layer set up by the fill

>
>
T
initial vertical effective stress

,
and Kp initial horizontal effective stress clay layer

in
»
pore pressure
on
of
This expression illustrates the dependence the change

stress as well as on the change
shear vertical stress though the latter
in
in
predominates most practical cases

of
To avoid the error the estimate

in
in
.
which may arise from the change void ratio on reconsolidation undis
of
in
A
may be deduced from the relation between the

of
turbed samples the value

A
,
undrained strength and the effective stress envelope using an assumed value
,
The relationship given by Bishop

is
of
1952
K
.
.
of (
dissipation pore pressure

is
of
is of
on
The estimate the rate based the

theory here that the greatest uncertainity arises es
of
consolidation
It
,
.
pecially stratified deposits and field observations pore pressure are

of
in
important works
on
advisable
.
Fig . 27 . -- Drawdown failure in Thames Valley gravel (after Bishop , 1952).
EMBANKMENT
18FT 18
FT
13
FT.
ROADWAY
- 20FT
. - TOPBANKLEVEL
191
FT. O.
BANK
FILLOFCOMPACTED
MARL
BROKEN
STONE
SELECT SELECT
PUDOLE WALLDRAIN
COREL L (a) LAIN
SANDBLANKET9IN. THICK
BY TRANSVERSE
UNDER
FRENCH TOE
STONE
DRAINSAT60FT, CENTRES (b) SAND
WELLS 9IN
. DIAM
AT10
FT. CENTRES
- Ear GL. 148
FT
. OD.
REDSANDY CLAY
GREY
SILTANDCLAY
VAVITI 72
CONCRETE
CUT-OFF KEUPER
MARL
133
FT.
Oo SCALE
30 40 50FEET
Fig . 28 . - Downstream slope of the Chew Stoke dam showing vertical sand
drains to accelerate dissipation of pore pressure in soft clay foundation
(after Skempton and Bishop, 1955 ) .
487
TRIAXIAL TEST
The values of c ' and p are taken from drained tests or consolidated un
drained tests with pore pressure measurement .
The downstream slope of the Chew Stoke dam ( Fig . 28 ) which had a factor
of safety against a foundation failure of 0 . 8 using the Q = 0 analysis was safe
ly constructed using a sand drain spacing designed to give a factor of safety
of 1. 5 (Skempton and Bishop , 1955 ) . Field observations of pore pressure indi
cated that the actual factor of safety was rather higher than 1. 5 owing to the
greater horizontal permeability resulting from stratification of the clay . The
Selset dam , founded on a boulder clay with little apparent stratification ,
showed a smaller difference between predicted and observed pore pressure
values (Bishop , Kennard and Penman , 1960 ) .
(i) Some Special Cases . --

In the examples described above the variation in safety factor with time
was either a steady increase or a steady decrease during the period from the
end of construction until the pore pressures reached an equilibrium condition .
The following examples will illustrate that under certain conditions we may
temporarily encounter a lower factor of safety at an intermediate stage .
Such cases are obviously very dangerous, as a failure might well occur
some weeks or months after the completion of construction , in spite of the
fact that it had been ascertained that the factor of safety was adequate in both
the initial and final stages . The basic reason in each case is that the re
distribution of excess pore pressure which occurs during the consolidation
process may lead to a temporary rise in pore pressure outside the zone where
the load is applied .
An interesting example is the stability of a river bank in a clay stratum
under the action of the excess pore pressure set up by pile driving for a
bridge abutment in the vicinity ( Bjerrum and Johannessen , 1960 ) . The chang
es in pore pressure with time are shown diagrammatically in Fig . 29a for two
points , one in the centre of the pile group and one outside it , but beneath the
slope . The excess pore pressures set up by pile driving will dissipate later
ally as well as vertically , particularly in a water - laid sediment where the
horizontal permeability tends to be greater than the vertical .
The exact magnitude of the effect is difficult to predict theoretically , and
in this case field measurements of pore pressure were used to regulate the
progress of pile driving . The most critical distribution of pore pressure oc
curred shortly after piling was completed ( Fig . 29b ) . The factor of safety of
the slope was calculated using the effective stress analysis with c' = 200 lb . /
sq . ft. and Ø ' = 27° , and dropped from an initial value of 1. 4 to 1. 15 after pil
ing , assuming low water in the river in each case .
A case in which the spread of pore pressure led to an embankment failure

some time after construction has been analysed in detail by Ward , Penman
and Gibson (1955 ) . The clay foundation on which the embankment was built
included two horizontal layers of peat . Due to the relatively higher perme
ability of the peat , the redistribution of pore pressure after the end of con
struction resulted in temporary high pore pressures in the peat on both sides
of the embankment , where initially the pore pressures were low . Because of
the difference between the void ratio pressure curves for consolidation and
swelling , the reduction in effective stress due to the presence of an additional
volume of pore water is much greater than the increase in effective stress in
the zone from which this volume has migrated . An overall decrease in ef
fective stress along a critical slip path can thus occur ( Fig . 30 ) at an
SAFETYFACTOR EXCESSPOREPRESS SHEARSTRESS
. -
-
-
TIME
-
-
-
-
-
-
-
- TIME
TIME
TIME
-
PERIODOF POREPRESSUREDISSIPATION
PILE DRIVING
a
(
)
(b
)
CRITICAL SLIP CIRCLE ABUTMENT

12
LLW.
M
1
-
-
/M
5
T
-
-
-
-
-
10
TIMP
-
-
-
M2
15T
-
-
-
-
-
-
/
M2
S20T
/
MG
OF
80
TOTALVOLUME PILES
=
10M
5
up
by
pore pressure set pile driving

29
on
Fig The effect the stability

of
--
.
.
an adjacent clay slope Changes pore pressure and factor

of
of
in
(a
:
safety with time diagrammatic observed pore pressures just

(b
,
:
after the completion pile driving after Bjerrum

of
and Johannessen 1960

,
).
(
TRIAXIAL TEST 489
intermediate stage , although in the long term equilibrium state the bank foun
dation would have been stable . u
Other engineering operations may result in a similar danger , such as the
rapid construction of an embankment or stockpile even some way back from
a river bank , cut or quay wall close to limiting equilibrium , and the driving of
ordinary piles or screw piles through the clay slopes of rivers or harbours .
In such cases an awareness of the danger will either lead to a modification of
the operation , or to the use of field measurements of pore pressure as a con
trol while it is carried out .
(a) Embankment on a clay

foundation with a horizontal peat layer
diagrammatic ;
LE - INITIALGWL
w drmmmmmm PEATLAYER
+-
( b) changes in pore pressure with time,

as a fraction of y . H where y is density
0-2 0-4 0-6 0-8 10 12 14 16 18 20 22 24
of fill ;
AVERAGEVALUEOF
U ACROSSBASE OFBANK
80
VALUEOF HAI
CENIRE OF BANK
7 ( c) effect of redistribution on maximum

and average pore pressure for
colc = 1/4 where ç, is coefficient of
0 O5 consolidation and c . is coefficient of
01 02 03 04 swelling (after Ward , Penman, and
Gibson , 1954).
Fig . 30 .
u. Other failures of this type are described by Terzaghi and Peck ( 1948 ) .
It is probable that a number of failures attributed to ' creep ' could more
correctly be attributed to the redistribution of pore pressure which occurs
after construction .
(j) Design in Earthquake Areas. –
The analysis of the stability of a structure or dam in an area subject to
earthquakes raises special problems which are outside the scope of this
paper .
It is , however , known that a transient load will leave residual excess pore
pressures which may be positive or negative depending on the void ratio and
stress history of the soil ( Bishop and Henkel , 1953) . A possible way of evalu
ating the stability under earthquake conditions may therefore be to use a
series of consolidated - undrained tests in which the stress ratio during con
solidation is chosen to represent the conditions prevailing in the field before
the earthquake . The sample is then subjected to a series of small variations
in deviator stress under undrained conditions corresponding to the additional
seismic stresses . The magnitude of the residual pore pressure and the ad
ditional strain will indicate the likelihood of failure under field conditions .
A discussion of the additional shear stresses likely to be set up in earth
quake areas is given by Ambra seys ( 1959) .
7. CONCLUSIONS
The discussion and ca se records presented in this paper point to four main
conclusions :
( 1) The effective stress analysis is a generally valid method for analysing
any stability problem and is particularly valuable in revealing trends in sta
bility which would not be apparent from total stress methods .
Its application in practice is limited to cases where the pore pressures
are known or can be estimated with reasonable accuracy . These include all
the class ( a ) problems , such as long term stability and drawdown in incom
pressible soils , where the pore pressure is controlled by ground water con
ditions or by a flow pattern . It is also applicable to both class (a ) and class
( b ) problems where field measurements of pore pressure are available .
Those class (b ) problems where the magnitude of the pore pressure has
to be estimated from the stress distribution and the measured values of the
pore pressure parameters can often be solved more simply by the u = 0
analysis . However , this alternative gives no indication of the long term sta
bility and does not enable account to be taken of dissipation of pore pressure
during construction , which may contribute greatly to economy in design .
(II ) Where a saturated clay is loaded or unloaded at such a rate that there
is no significant dissipation of the excess pore pressures set up , the stability
can be determined by the Qu = 0 analysis , using the undrained strength ob
tained in the laboratory or from in - situ vane tests .
This method is very simple and reliable if its use is restricted to the con
ditions specified above . It is essentially an end of construction method , and
in the majority of foundation problems , where the factor of safety increases
with time, it provides a sufficient check on stability . For cuts , on the other
hand , where the factor of safety generally decreases with time, the Øu = 0
method can be used only for temporary work and the long term stability must
be calculated by the effective stress analysis .
TRIAXIAL TEST 491
analysis require the measurement

III
The two methods the shear
of
of
(
)
strength parameters
of
on
terms effective stress the one hand and
in
and
'
'
cy
the undrained shear strength under the stress conditions obtaining the
in
field on the other
of
For saturated soils the values and are obtained from drained tests
p
c
'
'
consolidated undrained tests with pore pressure measurement carried out
to on or
,
samples The range stresses at failure should be chosen
of
undisturbed
.
correspond with those the field Values measured the laboratory ap
in
in
.
pear be satisfactory agreement with field records with two exceptions
to
in
.
stiff fissured clays the field value lower than the value given by
of
is
In
c
'
standard laboratory tests some very sensitive clays the field value
of
in
♡
;
lower than the laboratory value

is
un
For partly saturated soils the values are obtained from
of
and
Q
c
'
drained undrained tests with pore pressure measurement or
or
consolidated
,
-
from drained tests Provided comparable testing procedures are used the
.
obtained appear not be significant from

of
differences between the values
to
Q
'
by
practical point will be slightly
of
The values
of
view influenced
a
c
.
'
moisture content differences resulting fromthe different procedures
.
cy
shear strength obtained from undrained triaxial tests

is
The undrained
on undisturbed samples or from unconfined compression tests except on
,
(
fissured clays and from vane tests situ cannot be obtained without
It
in
,
)
.
on
error undrained tests where the

of
risk the unsafe side from consolidated

,
under the overburden pressure The error
is
sample reconsolidated seri
is
.
normally consolidated clays low plasticity and though
of
ous can be
it
in
minimised by consolidating under the stress ratio obtaining the field the in
,
effect of reconsolidation on the void ratio cannot be avoided
.
For this same reason probably more realistic calculate the value
to
is
it
the pore pressure parameter for undisturbed soil from the relationship
of
of A
between the undrained strength undisturbed samples and the values

of
c
'
and rather than measure consolidated undrained test
to
in
Ø
it
a
IV ',
The reliability any method can ultimately be checked only

by
mak
of
(
)
ing the relevant field measurements when failures occur or when construction
operations are likely bring soil mass near limiting equilibrium The
to
to
a
number published case records which the data sufficiently complete

is
of
in
critical comparison still regrettably small

is
for methods
of
a
ACKNOWLEDGMENTS
Terzaghi Arthur Casagrande

of
The work Skempton and the late

W
,
,
A
K
.
.
.
Taylor has contributed

of
any study
of
the background shear

so
much
to
W
D
.
.
strength and stability that specific references the text are inadequate
to in
acknowledgment The authors would also like express their gratitude

to
.
their colleagues Imperial College and the Norwegian Geotechnical Institute

at
for valuable comments the manuscript

on
APPENDIX THE USE OF THE PARAMETER Øcu

I.
-
errors likely
4b
section reference has been made arise applying

In
to
to
in
the field the relationship between undrained strength and consolidation

in
pressure obtained the laboratory from the consolidated undrained test

in
.
-
Two inherent errors have been referred to : The effect of reconsolidation
after sampling on the void ratio and on the value of the pore pressure parame
ter A ; and the error arising from consolidation under a stress ratio different
from that obtaining in the ground . A further error may arise from the way
in which the results are introduced into the stability analysis .
This point is illustrated in Fig . 31 (after Bishop and Henkel , 1957 ) . The
test is usually performed by consolidating the sample under a cell pressure
p , and then causing failure under undrained conditions by increasing the axial
stress . The total minor principal stress at failure (03 ) is thus equal to p; the
total major principal stress is (01) cu : The slope of the envelope to a series
of total stress circles obtained in this manner ( Fig . 31a ) is denoted cu , the
angle of shearing resistance in consolidated undrained tests , and is about one
half of the slope of the effective stress envelope (denoted by 0 ' ) for normally
consolidated samples . This relationship between shear strength and total
normal stress can only be used in practice if the identity between consoli
dation pressure and total minor principal stress imposed in the test also
FAILUREBY INCREASING
DRAINEDTEST- EFFECTIVESTRESSES
(a )
cu
C-U TEST- EFFECTIVESTRESSES
C- U TEST- TOTALSTRESSES
(UCU
(obcu
FAILUREBY DECREASING
0)
C- U TEST- TOTAL STRESSES
C- U TEST- EFFECTIVESTRESSES
- DRAINEDTEST- EFFECTIVESTRESSES
( EU (55), Voicu ( 1)cu

(u)CUNEGATIVE
Fig . 31. – The consolidated -undrained test on a saturated cohesive soil in

terms of total and effective stresses : (a ) Failure by increasing major
principal stress 01 ; (b ) failure by decreasing minor principal stress
03 (after Bishop and Henkel , 1957 ) .
TRIAXIAL TEST 493
applies around the slip surface considered . Passive earth pressure appears
to be the only case in which this is approximately true .
Had the failure been caused by holding 01 constant and equal to p and de
creasing the total minor principal stress 03 , the undrained strength would
have remained the same, and a radically different value of Ocy would have
been obtained , Fig . 31b . The relationship between shear strength and total
normal stress would then approximate to the case of active earth pressure .
The general use of cy (defined in Fig . 31a ) as an angle of shearing re
sistance in conventional stability analyses is therefore likely to lead to very
erroneous results , even if the samples are anisotropically consolidated . If
03 increases during the undrained loading (as in foundation problems ) the
factor of safety will be overestimated ; if 03 decreases , as in the excavation
of a cutting , the error may lead to an under - estimate of the factor of safety .
The most logical solution appears to be to plot contours of undrained
strength in terms of the consolidation pressure in the ground prior to the un
drained loading to be examined , and then to use the ºu = 0 analysis . This
method is of course limited to the end of construction analysis , in which it is
assumed that insufficient time has elapsed for consolidation or swelling to
occur .
In rapid drawdown analyses suggested by Terzaghi (1943 ) and Lowe and
Karafiath ( 1959 ) the undrained strength is related to the effective normal
stress on the potential failure plane before drawdown . However , unless the
samples are failed by reducing the stresses , there is a danger of overesti
mating the undrained strength of compacted samples which are difficult to
saturate fully in the laboratory .
APPENDIX II . - BIBLIOGRAPHY ON SHEAR STRENGTH AND STABILITY
1. Ambraseys , N . N . ( 1959 ), The seismic stability of earth dams. Thesis .

( University of London ). London . 2 vol.
2. Bazett, D . J. ( 1958 ) , Field measurement of pore water pressures . Canadi
anSoil Mechanics Conference , 12 . Saskatoon . Proceedings , p. 2 - 15.
3. Berger ( 1951), Unpublished report .

4. Bishop , A . W . (1948 ) , Some factors involved in the design of a large earth
dam in the Thames valley . International Conference on Soil Mechanics
and Foundation Engineering , 2. Rotterdam . Proceedings , Vol . 2, p . 13
18 .
5. Bishop , A . W . ( 1952 ) , The stability of earth dams. Thesis . (University of

London ). London . 176 p .
6. Bishop , A . W . ( 1954 a ) , The use of pore pressure coefficients in practice .

Geotechnique , Vol. 4 , No . 4, p. 148 - 152 .
7. Bishop , A . W . (1954 b ), The use of the slip circle in the stability analysis
of slopes . European Conference on Stability of Earth Slopes, Stockholm .
Proceedings , Vol . 1, p . 1 - 13. Geotechnique , Vol. 5 , No . 1, 1955 , p . 7 - 17.
8. Bishop , A . W . ( 1957 ) , Some factors controlling the pore pressures set up

during the construction of earth dams . International Conference on Soil
Mechanics and Foundation Engineering , 4. London . Proceedings , Vol . 2,
p . 294 - 300 .
9. Bishop , A . W. (1958 a ), Test requirements for measuring the coefficient
of earth pressure at rest . Brussels Conference on Earth Pressure
Problems . Proceedings , Vol. 1, p . 2 - 14.
10 . Bishop ,
A . W . (1958 b ) , Discussion on : Terzaghi , K . Design and per
formance of the Sasumua dam . Institution of Civil Engineers . Proceed
ings, Vol . 11 , November , p. 348 - 352 .
11 . Bishop , A . W . ( 1959 ) , The principle of effective stress . Teknisk ukeblad
106 ,
Vol. Norwegian Geotechnical Institute
39
No 859 863 Publ
.p
,
.,
-
.
.
(
32
.)
Bishop pore pressure

12
of
The measurement triaxial
W
1960
in
the
A
,
),
.
(
.
.
52
60
test Pore Pressure and Suction Soil Conference London
p
in
,
.
.
-
.
Bishop Alpan Blight Factors con

13
and Donald
W
1960
,
.,
,
.,
,
G
),
A
V
.
.
(
of J
trolling the strength partly
on
saturated soils Research Conference
.
Shear Strength Soils Proceedings
of
Cohesive
.
Bishop Undrained triaxial tests
14
on
and Eldin saturated
W
1950
,
,
G
),
A
.
.
.
.
(
sands and their significance the general theory
of
shear strength
in
.
No
Geotechnique
13
Vol 32
.p
,
,
2
-
.
.
Bishop stress history
15
on on
The effect
of
and Eldin
W
1953
,
),
A
K
.
.
.
.
.
.
(
the relation between and porosity sand International Conference

in
$
Soil Mechanics and Foundation Engineering Zürich Proceedings , .
,
3
.
.
Vol 100 105
p
1,
.
-
.
.
Bishop Pore pressure changes during

16
and Henkel
W
1953
,
),
D
A
.
.
.
.
. J.
(
two undisturbed clays Soil Me
on
shear International Conference
in
chanics and Foundation Engineering Zürich Proceedings Vol
p
,
,
3
1,
.
.
94 99
-
Bishop of soil proper

17
and Henkel 1957 The measurement

W
,
),
D
A
, (
.
.
.
.
J.
ties the triaxial test London Arnold 190

in
.p
.
Bishop
18
Kennard and Penman 1960 Pore

W
M
,
.,
),
D
A
A
F
.
.
.
.
.
.
(
pressure observations Selset Pore Pressure and Suction

in
Soil
at
dam
.
47
36
Conference London
.p
,
-
.
Bishop Morgenstern Stability

19
1960 coefficients for earth

W
and
,
),
N
A
(
.
.
.
slopes preparation
In
.
Bjerrum Geotechnical properties Norwegian marine

20
of
1954
,
, ),
a
L
.
.
(
clays
69
Geotechnique Vol Norwegian Geotechnical

49
No
.p
,
,
4
-
.
.
(
Institute Publ
.,
4
.
).
Bjerrum Theoretical and experimental investigations

21
on
1954 the
,
),
b
L
(
.
shear strength soils Thesis Oslo 113 Norwegian Geotechnical

of
.p
.
Institute Publ
.,
5
.
).
Bjerrum Stability natural slopes quick clay European

of
22
1954
in
,
),
L
c
on (
.
Stability Earth Slopes Stockholm Proceedings Vol

of
Conference
,
2
,
- .
Norwegian
40
Geotechnique
16
Vol No 1955 101 119

p
p
,
1,
,
5
-
.
., .
.
(
10
Geotechnical Institute Publ

,
)
.
Bjerrum and Eide Stability clay

2
of
1956 strutted excavations

in
No O
,
, ,
32 ,
L
.
. .
. .
1 (
Geotechnique Norwegian Geotechnical Institute

47
Vol
.p
,
,
6
-
.
.
(
19
Publ
.,
).
TRIAXIAL TEST 495
24 . Bjerrum , L. and Johannessen , I. (1960 ), Pore pressures resulting from

driving piles in soft clay . Pore Pressure and Suction in Soil Conference ,
London , p. 14 - 17 .
25 . Bjerrum , L . and Kirkedam , R . ( 1958 ), Some notes on earth pressure in

stiff fissured clay . Brussels Conference on Earth Pressure Problems .
Proceedings , Vol . 1, p . 15 - 27 . (Norwegian Geotechnical Institute . Publ. ,
33 . )
26 . Bjerrum , L . and Kjaernsli, B. (1957 ), Analysis of the stability of some

Norwegian natural clay slopes . Geotechnique , Vol . 7 , No . 1, p . 1- 16.
(Norwegian Geotechnical Institute. Publ . , 24 ) .
27 . Bjerrum , L . , Simons , N . and Torblaa , I. (1958 ) , The effect of time on the
shear strength of a soft marine clay . Brussels Conference on Earth
Pressure Problems . Proceedings , Vol. 1, p . 148 - 158 . (Norwegian Geo
technical Institute. Publ ., 33 .)
28 . Bjerrum ,
L . and Øverland ,
A . ( 1957 ), Foundation failure of an oil tank in
Fredrikstad , Norway . International Conference on Soil Mechanics and
Foundation Engineering , 4 . London . Proceedings , Vol . 1, p . 287 - 290 .
(Norwegian Geotechnical Institute . Publ. , 26 ) .
29 . Bruggeman , J. R . , Zangar , C . N . and Brahtz , J. H . A . (1939 ) , Notes on

analytic soil mechanics . Denver , Colo . (Department of the Interior ,
Bureau of Reclamation . Technical memorandum , 592 ).
30 . Cadling , L . and Odenstad , S . (1950 ) , The vane borer . Sthm . 87 p . ( Royal

Swedish Geotechnical Institute . Proceedings , 2 ) .
31 . Campion , F . E . ( 1951 ), Part reconstruction of Bo - Peep tunnel at St.

Leonards - on - Sea . Institution of Civil Engineers . Journal , Vol. 36 , p .
52 - 75 .
32 . Casagrande , A . ( 1934 ) , Discussion of Dr . Jürgenson ' s papers, entitled

“ The application of the theory of elasticity and theory of plasticity to
foundation problems " and " Research on the shearing resistance of soils . "
Boston Society of Civil Engineers . Journal , Vol. 21 , p . 276 - 283 . Boston
Society of Civil Engineers . Contributions to soil mechanics (925 - 940 .
Boston 1940 , p . 218 - 225 .
33 . Casagrande , A . (1949 ) , Soil mechanics in the design and construction of
the Logan airport . Boston Society of Civil Engineers . Journal , Vol . 36 ,
p. 192 - 221 . (Harvard University . Graduate School of Engineering .
Publ . , 467 - Soil mechanics series , 33) .
34 . Casagrande , A. and Albert , S . G . (1930 ) , Research on the shearing re
sistance of soils . Cambr ., Mass . Unpubl . (Massachusetts Institute of
Technology . Report ) .
35 . Cassel , F . L . ( 1948 ) , Slips in fissured clay . International Conference on

Soil Mechanics and Foundation Engineering , 2. Rotterdam . Proceedings,
Vol. 2, p. 46 -50 .
36 . Coates , R . H. and Slade , L . R . ( 1958 ) , Construction of circulating -water
pump house at Cowes Generating Station , Isle of Wight. Institution of
Civil Engineers . Proceedings , Vol . 9, p . 217 - 232 .
37 . Cornforth , D . ( 1960 ) , Thesis . ( University of London ) . In preparation .
, L. F. Golder , ( 1942 ), The analysis of
the
. Cooling . Q.
of
, an
38 and H failure
earth dam during construction Civil Engineers Journal
of
Institution
.
Vol
19
38
55
.p
,
-
.
.
39
Daehn and Hilf 1951 Implications pore pressure

W
de
of
W
,
in
,
),
.
.
.
J.
.
(
sign and construction International Congress
on
rolled earth dams
of
, .
Large Dams New Delhi Transactions Vol 259 270
.p
,
1,
4
-
.
.
De Lory Long stability slopes
40
of
1957 term over consolidated
in
,
),
A
L
-
.
.
. .
clays ( University
of
Thesis London London
.
).
.
(
Di
41 Biagio and Bjerrum 1957 Earth pressure measurements
in
,
),
E
L
.
.
(
stiff marine clay
on
trench excavated International Conference Soil
in
a
, .
Mechanics and Foundation Engineering London Proceedings Vol
2
,
,
.
.
Norwegian Geotechnical Institute
26
196 202 Publ
p
.,
-
.
).
(
42
compressibility
of
on
Fraser 1957 The influence stress ratio
,
M
),
A
.
.
.
(
and pore pressure coefficients compacted soils in Thesis University
(
.
.
of
London London
).
43 ca
of
Garde Hansen and Thernõe 1960 Grain silo 100 000 tons
, ,
, ),
,
S
P
, -
.
.
(
No
Turkey
48
CN
pacity
14
22
Mersin Post Cph
.p
,
(
.)
-
.
in
The progress clay layer
44
of
Gibson 1958 consolidation
in
,
),
R
a
.
.
.
(
creasing thickness with time Geotechnique Vol No 171 182

in
.p
,
,
8
-
.
.
45
Gibson and Marsland 1960 Pore water observations
in
,
),
R
A
E
a
.
-
.
.
.
(
saturated alluvial deposit beneath loaded oil tank Pore Pressure and
a
.
. 78
W 84
Suction Soil Conference London

in
.p
,
-
. .
Deformability
46
Glover Gibbs
of
and Daehn
W
1948
,
.,
),
H
R
E
.
.
J.
the stability earth dams following

on
earth materials and its effect

of
a
rapid drawdown
on
International Conference Soil Mechanics and Foun

.
dation Engineering Proceedings Vol

77
80
Rotterdam
p
,
, a ,
2
5 of 5
-
.
at
47
Golder and Palmer 1955 Investigation bank failure

, Q
,
),
D
H
.
.
.
.
J.
(
No
Scrapsgate Isle Geotechnique

of
55
73
Sheppey Kent Vol
,
,
.p
1
-
.
.
pore pressures observed
48
Gould 1959 Construction rolled earth

in
,
),
P
.
J.
.
(
Department
97
dams Denver the Interior

of
of
Colo Bureau
p
,
.
.
(
Reclamation Technical memorandum 650

,
).
.
internal hydrostatic pressure

on
49
of
Hamilton The effects

W
1939
,
),
L
.
.
.
(
the shearing strength soils American Society for Testing Materials

of
- .
Proceedings Vol
39
1100 1121
.p
,
,
.
.
50
of
Hansen and Gibson 1949 Undrained shear strengths

,
),
R
B
E
.
J.
.
.
.
(
No
anisotropically consolidated clays Geotechnique Vol 189

.p
,
,
1
3
.
204
.
Investigations
of
51 Henkel 1957 two long term failures London

in
,
),
D
-
.
.
J.
(
on
Soil
at
clay slopes Wood Green and Northolt International Conference

, .
Mechanics and Foundation Engineering London Proceedings Vol

,
,
4
2
.
315 320
p
.
-
.
The relationships between the strength pore water

52
Henkel 1959
,
),
,
D
-
.
.
J.
(
pressure saturated clays Geo

of
and volume change characteristics

,
.
-
technique Vol No 119 135

.p
,
,
9
-
.
.
TRIAXIAL TEST 497
53 . Henkel , D . J ., ( 1960 ), The strength of saturated remoulded clay . Re

search Conference on Shear Strength of Cohesive Soils . Proceedings .
54 . Henkel , D. J. and Skempton , A . W. (1955 ) , A landslide at Jackfield , Shrop

shire , in a heavily over - consolidated clay . Geotechnique , Vol . 5 , No. 2,
p . 131 - 137 .
55 . Hilf , J .
W . (1948 ) , Estimating construction pore pressures in rolled earth
dams. International Conference on Soil Mechanics and Foundation Engi
neering , 2 . Rotterdam . Proceedings , Vol . 3, p . 234 - 240 .
56 . Hilf , J.
W. (1956 ) , An investigation of pore -water pressure in compacted
cohesive soils . Denver , Colo . 109 p . (Department of the Interior .
Bureau of Reclamation . Technical memorandum , 654 ) .
57 . Holmsen , P . ( 1953 ), Landslips in Norwegian quick - clays . Geotechnique ,

Vol. 3, No . 5 , p . 187 - 194 . (Norwegian Geotechnical Institute . Publ ., 2) .
58 . Hvorslev , M . J . ( 1937 ) , Über die Festigkeitseigenschaften gestörter bin
diger Böden . Kbh ., (Gad ). 159 p. ( Ingeniørvidenskabelige skrifter , A
45 ) .
59 . Ireland , H . O . (1954 ), Stability analysis of the Congress street open cut

in Chicago . Geotechnique , Vol . 4 , No . 4 , p . 163 - 168 .
60 . Janbu , N . (1954 ) , Application of composite slip surfaces for stability

analysis . European Conference on Stability of Earth Slopes , Stockholm .
Proceedings , vol . 3, p . 43 - 49 .
61 . Janbu , N . (1957 ), Earth pressure and bearing capacity by generalized
procedure of slices . International Conference on Soil Mechanics and
Foundation Engineering , 4. London . Proceedings , Vol . 2, p . 207 - 212 .
62 . Kallstenius , J . and Wallgren , A . ( 1956 ) , Pore water pressure measure
ment in field investigations . Sthm . 57 p . (Royal Swedish Geotechnical
Institute . Proceedings , 13 ) .
63 . Kenney , T . C. (1956 ) , An examinationof the methods of calculating the

stability of slopes . Thesis . (University of London ) . London .
64 . Kjaernsli , B . (1958) , Test results , Oslo subway . Brussels Conference on

Earth Pressure Problems . Proceedings , Vol. 2, p. 108 - 117 .
65 . Larew , H . G . ( 1952 ) , Analysis of landslides . Wash . D . C. 39 p . (Highway
Research Board . Bulletin , 49 ) .
66 . Laughton , A . S. ( 1955 ) , The compaction of ocean sediments . Thesis .

( University of Cambridge ) . Cambr .
67 . Little , A . L . and Price , V . E . (1958 ) , The use of an electronic computer
for slope stability analysis . Geotechnique , Vol. 8, No . 3, p. 113 - 120 .
68 . Lowe, J. and Karafiath , L . (1959 ), Stability of earth dams upon drawdown .

Panamerican Conference on Soil Mechanics and Foundation Engineering ,
1 . Mexico . Paper 2 - A , 15 p.
69 . McLellan , A . G . ( 1945 ), The Hollowell reservoir scheme for Northamp

. Water and water engineering , Vol. 48 , p. 7 - 26 .
ton
70 . Morgan , H. D . ( 1944 ), The design of wharves on soft ground . Institution

of Civil Engineers . Journal , Vol . 22 , p . 5 - 25 .
71 . Murphy , V. A . ( 1951 ) , A new technique for investigating the stability of
slopes and foundations . New Zealand Institution of Engineers . Proceed
ings , Vol. 37 , p. 222 - 285 .
72 . Nixon , J. K . (1949 ), p = 0 analysis . Geotechnique , Vol. 1, No . 3, 4 , p .

208 - 209 , 274 - 276 .
73 . Nonveiller , E . and Suklje , L . ( 1955 ) , Landslide Zalesina . Geotechnique ,

Vol. 5, No . 2 , p . 143 - 153 .
74 . Odenstad , S. (1949 ) , Stresses and strains in the undrained compression
test . Geotechnique , Vol . 1, No . 4, p . 242 - 249 .
75 . Paton , J. and Semple , N . G . ( 1960 ) , Investigation of the stability of an
earth dam subject to rapid drawdown including details of pore pressure
recorded during a controlled drawdown test . Pore pressure and Suction
in Soil Conference , London , p . 66 -71.
76 . Peck , R . B. ( 1942 ), Earth pressure measurements in open cuts , Chicago

(Ill . ) subway . American Society of Civil Engineers . Proceedings , Vol.
68 , p. 900 - 928 . American Society of Civil Engineers . Transactions ,
Vol . 108 , 1943 , p . 1008 - 1036 .
77 . Peck , R . B . and Bryant , F . G . ( 1953 ) , The bearing capacity failure of the
Transcona elevator . Geotechnique , Vol. 3, No . 5 , p . 201 - 208 .
78 . Penman , A. D . M . ( 1956 ) , A field piezometer apparatus . Geotechnique ,
Vol . 6, No . 2, p. 57 -65 .
79 . Peterson , R . ( 1952 ) , Studies - Bearpaw shale at damsite in Saskatchewan .
N . Y. 53 p. (American Society of Civil Engineers . Preprint, 52 ).
80 . Reinius , E . ( 1948 ) , The stability of the upstream slope of earth dams.

Sthm . 107 p . ( Swedish State Committee for Building Research . Bulletin ,
12 ) .
81 . Rendulic , L . (1937 ) , Ein Grundgesetz der Tonmechanik und sein experi

menteller Beweis . Bauingenieur , Vol. 18, No . 31 / 32, p . 459 - 467 .
82 . Rosenqvist , I. T . (1953 ) , Considerations on the sensitivity of Norwegian
quick - clays . Geotechnique , Vol . 3 , No . 5 , p . 195 - 200 . (Norwegian Geo
technical Institute . Publ . , 2).
, ( 1956 ), The slide ,
6th
83. Sevaldson R. A. in Lodalen October 1954 Geo

,
.
No
technique Norwegian
16
Vol Geotechnical Institute

.p
,
,
6
(
-
.
.
24
Publ
.,
)
.
84 Shon
Sheppard and Aylen 1957 The Usk scheme for the water
.
,
G
),
B
A
(
.
.
. R.
.
.
supply Institution Civil Engineers Proceedings Vol

of
of
Swansea
,
7
,
.
paper 6210 246 274

p
,
.
-
.
General theory earth pressure

on
Discussion
85
Simons
of
1958
N
,
),
.
:
(
.
.
Earth Pressure Problems Proceedings Vol

on
Brussels Conference
3
,
,
.
.
53
Norwegian Geotechnical Institute

50
33
Publ
p
.,
(
.)
-
.
.
An
Skempton investigation the bearing capacity

of
- of
86
soft
W
1942
,
of ),
A
a
(
.
.
.
clay soil Civil Engineers

18
Institution Journal Vol 307 321

.p
,
,
.
Skempton slip
. of
87
the West Bank the Eau Brink cut

W
1945
in
,
),
A
A
(
.
of .
.
Civil Engineers Journal Vol

24
Institution 267 287

p
,
,
.
.
-
TRIAXIAL TEST 499
88 . Skempton , A . W . (1948 a ), The 0 = 0 analysis of stability and its theo

reticalbasis . International Conference on Soil Mechanics and Foun
dation Engineering , 2 . Rotterdam . Proceedings , Vol. 1, p . 145 - 150 .
89 . Skempton , A . W . ( 1948 b ) , A study of the
immediate triaxial test on co
hesive soils . International Conference on Soil Mechanics and Foundation
Engineering , 2 . Rotterdam . Proceedings , Vol. 1, p . 192 - 196 .
90 . Skempton , A . W . (1948 c ), The rate of softening in stiff fissured clays ,

with special reference to London clay . International Conference on Soil
Mechanics and Foundation Engineering , 2. Rotterdam . Proceedings ,
Vol. 2, p. 50 - 53 .
91 . Skempton , A . W . (1948 d ), The geotechnical properties of a deep stratum
of post - glacial clay at Gosport . International Conference on Soil Me
chanics and Foundation Engineering , 2. Rotterdam . Proceedings , Vol .
1, p . 145 - 150 .
92 . Skempton , A . W . ( 1950 ), Discussion on : Wilson , G . The bearing ca
pacity of screw piles and screwcrete cylinders . Institution of Civil
Engineers . Journal , Vol. 34 , p. 76 .
93 . Skempton , A . W . ( 1951 ) , The bearing capacity of clays . Building Re
search Congress , London . Papers , division 1, part 3, p . 180 - 189 .
94 . Skempton , A . W . ( 1954 ) , The pore pressure coefficients A and B. Geo

technique , Vol. 4 , No . 4, p . 143 - 147 .
95 . Skempton , A. W . (1959), Cast in - situ bored piles in London clay . Geo
technique , Vol. 9 , No . 4, p . 153 - 173 .
96 . Skempton , A . W . and Bishop , A . W . (1954 ) , Soils . Building materials ,
their elasticity and inelasticity . Ed . by M . Reiner with the assistance
of A . G . Ward . Amsterdam , North - Holland Publ . Co . Chapter X , p . 417
482 .
97 . Skempton , A . W. and Bishop , A . W . ( 1955 ) , The gain in stability due to

pore pressure dissipation in a soft clay foundation . International Con
gress on Large Dams , 5. Paris. Transactions , Vol . 1, p . 613 - 638 .
98 . Skempton , A . W . and DeLory , F . A . ( 1957 ), Stability of natural slopes in
London clay . International Conference on Soil Mechanics and Foun
dation Engineering , 4. London . Proceedings , Vol. 2, p . 378 - 381.
99 . Skempton , A . W. and Golder , H . Q . ( 1948 ), Practical examples of the

P = 0 analysis of stability of clays . International Conference on Soil
Mechanics and Foundation Engineering , 2. Rotterdam . Proceedings ,
Vol . 2 , p . 63 - 70 .
100 . Skempton , A . W . and Henkel , D. J. ( 1960 ), Field observations on pore
pressures in London clay . Pore Pressure and Suction in Soil Confer
ence , London , p. 48 -51.
101 . Skempton , A . W . and Ward , W . H . ( 1952 ) , Investigations concerning a

deep cofferdam in the Thames estuary clay at Shellhaven . Geotech
nique , Vol . 3, No . 3 , p . 119 - 139 .
102 . Suklje , (1953 a ), Discussion on : Stability and deformations of slopes
L.
and earth dams, research on pore - pressure measurements , ground
water problems . International Conference on Soil Mechanics and Foun
dation Engineering , 3. Zürich , Proceedings , Vol. 3, p . 211 .
103 . Suklje , L . ( 1953 b ), Plaz pri Lupoglavu v ecocenskem flisu . ( Landslide

in the eocene flysch at Lupoglav . ) Gradbeni vestnik , Vol . 5 , No . 17 / 18,
p. 133 - 138 .
104 . Taylor , D . W . ( 1944 ) , Cylindrical compression research program on

stress -deformation and strength characteristics of soils ; 10 progress
report . Cambr ., Mass . 46 p . Publ . by Massachusetts Institute of
Technology . Soil Mechanics Laboratory .
105 . Taylor , D. W. ( 1948 ), Fundamentals of soil mechanics . N . Y ., Wiley .

700 p.
106 . Terzaghi , (1923 ) , Die Berechnung der Durchlässigkeitsziffer des

K.
Tones aus dem Verlauf der hydrodynamischen Spannungserscheinungen .
Akademie der Wissenschaften in Wien . Mathematisch - naturwissenschaft
liche Klasse . Sitzungsberichte . Abteilung II a, Vol. 132 , No. 3 / 4 , p .
125 - 138 .
107 . Terzaghi , K . ( 1925 ), Erdbaumechanik auf bodenphysikalischer Grund

lage. Lpz ., Deuticke . 399 p .
108 . Terzaghi , K . ( 1932 ), Tragfähigkeit der Flachgründungen . International
Association for Bridge and Structural Engineering . Congress , 1.
Paris . Preliminary publ., p. 659 -683 , final publ., 1933 , p. 596 - 605 .
109 . Terzaghi , K . (1936 a ) ,
Stability of slopes of natural clay . International
Conference Soil Mechanics and Foundation Engineering . 1. Cambr . ,
on
Mass . Proceedings , Vol. 1, p . 161 - 165 .
110 . Terzaghi , K . (1936 b ), The shearing resistance of saturated soils and
the angle between the planes of shear . International Conference on Soil
Mechanics and Foundation Engineering , 1. Cambr . , Mass . Proceed
ings , Vol . 1, p. 54 - 56 .
111. Terzaghi , K. ( 1943 ), Theoretical soil mechanics . N . Y ., Wiley . 510 p.
112 . Terzaghi , K . ( 1958 ), Design and performance of the Sasumua dam . In

stitution of Civil Engineers . Proceedings , Vol . 11, November , p . 360
363 .
113 . Terzaghi , K . and Peck , R . B . (1948 ), Soil mechanics in engineering

practice . N. Y . , Wiley . 566 p .
114 . Toms , A . H . ( 1948 ) , The

present scope and possible future development
of soil mechanics British railway civil engineering construction and
in
maintenance. International Conference on Soil Mechanics and Foundation
Engineering , 2 . Rotterdam . Proceedings , Vol. 4 , p. 226 - 237 .
115 . Toms , A . H . (1953 a ) , Discussion on : Conference on the North Sea

floods of January 31st - February 1st, 1953 . Institution of Civil Engi
neers , Publ. , p . 103 - 105 .
TRIAXIAL TEST 501
116 . Toms , A . H . (1953 b ) , Recent research into coastal landslides at Folk
stone Warren , Kent , England . International Conference on Soil Me
chanics and Foundation Engineering , 3. Zürich . Proceedings , Vol . 2 ,
p . 288 -293 .
117 . Tschebotarioff , G . P . ( 1951), Soil mechanics , foundations , and earth

structures ; an introduction to the theory and practice of design and con
struction . N . Y., McGraw - Hill . 655 p .
118 . U . S. Department of the Interior . Bureau of Reclamation ( 1951 ), Earth
manual ; a manual on the use of earth materials for foundation and con
struction purposes. Tentative ed . Denver , Colo . 332 p .
119 . Ward , W . H ., Penman , A ., and Gibson , R . E . (1954), Stability of a bank
on a thin peat layer . European Conference on Stability of Earth Slopes ,
Stockholm . Proceedings , Vol. 1, p. 122 - 138 , Vol . 3, p. 128 - 129 . Geo
technique , Vol. 5 , No . 2 , 1955 , p . 154 - 163 .
120. Waterways Experiment Station , Vicksb ., Miss . ( 1947 ), Triaxial shear

research and pressure distribution studies on soils . Vicksb . , Miss .
332 p.
121 . Waterways Experiment Station , Vicksb . , Miss . ( 1950 ), Potamology in

vestigations . Triaxial tests on sands , Reid Bedford Bend , Mississippi
river . Vicksb ., Miss. 54 p. (Report , 5 - 3) .
122 . Wilson , G . ( 1950 ), The bearing capacity of screw piles and screwcrete
cylinders . Institution of Civil Engineers . Journal , Vol. 34, p . 4 - 73 .
123 . Wood , C. C . ( 1958 ), Shear strength and volume change characteristics

of compacted soils under conditions of plane strain . Thesis . (Uni
versity of London ). London .
FACTORS CONTROLLING THE STRENGTH OF PARTLY SATURATED
COHESIVE SOILS
By Alan W . Bishop , 1 I. Alpan, 2 G. E. Blight, 3 and I . B. Donald
1. INTRODUCTION
This paper describes some aspects of research at present being carried

out at Imperial College into the deformation and strength characteristics of
partly saturated soils . Owing to the limited amount of reliable test data so
far available , only tentative conclusions can be drawn at this stage . However ,
it is felt that the work presented in this paper will serve to illustrate the
relative importance of the various factors controlling the strength of partly
saturated soils . It may also provide an experimental background for the dis
cussion of basic principles and for the planning of further investigations .
2. THE PRINCIPLE OF EFFECTIVE STRESS IN SATURATED AND IN

PARTLY SATURATED SOILS
Soil , as an engineering material, may be visualized as a compressible

skeleton of solid particles enclosing voids, which in saturated soil are filled
with water , and in partly saturated soil with both air and water . Shear
stresses are carried only by the skeleton of solid particles , except at very
high rates of strain . On the other hand the total normal stress on any plane
is in general the sum of two components - the stress carried by the solid
particles and the pressure of the fluid in the void space .
It is evident that it is only that part of the cotal normal stress which is
carried by the soil skeleton that controls volume change and the frictional
component of shear strength .
For soils containing a single fluid in the pore space , either water or air ,
the effective normal stress (denoted o ' ) is in each case given by Equa
tion 1. a
a. Modern views on the principle of effective stress originate in the work of

Terzaghi on the theory of consolidation ( 1923 ) et seq . and on shear
strength ( 1932 ) . The validity of the principle has been demonstrated , for
saturated soils , by the experimental work of Rendulic ( 1937 ), Taylor
( 1944 ), Bishop and Eldin (1950), and Laughton ( 1955 ).
It should also be noted that Equation ( 1) holds true irrespective of the con
tact area between the soil particles . The theoretical and experimental
grounds for this view are discussed by Bishop and Eldin ( 1950 ) , Laughton
(1955 ) , Bishop (1959 ) and Skempton ( 1960 ) .
1. Reader in Soil Mech . , Imperial College of Science and Tech . , Univ . of

London , England .
2. Imperial College of Science and Tech ., Univ . of London , England .
3. Imperial College of Science and Tech . , Univ . of London , England .
4. Imperial College of Science and Tech . , Univ . of London , England .
503
50 % SHEAR STRENGTH CONFERENCE
ơ' = 0 - u . . . . . . . . . . . . . . . . . . ( 1)
where o denotes total normal stress and u denotes fluid pressure

in the pore
space .
The case of partly saturated soil , however , calls for special treatment
,
since the pore space contains two fluids , air and water , which may be in
equilibrium at pressures which differ considerably due to surface tension .
A tentative expression for effective stress for the general case was put for
ward by Bishop ( 1955 ) :5
O' = 0 - Ua + x (ua - Uw ) . . .. .. . . . . . . . ( 2)
where ug denotes pressure in the gas and vapour phase and uw denotes
pressure in the pore water . ( The lecture in Oslo in which this equation was
discussed appears in Teknisk Ukeblad October 22 , 1959 . )
The value of the parameter x is unity for saturated soils and is zero for
dry soils . Intermediate values will depend primarily on the degree of satu
ration S, but will also be influenced by factors such as soil structure , the
cycle of wetting and drying or stress change leading to a particular value of
S . It was anticipated that the general form of the relationship would be as
illustrated in Fig . 1. Values deduced from test results are given in subse
quent sections of this paper .
It is evident that some concept of this type must be accepted in extending
the principle of effective stress to soils which are not close to full saturation .
For the special case when the pore air pressure is atmospheric and the pore
water pressure is negative , expressions containing a similar parameter have
been put forward by Croney , Coleman and Black ( 1958 ) , Aitchison ( 1960 ) and
Jennings ( 1960 ) .
It is also generally recognized that the value of the parameter x in the re
lationship between volume change and effective stress will not necessarily be
identical with that in the strength - effective stress relationship , though there
is little quantitative information on this point as yet . The present paper is ,
however , concerned primarily with the application of the principle of effective
stress to the strength relationships .
OO 100 %
Fig . 1. - Anticipated Relationship Between the Parameter x and the

Degree of Saturation S.
5. Items indicated thus , Bishop ( 1955 ) , refer to corresponding entries ,

arranged alphabetically , in the Appendix References .
CONTROLLING FACTORS 505
The effective normal stress , with respect to shear strength , may be de

fined as that stress which , with zero pore air pressure and pore water
pressure , would give the same shear strength as the combination of total
normal stress and pore air and water pressure under consideration . In
terms of the modified Coulomb failure criterion the maximum shearing re
sistance on a given plane is then T where :
' [ o - ua
tan
( – uw )]
ol
T= c + + x ua
..
..
..
..
..
)
(3
where the apparent
is is
cohesion
c
terms stress
'
of
effective
in
.
the angle shearing resistance
of
'
)
evaluatingthe parameter test results and applying the re
In
from
in
,
x
practice simplest and most convenient

is
sults assume that for
to
in
it
a
given initial void ratio the values are independent the degree
of
of
and
,
c
'
fully
of
on
saturation The values and can then be obtained from tests
of
O
c
.
'
'
saturated samples and the parameter will represent the overall effect
on
x
,
the expression for shear strength of partial saturation This the procedure
is
.
adopted the present paper dealing with compacted soils
in
in
.
However there some evidence that the presence of air the voids may
is
in
,
modify the structural behaviour the soil skeleton itself under change
of
in
stress and
of
influence the values and volume change during shear

so
x o
',
c
'
.
This would mean that apparent values obtained by the simple procedure
of
outlined above would not be related quite as directly the degree satu
of
to
ration as simple mechanical model would indicate The apparent value
of
a
would still remain the most convenient practical parameter

x
TYPES OF STRENGTH TEST USED

3
.
All cylindrical samples brought

on
tests have
carried out failure
to
been
by axial compression drainage have been used
of
Four different conditions

.
with compacted samples

.
Undrained Tests
i.
these tests the sample rubber membrane after preparation

is
sealed
In
in
a
prevent any change During

of
the content water and air the sample

to
in
in
the application the all round pressure and the axial load the triaxial cell
of
in
both the pore water pressure and pore air

of
measurements are made

pressure
.
This test subject several limitations

is
pore
of
The measurement
to
water pressures below over long periods tends

13 give trouble ow
to
p
in s
-
.
.
i.
ing cavitation the water the measuring device Since the equilibrium
in
to
air pressure the unstressed sample generally also slightly negative

is
in
samples where the initial differ

be
undrained tests can only

on
carried out
ence between pore air and water pressure
10
less than about

is
p
s
.
i..
strictly undrained
In
addition difficult
is
maintain state with

to
it
a
'
'
rubber membrane since the air the sample diffuses through the membrane
in
Uw
this means that the capillary term

ua
Since zero this state

is
x
,
.
-
(
would general occur only the fully saturated soil with zero pore water
in
in
pressure
.
into the water in the cell . This leads to significant reduction in the pore air
pressure , and hence in the pore water pressure . Though this does not lead to
an error in the measured effective stress characteristics , it will give a mis
leading impression of the change in 'undrained ' strength with time. Tests in
which the rubber jacketed sample is surrounded in mercury in an intermedi
ate cell are now being used for partly saturated soil where strictly undrained
conditions are necessary .
ii . Constant Water Content Tests

the air pressure in the pore space is controlled independent
In these tests
ly . It is either allowed to equalize with atmospheric pressure or is boosted
through an accurate pressure valve so as to raise the pore water pressure ,
which differs from it by the capillary term (ug - Uw ) , into the range in which
it can be measured without cavitation in the pore water pressure system .
This necessitates the use of a saturated porous disc or probe to measure the
air
pore water pressure having an
ua
in uw
entry value greater than
A
-
.
(
)
ceramic having air entry value
of
an
about 60 has been used some

of
p
s
.
.i
.
the present tests
.
This procedure enables the tests
of
be extended into the range lower
to
water contents and higher clay fractions where the difference between air and
water pressure the sample major importance
is
of
in
.
iii Consolidated Undrained Tests with Full Saturation
.
by
these tests the sample progressively increasing the cell

is
saturated
In
pressure and admitting water under pressure the pore space through
to
a
volume gauge High back pressures the range 60 100 in2 are gener
lb
in
)
-
(
.
ally required short term basis Full saturation

on
ensure full saturation

to
.
can be checked by observing the rise pore pressure under undrained con
in
ditions for further increase cell pressure

in
a
is .
the sample while still the partly saturated

in
of
Over consolidation
it
-
state can be avoided by increasing the cell pressure and pore pressure
in
a
series of small steps with time for equalization allowed at each stage
.
After saturation the sample sheared under undrained conditions with the
is
measurement of pore water pressures

.
Drained Tests with Full Saturation

iv
.
The saturation procedure follows that the previous test The sample
of
is
.
then sheared under drained conditions with constant back pressure The
a
any significant
up
give ample time for the build

of
of
rate strain chosen

is
to
excess pore water pressure be avoided

to
Full saturation has been adopted for the consolidated undrained and drained
avoid the difficulty interpreting results from samples having
of
an
tests
to
arbitrary and possibly non uniform degree

of
saturation also enables

It
.
-
applied
be
investigating true cohesion and friction

of
to
Hvorslev
to
method
's
the results
.
re
on
of
On
two the materials tests have also been carried out samples
approximately full
at
moulded saturation
.
Drained test results are also quoted for silt placed as saturated slurry
a
rubber membrane enclosed former small negative pore pressure

in
in
A
a
.
top
applied after sealing the cap give the sample sufficient rigidity for
is
to
,
the former to be removed and the triaxial cell assembled . Tests are then
run either as saturated drained tests with a positive cell pressure and zero
pore water pressure ; or as partly saturated drained tests in which a negative
pore water pressure is applied to the fine porous disc at the base of sample
of sufficient magnitude to cause drainage and entry of air into the sample
through a connection to the top cap . Measurements of the amount of water
draining from the sample and of
its
overall volume change enable the average
degree of saturation be calculated throughout the test
to
.
APPARATUS FOR THE MEASUREMENT AND CONTROL OF PORE
4
.
WATER AND PORE AIR PRESSURE
pore pressure triaxial

of
of
detailed discussion the measurement the
in
A
tests has been given by Bishop paper
on
the Conference Pore Pressure
to
in
a
and Suction Soils held London April 1960 and only the main points
in
in
in
,
,
need be summarized here
.
The basic layout the apparatus for the direct measurement pore water
of
of
pressure the present series shown diagrammatically Fig
sy of
tests
in
is is
in
2
.
.
The pressure the measuring stem balanced against the water pressure
in
the pore space the sample using

of
an
accurate null indicator show that
to
in
no
satisfied
is
the flow condition

.
-
water filled small bore copper tube connects saturated porous cera
A
a
at -
glass capillary
of
mic disc the soil sample the upper end

of
the base
to
about mm internal diameter The lower part

of
the glass tube

of
tube
of is
1
filled with mercury and the tube dips below the surface
an
of
extension
,
a
SMALL BORE
WATER
COPPER TUBE
MERCURY
-
ZZZ
LZZA
2
WATER
22
00000000000000
Fig Layout the Apparatus for the Direct Measurement Pore

of
of
2
-
.
.
Water Pressure Using Null Indicator

a
small enclosed trough of mercury . Water pressure is applied to this surface
by the screw - controlled plunger so as to balance the pore pressure changes
in the sample , the no - flow condition being observed from the level of the
water -mercury surface in the capillary tube . When measurements are not
being made the apparatus is shut - off by a valve near the cell base , a piston
type valve being used to avoid displacement of water . Details of the apparatus
and measuring technique are given by Bishop and Henkel (1957 ).
Two special pieces of equipment have also been used in the present tests .
A ceramic probe placed horizontally at the mid - height of the sample and con
nected by a small - bore nylon tube to an additional null - indicator has been
used to study pore pressure gradients in the undrained test . A more compact
null indicator made from a block of Perspex which can be placed close to the
base of the triaxial cell has also been used in some tests . This minimizes
the length of tubing subject to temperature effects , etc. , and also facilitates
the initial de - airing of the system .
For the measurement of pore water pressure in partly saturated soils a
ceramic disc of high air entry value is used . Two commercial porous
ceramics , Aerox 'Celloton ' grade VI and Doulton grade P6A have proved
satisfactory and their relevant physical properties are given in Table I .
TABLE I. -- Properties of Two Porous Ceramics
Aerox ' Celloton ' | Doulton Grade PGA

Grade VI
Porosity 46 per cent 23 per cent
Air entry value 30 p . s . i. 22 p . s . i.

Permeability 2 . 9 x 10 - 6 .
2 1 x 10 7 -
cm. / sec . cm. / sec .
A disc of higher air entry value has been made by consolidating kaolin in
an oedometer , and then air drying and firing it . The air entry value is initi
ally about 45 p . s . i. , but has risen with use to about 60 p . s .i . This represents
the upper limit of the difference (ua - Uw ) which has been measured in the
present series of tests .
The drainage characteristics of the Aerox and kaolin discs are given in
Fig . 3.
The mountings used for the discs in the standard cells for 4 inch diameter
and 1- 1 / 2 inch diameter samples are illustrated in Figs . 4a and 4b re
spectively . It is most important to have a rigid mounting and a perfect seal .
Each disc is ground flat and sealed in with “ Twin - bond ', an epoxy resin which
has proved most suitable for this purpose .
The probe is cut from a disc of the same ceramic and is cylindrical in
shape so that it can be inserted in a hole drilled in the side of the sample .
The probe is sealed into a brass connector with ' Twin - bond '. The connector
passes through the rubber membrane which is sealed against it by a screw
down collar .
The measurement of pore air pressure calls for a porous element which
will give free communication between the air voids and the air pressure
mea suring system . It must therefore have a lower attraction for water than
,
BLOW THROUGH
UNIT AREA Cd
AEROX CELLOTON VI
VOLUMEOF WATERDRAINEDFROM
.
BLOW THROUGHT
COMPRESSED FIRED
KAOLIN DISC
oto od
o
o
o
-
-
-
-
-
-
-
-
-
-
10 15 20 25 30 45 50
AIR WATER PRE SURE DIFFERENCE ACROS D
-
Drainage Characteristics
of
Fig Two Porous Ceramic Discs
in
3
-
.
.
Long Term Tests
represented the capillary differ

by
the soil sample where the attraction

is
,
Uw
air filled connection

its
Its volume together with that

ua
to
of
ence
).
-
(
compressible fluid and

to be
the null indicator should since air

is
small
,
a
compensating adjustment the null position would otherwise have be made
to
.
single layer glass fibre cloth woven from coarse dressed yarn has
of
A
proved satisfactory for this purpose used some tests the form of
is
it in
in
). It
.
disc at the top the sample Fig necessary

of
4a
Where measure
is
to
a
.
(
air and water pressure the sample as dissipation test

at
the same end

of
in
,
with drainage the opposite end length glass fibre tape about inch
of
at
in
1
a
,
width for inch high samples placed round the sample just above the base
is
8
,
)
connection made through the rubber membrane the null indicator

is
to
and
a
The possibility measuring directly large negative pore pressures

is
of
limited by the inability the measuring system carry tension

of
the water
to
in
for more than brief period This places lower limit direct measure
. to
a
a
.
atmosphere over long periods about at sea level

13
ments
or
, of
p
1
s
.i
.
.
-
artifi
by
extended beyond this range

be
However indirect measurements can

cally increasing the tmospheric pressure under which the test run This
is
.
'
'a
technique has been described by Hilf 1956 who applied tests under equal
to
it
(
all round pressure and similar the pressure membrane method de

is
to
, ,
scribed by Croney Coleman and Bridge 1952

.
(
Two developments this technique have been used carrying out constant
of
in
water content tests Unconfined compression tests have been performed on

'
in .
a '
samples enclosed rubber membrane prevent evaporation but not

to
glass fibre tape

of of
sealed wick placed between the rubber membrane

is
,
a
(
and the surface the top cap and subjected all round air pressure applied
to
)
through Sunvic constant pressure valve The air pressure applied with
is
a
.
'
'
prevent loss water from the porous disc into

of
delay
of
the minimum
to
POLYTHENE TUBE FOR AIR
510
PRESSUREMEASUREMENT
CONTROL
OR
STAINLESS STEEL CAP
_
RUBBER RINGS
DISC OF GLASS FIBRE CLOTH

WITH EXTRA LAYER AROUND
CONNEXION
IN
4°
.
_
'
' OR
SAMPLE ENCLOSED PERSPEX NYLON CAP
DIA
IN
-0 "01
RUBBERMEMBRANE
.
O
THICKNESS RUBBER RINGS

.
SAMPLE ENCLOSED
DIA
of IN
SHEAR STRENGTH
RUBBERMEMBRANE
IN
oor THICKNESS
at
.
dia
RADIALGROOVES CERAMIC DISC
S
WIDExã DEEP
SEAL OF RUBBEROR EPOXY
RESIN BONDED CERAMIC
TO
+
á
. X
DIA
CERAMICDISC
&
DISC BRASSMOUNTING
TO
FILLER USED TO
LEVEL OFF RECESS
OF
BRONZE SCREWS SEAL EPOXY RESIN
SIX
BONDED CERAMIC
TO
BRASSMOUNTING
CONFERENCE
O
IN
RUBBER RINGS GROOVE DISC AND

TO
BRONZE
%
HAVINGABOUT 105 OF PEDESTAL

BRONZE PEDESTAL
O
SECTIONAL AREA OF RING

IN
SHALLOWRECESS
)0(
CERAMICDISC
PORE PRESSURECONNEXION
)(a
of
Ceramic Discs for Samples
.
-.4
Fig Details Triaxial Cell Bases with High Air Entry Value
in
.
for
/-
1
1
2
Dia
in
)b(
.
Samples Inches
4 Inches Dia and
the sample . The test is then run with the pore air pressure equal to the air
pressure acting all round the sample , i . e . 03 = up , and represents the lower
limit of effective stress for the given water content .
Higher effective stresses are obtained by sealing the sample and applying
an air pressure less than the cell pressure to the glass fibre disc at the top
of the sample . To avoid delay in the equalization of air pressure in the less
permeable soils , several tapes of glass fibre are placed down the sides of
the sample to the base (or alternatively air pressure is applied to a tape
placed around the sample just above the base ). The test is then run either at
constant pore air pressure , or ' undrained ' during the shearing stage with
both pore water and pore air pressure measured .
In the consolidated undrained and drained tests in which full saturation is
ensured by an applied pore water pressure , a water - kerosene surface is used
in the volume gauge to measure the volume of water entering or leaving the
sample during the test . The constant water pressure is applied by the self
compensating mercury control described by Bishop and Henkel ( 1957 ) .
5. THE EFFECT OF TIME ON THE RESULTS OBTAINED
Time has an important influence on the observed values of pore water and
pore air pressure . Several independent factors are involved . (a ) Change in
pore pressure at constant water content and constant all round stress . In
Fig . 5a is shown the relationship between negative pore pressure and time
for a freshly remoulded sample of London Clay at constant water content
( enclosed in a pre - soaked rubber membrane and standing in a triaxial cell
with a little free water around the base of the pedestal to maintain a saturated
atmosphere) . Since the time effect cannot be accounted for by the response
time of the apparatus or non - uniformity of the sample , its origin is assumed
at
rearrangement
of
lie
to the clay water system constant volume and

in
a
may represent the mechanism of the thixotropic regain strength

in
Fig the changes pore water and pore air pressure are shown for
5b
in
In
as
partly saturated compacted sample Earlier tests using water the cell
a
fluid had indicated large drop pore air pressure with time but this ap
in
,
a
pears air through the rubber membrane Tests using

of
be due diffusion
to
to
hollow porous ceramic cylinder place sample Fig

of
show that the

in
6
a
a
.
(
air pressure below its initial value

po
drop gauge under undrained con

in
greatly reduced aluminium foil jacket sealed with plastic

an
ditions
is
if
This clearly points

of
tape placed beneath the rubber membrane

is
loss
to
.
by )
air diffusion into the cell fluid and special tests are now being run under
,
mercury minimize this error as metal foil jacket impracticable

is
to
in
a
,
most soil tests

.
Time lag equalization pore pressure throughout the sample during

of
in
b
(
. )
single measurement pore pressure either

of
The validity using

of
shear
a
probe interpreting test data depends the as

on
from the base

or
from
in
a
sumption that the pore pressure sensibly constant over the whole
of
. is
the
which failure occurring Pore pressure gradients may arise due
is
zone
in
non uniformity stress during the shear stage due end restraint and
of
to
to to
-
-
narrow zone failure which appears

of
tendency be characteristic
to
to
a
some samples This can be best ensured by running the test at rate which
a
)
.
permits the pore pressure equalize over the whole sample when
ba
se
to
measurement will provide the required value

.
TIME AFTER REMOULDING MINS )

100 1000 10000 100000
LONDON CLAY
L. L. - 607.
P.L . - 27%
Ja
3w - 34.5 /
/(
INY
LB
PORE PRESSURE
3
4
5
6
7
IN
TIME DAYS
Fig Change Negative Pore Pressures Unconfined Sample

in
in
5
a
-
.
).
(
withTime after Remoulding
TIME AFTER SEALING MINS

(
)
000
10
10 100 000 100 000
1,
,
2
+
-
-
Uo
lb
in
=
,
1
5
.
(
)
/
blin
COMPACTED BOULDER CLAY

%
CLAY FRACTION
PRESSURE
OPTIMUM WATER CONTENT

9
3
.
COMPACTION WATER CONTENT

7
9
.
PORE
lb
in
=
U
(0
1
5
)
.
/
?
3
IN
TIME DAYS
Fig Change Pore Water Pressure and Pore Air Pressure with
in
5
(b
.
).
--
Time after Sealing Rubber Membrane under Mercury ompacted

in
a
.
(C
Clay Dia Sample

1
1
2
"
.
-
/
)
This necessitates a rate of testing which is much slower than is generally

recognized . This is conveniently illustrated by tests of 4 inch diameter x 8
inch high samples of compacted slay shale having a coefficient of consoli
dation of about 2. 7 x 10 - 4 cm2 / sec . since the equalization time is very large
compared with the response time of the pore pressure apparatus . Test re
sults are given in Figs . 7 and 8 illustrating the difference between base and
mid -height probe measurements of pore water pressure for strain rates giv
ing 20 % axial strain in 8 hours and 120 hours respectively . The marked non
uniformity of pore pressure at the higher strain rate will be noted . It is also
of interest that the base measurement gives too high a value even at the stage
when the pore pressure is rising rapidly with stress .
In Fig . 9( a ) a theoretical relationship is given between time to failure t,
n - Cyti for
expressed as time factor T = height
2H
undrained sample
an
and
of
,
H2
of up
the degree equalization the non uniformity pore pressure set dur
of
of
in
-
ing shear the simplifying assumptions that the rate

on
This
is
based load
.
ing constant that the non uniformity pore pressure approximately

in
is
is
,
parabolic form over the height the sample and proportional

of
the stress
in
to
consoli
of
and that the coefficient the same for

is
increment consolidation
,
dation and swelling

.
DAY WEEK MONTH

1
1
I
SEALED ALUMINIUM
(D
)
FOIL RUBBER
&
MEMBRANE
TO
AIR PRESSURECHANGESDUE DIFFUSION

OF AIR THROUGHLATEX RUBBER
MEMBRANES LATEX RUBBER
(9
)
DROP
MEMBRANEONLY
SAMPLE HOLLOW CERAMIC CYLINDER
:
IN
STANDARDTRIAXIAL CELL
PERCENTAGE AIR PRESSURE
MEMBRANES RUBBERMEMBRANEONLY
töö
(0
)
BY
ALUMINIUMFOIL SEALED
(b
)
PVC TAPE SURROUNDEDBY

&
.
Ô RUBBERMEMBRANE
Ô
10
100 1000 000 100 000

,
TIME MINS
(
Fig
6
.
514
20° 0 - 0,
-3
U
..
T
S
P
PRINCIPAL STRESS RATIO

o
-
VIATOR STRESS
Brorogoto
OO
U
Bo
DE
10
15
20
5
%
STRAIN
.
..
MEASURED AT BASE
S OF SAMPLE
PRESSURE UP
AT
LU
MEASURED PROBE
.,
WATER
AT MID HEIGHT OF
SAMPLE
PORE
25
Stains
s
20
'
CHANGE
,
in
10
15
20
25
0
FOTO
VOLUME
Fig
on
Effect Pore Pressures Measured

at
" of
8 of
Rate Strain End and

7
.
--
.
High Sample Compacted Shale

of
of
of
Centre Dia Rate

4
x
a
"
.
.
20
Strain Hours
%
in
8
:
- 0
.IS.
M
-6
.P,su
-0 se
0
STRESS RATIO
W
STRESS
-
N
PRINCIPAL
DEVIATOR
Ō -
.
02
25°
20
.7
15
STRAIN
.ISP.
U
PRESSURE
PORE WATER
AT
lu
MEASURED BASE
à
OF SAMPLE
AT Uw
MEASURED
,
AT
MID
Apart
PROBE
T
HEIGHT OF SAMPLE
15
20
10
STRAIN
%
IO
CHANGE
10
15
IS
20 25
VOLUME
of
at
Fig
on
Effect Rate Strain Pore Pressures Measured End and

4 of
8
.
.
-
Centre High Sample Compacted Shale

of
of
of
Dia Rate
8
a
x
"
"
.
.
20
Strain Hours
%
120
in
:
OF PORE PRESSURE at 120 HOUR TEST
PROBE
PERCENTAGE EQUALIZATION
001 01 10 10
.
0
.1
.
Fig Relationship Between Percentage Equalization Undrained
in
9
(a
-
.
).
is
Test and Time Factor where Time Failure
to
=
T
.lSP.
IN
IN
CIRCLE PLOTTED TERMS CIRCLE PLOTTED
AT
OF
UW AT BASE TERMS OF MID
U
|
HEIGHT
SHEAR STRESS
23 59
.
23
.5
60
70
50
30
10
20 40 80
.S
P
.I
.
Fig Effect Unequalized Pore Pressures Apparent Position

on
of
of
9
(b
.)
.
Failure Envelope for Compacted Shale

20
Hours
of
Rate Strain
%
in
8
.
Using this curve

Fig
of an
results indicate instantaneous

the difference
in
7
.
'
'
of
of 20
the sample at failure

at
between the base and middle

=
p
0
s
t
(
.
.i
.
The degree equalization represented by the 120 hour test Fig calculated
8
.
(
from this value corrected for the difference deviator stress be

is
of in
to
seen
(
quite close the theoretical curve spite the approximations involved

to
in
,
Fig illustrates the magnitude the cohesion intercept that would

of
9
(b
.
have been inferred from the base measurement of pore water pressure using
(
This error probably

of
the observed value from other tests cause

is
O
a
).
'
the marked differences between many the published values the ef

of
is of
of
fective cohesion intercept This view supported by the decrease

, in
c
c
'.
'
testing found with compacted boulder clay Alpan

of
with increased time

(
1959 Bishop 1960

;
)
.
6. EXPERIMENTAL DETERMINATION OF THE PARAMETER X
The series of tests most easily interpreted from this point of view was
that on the Braehead silt, placed as a slurry and drained by suction , since the
material showed zero effective cohesion intercept c ' and was also relatively
permeable (K = 3 . 3 x 10 - 5 cm / sec . ) when saturated . The values of the cell
pressure oz and of the pore water pressure Uw and the pore air pressure ua
were controlled . The deviator stress at failure and the average degree of
saturation at failure were obtained from observations of load and volume
change respectively . The value of x could then be calculated from Equation 3
using the value of o ' from the tests with full saturation
The results are shown in Fig . 10 and are compared with a theoretical
curve calculated by Donald ( 1960 ) from an equation given by Aitchison ( 1960 )
for effective stress in a soil where the compression of the soil is small com
pared with the volume of water draining from it .
The satisfactory agreement will be noted . The rate of volume change at
failure is found to vary with degree of saturation even for the same stress
history ( Fig . 11) and this will influence the relevant value of 0 ' ( Bishop and
Eldin , 1953 ) . The corrected value of 0' has been used in the calculation of
the x values given in Fig . 10.
Values of x have also been calculated from tests on three compacted soils
having clay fractions (< 2u ) of 22 %, 18 % and 4 % respectively . Here the exis
tence of a small amount of true cohesion makes the uniqueness of the failure
envelope less easy to accept , and may also contribute to the scatter of the
results .
Both pore water and pore air pressures were measured throughout the
tests on partly saturated material . A typical undrained test on the compacted
clay shale with 22 % clay fraction is illustrated in Fig . 12 .
In Fig . 13 (a ) a method of plotting the results is given which avoids the con
fusion of a large number of superimposed Mohr circles and enables the value
of x to be read off directly . For the saturated soil one half of the deviator
stress , 1/ 2 (01 - 03 ), is plotted against the mean effective stress ,
1 / 2 (01 + 03 ) - Uy . The slope v of the line through these points is related to
the slope of the Mohr envelope by the expression tan = sin o ' , and the inter
y
cept do is equal to c
de
' . tan
tano ' .
The tests on partly saturated samples are plotted against both 1 / 2 (01 +
0z ) -uw and 1/ 2 (04 + 03 ) - un , and the value of x is obtained directly in
terms of the relative distance of these two failure lines from the effective
stress failure line for fully saturated samples ( Fig . 13 (a )).
The lowerlimit of the term 1 / 2 (01 + 03 ) - ua is reached when ua = 03
and the termequals 1 / 2 (01 - 03 ). For a series of tests having ua = 03 (un
confined tests at a range of water contents ), the distance d replaces c in the
expression for x ( Fig . 13 (b )).
The x values for the compacted soils are discussed in the following para
graphs .
( a ) Clay shale (22 % < 2M ). - In Fig . 14 the test results from a series of un
drained and constant water content tests are presented in terms of the plot
described in Fig . 13(a ) . As a first assumption the true effective stress en
velope has been fitted on the basis of c ' = 0 . This leads to the relationship
between x and the degree of saturation S plotted to a large scale in Fig . 15
and described as line ( 1) .
1 0
0 .8
X =S
FACTOR
THEORETICAL
TO
MEAN POROSITY PRIOR
SHEAR 45
%
=
.5
0
2
.
20
80
40 60 100
DEGREE OF SATURATION
%
5
-
BRAEHEAD SILT
4g
ug
OF
Oʻ
uw
IN
VALUES
X
-
o
-
(
+
-
x
INFERRED FROM DRAINED TRIAXIAL TESTS WITH

Uw
CONTROLLED NORMALLY CONSOLIDATED

,
U
)
E
Fig
10
Braehead Silt
-
.
.
.
its
o en
80s
MEAN POROSITY PRIOR
=
.
. %
5
TO SHEAR 45
CONTROLLING
Uw
=
.01
0
VARIOUS
Uw
-0 o
,-
NEGATIVE
FACTORS
°
0
20
BO Тоо
40
60
%
S
,U , SATURATION
BRAEHEAD SILT
OF
AT
DEGREE SATURATION FAILURE VERSUS RATE

OF CHANGE OF VOLUME TRIC STRAIN
AT WITH AXIAL
STRAIN FAILURE
–
NORMALLY CONSOLIDATED DRAINED

11
.
-.
Fig Braehead Silt

519
50 2. 5
.lS.P
- Loo
)
) (-
Uw
-(
03 (0 - 0 )
10 - UW)
RATIO
50
STRESS
STRESS
DEVIATOR
PRINCIPAL
10
% 15 20
STRAIN
10
15
20
25
I...S
P
-
Ka
PRESSURE
PORE
15
-
.
%
% STRAIN
20
10 15 20 25
CHANGE
LO
VOLUME
1
.5
-
Fig Typical Undrained Triaxial Test Compacted Shale Showing

12
on
a
-
.
.
Variations Pore Air and Water Pressures

in
į ( + 03) -U
(-
;)
< (OT+ 03) - Uw
60
0
do - C
TRUE EFFECTIVE STRESS LINE

TAN
JAN
SIN
a
';
ido
=
.=
O
'
*
*
c
d
Uw
,
,
+
0
+
(6
0
i
-u
)
o
§
(0
–
AT
TESTS CONSTANT WATER CONTENT VARIOUS CONFINING

&
a
)
PRESSURES
La
TRUE EFFECTIVE STRESS LINE
(-
')}
60
Uw
+
0
¿
(
)
d
-
b
-
X
d
o
Uw
10
+
į
)
–
AT
TESTS VARIOUS WATER CONTENTS

,
WITH
=
U
0
b
)
Calculating
of
13
Fig Curves Failure Stresses

of
Method from
x
–
.
.
522
(
STRESSES AT MAX
10
-T -
Uw
.0
00
:
TO
TIME FAILURE 120 HOURS
10
0
0
I..S
.
P
03
-
,
+
LOT
FAILURE
AT
SHEAR STRENGTH
0
,
.3-2 p5
-
%
-c
8
ASSUMED TRUE EFFECTIVE
2489
(
LINES SEE TESTS
X
STRESS
27
-.-
CONFERENCE
)
70 ON SATURATED SAMPLES
(
LINE
:. 2 :1
LINE
50
40
35
30
20
15
5
10
09 25 45
+
-
0g
O
E
+
-
FAILURE
0
Uw AT
at
a
22
%
I.SP. )
,U (
14
Compacted Water
Fig
-.
.
Triaxial Tests Compacted Shale Clay Fraction
%
at
on of
Content
.6
a18 and Sheared Constant Water Content
However , a detailed series of consolidated - undrained tests on 1 - 1/ 2 inch

diameter samples saturated by a back pressure ( Fig . 16 ) indicated a cohesion
intercept c ' of 2. 3 p . s . i. and a lower value of o ' equal to 23 . 5°. This envelope
lies slightly below the effective stress points obtained from the 4 inch diame
ter samples which had reached saturation under undrained compression . An
envelope through these points and having c ' = 2 . 3 leads to the x values denoted
"line 2' .
(b ) Boulder clay (18 % < 2u ) . - The results of a series of undrained and
constant water content tests are given in Fig . 17 and are compared with the
effective stress line based on the assumption that c ' = 0. This leads to the
relationship between x and S shown in Fig . 18 and denoted line 1.
Using an effective cohesion intercept of 1 .4 p. s .i . obtained from drained
tests on saturated remoulded clay samples from the same stratum , the re
lationship denoted line 2 is obtained .
( c ) Boulder clay (4 % < 2u ) . - The results of a set of ' unconfined ' com
pression tests (ua = 03 ) at various water contents are presented ( Fig . 19) in
terms of the plot illustrated in Fig . 13 (b) . The relevant effective stress line
is taken from the slope of the initial straight section of the relationship .
This leads to the x - S relationship given in Fig . 20 .
.95 L
CALCULATED FROM LINE 1

. O CALCULATED FROM LINE 2
.907
•85 H
X
FACTOR
•75
r
65
80
.
75 85 95 ОО
.
-
-
-
-
-
-
-
-
5
1
0
OF
DEGREE SATURATION
5
Between Degree
of
Saturation and Factor

15
Fig Relation for Tests

x
--
at .
.
22
Compacted Shale Clay Fraction

on
Constant Water Content

a
)
(
ODI
OEI
oli
011
001
00
..P
V
S
08
FAILURE
WATER CONTENTS
TESTS
.C. _0-
04AT
FAILURE ZONE
0
WN
TESTS
.C.
O
09
os
o
o
OVERALL WATER CONTENTS
TESTS
. . _
OE
NC
AT END OF TEST
_
..
TESTS
02
C
O
H
SO
22 21
%
CONTENT WATER
Fig Shale Compacted

16
on
Consolidated Undrained Tests

at
Wate
a
a
.
.
-
High Samples Jacked from Dia Samples Side and End Filter Drai
4
"
Principal
ere
ontent
Used
of
50
ress Ratio
.
. 18
.6
%
Time
to
JO NORMALLY CONSOLIDATED SAMPLES
.I . S.P FAILURE
Failure
A
TO
: SAMPLES OVER CONSOLIDATED

LINES ASSUMED FOR ESTIMATION
N .P
..C. .S
OF
MAXIMUM EFFECTIVE STRESS 100
X
IN
OF UNSATURATED
.
"
+
4
BY
50 DIA SAMPLES SATURATED UNDRAINED
-.
SAMPLES
.
COMPRESSION
and Saturated by Back

LINE
Hours
.
10
23
5S
. ..2
3
I
0Ć
==
P
T .
LINE
CONTROLLING
Pressure
.
10
20
30
70
90
40 50 00 80 100 120 130 140
oo
110
esS..P
OoJ.
FAILURE
,
und
fall
ir
,to
Stresses are
A
A
%
at
. 6
,
18
BY
CONSOLIDATED UNDRAINED TESTS ON SHALE COMPACTED AT WATER CONTENT OF SATURATED BACK PRESSURE TESTS
"5.
1
Tests on
X
"3
E
.
ON
:
DIA
.
HIGH SAMPLES JACKED

50
FROM 4°DIA SAMPLES SIDE END FILTER DRAINS WERE USED

TO
TIME FAILURE HOURS

FACTORS
, .
AT
.5 STRESSES ARE MAXIMUM PRINCIPAL STRESS RATIO
"
Maximum
Dia
.
x
3
"
525
60
526
%
15
STRESSES AT STRAIN
TO
TIME FAILURE 120 HOURS
I..
S
.
P
FAILURE
SHEAR STRENGTH
gi
-
+
AT 03 Uw
.O
.A.S.P
-
4.1=
01
C/
2703
CONFERENCE
ASSUMED TRUE EFFECTIVE
O
. 1°
STRESS LINES
:2
==
28
/' 17
O
LINE
LINE
N
30
20
60
40
70
50
80
110
90
100
AT
S
+
+
Uw
,0
1
P
;O
E
.-U
-0
...
FAILURE
02
17
.
on
Fig
(
Triaxial
18
Tests
)%
at
Compacted Boulder Clay Clay Fraction Compacted

%
aof
-.a
11
. 6
at
Water Content and Sheared Constant Water Content

The results of consolidated undrained and drained tests on samples of the

same size (4 inch diameter ) saturated by a back pressure are given in Fig . 21 .
It will be noted that they lie very close to a common failure line correspond
ing to c ' = 0 and O' = 34 .5º. This value of O' is lower than the value of 41 . 40
19
on
showing the best with the test results the partly saturated ig
fit
soil
),
.
(F
but the failure strains and mode of failure were markedly different The
.
shape plot may reflect the difference between the assumed and
of
the
x
s
-
actual different degrees

at
value
of
saturation
d
.
'
CONCLUSIONS
clear the test results presented this paper that

is
is
from
it
It
in
(a
)
necessary modified expression for effective stress when the degree

use
to
saturation drops appreciably below 100

of
%
.
This follows directly from the divergence the plot
03
01
of
of
1
/2
-
(
)
Uw
against for partly saturated soils from that for saturated

in 03
01
+
1
2
-
/
(
Figs interest of
14
17
soils illustrated and note that

is
to
both
in
It
,
.
.
cases these plots would extrapolated give negative cohesion intercept

if
,
for the partly saturated samples This also indicates that necessary
is
to
it
.
take account the air pressure term realistic failure envelope

of
to
obtain
a
- .
the more general expression for effective stress
of
The use
=
0
+ (b
o
,
x )
'
both pore water and pore air pressure
Uw
of
takes account
ua
The
),
u
.
-
(
.1
0
CALCULATED FROM
O
LINE
I
CALCULATED FROM
O
LINE
95
2
0
.
90
0
.
X
FACTOR
85
0
-
80
0
-
75
0
.
70
0
.
80
85
% 95
75 90 100
DEGREE OF SATURATION
5
for
Fig Relation Between Degree

18

) at
of
Tests
x
-
.
.
Compacted Boulder Clay Clay Fraction

18
on
Constant Water Content

%
a
(
528
ASSUMEDTRUEEFFECTIVE
0
)4".
(-
41
STRESSLINE
STRESSES FAILURE
AT
0
5
10
15
20
25
30
35
40 45
55
50 60 65
*
U
-.
.IS.P
SHEAR STRENGTH
12
%
OPTIMUMWATERCONTENT
VARIATIONOF INITIAL PORE PRESSURES

WITH COMPACTION WATERCONTENT
CONFERENCE
CONTENT WATERCOMPACTION
S
10
15
25
20
30 35 40
50
55
60
45
65
.I.S.
P
BEFORE SHEARING
19
03
ua
on
-.
(
a
=
)
(
Fig Unconfined Compression Tests Compacted Boulder Clay Clay Fraction
%
.
.4).
Tests at Various Water Contents Sheared with Controlled Pore Air Pressures
values of x obtained from tests on a soil drained by suction ( Fig . 10 ) and

tests on compacted soil ( summarized in Fig . 22 ) indicate
is its
from dependence
degree simple general relationship apparent
no
at
on
saturation though
of
,
this stage for compacted soils
.
obtaining reliable values pore water
of
The difficulties encountered
in
c
(
)
and air pressure and of point the care which must be used
in
and
to
',
d
,
'
drawing conclusions from laboratory tests Test results for example may
,
.
indicate cohesion intercepts considerable magnitude not only
of
insufficient
if
fine porous discs are used measure pore water pressure partly satu
ly
to
in
rated soil but also tests with base measurement pore pressure are run
of
if
,
too rapidly for pore pressure equalization occur
to
.
support the view that the cohesion inter
do
The tests results not however

,
of ,
cept general zero clearly indicated for the
is
value
is
in
p
A
2
3
c
s
.
.
.
i.
'
compacted clay shale when saturated other tests indicate
of
value about
18 ;
a
for the boulder clay with per cent clay fraction Only the boulder
.p
1
.4
s
.
.
.i
clay with per cent clay fraction appears have negligible cohesion inter
to
4
a
cept least the fully saturated state
at
in
,
Of more practical significance however is the fact that these values as

,
,
anticipated values currently used
by
Hilf 1956 are much lower than the
in
),
(
much design work

.
APPENDIX REFERENCES
.
-
Relationships
of
Aitchison 1960 moisture stress and effective

,
G
),
D
1
“
.
.
.
stress functions Proc

of
soils
on
unsaturated Conference Pore
in
”
.
Pressure and Suction Soils Butterworth London

in
.
-
X
FACTOR
Oo20 20 40 60 80 100
OF
DEGREE SATURATION
S
Fig Between Degree

20
Saturation and Factor for Unconfined

of
Relation
x
-
.
.
Compacted Boulder Clay Clay Fraction

on
Tests
%
4
(
)
60
530
TO
: 75
TIME FAILURE HOURS
JO
.1 DRAINED TESTS
.S
b
.P CONSOLIDATED UNDRAINED
TESTS AT MAXIMUM
PRINCIPAL STRESS RATIO
FAILURE
AT
=
.543
Oʻ
03
-
01
0
10
20
30
50 60 70 80 90 100 110
P
.IS.
AT FAILURE
on
21
of
(
.
-.
Fig Comparison Drained and Consolidated Undrained Triaxial Tests Boulder Clay Clay
%
%
by
of
)4
7
2
at
a
.
Fraction Compacted Water Content and Saturated Back Pressure

2. Alpan , I. ( 1959 ) , " A study of the principle of effective stress in partly

saturated soils " . Ph . D. Thesis , University of London .
3. Bishop , A . W . ( 1959 ), « The principle of effective stress ” . Teknisk Uke
blad , 39 , 859 - 863 .
4. Bishop , A . W . (1960 ) , " The measurement of pore pressure in the triaxial
test " . Proc . of Conference on Pore Pressure and Suction in Soils - Butter
worth , London .
5. Bishop , A . W . and Eldin , G . (1950 ) , “ Undrained triaxial tests on saturated
sands and their significance in the general theory of shear strength " .
Geotechnique , 2 , 13 - 32 .
6. Bishop , A . W . and Henkel , D . J . (1957 ) , « The measurement of soil proper

ties in the triaxial test ” . Edward Arnold , London .
7. Croney , D ., Coleman , J. D . and Black , W .
P . M . (1958 ) , « The movement
and distribution of water in soil in relation to highway design and per
formance ” . Highway Research Board , Sp , Report No . 40 . Washington ,
D . C.
Ő
X Ö
FACTOR
S
=
X
20
40 60 80 100
OF
DEGREE SATURATION
5
UNCONFINED TESTS ON COMPACTED

BOULDER CLAY CLAY FRACTION
4
(
)
AT
TESTS CONSTANT WATER CONTENT

ON COMPACTED BOULDER CLAY
18
%
CLAY FRACTION
(
)
AT
TESTS CONSTANT WATER CONTENT

ON COMPACTED SHALE
22
%
CLAY FRACTION
(
Fig Relations Between Degree

22

of
for
x
.
.
-
Three Compacted Soils

8. Croney, D . , Coleman , J. D . and Bridge , P . M . (1952 ), “ The suction of
moisture held in soil and other porous materials ” . Road Research
Technical Paper No . 24 . London (H . M . Stationery Office ).
9. Donald , I. B . (1960 ) , “ Discussion - Proceedings of Conference on pore
pressure and suction in soils " . Butterworth , London .
10 . Hilf , J .
W . (1956 ) , " An investigation of pore water pressure in compacted
cohesive soils ” . Bureau of Reclamation Tech . Mem . 654 .
11. Jennings , J . E . (1960 ), “ A revised effective stress law for use in the
prediction of the behaviour of unsaturated soils " . Proceedings of Con
ference on Pore Pressure and Suction in Soils . Butterworth , London .
12 . Laughton , A . S. (1955 ) , “ The compaction of ocean sediments ” . Ph . D.

Thesis , University of Cambridge .
13 . Rendulic , L . ( 1937 ), “ Ein Grundgesetz der Tonmechanik und sein experi
mentaler Beweis " . Bauingenieur , 18, 459 - 467 .
14 . Skempton , A . W . ( 1960 ) , " Inaugural Address - Proceedings of Conference
on pore pressure and suction in soil ” . Butterworth , London .
15 . Taylor , D . W . (1944 ) , « Tenth progress report on shear research to U . S.

Engineers " . M . I. T . Publication .
16. Terzaghi , K . (1923), “ Die Berechnung der Durchlässigkeitsziffer des

Tones aus dem Verlauf der hydrodynamischen Spannungserscheinungen ."
Sitz . Akad . Wissen . Wien Math - naturw . Kl. Abt. 11a , 132 , 125 - 138 .
17 . Terzaghi , K . ( 1932 ) , “ Trägfahigkeit der Flachgrundungen ". Int. Assoc .

Bridge Struct . Eng . Prelim . Publ . , 659 -683 .
THE SHEAR STRENGTH OF SATURATED REMOULDED CLAYS
By D . J. Henkel ?
INTRODUCTION
A fundamental appreciation of the shear strength behavior of saturated clays

requires a full knowledge of the way in which individual clay particles behave
at their contacts with each other and also the role played by water , absorbed
ions and electrolytes in affecting the forces between particles . However , for
many engineering applications it is sufficient to deal with the problem of shear
strength in a less fundamental way and the work of Hvorslev (1937 ) “ , who in
troduced the concepts of " true cohesion ” and “ true angle of internal friction , "
has helped to explain many of the shear strength phenomena associated with
both normally and overconsolidated clays . In practice even the Hyors lev ap
proach to the shear strength problem becomes very complicated and it is more
convenient to work in terms of shear strength parameters related directly to
the effective stresses in the clay .
A considerable amount of experimental work has been carried out on the
measurement of the properties of saturated clays . An important contribution
was made by the work of the Co - operative Triaxial Shear Research Program
me of the Corps of Engineers and the review of this work prepared by Rutledge
(1947 ) was a significant step in increasing our understanding of shear strength .
In later work on shear strength much emphasis has been placed on the mea
surement of the pore water pressure in undrained tests and at the present time
sufficient data has been accumulated to enable conclusions to be drawn about
the general behaviour of remoulded saturated clays under various conditions
of loading and drainage .
In most cases natural deposits of clay are far from homogeneous and the
variations that occur between individual test specimens make the task of ob
taining a satisfactory overall picture of their shear strength behaviour very
difficult . For this reason many investigators have used large batches of thor
oughly mixed remoulded clays in order to obtain sufficiently uniform samples
to ensure that valid comparisons between different stress histories and types
of test can be made . It is recognised that important differences , particularly
in the structure , exist between natural clay deposits and remoulded and re
consolidated samples of clay . Nevertheless it is considered that the detailed
study of remoulded clays provides data which enables many of the shear
strength phenomena of undisturbed clays to be understood .
Over the last ten years a comprehensive series of tests on remoulded clays
have been carried out at the Imperial College of Science and Technology in
London . As a result of this work many of the earlier ideas on the behaviour of
saturated clays have been confirmed . In addition it has been possible to obtain
a general picture of the behaviour of the clays under various combinations of
stresses and also to correlate the results of consolidated undrained tests ( Con
solidated Quick Tests ) and drained tests (Slow Tests ) (Henkel 1958 ) . Some of
1. Sr . Lecturer in Civ . Engrg ., Imperial College , London , England .

2. Items indicated thus , Hvorslev (1937 ) , refer to corresponding entries ar
ranged alphabetically in the Appendix .
533
these results have already been published (Henkel 1959) and an additional pa
per has already been submitted for publication . However , it is believed that a
short account of the conclusions reached in this work will be a useful contri
bution to the conference.
CLAYS TESTED
Two batches of remoulded clay have been tested and details of their Atter
berg Limits and moulding water contents are tabulated below :
Mounting Water
CLAY LL PL Content %
Weald Clay 18 34
London Clay 26 56
TESTS
Cylindrical samples of
clay 3 inches in length and 1 1/ 2 inches in diameter
were formed in in the majority of cases these samples were
brass moulds and
subsequently consolidated in triaxial cells under a hydrostatic pressure . The
hydrostatic consolidation pressures used varied between 15 and 120 lb . / in .
and these pressures were usually applied in a single increment . Samples pre
pared in this manner were considered to be essentially normally consolidated .
In order to produce overconsolidated samples , specimens already in equi
librium under a certain hydrostatic pressure were allowed to swell under a
lower hydrostatic pressure before the shear stage of the test was started .
Series of overconsolidated samples were produced having maximum consoli
dation pressures of 120 , 60 and 30 lb . / in . 2
After consolidation was complete either drained tests , in which the volume
change was measured , or undrained tests , in which the pore pressure was
measured , were carried out . Both compression and extension tests were per
formed using various types of loading . In the compression tests failure was
brought about either by increasing the axial stress , keeping the cell pressure
constant , or by reducing the cell pressure , keeping the axial stress constant .
In a few cases, however , the loading was arranged so that the sum of the three
principal stresses was maintained constant throughout the test . In the exten
sion tests , which were confined to the Weald clay samples , failure was pro
duced either by decreasing the axial stress , keeping the cell pressure con
stant , or by decreasing the cell pressure , the axial stress being kept constant .
TEST RESULTS
The analysis of the test data showed that the basic shear strength behaviour
could be adequately described by the three following conclusions :
(i ) There was a unique relationship between the strength (or maximum de
viator stress ) measured in the test and the water content of the specimen
at failure for all types of test on normally consolidated samples , provided
compression and extension tests were considered separately . A similar
result was obtained for overconsolidated specimens having a common
maximum consolidation pressure .
REMOULDED CLAYS 535
( ii ) Common effective stress Mohr failure envelopes were obtained for all
types of compression test on normally consolidated samples . Common ef
fective stress Mohr failure envelopes were also obtained for all types of
compression tests on overconsolidated samples having the same maximum
consolidation pressure . Similar results were obtained for the extension
tests . The envelopes for compression and extension tests were in general
slight ly different , but for normally consolidated samples they were practi
cally identical . It was also found that in all compression tests the ratio of
the principal effective stresses at failure depended only on the ratio of the
maximum consolidation pressure to the average effective pressure at fail
ure , and was independent of the value of the maximum consolidation pres
sure . A similar result was obtained for the extension tests .
At all stages unique
iii
( both drained
tests there was and undrained
in
a
)
relationship between the effective water content provid

stresses and the
,
normally consolidated samples and samples having common maximum
ed
consolidation pressures are considered separately
.
these conclusions are not basically new but involve extensions
of
of
Most
the first two are implied principle
of
ideas published the past Parts

in
in
in
.
the Working Hypothesis given by Casagrande the Third Harvard Report

in
A
.
the Cooperative Triaxial Shear Research Programme of the Corps En
of
in
gineers and summarized by Rutledge 1947 They were also implied by

(
)
.
Taylor the shear strength clays The ideas be

of
his treatment
of
1948
in
.
(
by
hind the third conclusion were developed Rendulic 1936 and 1937 based
)
(
very elegant series remoulded samples
of
of
on
of
tests
on
the results
a
Wiener Tegel Rendulic his tests used samples which had been subjected
in
.
increasing effective pressures only

so
during consolidation third con

to
the
,
Clusion given above slight extension

of
his work
is
a
The experimental evidence for the conclusions given above shown the
in
is
following figures
.
The Unique Relationships between Water Content and Strength

.
-
Figs give the results normally consolidated samples

on
tests
of
and
1
,
2
3
.
Weald and London Clay and also the results specimens having
of
on
tests
of
consolidation pressure certain amount

of
common maximum
lb
, in
120
A
2
a
.
/
remoulded clays but for practical

on
scatter inevitable even tests

is
of
in
will seen that the assumptions unique relationships between

be
of
purposes
it
the water content and the strength provides satisfactory working hypothesis
a
The Unique Effective Stress Failure Envelopes

.
The use Mohr envelopes give the relationship between the normal and
to
of
failure requires that separate envelopes should

be
shear stresses drawn

at
the
for each maximum consolidation pressure when the results over
on
of
tests
consolidated samples are be considered however the more general
to
If,
,
.
relationship between the ratio of the major the minor principal effective
to
stress and the ratio of the maximum consolidation pressure the average
to
all tests may be expressed

of
effective stress failure used the results

at
is
single diagram
in
a
Figs these general relationships are shown for the compres

In
and
4
5
6
,
.
Weald Clay and for the compression tests

on
on
sion and the extension tests

London Clay consolidation pressure and
Jm
used for the maximum for

Jf
is
.
26
NORMALLY CONSOLIDATED
UNDRAINED
DRAINED AXIAL STRESS INCREASED
24 OVERCONSOLIDATED MAX PRESSURE
120 LBIIN ?
O UNDRAINED
O DRAINED AXIAL STRESS INCREASED
007
WATER
CONTENT
18
16
60
70
20
30
40
50
15
8090 100 150 200

7
8
9
0
5
IN²
03
LB
-,
),
(6
Fig Weald Clay Compression Tests

1
-
.
will noted that each

of
at
be
the average effective stress failure the dia

in
It
.
grams the results the undrained tests and the various types
of
drained test
of
The actual points shown the figures were deriv

on
on
lies common curve

a
normally consolidated samples over wide range consoli

of
ed
on
from tests
a
dation pressure and from tests samples having maxi

on
overconsolidated
a
consolidation pressure The results from tests with other

of
lb
in
mum 120
.2
/
.
maximum consolidation pressures agree closely with the curves shown the
in
figures but the actual points have not been plotted

,
course only more general

ex
of
The results given Figs and are

in
a
5
6
4
.
pression
of
the fact that the effective stress Mohr envelopes for drained and
undrained tests are identical
.
The Unique Relationships between the Effective Stresses and the Water
.
Content
order demonstrate the unique relationships between water content and

to
In
the effective stresses follow Rendulic 1937 and plot the

is
to
convenient
it
the stress space appropriate the stress plane that can

be
results investi
to
in
the vertical stress

as
the triaxial cell

01
gated taken the limitations

is
in
If
.
the conventional triaxial cell require that the two horizontal stresses
On
of
REMOULDED CLAYS 537
UNDRAINED
• DRAINED AXIAL STRESS DECREASED
• DRAINED RADIAL STRESS INCREASED
OVERCONSOLIDATED MAX
PRESSURE 120 LBLIN ?
O UNDRAINED
O DRAINED AXIAL STRESS DECREASED
NO + DRAINED RADIAL STRESS INCREASED
A DRAINED J CONSTANT
WATER
CONTENT lo
20
? 6 7 8 9 10 15 20 30 40 50 60 70 80 90 100 150
(6, - 03) , /
LB IN ?
Fig . 2 - Weald Clay Extension Tests
and oz must be equal. The stress plane that satisfies these conditions of an
axially symmetrical stress system is shown in Fig . 7, and it is convenient to
plot the results in this stress plane .
In order to avoid confusion about major and minor principal stresses the
symbol oa ' has been used for the axial effective stress , while or ' has been
used for the radial effective stress . The geometry of the stress plane re
quires that values of the radial stress plotted along the horizontal axis should
be multiplied by V2 .
The plane is shown in Fig . 8. The line representing a hydrostatic state of
stress rises at a slope of 1 / v2 from the origin and the points representing
the stresses in the specimens after consolidation and before the commence
ment of the shear test will all lie on this hydrostatic pressure line . As shear
ing forces are applied to the specimens the points representing the effective
stresses will move away from the hydrostatic pressure line and the path
traced out in the stress space by these points is called a stress path . The
stress paths for compression tests appear above the hydrostatic pressure
line , while the stress paths followed in the extension tests below the hy
lie
drostatic pressure line

.
On Fig the stress paths followed compression and exten

of
series
in
on 8
a
.
sion tests samples Clay consolidation pressure

of
Weald with maximum

,
a
38
I UNDRAINED
• DRAINED AXIAL STRESS INCREASED
DRAINED RADIAL STRESS DECREASED
OVERCONSOLIDATED MAX PRESSURE
120 LB / IN ?
UNDRAINED
O DRAINED AXIAL STRESS INCREASED
+ DRAINED RADIAL STRESS DECREASED
WATER
CONTENT
. e
.
.
.
4 5 6 7 8 9 10 15 20 30 40 50 60 70 80 90 100 150 200
(9 - 0 ), /
LB IN ?
Fig . 3 - London Clay Compression Tests
of 120 lb . / in . , have been shown . The magnitude of the hydrostatic pressure

at the starting point of each stress path is given directly by the vertical ordi
nate , while the appropriate overconsolidation ratio is also shown .
In the drained tests in which the excess pore water pressure is zero the
total stresses applied to the sample are also the effective stresses and the
stress path followed in the test can thus be controlled . In Fig . 8 the stress
paths followed in the various types of drained tests are the straight dashed
lines . In undrained tests , however , the stress path is predetermined by the
pore water pressure characteristics of the specimen . In spite of variations
in the applied total stresses , for a specimen with a particular consolidation
history only one effective stress path can be obtained . These are shown in
the diagram by the full curved lines .
The stress paths have been continued until maximum deviator stress is
reached and the final points has already
lie
as
the failure envelopes which

on
been shown are common for both the undrained tests and the various types
of
,
drained tests
.
5
DRAINED CONSTANT
DRAINED AXIAL STRESS DECREASED
J- – –
DRAINED RADIAL STRESS INCREASED
A•+O
UNDRAINED
off
30
r'o
REMOULDED
.
2
5
CLAYS
C.
N
.2
0
.
1
5
6
7
8
9
1
0
10
•5
.
.
..
.
4
6
7
8
9
15
20
30
50
60
5 um 910 40
-4
.
Fig Weald Clay Compression Tests
539
540
3
5
.
DRAINED CONSTANT
J- - –
A • +O
UNDRAINED
0
3
.
SHEAR STRENGTH
.
2
5
CONFERENCE
.
.C
.
2
0
50
40
30
20
10
15
7
9
6
8
3
4
60
5
2
. 1
.8.
7
.6.
5
910
.
Fig
-5
Weald Clay Extension Tests
--
•+ O
UNDRAINED
3
0
-
.25
REMOULDED
.02
NICI
NO
CLAYS
5
..5
.1°
6
7
...
8
9 10
5
2
3
4
5
6
7
8
9
10
15
20
30
50
40
. -6
Fig London Clay Compression Tests
541
STRESS PLANE THAT CAN BE

INVESTIGATED IN STANDARD
TRIAXIAL APPARATUS
W104
02
03
-
Fig
7
.
by
by
Thestress paths followed the specimens undrained tests are def

in
of
inition contours constant water content drained tests however changes

In
,
.
volume and water content occur during the test From knowledge
in
in of
the
a
.
final water content and the variation sample volume with the change
in
stress drained test the water content any point the stress path may
on
at
in
a
series points
of
of
be calculated From set drained tests which the

at
a
a
.
water content the same can be plotted the stress plane and these points
in
is
if
are joined contours

of
constant water content for drained tests may be

ob
,
tained
.
Fig the stress paths followed undrained compression and extension

In
in
9
.
samples Weald Clay normally consolidated under

30
60
on
of
tests and 120

,
have been plotted

of
addition contours constant water content have

In
lb
in
,
.2
.
./
been constructed using the results of drained compression and extension tests
normally consolidated samples will be noted that the undrained stress
on
It
.
paths very well with the contours obtained from the drained tests and
fit
in
the unique relationship between the effective stresses and the water content
for normally consolidated samples Weald Clay satisfactorily demon
of
is
strated
.
REMOULDED CLAYS 543
280
- - - · - - - • DRAINED - AXIAL STRESS INCREASED
OR DECREASED
- - - - - + DRAINED - RADIAL STRESS INCREASED
OR DECREASED
| -- , - A DRAINED - J CONSTANT
240
- O UNDRAINED
COMPRESSION
-
200
ENVELOPE
160
FAILURE
oo
LBIIN
?
120 YANC
-
--
-
80
EXTENSION
.
1
/
O
.R
.
.C
-
l'
ENVELOPE
FAILURE
word
ol 40 80 120 100 200
,
IN2
LB
vz
/
O
'.
.
Fig Weald Clay Stress Paths

8
-
.
Fig samples
of
of
10
on
shows the results obtained from series tests

a
.
Weald Clay having common maximum consolidation pressure

of
lb
in
120
a
/
.
The shape the contours of constant water content are very different from
of
normally consolidated samples but

of
on
those obtained from the results tests

,
here again very satisfactory agreement between the results

of
the undrained
tests and the drained tests obtained
is
Figs give respectively the contours

11
12
constant water content

of
and
.
obtained from drained compression tests normally consolidated samples

on
220
- - - - - WATER CONTENTCONTOURS
FROM DRAINEDTESTS
129
-
-
_
200
IN
STRESSPATHS UNDRAINED
TESTS
17
5
-
-
.
180
--
-
--
--
%
18
-
160
-
-
%
18.
3
140
-
COMPRESSION
%
18.
5
120 94
i
-
or
-
FAILURE
oo
i
%
19.
80 5
20 592
77 .
% 20
--
L
.
-
-
-
21
.5 -
-
-
-
21
%
_
r
222
224
-
-
54
22
P
)
:
23 EXTENSION
.
SAILURE
Oí
v2
lb
in
?
/
Fig Weald Clay Normally Consolidated

9
-
.
Clay and samples having pressure

on
of
of
London maximum consolidation

a
The undrained stress paths both the figures well

fit
120 with the

lb
in
in
in
of .?
.
/
contours constant water content obtained from drained tests

.
APPLICATION OF THE UNIQUE RELATIONSHIPS
Figs fact combine all the three basic conclusions reached from
12
in
of 9
.
study the test data The common failure envelopes for all tests and the
a
unique relationships between the effective stresses and the water content
at
all points the stress plane mean that once these relationships have been
in
any type triaxial test may be predicted Either

of
of
established the result

.
undrained tests may be used establish these basic relationships

or
drained
to
any drained test required the direction

of
the result
of
the proposed
is
If
stress path the diagrams starting the initial consolidation con

on
drawn
at
is
The drained strength will be given by the vertical distance

of
ditions the
.
REMOULDED CLAYS 545
220
WATERCONTENTCONTOURS
200 FROM DRAINEDTESTS
STRESSPATHSIN UNDRAINED
TESTS
180
17
. 5%
.
COMPRESSION
160
182
140
FAILURE
18
%
-3
120 18
%
.5
,
18
16
74.
in
100
/
?
.
18
%
.9
19%
.
80
%
119
.2
%
119
.5
202
EXTENSION
FAILURE
19.
7
LLLLLLLL
.
-
L212
%%
21-
2
20
40
60
80
100 120 140 160 180

0
200
of
12
16
in
?
/
Fig Clay Maximum

10
Weald Consolidation Pressure in2

lb
120
/
-
.
.
point intersection the stress path and failure envelope from the line rep
of
of
resenting the hydrostatic state

at
failure may
of
stress The water content

.
be interpolated between the points where the constant water content lines cut
the failure envelope while the change water content during the test may be
in
,
obtained from the points which the stress paths cut the contours
of
constant
at
water content
.
undrained test may be obtained by following the contour

of
of
an
The result
constant water content which leaves the hydrostatic pressure line
at
the point
by
representing the consolidation conditions The strength will

be
given the
.
vertical distance the point the particular constant water

of
of
of
intersection
content contour and the failure envelope from the hydrostatic pressure line
.
The pore water pressure any point the test may be obtained compar
by
at
in
ing the effective stress path the constant water content line with the total
,
is .e
.
i
This procedure best demonstrated by reference Fig

13
stress path
to
.
Fig the sample hydrostatic

13
considered be consolidated under

is
to
In
a
.
pressure corresponding point the hydrostatic pressure line

on
The
to
'o
.
'
- - WATERCONTENTCONTOURS 257
FROMDRAINEDTESTS
STRESSPATHSIN UNDRAINED
TESTS
- - - 26
COMPRESSION
27%
FAILURE
%
28
%
27
.5
29
%
1b
-
-
llos
-
in
?
/
30
%
C
-
-
-
-
%
29
3
.
31%
.
322
E
31.
6
-
34 -
-
-
I
%
L
-
-
-
35
8
.
2000
OKT
in² 20
60
do
ok
do
!
or o
o
1b
/
z
Fig London Clay Normally Consolidated

11
-
.
contour constant water content through obg and this also the ef
is
of
is
If o
'
'
'
'
fective stress path the undrained test failure be brought about by

in
is
to
.
increasing the axial stress keeping the radial stress constant the total stress
,
path will vertical line through The total stress path will termin
be
on
at a
'o
'.
failure the point which the compression strength terms total

at
at
of
ate
in
stresses equal the strength represented by the point By drawing

to
is
'g
a
.'
line through parallel the hydrostatic pressure line cut the total stress
to
to
'g
'
the point giving the total stresses failure will be obtained

at
path The
,
.
'
at 'f
pore pressure failure Ue the difference between the total stresses

is
at
,
and the effective stresses and given by the vertical distance be

is
at
'g
'
'f
'
tween the points and

'g
'.
'
'f
intermediate points during the test the pore water pressure may also
At
be obtained by comparing points the total and effective stress paths for
on
which the shear stress the same These points are equidistant from the
is
hydrostatic pressure line obtain the pore water pressure point

To
at
a
a
.
'
'
REMOULDED CLAYS 547
220
- - WATER CONTENTCONTOURS
FROMDRAINED TESTS
200
- STRESSPATHSIN UNDRAINED
TESTS
26%
.
180
160
%
27
LIIII
COMPRESSION
120
28°7
%
20
%.
5
.
-
-
_
lin
16
FAILURE
?
IOO
- 0
BO
0
|
%
29
%
130
%
131
132%
%
30
.
20
do
/ do
zdo
do
do
Oí do
20 60
ho
bo
bo
16
12
in ?
Fig London Clay Maximum

12
Consolidation Pressure 120

lb
in
?
-
.
the total stress path line parallel the hydrostatic pressure line
on
to
is
,
a
drawn through cut the effective stress path The vertical distance
to
at
'b
'a
'
'.
the diagram will then give the pore water

on
between and marked Ua

',
,
b
'a
'
'
pressure the point the total stress path

on
at
'a
.
'
ANISOTROPIC CONSOLIDATION
The tests used define the relationships between water content and the
to
ef
samples consolidated
on
fective stresses given above have been carried out

under hydrostatic pressure However few undrained tests have been per
a
a
,
.
samples Weald Clay consolidated under anisotropic stress

' of
on
an
formed
. ef
system with was found that when the water content and the
or
'>
It
o
,
process were plotted Fig

on
fective stresses
at
of
the end the consolidation

8
200
160
- 1
120 WATER CONTENT
CONTOUR
:'
od
LB / IN FAILURE
ENVELOPE
80
0
0=
40
of
go 80 120 160 200
O: 12 LBIIN ?
.Fig
13
they fitted with the water content contours addition during the undrain
in
In
,
.
stress path followed was parallel

ed
test the the water content contours

to
,
and failure the water content strength and effective stress relationships
at
agreed with those obtained from the tests the samples consolidated under
on
hydrostatic pressure
a
APPLICABILITY TO UNDISTURBED CLAY SAMPLES

rig
as
As far the author aware insufficient test data available for

is
is
,
samples
of
orous examination
of
undisturbed clay
of
the behaviour However

,
is .
some natural deposits have been published and

on
limited data useful

to
it
on
see how far this information conforms with the results obtained remoulded
clays
.
the Triaxial Shear Report 1947 results have been given for immediate
In
on
undrained tests consolidated undrained Qc tests and drained tests

(Q
N S
)
(
. )
undisturbed soft clays from the Chicago Subway and from Massena
Y
.
These tests show quite clearly that for practical purposes unique relation
at a
ship between the compression strength and water content failure obtained
is
of
independent the type test used the investigation

of
in
.
REMOULDED CLAYS 549
The evidence for the uniqueness of the effective stress failure envelopes
for both drained and undrained tests on undisturbed soils is rather conflicting .
Data given by Casagrande ( 1953 ) , Casagrande and Wilson ( 1953) and Hirschfeld
(1959) suggests that for undisturbed clays the undrained tests give rather
higher effective stress failure envelopes than the drained tests . A feature of
the experimental data on which these findings are based are the comparative
ly short loading times used in the undrained tests and the very long loading
times used in the drained tests .
If , however , valid comparisons between the results of drained and undrain
ed tests are to be made , the measured strength values must be corrected to
the same rate of strain or time of loading . In addition , when end measure
ment of pore pressure is used , the rate of strain in undrained tests must be
sufficiently slow to ensure that the pore water pressure measured at the end
of a specimen represents the correct pore water pressure in the failure zone .
In the author ' s experience, localized thin failure zones often occur in un
disturbed clay samples , while in remoulded samples , even when overconsoli
the failure zone is distributed over a much larger volume of the sam
dated ,
ple . In these circumstances it has been found that the satisfactory measure
ment of pore
water pressure in undisturbed samples requires a very much
lower rate of testing than is required on remoulded samples of the same clay .
Recent experience gained on the measurement of pore water pressures in
consolidated undrained tests on undistrubed London Clay specimens has shown
the importance of this effect . A series of these tests was carried out on sam
ples consolidated under a hydrostatic pressure of 30 lb. / in . In spite of the
heavy overconsolidation of the clay in - situ the pore water pressure showed a
continuous increase until just before failure was reached .
In these tests failure took place in a very thin zone and when the failure
points for tests lasting only a few hours were plotted , they lay well above the
effective stress envelopes obtained from drained tests which had lasted be
tween 3 and 5 days .
tests lasting about 3 days were then carried out and it was found
Undrained
that the measured pore water pressures at failure were lower than those in
the more rapid undrained tests , and that failure points now lay very close to
the envelope obtained from the drained tests .
Close agreement between the effective stress envelopes obtained from
drained and consolidated undrained tests , carried out at similar strain rates ,
has also been found for an undisturbed alluvial silty clay (LL =40 , PL =20 ) from
Chew Stoke .
would appear therefore that before final conclusions can be drawn about
It
the effective stress envelopes obtained from drained and undrained tests ,
further experimental evidence on the effect of rate of testing on both the mea
sured pore water pressure and the effective stress shear parameters , is re
quired . Bjerrum , Simons and Torblaa (1958 ) , for example , have reported that
for a normally consolidated marine clay the effective stress envelope obtain
ed from undrained tests , carried out at slow rates of deformation , lies some
what below the effective stress envelope measured in drained tests at the
same rate . Existing evidence does , however , suggest that for some undis
turbed clays almost identical shear stress parameters are obtained from both
drained and undrained tests , provided comparable rates of testing are used
and due allowance is made for the effects of overconsolidation .
No published data is available for sufficiently uniform samples of undis
turbed clay to enable the concept of a unique relationship between water con
tent and the effective stresses to be examined . However , the fact that in some
uniform undisturbed clays the strength is a unique function of the water con
tent and the effective stress envelopes for drained and undrained tests are
very similar , does suggest that the unique relationships between water content
and effective stress may well apply for other combinations of stresses .
The manner in which highly sensitive clays , in which structure plays a very
important part, behaye is likely to be rather more complicated than that sug
gested by the simple relationships which apply to remoulded clays . Complex
behaviour may also be expected in clays for which the secondary compression
is large .
It is , however , considered that the results obtained from tests on remould
ed clays will provide a valuable guide to the behaviour of those natural clays
in which structure and secondary compression do not play an important role .
APPLICATION TO FIELD PROBLEMS
The fact that in many field problems plane strain conditions are imposed
on the clay while in the conventional triaxial cell the behaviour of clays can
be examined only under axially symmetrical stress conditions , must be taken
into account when transferring laboratory results to the field .
Some guidance on the way differing stress systems affect the results of
shear tests on clay may be obtained by a study of the differences between com
pression and extension tests . Reference to figs . 4 and 5 show that for normal
ly consolidated samples (the points marked N , C in the figures ) the principal
effective stress ratios at failure are approximately the same in compression
and extension . In undrained tests , however , the strengths obtained in exten
sion tests are some 14 % lower than those obtained in compression tests and
this lower strength is due to the higher pore water pressures set up in the
extension tests .
In undrained tests part of the change in pore water pressure is due to the
change in the average total stress applied to the sample , while the other part
is related to the shear deformation associated with the shear stress . In an
undrained compression test in which the axial total stress is increased the
increase in the average total stress on the sample will be equal to one third
of the increase in axial stress . In an undrained extension test in which the
radial stress is increased , the average increase in total stress is equal to
two thirds of the increase in radial stress . If a valid comparison between
these two types of test is to be made , the difference in the changes in aver
age total stress must be taken into account .
Skempton (1954) has introduced the concept of pore pressure parameters
and for saturated soils he has suggested that the pore pressure change ac
companying changes in total stress may be expressed as :
Au = 40% + A 001 - 40z )
in which 9 and 03 are respectively the major and minor principal total stres
es , and Au is the change in pore water pressure .
This expression does not take into account the value of the intermediate
principal stress . If it is applied to failure conditions in undrained tests on
REMOULDED CLAYS 551
normally consolidated samples of Weald Clay , the value of Af in compression
tests is found to be 0 . 94 , while in extension tests Ap is found to have a value
of 1 . 28 .
For a fundamental appreciation of the problem it is better to express the
pore pressure change as functions of the change in the average total stress
and the change in shear stress which is independent of changes in average to
tal stress . For the shear stress term it is appropriate to use a stress invari
ant (Henkel 1958 ) and a possible expression for the change in pore water pres
sure for saturated soils is of the form :
103
soz
.
104
au
+
+
20
A0
oz
,
602
,
603 903
12
+
/
a
-
)2
+
(
-
)2
(
+
-
are the changes
02
03
where A01 and the total stresses and

in
is
,
'a
a
'
pore pressure parameter
*
.
triaxial ex
02
01
02
test where
or
the conventional A03
In
the
=
=
,
,
A
A
4
pression for the change pore water pressure reduces
to
in
:
Ao
so
A0
,
,
+
*+
Au
rz
403 404
1
10z
.
a
-
-
cell pressure
)
For compression test where kept constant
is
the the
The pore pressure
01
A02 A03 and 40 03 change may there

=
=
,
0
.
(
fore be written
:
Au
oz
=
,
(
2
/
)
0
(
For the extension test where the cell pressure increased and the axial
is
The pore pres

03
0g
kept constant
01
01
stress and A03

is
=
be A
0
-
.
(
sure change may thus written

:
Au
35
0z
+
,
-
(
a
2
/
)
0
(
-
)
of
The parameter measures the contribution the shear stresses

to
the
'a
'
pore water pressure change and extension and compression tests are be
to
if
obtained from each test should be compared These

of
correlated values av
2
,
values are obtained by subtracting from Skempton value one third the
in
A
's
case of compression tests and two thirds

of
the case extension tests

in
done for failure conditions normally con

on
this
is
the undrained tests

in
If
solidated samples Weald Clay values

of
of
61 are obtained for both

av
of
,
0
.
the compression and extension tests This agreement for normally consoli
.
dated samples may well be fortuitous

as
found that for overconsolidated

is
it
,
samples the values compression tests are all somewhat higher

of
av
in
2
than those obtained from the extension tests However the difference the
in
,
.
test are smaller than the corresponding

of
values the two types

of
av from
2
difference the values

in
the important applications

of
of
One this more fundamental definition

of
pore pressure parameters that enables the pore pressure phenomena as

is
it
sociated with plane strain be considered logical manner

in
to
Reiner 1949 has used similar expression take into account the volume
to
a
*
to )
changes due dilatancy during shear

.
No information is as yet available for the value of the intermediate princi
pal stress in plane strain tests on saturated normally consolidated clays .
However , tests carried out by Wood ( 1958 ) on compacted clays showed that the
intermediate principal stress in both total and effective stress terms was
somewhat less than the mean of the major and minor principal stresses . In
these circumstances it would not be unreasonable to assume that , for normal
ly consolidated clay , the intermediate principal stress would be slightly be
low the mean of the major and minor principal stresses . For the purposes
of demonstration the intermediate principal stress has been assumed to be
0. 9 times the mean value of the major and minor principal stresses .
With this assumption the stress changes in a plane strain test starting from
a hydrostatic state of stress , would be :
),
and
- ( 04 - -
45
son 03 603 0 soz 104
-
.
0
as
The change pore water pressure may expressed
be
:
in
107
45001
Au
. 0
+
-
45
40
90
110oz 4540
)2;
a
.
?
-0
(0
)2
+
+
48
av
01
)03
505
+
•
(
.0
. 1
)
(
-
failure obtained from compression and extension tests

of
at
the value
If
'a
'
normally consolidated samples plane

be
on
Weald Clay apply

to
to
of
assumed
strain then the pore pressure change failure would be
at
:
,
48
oz
Aue
01
oz
53
,
+
-
0
,
(
0
.
0
)
(
)
-
1
0
.
(
the pore water pressure change known the strength may readily be
is
If
is ,
the angle shearing resistance The available evi

of
calculated known
if
dence suggests that clays obey the Mohr Coulomb failure criterion and that
-
shearing resistance only very slightly higher

of
plane strain the angle

is
in
axial symmetry
of
than tests carried out under conditions

in
From the triaxial tests Weald Clay the angle of shearing resistance was
on
be
plane strain
is
found this value used found that the

to
220 and
in
if
is
it
compression strength ordinary tri

an
likely
be
to
some lower than

is
in
4
axial compression test with the same consolidation pressure

.
of
The treatment the plane strain problem given above course rather
is
of
tentative and experimental work required

of
obtain the exact value the

to
is
principal stress was given however show the importance

to
intermediate
It
,
,
.
expressing ideas
of
the most fundamental form better understanding

of
of if
in
the various factors controlling the shear strength clays be obtained

to
is
nature however the clay would general be consolidated under aniso

an
in
In
,
,
tropic stress system and the treatment the stress changes would involve
of
taking the different initial stresses into account

.
REMOULDED CLAYS 553
CONCLUSIONS
The results of a series of tests on saturated remoulded clays have been

presented and it has been shown that a comprehensive picture of the behaviour
of these clays can be obtained in terms of simple relationships revealed by
the test results .
It is considered that in some circumstances in uniform clays of low sen
sitivity these fundamental relationships can also apply to undisturbed soils .
One of the difficulties of applying laboratory results to field problems is
the fact that in many practical cases conditions of plane strain are encounter
ed , whereas laboratory tests are normally carried out under axially symme
trical stress conditions . By the use of a pore pressure parameter , defined in
fundamental terms, a tentative approach to the problem of the undrained com
resssion strength in the standard test and in plane strain has been suggested .
The unique relationships indicated by the results of the tests on the re
moulded clays suffer in fact from a major defect from a fundamental point of
view . While these relationships provide a satisfactory framework in which
the effective stresses , the shear stresses and the volume changes or water
contents can be correlated , it has not yet been possible to bring the shear
strains into the general picture .
In many of the problems encountered in soil mechanics deformations play
a very important part and until the relationships between the shear strains and
the stress components are better appreciated , no basic understanding of the
validity of many design methods will be obtained .
Bjerrum , L . , Simons , N . and Torblaa , I. (1958) “ The effect of time on the

shear strength of a soft marine clay ” Proc . Brussels Conf . on
Earth Pressure , 1: 148 - 158 .
Casagrande , A . ( 1953 ) “ Prestress induced in consolidated quick triaxial tests "

Proc . 3rd Int. Conf . Soil Mech . 1: 106 - 110 .
Casagrande , A . and Wilson , S. D . ( 1953) " Effects of stress history on the
strength of clays ” Harvard Soil Mechanics Series No . 43 .
Henkel , D. J. (1958 ) “ The correlation between deformation , pore water pres

sure and volume change characteristics of saturated clays "
Ph . D . Thesis , University of London .
Henkel , D. J. (1959 ) “ The relationships between the strength , pore water pres
sure and volume change characteristics of saturated clays ”
Geotechnique , 9 : 119 - 135 .
Hirschfeld , R . C . (1959) “ The relationship between shear strength and effec

tive stress " Proc . 2 st Pan - American Conf. Soil Mech .
Hvorslev , M . J (1937) « Über die festigkeitseigenschaften gestörter bindiger

Būden " Ingenioridenskabe lige Skrifter A , No . 45 . Copenhagen .
Reiner , M . ( 1949 ) Proc . Conf. Inst. Rheology , Amsterdam . Section IV , P . 55 .

Rendulic , L . (1937)
: der
« Ein Grundgesetz tonmechanik und sein experimental

Bauingenieur
18
ler beweis 459 467

,
”
-
Skempton , A . W . (1954 ) " The pore - pressure coefficients A and B”
Géotechnique , 4 : 143 - 147 .
Taylor , D. W. (1948 ) " Fundamentals of Soil Mechanics ” John Wiley and Sons ,
New York .
Waterways Experiment Station Vicksburg (1947) “ Triaxial shear research

and pressure distribution studies on soils "
Wood , C. C . (1958 ) “ Shear Strength and volume change characteristics of

compacted soil under conditions of Plane strain " Ph . D . Thesis ,
University of London .
A MECHANISTIC PICTURE OF SHEAR STRENGTH IN CLAY
By T . William Lambe ,' F . ASCE
I. INTRODUCTION
The soil engineer of today need hardly be reminded of the great importance
of shear strength knowledge . Among the most frequent and most serious prob
lems which he faces are usually those of stability , i . e ., those requiring know
ledge of the shear strength of soil. The se lection of the proper value of shear
strength to be employed in a stability problem is usually the most difficult
part of the problem , especially when the soils involved contain small particles .
Because of the great importance of shear strength knowledge to the practice
of soil engineering , coupled with unfortunate complexities of the principles
of the shear strength of fine grained soils , the Organizing Committee has in
deed done a major service to the profession by calling this Conference .
The Massachusetts Institute of Technology has played a long and active role
in research on the shear strength of soils . During the late 1920 ' s and early
1930 ' s , Terzaghi , Gilboy , Casagrande , and Jurgenson , working at M . I. T. , made
important contributions to our knowledge . The late 1930 ' s and the early
1940 ' s were the years in which the late D . W . Taylor was doing his pioneer
work at M . I . T. on the shear strength of clay . His shear tests in which pore
water pressures were measured are now widely known and recognized .
During recent years the M . I . T . research has been concentrated in the fol
lowing three areas :
1. The improvement of the shear strength of soils , i . e., soil stabilization .

2. The development of a mechanistic picture of the nature of shear
strength in soils , especially in cohesive soils .
3. A study of the strength of soils subjected to rapid rates of loading .
Research on soil stabilization was undertaken at the request of the Corps
of Engineers , U . S. Army since military operations during World War II had
emphasized the great need for techniques to give weak soils the ability to
support whee led loads . Soil stabilization research at M . I . T. has been under
way continuously since its start in 1946 .
While military considerations were the sole reason for initiating the re
search , and have remained a positive factor in the continuation of the re
search , two other factors have become important reasons for the continu
ation , and in fact , expansion of the soil stabilization research . First , as will
be discussed later in this paper , soil stabilization research can give insight
into soil behavior , especially shear strength . Second , is the belief that more
and more in the future the soil engineer will be required to tailor , to some
extent at any rate , the properties of the soils involved in a particular problem
to meet the demands of the problem at hand . Certainly appealing to the im
aginative soil engineer is the approach of : making soil serve the desired en
gineering project in a cooperative manner , rather than letting the soil behave
1. Prof . and Head , Soil Engrg ., Massachusetts Inst. of Tech ., Cambridge ,

Mass .
555
in a contemptuous fashion and determine the nature of the above - ground pro
ject and , in fact , sometimes even the feasibility of the project in any form .
Early in the research on soil stabilization the test results obtained made
it quite clear that substantial progress depended on a better understanding of
the mechanism of strength generation in soils , especially in fine grained soils .
This fact becomes especially true when one considers techniques of soil sta
bilization which effect a relatively modest improvement in soil strength , as
opposed to a solidification . In the nonsolidification type of stabilization , the
soil must be considered as a participant in the reaction , not an inert filler .
How can effective and economical means of increasing the strength of soils be
deve loped in the nature of shear strength generation in untreated soils is not
kurown ? Another fact that emphasized the need for a consideration of soil as
a partner is the great dependence of stabilizer effectiveness on compaction
and cure conditions . The M . I. T . Research was thus forced into a considera
tion of the physical and chemical nature of soils , the arrangement of, and the
forces between , soil particles and the influence of these factors on the
strength of soils .
As was the case with the soil stabilization research , the dynamic research
was initiated to furnish knowledge for a particular practical problem , i. e. ,
the design of structures to resist dynamic loads . The initial objectives of this
dynamic work were obviously military in nature . As the work continued , it
became apparent that it was of more than practical value in a limited area .
The dynamic shear studies have thrown light on " time effects , " one of the
more important and less understood areas of shear strength . As a technique
for obtaining fundamental information on the shear behavior of soils , dynamic
tests have the great advantage of permitting considerable data to be gathered
in a short period of time . This expedites testing and prevents , or at least
minimizes , the effects of physical and chemical changes of the soil during
testing .
At the present time there is a concerted effort being directed at M . I. T . to
ward elucidating the nature of shear resistance mobilization in fine grained
soils . The effects of environmental factors - nature of pore fluid , exchange
able ion complex , temperature , time, etc . - on shear strength are being stud
ied . While these factors are being considered at a submicroscopic scale ,
their effects are being determined by standard engineering tests , especially
undrained triaxial tests with measured pore water pressures . It is hoped that
this experimental and theoretical research will identify the important factors
influencing shear strength and indicate how important each one is . The fol
lowing paper presents some of the recent theoretical work which is guiding
the experimental investigation .
The M . I . T . Soils Group feels that the effective stress principle is the most
powerful tool available to the engineer who must work with the shear strength
of soils . It has been pointed out (Hvorslev , 1938 , and others ) however , that
shear strength is not uniquely related to the effective stress but rather a
number of other factors can enter into the relationship . This point is discuss
ed in a companion paper to this Conference by R . V. Whitman . It is the in
tention of the following paper to present a mechanistic picture of the deve lop
ment of shear resistance of soils by considering the nature of shear resist
ance of soils by considering the nature of shear resistance between individual
soil particles .
The following concepts of shear resistance on a submicroscopic scale are
not intended as an alternate , or substitute , approach to the macroscopic
MECHANISTIC PICTURE 557
approach . The mechanistic concepts are complimentary to the macroscopic

approach in that they help one in understanding and using macroscopic con
cepts , especially the effective stress principle . An analogous situation exists
in the field of metallurgy where studies of the crystal structure have helped
considerably in understanding metal behavior , but have not invalidated or re
duced the usefulness of the principles of mechanics as applied to meta ls .
While much progress has been made in the development of submicroscopic
scaled explanations for the behavior of fine grained soils , both natural and
compacted , these mechanistic pictures are still far from precise or fully sub
stantiated . On the basis of particle arrangements and electrical forces be
tween particles , mechanistic pictures which help explain many of the complex
phenomena in soils such as thixotropy , structure , effect of type of compaction
and molding water content on compacted clays , etc . , have been presented .
These principles are , unfortunately , quite general and nonquantitative .
The main reason that these concepts are still general and not fully sub
stantiated lies in the absence of methods of actually observing the arrange
ment of individual clay particles and of measuring the forces between them .
As better techniques become available , marked progress in the submicroscop
ic study of soil behavior will be made . It is doubtful, however , if the soil engi
neer will ever be able to develop quantitative expressions for soil behavior
based on atomic structures of minerals to the extent that the metallurgist has
with metals because of the complex and heterogeneous nature of soil .
The remaining portion of this paper will first give a brief description of
the nature of soil mineral surfaces , followed by a consideration of forces be
tween particles , and , finally a description of the various components of shear
resistance . The Author gratefully acknowledges his colleagues ' assistance
in preparing this paper . Professor R . V . Whitman , Dr . R . T . Martin , Mr .
C . C . Ladd and Mr . R . M . Harkness read a draft of the paper and made many
helpful suggestions which were incorporated in the final copy .
II . NATURE OF CLAY PARTICLES
The logical starting point for a consideration of two or more adjacent clay
particles is an examination of an individual particle . An isolated clay particle
is indeed a colloid in the classical sense of the definition in that its behavior
is controlled by surface electrical forces . A dilute suspension of clay in
water exhibits all of the properties of an electrocratic sol. Clay particles in
a typical element of natural soil , however , cease to behave as true colloids .
While electrical forces emanating within the soil particles play a critical role
in the behavior of the particles , externally derived forces are also of major
importance . Thus while an isolated clay particle in water exhibits Brownian
movement - its mass is of no concern - the behavior of the same particle in
an element of natural soil is greatly influenced by the weight of over lying soil
particles and by loads applied to the soil element by an engineering structure .
The soil engineer must , therefore , combine these forces - ones derived from
a source outside of the individual particles under consideration - with those
emanating within - colloidal -type electrical forces . Let us first consider the
nature of the individual clay particles as a first step to an examination of the
electrical forces which arise within the particle .
The crystal structures of the common minerals which compose clay have
been approximately delineated and are widely illustrated in the literature -
e . g ., Grim (1953 ) and " Clays and Clay Minerals ” (1953 - 1958 ). The individual
clay particles are , in general , sheet mineral crystals made up of various
combinations of silica , brucite , and gibbsite type structures . Their lateral
surfaces are oxides or hydroxides . Because of isomorphous substitution oc
curring within the mineral lattice (and other factors ) the individual crystals
carry net electrical charges . These net electrical charges are balanced by
ions at the surfaces of the minerals ; since these ions can readily be replac
ed with others , they are termed " exchangeable ions . "
To help give a picture of individual clay particles two typical clay particles
will be examined in more detail . These two particles are shown , to scale , in
Fig . 1. Fig . la shows a typical particle of montmorillonite , one of the small
est and most water sensitive minerals encountered in clay ; Fig . 1b shows a
typical particle of kaolinite , one of the larger and less water sensitive min
erals encountered in clays . These two clay particles carry net negative
charges because of isomorphous substitution within the lattice - a typical sub
stitution in kaolinite is , one aluminum ion for every four hundredth silicon
ion and in montmorillonite , one magnesium ion replacing every sixth alumi
num ion . Because of this isomorphous substitution the montmorillonite parti
cle has a total net negative charge of the order of 15 , 000 ( ions per particle )
and the kaolinite particle a charge of the order of 4, 000 , 000 ( ions per parti
cle ) . In the following discussion sodium has been chosen as the exchangeable
ion for illustrative purposes . The montmorillonite particle in Fig . 1 would
thus carry of the order of 15, 000 sodium ions and the kaolinite particle carry
4 , 000 , 000 sodium ions .
In Fig . 2 are shown small sections of the surfaces of the two soil particles
when they are completely dry . The spacing of the exchangeable sodium ions
are shown to scale to give an idea of the approximate spacing of these ions .
It should be noted that the charge intensities are statistical numbers and the
ion spacings shown in Fig . 2 are thus more regular than would actually occur .
If the individual clay particles are now dropped into water both the mineral
surfaces and the exchangeable ions pick up water , i . e ., hydrate . Upon hydra
tion the sodium ion grows some seven fold as is illustrated in Fig . 2c . As the
scaled drawings shown in Fig . 2 show , the hydrated sodium ions are too large
monoionic layer on the mineral particles even they wanted
fit
into
to
to
if
a
.
Actually the exchangeable ions with their shells water move away from
of
,
the mineral surfaces positions equilibrium The ions are attracted

to
of
to
the mineral surface satisfy the negative charge existing within the surface
to
they also desire move away from each other because their thermal en
to
of
ergy the actual positions they occupy are compromises between these two
;
types forces Thus when the individual particles are dropped into water the
of
ions move away from the surface form what termed double layer
is
In
to
."
"
Fig are shown the clay particles with the fully developed double layers they
3
.
would have pure water Fig shows three dimensions the same surface
in
4 in
4
of . .
sections presented Fig From Fig one can get some idea even though
in
,
.
the drawings are crude the approximate spacing the hydrated ions
of
the
in
,
double layer This spacing represents maximum since the pore fluid this
in
a
.
distilled water Fig are shown the double layers

of
case
is
the sodium
In
5
.
kaolinite particle and the sodium montmorillonite particle the same scale
to
Fig Fig
5a
the ions around our selected surface sections are

as
in
In
4
.
.
shown as point charges Fig are plots

of
of
5b
the concentration sodium

In
.
ions vs distance from the particle surface At distance of approximately

a
.
the double layer the concentration

of
of
400 the extent sodium ions has

,
,
A
0. Montmorillonite
1000 A by 10Athick
b. Kaolinite
10, 000 A by 1000A thick
Typical Clay Particles
Fig . 1
+ 12 A
a. Kaolinite b. Montmorillonite
Dry Clay Surfaces with Exchangeable Sodium

420
Hydroted
71
8A
8A
Nô
Hydrated
77
R•
,
0+
+
.
-
-
.
No
98
Un
hydroted
A
R
0
,
*
.
-
"
lon
Sodium
c
.
Fig
2
.
SodiumMontmorillonite
a
.
SodiumKoolinite
b
.
Fig
3
.
a. Sodium Kaolinite
--
b. Sodium Montmorillonite
Fig . 4
it ++ + + + + + +
.
a 13 * + + + + +
+
+ +
+
+ + + + +
++ + + +
+
+
+
+
+ +
Concentration Nat
of of
Concentration
Distance from Surface

in

→
in
A
A
,
-
Electrical Potential
Electrical Potential

in

→
in
→
A
,
A
,
SODIUM KAOLINITE SODIUM MONTMORILLONITE
Fig
5
.
become equal to that in the " pore" or " free " water . In Fig . 5c are plots of
electrical potential vs . distance from the surface . The double layer is thus
the distance required to neutralize the net charge on the particle , i . e. , the dis
tance over which there is an electrical potential around the soil particle .
When two charged particles approach each other close enough for the double
layers to interact significantly , the two particles exert an electrical repulsive
force one on the other .
Figs . 3, 4 and 5 show about the same thickness of double layer for kaolinite
and for montmorillonite . The potential in the kaolinite double layer at any
distance is larger than that at the corresponding distance from the surface of
the montmorillonite . The higher potential results from the higher charge den
sity for kaolinite . In other words , on the basis of extent from surface , the
double layers are about equal; however , on a basis of potential the double lay
er of kaolinite is much “ larger " than that of montmorillonitę . The double lay
er water per mass of clay mineral is very much greater in montmorillonite
than in kaolinite . For dillute ionic concentrations , the extent of double layer
as well as the potential is directly related to charge density . For typical clay
water systems , in which the particle surface potentia ls tend to exceed 150
millivolts , the extent of the double layer is approximate ly independent of sur
face potential . This principle was used in preparing Figs . 3, 4 and 5 .
In the preceding discussion of the nature of the soil particle , the electrical
nature of the ends of the particles has been neglected . There is convincing
evidence that the ends of clay particles carry not negative but positive charges .
This fact would mean that the exchangeable ions are somewhat more closely
spaced than shown in the figures and that the soil particles are able to attract
each other in an edge - to - face alignment . This attraction will be considered
later .
We can generalize our picture of an individual wet clay particle as : a min
eral particle surrounded by a double layer consisting of mobile ions in water .
The size and nature of the double layer is very sensitive to environmental
factors - temperature , concentration and type of ions in the double layer , and
dielectric properties of the double layer fluid .
PARTICLES
III
FORCES BETWEEN SOIL

.
overwhelming mass experimental data prove that the

an
to
of
There
is
of
shear resistance which can be mobilized between two bodies function

is
In a
the pressure acting normal the contact between the bodies soils
to
,
.
pp
flopp
"
where
on
Tff stress
shear the failure
=
surface failure
at
on
Off normal effective stress

=
failure surface failure

at
attempt this expression

an
learn more about use and limitations

of
be an
to
In
the forces acting between adjacent soil particles will next

of
examination
undertaken Since shear soil occurs between particles and not through
in
.
them the forces normal the potential shear surface between particles
to
is
it
,
that will be treated

.
In Fig. 6a is shown an average interparticle condition for the soil mass . In

other words , the condition shown is selected to represent the whole potential
shear surface , thus an element d . d corresponds to a unit cell in a crystal; it
is , in effect , the repeating unit . The condition is general as it includes all the
possible interfaces , namely : mineral - mineral, air - mineral, water - mineral
and air - water .
In Fig . 6a are shown the forces acting between the particles ; they are:
Fm = force where contact is
mineral - mineral
F. contact is
= force where
air - mineral
F ... = force where contact is
water - mineral or water
water
R ' = electrical repulsion between
particles
A' = electrical attraction between
particles .
These forces acting across the potential shear surface must be equal to com
bined normal stress times the area , i . e . ,
odd - F
ma +F +F
' w
+R '- A '
or
odd
CAPĄ
=
-'
R
' A
where
,
Am
Fa
Fw
m
uxx
-
P
-
.
-
or
oamp
aa
ua
+
A
R
+
-
=
o
where
,
amanda anadan olanda Ravel

.
A
Eq the pressures acting normal

relates the potential shear surface
to
;
it 1
.
statics correct The limita

of
since accordance with the laws

is
is
in
1 it
. ,
Eq
in
tions and they are considerable the use lie the evaluation
to
of
in
,
)
(
fact the interpretation the terms Let therefore consider turn the
us
of
in
,
,
various terms Eq
in
1
.
.
Shear Surface
Potential
YZWOIL
air
Fig Forces between Adjacent Particles

6
-
.
.
Fig Arrangements Adjacent Particles

of
7
-
.
.
Fig
8
.
1. Mineral -Mineral Contact . –
The term õam is easily visualized in granular soils . Bishop and Elden
( 1953),” Roberts and de Souza (1958 ) and others have shown that the compres
sive contact stresses in granular soils can be large enough to cause crushing
of mineral particles . Terzaghi (1936 ) showed many years ago that am is very
small in granular soils . It is thus pretty clear that there is an actual mineral
contact in granular soils and that this value depends upon the shape of the par
ticles and the magnitude of the contact pressure . At most it is less than a few
per cent of the total area under consideration .
The meaning of contact between clay particles is , unfortunate ly , far from
clear . Consider for a moment forcing the two particles shown in Fig . 4 into
" contact . " Before the mineral surfaces of the two particles can touch each
other the mass of hydrated ions must either be extruded or must have their
shells of hydration water extruded from between the particles . One would
thus suppose that a mineral to mineral contact between clay surfaces in the
sense that we had contact between sand particles or silt particles is not easily
attained except in highly flocculated systems. Considerable time and pressure
would be required to extrude the water from the zone between approaching
particles . Soils which have been flocculated , heavily overconsolidated , desic
cated or cemented (cementation is discussed later in this paper ) undoubtedly
have true areas of mineral to mineral contact . Rosenqvist ( 1959 ) showed , for
example , complete removal of water from a flocculated soil resulted in es
sentially no volume change . He suggested values of am less than 1% In high
ly plastic , freshly remolded clays, and clays compacted wet of optimum , it
would seem logical that the contact area between clay particles is essentially
zero .
2. Air -Mineral Contact .

In soil - water systems where there is
an availability of water it is difficult
to imagine the existence of an air - mineral contact surface . Since both miner
al surfaces and exchangeable ions adsorb water , having a polar molecule , in
preference to air there should be no air - mineral surface if there is a suffi
cient supply of water available . The thirst that a clay has for water is clearly
illustrated by the fact that an oven - dry clay will pick up moisture in a humid
atmosphere .
3. Water - Mineral Contact :

There is , of course , no question as to the existence of water -mineral con
tact zones in all types of soil . There is , however , considerable uncertainty as
to the nature of the water near the surfaces of clay particles . As already has
been discussed in this paper , all of the water in the double layer around a clay
particle is under some attractive pull to the soil surface. Not only is water
held to the soil surface by being hydrated onto the exchangeable ions which in
turn are held to the mineral surfaces , but the water is attracted to the surface
itself because of the polar nature of the water and the stray electrical charges
on the mineral surfaces . While there is considerable controversy as to the
exact nature of the water immediately next to the mineral surface , it is gener
ally recognized that at least the first few molecular layers around the soil
2. Items indicated thus , Bishop and Elden ( 1953 ) , refer to corresponding en

tries arranged alphabetically in the Appendix Bibliography .
surface are attracted with some significant force to the mineral surface. The
evidence avai lable , especially Martin ' s work (1959 ) which shows more dis
order in the water next to the mineral surface than in free water , indicates
that the water immediately next to the mineral surfaces is not " ice - like . ” All
of the unusual behavior of clay systems can be explained in terms of the pic
ture which has been presented herein without resorting to any postulation of
an ice - like layer of water .
The real question which needs answering is : can the water immediately
next to the mineral surface transmit both compressive and hydrostatic pres
sures ? Can the water between two particles which are very close together
transmit compressive stresses between particles and thus act as if it were
mineral , while at the same time transmit hydrostatic pressures and thus act
as if it were water ? This is a basic question which can not be positively an
swered at this time.
The Author is of the opinion that the water immediately next to the mineral
surfaces is free to move laterally along the surface of the mineral and can
thus transmit hydrostatic pressures . The double layer theory would appear
to account for the unusual properties of soil . At the present state of knowledge
the Author takes the view that all of the soil water , even that immediately next
to the mineral surfaces , behaves as " water " and transmits only hydrostatic
pressures .
4. Electrical Attraction and Repulsion .

In Eq . 1, A is the electrical attractive pressure and R is the electrical re
pulsive pressure acting between the particles ; they are similar to the pres
sures which can act between bar magnets . These electrical pressures have
two important characteristics : 1) they originate within the particles ; and
2) they can act over relatively large distances , i . e. , greater than 100 A . The
first characteristic means , for example , that significant pressures could be
acting between the isolated pair of particles shown in Fig . 7a even though no
externally derived pressure is applied to the pair . The second characteristic
means that A and R can act , even though the particles are not in contact .
The electrical forces between particles in the simplified system shown in
Fig . 7 have been treated extensively by the colloid chemists . R is a Coulombic
pressure arising from the like charges on the two particles ; A is the van der
Waals or secondary va lence pressure . Formulas showing the variation of A
and R with distance and with system characteristics have been deve loped (see
Kruyt , 1952 ) . Both A and R are very sensitive to distance ; for most soils ,
R - A apparently increases with decreasing interparticle spacing . These for
mulas also show that Ris relatively sensitive , but A insensitive , to the system
characteristics .
The colloidal formulas have been proved qualitatively , and to some extent
quantitatively , correct in very dilute soil - in -water suspensions . One can show
experimentally that any of the following system changes will tend to cause the
flocculation of a soil water suspension :
1. Increase electrolyte concentration
2. Change the cations in the fluid to ones of higher va lence
3. Reduce the die lectric properties
4. Increase the temperature .
Flocculation indicates a reduction in R - A ; or since A is relatively insensitive
to environment , flocculation indicates a reduction in R.
The colloidal formulas are , unhappily , of very limited , quantitative value
in typical soil systems with which the engineer deals . Among the reasons for
this situation are :
1. Soil particles , while carrying net negative charges , can have positive
charges at their ends ; these end charges are not covered in the col
loidal theories ;
2 . The particles can be much closer spaced and more concentrated than
assumed in the theories ;
3. Externally derived forces exist ;
4 . Soil particles can change orientation , as well as spacing .
Consider , for example , the particles shown in Fig . 7b . The electrical

forces acting between Particle 1 and Partic le 2 are of the same type as be
tween the particles in Fig . 7a , but their evaluation is much more difficult .
Further , forces from Particle 3 ( Fig . 7b ) act on both of the others .
In Fig . 7c are shown two particle pairs with the same inter - particle spac
ing . Since for both A and R the force - distance curve is steeper than linear ,
both A and R are changed by altering the orientation of the particles . In going
from the perpendicular to the parallel arrangement , both R and A would tend
to decrease . In a system of parallel particles theoretical pressure - distance
expressions have been shown closely related to pressure - void ratio consoli
dation curves ( Bolt , 1956 ) . With natural soils , with their more random parti
cle arrangements , such checks have not been obtained .
In summary , A and R in Eq . 1 are electrical pressures between particles ;
however , only in very carefully selected soil - water systems can A and R be
evaluated . There are undoubtedly components to A other than van der Waals
forces and to R other than Coulombic repulsion , particularly at very close in
terparticle spacing , ( See , for example , Low and Deming ( 1953 ) , Norrish (1954 ) ,
and MacEwan (1954 ) ] .
A further look into the pore pressure term , u , which appears in Eq . 1 would
seem to be warranted . The value of u which should go in Eq. 1 is the pore
pressure as would be measured by a piezometer . Specifically , u , is the pres
sure in the fluid at Point 2 in Fig . 8, as measured by a piezometer inserted
at Point 2 and containing a fluid of the same composition and at the same tem
perature as that at Point 2 . At equilibrium the pressure in the fluid of this
composition is the same throughout the system ; if it were not , water would
flow to equalize any differences in pressure . As has been discussed and illus
trated in the preceding pages , however , there is a difference in cation con
centration between the Points 2 and 1. Thus in addition to the pore water pres
sure as measured at Point 2 , there is at Point 1 pressure (which can be con
sidered a cation partial pressure ) resulting from the higher concentration of
cations at Point 1. In other words , one can consider the total fluid pressure
between Points 1 and 2 being unequal by an amount of the partial pressure of
the excess cations at 1. This difference in total fluid pressure or osmotic
pressure between Points 2 and 1 is numerically equal to the electrical repul
sion between particles . In other words , the electrical repulsive pressure , R ,
plus the pore water pressure , u, is the total fluid pressure existing at the mid
plane between the adjacent particles , Were it not for the fact that these two
pressures act over different areas , it would make good sense to combine the
term uaw + R in Eq . 1 into one term representing the total fluid pressure be
tween the particles .
As Eg. 1 is a general and correct expression of equi

has been pointed out,
librium of pressures acting across the plane of potential shear between two
particles . For certain soil systems , or for systems where we can justifiably
make approximations , there are special cases of Eq . 1. Some of these special
cases are as follows :
.
1 No Air -Mineral Contact .
For this case Eq. 1 simplifies to
0 = gan
m tua w + R -A .
This condition holds for soils containing no air , i.e ., saturated ; or for soils
where all of the air exists as air bubbles surrounded by water .
.
2 No Air - Mineral Contact or Mineral-Mineral Contact .
Equation 1 simplifies to
0 : u +R - A
or
0 - -A ( 3a )
where Pe is the total fluid pressure as discussed above and equal to u + R. A

very plastic , saturated , dispersed soil would meet the conditions of Eq . 3. A
system of sodium montmorillonite in water would , for example , appear to
meet the conditions of Eq . 3 and 3a , as shown by Bolt (1956 ) .
3. No Air - Mineral Contact and No Net R-A .

For this condition Eq. 3 simplifies into
5am
o = tuam
or
4a
)am
mtu
za
)
=
(
o
(
-1
.
would be met granular soils where the degree

of
This condition
in
saturation
also possible that Eq certain clay
by
relatively high could be met

It
is
is
4
at .
water systems specific void ratios

.
On the basis the preceding discussion we can see

of
at
least two circum

stances where we can mechanistically visualize the effective stress
u
;
o
,
-
no
of
air mineral contact and mineral mineral con

no
for the condition

.,
e
-
i.
tact
,
=
Ō
R
A
-
(5
)
air
the
for
no
or
no
mineral contact
of
and condition net
,
R
A
-
-
oam
īz
or
17
uam
(6
)
=
-
õ
other words highly plastic saturated dispersed clay the effective
In
in
,
,
a
stress the net electrical stress transmitted between particles and
is
in
is ;
a
granular soil high degree approxi
of
to atsaturation the effective stress
mate equal a
the load carried by the soil ske leton divided by the total area
ly
under consideration
.
COMPONENTS OF SHEAR RESISTANCE

IV
.
shear resistance will
of
this section the mechanistic components
, be
In
dis
the discussion will
be
cussed Used three terms cohesion dilatancy and
in
,
.
-
friction which are widely used soil mechanics literature Unfortunately
in
,
-
.
these words mean different things different people Serious consideration
to
.
was given avoiding their use this paper and coining new terms denote
to
to
in
the mechanisms described the following pages The decision was made
in
,
.
however employ these terms principally because they best express what
to
to ,
,
meant and because attempts alter common usage are usually futi The
le
is
.
words cohesion dilatancy and friction are used herein the classical
in
,
,
-
dictionary sense denote the mechanisms described This use does not sug
to
gest that there now any standard laboratory test which can evaluate these .
is
strength components As matter fact the boundaries between the three

be of
,
a
.
will
as
mechanisms are not sharp discussed further any given case

in
,
;
the actual shear resistance may be furnished by two

of
more the compo
or
nents acting together

.
Cohesion
.
1
.
By cohesion meant shear resistance which mobilized between

is
can be
”
"
two adjacent particles which stick cohere each other without the necessity
i. to
,
any externally derived normal pressure

of
the linkage between the ad

;
e
.,
jacent particles from forces arising within the particles themselves Vari
is
ous types of cohesion are illustrated Fig

in
9
.
.
Fig and Fig

9b
9a
show flocculation cohesion Salt type flocculation

.
arises when enough electrolyte added soil water suspension cause

to
to
is
the clay particles approach each other and stick by van der Waals forces
to
;
'
Eq
by
exceeds Salt type flocculation treated extensively

in
is
R
.,
1,
A
e
i.
the classical theories colloidal chemistry The particles natural soil

of
in
.
which are deposited salt water marine clays many lake clays would
in
,
-
-
9b be
by
this type electrical force

of
to
tend linked
.
Fig one arising from

of
shows another type electro

an
flocculation
-
.
static attraction between the positively charged end one clay particle the
to
of
negatively charged face another particle As Schofield and Sampson 1953

of
,
.
have shown this type edge face linkage tends kaolinite

of
occur
to
to
1954
in
,
)
water systems low electrolyte concentrations

at
flocculation probably relatively weak and sensitive

en
Cohesion from
is
to
caused by altering environmental fac

as
be
vironment Just flocculation can

., .
by
or
heating adding salt destroyed

be
tors can doing the reverse

.g
it
e
-
cooling or removing salt

e
,
i.
.
.
a Salt Flocculation
b. Edge - Foce Flocculation
S C. H - Bonding between Kaolinite Sheets
d. K - Bonding
e. Cementation
Fig . 9. - Various Types of Cohesion
There are several ways in which individual clay crystals can become more
or less linked one to another , to in effect grow into larger parti
permanently
cles . Two of these possibilities are shown in Figs . 9c and 9d . The potassi
fit
um ion is of such size and coordination number that it can easily into sur
face holes the silica sheets Drying can bring the adjacent clay sheets
in
close enough together for the exchangeable potassium ions tightly into
fit
to
the holes and thus become nonexchangeable and serve as permanent linkage
a
the clay crystals This type intercrystal linkage amounts ef

of
between
in
,
.
crystal growth
to
fect
,
Adjacent soil particles can become cemented one another as illustrated

to
Fig Naturally occurring materials which are

9e
known be cements are

in
to
:
.
carbonates iron oxides silicates aluminates and certain organic matter

,
These materials can cement small particles another produce larger

to
to
one
particles
.
particle cementation plot

by
suggested Fig
of
10
The wide occurrance

is
a
,
.
experimental data natural soils Lambe and Martin 1956 Fig

on
of
10a
,
).
.
(
shows per cent clay composition clay size all the clay material existed
on vs
If
.
clay size all points would plot As can be seen

or
above the 45° line

in
as ,
.
(
most the points fall low the 45° line indicating clay material existing
of
be
particles larger than Figs 10b and 10c

as
The test data are plotted

24
by in
per cent clay size and per cent clay composi

by
divided divided
P
P
, .
.
I.
.I
tion respectively
vs
composition The hatched areas indicate the theoretical

,
.
.
100
0 - lllito - Montmorillonoid
A - Illite - Chlorite
SOL
20
MINUS
100
%
o
40
*
no
40 6080 IOO
20
CLAY CONTENT
(
)
Fig 10a Clay Content versus Clay Size

-
.
ACTIVITY Plastidity Index minus

2
.
.
On
·
MONMORILLONOID
ILLITE
60
IN
0
ole
1
60
8
40
IN
% 0
80
20
.com
Kusual Ronge Depondino

,
Exchangeable
of
Nature lons
Fig 10b Plasticity versus Composition

-
.
.
Exchangeablelons
on
Kusual Rangs Depending

of
CLAY CONTENT
ONTENT
Nature
,*
INDEX
INDEX
PLASTICITY
2
PLASTICITY
+
+
+ 10
oo
20 40
o
80
N20
100
IN
x
60
MONTMORILLONGUP
0
IN
%
ILLITE
Fig 10c Plasticity versus Composition

-
.
test results from pure minerals Fig 10b and Fig 10c show
on
values based
.
actual plasticities far below what would be expected from compositional data
.
These data indicate that clay minerals can intergrow and

or
aggregate
to
form
/
larger and less plastic particles Similar work on other soils Lambe and
(
.
Martin 1954 showed all the natural soils which had aggregated fines con
of
,
tained one or more of carbonates iron oxides and organic matter

,
:
Particle cementation aided by drying Not only does drying bring the
is
adjacent particles close enough together permit cementing but also tends
to
precipitate soluble cementing materials from the pore water One led
is
to
to
.
suppose that some drying which are commonly attributed

of
the effects
, to
of
capillary precompression may actually be due chemical weathering

to
e
,
i.
.
cementing Drying for example tends produce ferric oxides which are
to
,
,
.
relative active cementing agents Thus brown or red colored soils produc
ly
by desiccation often contain fines which are cemented together

ed
Soil stabilization studies have shown the potency aluminum and silicon
of
materials as cements Many of the stabilizers sodium silicate phosphoric

,
(
.
by
acid sulfuric acid Portland cement part

/ at
lime etc function least

in
,
,
.)
supplying directly indirectly amorphous aluminates and silicates

or
or
to
weld adjacent soil particles These same aluminates and silicates are com
.
mon weathering products which are available usually very low concentra
in
,
tions ground waters Many sedimentary rocks are fact made by the
in
in
,
,
.
cementation of soil particles by weathering products carried in ground water .

general , cohesion , unlike the other two components of shear resistance ,
In
is characterized by very low strains needed to mobilize the resistance . This
important characteristic is evidenced in strength tests on soils high in cohe
sion , e. g .:
1. The stress - strain modulus of clay compacted on the dry side of opti
mum (where there is a tendency toward flocculation ) is much greater
than that for the same clay compacted wet of optimum . The stress
strain modulus of dry - side compacted clay may typically be 5 to 10
times that of the same clay compacted on the wet side .
2. Soil - cement and soil - lime are much more rigid than the same soils
without the stabilizers . A soil stabilized with cement may typically
have a stress - strain modulus 10 times that of the untreated soil.
In the preceding discussion three important characteristics of cohesive
strength have been brought out , namely :
1. Cohesion can be variable - van der Waals and electrostatic cohesion
can be sensitive to environment .
2 . Cohesion may merely make larger soil particles out of smaller ones ;
thus a highly cemented soil may show its improved strength by an in
crease in the friction angle rather than the cohesion intercept on an
effective stress - strength plot . ( This fact is illustrated in Fig . 14 . )
3. Cohesion is generally mobilized at relatively low strains . Thus in a
natural soil, cohesion can be mobilized and destroyed before the other
components of strength become active . This point will be discussed
later .
From the preceding discussion one can easily see why mechanistic cohe
sion may not be easily detected by standard strength tests run by a soil engi
neer .
2. Dilatancy and Friction .

The next two shear components - di latancy resulting from particle inter
ference and friction are treated together because of their close relationship .
During shear displacements the moving particles tend to interfer with each
other both electrically and physically . If the interference results in a ten
dency toward a volume increase , a higher shear resistance can be mobilized .
Eq . 1 expressed an equilibrium of pressures across a shear surface . Moving
the particles can cause a change in the electrical forces , thus requiring a
change in volume or change in some other pressure , such as oor u , to main
tain equilibrium . This phenomenon can result in a di latancy component to the
shear resistance .
Dilatancy in granular soils is well understood . If the three particles in
Fig . 11 a lie on a horizontal shear surface , Particle 1 must move vertically in
order to move horizontally relative to Particle 2 . For Particle 1 to move
vertically against the normal pressure requires a larger applied horizontal
stress than for it to move along a smooth horizontal surface . Thus the soil
in Fig . 11a has a higher shear resistance because of a physical or geometric
interference of particles resulting in a tendency toward volume increase or
di latancy . The added shear resistance from the di latancy can be computed ,
as done by Taylor (1948 ) .
CHANISTIC PICTURE
Particle
Porticle
a. Interference in Silt
b. Interference
in Clay
Friction between
C
.
Rough Clay Surfaces

bolli
toshorstof Friction between

d
.
Smooth Clay Surfaces
Fros rasi
,
Fig
11
Friction Soil
in
-
.
.
as
Combined measured
,
v
Dilatency
STRESS
Friction Interference
+
STRAIN
→
of
Fig Components Shear Resistance

12
.
-
.
Figs . 11b through 11d show various ways friction between bodies may be
mobilized . Fig . 11b illustrates physical interference of clay particles , one on
another . In this system , however , the sliding of particles becomes of more
relative importance than for the system in Fig . 11a .
Fig . 11c shows two particles with their contact parallel to the direction of
shear . After sliding a distance x , however , there will be a physical interfer
ence arising from roughness on the particles .
Fig . 11d illustrates , on an atomic scale , the frictional resistance between
two clay mineral surfaces , e. g . , adjacent montmorillonite surfaces . The lat
tice oxygen anions are seats of negative charge while the lattice silicon ca
tions are seats of positive charge . Thus while the system is electrically neu
tral in toto , there are local variations in charge . The sites of isomorphous
substitution , exchangeable ions , impurities , etc . , are other sources of elec
trical charge - variation . To move the surfaces relative to one another re
quires a shear force because the local e lectrical charge pairs exert forces
one on another .
A most important characteristic of both dilatancy
and friction is that the
shear resistance which can be mobilized is direct function of the force act
a
ing normal to the shear surface . This fact is readily evident for all of the
systems shown in Fig . 11, with the possible exception of that in Fig . 11d . One
can reason this case as follows : the greater the normal force , the closer the
surfaces , and thus the greater the influence of electrical force fields , one on
another . The experimental evidence ( e. g ., Bowden and Tabor , 1942 ) surely
proves that , even between truly " smooth " surfaces , frictional resistance is
directly related to the normal stress acting across the shear plane .
3. Combination of Cohesion , Dilatancy and Friction .

The possible components of shear resistance have been considered individ
ually . In general , however , the actual shear resistance is a combination of
components . Fig . 12 suggests an addition of all of the components to give the
one stress - strain curve which would be measured in a shear test . The cohe
sion is mobilized at very small strains ; after cohesion is destroyed it con
tributes no further to shear resistance . The interference of particles causes
a volume increase tendency - or added shear force to overcome . After some
straining , there is no further tendency toward volume increase , thus the dila
tancy component disappears . When the stress - strain curve becomes horizon
tal, the only component to shear is friction , including particle interference .
As shear occurs clay particles tend to align themselves with their long axes
parallel to the direction of shear . Thus with increased shear in clay , friction
between parallel plate surfaces increases in importance .
While this situation shown in Fig . 12 seems reasonable , cohesion can exist
and manifest itself as increased friction and di latancy . This situation is dis
cussed in the next section .
V. EXAMINATION OF SHEAR MECHANISM
In the preceding pages the mechanism of shear resistance in soils has been
discussed . First the nature of clay particles was considered , next an equa
tion of equilibrium of pressures across a shear surface was presented , and
finally the components of shear resistance were discussed . A reason for this
submicroscopic treatment of shear is to give the soil engineer insight into
shear phenomena . To become of full value to the engineer these concepts

must be examined experimentally and made more " operational . "
There is a considerable amount of experimental data (e . g . , Seed and Chan ,
1959 ; Lambe , 1953) which shows that many of the principles presented here
in and in other papers by the Author are not incorrect . It would thus appear
most useful in the limited space available herein to examine the concepts not
previously examined in other publications . Let us start with Eq . 1,
o = o am + Paaa + uaw + R - A.
In Eq . 1 only twoterms , o and u , can in general be measured ; therefore , a
thorough evaluation of Eq . 1 is not possible . The results of various tests lend
support to the validity of Eq . 1. Three of these tests are next described :
1. A group of samples of plastic clay compacted wet of optimum were

soaked , some in pure water and some in salt water . The amount of
swell upon immersion was inversely related to the concentration of salt
in the water . In fact , at very high salt concentration , actual shrinkage
occurred . Typical test data are (Ladd , 1959) :
Soaked in distilled water , swell = 5 % of initial volume .
Soaked in 0. 5 M CaCl2 , swell = 4 % of initial volume .
Soaked in 5 M CaCl , swell = - 0 . 3% of initial volume , i . e. , a volume
decrease of 0 . 3%
Increasing the salt concentration in the pore fluid reduces R , so the in
by
terparticle spacing reduces enough
for
the equilibrium expressed

Eq re established
be
to
1
.
-
.
saturated plastic clay was sheared failure triaxial

of
sample
to
in
a
A
2
.
which point the temperature was reduced while the volume was
at
test
,
held constant The following data this test was run by Harkness
M
R
(
.
.
.
were obtained
at
M
T
.
.I
:
.)
°C
Temperature
kg
om cm2
,
,
./
-u
13 18 . 25
260
.
22 300
3
348
.
375
.
These data show reduction with cooling Cooling increases

in
in
R
u
a
Eq the pore pressure therefore decreases maintain equili

to
u
,
,
1
;
.
brium Au AR
=
+
0
.ie
.,
,
Many tests have been run which samples clay have been cooled
of
in
3
.
.
When the confining pressure was held constant sample expansion

oc
a
,
curred Salas and Serratosa 1957 when the volume was held con
g
;
,
e
.,
)
.
(
stant the confining pressure increased

.
These and other tests show that Eq not incorrect and fact sug
in
is
it
,
,
1
.
gests that for the terms investigated correct

is
,
it
Eq most helpful
of
giving fundamental understanding volume

in
is
a
1
.
changes The preceding tests illustrate volume changes pore pressure

or
.
of
changes needed change Consider now the effect

in
accommodate
to
R
a
.
)
the following ways

so
applying Ao sample This can be carried

to
in
a
:
.
am
Increase soils with mineral mineral contact very little

In
in
A
1
-
.
required pick Ao
up
volume
to
is
.
air
2. Increase in Paaa : Any pressure increment applied to continuous
is
is by
dissipated
up
quickly air flow Discontinuous air can pick Ao but
a
.
relatively large Ay required because the compressibility the air
of
of
.
pore pressure can pick
up
Increase An increase Ao
no in
es
uaw
at
no in
3
.
air
sentially volume change bubles are
to
the water make
in
it
,
if
compressible
.
carried by but reduction
be
The Ao can
in
Increase
in
(R
A
A
R
a
A
,
4
)
-
:
-
.
average particle spacing needed This reduction can be accomplished
is
.
by volume decrease very common by change particle orienta
or
in
a
a
(
)
tion less parallel array This rearrangement hard visualize
is
to
to
a
.
as happening As matter of fact the reverse common dur
is
g
e
a
,
:
.
.
.
ing shear clay particles tend be aligned reducing thus causing
an to
R
A
,
,
-
either volume decrease or increase pore pressure
in
a
.
Eq study shear strength straight forward
to
so
of
The use not The
1 is
1
-
.
.
exact relationship between the terms Eq and shear strength somewhat
in
is
, .
obscure Certainly the closer the particles the lower the void ratio for
e
.,
.
.i
the same particle arrangement the higher the friction and interference
In
,
.
creasing and reducing tend permit the particles approach thus tend
to
to
R
,
A
ing the interrelationship
of
increase contact stress Because however
to
,
.
of
these terms simple tests which only one term varied are not easy
is
in
,
.
for example tests strength
of
varying
on
Consider show the effect
to
,
R
.
sample saturated clay
of
cooled increases As increases the

is
If
R
a
,
.
increases with
of
volume reduction both and fact In

R
,
A
- a
ARfrom cooling Afrom expansion Thus for the cooled sample the
,
=
(R
A
:
-
volume greater but the net change terms Eq slight The change
is
is
in
in
1
.
.
in by
strength resulting from cooling followed expansion may insignificant
. be
If in
.
sample expansion prohibited change other terms Eq neces
is
in
is
a
,
1
sary maintain equilibrium
to
Figs are presented the results laboratory tests run by

of
13
) 14
In
and
.
(
Massachusetts clayey silt
on
Anwar Wissa 21 finer than

%
%
M
=
8
a
T
,
I
(
.
I.
in .
.
by .
Fig and Fig

.i 2M
13
14
The notation are those used Whitman 1960

),
.e, ).
shear stress where Mohr circle tangent envelope

is to
Tre
= =
is
off normal effective stress where Mohr circle tangent

envelope
to
03c consolidation chamber pressure

=
.
01
03
-
slope compressive stress axial strain curve

of
=
E
.
of
Fig three types undrained triaxial

13
shows the results

of
shear tests
.
with measured pore pressures drained triaxial and unconfined The silt
;
;
, .
hydrated lime cured relative humidity soaked un

at
was mixed with

%
100
,
5
confined water for days and then sheared The effective stress envelope
in
of 4
shows an inclination 450 The drained tests fall above the envelope pro
.
jected back the origin The results of unconfined tests indicate the exist
to
negative pore pressures during these tests

of
ence
.
silt plus lime are replotted Fig along with shear data
14
D on
The data
in
obtained by the silt alone and silt plus

on
an
Davison obtained
M
M
,
T
.
.
.
I.
(
.)
aggregant quarternary ammonium chloride ARQ and silt plus disper

a
,
a
”
"
)
(
sant sodium tetraphosphate STP The silt was compacted and permeated
,
"
"
.
(
)
OU : Undroined Triaxial Test

with Measured Pore
Water Pressure
+D : Drained Triaxial Test
/.
cm2
kg
in
Ime
2A cos 45° from Uncontined Tests
-
16
>
Top
./kg
cm
in
Fig
13
on
Shear Tests Silt Lime

%
+
5
-
.
.
with water aggregant solution dispersant solution the case may

or
be
un
as
,
,
til
equilibrium period days

12
flow was obtained

to
of
7
a
-
The test results show three significant facts that already have been dis
cussed namely
,
The lime cemented the small particles into larger ones causing in
an
,
1
.
crease strength angle from 37° 45° The lime also gave the silt
to
in
some mechanistic cohesion

.
The aggregant caused strength and the dispersant caus

an
increase
in
ed decrease strength when plotted against consolidation pressure

in
;
a
but they had the strength angle

no
on
effect One suppose that

to
led
is
.
changes will have noticeable effects undrained strength

on
and
in
,
A
strength as The re
on
but little effect function of effective stress

a
compacted soils also lend support

on
sults this supposition sam

to
a
-
ple compacted dry optimum has higher strength constant volume

of
at
, a
)
(
optimum but both strengths will fall

on
than one compacted wet

of
the
same effective stress Mohr envelope Seed and Chan 1959 show
in
,
.
on (
fact compaction method can have significant effect undrained

a
,
strength but all

Au
of
be
the difference can accounted for terms

of
in
,
The increase cohesion and decrease resulted greater

3
in
in
in
both
.
R
A
-
stress strain moduli

.
-
VI
CONCLUSIONS
.
of
of
This paper presents theoretical consideration the mechanism shear

a
strength generation fine grained soils First given brief description

is
in
a
.
soil mineral surfaces followed by consideration

of
of
of
the nature forces

a
,
./.
ARQ
2
Woter
STP
cm
kg
in
in
Zip
Zop
/..
./
2
.
ccm2
hg
m
IInn
kg
in kg./ cm2
Water
max
PSTP
o Lime
T
13c in kg / cm2
./.
Lime
2
cm
kg
ARO
Water
STP
vain
.kg
Icm
In
Psc
.2
Fig
14
.
between adjacent soil particles and finally description the various com
of
a
ponents The description fine grained

of
shear resistance the nature

of
of
.
soils includes sketches which are approximate scale

ly
to
a .
The herein presented concepts of shear resistance on submicroscopic

substitute for the macroscopic approach
as
to of
scale are not intended the

a
soil engineer Rather the mechanistic concepts are complimentary the

,
.
macroscopic approach and can help give considerable insight into the engi
neering behavior of fine grained soils
.
this paper equation based

an
on
Presented the laws statics which

of
is
in
relates the various pressures acting across shear surface between adjacent
a
soil particles . This equation is very helpful to an understanding of soil yol

ume change and shear strength characteristics . It is not , however , " opera
tional ” since it contains terms that cannot at the present state of knowledge
be measured ; in fact , the interpretation of several of the terms is difficult .
Three components of shear - cohesion , dilatancy and friction - are dis
cussed . Presented is evidence to show that a major role of cohesion , both
natural and artificial , is to make larger particles out of smaller ones , thus
from the macroscopic view increasing the " friction angle . " While all of the
three shear components can , in general , act in a cohesive soil , their maximum
values may not all be effective at the peak shear resistance of the soil mass
since the strains required to develop each are probably different .
Continued study of shear mechanism would appear to be very promising .
The use of shear tests with measured pore water pressures on stabilized soils
is an interesting avenue of research . By means of stabilizers the cohesion
component of shear strength can be altered . The deve lopment of a technique
which will permit the visual observation of adjacent particles in any soil with
out disturbing the particles will be a powerful stimulus to mechanistic re
search .
APPENDIX . - BIBLIOGRAPHY
1. Bishop , A . W . and Eldin , G ., "Undrained Triaxial Tests on Saturated

Sands and Their Significance in the General Theory of Shear Strength , ”
Geotechnique , Vol. 2, 1950 .
2. Bishop , A . W . , " The Principle of Effective Stress, " Teknisk Ukeblad ,

October , 1959 .
3. Bishop , A . W . , " The Measurement of Pore Pressure in the Triaxial

Test; " Conference on Pore Pressures , London , 1960 .
4. Bolt , G . H ., " Physical - Chemical Analysis of the Compressibility of Pure
Clay, ” Geotechnique , Vol. 6 , No . 2 , June , 1956 .
5. Bowden , F . P . and Tabor , D ., “ Mechanism of Metallic Friction , ” Nature ,

Vol. 150 , p. 197 .
6. Davison , D. M ., “ The Effects of Aggregants and Dispersants on the

Strength Characteristics of a Compacted Soil , " S. M . Thesis , M . I . T . ,
June , 1959 .
7 . Grim , Ralph E . , “ Clay Mineralogy , ” McGraw - Hill, 1953 .
8 . Hvorslev , Juul , “ The Shearing Resistance of Remolded Cohesive Soils , "
Proc . Soils and Foundation Conference , U . S . Engr . Dept . Boston , Mass .
June , 1938 .
9. Kruyt , H . R ., Colloid Science , Vol. 1, Elsevier Publishing Co . , 1952 .

10 . Ladd, C . C. , " Mechanisms of Swelling by Compacted Clay , " Highway Re
search Board Annual Meeting , January , 1959 .
11 . Lambe , T . W . , " The Structure of Inorganic Soil, ” ASCE Separate 315 ,

October , 1953 .
Lambe , T . W . and Martin , R . T. , " Composition and Engineering Proper

ties of Soil, " Highway Research Board Annual Meeting , 1956 and 1957 .
13 . Lambe , T . W.,
" The Structure of Compacted Clay " and " The Engineering
Behavior of Compacted Clay , " Journal of the Soil Mechanics and Founda
tions Division , Proc . of the ASCE , May , 1958 .
14 . Lambe , T . W . and Whitman , R . V . , " The Role of Effective Stress in the
Behavior of Expansive Soils , " First Annual Soil Mechanics Conference ,
Colorado School of Mines , April , 1959 .
15 . Low , P. F . and Deming , J. M ., “ Movement and Equilibrium of Water in
Heterogeneous Systems with Special Reference to Soils , " Soil Science ,
Vol. 75 , 1953 .
16 . MacEwan , D. M . C ., " Short - Range Electrical Forces Between Charged
Colloid Particles , ” Nature , Vol . 175 , 1954
.
17 . Martin , R . T. , “Water Vapor Sorption on Kaolinite : Entropy of Adsorp
tion , " 8th National Clay Conference , Norman , Oklahoma, October , 1959 .
18 . National Academy of Sciences , "Clays and Clay Minerals , " First through
Fifth National Conferences on Clays and Clay Minerals , 1953 - 1958 .
19 . Norrish , K . , " Swelling of Montmorillonite ," Faraday Society Discussion ,
No . 18 , 1954 .
20 . Roberts , J. E . and de Souza , J. M ., « The Compressibility of Sands , " Annu

al Meeting ASTM , Boston , Mass ., June , 1958 .
21 . Rosenqvist , I. T., " Soil - Water Systems,” ASCE Journal of Soil Mechanics ,
Vol . 85 , 1959 .
22 . Salas , C. E . and Serratosa , J . M ., “ Foundations Clays , ” Proc .
on Swelling
of the Fourth Inter . Conf . on Soil Mechanics and Foundation Engineering ,
Vol. 1, London , 1957 .
23 . Seed , , C . K., " Structure and Strength Characteristics of
H . B . and Chan
Compacted Clays , " Journal of the Soil Mechanics and Foundations Divi
sion , Proc . of the ASCE , October , 1959 .
24 . Taylor , D . W . , Fundamentals of Soil Mechanics , Wiley , 1948 .
25 . Terzaghi , Karl, « Simple Tests Determine Hydrostatic Uplift , ” Engineer

ing News - Record , June 18 , 1936 .
26 . Whitman , R . V. , " Some Consideration and Data Regarding the Shear

Strength of Clays , " ASCE Research Conference on Shear Strength of
Cohesive Soils , Boulder , Colorado , 1960 .
SOME CONSIDERATIONS AND DATA REGARDING THE SHEAR
STRENGTH OF CLAYS
By Robert V . Whitman ,1 M . ASCE
1. INTRODUCTION
This paper consists of four more - or - less independent sections . The next
two sections present certain general considerations regarding the meaning of
the effective stress principle and regarding the way in which shear strength
should be defined . These considerations have developed out of the attempt to
present the subject of shear strength to students in such a way that they may
discern the most important and useful concepts and yet have a clear picture
of the limitations to our current knowledge . It is pertinent to lay these con
cepts before the Conference , since it is imperative that terms be defined and
that differences in fact be separated from differences in viewpoint . The
fourth section presents data which have been obtained at M . I . T . using con
ventional strain - rates , particularly data obtained under the guidance of the
late D . W . Taylor and which have thus far been available only in reports with
a limited distribution . The fifth section represents an effort to pinpoint the
most important implications of the extensive program of compression tests
at rapid strain - rates conducted at M . I. T . and earlier at Harvard University .
II . GENERAL COMMENTS OF SHEAR STRENGTH BEHAVIOR
Notation . -- The letter symbols adopted for use in this paper are defined
and arranged alphabetically , for convenience of reference , in Appendix I .
Several key thoughts form the basis for the material presented in this
section . They are as follows .
(a )Effective stress is by far the most important factor determining the

shear strength of cohesive soils .
(b ) Many factors in addition to effective stress also influence the shear
strength of cohesive soils to some extent .
( c ) It may prove possible to simplify this long list of additional factors ;
that is , to show that just a few primary factors exert a direct influence on
strength , while the secondary factors have an indirect effect . Herein lies the
hope of progress in the understanding of strength behavior .
(d) For laboratory research and for important earthwork and foundation
projects , the behavior of soil is best studied through use of a limited number
of carefully conducted triaxial tests in which special care is given to the
measurement of pore pressures .
( e ) For practical problems in which expediency must govern and which
also fulfill certain special conditions, it may be possible to determine shear
strength without haying to evaluate the effective stresses .
1. Assoc . Prof. of Soil Engrg . , Massachusetts Institute of Technology ,

Cambridge , Mass .
581
A. General Statement of Beliefs
The author believes in the effective stress principle for the shear strength
of soils . That is to say , if he were asked to evaluate the shear resistance
available at a particular point within a mass of soil , the first thoughts would
be to : ( 1) Estimate the effective stresses which exist at that point today , and
also the changes in effective stresses which might be expected in the future ,
and (2) determine , by the simplest means possible , the relationship between
effective stress and shear resistance . Having done so , the author would feel
that he could provide a sound first approximation to the problem which had
been presented to him , an answer which might well suffice for a great many
engineering projects . Thus , the effective stress principle is the most power
ful tool available to the engineer who must work with the shear strength of
soils .
At the same time, it must also be realized that : ( 1) There is not a unique
relationship between effective stress and shear resistance , and ( 2) in many
problems , it is difficult to predict with any high degree of certainty the chang
es in effective stress which may occur in the future . These two statements
in effect represent limitations upon the effective stress principle . The first
statement implies that factors other than effective stress must be taken into
consideration whenever an accurate evaluation of shear strength is required .
Thus , it is necessary to be very careful in the formulation of the effective
stress principle so that the exact meaning of the principle and its limitations
are made clear . The second statement represents a shortcoming in the
ability of engineers to apply the effective stress principle : that is , other
knowledge which is necessary before the principle can be applied is generally
not available . This situation has led to the development of an alternate ap
proach to shear strength evaluation , which can be called the undrained
strength principle .
B. First Formulation of the Effective Stress Principle
The simplest and most useful formulation of the effective stress principle
is :
õtf
TH = f(
)
(1
typical soil general the )
Fig
of
The form this relation for shown

is
In
, in
,
1
a
.
.
curve representing Eq will have slight curvature although with some

1
(
)
.
particular soil conditions the relation appears be precisely straight line

to
all cases the actual function can for engineering purposes be approximat
ed In
by straight line over whatever range

of
of
Otf values are interest

a
.
Tu
Off tan
+
0
=
(2
)
c
The values and used describe this straight line approximation will
of
to
6
c
of
depend upon the range õtf values selected and they will course also de
of
pend upon the composition the soil upon the proportion and type
of
of
;
e
i.
.
granular and colloidal particles

.
double subscript notation has been adopted because the shear and
A
point vary both with direction and with time

at
normal stresses acting

a
Since this paper concerned with shear strength

is
concerned with the

is
it
,
stresses which exist at failure and the second subscript the letter
,
,
"
f"
stands for failure The yariation stresses with direction

is
of
described
"
"
.
CLAYS 583
by the Mohr stress circle , as indicated in Fig . 2 . The particular stresses of

interest here are the shear and normal stresses at the point of tangency of
the envelope to all circles representing possible states of stress at failure .
This curve frequently is called the Mohr envelope . The first subscript , the
letter " t" , stands for " plane of tangency " . Eq . ( 1) thus describes the Mohr
envelope to the circles representing the states of stress at failure , and the
following definitions can now be stated :
Tuf: shear stress on the plane of tangency of the Mohr envelope at the time
of failure .
or : effective normal stress on the plane of tangency of the Mohr envelope
at the time of failure .
As will be discussed in Section Eq not the only possible definition
is
III
,
1
.
(
)
of
the relation between shear strength and effective stress and hence the use
,
of double subscript system
a
The quantity Thi has been defined adequately
of
for the notion shear stress
,
had had long usage applied mechanics The quantity ötf must be defined
in
a
.
by
further the following equation

:
otf
uf
-
õtf
3
(
)
=
of -
the plane
at
tangency Mohr envelope
on
total normal stress

of
Orfi the
time failure
of
pore pressure
of
the time failure

at
uf
.
:
Otf
The total stress the same stress which encountered applied me

is
is
in
properly designed stress gage inserted

by
chanics and can be measured
in
,
the soil at the proper orientation The pore pressure up the fluid pressure
is
.
as the soil The effective stress Ötf cannot be directly measured

by
any
.
system that say any measuring system would have be based upon the
is
to
to
;
is ,
assumption that Otf as defined by Eq can be shown that Örf

is
It
in
3
(
)
.
some way related the magnitude of the forces within the mineral skeleton
to
the soil that öff certainly has some physical significance but
of
is
is
it
;
of ,
really quite difficult present precise physical description this quantity

to
; a
see Lambe and Whitman 1959 and Lambe 1960

.2
,
,
)
(
More Complete Formulation the Effective Stress Principle

of
C
.
Eq
Otf
implies unique relationship between Thi and Actually this

,
)1
a
(
.
relationship far from unique and many other factors into the picture
is
enter
,
more complete formulation might

be
A
72f
Ttf
flot
,
ef
,
=
,
,
,
,
,
,
,
H
T
(4
E
C
S
)
F
)
at
void ratio failure

ef
:
time
T
:
stress history
H
:
structure flocculated dispersed

or
,
0
Items indicated thus Lambe 1960 refer corresponding references

to
,
),
2
.
arranged alphabetically Appendix

in
II
.
584
Tafa
Straight line approximation
Off
ton
C
+
=
Ifp
(f=
) Off
Ppp
Tata
(f=
off
)
Tff
within this ronge
of
Stress range interest
SHEAR STRENGTH
. 1
Fig
port
-
Envelope
off
and
Tff
. 3
CONFERENCE
Fig
Mohr stress
of
circle
failure
Fig
. 2
CLAYS 585
E : environmental conditions , such as nature of pore water , temperature,

etc .
s : degree of saturation
F : conditions at formation of soil

C : capillary tension
Özf : effective stress in direction normal to the plane of greatest shear
distortion
Fig . 3 illustrates that the effect of these additional factors is to give a band
rather than a simple line on the Ttf Vs Ott diagram . The width of the band
shown is believed to represent the extent of the uncertainty as it is under
stood today . With large values of otf , the band width is a relatively small
of
Tuf
fraction of the average small the range uncertainty ex

at
ofof
but
ofis
,
of
tremely important Table contains brief resume the effects asome
1
.
Eq obviously not useful for

of
these factors and the sources
is
the data
(4
)
.
.
analytic computations
of
but serves point out the complex nature shear
to
,
strength behavior
.
TABLE PARTIAL LIST OF FACTORS REGARDING EFFECTIVE

I.
-
STRESS STRENGTH RELATION
Effect Data
Stress history important some undisturbed clays

in in
drained HIRSCHFELD 1958

;
) (
)
:
not important normally consolidated

vs
undrained shear HENKEL 1958 remolded

;
(
.
clays some importance overconsolidated remolded clays

in
;
.
no
effect
in
TAYLOR 1955 and HIRSCHFELD 1958 undisturbed
oza
;
(
normally consolidated clay

HENKEL 1958 some effect with overconsolidated remolded clay
;
(
.
Structure KENNEY 1959 same for remolded and undisturbed clays
p
;
(
( )
by
HIRSCHFELD 1958 from undrained tests altered remolding

) p
;
)
by
SEED and CHAN 1959 structure imparted compaction affects

;
(
strength low strains but not at high strains

at
by
effects explained
at
Conditions formation BJERRUM 1954 and this paper Hvorslev

;
(
. )
parameters
Environmental factors LAMBE 1960 little effect except when cementing agents added
;
(
Time GIBSON and HENKEL 1954 decrease per log cycle

%
5
;
(
BJERRUM SIMONS and TORBLAA 1958 no effect drained shear

in
;
;
,
. (
effect undrained shear

in
Perhaps even more factors could be added Eq certainly there has

to
);
(4
.
tendency for the list parameters affecting shear strength grow

of
to
been
in
a
recent years As the list has grown there has inevitably been rise the
in
,
a
.
the engineer who must evaluate the shear resistance

of
level frustration
of
available practical problems Actually there believe

is
good reason
to
in
,
.
that all these additional factors are not independent

of
each other and that

of
may be possible express the effects brought about by all of these factors
to
it
just few
of
terms
of
the effects
in
Of all
the secondary factors the void ratio failure appears play
at
to
ef
,
an especially important role

ef
assumed for the moment that oth and

is
it
If
.
are the only factors which affect shear resistance then we would have
,
:
Ttf
õtf
(
)5
ef
,
=
f(
)
Certainly Ötf and are say once
some extent interrelated that of
of is
to
is , ef
to
,
;
has been specified fall
only within certain range
ef
can values On the
.
other hand there some independence between these two factors for more
,
,
single possible for any given The work Hvorslev 1938
āt
ef
of
than
if
a
)
and Bjerrum factors can be
of
of
1954 has shown how the effect number
a
(
)
explained through their effect upon
ef
of
Hvorslev showed that the effects
.
preconsolidation can be explained this way and Bjerrum showed that the
in
,
of
of
effect saturated soil the time
at
the water content
a in
formation can
a
similarly be explained this paper data will be present
of
later section
In
,
.
show the environmental conditions during consolidation clay affect
ed
of
to
a
the strength mobilized at given Ötf but that the effects could be explained
in ed
in a
,
terms the differences These observations suggest that the follow
of
ef
ing relation might exist .
:
(6
)
,
,
be ef
=
F
)
f(
possible
be
this were true would eliminate the factors and

) to
to
it
If
,
H
E
from Eq considerable simplification our concepts shear
of
with
to in
(4
It a
F
strength behavior perhaps asking too much hope that
af
is
and
H
is E
,
F
.
fect strength only through their effect upon Certainly there only scant
ef
.
evidence this regard today However would also seem that more at
in
it
,
.
tention should be paid this avenue of investigation Whenever found

to
is
it
.
that some factor affects the Tlf vs Otf relationship the first step should be
to
,
examine the effect upon use the method developed by Hyorslev

ef
and
to
to
,
determine the extent which the change ttf can be explained through
to
in
course necessary
ef
of
differences such efforts take into con

is
to
In
in
it
,
.
sideration the possibility that the void ratio may vary throughout sample at
the time that failure occurs a
.
Recent work regarding the effective stresses which exist within compacted
clay provides second example simplification the overall picture of
of
in
a
strength behavior had been suspected for years that the difference
It
the
in
.
Ötf relation for the same clay compacted dry

vs
of
Tlf were
or
wet optimum
the result of differences the capillary tensions the two conditions
in
in
If
.
these capillary tensions could be taken into account

of
the determination
in
uf was felt that the difference Otf would largely disappear Re

vs
Ttf
in
it
,
at .
cently this contention has been supported by excellent testing work Im

perial College
of
England and the Bureau Reclamation this country as

in
in
described other papers this conference These results suggest that

to
in
:
õt
)
(7
=
C
f(
)
removed from Eq
be
Thus appears
that another term can
of is it
,
4
.
(
.)
possible functional relationships along the

of
envision whole host

to
It
lines Eqs and For example where internal coherence between

,
(6
7
.
).
)
may be possible
ac
clay particles affects strength take this effect into

to
it
,
Eq
by
as
modifying the definition stress

of
count effective indicated

is
in
,
8
.
(
.)
OH
up
0t
(
)8
=
+
1
-
net interparticle attraction

I
:
CLAYS 587
Perhaps the effects of chemical and temperature environment can also be
taken into account in this way , and some of the effects of partial saturation
as well . This approach has received particular attention at M . I. T . in recent
years, and some of the results of the theoretical studies are discussed in the
companion paper by Lambe. Perhaps the effects of T and 72f can be explained
through their effect upon ef. These suggestions merely illustrate the possi
bilities might profitably be explored .
It is certainly true that the establishment of relationships such as those in
Eqs . (6 ) , ( 7) and (8 ) does not reduce the total number of variables which must
be considered . However , as the mechanism through which any factor affects
strength becomes more clearly defined , the easier it is to deal with this
factor . Delineation of these mechanisms aids the researcher , but also aids
the practicing engineer as well . For example , as soon as the work of
Hvorslev clearly demonstrated the importance of knowing ef , the effect of
sample disturbance upon the Ttf vs Õuf relation became much clearer . Today
the practicing engineer knows that : ( 1) With preconsolidated soils , the Tlf vs
Ötf will be affected by sample disturbance ; ( 2) in order to determine the
correct relation ; the stress history of a sample must be manipulated so as to
achieve the correct ef for any otf ; and ( 3) where it is not feasible to recreate
exactly the proper ef , the error can be estimated through the method of
Hvorslev . In a similar vein , the ability of the engineer to evaluate ttf for a
given Ōt in a compacted clay is much greater now that the role of capillary
tensions has been at least partially clarified . Thus the engineer can expect
to gain much from all investigations which seek to clarify the mechanisms
involved in shear resistance .
D. Undrained Strength Principle
The usefulness of the effective stress principle of course rests upon the
presumption that õtf can be determined in any problem . The values of õtf
which have existed in the past or which exist at the present at any point with
in a soil mass can be found from measurements . There are some situations
where the Otf which will exist in future time can be estimated with quite
reasonable accuracy , such as those situations in which loads are applied so
slowly that no excess pore pressures are generated , or the situations which
will exist a long time after load application when all excess pore pressures
have dissipated . In general , however , it is a very difficult matter to estimate
the pore pressures which will exist in future time and hence to estimate the
effective stresses .
Fig . 4 illustrates the situation which exists when a soil is sheared without
permitting drainage . The stresses which exist before the undrained shear
are otc and Ttc where the second subscript , the letter " c" , indicates that
these are the stresses " at the time of consolidation " . It is known that , if the
soil is sheared all the way to failure , the final stress conditions will fall on
the appropriate Tif vs Of curve . The trouble is , it is never completely
certain what path will be followed between the initial and final points nor just
where the final point will be . The path followed by the 7t and Tt values is
called the effective stress path .
The symbol dõt will be used to denote the amount which ot changes during
an undrained shear process . The for which will occur in any particular case
depends : ( 1) Upon certain conditions which existed before the application of
shear ; and ( 2 ) upon certain conditions which exist during the application of
shear .
The factors which fall into the first group can be represented by the
following relation :
Aõt = fb ſõlc , 73c , ec , H, 0 , E , S , F , C) (9 )

Most of the symbols appearing in Eq . ( 9 ) have been defined previously . The
subscript " c" used with ō and e means " at the time of consolidation " i . e . be
fore the undrained shear process commences . The subscripts 1 and 3 denote
major and minor principal stresses . The factors which fall into the second
group can be similarly represented .
Aõt = fp ( T, R , 401 , 402 , 403 )

( 10 )
A new symbol appears here : R : rotation of principal stress directions . The

presence of R , 101 , 102 and 403 in the functional relation emphasizes that
Sot depends in general upon the whole process of shear application . Table 2
gives a brief resume of the effects of some of the factors appearing in Eqs .
(9 ) and ( 10) .
As is the case with Eq . (4 ) , it would seem likely that many of the factors
appearing in Eq . ( 9) exert an indirect influence upon Aõt rather than a direct
influence . ec is certainly the most significant of the variables , as has been
shown by the work of Rutledge (1947 ) . It would seem likely that factors such
as H , E and F exert their main influence on strength through their influence
upon ec
It is known that Eq . (10 ) can be simplified considerably in the case of soft ,
saturated clays, for it has been found that 401 and 103 exert very little in
fluence upon the Aõt developed when the sample is sheared to failure . Thus
Eq . ( 10) simplifies to :
Aõt = fp ( T, R , 402 ) (11 )
There are some data regarding the effect of 402 , but the data regarding the
effects of the other factors are scarce . It has been common for some time to
make the simplifying assumption that fp = 0; that is to say , that Aõt is de
termined solely by the conditions which exist prior to the undrained shear
process and is totally independent of the circumstances of the shear process .
Tuf
de
be
Since depends upon õtc and Aõt both which are now assumed
to
of
,
termined by the conditions which exist

at
of
the start the undrained shear

process the implication that ttf depends entirely upon the conditions which
is
,
exist at the start shear process

of
the undrained
.
THE
UNDRAINED FUNCTION OF CONDITIONS BEFORE UNDRAINED

=
an 12
SHEAR
, (
)
repeat the reasoning Ttf depends upon Ötf

To
depends upon ötc

ot
;
) (3 (1
2
) )
(
)
:
initial condition and Aõt under some circumstances Aõt also

is
function
(4 ;
a
,
entirely
of
the initial conditions hence under these same conditions

is
. Tt
,
;
the initial conditions The principle embodied Eq

of
12
function has
in
a
greatest meaning soft saturated clays but has general validity

of
the case
in
all saturated soils as shown by Bishop and Eldin 1950

to
)
.
Eq has great practical significance necessary

12
is
know the
to
it
If
.
.
(
)
be
of
undrained shear resistance saturated soil this resistance can evalu

,
a
by
ated simply shearing the soil its natural state by

an
undrained shear
in
process Any undrained shear process will suffice regardless

of
what
,
.
CLAYS 589
pattern of 401, 402 or 403 are imposed . The vane shear device can be used ;
the unconfined compression test likewise suffices . Thus the undrained
strength can be evaluated without the need of triaxial testing equipment or
without having to measure pore pressures during the shear process .
While the undrained strength principle thus has great practical value , it is
very important to realize the limitations of this principle . First and fore
most , of course , the principle does not provide a direct answer in those
practical problems where the water content of the soil changes as shear loads
are applied . Even in the case of soft , saturated clays, sheared with no drain
age , the principle is only approximate , for it has been found that the factors
indicated in Eq . ( 11 ) do have some effect upon the shear resistance . Unless
the soil is saturated , other factors such as those appearing in Eq . ( 10) must
be considered . In a sense it can be claimed that the undrained strength
principle still applies to unsaturated soils , but as soon as it is necessary to
be very careful about the manner in which the shear load is applied in tests
which are to determine the undrained strength , the principle loses its greatest
virtue : its simplicity .
The greatest criticism of the use of the undrained strength principle is
this : Using this principle means that a clear understanding of what is going
on within the soil is sacrificed for the sake of expediency . The greatest in
sight into any stability problem comes from a clear picture of the time vari
ation of the effective stresses which do and will exist within the soil . There
are many problems where the expediency of direct undrained strength
mea surements coupled with liberal safety factors may represent the proper
engineering course . However , in any problem where the cost of earthwork
or foundation treatment is significant , it becomes imperative to supplement
undrained strength measurements with tests which provide a clear picture of
the way in which the pore pressures within the soil mass will change with
time .
III
DEFINITION OF FAILURE
.
of
strength behavior
of
the understanding
of
Some the current confusion

in
cohesive soils arises from the fact that different research workers have de
fined strength different ways There appear be two separate problems
to
in
Of all the shear stresses which exist soil sample at the time
of
in
1
a
(
)
failure which stress be taken as the shear strength and what stage
is
to
2
;
,
(
)
the shear process represents failure

in
Choice Shear Stress Represent Failure

of
to
A
.
This problem illustrated Fig Assuming that the circle shown

in
is
5
.
.
gives the state triaxial sample the time when that sample
at
stress
of
in
a
fails there still remains the problem deciding which all the shear
of
of
,
stresses be called the shear strength Four possibilities are suggested

is
to
the figure
in
Tf defined previously
:
at
the failure plane

on on
Tff shear stress failure

:
TBf shear stress the plane maximum obliquity failure

at
of
:
Tmf shear stress failure

at
maximum
:
.
590
Ich
Iff
=
)
floff
Tpf
us
Off
Iff
,
Possible oft
Possible effective
stress paths Effective
stress
path
SHEAR
etc
oro
&
,
Possible at
. 6
Fig
. 4
Fig
er
Ay
Curve
Mohr
Foilure
8
envelope Curve
STRENGTH CONFERENCE
Poth
i
tra for
stage
looding
Effective stress
path for continuous
os
-
determined undrained shear

from observed
slip planes
7
5
.
.
Fig Fig
CLAYS 591
By failure plane is meant the plane of maximum shear distortion . It must
be remarked that it is not an easy matter to establish just where in a failed
sample this plane is located . The plane of maximum obliquity is the plane on
which the ratio of shear stress to normal effective stress is the greatest, and
is the point of tangency of a straight line through the origin .
The differences between Tt , Tff and TBf are seldom great, and usually all
points would fall within the range of uncertainty in the data . For that matter ,
the difference between any of these stresses and Tmf may be of no practical
significance . However , Gibson (1952 ) has shown that it is possible to dis
tinguish Trf and Tff . Taylor in his work insisted that if the failure plane could
be identified , then Tif should be used as the shear strength and not Tif . Faced
with conflicting suggestions as to how failure should be defined , it is hardly
satisfying even to the practicing engineer to avoid the question on the basis
that the differences are small .
The choice among Ttf, Tff, TBf and Tmf in research work is essentially a
matter of which theory of failure is believed to be correct for the soil in
question . The author has no suggestions to make in this regard , except that
until the matter is clarified , it would be unfortunate to inhibit thinking with
regard to shear strength by insisting that one of these quantities be adopted
as standard . Thus , in the research of M . I. T. , the author and his colleagues
are likely to plot up the data in whatever way seems to give the clearest
picture of shear strength behavior in the particular problem at hand . Using
the double subscript notation , it can at least be made clear what definition of
strength has been used .
On the other hand , when shear strength data is to be used in connection
with conventional methods of stability analysis , then it appears that there is
one definition of shear strength which is clearly correct . In discussing the
use of the undrained strength , Skempton ( 1948 ) pointed out that the convention
al method of analysis for the stability of an unsupported vertical cut predicts
incorrectly the surface along which slip will take place , but does give the
correct safety factor if Tmf is taken as the shear strength . In effect , the
conventional analysis assumed the " wrong ” plane on which to analyze stress
conditions , and also uses the " wrong ” definition of shear strength (wrong be
cause there is some plane on which the shear resistance , is less than Tmf),
and yet these two wrongs cancel to make a right as far as the safety factor is
concerned . The contentions of Skempton can be verified by theoretical analy
sis and are also backed up by analyses of actual failures . To the author ' s
knowledge , there is no theoretical proof for other than vertical cuts , although
there is some empirical evidence . For these reasons , it is most meaningful
for practical work to choose Tmf , which of course is one - half of the deviator
stress , to represent the undrained shear strength of a soil . The discussion
in section II D was in terms of Th , which is a useful way to express undrained
strength for the purpose of fundamental studies . If th is known , Tmf can of
course be found , and so the dual usage need cause no confusion .
A similar situation exists with regard to the use of the effective stress
strength relation . For the case of vertical cuts , it can be shown theoretically
that the conventional method of analysis predicts incorrectly the surface
along which slip will take place , but does give the correct safety factor if the
Ttf vs Otf relationship is used . Again , the author knows of no proof for cases
other than vertical cuts . Thus , there are sound reasons for the conventional
practice of expressing the effective stress strength relation in terms of the
Mohr envelope .
In summary , the main points of this subsection are :
(a ) When studying shear strength in the abstract without applying it to sta
bility analysis , there is at this time no logical basis for a choice among
Tlf , Tff , TBf or Tmf to represent shear strength . Soil engineers have
long assumed that the Mohr - Coulomb theory was the correct theory ,
but there is ample evidence to the contrary .
(b ) Whenever shear strength data is used in connection with one of the con
ventional methods of stability analysis , then there is a single correct
definition . When dealing with undrained strength , Tmf should be used ;
when dealing with shear strength in terms of effective stress , ttf ( i . e.
the Mohr envelope ) should be used . The conventional analyses assume
that the Mohr - Coulomb theory is correct , and it is logically incon
sistent to use in these methods data which have been interpreted by
some other theory .
B. Stage in Shear Process which Represents Failure

The second problem of this section concerns identification of the stage in
the shear process at which a soil is said to be failed . This is a very complex
question , and involves such problems as progressive action . It seems likely
that there is no general answer to this question , and so the following dis
cussion will be confined to several specific examples .
Suppose that the objective of an undrained test is to determine the un
drained shear strength which will exist in an actual practical problem . Pore
pressures are measured during the test , and the effective stress path is as
shown in Fig . 6. Now the question arises ; is Tmf to be taken as ordinate A or
ordinate B ? If it is really believed that the shear test represents the con
ditions which can be expected in the actual problem , then it seems that it is
logical to select ordinate B to be the shear strength : i . e. to define the peak
point of the effective stress path as representing the time at which the sample
is said to have failed . Perhaps it can be argued that it is conservative to use
ordinate A , and that provision can in this way be made for some of the un
certainties in the accuracy of the test conditions . Such an argument seems
to be dodging the issue , however .
But now suppose that the undrained test is being run for the purpose of de
termining the Ttf vs ötf relationship . Such a procedure might be adopted if
an embankment were to be built in stages upon a soft clay . The use of un
drained tests not only give the Tif vs Otf relationship , but gives the data which
are useful in predicting the pore pressures which will be set up in the clay at
various stages of construction . The curves of Fig . 6 have been redrawn in
Fig . 7, and again the question is : What point along the effective stress path
represents failure ? Again only one answer appears to be logical : Through
the end point as shown in Fig . 7 by curve A . This ambiguity would not arise
if the laboratory sample were subjected to cycles of shearing and consoli
dation as would actually occur in the embankment foundation , for then any
instability in the mineral skeleton would be destroyed by consolidation under
a stress system involving considerable shear stress , and the effective stress
path would reach its maximum point at the same time that it reaches the
Ttf vs ētf curve . There may be some question as to whether the undrained
test is a proper way to evaluate the Tif vs Otf relationship ( see Casagrande
and Wilson , 1953 ) for such a problem , but the statements above concerning
the way in which the failure point should be defined are not invalidated by
this question .
CLAYS 593
As a third example , again ,
suppose that the undrained test is being run for
the purpose of determining the Tlf vs Otf relationship , but that now the
problem to be solved is as shown in Fig . 8. Here the problem is one of
failure under constant total stresses , the effective stresses along the surface
of sliding being reduced by excess pore pressures migrating outwards from
under the newly constructed embankment . In Fig . 8, the effective stress path
which might be expected in the field situation is superimposed upon the ef
fective stress path from an undrained laboratory test . Although this case is
far from clear , it seems likely that failure will occur when the Ttf vs ētf re
lationship indicated by curve B is reached ; i . e. for such problems , the peak
points of the effective stress curve represent the failure condition . Such a
definition of failure leads , with some soils , to extremely low values of the
coefficient o in Eq. (2 ) .
These few examples illustrate that the stage of a shear process which
represents failure must be defined to fit each problem . The examples cited
are fortunately extreme , arising with sensitive clays and very loose sands .
These same considerations apply to fundamental strength studies in the labo
ratory . For example , suppose that the effect of rate - of - strain upon undrained
strength is being investigated . Fig . 9 shows a typical effective stress path
for a soft , saturated clay sheared undrained with a conventional rate - of
strain . The figure also shows an effective stress path which might be ob
tained at a slower strain - rate with some membrane leakage . If the Ttf vs Off
relationship were to be drawn through the peak points of the effective stress
paths, test results such as these would suggest that the coefficient o in Eq . (2)
is a function of strain - rate . These circumstances would be expected to lead
to a smaller strain at failure in the slow tests as compared to the tests at
conventional speeds . Thus , where membrane leakage is possible , undrained
tests may not provide a satisfactory means for determining the effect of
strain - rate upon the effective stress strength relationship . The r ?nts
suggested in this paragraph may provide a possible explanation foi che re
sults reported by Bjerrum , Simons and Torblaa ( 1958 ) .
IV . SOME RESULTS FROM TESTS AT CONVENTIONAL STRAIN RATES
The data which have been selected for presentation in this section shed
light on the following questions .
(a ) How much do the test conditions affect the undrained shear strength of
a soil ?
( b ) How unique are the effective stress paths obtained during undrained
shear ?
( c ) Is the consolidated undrained test with pore pressure measurements a
satisfactory means for establishing the effective stress strength re
lationship ?
( d) Does pore water migration occur during undrained shear ?
( e) How do environmental conditions at the time when a clay deposit is
formed affect its strength characteristics ?
A. Factors Affecting the Undrained Shear Strength of Boston Clay
The work described in this subsection was performed under the guidance
of the late D . W . Taylor in the years between 1949 and 1955 . The results are
New embankment
constructed
Existing - Excess pore

Slope pressures here
migrate laterally
and decrease stability
of existing slope
(0)
(b ) Curve A
Curve B
Effective
stress path
from laboratory
Poth in test
problem
Failure above
Fig . 8
for
off
Tff us
conventional speed
Tapparent
for
Off
Tiff
vs
slow speed
Effective
for
-
Path
for
-
slow speed stress path

-
conventional
-
speed
-
-
Fig
9
.
CLAYS 595
described in detail in a series of reports prepared by the M . I . T. Staff (M . I. T .

1949 , 1950 , 1951 , 1953 ) and summarized by Taylor (1955 ) . These studies
were aimed at finding out the degree of uncertainty involved in the use of the
undrained strength principle .
The testing program included both triaxial and direct shear tests , all run
under undrained shear conditions. The triaxial testing equipment and the
pore pressure measuring systems which were in use at this time have been
described by Lambe ( 1951 ) . Unless otherwise specified , pore pressures were
always measured in the central portion of the sample . From the reports , it
is apparent that Taylor was quite concerned by the question of time - lags in
the pore pressure measuring systems , and that the technique of suddenly in
creasing the chamber pressure and noting the response of the pore pressure
measuring system was used to check the performance of the measuring
system . Generally , with each test discussed in the report , there is a notation
telling the duration of the time - lag which was measured . Those of the M . I. T .
staff who worked with Taylor during this period recall that he was never
satisfied with the validity of test results unless the pore pressure rose to
better than 95 per cent of the applied chamber pressure change within one or
two minutes after the change occurred . Taylor worked out a technique for
performing direct shear tests in such a way that the chances of partial drain
age during the test were minimized . This technique involved varying the
normal load upon a sample in such a way that the thickness of the sample re
mained unchanged during the application of shear stresses .
Undisturbed Boston clay was used in these tests . Unless otherwise noted ,
the clay was consolidated to an all - around pressure of 110 lb / in2 , a stress
well above any stresses which had ever acted upon the clay in its natural
state , and then sheared . Triaxial samples were 2 . 8 inches in diameter and
at least 6 . 5 inches long . Direct shear samples were 3 inches square and 1/ 2
inch thick .
Fig . 10 shows some of the results obtained with triaxial compression tests .
Taylor preferred to plot his results in terms of the effective normal and
shear stresses which existed on the plane which at failure was the plane at
maximum obliquity , and this form of plotting has been preserved in Fig . 10.
Curves A and B represent , respectively the average of a number of effective
stress paths obtained from tests in which the time - lags were questionable
and from tests in which the time - lags were less than two minutes . The rela
tive position of these two curves is as would be expected . During the early
stages of the test , when both the shear stress and the excess pore pressure
were rising rapidly , the recorded pore pressure tended to be too low if there
were time - lags in the pore pressure measuring system , and the effective
stresses were hence too high . As failure was approached , however , the rate
of increase of the shear stress and the excess pore pressure slowed , and
there was an opportunity for the measuring system to catch up with and
measure the pore pressures actually existing within a sample .
Taylor was interested in the effects of sampling and laboratory handling
upon the undrained strength and effective stress path which might be obtained
from consolidated undrained tests . Hence , he subjected samples to varying
numbers of cycles of stress application ; that is to say , a small amount of
stress would first be applied in compression and then the stress would be re
moved . There were a few tests with stress reversals . Curve C represents
the average of several effective stress paths obtained when samples were
subjected to numerous cycles of stress and then sheared in one continuous
596
?
in
/.
Ib
,
origin
A
Curve
sample
through
30°
of
line Continuous straining
of slope
"
to"
failure questionable
at
time lags
axis
Curve
to
to
Final puints fell Continuous straining failure
"
30°
within this zone very good time lags
-
"-
B -
SHEAR STRENGTH
--
--
TWI
SHITET
FETTET
inclined
C
Curve LITTLE
–
to
-
Continuous straining failure
reversals
plone
following many strain
-.
LITLĪTOTITOM
on
CONFERENCE
D
Curve
-
Points during strain reversals

120
110
100
80
60
50
30
20
10
stress Shear
in
./. 90 ?
1b
70 of
,
to
sample
40 on
Effective stress plane inclined 30° axis
10
.
Fig
CLAYS 597
operation to failure . The stress cycles set up excess pore pressures and
eventually the effective stress decreased to the vicinity of 80 lbs / in2 . How
ever , when the sample was then sheared to failure , the effective stress paths
rose up and met the path obtained with undisturbed samples and yielded the
same undrained strength as in those samples . Curve D shows the location of
the effective stress paths obtained when samples were sheared to failure after
only a few stress cycles . These results demonstrate two points : (a ) The un
drained shear strength of Boston clay as measured in consolidated undrained
tests is insensitive to moderate amounts of disturbance as long as no further
consolidation is allowed ; and (b ) the location of the early portion of the ef
fective stress path is quite sensitive to time - lags in the pore pressure
measuring system and to disturbance of the sample after consolidation .
The end points (the points at which the obliquity ratio Ter / Obf was a maxi
mum ) of all effective stress paths fell within the zone indicated in Fig . 10 .
The location within this zone of the end point of any particular curve seemed
to bear no relation to whether the time - lag was " good " or " questionable " , or
to the stress cycle treatment which had been applied to the sample . If the
purpose of these undrained tests had been to determine the effective stress
strength relationship , it is clear that there would be a fair amount of un
certainty as to the exact location of the curve representing this relation . Note
that, on the diagram as drawn , the effective stress strength relation would be
expressed as TBf vs Obf . For normally consolidated Boston clay, there is es
sentially no difference between the plane of maximum obliquity , the plane of
failure and the plane of tangency . A line at a slope of 30° has been drawn in
Fig . 10 , since 30° is the most commonly accepted value for the angle of
drained shearing resistance in the Boston clay . These results illustrate the
point that the effective stress strength relation is not a definite line , but
rather that its location depends upon details in the makeup of a particular soil
sample and in the test procedure .
The following paragraphs summarize some of Taylor ' s other conclusions
concerning the factors which affect the consolidated undrained strength of
Boston clay .
( a ) The consolidated undrained strength obtained in axial compression was
found to be 10 per cent greater , for the same consolidation pressure ,
than the consolidated undrained strength in axial extension . From the
pore pressure measurements , it was clear that the difference in
strength was due to the differences in the pore pressures generated
during the two types of tests . These results have since been confirmed
by the very careful tests of Hirschfeld ( 1958 ) .
(b ) Using direct shear tests , the effect of the time of consolidation was
studied . Approximately 40 minutes was required to give essentially
100 per cent consolidation . Taking the undrained strength with this
consolidation condition as the norm , the following results were found .
Time of consolidation Increase over norm
1 day 16%
7 days 21 %
(c) Triaxial compression tests and a relaxation procedure were used to

study the effect of strain - rate on strength . It was found that decreasing
the strain - rate from 0 . 17 per cent strain per minute to 0. 001 per cent
per minute decreased the consolidated undrained strength by 12 per
cent .
(d ) Two tests were made upon samples consolidated anisotropically . The
results , compared to those for isotropic consolidation , were as follows
(all data in lb / in2 );
3C
Eml Ozc/
me
106
106
40
151 268
0
.4
.
150
41
150 276
0
3
.
.
35
76
119 298
0
.5
.
28
108 108 266
0
.
.
the last line represents
of
average
an
The data five tests For the Boston
in
.
by
clay the Tmf õlc ratio apparently not decreased the particular degree
is
,
anisotropic consolidation used

of
these few tests

in
.
The purpose
of
of
presenting these data illustrate the magnitude is the
uncertainties which one faces when using the undrained strength principle to
.
Items reflect uncertainties involved when one tries measure
to
and
in
(a
(c
)
situ strength using vane device or by using unconsolidated undrained tests

a
.
per cent would present even before the
of
10
be
on
Uncertainties the order

problem sampling disturbance
of of
on
considered Items and bear the
is
(b
(d
)
)
in 20 to .
problem using consolidated undrained tests estimate undrained strengths

field problems per cent are present when
of
available Uncertainties
in
. .
practice As was indicated

is
such followed Table there are other

a
II,
uncertainty as well
of
causes
.
TABLE PARTIAL LIST OF FACTORS AFFECTING UNDRAINED

II
.
-
STRENGTH
Erfect Data
Time TAYLOR 1955 effectconsolidation and
of
of
time
of
rate
;
(
of shearing summarized this paper

( in
,
moderate loss strength

of
BJERRUM SIMONS and TORBLAA 1958

,
;
)
soft undisturbed clay

in
) .
up
CASAGRANDEand WILSON 1951 strength loss

to
;
2
/3
(
when some clays sheared slowly

ö5c
BISHOP and HENKEL ratio Telon

if
1953 less
;
olcs
(
Oczc
>
1
this paper
Бор
TAYLOR 1955 summarized

in
;
(
Ag
10
20
17
strength Ag
%
HIRSCHFELD 1958 less

=
;
,
(
Structure SEED and CHAN 1959 soils having same water content and
;
(
volume have different strengths with different structures

.
by
Environmental factors LAMBE 1960 strength altered ion concentration

;
(
.
CLAYS 599
Many other ideas and data are discussed in Taylor ' s reports . Only those
data which can be presented with sufficient clarity in a limited space have
been included in this paper .
B. Migration of Pore Water during Undrained Shear
A thesis by Clough ( 1951 ) describes a series of tests which studied the mi

gration of pore water away from the central zone of a clay sample undergoing
undrained shear . These tests involved both direct measurement of the change
in water content in the various portions of the clay samples , and also
measurement of the pore pressure gradients between the center and ends of
the samples . There were only a half - dozen tests involved in this program ,
which was obviously exploratory in nature . 2. 8 inch diameter , undisturbed
samples of Boston clay were used , and in each case the samples were consoli
dated to 110 lb / in2 before testing .
It is not as easy matter to establish the water content patterns before and
after shear , particularly since the water content of Boston clay quite gener
ally will vary several per cent within a matter of an inch . To overcome this
problem , two samples were trimmed from immediately adjacent positions
within a block of clay and then consolidated under absolutely identical con
ditions . One sample was then cut up into slices for determination of the water
content pattern before shear , while the second sample was sheared to failure
undrained in about one hour and then similarly cut into slices for water
content determinations . Actually , it was impossible to know how much of the
water content change occurred during the shear process and how much oc
curred while the apparatus was being dismantled . The change in water
content due to shear was then taken as the difference in water content of
corresponding slices in the two samples . The results obtained for the sever
al pairs of samples have all been plotted in Fig . 11 . Considering the possible
sources of error with such a technique , the consistency of the data is really
quite remarkable .
The results of measuring , simultaneously , the pore pressures at the center
and one end of samples sheared at various rates have been summarized in
Fig . 12 . There was only one test of the faster strain - rate , and one test which
started out at the standard strain - rate and then was switched to the slow
strain - rate . There were several tests using the standard strain - rate , but
using samples of different lengths . The tendency of the end pore pressure to
lag behind the center pore pressure with the faster strain - rates is evident.
The results for the several samples of varying length but same overall
strain - rate have not been plotted , but the lag was found to be greater as the
sample length was increased . The tendency for the center pore pressure to
exceed the end pore pressure is of course consistent with the evidence that
water migrated from the center of the sample toward the ends .
It is impossible from these tests to conclude much concerning the effect
of strain - rate upon the pore pressures in the central part of the sample or
upon the strength , since any differences were within the expected data scatter .
Certainly there was no marked tendency for strain - rate to affect these quanti
ties . Thus , moderate changes in the strain - rate used for undrained shear
affects the accuracy of pore pressure measurements made at the ends of the
sample , but may have little effect upon the strength determination . The latter
point certainly requires much additional study , however .
1
--
L-
D
length
O + 4. 35
,
of
%
above base
D
Height
ŞIroo
Change Water Content

%
in
,
11
Fig
.
As this paper being written new study the effect strain rate upon
of
. of
is
,
a
pore pressures during undrained shear underway new pore

at
is
A
T
.I
.
pressure measuring system being developed which will use an electrical

is
pressure transducer located within the triaxial cell Preliminary tests with
.
clay samples has shown that overall time lags as short as second can be
1
2
/
-
paper describing the early stages this work will appear

of
obtained the
in
A
.
Proceedings Soil Mechanics and Foun

on
the 1961 International Conference

of
dation Engineering
.
Hvorslev Parameters for Backswamp Clay

C
.
The tests reported herein are described detail report prepared for
in
in
a
the Waterways Experiment Station MIT 1959 The soil was river back
a
,
(
).
swamp clay from the vicinity Vicksburg Mississippi the origin

of
of
which
,
of
liquid limit
70
described Kolb and Shockley The clay has

is
1957
in
a
,
a
(
).
plastic limit
on
and classifies as CH the Unified Classification system

of
45 30
,
Approximately per cent by weight smaller than microns and the clay
is
fraction largely illite and montmorillonite equal parts Organic matter

is is
in
content per cent by weight

1
slurry large consolidometer

of
the clay was first consolidated

to
in
A
approximately samples with

an
kg
10
cm2 and area cm2 and

of
then
,
1
a
/
4
of
-
Center Average tests
804
of
3
-
-
End Ratel Average tests
Bottomy
A
Center
- 1
2
-
-
End Rate test
?
in
. /.
in 1b
/
. ,
%
16
.0
Ib ihr Rate changed from
,
to
%
.0
016 per minute
:
Note
Pressure
16
Rotel per min
..
%%
2
.0:.0:
Rate 67 per min
CLAYS
Pore
11
Measured
Pressure Pore Measured

|
|
|
|
12
11
II
10
II
9
8
6
5
4
3
1
2
12
11
10
|9
8
4
5
3
2
7% ,
109
Strain
7%
,
Strain
12
.
Fig
length of 8 cm were cut for testing . The triaxial equipment and pore pressure
mea suring system developed by the Norwegian Geotechnical Institute were
used . Filter paper strips were placed along the sides of the samples to aid
in consolidation of the samples and in the pore pressure measurements .
After further consolidation to the desired pressure , the samples were back
pressured to approximately 0. 5 kg / cm2 . Then the response of the pore
pressure measuring system was checked by increasing the chamber pressure
without allowing the sample to drain , and observing the pore pressure re
sponse . The time - lags were generally on the order of 2 to 10 minutes, but
the final level of pore pressure response frequently fell well below the
amount of chamber pressure change . This type of behavior suggests the
presence of air within the sample . Hence , all but two of the tests were run
with constant pore pressure , varying the chamber pressure so as to maintain
a constant volume condition . A strain - rate of approximately 1 per cent strain
per hour was used , and the peak deviator stress was reached in from 4 to 9
hours . Piston friction proved troublesome in the tests , although it is believed
that this friction was largely eliminated by rotating the loading piston . The
standard area - ratio correction was applied . The program involved tests on
normally consolidated samples using pressures in the range between 2 and
8 kg / cm2 , and tests upon samples which were first consolidated to 8 kg / cm2
and then reconsolidated to various pressures in the range from 1 / 4 to 6 kg /
cm .
Fig .
13 shows the relationship between the consolidation pressure and the
water content of the samples . Three large chunks of the soil were consoli
dated , here labeled batch 2, batch 3 and batch 4. Batches 2 and 3 proved to
have substantially identical characteristics , but batch 4 showed a different
curve . In retrospect , it is known that the ambient temperature was much
higher while batch 4 was being consolidated than was the case with batches 2
and 3, but other unknown factors may also have influenced this result .
For the purpose of these tests , failure was defined as the time at which the
maximum deviator stress developed in the sample . With this soil, this is
also the time at which the maximum obliquity ratio was reached . Fig . 14
shows Mohr circles for the failure state , with the normally consolidated tests
represented by solid circles and the over - consolidated samples by da shed
circles . For the normally consolidated samples , the coefficients for Eq. (2 )
are found to be c = 0 and 0 = 22º . Fig . 15 (a ) is a plot of Tmf vs 03c , and hence
shows the effect of initial conditions upon the undrained strength . Since the
consolidation process involved no shear , 73c = Õlc = õfc , etc . Fig . 15 (b ) is a
plot of Tff vs Off , where the failure plane is assumed to make an angle of 33°
to the axis of the sample . In both parts of Fig . 15 , the points representing
batches 2 and 3 have been distinguished from those representing batch 4 , and
it is seen that there is a general tendency for the tests upon samples trimmed
from batch 4 to show a lower strength for a given 03c or Off .
The method of plotting suggested by Bishop and Henkel ( 1957 ) has been
used in preparing Fig . 16 . The quantity Pe is the equivalent consolidation
pressure , and is found for any sample by entering the virgin branch of the
consolidation curve for the soil with the measured ef. The Hvorslev shear
parameters can be calculated from the geometry of Fig . 16 , and are found to
be : e = 16° and K = 0 . 08 . A problem arose in the preparation of Fig . 16 ow
ing to the fact that there was not a unique consolidation curve. It can be
shown that , in order to retain the intent of Hvorslev , some one consolidation
curve must be taken as the basic curve . In this work , the curve obtained
CLAYS 603
,
% -Botches 2, 3
Content
Water
Batch
4
1
Consolidation Pressure
Kg cm2
/
Fig
13
.
for
Envelope normally
consolidated soil
/,
Fm
cme
stubos mo
,
o
c
Kg
for
Envelope over
consolidated soil
-
Failure
il
of
Stress
L
L
Sheor
Kg cm2
of
Effective Stress Foilure

,
/
Fig
14
.
/)
o Batches 2, 3 Overconsolidated
• Botch 4 to 8 Kg / cm
cm2
-(ż
Kg
Normally
031 consolidated
0f
/Kg
Ofic
Normal Effective Stress cm2
Batches
o •
2,
3
Batch
4
Failure Plane
/,
cm2
on Kg
Shear Stress
Failure
(0
ot
Kg cm2
on
of
Effective Stress Failure Plane Failure

,
Fig
15
.
40 Batches
2
,
3
o
0
.
Botch
4
•
ape
OS
00
55
.0
3
0
.6
.0
. .
2
0
4
.
-
Pe
3r
Fig
16
.
CLAYS 605
with batch 4 was used , extrapolated where necessary , and all values of pe
were reckoned from this curve .
It can be seen from Fig . 16 that the data points from all batches plot on the
same line . Thus , the effects of whatever conditions created the differences
in behavior of batches 2 - 3 and batch 4, as seen in Figs . 13 and 15 , are com
pletely explained by expressing the strength of the soil in terms of the
Hvorslev parameters ; i . e . the changed conditions exerted their influence
or
strength by influencing
off
upon 73c and
ef
ec
upon and
.
DATA FROM TESTS AT RAPID STRAIN RATES
V
-
.
Extensive testing using very rapid strain rates has been reported by
-
Harvard Casagrande and Shannon 1948 and by 1953 1954 1959
M
;
,
(
(
.
I.
.
).
These efforts have previously been reviewed
by
the author Whitman 1957a
,
(
and the discussion this section represents amplification
an
of
in
1957b the
),
points made the earlier papers

in
concerning the
of
There has been tremendous mass data accumulated

a
strength at very rapid strain rates Some ten different cohesive soils have
-
been tested and several different sands some cases the soils have been
In
,
.
range confining pressures this point

of
The problem
is
of
tested under
to
a
reach some general conclusions based upon this mass
of
try data Un
to
.
fortunately the field problems which such information might be applied are
to
,
extremely complex and difficult obtain field checks upon the validity
is
to
it
This situation makes imperative that other

of
the strain rate effect data

it
.
-
by
means be found for validating the test results as examining the internal
,
consistency of the assemblage data or by postulating mechanisms ex

of
to
plain the data
.
There have been only two instances which has proved possible pin
in
it
to
point any recorded behavior pattern which appears consistent with theoretical
considerations These two instances are discussed the following sub
in
.
sections
.
of
on
Effect Strain Rate Pore Pressure

A
.
There are two effects be considered The first that which was dis
is
to
cussed previous subsection Namely that as the strain rate increases

in
a
,
:
gradients pore pressure will be set up within triaxial sample which

is
of
sheared undrained and hence the pore pressure within the central zone may
be time ependent one were testing soft saturated clay axial com
in
,
If
a
.
, -d
pression would be expected that anything the pore pressures the

if
in
it
central zone would increase as the strain rate increases The second possi
.
-
bility that the mineral skeleton may under rapidly applied loads have
is
a
,
compression which approaches that

of
resistance water behavior

to
Such
a
.
would be the result structural viscosity then the excess pore

, so
of
If
,
.
during undrained shear

up
pressures set soft saturated soils would de

in
This possibility has been suggested by

se
as
crea the strain rate increases

.
-
Casagrande and Wilson 1953b

)
.
(
As was mentioned the earlier subsection development of pore

in
a
,
pressure measuring system with rapid response time

at
now underway
is
a
As preliminary step the development work the system has been

in
M
a
T
,
.I
.
to .
study the pore pressures soil sample under

at
used the bottom surface

of
a
going consolidation and draining the top surface only The experimental
at
.
arrangement is shown in Fig . 17 . The soil was a remolded backswamp clay
with LL = 77 and PL = 30 . Loads were applied using a Karol - Warner consoli
dation loading frame with which load increments were applied within less than
one second . Just a few tests have been completed as this paper is written ,
but it has been found that the measured pore pressure rose to better than 98
per cent of the applied pressure increment within 1 / 2 second . Such results
indicate no tendency for the amount of pore pressure developed to depend upon
structural viscosity .
M . I. T . (1954 ) has reported the results of some very rapid triaxial tests on
saturated sands , using a pore pressure transducer located at one end of the
sample . Some results of this investigation are shown in Fig . 18. The sand
was a fine sand and was tested at a void ratio of 0.77 +0. 03 . The chamber
pressure was 60 lb / in2 , and initially the pore water pressure was 30 lb / in2 .
All tests , both fast and slow , were undrained tests . When dry , fine sands had
been tested , little or no strain - rate effect was found . Here , however , the
deviator stress at a strain of 6 per cent was 150 per cent higher in the most
rapid test than in the slowest tests . When pore pressure data were introduced
by calculating the principal effective stress ratio , a large fraction of this
strain - rate effect, but not all of it , di sappeared . A very high peak of princi
pal effective stress ratio was found at low strains in the fastest tests .
Two conclusions can be drawn from the data in the previous paragraph :
( 1) The end pore pressures almost certainly differed from the pore pressures
in the center of the sample , and (2) there certainly was a change in the
central zone pore pressures with strain -rate . Since the pore pressures tend
to decrease as the sample is sheared , it would be expected that the end pore
pressures were larger than the pore pressures in the central zone . This
situation accounted for the fact that the calculated effective principal stress
ratios increased with strain - rate . From this reasoning , it would be expected
that the end pore pressures would have been larger for the faster strain
rates . The data show that just the opposite occurred . The actual behavior
suggests the type of structural viscosity mechanism postulated by Casagrande
and Wilson . It appears , although the data are certainly not conclusive , that
both of the strain - rate vs. pore pressure effects were present in these tests .
The data cited in this section and earlier in section IV B show that the ef
fects of strain - rate on pore pressure show up in some tests , but not in
others . Obviously no general conclusions can be drawn at this point , other
than that there is some direct evidence of the existence of the phenomena .
Curiously , the best evidence of these effects have been found in sands rather
than in a clay . This is because one can work with more rapid tests when
dealing with sands , and because structural viscosity effects become more im
portant with a relatively incompressible sand than with a very compressible
clay .
B. Strain - Rate vs . Strength Behavior in Clay Soils
One pattern of behavior has emerged from all of the test work on cohesive
soils , and this pattern has been summarized in Fig . 19. The data on which
this summary is based have been reported by M . I . T . (1954 ) , and some of the
data have been reproduced in a paper by the author (Whitman 1957b ) . None
of these data have been reproduced here , because , frankly , there has been no
single set of " crucial experiments ” which demonstrate this behavior beyond
all doubt . The pattern has been inferred from the entire mass of experiment
al data , and would appear to be reasonably well established . Recent tests
-
STONE
1
GAGE
SOIL
SOILIE
of
Deviatorstress
-
peak
-
Oil
ratio 031
CEMENT PLUG
%
6
- ot
strain
speedslowest
at
//
//
Deviator stress
%
6
at
value
strain
to
Ratio
101
10
-3
-2 10
+
+
10
10
2
103 104
of
T,
Rote stroin percentper second
CLAYS
120
Deviatorstress
%
/
.
in 50 sec
/
lb
,
Deviator
stress
/ %
.
001 sec
Pore
Pressure
Pore pressure
pressure pore
and stress Deviator
50
10
.0I ./
%
sec
2
,
Axial strain percent
. 17
Fig
. 18
Fig
607
(M . I. T ., 1959 ) add additional confirmation , but again provide no complete
proof .
It seems possible , as is indicated in Fig . 19, to distinguish between the
behavior of those samples which failed in a brittle fashion , as by cracking or
splitting , and those samples which deformed in a plastic fashion . When a
sample fails in a brittle way , the inference may be drawn that the bonds with
in the sample have been overcome , and once overcome have been destroyed
rather completely . On the other hand , in a plastic type of failure the material
remains intact .
With those soils which deformed in a plastic fashion , the pattern of strain
rate effect suggested some sort of viscous action . The stress - strain curves
for slow and fast tests showed the same general shape , and the stress at 1
per cent strain was increased by the same percentage by increasing the
strain - rate as was the stress at 10 per cent strain . The soils which deformed
in a plastic way generally showed moderate strain -rate effect ; that is , the
strength increased on the order of 50 per cent as the time - to - failure was de
creased from several minutes down to several milliseconds .
The soils which broke in a brittle way generally showed larger strain - rate
effects - perhaps 100 per cent or more strength increase for the same range
of time - to - failure . The stress - strain curves from fast tests appeared quite
different than those from slow tests , having much broader peaks . As the
strain - rate increased , so too did the strain at which the maximum deviator
stress was reached . This pattern suggests that the bonds which hold the
brittle samples together have time -dependent characteristics ; i . e. there is a
limit to the length of time that certain of the bonds can be preserved . Thus ,
the amount of bonding which acts at a given strain increases as the strain
rate increases ; or to say it another way , the faster the strain - rate , the larger
the strain to which the sample holds together and hence the higher the
strength .
The test data suggest that it was primarily the level of pore pressures
within a sample which determined the category into which the sample fell .
Where the pore pressures which existed at failure were positive or slightly
negative , the plastic type of behavior generally was encountered . On the
other hand , where large negative values of pore pressure existed at failure ,
the brittle type of failure generally occurred . Thus , strong undisturbed clays
tended , when tested in unconfined compression , to fail in a brittle way, while
soft , remolded clays showed a plastic type of behavior . When compacted
clays were tested in unconfined compression , the dry samples showed brittle
behavior and the wet samples showed plastic behavior . When enough confin
ing pressure was applied to ensure positive pore pressures at the time of
failure , all aspects of the plastic type of action were generally encountered .
In several instances , the same soil was tested both in unconfined compression ,
showing the pattern of behavior associated with brittle failure , and in triaxial
compression , showing the pattern of behavior expected with plastic shearing .
There was even one series of unconfined compression tests where the samples
split in slow tests , but simply bulged in fast tests.
The exact cause of the viscous - type behavior during plastic failures is un
known . It must in some way be associated with the structural viscosity of an
assemblage of mineral particles , but the effect may be direct (viscous action
in the resistance of two particles to relative shear motions ) or indirect
( viscous action in the resistance of two particles to compression , and hence
an effect upon the pore pressures caused by loading ) . The latter hypothesis
is the one which was discussed in section IV B.
CLAYS 609
BRITTLE FAILURE PLASTIC FAILURE
FAILURE BY SPLITTING OR
FAILURE BY BULGING
PRONOUNCED FAILURE PLANES
Occurs where there are large Occurs in triaxial tests with large
negative pore pressures in chamber pressures , or where there
unconfined compression tests : are small negative pore pressures
in unconfined compression tests :
( 1) Soils compacted dry of
optimum content . ( 1) Soils compacted wet of
water
optimum water content .
( 2) Stiff saturated soils .
( 2) Soft saturated soils .
STRAIN - AT - FAILURE AFFECTED STRAIN AT FAILURE INDEPENDENT

BY STRAIN - RATE OF STRAIN - RATE
Fast test 01 Fast test
Slow test
Slow test
LARGE STRAIN - RATE EFFECT MODERATE STRAIN -RATE EFFECT

Strength
Strength
Strain Rate Strain Rote

-
-
Fig
19
.
hypothesis has been developed explain the time effects associated

to
A
with brittle action This hypothesis depends upon the idea time lag
of
in
a
.
the tendency cavitate when large negative pore pressures exist

of
water
to
Soil whose strength depends upon the existence large negative pore
of
pressures tend break apart brittle way when the pore water certain
to
in
in
a
portions the soils cavitates thus reducing almost eliminating the ability
of
or
of , ,
the soil resist shear Or unsaturated soils with large capillary

to
of
in
.
tensions the sudden release negative pore pressure occurs zones where
in
,
there is a concentration of coarser particles which tend to dilate into a looser
packing upon shearing . In such zones , one could easily imagine a time - effect
in the ability of the soil to develop and sustain large capillary tensions . Obvi
ously these thoughts need to be developed in much more precise form , but it
does seem that they shed light upon the shear resistance mechanism in co
hesive soils having negative pore pressures .
A result has been obtained using the testing apparatus shown in Fig . 17
which bears somewhat upon this hypothesis . After a sample of the backswamp
clay had been consolidated to 20 lb / in ? , the entire load was quickly removed .
With one - half second , the gage recorded a negative pore pressure of 20 lb /
in ” , i . e . 5 lb / in below absolute zero . However , after a matter of several
seconds , the negative pressure suddenly decayed to 15 lb / in2 ; i. e . to absolute
zero . It is likely , of course , that the cavitation occurred within the small
chamber of water adjacent to the gage rather than within the soil itself . In
any case , the data offer direct proof of the time -dependence of the cavitation
process .
VI. CONCLUSIONS
Of all the factors which affect the shear strength of a soil, the effective
stress at the time of failure is by far the most important factor . However ,
the influence of all the other factors cannot generally be ignored . The hope
for improving our understanding of shear strength behavior lies in establish
ing which of these additional factors exert a direct influence upon strength
and which exert only an indirect effect .
In problems where shear of saturated soils takes place without movement
of pore water and where expediency must be the rule , the in situ undrained
shear resistance can be measured directly and no attention paid to the ef
fective stresses which exist within the soil during shear . The price of this
simple expedient is lack of a clear picture of the manner in which the soil is
behaving . In important earthwork problems , such direct measurements of
undrained strength should always be supplemented with tests which give an
indication of the effective stresses which will exist within the soil mass .
It is a difficult matter to define exactly what is meant by “ strength " of
many materials , and this is particularly true of soils . The definition of
strength must be suited to the problem which is being attacked . Use of con
ventional methods of stability analysis does , however , commit the user to de
fine strength according to certain specific rules .
Data have been presented to illustrate the dependence of consolidated un
drained strength upon numerous factors , and to show further that the effects
of some factors can be accounted for once the fundamental behavior of the
soil has been established . Particular attention has been given to the effect
of strain - rate upon the pore pressure generated during undrained shear , but
the surface of this important subject has merely been scratched . The data
from triaxial tests at very rapid strain - rates have shed light on the mecha
nisms through which shear resistance is developed in soils with negative pore
pressures .
CLAYS 611
APPENDIX I. - SYMBOLS
capillary tension
the
intercept of straight line approximations to flot
)
as
tempera
of
environmental conditions such nature pore water
,
ture etc
,
of
void ratio time
at
consolidation
at
void ratio time failure

of
formation soil
at
conditions
of
stress history
net interparticle attraction
equivalent consolidation pressure Hvorslev theory
principal stress directions in
of of
rotation
degree saturation
time
pore pressure time failure
at
of
change plane
of
tangency
on
effective normal stress Mohr
to
in
envelope
101 change major principal total stress

in
102 change intermediate principal total stress

in
change minor principal total stress

in
Δσα
dispersed
or
structure flocculated
,
coefficient Hvorslev theory

in
ouf
slope straight line approximation

of
Tu
to
=
f(
coefficient Hvorslev theory

in
of
effective normal stress plane

on
failure time
at at
of of
OFC consolidation
normal stress plane failure

on
failure time
of
effective
effective normal stress plane tangency Mohr envelope

on
to
of of
oto
effective normal stress plane tangency

on
at
Mohr envelope
to
time consolidation
of
plane tangency time

at
total normal stress envelope

on
Mohr
to
Otf
of
failure
of
plane
at
effective normal stress tangency Mohr envelope

on
of
to
of
the time failure

major principal stress time
at
of
010 effective consolidation

of
effective stress normal the plane greatest

to
direction shear
in
distortion
73c minor principal effective stress at time of consolidation

OBE effective normal stress on plane of maximum obliquity at time of
failure
shear stress on plane of tangency to Mohr envelope

shear stress on plane of tangency to Mohr envelope at time of
consolidation
TH shear stress on plane of tangency to Mohr envelope at time of

failure
Τgr shear stress on plane of maximum obliquity at time of failure
APPENDIX II. - REFERENCES
1. Bishop , A . W . and Eldin , A . K . G . , 1950 : " Undrained Triaxial Tests on

Saturated Sands and Their Significance in the General Theory of Shear
Strength ” , Geotechnique , Vol . 2 , p . 13.
2. Bishop , A . W . and Henkel , D . J ., 1953 : “ Pore Pressure Changes during

Shear in Two Undisturbed Clays ” , Proc . 2nd International Conference on
Soil Mechanics and Foundation Eng ., Vol . I , p . 94 .
3. Bishop , A . W . and Henkel, D . J., 1957 : " The Triaxial Test " , Edward
Arnold , London .
4. Bjerrum , L . , 1954 : “ Theoretical and Experimental Investigations on the

Shear Strength of Soils ” , Norwegian Geotechnical Institute , Publication
No . 5.
5. Bjerrum , L ., Simons , N . and Torblaa , I. , 1958 : “ The Effect of Time on

the Shear Strength of a Soft Marine Clay " , Proc . European Earth
Pressure Conference , p. 148 .
6. Casagrande , A . and Shannon , W . L . , 1948 : “ Research on Stress

Deformation and Strength Characteristics of Soils and Soft Rocks under
Transient Loading ” , Harvard University Soil Mechanics Series No . 31.
7. Casagrande , A ., and Wilson , S . D ., 1951: " Effect of Rate of Loading on
the Strength of Clays and Shales at Constant Water Content ” , Geo
technique, Vol. 2, p . 251 .
8 . Casagrande , A . and Wilson , S. D. , 1953a : “ Prestress Induced in

Consolidated - Quick Triaxial Tests” , Proc . 3rd Me
Int
on
Conf Soil
.
chanics and Foundation Eng Vol 106

.p
.,
1
,
.
.
Casagrande Stress History

on
and Wilson Effects

of
1953b
,
.,
D
of A
9
S
.
"
:
Clays Harvard University Soil Mechanics Series No

43
the Strength
,
”
.
.
Strength and Rigidity

of
Clough
An
Investigation
10
1951 the
,
H
R
“
.
.
.
of :
Characteristics Clay Sheared without Drainage thesis

at
ScD
M
,
T
”
.
.
.I
11
Experimental Determination
of
Gibson 1953 the True Cohesion

,
.,
R
E
.
"
:
and True Angle Clays

on
Internal Friction Proc 3rd Int Conf

of
in
,
”
Soil Mechanics and Foundation Engineering Vol 126

p
1
,
,
.
.
CLAYS 613
12 . Gibson , R . E . and Henkel , D . J . , 1954 : " Influence of Duration of Tests

on ‘Drained ' Strength " , Geotechnique , Vol. 4, p. 6.
13. Henkel , D . J., 1958 : “ The Correlation between Deformation , Pore Water
Pressure and Strength Characteristics of Saturated Clays ” , PhD thesis
at University of London .
14 . Hirschfeld , R . C. , 1958 : “ Factors Influencing Constant Volume Strength
of Clays ” , ScD thesis , Harvard University .
15. Hvorslev , M . J. , 1938 : “ The Shearing Resistance of Remolded Cohesive

Soils ” , Proc . of Soils and Foundations Conf. , U. S. Engineer Department ,
U . S. Army .
16. Kenney, T . C., 1959 : discussion of paper " Geotechnical Properties of

Glacial Lake Clays ” , by T . H . Wu ; ASCE Journal of Soil Mechanics and
Foundation Division , Vol. 85 , SM3.
17 . Kolb , C. R . and Shockley , W . G ., 1957 : "Mississippi Valley Geology -
Its
Engineering Significance
of
ASCE Journal Soil Mechanics and Foun
,
”
83
dation Division Vol SM3

,
,
.
Soil Testing for Engineers John Wiley

18
Lambe Sons
W
1951 and
,
.,
,
:
"
.
. . T.
”
New York
Shear Strength Clay
of
19
Lambe Mechanistic Picture

W
1960
in
,
.,
,
A
T
:
"
"
.
paper this Research Conference

to
The Role Effective Stress

20
Lambe and Whitman

of
W
1959
R
,
.,
V
,
T
“
.
:
.
.
.
of
Expansive
of
Soils Quart the Colorado School

of
the Behavior
in
,
”
Vol
54
Mines No
4
,
,
.
.
.
Shearing Clay Report

of
on
1949 Research Characteristics

to M
.,
,
T
:
.I
.
"
"

.
Triaxial Research Report Waterways Ex

to
Clay
22
on
1950
M
.,
,
T
:
.I
"
"
.
periment Station
.
Shearing Clay Report

23
on
Characteristics
of
1951 Research
to M
.,
,
T
.
"
"
:
.I

.
of
Shearing Characteristics Clay

24
on
Research Effects
of
1953
M
.,
, ;
T
.
"
:
.I
Duration Pre Shear Consolidation and Rebound Pressures Report

of
to
”
-

.
Report Wave Propagation and Strain Ef

25
on
1953 Interim Rate

M
.,
T
to "
.
-
.I
report Engineers Army

of
of
fect Office the Chief

U
,
S
"
.
.
Final Report Laboratory Studies report

on
26
Office
of
to
1954
M
.,
,
T
”
.
:
.I
Engineers Army
of
the Chief
,
U
S
.
.
First Report Dynamic Soil Tests report

27
on
Interim
to
1959
M
.,
,
T
.
"
"
:
.I

.
Rutledge Cooperative Triaxial Shear Research Program

28
1947
,
of .,
C
P
“
.
the Corps Engineers Waterways Experiment Station

of
,
”
Structure and Strength Character

29
Seed and Chan 1959

,
85 , .,
H
C
B
” K
:
"
.
.
.
istics Compacted Clays ASCE Journal Soil Mechanics and

of
of
Foundation Division Vol SM5

,
,
.
.
STRENG
,
Analysis
30
Skempton 19WAR Stability and its Theoreti
..
)4;
of
The
W
W
1948
=
,
.,
A
on 0
“
.
:
cal Basis
Basig Proc Int Conf Soil Mechanics and Foundation Engi
Cal
2nd
,
"
.
neering Vol
72
.p
,
I,
.
.
Taylor
31
Shearing Strength Clay
of
of
on
Review Research
W
1955
.,
D
:
.
"
1953 Report Waterways Experiment Station
to
1948
M
,
.R
”
.I
-
.
.
32
of
Whitman 1957a The Behavior Soils under Transient Load
R
,
.,
V
.
"
:
4th
ings Engineering
on
Proc Int Conf Soil Mechanics and Foundation
1 ”,
,
.
.
Vol 207
.p
,
.
.
33
Whitman 1957b Testing Soils with Transient Loads ASTM

,
.,
,
. R
V
«
:
.
"
Spec Tech Pub STP 232 242
.p
,
.
.
A STUDY OF SHEAR FAILURE
IN CERTAIN TERTIARY MARINE SEDIMENTS
By James P . Gould ,' M . ASCE
SYNOPSIS
Observations of a group of failures in preconsolidated clays

and analyses
are presented herein . data indicate that for these materials with a long
The
history of sliding the shear strength is reduced in proportion to the prior
shear displacement in the field . The lower limit of field strength is equal to
the probable true internal friction . Observations of initial creep of intact
slopes show that laboratory failure strain is exceeded by the shear strain
which accompanies a small movement at top of bank .
INTRODUCTION
This paper summarizes studies and analyses of landslides in a coastal

area of California which involve two heavily preconsolidated marine clays of
Tertiary age . ( These studies form part of a project carried out under con
tract with the Department of Public Works , State of California . ) The slide
phenomena vary from sudden massive failures on cliffs with slope angles as
great as 45 degrees , to more or less continuous deep - seated creep on flat
slopes with angles as small as 9 degrees . Many of the sudden slides and all
of the continuous creep movements are conditioned by the existence of an
cient slides at the same location . The slides were investigated extensively
by borings and the position of failure surfaces and rate of shear deformation
were determined by " slope indicator " measurements in special casing in
stalled in boreholes . Strengths obtained in laboratory tests are compared
with those derived from stability analyses of the actual sliding masses in the
field to reach conclusions concerning the shear characteristics of these stiff
Tertiary clays .
The clays involved are classified geologically in the “Modelo " and the
"Martinez ” formations. The more troublesome material, the Modelo , is equi
valent to the Monterrey series of bituminous clay - sha les which are found at
many locations along the California coast and which have been involved in a
number of slides . While the data presented are derived from a particular
setting , it is possible that certain conclusions are of wider interest . One of
the most difficult problems encountered was that of creep movements in the
clays , a matter of increasing concern in recent investigations of strength and
stability of cohesive soils .
Physical Setting : -
The lands lides considered in this study occurred within a coastal strip 15
miles long , which lies on the periphery of the Los Angeles Basin at the foot
of the east -west trending Santa Monica Mountains on the north . In Cretaceous
and Tertiary times the area experienced periods of marine intrusion and
1. Assoc ., Moran , Proctor , Mueser & Rutledge, Cons . Engrs ., New York , N . Y .
615
deposition alternating with uplift of the northern mountains and erosion of the
marine sediments . Uplift was accompanied by extensive faulting and folding
of the Tertiary deposits , the axis of deformation aligned parallel with the
east -west trend of the coast . In Pleistocene times the coastal piedmont rose
in several stages exposing flat - topped marine terraces cut in the Tertiary
deposits at various levels on the south flanks of the mountains. Coarse grain
ed alluvium washed from the mountains was laid down on these terraces .
Towards the end of the Pleistocene the ocean advanced several miles in
land cutting cliffs along the ocean front and causing incision of streams flow
ing from the north . From the shelving beach the cliffs rise on slopes as steep
as 45 degrees to mesas formed by the seaward terrace at elevations between
200 and 300 feet above sea level . At certain points along the shore wave at
tack on the cliffs produced enormous slides in prehistoric times which left
bowl - shaped areas of gentle slope bordering the mesas . Similar ancient
slides were caused by undercutting in the stream canyons . Cross - sections
typical of prehistoric and recent slide areas are shown in Figures 1 through
4.
The tectonic of Tertiary times distorted the older sediments
disturbances
in a complex manner . Movements on the main east -west fault system during
the Pleistocene offset the alluvial cover of the mesas . There is no evidence
of recent fault movements in locations studied but the whole region is seismi
cally active . The largest shock recorded within the study area was of Mer
calli magnitude VIII , while shocks of magnitude VI occur at intervals averag
ing ten years . The epicenters are situated to the south and east within the
Los Angeles Basin .
The region is semi - arid with total annual rainfall averaging 14 . 7 inches , of
which 93 per cent occurs in the six months from November to April . Precipi
tation varies from place to place over short periods of time and the variability
is accentuated by recurring sequences of wet or dry years . In residential
areas the net domestic water used (total water consumption less sewage out
flow ) is equivalent to 10 inches of rainfall annually spread over the area of
use . It is estimated that almost one - half of the infiltration in such areas is
supplied by domestic water used in irrigation of grounds .
In the last 40 years the area has changed from occupancy by scattered
ranches to extensive residential developments on the mesa surfaces and
slopes . At the same time the number of slides , the frequency of movement
and the damage resulting therefrom has increased even though beaches and a
highway built up along the shore have eliminated wave attack on the cliffs .
The stage was set for instability by retreat of the coastline and incision of
the streams which caused slides at some locations and brought others to the
limit of stability . Slides recorded during the last 60 years , studied in con
junction with rainfall and seismic records , provided no evidence of the direct
influence of earthquake shocks on stability of the Tertiary clays . The land
slide history showed that the majority of failures were associated with periods
of heavier than average rainfall . Certain slide masses of prehistoric origin
have been moving intermittently for more than 50 years .
Since the number of failures increased with development of the area though
there was no long term change in rainfall, it appeared that the works of man
were of direct significance . Activities such as fills at top of bank , toe cutting ,
concentration of runoff on streets above the slide area , blocking of natural
drainage , etc ., contributed to slide occurrences . Piezometer observations
showed a relatively higher ground water table in sett led versus undeveloped
MARINE SEDIMENTS · 617
202
1927
PRIORTO1958 --
SLIDE-
MIXED AND B7 215 A

100
..
Wom SCALES
50
100
LOCATION I 0 50 150
*** SECTIONS
HORIZONTALO
MAXMUM
PIEZOLEVELS FOR1959 MOVEMENT
LEGENDFOR SECTIONS
m
GROUNDLINE1959
146 FORMER
GROUNDLINES
-
--
-
1924 STEATA
LINES
FILL FAILURESURFACE
.
.
.
PIEZOMETRIC
LEVEL
ON
SURFACE
FAILURE
..
OF..
.
..
MOVEMENT
HORIZONTAL
.
..
SLOPEINDICATOR
..
:
...
LOCATION PERIODOF
2
OBSERVATION
DAYS
,
50
-
Fig
1
.
.
250 250
48
FILL
1200
-
150 150
1927
100
WANIE LOCATION
3
200
194
100 188
50 LOCATION
4
a
300
300
296
9
.
250 192
200
AFTER
AFTER
200
O
F
I
SLIDE
1958 SUDE
-
LOCATION
5
Fig
2
.
.
sia SHEAR STRENGTH CONFERENCE
de w
FILL
MOVEMENTSHOWNFOR 11 - 8-58
70 SEE FIGURE5 FOR FULL RECORD
s
FILL
LOCATION 6
°,
LOCATION 7 /c0
SO
7
|
FILL LOCATION 8
D
•50 FILL
- FILL
°
s
FILL
LOCATION 9
•
•
.Fig
3
.
20
.
..
o
I.
TOEOF SLIDE
75
SOUTH
'
5
775
7
10
LOCATION
50
6
FILL
12
11
LOCATION LOCATION
4
/
s 5
-
to
FILL
5
ZD
..
a
Fig
,
4
.
MARINE SEDIMENTS 619
areas and it was concluded that this was the principal overall effect of devel
opments .
Identification Properties of the Tertiary Clays . –
The Mode lo exposed here is the lowest part of a thick column of Miocene
marine sediments of which several thousand feet have been eroded . Precon
solidation brought the natural water content 5 to 10 per cent below the plastic
limit . While the Modelo is a compaction rather than a cementation material,
it is described as a shale in geological studies . In engineering terms it is a
hard dark gray silty clay thin bedded with laminations of silty fine sand , which
at their thinnest are only sand partings . Though failures studied in the Mode lo
are separated by as much as 12 miles , the average material at the various
sites is similar . At some points the clay is more nearly massive and highly
bituminous while occasionally a sandy phase is identified . Because of the
differing percentages of fine sand in laboratory samples the limits and grada
tion vary considerably . Intense tectonics left the Mode lo with intricately dis
torted bedding and innumerable small fractures at random orientations and
occasional slickensides which are revealed in breaking or shearing the sam
ple . The weathered Mode lo is characteristically lighter colored , iron stain
ed , containing growths of gypsum crystals , with lower sampler penetration
resistance and higher water content than the parent material.
The Paleocene Martinez consists principally of a hard dark gray - green
clay with pockets of medium to fine sand and occasional calcareous nodules
and cemented fragments . It has been preconsolidated to below the plastic
limit by the weight of sediments since eroded , highly distorted by tectonics ,
fractured and occasionally slickensided . A predominately clayey phase and
a hard and partially cemented sandy phase were identified . The clayey stra
tum which is of importance to stability lacks the thin - bedded character of the
Modelo and actually has been so distorted that it is marbled with sworls and
pockets of sand . The parent materials weather to lighter and brighter color
ed soils of lower penetration resistance and higher water content .
The average properties of these weathered and unweathered Tertiary ma
terials are given in Table 1. The symbols distinguish the various soil strata
on the cross sections in Figures 1 through 4 . Soils of
principal concern are
the Strata " B2 " and "C4 " , the unweathered Modelo and Martinez clays . For
practical purposes the physical characteristics of the Clayey Mode lo and the
clayey Martinez are identical in the unweathered condition and are closely
similar in the weathered condition . The alluvium , Stratum " A " , found capping
the mesas or lying on slide masses at lower levels , grades upward from very
compact sand and gravel at the base of the deposit to hard brown silty clay
at the surface .
Influence of Physico - Chemical Changes .
Samples of weathered , partially weathered and unweathered Mode lo and
Martinez clays were subjected to compositional analysis to determine the
types
effects of continued leaching and weathering (1) ” . Montmorillonoid
constitute 60 to 70 per cent of the " clay size " fraction (minus 2 micron ) in
the parent materials , with 30 per cent Illite and the remainder Kaolin . The
unweathered soils include a comparatively large amount of organic material,
3 to 4 per cent , and the pore water contains almost no soluble salts .
The analysis indicated that the following changes take place in weathering :
the ratio of sodium ions to divalent calcium and magnesium ions in the pore
2. Numerals in ' parentheses refer to corresponding references in the Appendix .

1
TABLE
620
IDENTIFICATION PROPERTIES OF THE TERTIARY FORMATIONS
1
%
%
Dry Unit Natural Smaller Smaller
Weight Water Liquid Plastic Plastic Than Than
%
Material Symbol Description PCF Content Limit Index Limit 002mm 200 Sieve Activity
Bi
.
|9
. 7
Weathered Clayey
89
28
Stiff brownand gray
.
0
66
31
35
34
85
91
Modelo silty claythinbedded

with some lightgray
fine sand
|
Unweathered
.
.
98
2
23
3
|
|
|
69
37
32
31
.
84
1
19
B2 Hard dark gray silty

ClayeyModelo claythinbeddedwith
some light grayfine
sand
IP
|
.
.
4
16
5
-
Compactlight gray
6
SandyModelo B3
.
26
NP NPL
110
to
andtan fine medium
,
,
.
sand somesilt occ
clayseams
to
|6
1
.
.
|
|
|
|
WeatheredClayey CI
|
|
22
Soft mediumstiff
1
.|
62
38
24
30
103
73
26
Martinez brown clayeysilt and

silty clay
. 6
. 2
Weathered Sandy
T
Medium compactrust 108

17
14
NP NP
Martinez brown and yellow
to
silty fine medium
sand
3
Unweathered
.
.
|7 12
8
|
-|
|
|
|
Very compactgray
16
.|
38
120 NP NP
to
SandyMartinez greenclayeyfine
medium sandwith
calcareousconcretions
-
||
.. 6
Unweathered
.
6
19
2
|
|
|
68
40
|
|
28
30
.|
Hard dark gray green 105

80
1
33
105
ClayeyMartinez claywith some pockets
of
to
graymedium fine
,
sand occasionalshale
fragments
fluid decreases ; the quantity of organic matter decreases ; the ethylene

glycol retention and , therefore , the surface area per unit mass , decreases ;
the free iron oxide increases ; the cation exchange capacity of both the entire
soil and the clay fraction decreases ; the amount of montmorillonite de
creases . Because of the influence of varying content of fine sand in the test
samples the effect of weathering on limit properties was not entirely con
sistent , but , activity appeared to decrease with weathering . All of these
trends suggest that the shear strength at equal effective stress should in
crease rather than decrease as weathering proceeds . Although these tests
did not include a sufficient number of samples to be comprehensive the re
sults indicate that long - term physico - chemical changes would not of the m
selves , decrease stability . The weathered materials naturally are found at
shallower depths under the lower overburden pressure . The ordinary effects
on strength of release of pressure and swell would be working in opposition
to the physico - chemical changes in weathering .
Field Investigations .
As a part of the study , a program of borings , sampling and field observa
tions was carried out in sliding areas . Borings were positioned on lines in
the direction of movement to establish geological sections for analysis . Sam
ples were obtained in the standard split spoon and in 3- inch diameter , 16 gage
She Iby tubes employing a fixed piston . Because of the apparent effects of
ground water on stability , devices for observing water levels were provided
in every boring . These included Casagrande porous tube piezometers plus
small diameter wellpoint piezometers in areas of most active movement . To
measure pore pressures at different elevations usually more than one and as
many as four piezometers were placed in individual boreholes separated by
cement -bentonite seals . Plastic casing for slope indicators were installed
in 28 borings at active or potential slide locations.
A minor amount of field work was performed and a few piezometers were
installed in the early spring of 1958 . The winter and spring of 1957 - 1958
experienced 8 inches of rainfall above average and slide activity was second
only to that of the record rainy season of 1940 - 1941 . A number of areas
which had been moving in the past were again active and the largest slide of
record occurred . The main exploration was carried out from August to De
cember 1958 . Piezometers and slope indicators then installed were observed
from September 1958 to May 1959 . It was expected that piezometer obser
vations combined with surface and subsurface movement in 1958 - 1959 would
reveal a direct relation between changes in driving forces and shear strain .
However , this expectation was largely frustrated since the rainy season
1958 - 1959 had 8 inches of rain less than the average , a near record dry year .
Slope Indicator Observations .

Profiles of total horizontal movement of slope indicator casings are shown
at the position of the casings on the cross sections of Figures 1 through 4 ,
with notation of the number of days over which the observations extend . The
horizontal scale of movements is distorted to 320 times the scale of the sec
tions . Detailed profiles from three of the casings are shown on Figure 5.
The installation is made by coupling 5 - foot lengths of 3- 3 / 8 - inch O . D . , 1 / 4
inch wall, plastic pipe grooved lengthwise on two orthogonal diameters , and
lowering it in a completed borehole . (Equipment for the slope indicator in
stallation was provided by Slope Indicator Company of Seattle . ) The inclina
tion of the casing from plumb is measured at points spaced 1 '- 8" vertically
10- 17- 58
11- 11- 58 8 - 22 -58
' 12- 2-58
258
t6 -59 HEIM ELEV. *
5
.
ZIM ELEV 55
./
-3 -
.
59
24
59
26
59
12
30
58
-17
58
58
51
710
<
3
8
.
-
-
-
-
-
-
(
40
ELEVATION
OF
PROFILE
HORIZONTAL MOVEMENT
591
5
3 7
-
-
591
- -2-
591
12
21012
-
+
2001 +
to
+
-
--
--
! CS2
LOCATION
I
.
LOCATION
7
UPPER SLIDE
(
0
3
OF
OBSERVED HORIZONTAL MOVEMENT SINCE DATE INITIAL OBSERVATIONS INCHES
,
40
58
18
58
%
-3
-
8
-
-
- -4
6 9
-
33
10
58 ELEV
+
RIM
.4
,
.
-
30
SLOPE INDICATOR CASING

+
HE OBSERVED
HORIZONTAL MOVEMENT
ELEVATION
RESULTANT MOVEMENT
T
-
FAILURE ZONE
59
- 24
3
)
-
12-
26
58
.
10124 58 STATIONARY
-
1
LOCATION
9
SHEAR STRAIN
=
,
,
8
T -
LOWER SLOPE INDICATOR RADIANS Cosao

)
(
)
?
(
Fig Typical Slope Indicator Observations

5
.
.
its
by a pendulum device lowered into the casing with orientation controlled
by rollers riding the grooves
in
(2
)
.
integrated from the bottom the pipe
of
Successive values the inclination
of
,
,
assumed stationary give profile
of
which horizontal movement Total
is
a
,
by .
the tops of the casings thus computed were checked ground
of
movements
surveys The longest period observation extended over 275 days and
of
in
.
that time eight the 28 casings had been crimped or sheared so that the pen
of
dulum device would not pass the failure zone Such condition indicated
is
a
.
the sections by observation periods less than days
50
on
.
continuous creep was observed intact slopes and movements
of
No case
in
fall within the range
as
of
of
intact materials were small instrument
so
to
error else where accelerated and then slowed
of
Each location continuous
or
.
the past These are examples
an
movement was area which had failed

in
in
.
mass creep occurring
as
gravity forces
of
of
depth under the influence
at
”
"
distinguished from seasonal creep surficial materials the majority

of
of
In
is .
cases typified by Locations Figure occurring
on
and shear nar
in
,
5
,
a
row zone feet thick with comparative little distortion within the over
ly
to
2
6
,
lying mass The lower limit shear abrupt with little strain produced
of
in
is
.
the under lying material some instances the upper boundary the shear
In
of
.
equally abrupt The magnitude radians comput

of
zone shear strain
is
in
is
.
ed
follows for small angle changes

as
wc
H
Toos20
where
:
shear strain radians

in
= = =
T H y
observed horizontal movement

vertical dimension
of
shear zone
angle
of
shear zone with horizontal

=
0
spring 1958 1959 the rate shear strain was usually con
of
From fall
to
decreasing with time The average time rate

of
stant strain for individual

or
slope indicators ranges from

.6 an
radians per day with

10
10
to
,
1
5
x
2
”
x
radians per day equivalent

of
overall average
10
This value
to
is
0
4
4
x
on .
-
horizontal movement per month shear plane

of
inches feet thick which

4
a
makes flat angle with the horizontal

a
Ground Water Conditions

.
Two general ground water situations were encountered one the ancient
in
,
slide areas lying low

of
on
the sides the mesas the other slides

or
intact
in
,
on
gentle slopes
up
slopes extending the top the mesa Within old slides

a of
to
moving flat gradient and piezometric levels are

on
the ground water

is
nearly equal different elevations single borehole Old slide movements

at
in
a
strata which might have survived earlier tec

of of
have disrupted any continuity

excess hydrostatic pressures transmitted
no
tonics
in
There evidence
is
.
aquifers produced failure zones by very slow shear strain At these

or
in
piezometric levels the rainy season

of
of
locations the fluctuations 1958

in
1959 amounted only feet

or or
to
, of
those intact slopes slides which extend the tops the mesas there
to
In
usually steep seepage gradient beneath the slope Long deep canyons
is
isolate the mesas from distant flow and the ground water supplied
is is
tend
to
chiefly by rainfall and irrigation the flat mesa surface directed

on
Flow
.
624
primarily from the surface downward and piezometric levels decrease with
decreasing elevation particular boring . In an extreme case , on an iso
in a
lated mesa at Location 11 , there is an 80 - foot difference in head between the
highest and lowest piezometer in a single boring . Infiltration from intense
rains have set up perched water tables in the pervious alluvial capping of the
mesas on top of the Tertiary clays . In 1957 - 1958 , 10 feet or more of water
pressures were built up in certain piezometers at the top of bank by perched
water tables . In 1958 - 1959 only 7 inches of rain fell , almost no instance of
perched water was observed and the permanent water table generally respond
ed with several feet rise in pressures . At a few locations where runoff from
streets was concentrated the piezometric levels rose 3 to 5 feet .
Characteristics of the Failures .

Information on twelve slide locations is summarized on the cross sections
of Figs . 1 through 4 and in Table 2 . Failure limits were estimated from sur
face evidence , slope indicator observations , sampler penetration resistance
and sample characteristics . Ten of these cases involve Mode lo materials ,
two are within the Martinez . As a group they are typical of the performance
of the Tertiary clays large number of areas investigated . At three lo
in the
cations , Nos . 3, 5 and 12 , the
materials on the present s lope have undergone
limited shear strain without having been involved in massive failures . At
three other locations , Nos . 1, 4 and 6 , intermittent slides and creep occurred
in historic times on sections which were probably intact prior to development .
The other six cross sections are at locations where massive slides occurred
in prehistoric times either on or below the present active failure surface .
Since deve lopment these have continued to creep at depth but have experienced
no sudden slides . Here the failure surfaces are approximated by a straight
line or flat arc with active and passive wedges . On the plane surface the
creep consists of translation on a narrow zone with litt le distortion within the
failure mass .
Movements at Location 1 may have commenced as early as 1889 and con
tinued intermittently , until by 1958 a scarp averaging 25 feet in height had
developed at the top of bank . From 1954 to March , 1958 the mass moved
downward at a nearly constant rate of 2 inches per month , and failed sudden
ly on March 31 , 1958 , during heavy rains . At this time a perched water table
resulting in at least 10 or 15 feet of pressure had built up to within 4 feet of
the surface in a zone of cracks at the top of bank . In 1959 the permanent
water table was 25 feet deep at this location , no perched water was observed ,
and the main slide mass was stationary . In 1958 a secondary failure deve lop
ed behind the scarp which moved at an essentially constant rate in 1959 .
At Location 2 in an ancient slide , creep has continued intermittently since
it was first noted in 1922 . The bottom limit of shear extends to the lowest
elevation of any failure observed .
Location 3 is in intact material which first moved with a displacement of
8 to 10 inches at top of bank during the rains of 1958 . In 1959 a small addi -
tional strain accompanied a gradual 3 - foot rise in ground water pressures .
The fill shown lies in a narrow declivity and makes the cross section even
with adjacent slopes in natural soil .
Location 4 was first noted as active in 1924 and has been the scene of creep
movements and repeated slides , the latest of which was in April 1958 . The
change in sampler penetration resistances is abrupt at the failure surface
which probably has been the lower limit of past slides . Repeated failures of
the past probably are responsible for the uniform shear strain observed with
in the moving mass .
Large slides occurred in 1956 and 1958 on the cross section at Location 5 .
Although a pavement behind the top of bank is checked and cracked , no de
finite evidence of creep was observed in 1959 in the existing intact slope .
At Location 6 intermittent movements and small sudden slides have oc
curred since the earliest record in 1933 . Penetration resistance indicates
that the current active shear surface is also the lower limit of old failures ,
and movement at the top of bank has not spread back of the old slide mass .
Location 7 shows two distinct shear planes , one in weathered Mode lo be
low a fill and the lower in unweathered Modelo , probably at the base of an
ancient slide . Soil between the two surfaces is being dragged by the upper
failure. Surface movement , which tota led 3 feet during April 1958 , s lowed to
1 inch per month in 1959 .
Location 8 is within an ancient slide area which has undergone a total dis
placement of 90 feet in prehistoric times and has moved almost imperceptibly
in 55 years . Creep was accelerated by the rains of 1958 and is occurring on
a thin shear zone with almost no distortion within the mass and little surface
evidence of movement .
At Location 9 creep has continued within a prehistoric slide for at least
45 years . The rate of surface movement decreased from 5 inches per month
in April 1958 to 1 / 2 inch per month in March 1959 . Although the horizontal
movement in 20 years since the first development has exceeded 5 feet , struc
tures in the center of the sliding area have suffered no distress , so small is
the distortion within the mass . The upper slope indicator was positioned off
the axis of the slide and is not recording the main movement .
Location 10 is in an ancient slide which has moved 30 feet horizontally
since 1937 , with acceleration in heavy rains . In the spring of 1958 it moved
an additional 3 to 4 feet whereas in 1959 only small creep was observed .
Water levels at one piezometer were 7 feet higher in 1958 than in 1959 .
Location 11 has undergone intermittent movements since 1932 , probably on
the lower limit of an ancient slide mass . Shear on the long straight plane oc
curs in a zone 3 feet thick while at the upper part of the slide an active wedge
of material is tilting downs lope .
At Location 12 a small spread and settlement of the top of bank occurred
prior to 1958 . Several inches of movement were added in 1958 and ceased
with the end of the rains . One inch of horizontal movement took place abrupt
ly in February 1959 with a sudden rise of 3 feet in piezometric levels .
Stability Analysis Procedures .

Steep intact slopes were analyzed by trial circular arcs using conventional
effective stress methods . For active or recent movements failure surfaces
incorporating circular arcs , straight planes , and active and passive wedges
which matched the observed limits of shear were studied . Where a circular
arc dominated the failure surface , moments were summed about the center of
rotation . Where a long straight plane predominated , forces were summed in
the direction of the plane . For current movements the maximum pore pres
sures of 1959 were applied to the failure surface . In analyzing past slides or
the stability of intact slopes the estimated maximum seepage pore pressures
were applied . Pore pressures developed during undrained shear were not in
cluded in the computations .
626
2
TABLE
IN
CHARACTERISTICS OF FAILURES THE TERTIARY CLAYS
SHEAR
Average
in
of
Dimensions Suide Plan Average Average Rate of
of
(
)
Volume
of
Foet Depth of Height Average Shear Strain
of
Sude Dates Slides
in
Failure Mass Total Prior Direction Perpendicular Failure Surface Slope Slope Angle 1959
(
.
La of
)
|
Location
to
or Main Movements
Cu
(
)
Yds
(
Movement
)
(
Movement
)
(
Movement
)
Feet Feet Degrees Radians per day

STRENGTH
to
Creep 1889 1958
28
Slide 1958 780 000 730 750 100 230
, ,
Failed mass 1959 780 000 120 Stationary
88
ft & ft
. . .6
(
!
,
x
8
.3
Creep 1959 upper scarp 000 350 10°
200
CONFERENCE
Prehistoric slide
to
1
- 5
570
,
.ft
x
10
10
4
Creep 1922 1959 11

.6
600 000 110 500 100 125
First movement 1958
ia
.
,
x
Creop 1959
A
8
360
.5
10
150 000 soo 185 1005
Intermittent slides and

movement since 1924
20
Slide 1944 168 000 140 185
220 195
40
sudo 1958 85 000 450

"
2
x
2
.
10
220 180
.ft .ft .
ft
55 430
65
Cren 1959
, ,,
to
in
1.
2
.I
Stationary
|
Intact slope 1959 100210
Intermittent slides and

movement since 1933
3 -
x
10
|
3
200
1. 1
200
.ft
35
2
,
380 750
.|
Creep 1959 370 000 33100
,
Prehistoric slide
continuous croep
10
- 4
x
1
.
6
15
.A
,
52
Creep 1959 000 110 250 280
,
Prehistoric slide slight
movement since 1905
0
x
9
.
.t
,
60
10
Creep 1959 110 000 90 360 280
,
Prehistoric slido con
MARINE
tinuous creep since 1915
1
120
400 20
580
l
.ft
,
,
. 3
-
4
Creep 1959 220 000 6210
160
,
Prehistoric slide con
10
tinuous croep since 1937
10
- 4
x
3
|
.4
|
SEDIMENTS
165
,
10
.ft
930
15
Creep 1959 265 000 180 330
,
Prehistoric slide inter
mittent creep since 1932
O
,
22
0
.
1
.ft
140 000 730 103
25
Creep 1959 250
270
Slightmovement since
-
1955 1956
3
&
5
.
3
.
in
10 38
,
65 104
50
Creep 1959 000 210 180
627
The results of the analyses were expressed in terms of values of C , cohe
sion and , tan , tangent of the angle of shearing resistance mobilized by move
ment or failure . That is , the C required for a factor of safety of one was de
termined setting tan º equal to zero , and similarly the tan º required with C
equal to zero . For one specific circle and assumed conditions the usual anal
ysis yields :
(K.,)+
tan
'Kz
C
=
(
K
??
Kn
K1 are
to
where constants
.
The corresponding result for planar surface
is
a
:
)Kg
tano Active Force Passive Force
+
,
C
K
(
)
-
where active and passive forces are not general linear functions
of
and
in
,
C
tan
Q
.
particular circular requir
of
Thus for arc the combinations and tan
o
a
straight line joining the limiting values

ed
safety
of
on
lie
for factor one

of
a
a
of and tan such as Fig For long plane plus active and passive
®
in
C
a
,
.
.
wedges the straight line relationship only fair approximation The locus
is
a
,
.
of strength combinations for all possible failure circles an intact slope
on
is
straight lines one for each trial circle
of
curve tangent number

to
at ,
a
.
tension cracks are present of
or
Invariably the top recent slides active
,
Where vertical line the upper end of the failure sur

at
movements shown
is
a
. .
face Figs through the sliding mass limited by main tension crack
is
in
a
1
no 4
beyond which there apparent spreading active wedge

an
Where arc
or
is
continues the surface there zone of spreading without one dominant ten
to
is
a
Depths crack computed from strength tests range between

of
15
sion crack
.
Field evidence indicated that the actual cracks are no larger than
30
and feet
.
cracking depth were assumed for

of
10
20
the lower value

to
Zones feet
in
.
the analyses containing pore pressures where this was appropriate No al

,
lowance was included for negative pore pressures the zone between the bot
in
the tension crack and the upper flow line

of
tom
.
by
as
The slides investigated are anything but two dimensional shown the
,
many cases the width the slide mass along the bank
of
of
Table
is
data
In
2
.
less than the plan dimension the direction movement To account for the
of
in
three dimensional effect correction was applied the limiting and tan
to
C
,
a
•
-
values computed the two dimensional analysis The limiting value was
in
C
of .
divided by the ratio of the failure surface area an equivalent mass with ver
tical sides the section analyzed times the width
of
the failure surface area

to
required was divided by the ratio

of
slide total normal force

of
on
The tan
o
.
the equivalent mass those on the base of the section

of
the base and sides

to
slide For example the main slide

re
of
times the width Location

at
in
. C
1 1
,
,
.
quired by the conventional analysis was divided by by

10
19
the tan
o
1
.
.
Results Stability Analyses

.
of
of
safety
12
one for the

at
Combinations locations
of
of
and tan factor

C
a
o
discussed are shown Fig Since failure surfaces had been defined by field
in
6
.
.
observations single straight lines represent the limiting combinations

of
C
,
DATE AND PRIOR CURRENT
54 SLIDE 1956
2- 4
24
CONDITION VOVEMENT MOVEMENT
Ta
RADDAY
58 SLIDE 1958
CREEP 1959 1 5.8x105
/
=
SLIDE 1958
3.5x 10 "
CREEP 1959 38x10 +

SLIDE 1958
AT FACTOR OF SAFETY
CREEP 1959 163x10 3

SLIDE 1944 20
CREEP 1959 40x10°3
CREEP 1959 110 6.4x10 -3
FAILED MASS 120' O
CREEP 1959 180' 143x10
2.6 x10
6. / 10
9.0x10
REQUIRED
STABLE
.
Sorteadet
UNSTABLE
SQUARE FOOT
---
TONS PER
, --
COHESION
---
--
=
==
=
rån
OF
AT
REQUIRED FACTOR SAFE

Ø
º
Analysis
Of
Fig Past Slides And Active Movements

6
.
.
Average rates shear strain for those slides which were moving
of
and tan
, 0
.
1959 as determined from lope indicator observations are also given

in
,
s
Prior movement the total displacement which has occurred

on
the failure
is
"
"
surface before the particular slide or prior the 1959 creep movements
es
of to
timated from surface measurements offset strata borings and old topo
in
,
graphy
.
5A
The slide mobilized the highest strengths

at
of
line Location 1956

in
in 5
Tertiary clays
of
any movement The lowest value

observed thetan
is
º
.
mobilized by failure masses which are not continuing creep They are
to
.
represented by the failed mass of the 1958 slide at Location 1 ( line 1 C ) . Al
though cohesion required for such stationary slide masses varies , the tan 0
at cohesion equal to zero coincides at a value of about 0. 19 . The strength loci
of Fig . 6 fall in three fairly distinct groups depending on the amount of move
ment which has taken place prior to the event analyzed : those which have ex
perienced up to one foot displacement , 5 to 50 feet , and 100 to 200 feet of dis
placement . While there are the usual anomalies , these three groups have the
following ranges of tan Ø values at cohesion equal to zero : 1. 0 to 0. 7, 0 .6 to
0 . 4 , and 0 . 3 to 0 . 2 . The division in values of cohesion is less distinct .
The lower limiting tan from stationary failed masses , 0. 19 for an angle
of 11 degrees , is equivalent to the " true angle of internal friction " to be ex
pected from clays of the plasticity and activity of the Mode lo and Martinez .
Data summarized by Bjerrum indicate that the true friction angle should fall
between 10 and 14 degrees (3) . True friction is that component of strength
which depends solely on effective normal stress , as distinguished from true
cohesion which is a function of the denseness of particle arrangement acquir
ed under past or present loads .
Strengths mobilized by the rapid slides are the peak values available at the
time of failure , if transient strengths are ignored . Strengths mobilized by the
creep movements of 1959 are not necessarily the maximum va lues available .
Nevertheless , small percentage increases in driving force have produced much
more rapid creep , roughly 100 times as fast as that of 1959. The differences
in shear resistance at various rates of strain are not nearly as large as the
differences in strength associated with various amounts of prior displacement .
Since it is apparent that strengths mobilized decrease to some extent with de
creasing speed , the shear resistance available to halt creep entirely would be
less than that mobilized by the slow movements of 1959 .
Strength Test Procedure .

Triaxial tests were performed on 2. 8" diameter , 6 " long , undisturbed sam
ples of the Tertiary clays by the consolidated -undrained procedure with meas
urement of pore pressures . A limited number of consolidated - drained tri
axial tests were included . Specimens were prepared by lightly scraping the
sides to remove smear resulting from sampling and extrusion and to expose
sandy lenses . Samples were surrounded with strips of light filter paper con
nected to the end stones .
' The tests were intended to duplicate conditions encountered in situ . Be
cause of the high degree of preconsolidation , chamber pressures were set
equal to the total stress of overburden on the particular sample . After de
airing the test apparatus , and as the chamber pressure was imposed , a pore
pressure was applied at both ends of the sample equal to the measured or es
timated ground water pressures on the sample in the field . Swell of specimens
in sampling and extrusion appeared to be small since the replacement of in
situ effective stresses generally resulted in little consolidation . To define
better the failure envelope , certain samples were consolidated at higher ef
fective stresses than those acting on the material in the field .
Before testing the clay samples averaged about 90 per cent saturation . No
attempt was made to saturate the sample beyond replacement of field pore
pressures . If little or no weight of gas was lost from the sample in the re
duction of pore pressures from in situ values to atmospheric or less , the re
placement of field pressures should return the sample to field saturation . The
saturation achieved by this procedure was generally between 95 and 98 per
cent. Passing de -aired water through the sample could result in either an in
crease or decrease of saturation from field values , depending on the quantity
of flow and pore pressures used .
Pore pressures at the end stones were measured in a capillary size spa
ghetti tube which magnified any tendency for volume change during shear .
Back pressure from an air reservoir , varied to maintain the air -water con
tact at a fixed position in the spaghetti tube , was taken as the sample pore
pressure (4) . The use of positive boundary pressures at the start of the test
avoided the complication of measuring negative gage pressures during shear .
In any case , since negative gage pressures are undoubtedly influenced by the
size and shape of pore air spaces , it was assumed that they would not be
realistic if the initial in situ pore pressures were omitted .
To evaluate the influence of time rate of shear , series of tests were run at
various rates of strain with the time to failure ranging between 5 and 6000
minutes . The more rapid tests were conducted with constant strain rate by
applying an axial load mechanically through a gear train and proving ring . The
slower rates of shear were obtained for the most part by dead weight loading
using comparatively small load increments , equivalent to a constant rate of
stress increase . A more valid model test would be obtained by increasing pore
pressures to failure after consolidating under a ratio of minor to major total
principal stress somewhat less than one . Because of the uncertainties in con
trolling the time rate of shear in such a procedure it was not employed .
An effort was made to determine the load after failure at which samples
would reach equilibrium without continuing plastic flow . This was accomplish
ed by allowing the combination of sample and proving ring to relax after failure
by expansion of the proving ring and shortening of the sample until stress
strain equilibrium was reached . As a check on this procedure , the apparent
equilibrium load was replaced with dead weight loading maintained for at least
24 hours .
Strength Test Results .

Test data for the Tertiary clays with values of cohesion and tan ø from ef
fective stress envelopes are summarized in Table 3. Individual test results
are plotted in Figs . 7 and 8 as effective normal and shear stress on the failure
plane at the condition of maximum ratio of principal effective stress . The in
crement of pore pressure developed during shear is indicated by a horizontal
bar . Boundary pore pressures applied at the beginning of shear do not appear
in these diagrams . Preconsolidation exceeds any normal stress applied ; as
an approximation the failure envelope is taken as a straight line .
Axial strain at failure ranges from 2 to 6 per cent for the various test
series . Shear strain on the failure plane in terms of normal strain and volume
change is as follows :
cos
en
Yp
@
=
(
)
♡
where
:
Yo
failure plane
on
shear strain radians

in
=
angle shearing resistance

of
=
Ø
INTACT SAMPLES
WEATHERED MODELO CLAY (81) 5 MIN . TESTS ,
DRAINED TESTS : AV
5MIN .
150MIN. SHEAR
ZONE
..
SAMPLES
6000MIN.
'EQUILIBRIUM
FT STRENGTH
SQ
TONS PER
PROBABLE TRUE FRICTION
, UNWEATHERED
UNDRAINED TESTS
MODELO CLAY
(
82
)
:
STRENGTH
EFFECTIVE NORMAL AND SHEAR _

_
o
)
(-
STRESS ON FAILURE PLANE INTACT SAMPLES
KA MIN TESTS
.
SHEAR PODE PRESSURE
ZERO
SHEAR
PORE PRESSURE WOT

MEASURED
MIN
60 7
.
MIN
MIN
.
4500
SA
AE . EQUILIBRIUM
od STRENGTH
A
ODO
2
EFFECTIVE NORMÁL STRESS TONS PER SQFT

.
Fig Strength Test Results For Modelo Clays

7
.
. Eg
axial and radial unit strain

Eh
=
arithmetic where compression

is
,
sum
=
+
e
b
a
)
(
en
and extension
AV
volume change test specimen

of
=
Assuming negligible volume change the undrained test this expression for
in
small angle changes simplifies

to
:
cos
ca
YA
O
=
For the test series of Table 3 the shear strain on the failure plane at peak
strength ranges from about 0 . 04 to 0 . 09 radians .
.A WEATHERED AND UNWEATHERED MARTINEZ ( C124 )
.
TONS PER SQ FT
AO UNDRAINED TESTS
PORE PRESSURE NOT SHEAR ZONE
MEASURED SAMPLES
C1(45 MIN.)
N
. 04180 MIN
SHEAR STRENGTH
.)
Hono
?
EFFECTIVE NORMAL STRESS TONS PER SQUARE

SQUARE FOC
FOOT
.
Fig Strength Test Results For Martinez Clays

8
.
.
As testing proceeded became evident that there was substantial dif

it
ference strength between samples obtained from intact slopes and those re
in in
or near zones
of
covered shear distortion Nevertheless was difficult
to
it
,
.
distinguish by appearance alone intact samples from those sheared situ

in
.
All specimens had pattern irregular fractures with distortion and offset
of
a
bedding but those previously sheared contained irregular slick surfaces

of
on
,
as
which the clay has graphite like appearance These surfaces are not
a
.
-
the usual polished and striated slickensides

as
well developed homogeneous

in
material They are similar fault gouge the Tertiary formations There
to
in
.
ex
or
apparent opening softening

of
no
was fractures nor reasonable

to
is
it
,
pect that clay deep failure planes would support actual fissures
on
Water
.
samples any test series varied considerably because the dif

of
of
content
in
ferent proportions of fine sand pockets and lenses The average natural
in
or
of
water contents materials which had been sheared the field were
in
per cent higher than those the intact samples Presumably this was the
of
,
.
result expansion accompanying shear distortions Standard penetration re

of
sistance of the intact clays ranged from more than 100 blows per foot
80
as to
.
as
as
high
40
50
Materials
or
as
shear zones had blow counts low and

in
of
that the intact clays

.
Test failures occurred on angles between 50 and 55 degrees from the hori
zontal often following irregular fractures slick surfaces Failure angles
or
were not particularly consistent with the slope the effective stress envelope
of
bedding but the rather erratic test re

by
Shear planes were seldom controlled

sults are probably caused by the variation constituent materials from sam
in
ple sample
to
summarizing the strength test results should be noted that four dif
In
it
ferent categories
of
materials are involved weathered and unweathered

:
634
3
TABLE
SUMMARY OF STRENGTH TEST DATA FOR THE TERTIARY CLAYS

SHEAR STRENGTH
Average
of
Angle Initial
Cohesion Shearing Water Average Time
of
Average Failure Strain Average Rate Strain

to
Resistance Content Failure
(
%
(
)
с sq
|).
%
/
)
|.
/
(
Material Symbol tons Tan Axial Shear Radians Axial min Radians day Minutes
.
.4
)( 6
.
1
.
,
2
./ft .
064
40
25
22
19
Weathered Clayey BI 62 92
Modelo
CONFERENCE
(
)
Intact
1
1
.
..
..
5. 5
1
22
|
32 068
20
Weathered Clayey BI
068
32
L
.
27
51
27
.
.
.
4
1
|
Modelo
(
)
Failed
.
.
.
| 150
|
.
.
.
24
5
46
27
1
1
3
6
25 024
BI
Weathered Clayey 049
Modelo
(
)
Failed
36.3
..
.|.
BI
.
|
22 32
.
|
.
1
L 052
|
.
.
.
22
17
5
32
26
0
1
Weathered Clayey
012
00060 6000
.1.
1
Unweathered
.
.6
.
6
5
.
000
1
0
3
.
21
22
38
Clayey Modelo 091 26
(
)
Failed
Unweathered B2
14
.7
25
80
6
.
3
.
23
.
.
Clayey Modelo 092 09 19
(
)
Failed
B2
.5
35
.
Unweathered 12
.5
.
22
2
22
.
5
.
.
036 042 85 60
Clayey Modelo
(
)
Failed
. 5
.
Unweathered
.
.6
B2
32
.
8
19
10
.
23
.
086 0013
.
ClayeyModelo 028 4500
(
)
Failed
Weathered Clayey CI
.1
3
49
.
2
26
.
22
.
4
1
.
043
la
2255
071
26
45
.
4.9
|
22
.|
Martinez
.1
3
2
.
071
.| 043
.|
(
)
Failed
Unweathered
20
14
.5
26
26
|.
2
.
19
|
. 4
6
.
065
.
1
2
ClayeyMartinez 058 80
MARINE SEDIMENTS
Failed
) C
( :
Note and
-
values determined from effective stress envelopes for
consolidated undrained
.
tests with pore pressure measurements
635
clays ; samples from intact locations and those from zones of shear distortion
in the field . There was no essential difference in results between Mode lo and
Martinez clays in the various categories . In Figs . 7 and 8 the tests of intact
materials are indicated by solid circle symbols , all other symbols apply to
clays from shear zones .
This paragraph concerns intact Modelo materials , that is samples obtain
ed from locations which have undergone practically no shear distortion . Test
strengths of the weathered material are substantially greater than those of the
unweathered clay, confirming the results of the compositional analysis . Intact
samples failed in a brittle manner with 80 to 90 per cent of the peak strength
at two to three times the failure strain . Strengths of samples remolded and
recompacted at natural water content were about 80 per cent of those of un
disturbed samples . Pore pressures developed during shear of both weathered
and unweathered materials were small and erratic , either positive or nega
tive values , apparently influenced by the degree of saturation . It was conclud
ed that the increment of pore pressures set up by rapid slides in intact ma
terials was not significant . Therefore , the values of cohesion and tan o as
computed in the effective stress analysis of intact slopes can be compared
directly with the test envelopes for effective stresses in the same materials .
Strengths in five minute tests of intact Modelo , both weathered and unweath
ered materials , were as much as double those of the samples of the same ma
terials sheared in the field .
Weathered and unweathered Modelo samples from shear zones were tested
at various speeds of shear to evaluate continuing creep . No distinction could
be made in these samples between different magnitudes of prior shear dis
tortion in the field . Both weathered and unweathered clays showed a decrease
in cohesion and tan Q with decreasing strain rate . Substantial positive pore
pressures developed in shear of the weathered Modelo , probably because of the
higher initial void ratio , whereas small negative pressures predeominate in
the unweathered material. Samples of these previously sheared materials
failed in a plastic manner with essentially constant strength at increasing
strain after failure . There is no tendency for the pore pressures to rise after
failure with continuing gradual strain as in certain normally consolidated , sen
sitive clays . Strengths in five minute tests of samples remolded and recom
pacted at natural water content were 1. 5 to two times the test strength of sam
ples sheared in the field .
Modelo samples from shear zones were allowed to relax after failure to
stress - strain equilibrium . The " equilibrium " strengths were so erratic that
only a limiting range is shown in Fig . 7. Rowe (5 ) has suggested that creep
occurs by the successive breakdown of inter - particle bonds in the order of
their strength , and that equilibrium would be attained only if applied stress
does not exceed the strength from true friction . The equilibrium test strengths
are higher than the probable true friction envelope , due in part to insufficient
duration of test . Many of the tests were stopped if axial movement was less
than . 0005 inches in 24 hours ( shear strain less than 1 x 10 - 4 radians per
day ) , which is only somewhat slower than the average creep rate observed in
the field .
Comparison of Field and Laroratory Data .

The location of the failure surface in ancient slides was established by
shear under an ear lier profile . This position would not necessarily be consis
tent with the theoretical surface for laboratory strengths on the present profile .
Certain recent slides on intact slopes are shallower than would be expected
for the test envelopes with high apparent cohesion (for example , Locations 3
and 5) . This may be due to the tendency for failure to occur on planes extend
ing down from the deepest tension crack which usually forms a short distance
behind the top of bank .
The cross sections of Figs . 1 through 4 show that the lower limit of shear
lies in unweathered rather than weathered clay . Bedding attitude has influ
enced surficial sloughing but bears no consistent relation to the direction of
the deep failure planes . Tectonic distortions and the absence of strata with
greatly differing strength in the intact materials tend to minimize the impor
tance of bedding . This is the marked contrast to the performance of related
clays north of the Santa Monica Mountains and in the Palos Verdes peninsula
where failures have followed bedding planes with great persistence . This con
trast could be explained by the possibility that the Mode lo in the study area
lies further from the shore of the basin of marine deposition and thus includes
less variation in constituent materia ls .
Test strengths from effective stress envelopes are plotted in Fig . 9 along
with the possible combinations of C and tan º mobilized by slides and creep
movements . The upper most curve is the locus of strengths required by the
steepest slope presently stable in the Tertiary clays . The probably strengths
mobilized by failures in the unweathered clays , Strata B2 and C4 ,
on
lie
line
a
intersecting the laboratory values and the loci possible field strengths
of
10 .
These probable field strengths are plotted as failure envelopes Fig The
in
.
field and laboratory envelopes range over roughly the same band of shear re
sistance from maximum value for the rapid shear intact clay mini
of
to
a
a
for the gradual creep old slide material
of
mum
.
strengths mobilized by creep not vary systematically with the

do
Field
speed shear apparently because the effect of prior movement predominates
of
.
this respect the Tertiary clays may be similar the Gault clay Folke
of
to
In
stone Warren or the Cucaracha clay sha the Panama Canal

of
the latter
In
le
-
material creep continuing after slides 300 feet displacement with strength
of
to is
mobilized equal one quarter the original strength of the intact clay
of
(6
by )
.
-
seems reasonable that all or part

of
true cohesion should be eliminated

It
continuing creep which pervents the reestablishment inter particle bonds

of
not clear why large prior displacements such as 200 feet should be more
is
It
effective this respect than smaller ones say 50 feet This could involve
in
a
,
in .
Ma
of
more complete development slickensided surfaces the shear zones

.
terial bank usually alluvium and weathered clays

, at
top
of
the tension zone

in
may be fissured laboratory evidence reveals fissures

no
or
However field
in
.
clays The process strength reduction would

of
the deep shear surfaces

on
differ from that repeatedly cited for fissured British Tertiary clays
.
radians per day

of
The average rate creep

10
is
the field about

in
,
4
-4
x
five cycles slower than the five minute laboratory test

of
The fastest rate

.
creep measured radians per day the speed

of
10
about one tenth

is
of
1
3
,
,
x
the 5000 6000 minut laboratory test For the latter case the field and labo
to
ratory strengths for shear zone samples are fairly similar For slide masses
.
which have undergone displacement exceeding 100 feet probable that creep
is
it
will not be stopped unless remedial measures provide factor of safety of one
a
for the true angle of internal friction alone For most of the slides with inter
.
would be quite impractical meet this

of
mediate displacement 50 feet

to
to
it
5
requirement expected that measures which will limit the rise piezo
of
is
It
.
the slope will prevent sudden slides but will

at
metric levels the upper part

of
not necessarily halt creep movements

.
SYMBOL FOR LAB STRENGTHS

STRATA INTACT FAILED
BI
B2 90
CI
C4
GTEST
TIME TO
FAILURE
MINUTES
FOOT
- STRENGTHS REQUIRED FOR

F $ =1 OF STEEPEST INTACT
SLOPE IN TERTIARY CLAYS
SQUARE
TONS PER
MOST PROBABLE FIELD

STRENGTHS
,
UAWEATHERED CLAYS
COHESION
it MOVEMENT PRIOR
TO
de
SLIDE CREEP
FIGURE
6
(
SA
R0260000
ooto Loti
so
'to
]
1200
ol
Tz
1
.0
0
Tanilo
8
TAN
A
Of
Fig Comparison Field And Laboratory Strength

9
.
.
steep intact slopes the Tertiary clays reported and observed

of
Failures
in
the study area have taken place as result of some specific change driv
in
in
resisting forces Apparently the most important triggering mechanism

or
ing
.
the rise of piezometric levels heavy rains definite episode

no
There
is
is
in
recorded which failure has occurred spontaneously with the steady applica
in
of
tion constant force

a
Initial creep intact slopes was observed occur narrow shear zones
of
to
in
about the same thickness as those ancient slides The profile move
of
of
in
initial creep for example Location Fig almost exactly that

12
is
ments
in
4
,
,
)
.
(
by
predicted Ter Stepanian Tentative surface evidence indicates that the

(7
)
.
-
..
FT
ID MOST PROBABLE FIELD
FIG
STRENGTHS FROM
9
.
SQ
EST
TONS PER
.
MIN
)5
SLOPE
152
-
TEST INTACT CREEP OTO
.
&
MIN
TEEPEST SLIDES
78425
STRENGTH
OBSERVER
I57050
.
-
TooʻTO 200
--
--
-
--
IC
--
--
PRIOR
SHEAR
MOVEMENT
)
B2
4500 MIN TEST

L
(
)
5
2
, 3
EFFECTIVE NORMAL STEESS TONS PER SQUARE FOOT
Fig Comparison Field And Laboratory

10
Of
.
.
Failure Envelopes
overlying mass rotating translating more less rigidly with about equal
or
or
is
shear strain along the entire failure surface initial
an
several instances
In
at .
displacement
of
top bank during two rainy

to
of
inches was observed

4
These displacements
is on
seasons without massive failure narrow shear

a
.
as
as
of
zone would amount

Research Conference On Shear Strength of Cohesive Soils (1960 University of Colorado)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Research Conference On Shear Strength of Cohesive Soils (1960 University of Colorado)

Uploaded by

Copyright:

Available Formats

Research Conference on Shear Strength of Cohesive Soils.

New York, N.Y. : American Society of Civil Engineers, 1960.

Public Domain, Google-digitized

We have determined this work to be in the public domain,

SOIL MECHANICS AND FOUNDATIONS DIVISION , ASCE

UNIVERSITY OF COLORADO JUNE , 1960

SHEAR STRENGTH OF COHESIVE SOILS

SOIL MECHANICS AND FOUNDATIONS DIVISION , ASCE

UNIVERSITY OF COLORADO JUNE, 1960

SOIL MECHANICS AND FOUNDATIONS DIVISION

BRC EXECUTIVE COMMITTEE

J. O. Osterberg , Chairman , T . M . Leps , S. J. Johnson ,

TASK COMMITTEE ON SHEAR STRENGTH OF SOILS

W. J . Turnbull , Chairman , M . J . Hvorslev , Vice - Chairman ,

LOCAL COMMITTEE ON ARRANGEMENTS

Wesley G . Holtz , Chairman , Milton E . Bender , Jr . , Hugh W .

AMERICAN SOCIETY OF CIVIL ENGINEERS

SHEAR STRENGTH OF COHESIVE SOILS

Program of Conference .... .. .... ..

Reports by Moderators of Panel Sessions . . . . . . . . . . . . . . . . . . . . 1121

Table of Contents - Subject Index ......... . . . . .. 1157

Index of Authors and Discussors .... ... . . .. .. .. 1163

Copyright 1961 by the American Society of Civil Engineers .

PURPOSE AND ORGANIZATION OF CONFERENCE

The procedure of the ASCE Research Conference on Shear Strength of Co

1. Chf . , Soils Div . , Corps . of Engrs . , Waterways Experiment Sta . , Vicks

MONDAY AFTERNOON , JUNE 13

Chairmen : S . Mark Davidson , President , Colorado Section , ASCE , and

Assistants : M. J. Hvorslev and J. W. Hilf

Ralph Engineering Universi

Discussion Peck Professor Foundation

TUESDAY MORNING JUNE

Topic Testing Equipment Techniques and Errors

Moderator summary panel discussion and general discussion

Moderator Arthur Casagrande Professor Soil Mechanics Founda

tion Engineering Harvard University Cambridge Mass

Associate Moderator Stanley Wilson Partner Shannon and Wilson

Bishop Reader Soil Mechanics Imperial College

Technology London England

Wesley Chief Earth Laboratory Reclamation Denver

Woodland Shockley Assistant Chief Soils Division USAE Waterways

Experiment Station Vicksburg Miss

Consulting Engineer Hastings

Associate Professor Soil Engineering

Massachusetts Institute Mass

TUESDAY AFTERNOON , JUNE 14

Associate Moderator : Robert V. Whitman , Associate Professor of Soil

Laurits Bjerrum , Director , Norwegian Geotechnical Institute , Oslo ,

M . Juul Hvorslev , Consultant , USAE Waterways Experiment Station ,

Gerald A. Leonards , Prof . of Soil Mechanics , Dept . of Civil Eng . , Purdue

Jorj Osterberg Professor Civil Engineering Northwestern Uni

Shear Undisturbed Cohesive Soils

Moderator summary panel discussion and general discussion

Moderator Ralph Peck Professor Foundation

versity of Illinois Urbana Ill

Associate Moderator John Lowe Associate Partner Tippetts Abbett

McCarthy Stratton New York

Laurits Bjerrum Director Norwegian Geotechnical Institute Oslo

Harold Gibbs Supervisory Civil Engineer Soil Mechanics Earth Labo

Reclamation Denver Colo

James Gould Associate Moran Proctor Mueser Rutledge New York

Ronald Hirschfeld Instructor Soil Mechanics Harvard University

THURSDAY MORNING , JUNE 16