Full Ebook of Differential Geometry of Curves and Surfaces Third Edition Thomas F Banchoff Online PDF All Chapter

Differential Geometry of Curves and
Surfaces: Third Edition Thomas F.

Banchoff
Visit to download the full and correct content document:
https://ebookmeta.com/product/differential-geometry-of-curves-and-surfaces-third-edit
ion-thomas-f-banchoff/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Topological, Differential and Conformal Geometry of

Surfaces (Universitext) Norbert A'Campo
https://ebookmeta.com/product/topological-differential-and-
conformal-geometry-of-surfaces-universitext-norbert-acampo/
Tensor Algebra and Analysis for Engineers: With

Applications to Differential Geometry of Curves and
Surfaces 1st Edition Paolo Vannucci
https://ebookmeta.com/product/tensor-algebra-and-analysis-for-
engineers-with-applications-to-differential-geometry-of-curves-
and-surfaces-1st-edition-paolo-vannucci/
Parametric Geometry of Curves and Surfaces

Architectural Form Finding 1st Edition Alberto Lastra
https://ebookmeta.com/product/parametric-geometry-of-curves-and-
surfaces-architectural-form-finding-1st-edition-alberto-lastra/
Differential Geometry of Plane Curves 96th Edition

Hilário Alencar
https://ebookmeta.com/product/differential-geometry-of-plane-
curves-96th-edition-hilario-alencar/
A First Course in Differential Geometry Surfaces in
Euclidean Space Instructor Solution Manual Solutions
1st Edition Lyndon Woodward
https://ebookmeta.com/product/a-first-course-in-differential-
geometry-surfaces-in-euclidean-space-instructor-solution-manual-
solutions-1st-edition-lyndon-woodward/
Differential Geometry 1st Edition Victor V. Prasolov
https://ebookmeta.com/product/differential-geometry-1st-edition-
victor-v-prasolov/
Manifolds, Vector Fields, and Differential Forms: An

Introduction to Differential Geometry 1st Edition Gal
Gross
https://ebookmeta.com/product/manifolds-vector-fields-and-
differential-forms-an-introduction-to-differential-geometry-1st-
edition-gal-gross/
CRC Standard Curves and Surfaces with Mathematica 3rd

Edition Von Seggern David H
https://ebookmeta.com/product/crc-standard-curves-and-surfaces-
with-mathematica-3rd-edition-von-seggern-david-h/
Differential Geometry and General Relativity Volume 1

1st Edition Canbin Liang
https://ebookmeta.com/product/differential-geometry-and-general-
relativity-volume-1-1st-edition-canbin-liang/
Differential Geometry
of Curves and Surfaces
Differential Geometry of
Curves and Surfaces
Third Edition
Thomas F. Banchoff
Stephen Lovett
Third edition published 2023
by CRC Press
6000 Broken Sound Parkway NW, Suite 300, Boca Raton, FL 33487-2742
and by CRC Press
2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN
© 2023 Stephen Lovett and Thomas F. Banchoff
First edition published by CRC Press 2010
Second edition published by CRC Press 2015
CRC Press is an imprint of Taylor & Francis Group, LLC
Reasonable efforts have been made to publish reliable data and information,
but the author and publisher cannot assume responsibility for the validity of
all materials or the consequences of their use. The authors and publishers have at-
tempted to trace the copyright holders of all material reproduced in this publication
and apologize to copyright holders if permission to publish in this form has not
been obtained. If any copyright material has not been acknowledged please write
and let us know so we may rectify in any future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be
reprinted, reproduced, transmitted, or utilized in any form by any electronic,
mechanical, or other means, now known or hereafter invented, including photo-
copying, microfilming, and recording, or in any information storage or retrieval
system, without written permission from the publishers.
For permission to photocopy or use material electronically from this work, access
www.copyright.com or contact the Copyright Clearance Center, Inc. (CCC), 222
Rosewood Drive, Danvers, MA 01923, 978-750-8400. For works that are not
available on CCC please contact mpkbookspermissions@tandf.co.uk
Trademark notice: Product or corporate names may be trademarks or registered
trademarks and are used only for identification and explanation without intent to
infringe.
ISBN: 9781032281094 (hbk)

ISBN: 9781032047782 (pbk)
ISBN: 9781003295341 (ebk)
DOI: 10.1201/9781003295341
Typeset in Times
by codeMantra
Contents
Preface vii
Acknowledgments xiii
Authors xv
1 Plane Curves: Local Properties 1

1.1 Parametrizations . . . . . . . . . . . . . . . . . . . . . 1
1.2 Position, Velocity, and Acceleration . . . . . . . . . . . 11
1.3 Curvature . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.4 Osculating Circles, Evolutes, Involutes . . . . . . . . . 31
1.5 Natural Equations . . . . . . . . . . . . . . . . . . . . 38
2 Plane Curves: Global Properties 43

2.1 Basic Properties . . . . . . . . . . . . . . . . . . . . . 43
2.2 Rotation Index . . . . . . . . . . . . . . . . . . . . . . 48
2.3 Isoperimetric Inequality . . . . . . . . . . . . . . . . . 55
2.4 Curvature, Convexity, and the Four-Vertex
Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3 Curves in Space: Local Properties 65

3.1 Definitions, Examples, and Differentiation . . . . . . . 65
3.2 Curvature, Torsion, and the Frenet Frame . . . . . . . 72
3.3 Osculating Plane and Osculating Sphere . . . . . . . . 81
3.4 Natural Equations . . . . . . . . . . . . . . . . . . . . 88
4 Curves in Space: Global Properties 91

4.1 Basic Properties . . . . . . . . . . . . . . . . . . . . . 91
4.2 Indicatrices and Total Curvature . . . . . . . . . . . . 94
4.3 Knots and Links . . . . . . . . . . . . . . . . . . . . . 101
5 Regular Surfaces 111

5.1 Parametrized Surfaces . . . . . . . . . . . . . . . . . . 111
5.2 Tangent Planes; The Differential . . . . . . . . . . . . 118
v
vi Contents
5.3 Regular Surfaces . . . . . . . . . . . . . . . . . . . . . 128

5.4 Change of Coordinates; Orientability . . . . . . . . . . 139
6 First and Second Fundamental Forms 149

6.1 The First Fundamental Form . . . . . . . . . . . . . . 149
6.2 Map Projections (Optional) . . . . . . . . . . . . . . . 162
6.3 The Gauss Map . . . . . . . . . . . . . . . . . . . . . . 172
6.4 The Second Fundamental Form . . . . . . . . . . . . . 178
6.5 Normal and Principal Curvatures . . . . . . . . . . . . 188
6.6 Gaussian and Mean Curvatures . . . . . . . . . . . . . 199
6.7 Developable Surfaces; Minimal Surfaces . . . . . . . . 207
7 Fundamental Equations of Surfaces 219

7.1 Gauss’s Equations; Christoffel Symbols . . . . . . . . . 220
7.2 Codazzi Equations; Theorema Egregium . . . . . . . . 229
7.3 Fundamental Theorem of Surface Theory . . . . . . . 237
8 Gauss-Bonnet Theorem; Geodesics 241

8.1 Curvatures and Torsion . . . . . . . . . . . . . . . . . 242
8.2 Gauss-Bonnet Theorem, Local Form . . . . . . . . . . 251
8.3 Gauss-Bonnet Theorem, Global Form . . . . . . . . . 261
8.4 Geodesics . . . . . . . . . . . . . . . . . . . . . . . . . 271
8.5 Geodesic Coordinates . . . . . . . . . . . . . . . . . . 286
8.6 Applications to Plane, Spherical, and Elliptic
Geometry . . . . . . . . . . . . . . . . . . . . . . . . . 297
8.7 Hyperbolic Geometry . . . . . . . . . . . . . . . . . . 303
9 Curves and Surfaces in n-Dimensional Space 313

9.1 Curves in n-Dimensional Euclidean Space . . . . . . . 313
9.2 Surfaces in Euclidean n-Space . . . . . . . . . . . . . . 322
A Tensor Notation 335

A.1 Tensor Notation . . . . . . . . . . . . . . . . . . . . . 335
Bibliography 359
Index 363
Preface
What Is Differential Geometry?

Differential geometry studies properties of curves and surfaces, as well
as their higher-dimensional generalizations, using tools from calculus
and linear algebra. Just as the introduction of calculus expands the
descriptive and predictive abilities of nearly every field of scientific
study, so the use of calculus in geometry brings about avenues of
inquiry that extend far beyond classical geometry.
Before the advent of calculus, much of geometry consisted of prov-
ing consequences of Euclid’s postulates. Even conics, which came
into vogue in the physical sciences after Kepler observed that planets
travel around the sun in ellipses, arise as the intersection of a dou-
ble cone and a plane, two shapes which fit comfortably within the
paradigm of Euclidean geometry. One cannot underestimate the im-
pact of geometry on science, philosophy, and civilization as a whole.
The geometric proofs in Euclid’s Elements served as models of math-
ematical proof for over two thousand years in the Western tradition
of a liberal arts education. Geometry also produced an unending flow
of applications in surveying, architecture, ballistics, astronomy, and
natural philosophy more generally.
The objects of study in Euclidean geometry (points, lines, planes,
circles, spheres, cones, and conics) are limited in what they can de-
scribe. A boundless variety of curves, surfaces and manifolds arise
naturally in areas of inquiry that employ geometry. Various branches
of mathematics brought their tools to bear on the expanding horizons
of geometry, each with a different bent and set of fruitful results.
Techniques from calculus and analysis led to differential geometry,
pure set theoretic methods led to topology, and modern algebra con-
tributed the field of algebraic geometry.
The types of questions one typically asks in differential geometry
extend far beyond what one can ask in classical geometry and yet
the former do not entirely subsume the latter. Differential geometry
questions often fall into two categories: local properties, by which one
means properties of a curve or surface defined in the neighborhood
vii
viii Preface
of a point, or global properties, which refer to properties of the curve

or surface taken as a whole. As a comparison to functions of one
variable, the derivative of a function f at a point a is a local property
since one only needs information f near a whereas the integral of f
between a and b is a global property. Some of the most interesting the-
orems in differential geometry relate local properties to global ones.
The culminating theorem in this book, the Gauss-Bonnet Theorem,
relates global properties of curves and surfaces to the topology of a
surface and leads to fundamental results in non-Euclidean spherical
and hyperbolic geometry.
Using This Textbook

This new edition is intended as a textbook for a single semester un-
dergraduate course in the differential geometry of curves and surfaces,
with only multivariable calculus and linear algebra as prerequisites.
The interactive computer graphics applets that are provided for this
book can be used for computer labs, in-class illustrations, exploratory
exercises, or simply as intuitive aids for the reader. Each section con-
cludes with a collection of exercises which range from perfunctory to
challenging, suitable for daily or weekly problem sets.
However, the self-contained text, the careful introduction of con-
cepts, the many exercises, and the interactive computer graphics also
make this text well-suited for self-study. Such a reader should feel free
to primarily follow the textbook and use the software as supporting
material or primarily follow the presentation in the software package
and consult the textbook for definitions, theorems, and proofs. Either
way, the authors hope that the dual nature of software applets and
classic textbook structure will offer the reader both a rigorous and
intuitive introduction to the field of differential geometry.
This book is the first in a pair of books which together are intended
to bring the reader through classical differential geometry into the
modern formulation of the differential geometry of manifolds. The
second book in the pair, by Lovett, is entitled Differential Geome-
try of Manifolds[24]. Neither book directly relies on the other but
knowledge of the content of this book is quite beneficial for [24].
Each section ends with a list of Problems, ranging from staightfor-
ward applications of formulas to proofs of general results. Problems
marked with (*) indicate difficulty, which may be related to technical
ability, insight, or length. A few problems require the use of differ-
ential equations but we have marked these explicitly with the code
(ODE). Occasionally a problem involves calculations or integrals that
are particularly challenging; we marked these exercises with the code
(CAS) to indicate that the student should use a computer algebra
system to help calculate the desired number or determine an integral.
Preface ix
Computer Applets
An integral part of this book is the access to on-line computer graph-
ics applets that illustrate many concepts and theorems introduced
in the text. Though one can explore the computer demos indepen-
dently of the text, the two are intended as complementary modes
of studying the same material: a visual/intuitive approach and an
analytical/theoretical approach. The text always motivates various
definitions and topics, but the graphical applets allow the reader to
explore examples further, and give a visual explanation for defini-
tions or complicated theorems. The ability to change the choice of
the parametric curve or the parametrized surface in an applet, or to
change other properties allows the reader to explore the concepts far
beyond what a static book permits.
Any element in the text (Example, Problem, Definition, Theo-
rem, etc.) that has an associated applet is indicated by the symbol
shown in this margin. Each demo comes with some explanation text.
The authors intended the applets to be intuitive enough so that after
using just one or two (and reading the supporting text) any reader
can quickly understand their functionality. At present the applets
run on a Java platform but we may add other formats possibly based
on standard computer algebra systems. As an interest to an instruc-
tor of this course, the Java applets are extensible in that they are
designed with considerable flexibility so that the reader can often
change whether certain elements are displayed or not. Often, there
are additional elements that one can display either by accessing the
Controls menu on the Demo window or the Plot/Add Plot menu on
any display window. Applets given in a computer algebra system are
naturally extensible through the capabilities of the given program.
In previous editions, the applets were delivered via Java embedded
into a browser. That is no longer the case. Now all the applets are
delivered via a single Maple workbook, using embedded components.
The user does not need the full Maple software, but only needs to
download the free MaplePlayer. The reader will find instructions to
download the MaplePlayer software and the workbook that accom-
panies this text at:
http://cs.wheaton.edu/∼slovett/diffgeo
Organization of Topics
Chapters 1 through 4 cover alternately the local and global theory
of plane and space curves. In the local theory, we introduce the
fundamental notions of curvature and torsion, construct various asso-
ciated objects (e.g., the evolute, osculating circle, osculating sphere),
and present the fundamental theorem of plane or space curves, which
x Preface
is an analogue of the fundamental theorem of calculus. The global

theory studies how local properties (esp. curvature) relate to global
properties such as closedness, concavity, winding numbers, and knot-
tedness. The topics in these chapters are particularly well suited for
computer investigation. Students often make discoveries on their own
by manipulating curves or surfaces and various associated objects.
Chapter 5 rigorously introduces the notion of a regular surface,
the type of surface on which the techniques of differential geometry
are well-defined. Here one first sees the tangent plane and the concept
of orientability.
Chapter 6 introduces the local theory of surfaces in R3 , focus-
ing on the metric tensor and the Gauss map from which one defines
the essential notions of principal, Gaussian, and mean curvatures. In
addition, we introduce the study of surfaces that have Gaussian cur-
vature or mean curvature identically 0. One cannot underestimate
the importance of this chapter. Even a reader primarily interested in
the advanced topic of differentiable manifolds should be comfortable
with the local theory of surfaces in R3 because it provides many visual
and tractable examples of what one generalizes in the theory of man-
ifolds. Here again, as in Chapter 8, the use of the software applets is
an invaluable aid for developing a good geometric intuition.
Chapter 7 first introduces the reader to the component notation
for tensors. It then establishes the famous Theorema Egregium, the
celebrated classical result that the Gaussian curvature depends only
on the metric tensor. Finally, it outlines a proof for the fundamental
theorem of surface theory.
Another title commonly used for Chapter 8 is Intrinsic Geometry.
Just as Chapter 1 considers the local theory of plane curves, Chapter
8 starts with the local theory of curves on surfaces. Of particular im-
portance in this chapter are geodesics and geodesic coordinates. The
highlight of this chapter is the famous Gauss-Bonnet Theorem, both
in its local and global forms, without which no elementary course in
differential geometry is complete. The chapter finishes with applica-
tions to spherical and hyperbolic geometry.
The book concludes in Chapter 9 with a brief discussion on curves
and surfaces in Euclidean n-space. This chapter emphasizes in what
ways definitions and formulas given for objects in R3 extend and
possibly change when adapted to Rn .
A Comment on Prerequisites
The mathematics or physics student often first encounters differential
geometry at the graduate level. Typically, at that point, one is imme-
diately exposed to the formalism of manifolds, thereby skipping the
intuitive and visual foundation that informs the deeper theory. The
Preface xi
advent of computer graphics has renewed the interest in classical dif-

ferential geometry but this pedagogical habit remains. The authors
wish to provide a book that introduces the undergraduate student to
an interesting and visually stimulating subject that is accessible with
only the full calculus sequence and linear algebra as prerequisites.
In calculus courses, students usually do not study all the analysis
that underlies the theorems. Similarly, in keeping with the stated
requirements, this textbook does not always provide all the topolog-
ical and analysis background for some theorems. The reader who is
interested in all the supporting material is encouraged to consult [24].
A few key results in this textbook rely on theorems from the the-
ory of differential equations. We either spell out the calculations or
provide a reference to the appropriate theorem. Therefore, experience
with differential equations is occasionally helpful though not neces-
sary. A few exercises require some skills with differential equations
but these are clearly marked with the prefix (ODE).
Notation
As a comment on vector notation, this book consistently uses the fol-
lowing conventions. A vector or vector function in a Euclidean vector
⃗
space is denoted by ⃗v , X(t), ⃗
or X(u, v). Often γ indicates a curve
⃗ ⃗
parametrized by X(t) while writing X(t) = X(u(t),⃗ v(t)) indicates a
curve on a surface. The unit tangent and the binormal vectors of a
curve in space are written in the standard notation T⃗ (t) and B(t)
⃗ but
⃗ ⃗
the principal normal is written P (t), reserving N (t) to refer to the
unit normal vector to a curve on a surface. For a plane curve, U ⃗ (t)
⃗
is the vector obtained by rotating T (t) by a positive quarter turn.
Furthermore, we denote by κg (t) the curvature of a plane curve since
one identifies this curvature as the geodesic curvature in the theory
of curves on surfaces.
In this book, we often work with matrices of functions. The func-
tions themselves are denoted, for example, by aij , and we denote the
matrix by (aij ). Furthermore, it is essential to distinguish between
a linear transformation between vector spaces T : V → W and its
matrix with respect to given bases in V and W . Following common
notation in current linear algebra texts, if B is a basis in V and B ′ is a
B
basis in W , then we denote by T B′ the matrix of T with respect to
these bases. If the bases are understood by context, we simply write
[T ] for the matrix associated to T .
Occasionally, there arise irreconcilable discrepancies in definitions
or notations (e.g., the definition of a critical point for a function
Rn → Rm , or how one defines θ and ϕ in spherical coordinates).
In these instances the authors made a choice that best suits their
purposes and indicated commonly used alternatives.
xii Preface
Section Dependency
The topics in this book primarily oscillate back and forth between lo-
cal theory and global theory, first of plane curves, then space curves,
and then surfaces in R3 . In the authors’ perspective, the Gauss-
Bonnet Theorem (presented in Sections 8.2 and 8.3) serves as the
climax theorem of the textbook. Sections depend on each other ac-
cording to the following chart.
1.1–1.5
2.1–2.4 3.1–3.4
4.1–4.3 5.1–5.5
6.1, 6.3–6.6 A.1
6.2 6.7 7.1–7.2
7.3 8.1–8.5
8.6–8.7 9.1–9.2
Note that Appendix A.1 on tensor notation is valuable for back-

ground on the indices of tensor notation but it is not essential for
Chapter 7 and subsequent chapters.
Changes from the Second to the Third Edition

Feedback from faculty who used this textbook has led to the improve-
ments for the third edition. These include:
adding more exercises, both introductory and advanced;
improving the explanations in select areas;
rewriting the computer demos into Maple, as described earlier.

Acknowledgments
We would like to thank Dawson Bremner, Jonathan Higgins, and

Kevin Tully for their many suggestions for more exercises.
With the transition from the second edition to the third edition,
the interactive demos changed from being Java applets embedded in a
browser to a self-contained workbook in Maple. We are very grateful
to Dawson Bremner, Dave Broaddus, Lindsay Fadel, Dawson Miller,
and Caitlin Smith for programming all of the demos in the new Maple
format.
Thomas Banchoff
Our work at Brown University on computer visualizations in differen-
tial geometry goes back more than forty-five years, and I acknowledge
my collaborator computer scientist Charles Strauss for the first fifteen
years of our projects. Since 1982, an impressive collection of students
have been involved in the creation and development of the software
used to produce the applets for this book. The website for my 65th
birthday conference
http://tombanchoff.com
lists dozens of them, together with descriptions of their contributions.

Particular thanks for the applets connected with this book belong to
David Eigen, Mark Howison, Greg Baltazar, Michael Schwarz, and
Michael Morris. For his work as a student, an assistant, and now as a
co-author, I am extremely grateful to Steve Lovett. Special thanks for
help in the Brown University mathematics department go to Doreen
Pappas, Natalie Johnson, Audrey Aguiar, and Carol Oliveira, and
to Larry Larrivee for his invaluable computer assistance. Finally, I
thank my wife Kathleen for all her support and encouragement.
Stephen Lovett
I would first like to thank Thomas Banchoff my teacher, mentor,
and friend. After one class, he invited me to join his team of stu-
xiii
xiv Acknowledgments
dents on developing electronic books for differential geometry and

multivariable calculus. Despite ultimately specializing in algebra, the
exciting projects he led and his inspiring course in differential geom-
etry instilled in me a passion for differential geometry. His ability to
introduce differential geometry as a visually stimulating and mathe-
matically interesting topic served as one of my personal motivations
for writing this book.
I am grateful to the students and former colleagues at Eastern
Nazarene College. In particular I would like to acknowledge the un-
dergraduate students who served as a sounding board for the first few
drafts of this manuscript: Luke Cochran, David Constantine, Joseph
Cox, Stephen Mapes, and Christopher Young. Special thanks are due
to my colleagues Karl Giberson, Lee Hammerstrom, and John Free.
In addition, I am indebted to Ellie Waal who helped with editing and
index creation.
The continued support from my colleagues at Wheaton College
made writing this book a gratifying project. In particular, I must
thank Terry Perciante, Chair of the Department of Mathematics and
Computer Science, for his enthusiasm and his interest. I am indebted
to Dorothy Chapell, Dean of the Natural & Social Sciences, and to
Stanton Jones, Provost of the College, for their encouragement and
for a grant which freed up my time to finish writing. I am also grateful
to Thomas VanDrunen and Darren Craig for helpful comments.
Finally, I cannot adequately express in just a few words how grate-
ful I am to my wife Carla Favreau Lovett and my daughter Anne.
While I was absorbed in this project, they provided a loving home,
they braved the significant time commitment and encouraged me at
every step. They also kindly put up with my occasional geometry
comments such as how to see the Gaussian curvature in the reflection
of “the Bean” in Chicago.
Second Edition Acknowledgments

In preparation for this second edition, we wish to thank the many
people who contributed corrections, improvements, and suggestions.
These include but are not limited to Teddy Parker for careful editing,
Daniel Flath for many suggestions for improvements, Judith Arms
for errata and suggestions, and Robert Ferréol for bringing to our
attention an excellent on-line encyclopedia of curves at
http://www.mathcurve.com/ (in French).
We would also like to thank Nate Veldt, Gary Babbatz, Nathan Bliss,
Matthew McMillan, and Cole Adams.
Authors
Thomas F. Banchoff is a geometer and a professor at Brown Uni-

versity. Dr. Banchoff was President of the Mathematical Association
of America (MAA) from 1999 to 2000. He has published numerous
papers in a variety of journals and has been the recipient of many
honors, including the MAA’s Deborah and Franklin Tepper Haimo
Award and Brown’s Teaching with Technology Award. He is the au-
thor of several books, including Linear Algebra Through Geometry
with John Wermer and Beyond the Third Dimension.
Stephen Lovett is a professor of mathematics at Wheaton College.

Dr. Lovett has taught introductory courses on differential geometry
for many years. He has given many talks over the past several years
on differential and algebraic geometry as well as cryptography. In
2015, he was awarded Wheaton’s Senior Scholarship Faculty Award.
He is the author of Abstract Algebra: A First Course, Differential
Geometry of Manifolds, Second Edition, and Transition to Advanced
Mathematics with Danilo Diedrichs, all published by CRC Press.
xv
CHAPTER 1
Plane Curves: Local Properties
Just as calculus courses introduce real functions of one variable be-

fore tackling multivariable calculus, so it is natural to study curves
before addressing surfaces and higher-dimensional objects. This chap-
ter presents local properties of plane curves, where by local property
we mean properties that are defined in a neighborhood of a point on
the curve. For the sake of comparison with calculus, the derivative
f ′ (a) of a function f at a point a is a local property of the function
since we only need knowledge of f (x) for x in (a − ε, a + ε), where ε
is any positive real number, to define f ′ (a). In contrast, the definite
integral of a function over an interval is a global property since we
need knowledge of the function over the whole interval to calculate
the integral. In contrast to this chapter, Chapter 2 introduces global
properties of plane curves.
1.1 Parametrizations
Borrowing from a physical understanding of motion in the plane, we
can think about plane curves by specifying the coordinates x and y
as functions of a time variable t, which give the position of a point
traveling along the curve. Thus we need two functions x(t) and y(t).
Using vector notation to locate a point on the curve, we often write
⃗
X(t) ⃗
= (x(t), y(t)) for this pair of coordinate functions and call X(t)
2
a vector function into R . From a mathematical standpoint, t does
not have to refer to time and is simply called the parameter of the
vector function.
Example 1.1.1 (Lines) Euclid’s first postulate of geometry is that

through two distinct points there passes exactly one line. The point
slope formula gives a Cartesian equation of a line through two points.
In analytic geometry, the approach to describing a line through two
points is slightly different. Given two distinct points p⃗1 = (x1 , y1 )
DOI: 10.1201/9781003295341-1 1
2 1. Plane Curves: Local Properties
⃗v
p⃗ + 2⃗v
p⃗
Figure 1.1: A line in the plane.
and p⃗2 = (x2 , y2 ), the vector
⃗v = p⃗2 − p⃗1 = (x2 − x1 , y2 − y1 )
is called a direction vector of the line because all vectors along this
line are multiples of ⃗v . Then every point on the line can be written
with a position vector p⃗1 + t⃗v for some t ∈ R. Figure 1.1 shows
an example with t = 2. Therefore, we find that a line can also be
described by providing a point and a direction vector.
Using the coordinates of vectors, given a point p⃗ = (x0 , y0 ) and a
direction vector ⃗v = (v1 , v2 ), a line through p⃗ in the direction of ⃗v is
the image of the following vector function:
⃗
X(t) = p⃗ + t⃗v = (x0 + v1 t, y0 + v2 t) for t ∈ R.
Note that just as the same line can be determined by two different
pairs of points, so the same line may be specified by different sets of
these equations. For example, using any point p⃗ on the line will ulti-
mately trace out the same line as t varies through all of R. Similarly,
if we replace ⃗v with any nonzero multiple of itself, the set of points
traced out as t varies through R is the same line in R2 plane.
Example 1.1.2 (Circles) The pair of functions

⃗
X(t) = (R cos t + a, R sin t + b)
traces out a circle of radius R > 0 centered at the point (a, b). To
see this, note that by the definition of the sin t and cos t functions,
(cos t, sin t) are the coordinates of the point on the unit circle that
is also on the ray out of the origin that makes an angle t with the
positive x-axis. Thus,
⃗ 1 (t) = (cos t, sin t),
X with t ∈ [0, 2π],
1.1. Parametrizations 3
traces out the unit circle in a counterclockwise manner. Multiplying

both coordinate functions by R stretches the circle out by a factor of
R away from the origin. Thus, the vector function
⃗ 2 (t) = (R cos t, R sin t),
X with t ∈ [0, 2π],
has as its image the circle of radius R centered at the origin. Notice
also that by writing X ⃗ 2 (t) = (x(t), y(t)), we deduce that x(t)2 +
2 2
y(t) = R for all t, which is the algebraic equation of the circle. In
order to obtain a vector function that traces out a circle centered at
the point (a, b), we must simply translate X ⃗ 2 by the vector (a, b).
This is vector addition, and so we get
⃗
X(t) = (R cos t + a, R sin t + b), with t ∈ [0, 2π].
Two different vector functions can have the same image in R2 .
For example, if ω > 0,
⃗
X(t) = (cos ωt, sin ωt), with t ∈ [0, 2π/ω],
also has the unit circle as its image. Referring to vocabulary in
physics, this latter vector function corresponds to a point moving
around the unit circle at an angular velocity of ω (lower case Greek
“omega”).
When trying to establish a suitable mathematical definition of
what one usually thinks of as a curve, one does not wish to consider
as a curve a set of points that jump around or pieces of segments. We
would like to think of a curve as unbroken in some sense. In calculus,
one introduces the notion of continuity to describe functions without
“jumps” or holes, but one must exercise a little care in carrying over
the notion of continuity to vector functions. More generally, we need
to define the notion of a limit of a vector function as the parameter
t approaches a fixed value. First, however, we remind the reader of
the Euclidean distance formula.
Definition 1.1.3 Let ⃗v be a vector in R2 with coordinates ⃗v =

(v1 , v2 ) in the standard basis. The (Euclidean) length of ⃗v is given by
√ q
∥⃗v ∥ = ⃗v · ⃗v = v12 + v22 .
If p and q are two points in Rn with coordinates given by vectors ⃗v

⃗ then the Euclidean distance between p and q is ∥w
and w, ⃗ − ⃗v ∥.
Definition 1.1.4 Let X ⃗ be a vector function from a subset of R into

⃗
Rn . We say that the limit of X(t) as t approaches a is a vector w,
⃗
and we write
⃗
lim X(t) = w,
⃗
t→a
if for all ε > 0 there exists a δ > 0 such that 0 < |t − a| < δ implies
⃗
∥X(t) − w∥
⃗ < ε.
Definition 1.1.5 Let I be an open interval of R, let a ∈ I, and let

⃗ : I → R2 be a vector function. We say that X(t)
X ⃗ is continuous at
⃗
a if the limit as t approaches a of X(t) exists and
⃗
lim X(t) ⃗
= X(a).
t→a
The above definitions mirror the usual definition of a limit of a real

function but must use the length of a vector difference to discuss the
⃗
proximity between X(t) and a fixed vector w.
⃗ Though at the outset
this definition appears more complicated than the usual definition of
a limit of a real function, the following proposition shows that it is
not.
Proposition 1.1.6 Let X ⃗ be a vector function from a subset of R

2
into R that is defined over an interval containing a, though perhaps
⃗
not at a itself. Suppose in coordinates we have X(t) = (x(t), y(t))
⃗
wherever X is defined. If w
⃗ = (w1 , w2 ), then
⃗
lim X(t) =w
⃗ if and only if lim x(t) = w1 and lim y(t) = w2 .
t→a t→a t→a
⃗
Proof : Suppose first that limt→a X(t) = w.
⃗ Let ε > 0 be arbitrary
and let δ > 0 satisfy the definition of the limit of the vector function.
⃗ − w∥
Note that |x(t) − w1 | ≤ ∥X(t) ⃗ − w∥.
⃗ and that |y(t) − w2 | ≤ ∥X(t) ⃗
Hence, 0 < |t − a| < δ implies |x(t) − w1 | < ε and |y(t) − w2 | < ε.
Thus, limt→a x(t) = w1 and limt→a y(t) = w2 .
Conversely, suppose that limt→a x(t) = w1 and limt→a y(t) = w2 .
Let ε > 0 be an arbitrary positive real number. By definition, there √
exist δ1 and δ2 such that 0 < |t − a| < δ1 √ implies |x(t) − w1 | < ε/ 2
and 0 < |t−a| < δ2 implies |y(t)−w2 | < ε/ 2. Taking δ = min(δ1 , δ2 )
we see that 0 < |t − a| < δ implies that
r
⃗
p
2 2
ε2 ε2
∥X(t) − w∥
⃗ = |x(t) − w1 | + |y(t) − w2 | < + = ε.
2 2
This finishes the proof of the proposition. □
Corollary 1.1.7 Let I be an open interval of R, let a ∈ I, and con-

sider a vector function X⃗ : I → R2 with X(t)⃗ = (x(t), y(t)). Then
⃗
X(t) is continuous at t = a if and only if x(t) and y(t) are continuous
at t = a.
Definition 1.1.5 and Corollary 1.1.7 provide the mathematical

framework for what one usually thinks of as a curve in physical intu-
ition. This motivates the following definition.
Definition 1.1.8 Let I be an interval of R. A parametrized curve (or

parametric curve) in the plane is a continuous function X⃗ : I → R2 .
⃗
If we write X(t) = (x(t), y(t)), then the functions x : I → R and
y : I → R are called the coordinate functions or parametric equations
of the parametrized curve. We call the locus of X(t)⃗ the image of
⃗ 2
X(t) as a subset of R .
It is important to note the distinction in this definition between

a parametrized curve and its locus. For example, in Example 1.1.1
we show that the parametrization X(t) ⃗ = t⃗v + p⃗ traces out a line L
that goes through the point p⃗ with direction vector ⃗v . However, the
line is the locus of X ⃗ : R → R2 and not the vector-valued function
⃗
X itself. The vector-valued function is the parametrized curve. Since
the function t 7→ t3 +t is a bijection from R to itself, the parametrized
curve Y⃗ (t) = (t3 + t)⃗v + p⃗ has the same locus, the line L, but is quite
different as a function.
The following examples begin to provide a library of parametric
curves and illustrate how to construct parametric curves to describe
a particular shape or trajectory.
Example 1.1.9 (Graphs of Functions) The graph of a continu-

ous function f : [a, b] → R over an interval [a, b] can be viewed
as a parametric curve. In order to view the graph of a continuous
function as a parametrized curve, we use the coordinate functions
⃗
X(t) = (t, f (t)), with t ∈ [a, b].
Example 1.1.10 (Circles Revisited) The parametrization given

for circles in Example 1.1.2 utilizes trigonometric functions. Inspired
by Euclid’s formula for Pythagorean triples, we can get a parametri-
zation for the unit circle using rational functions:
1 − t2 2t
⃗
X(t) = (x(t), y(t)) = , for t ∈ R. (1.1)
1 + t2 1 + t2
It is easy to see that for all t ∈ R, x(t)2 + y(t)2 = 1. This means that
the locus of X ⃗ is on the unit circle. However, this parametrization
does not trace out the entire circle as it misses the point (−1, 0). We
leave it as an exercise to determine a geometric interpretation of the
parameter t and to show that
⃗
lim X(t) ⃗
= lim X(t) = (−1, 0).
t→∞ t→−∞
This example emphasizes that it is not possible to tell what the

parametrization is simply by looking at the locus. A counterclockwise
circle is indistinguishable from the circle covered twice, or from the
circle traced out in a clockwise fashion, or, as in this example, from
a circle covered in a completely different way.
Example 1.1.11 (Ellipses) Without repeating all the reasoning of

the previous exercise, it is not hard to see that
⃗
X(t) = (a cos t, b sin t)
provides a parametrization for the ellipse centered at the origin with

axes along the x- and y-axes, with respective half-axes of length |a|
and |b|. Note that these coordinate functions do indeed satisfy
x(t)2 y(t)2
2
+ 2 =1 for all t.
a b
Example 1.1.12 (Lissajous Figures) It is sometimes amusing to

see how cos t and sin t relate to each other if we change their respective
periods. Lissajous figures, which arise in the context of electronics,
are curves parametrized by
⃗
X(t) = (cos mt, sin nt),
where m and n are positive integers. See Figure 1.2 for an example
of a Lissajous figure with m = 5 and n = 3.
Figure 1.2: A Lissajous figure.
Example 1.1.13 (Cycloids) We can think of a usual cycloid as the

locus traced out by a point of light affixed to a bicycle tire as the
bicycle rolls forward. We can establish a parametrization of such a
curve as follows.
Assume the wheel of radius a begins with its center at (0, a) so

that the part of the wheel touching the x-axis is at the origin. We
view the wheel as rolling forward on the positive x-axis. As the wheel
rolls, the position of the center of the wheel is f⃗(t) = (at, a), where t
is the angle measuring how much (many times) the wheel has turned
since it started. At the same time, the light—at a distance a from
the center of the wheel and first positioned straight down from the
center of the wheel—rotates in a clockwise motion around the center
of the wheel. See Figure 1.3. The motion of the light with respect to
the center of the wheel is
π π
⃗g (t) = a cos − t − , a sin − t − = (−a sin t, −a cos t).
2 2
The locus of the cycloid is the vector function that is the sum of f⃗(t)
and ⃗g (t). Thus, a parametrization for the cycloid is
⃗
X(t) = (at − a sin t, a − a cos t).
Figure 1.3: Cycloid.
One can point out that reflectors on bicycle wheels are usually not
attached directly on the tire but on a spoke of the wheel. We can
easily modify the above discussion to obtain the relevant parametric
equations for when the point of light is located at a distance b from
the center of the rolling wheel. We obtain
⃗
X(t) = (at − b sin t, a − b cos t).
⃗
Whether or not b < a, the curve traced out by X(t) is called a
trochoid .
Example 1.1.14 (Heart Curve) The most popular curve around

Valentine’s Day is the heart shape. Here are some parametric equa-
tions that trace out such a curve:
⃗
X(t) = ((1 − cos2 t) sin t, (1 − cos3 t) cos t).
We encourage the reader to visit this example in the accompanying
software and to explore ways of modifying these equations to cre-
ate other interesting curves. The Online Encyclopedia of Curves at
http://www.mathcurve.com/ credits this curve to Raphaël Laporte
who designed it for his “petite amie” (translated “girl friend”).
Example 1.1.15 (Polar Functions) When working with polar co-

ordinates, we usually consider functions of the radial distance r as a
function of the angle θ, written r = f (θ). The graphs of such functions
are as parametrized curves in a natural way. Recall the coordinate
transformation (
x = r cos θ,
y = r sin θ.
Then take θ as the parameter t, and the parametric equations for the
graph of r = f (θ) are
⃗
X(t) = (f (t) cos t, f (t) sin t).
As an example, the polar function r = sin 3θ traces out a curve

that resembles a three-leaf flower. (See Figure 1.4.) As a parametric
curve, it is given by
⃗
X(t) = (sin 3t cos t, sin 3t sin t).
Figure 1.4: Three-leaf flower.
Example 1.1.16 (Cardioid) Another common polar function is a

cardioid, which is the locus of r = 1 − cos θ. In parametric equations,
we have
⃗
X(t) = ((1 − cos t) cos t, (1 − cos t) sin t).
As mentioned in Example 1.1.2, for a parametric curve X ⃗ :I →

2 ⃗ 2
R , the set of points C = {X(t) | t ∈ I} as a subset of R does not
depend uniquely on the functions x(t) and y(t). In fact, it is important
to make a careful distinction between the notion of a parametrized
curve as defined above, the image of the parametrized curve as a
subset of R2 (also called the locus of the curve), and the notion of
a curve, eventually defined as a one-dimensional manifold. (See [24,
Chapter 3].)
Definition 1.1.17 Given a parametrized curve X ⃗ : I → R2 and any

continuous functions g from an interval J onto the interval I, we can
produce a new vector function ξ⃗ : J → R2 defined by ξ⃗ = X ⃗ ◦ g. The
⃗ ⃗ ⃗
image of ξ is again the set C, and ξ = X◦g is called a reparametrization
⃗
of X.
If g is not onto I, then the image of ξ⃗ may be a proper subset of

C. In this case, we usually do not call ξ⃗ a reparametrization as it does
not trace out the same locus of X.⃗
Problems
In Problems 1.1.1 through 1.1.4, sketch the curve with the given parametric
equations. Indicate with an arrow the direction in which t increases.
⃗
1. X(t) = (t3 , t2 ).
⃗
2. X(t) = (t2 , cos t).
⃗
3. X(t) = (t cos t, t sin t).
⃗
4. X(t) = (t2 − 1, t3 − t).
5. Equation (1.1) gave the following parametric equations for a circle:
1 − t2 2t
⃗ = (x(t), y(t)) =
X , , for t ∈ R.
1+t 2 1 + t2
Prove that the parameter t is equal to tan θ, where θ is the angle

shown in the following picture of the unit circle.
6. Let C be the circle of center (0, 2) and radius 1. Let S be the set of
points in R2 that are the same distance from the outside of circle C
as they are from the x-axis. Show the following.

⃗ 3 sin t 3 cos t
(a) S has a parametrization X(t) = ,2 − .
1 + cos t 1 + cos t
[Hint: Use the parameter t as the angle as shown below.]
(b) S is a parabola. Find the Cartesian equation for it.
(0, 2)
t
⃗
X(t)
7. An epicycloid is defined as the locus of a point on the edge of a circle

of radius b as this circle rolls on the outside of a fixed circle of radius
a. Supposing that at t = 0, the moving point is located at (a, 0).
Prove that the following are parametric equations for an epicycloid:

⃗ a+b a+b
X(t) = (a + b) cos t − b cos t , (a + b) sin t − b sin t .
b b
8. A hypocycloid is defined as the locus of a point on the edge of a circle
of radius b as this circle rolls on the inside of a fixed circle of radius a.
Assuming that a > b and that at t = 0, the moving point is located
at (a, 0), find parametric equations for a hypocycloid.
9. Revisit Example 1.1.13 and find parametric equations for a cycloid
resulting from a circle of radius 3 rolling on the line y = x/2.
⃗ 2 ⃗
10. Consider the parametrized
1 t −t 1 t −t
curve X : R → R with equations X(t) =
2
(e + e ), 2
(e − e ) . Prove that the locus is one branch of a
unit hyperbola x2 − y 2 = 1. Give a parametrization for the other
branch of the hyperbola.
11. Consider the curve C in R2 that is the solution to x3 + y 3 = 1.
(a) Let√ X⃗ : R − {−1} → R2 be the parametrized curve with X(t) ⃗ =
√
3 3
(1/ t + 1, t/ 3 t3 + 1). Prove that the locus of X ⃗ is C except
for one point.
(b) Show that the “missing” point arises as the limit limt→∞ X(t) ⃗
⃗
or limt→−∞ X(t).
(c) Let √Y⃗ : R → R2 be the parametrized curve with Y ⃗ (u) = ((1 +
√
3 3
u)/ 2 + 6u , (1 − u)/ 2 + 6u ). Prove that the locus of Y
2 2 ⃗ is
on and is in fact all of C.
(d) Show that (over most of its domain) Y ⃗ is a reparametrization
⃗ with t = (1 − u)/(1 + u) and discuss the differences in the
of X
domains of X ⃗ and Y⃗ and why one covers all of C, while another
doesn’t.
12. Consider the vectors ⃗a = (3, 3) and ⃗b = (−1, 1). Explain why the
parametric curve X(t)⃗ = (cos t)⃗a +(sin t)⃗b is an ellipse. Furthermore,
give a Cartesian equation of the locus of this parametric curve.
13. Consider the parametric curve X(t) ⃗ = (t3 − 5t, 3t2 ). Graphing it
shows that it intersects itself. By solving for t1 and t2 the equation
⃗ 1 ) = X(t
X(t ⃗ 2 ), find the parameters where the curve intersects itself
and give the coordinates of the point on the locus where this occurs.
1.2. Position, Velocity, and Acceleration 11
14. Consider the parametrized curve X⃗ : [0, 2π] → R2 that gives a defor-
⃗
mation of a cardioid X(t) = ((a + cos t) cos t, (a + cos t) sin t), where
a is a real number. Show that the curve intersects itself for |a| < 1.
Describe what happens for a = 0.
15. Let Γ be a circle and let O a point on Γ. Set L to be the line tangent
to Γ such that its point of tangency is diametrically opposite from O.
The cissoid of Diocles relative to Γ and O is the set of points P such
←→
that OP = AB, where A is the intersection of OP and Γ and where
←→
B is the intersection of OP and L. Using the setup in the following
diagram, find parametric equations for the cissoid of Diocles.
y
B
A
O a x
Γ L
1.2 Position, Velocity, and Acceleration

⃗
In physics, one interprets the vector function X(t) = (x(t), y(t)) as
providing the location along a curve at time t in reference to some
fixed frame. A frame is a (usually orthonormal) basis attached to a
fixed origin. The point O = (0, 0) along with the basis {⃗ı, ⃗ȷ}, where
⃗ı = (1, 0) and ⃗ȷ = (0, 1),
⃗
form the standard reference frame. We call the vector function X(t)
the position vector. When one uses the standard reference frame, it
is not uncommon to write
⃗
X(t) = x(t)⃗ı + y(t)⃗ȷ.
Directly imitating Newton’s approach to finding the slope of a

curve at a certain point, it is natural to ask the question, “What is
the direction and rate of change of a curve X ⃗ at point t0 ?” If we
⃗ ⃗ 1 ), the change in
look at two points on the curve, say X(t0 ) and X(t
⃗ ⃗
position is given by the vector X(t1 ) − X(t0 ), while to specify the
rate of change in position, we would need to scale this vector by a
⃗ 0)
factor of t1 − t0 . Hence, the vectorial rate of change between X(t
⃗ 1 ) is
and X(t
1 ⃗
⃗ 0) .
X(t1 ) − X(t
t1 − t0
To define an instantaneous rate of change along the curve at the
point t0 , we need to calculate (if it exists and if it has meaning) the
limit
⃗ ′ (t0 ) = lim 1 X(t

X ⃗ 0 + h) − X(t
⃗ 0) . (1.2)
h→0 h
⃗
According to Proposition 1.1.6, if X(t) = (x(t), y(t)), then the limit in
Equation (1.2) exists if and only if x(t) and y(t) are both differentiable
⃗ ′ (t0 ) = (x′ (t0 ), y ′ (t0 )).
at t0 . Furthermore, if this limit exists, then X
This leads us to the following definition.
Definition 1.2.1 Let X ⃗ : I → R2 be a vector function with coordi-

⃗
nates X(t) = (x(t), y(t)). We say that X ⃗ is differentiable at t = t0
if x(t) and y(t) are both differentiable at t0 . If J is the common
domain to x′ (t) and y ′ (t), then we define the derivative of the vec-
⃗ as the new vector function X
tor function X ⃗ ′ : J → R2 defined by
⃗ ′ ′ ′
X (t) = (x (t), y (t)).
If x(t) and y(t) are both differentiable functions on their domain,

the derivative vector function X ⃗ ′ (t) = (x′ (t), y ′ (t)) is called the ve-
locity vector. Mathematically, this is just another vector function
and traces out another curve when placed in the standard reference
frame. However, because it illustrates the direction of motion along
the curve, one often visualizes the velocity vector corresponding to
t = t0 as based at the point X(t ⃗ 0 ) on the curve.
Following physics language, we call the second derivative of the
vector function X⃗ ′′ (t) = (x′′ (t), y ′′ (t)) the acceleration vector related
⃗
to X(t).
In general, our calculations often require that our vector functions
can be differentiated at least once and sometimes more. Consequently,
when establishing theorems, we like to succinctly describe the largest
class of functions for which a particular result holds.
Definition 1.2.2 Let I be an interval of R, and let f : I → R be a

function or X ⃗ : I → R2 a parametrized curve. We say that f or X ⃗
r ⃗
is of class C on I if the rth derivative of X exists and is continuous
on I. We denote by C ∞ the class of functions or parametrized curves
that have derivatives of all orders on I. It is also common to write
f ∈ C r (I, R) to say that f is a real-valued function of class C r and
to write X ⃗ ∈ C r (I, R2 ) to say that X
⃗ is a parametrized curve in the
r
plane of class C .
To say that a function is of class C 0 over the interval I means that

it is continuous. By basic theorems in calculus, the condition that a
function be of a certain class is an increasingly restrictive condition.
In other words, as sets of functions defined over the same interval I,
the classes are nested according to
C 0 ⊃ C 1 ⊃ C 2 ⊃ · · · ⊃ C ∞.
Whenever we impose a condition that a function is of class C r for some

r, such a supposition is generically called a “smoothness condition.”
Proposition 1.2.3 Let ⃗v (t) and w(t)

⃗ be vector functions defined and
differentiable over an interval I ⊂ R. Then the following hold:
⃗
1. If X(t) ⃗ ′ (t) = c⃗v ′ (t).
= c⃗v (t), where c ∈ R, then X
⃗
2. If X(t) = c(t)⃗v (t), where c : I → R is a real function, then
⃗ ′ (t) = c′ (t)⃗v (t) + c(t)⃗v ′ (t).
X
⃗
3. If X(t) = ⃗v (t) + w(t),
⃗ then
⃗ ′ (t) = ⃗v ′ (t) + w
X ⃗ ′ (t).
⃗

4. If X(t) = ⃗v f (t) is a vector function and f : J → I is a real
function into I, then
⃗ ′ (t) = f ′ (t)⃗v ′ f (t) .

X
5. If f (t) = ⃗v (t) · w(t)

⃗ is the dot product between ⃗v (t) and w(t),
⃗
then
f ′ (t) = ⃗v ′ (t) · w(t) ⃗ ′ (t).
⃗ + ⃗v (t) · w
Example 1.2.4 Consider the spiral defined by X(t)⃗ = (t cos t, t sin t)

for t ≥ 0. The velocity and acceleration vectors are
X⃗ ′ (t) = (cos t − t sin t, sin t + t cos t),

⃗ ′′ (t) = (−2 sin t − t cos t, 2 cos t − t sin t).
X
If we wish to calculate the angle between X ⃗ and X ⃗ ′ as a function of

t, we use the dot product as follows. We calculate
p
⃗
∥X(t)∥ = t2 sin2 t + t2 cos2 t = |t|,
p
⃗ ′ (t)∥ = (cos t − t sin t)2 + (sin t + t cos t)2 = 1 + t2 ,
p
∥X
⃗
X(t) ·X⃗ ′ (t) = t cos2 t − t2 cos t sin t + t sin2 t + t2 cos t sin t = t.
⃗
Thus, the angle θ(t) between X(t) ⃗ ′ (t) is defined for all t ̸= 0
and X
and is equal to
!
⃗
X(t) ·X⃗ ′ (t)
t

−1
θ(t) = cos = cos−1 √ .
⃗
∥X(t)∥ ⃗ ′ (t)∥
∥X |t| 1 + t2
⃗ ′ (t0 )
X
⃗
∆X
⃗ 1)
X(t
⃗ 0)
X(t
O
Figure 1.5: Arc length segment.
⃗
Let C be the locus of a vector function X(t) for t ∈ [a, b]. Then
we approximate the length of the arc l(C) with the Riemann sum
n
X n
X
l(C) ≈ ∥X(t ⃗ i−1 )∥ ≈
⃗ i ) − X(t ⃗ ′ (ti )∥∆t.
∥X
i=1 i=1
⃗ i) −
(See Figure 1.5 as an illustration for the approximation ∥X(t
⃗ ⃗ ′
X(ti−1 )∥ ≈ ∥X (ti )∥∆t.) Taking the limit of this Riemann sum, we
obtain the following formula for the arc length of C:
Z bp
l= (x′ (t))2 + (y ′ (t))2 dt. (1.3)
a
A rigorous justification for the arc length formula, using the Mean
Value Theorem, can be found in most calculus textbooks.
In light of Equation (1.3), we define the arc length function s :
⃗
[a, b] → R related to X(t) as the arc length along C over the interval
[a, t]. Thus,
Z t Z tp
s(t) = ⃗ ′ (u)∥ du =
∥X (x′ (u))2 + (y ′ (u))2 du. (1.4)
a a
By the Fundamental Theorem of Calculus, we then have

⃗ ′ (t)∥ = (x′ (t))2 + (y ′ (t))2 ,
p
s′ (t) = ∥X
which, still following the vocabulary from the trajectory of a moving

⃗
particle, we call the speed function of X(t).
Example 1.2.5 Consider the parametrized curve defined by X(t) ⃗ =

(t2 , t3 ). The velocity vector is X⃗ ′ (t) = (2t, 3t2 ), and thus the arc
length function from t = 0 is
Z t Z tp Z t r
⃗ ′ 2 4
4
s(t) = ∥X (u)∥ du = 4u + 9u du = 3|u| + u2 du
0 0 0 9
3/2 !
4 2 8
= sign(t) +t − ,
9 27
where sign(t) = 1 if t > 0 and −1 if t < 0 and sign(0) = 0. The

⃗
length of X(t) between t = 0 and t = 1 is
1
133/2 − 8 .

s(1) − s(0) =
27
The speed function allows us to understand that a parametrized
curve X⃗ : I → R2 does not only contain the information that describes
the locus of the curve but it also contains information about the speed
of travel s(t) along the locus. By looking at the locus, it is impossible
to discern the speed function.
Consider a reparametrization of the parametrized curve X ⃗ ◦ f,
where f : J → I is a differentiable surjective (i.e., onto) function. If
⃗ (u0 )) =
t0 = f (u0 ) where f is a differentiable function at u0 , then X(f
⃗
X(t0 ) and
d ⃗ ⃗ ′ (f (u0 ))∥ = |f ′ (u0 )| ∥X

⃗ ′ (t0 )∥
X(f (u0 )) = ∥f ′ (u0 )X
dt
= |f ′ (u0 )|s′ (t0 ).
Consequently, under a reparametrization, the unit vector associated

⃗ ′ (t) only changes by a factor of +1 or −1. More precisely, if
with X
⃗ ⃗ ⃗ ′ (f (t)) ̸= ⃗0,
ξ = X ◦f , at any parameter value t where f ′ (t) ̸= 0 and X
we have
ξ⃗ ′ (t) ⃗ ′ (f (t))
f ′ (t) X
= ′ , (1.5)
∥ξ⃗ ′ (t)∥ ⃗ ′ (f (t))∥
|f (t)| ∥X
where f ′ (t)/|f ′ (t)| is +1 or −1. It bears repeating that in order for
the right-hand side to be well defined, we need f ′ (t) ̸= 0.
Definition 1.2.6 Let X ⃗ : I → R2 be a parametrized curve. For a

continuously differentiable function f from an interval J onto I, we
call ξ⃗ = X
⃗ ◦ f a regular reparametrization if for all t ∈ J, f ′ (t) is
well defined and never 0. In addition, a regular reparametrization is

called positively oriented (resp. negatively oriented ) if f ′ (t) > 0 (resp.
f ′ (t) < 0) for all t ∈ J.
By the Mean Value Theorem, if f ′ (t) ̸= 0 for all t ∈ J, then f :

J → I is an injective function. Thus since f is by definition surjective,
regular reparametrizations involve a bijective function f : J → I.
Equation (1.5) shows that the unit vector X ⃗ ′ (t)/∥X
⃗ ′ (t)∥ is invari-
ant under a positively oriented reparametrization and simply changes
sign under a negatively oriented reparametrization. Consequently,
this unit vector is an important geometric object associated to a curve
at a point.
Definition 1.2.7 Let X ⃗ : I → R2 be a plane parametric curve. A

⃗
point t0 ∈ I is called a critical point of X(t) ⃗
if X(t) is not differentiable
⃗ ′ ⃗
at t0 or if X (t0 ) = 0. If t0 is a critical point, then X(t ⃗ 0 ) is called
a critical value. A point t = t0 that is not critical is called a regular
point. A parametrized curve X ⃗ : I → R2 is called regular if it is of
1
class C (i.e., continuously differentiable) and X ⃗ ′ (t) ̸= ⃗0 for all t ∈ I.
⃗
Finally, if t is a regular point of X, we define the unit tangent vector
T⃗ (t) as
⃗ ′ (t)
X
T⃗ (t) = .
⃗ ′ (t)∥
∥X
It is not uncommon to call a property of a curve C or a property
of a point P on a curve C a geometric property if it does not depend
on a parametrization of the locus near the point. In particular, a
geometric property does not depend on the speed of travel along the
curve through the point.
At any point of a curve where T⃗ (t) is defined, that is to say at
any regular point, we can write the velocity vector as
⃗ ′ (t) = s′ (t)T⃗ (t).
X (1.6)
This expresses the velocity vector as the product of its magnitude

(speed) and direction (unit tangent vector). In most differential ge-
ometry texts, authors simplify their formulas by reparametrizing by
arc length. From the perspective of coordinates, this means picking
an “origin” O on the curve C occurring at some fixed t = t0 and us-
ing the arc length along C between O and any other point P as the
parameter with which to locate P on C. In fact, this can always be
done.
⃗
Proposition 1.2.8 If X(t) is a regular parametrized curve, then there
is a regular reparametrization of X⃗ by arc length. Furthermore, if X ⃗
is of class C k , then the arc length reparametrization is also of class

Ck.
Proof : Let s = f (t) be the arc length function with s = 0 corre-
sponding to some point on the curve. Since X ⃗ is regular over its
′ ⃗ ′
domain, then f (t) = ∥X (t)∥ > 0 for all t. Since f (t) is strictly in-
creasing, it has an inverse function t = h(s). By the Inverse Function
Theorem, since f ′ (t) ̸= 0, then h(s) is differentiable with
1 1
h′ (s) = = . (1.7)
f ′ (h(s)) f ′ (t)
⃗ (s) = X(h(s))
Note that the composite function Y ⃗ satisfies
⃗′
⃗ ′ (s) = X
Y ⃗ ′ (t) 1 = X (t) .
⃗ ′ (h(s))h′ (s) = X
f ′ (t) ⃗ ′ (t)∥
∥X
⃗ is of class C k , then s′ (t) = f ′ (t) = ∥X
If X ⃗ ′ (t)∥ is of class C k−1 .
k
Hence s(t) itself is of class C . The Inverse Function Theorem states
that the inverse function is also of class C k . Consequently, the arc
length parametrization ⃗y (s) is of class C k since it is the composition
⃗
of X(t) of class C k and t = h(s) of class C k . □
The Inverse Function Theorem as used in the above proof (and its
more general multivariable counterpart) appear in most introductory
texts on analysis. (See [7] for example.) However, the key derivative
in Equation (1.7) follows from the chain rule for invertible functions
and is a standard differentiation formula.
The habit of reparametrizing by arc length has benefits and draw-
backs. The great benefit of this approach is that along a curve para-
metrized by arc length, the speed function s′ is identically 1 and the
velocity vector is exactly the tangent vector
⃗ ′ (s) = T⃗ (s).
X
This formulation simplifies proofs and difficult calculations. The main
drawback is that, in practice, most curves do not admit a simple for-
mula for their arc length function, let alone a formula that can be
written using elementary functions. For example, given an explicit
⃗
parametrized curve X(t), it is often very challenging to find s(t) as de-
fined in Equation (1.4). Furthermore, to reparametrize by arc length,
it would be necessary to find the inverse function t(s), representing
the original parameter t as a function of s. Even if it is possible to
find s(t), determining this inverse function usually cannot be written
with elementary functions. Then the parametrization by arc length
⃗
is X(s) ⃗
= X(t(s)). Even using a computer algebra systems (CAS),
reparametrizing by arc length remains an intractable problem.
The notion of “regular” in many ways mirrors geometric properties

of continuously differentiable single-variable functions. In particular,
⃗
if a parametrized curve X(t) is regular at t0 , then locally the curve
looks linear. As we will see again in Chapter 3, with the formalism of
vector functions it becomes particularly easy to express the equation
of the tangent line to a curve at a point in any number of dimensions.
⃗
If X(t) is a curve and t0 is not a critical point for the curve, then the
equation L⃗ t (t) of the tangent line at t0 is
0
⃗ t (t) = X(t
L ⃗ 0 ) + (t − t0 )X
⃗ ′ (t0 ), with t ∈ R,
0
or alternatively
⃗ t (u) = X(t
L ⃗ 0 ) + uT⃗ (t0 ), with u ∈ R, (1.8)
0
if we do not wish to confuse the parameter of the parametrized curve

⃗
X(t) and the parameter of the tangent line.
On the other hand, if t0 ∈ I is a critical point for the curve
⃗
X(t), the curve may or may not have a tangent line at t = t0 . If
the following one-sided limits exist, we can define two unit tangent
⃗ 0 ):
vectors at X(t
T⃗ (t+ ⃗
0 ) = lim+ T (t) and T⃗ (t− ⃗
0 ) = lim− T (t).
t→t0 t→t0
Consequently, we can determine the angle the curve makes with itself
at the corner t0 as

α0 = cos−1 T⃗ (t+ ⃗ −
0 ) · T (t0 ) .
To be more precise, the angle from T⃗ (t− ⃗ +

0 ) to T (t0 ) is the exterior
angle of the curve at the corner t = t0 . One can only define a tangent
⃗
line to X(t) at t = t0 if T⃗ (t− ⃗ +
0 ) = T (t0 ).
Example 1.2.9 Let X(t) ⃗ = (t2 , t3 ). We calculate that X ⃗ ′ (t) =

2
(2t, 3t ), and therefore t = 0 is a critical point because X⃗ (0) = ⃗0.
′
However, if t ̸= 0, the unit tangent vector is

1 t
T⃗ (t) = √ (2t, 3t2 ) = √ (2, 3t).
2
4t + 9t 4 |t| 4 + 9t2
Then the right-hand and left-hand side unit tangent vectors are
T⃗ (0− ) = (−1, 0) and T⃗ (0+ ) = (1, 0).
⃗
These calculations indicate that as t approaches 0, X(t) lies in the
fourth quadrant but approaches (0, 0) in the horizontal direction (−1, 0).
From an intuitive standpoint, we could say that X ⃗ stops at t = 0,
◦
spins around by 180 , and moves away from the origin in the direction
(1, 0), remaining in the first quadrant.
Example 1.2.10 Let X(t) ⃗ = (t, | tan t|). We calculate that X⃗ ′ (t) =
2
(1, sign(t) sec t) and deduce that t = 0 is a critical point because
sign(t) is not defined at 0, and hence X ⃗ is not differentiable there.
However, we also find that
1 1
T⃗ (0− ) = √ (1, −1) and T⃗ (0+ ) = √ (1, 1).
2 2
Thus,
T⃗ (0− ) · T⃗ (0+ ) = 0,
which shows that X ⃗ makes a right angle with itself at t = 0. However,
one can tell from the explicit values for T⃗ (t− ⃗ +
0 ) and T (t0 ) that the
π
exterior angle of the curve at t = 0 is 2 .
If t0 is a critical point, it may still happen that
lim T⃗ (t)
t→t0
exists, namely when T⃗ (t+ ⃗ −

0 ) = T (t0 ), which also means α0 = 0. Then
through an abuse of language, we can still talk about the unit tangent
vector at that point. As an example of this possibility, consider the
⃗
curve X(t) = (t3 , t4 ). We can quickly calculate that
lim T⃗ (t) = (1, 0) = ⃗ı.

t→0
When this limit exists, even though t0 is a critical point, one can
use Equation (1.8) as parametric equations for the tangent line at t0 ,
replacing T⃗ (t0 ) in Equation (1.8) with limt→t0 T⃗ (t).
Though we will not make use of the following comments until
Chapter 2, we finish this chapter on calculus concepts with a quick
review of notation for integration.
In geometry, physics, and other applications, we must sometimes
integrate a function along a curve. To this end, we use path and
line integrals of scalar functions or vector functions, depending on
the particular problem. Since we will make use of them, we remind
the reader of the notation for such integrals. Let C be a curve para-
⃗
metrized by X(t) over the interval [a, b]. Let f : R2 → R be a real
(scalar) function and F⃗ : R2 → R2 a vector field in the plane. Then
the path integral of f and the line integral of F⃗ over the curve C are
respectively
Z Z b
def p
f ds = f (x(t), y(t)) (x′ (t))2 + (y ′ (t))2 dt,
C a
Z Z b
def
F⃗ · d⃗s = F⃗ (x(t), y(t)) · T⃗ (t) dt.
C a
Problems
In Problems 1.2.1 through 1.2.4, calculate, the velocity, the acceleration,

the speed, and where defined, the unit tangent vector function of the given
parametrized curve.
⃗
1. X(t) = (t2 − 1, t3 − t) for t ∈ R.
⃗
2. The circle X(t) = (R cos ωt, R sin ωt).
3. The circle as parametrized in Equation (1.1) in Example 1.1.10.
4. The epicycloids defined in Problem 1.1.7.
5. Using linear algebra, it is possible to prove that the shortest distance
between a point (x0 , y0 ) and a line with equation ax + by + c = 0 is
|ax0 + by0 + c|
d= √ .
a2 + b 2
However, a calculus proof of this result is also possible using para-
metrized curves.
(a) Assuming a ̸= 0, show that X(t)⃗ = (−bt − c/a, at) with t ∈ R
is a parametrization of the line with equation ax + by + c = 0.
(b) Find the value t0 of t that minimizes the function
⃗
p
f (t) = ∥X(t) − (x0 , y0 )∥ = (−bt − c/a − x0 )2 + (at − y0 )2 ,
which gives the distance between a point X(t) ⃗ on the line and
the point (x0 , y0 ).
(c) Show that f (t0 ) simplifies to the distance formula given above
for d.
⃗ be a fixed point, and let ⃗l(t) = ⃗at+⃗b be the parametric equations
6. Let p
of a line. Prove that the distance between p ⃗ and the line is
s
(⃗a · (⃗b − p
⃗))2 ∥⃗a × (⃗b − p
⃗)∥
∥⃗b − p
⃗∥2 − = ,
∥⃗a∥2 ∥⃗a∥
where for vectors ⃗v = (v1 , v2 , 0) and w

⃗ = (w1 , w2 , 0) in the plane, we
call ⃗v × w
⃗ = (0, 0, v1 w2 − v2 w1 ), which is the cross product between
⃗v and w ⃗ when viewed as vectors in R3 . [See the previous problem.]
7. Find the closest point to (16, 0.5) on the curve X(t) ⃗ = (t, t2 ).
8. A quadratic Bézier curve with non-collinear control points p ⃗0 , p
⃗1 ,
and p ⃗2 is a curve X ⃗ : [0, 1] → R2 with end points X(0) ⃗ = p⃗0 and
⃗
X(1) =p ⃗2 , whose component functions are quadratic polynomials,
and such that the control point p ⃗1 satisfies X⃗ ′ (0) = λ(⃗
p1 − p
⃗0 ) and
X⃗ ′ (1) = µ(⃗ p2 − p⃗1 ), for some positive constants λ and µ.
(a) Prove that the above conditions imply that λ = µ = 2.
(b) Prove that a parametrization for the quadratic Bézier is X(t) ⃗ =
2 2
(1 − t) p ⃗0 + 2t(1 − t)⃗ p1 + t p⃗0 with t ∈ [0, 1].
9. Find the tangent line to X(t) ⃗ = (cos(2t), sin t) at t = π/6. Also give
the components of the unit tangent vector there.
10. We say that a plane curve C parametrized by X ⃗ : I → R2 intersects

⃗
itself at a point p if there exist u ̸= t in I such that X(t) ⃗
= X(u) = p.
Consider the parametric curve X(t) ⃗ = (t2 , t3 − t) for t ∈ R. Find the
point(s) of self-intersection. Also determine the angle at which the
curve intersects itself at those point(s) by finding the angle between
the unit tangent vectors corresponding to the distinct parameters.
11. The parametrized curve X(t) ⃗ = ((1 + 2 cos t) cos t, (1 + 2 cos t) sin t)
intersects itself at one point. Find this point of intersection and
find the angle of self-intersection (i.e., the acute angle between the
tangent lines corresponding to the two different parameters t1 and
t2 of self-intersection).
12. For how many points on the Lissajous curve X(t) ⃗ = (cos(3t), sin(2t))
does the tangent line go through the point (3, 0)?
13. Consider the Lissajous figure X(t) ⃗ = (cos(mt), sin(nt)). Prove that
this curve intersects itself 2mn − (m + n) times.
14. What can be said about a parametrized curve X(t) ⃗ that has the
⃗ ′′
property that X (t) is identically 0?
15. Consider the linear spiral given in Example 1.2.4. Suppose that we
reparametrize the spiral by t = f (u) = tan(u) with 0 ≤ u < π/2. (a)
Show that this reparametrization is regular and positively oriented.
(b) Prove that with this reparametrization the angle between the
position and the velocity vectors is exactly u.
16. Find the arc length function along the parabola y = x2 , using as the
origin s = 0.
17. Sketch the parametrized curve X(t) ⃗ = (3t2 , 3t − t3 ) near the origin
and calculate the length of the loop in the curve.
⃗
18. Consider the curve X(t) = (t2 , t3 ) with t ∈ R. Show that the repa-
rametrization of this curve by arclength is
3/2 !
⃗ 1 2/3 4 1 2/3 4
X(s) = (27s + 8) − , sign(s) (27s + 8) − .
9 9 9 9
19. (*) Consider the cycloid introduced in Example 1.1.13 given by X(t) ⃗ =
(t − sin t, 1 − cos t). Prove that the path taken by a point on the edge
of a rolling wheel of radius 1 during one rotation has length 8.
20. Calculate the arc length function of the curve X(t) ⃗ = (t2 , ln t), de-
fined for t > 0.
21. Let X⃗ : I → R2 be a regular parametrized curve, and let p ⃗ be a fixed
point. Suppose that the closest point on the curve X ⃗ to p
⃗ occurs at
t = t0 , which is neither of the ends of I. Prove that the line between
p
⃗ and the point closest to p ⃗ 0 ), is perpendicular to the
⃗, namely X(t
curve X ⃗ at t = t0 .
22. Prove the differentiation formulas in Proposition 1.2.3.
⃗
23. Prove that if X(t) ⃗ ·X
is a curve that satisfies X ⃗ ′ = 0 for all values of
⃗ is a circle. [Hint: Use ∥X∥
t, then X ⃗ =X
2 ⃗ ·X
⃗ and apply Proposition
⃗ 2 .]
1.2.3 to calculate the derivative of ||X||
⃗
24. Consider the ellipse given by X(t) = (a cos t, b sin t). Find the ex-
tremum values of the speed function.
25. Consider the linear spiral of Example 1.2.4. Let n ≥ 0 be a non-
negative integer. Prove that the length of the nth derivative vector
function is given by
p
⃗ (n) (t)∥ =
∥X n2 + t 2 .
x(t) = (aebt cos t, aebt sin t) where a

26. Consider the exponential spiral ⃗
and b are constants. Calculate the arc length s(t) function of ⃗ x(t).
Reparametrize the spiral by arc length.
x(t) = (aebt cos t, aebt sin t),
27. Consider again the exponential spiral ⃗
with a > 0 and b < 0.
⃗
(a) Prove that as t → +∞, lim X(t) = (0, 0).
⃗ ′
(b) Show that X (t) → (0, 0) as t → +∞ and that for any t0 ,
Z t
lim ⃗ ′ (u)∥ du < ∞,
∥X
t→∞ t0
i.e., any part of the exponential spiral that “spirals” in toward

the origin has finite arc length.
28. Recall that polar and Cartesian coordinate systems are related as
follows:
( ( p
x = r cos θ, r = x2 + y 2 ,
and
y = r sin θ, tan θ = xy
Suppose C is a curve in the plane parametrized using polar coordi-

nate functions r = r(t) and θ = θ(t) so that one has a parametriza-
tion in Cartesian coordinates as
⃗
x(t) = (x(t), y(t)) = (r(t) cos(θ(t)), r(t) sin(θ(t))).
(a) Express x′ and y ′ in terms of r, θ, r′ , and θ′ .

(b) Express r′ and θ′ in terms of x, y, x′ , and y ′ .
(c) Express ||⃗ x′ || in terms of polar coordinate functions.
x|| and ||⃗
(All coordinate functions are viewed as functions of t and so r′ , for
example, is a shorthand for r′ (t), the derivative of r with respect to
t.)
29. (ODE) Suppose a curve C is parametrized by ⃗ x(t) such that ⃗x(t)
and ⃗x′ (t) always make a constant angle with each other. Find the
shape of this curve. [Hint: Use polar coordinates and the results of
Problem 1.2.28.]
1.3. Curvature 23
1.3 Curvature
Let X⃗ : I → R2 be a twice-differentiable parametrization of a curve
C. As we saw in the previous section, the decomposition of the ve-
⃗ ′ = s′ T⃗ into magnitude and unit tangent direction
locity vector X
separates the geometric invariant (the unit tangent T⃗ ) from the dy-
namical aspect (the speed s′ (t)) of the parametrization. Taking one
more derivative, we obtain the decomposition
⃗ ′′ = s′′ T⃗ + s′ T⃗ ′ .
X (1.9)
The first component describes a tangential acceleration, while the

second component describes a rate of change of the tangent direction,
or in other words, how much the curve is “curving.” Since T⃗ is a unit
vector, we always have T⃗ · T⃗ = 1. Therefore,
(T⃗ · T⃗ )′ = 0 =⇒ 2T⃗ · T⃗ ′ = 0,
=⇒ T⃗ · T⃗ ′ = 0.
Thus, T⃗ ′ is perpendicular to T⃗ .
Just as there are two unit tangent vectors at a regular point of the
curve, there are two unit normal vectors as well. Given a particular
parametrization, there is no naturally preferred way to define “the”
unit normal vector, so we make a choice.
Definition 1.3.1 Let X ⃗ : I → R2 be a regular parametrized curve

and T⃗ = (T1 , T2 ) the tangent vector at a regular value X(t).
⃗ The unit
⃗
normal vector U is
⃗ (t) = (−T2 (t), T1 (t)).

U
Equivalently, U ⃗ is the vector function obtained by rotating T⃗ by π .

2
In still other words, if we view the xy-plane in three-dimensional
space with x-, y-, and z-axes oriented in the usual way, and if we call
⃗k = (0, 0, 1) the unit vector along the positive z-axis, then U
⃗ = ⃗k × T⃗ .
Since T⃗ ′ ⊥ T⃗ , the vector function T⃗ ′ is a multiple of U

⃗ at all t.
This leads to the definition of curvature.
Definition 1.3.2 Let X ⃗ : I → R2 be a regular twice-differentiable

parametric curve. The curvature function κg (t) is the unique real-
valued function defined by
T⃗ ′ (t) = s′ (t)κg (t)U

⃗ (t). (1.10)
This definition only gives κg (t) implicitly, but we can obtain a

formula for it as follows. Equation (1.9) becomes
⃗ ′′ = s′′ T⃗ + (s′ )2 κg U
X ⃗. (1.11)
⃗ = ⃗k.
Viewing the plane as the xy-plane in three-space, we have T⃗ × U
Thus,
⃗ ′×X
X ⃗ ′′ = (s′ T⃗ ) × (s′′ T⃗ + (s′ )2 κg U
⃗)
= s′ s′′ T⃗ × T⃗ + (s′ )3 κg T⃗ × U⃗
= (s′ )3 κg⃗k.
⃗ ′ (t)∥, which leads to

But s′ (t) = ∥X
⃗ ′ (t) × X
(X ⃗ ′′ (t)) · ⃗k x′ (t)y ′′ (t) − x′′ (t)y ′ (t)
κg (t) = = . (1.12)
∥X ⃗ ′ (t)∥3 (x′ (t)2 + y ′ (t)2 )3/2
As straightforward as Formula (1.12) may seem, we offer a word

of wisdom to the student. As with many calculations in differential
geometry, it is very useful to simplify algebraic and trigonometric
expressions as much as possible. Not doing so can quickly lead to
intractable expressions.
The reader might well wonder why in the definition of κg (t) we
include the factor s′ (t). It is not hard to confirm (see Problem 1.3.20)
that including the s′ (t) factor renders κg (t) independent of any regular
reparametrization (except perhaps up to a change of sign).
⃗
One should note at this point that if a curve X(s) is parametrized
by arc length, then by Equations (1.6) and (1.9), the velocity and
acceleration have the simple expressions
X⃗ ′ (s) = T⃗ (s),
⃗ ′′ (s) = κg (s)U
X ⃗ (s).
The curvature of a curve at a point has an interpretation that is

easy to visualize. We explore this interpretation in the next three
examples.
Example 1.3.3 Consider the vector function that describes a par-

ticle moving on a circle of radius R with a nonzero and constant
⃗
angular velocity ω > 0 given by X(t) = (R cos ωt, R sin ωt). In order
to calculate the curvature, we need
⃗ ′ (t) = (−Rω sin ωt, Rω cos ωt),
X
⃗ ′′ (t) = (−Rω 2 cos ωt, −Rω 2 sin ωt).
X
1.3. Curvature 25
Thus, using the coordinate form (right-most expression) in Equation

(1.12), we get
R2 ω 3 sin2 t + R2 ω 3 cos2 t R2 ω 3 1
κg (t) = 2 2 2 2 2 2 3/2
= 3 3
= .
(R ω sin t + R ω cos t) R ω R
The curvature of the circle is a constant function that is equal to the

reciprocal of the radius for all t, regardless of the nonzero angular
velocity ω.
The study of trajectories in physics gives particular names for
the components of the first and second derivatives of a vector func-
tion in the basis {T⃗ , U ⃗ }. We already saw that the function s′ (t) in
Equation (1.6) is called the speed. In Equation (1.11), however, the
function s′′ (t) is called the tangential acceleration, while the quan-
tity s′ (t)2 κg (t) is called the centripetal acceleration. Example 1.3.3
connects the curvature function to the reciprocal of a radius, so if
κg (t) ̸= 0, then we define the radius of curvature to X(t)⃗ at t as the
function R(t) = κg (t) . Using the common notation v(t) = s′ (t) for
1
the speed function, one recovers the common formula for centripetal
acceleration of
v2
(s′ )2 κg = .
R
In introductory physics courses, this formula is presented only in the
context of circular motion. With differential geometry at our disposal,
we see that centripetal acceleration is always equal to v 2 /R at all
points on a curve where κg ̸= 0, where by R one means the radius of
curvature.
In contrast to physics where one tends to refer to the radius of cur-
vature, the curvature function is a more geometrically natural quan-
tity to study. Indeed, the radius of curvature of a line segment is
undefined even though line segments are such useful geometric ob-
jects. On the other hand, a radius of curvature is 0 (and hence the
curvature is undefined) at degenerate curves that are points or at
critical points.
Example 1.3.4 Let I be a real interval and consider the plane curve
given as the graph y = f (x) of a twice-differentiable function f : I →
R. We can parametrize this as X(t)⃗ = (t, f (t)) for t ∈ I. A quick
application of (1.12) gives a curvature function of
f ′′ (t)
κg (t) = .
(1 + f ′ (t)2 )3/2
Expressing the same function back in the variable x, we have κg (x) =

f ′′ (x)/(1 + f ′ (x)2 )3/2 . This calculation underscores that κg (x) = 0
B H
B
C A C
F G
A
E
D E
G
H D F
Figure 1.6: Curvature function example.
p
precisely when f ′′ (x) = 0. Recalling from calculus that 1 + f ′ (x)2 is
the speed function s′ (x) along the curve, we see from yet another angle
that the curvature function of a function graph is κg = f ′′ (x)/s′ (x)3 .
Note that the curvature function κg may be positive or negative.

The previous example along with the typical calculus terminology
motivates the following definition.
Definition 1.3.5 Let C be a regular curve of class C 2 . An inflection

point of C is any point where the curvature function κg changes sign.
Example 1.3.6 To further develop an intuition for curvature, con-

sider the parametric curve X(t)⃗ = (2 cos t, sin(2t) + sin t) for t ∈
[0, 2π].
Figure 1.6 shows side by side the locus of the curve and the graph
of its curvature function, along with a few labeled points to serve as
references. First, we see that the curvature function is positive when
the curve turns to the left away from the unit tangent vector T⃗ and
negative when the curve turns to the right. Though it is possible
for the curvature to become zero but not change signs, at the points
labeled C and G, the curvature function changes sign. These are
inflection points. Precisely here, the curve changes from curving to
the left of T⃗ to curving to the right of T⃗ , or vice versa. The curve is
locally curving the “most” when κg (t) has a maximum and κg (t) > 0
or when κg (t) has a minimum and κg (t) < 0. These cases correspond
respectively to curving the most to the left and to the right. Besides
a situation when the curvature is 0, the curve is locally curving the
“least” when κg (t) has a minimum and κg (t) > 0 or when κg (t) has a
maximum and κg (t) < 0. Figure 1.6 illustrates all five of the different
possibilities just described.
1.3. Curvature 27
Proposition 1.3.7 A regular parametrized curve X ⃗ : I → R2 has

⃗ is a line
curvature κg (t) = 0 for all t ∈ I if and only if the locus of X
segment.
Proof : If the locus of X ⃗ traces out a line segment, then (perhaps
after a regular reparametrization) we can write X(t) ⃗ = ⃗a + φ(t)⃗b,
′
where φ(t) is a differentiable real function with φ (t) ̸= 0. Then
X⃗ ′ (t) = φ′ (t)⃗b,
⃗ ′′ (t) = φ′′ (t)⃗b.
X
Thus, by Equation (1.12),
(φ′ (t)φ′′ (t)⃗b × ⃗b) · ⃗k

κg (t) = =0
(|φ′ (t)| ∥⃗b∥)3/2
because ⃗b × ⃗b = ⃗0.
Conversely, if X(t)⃗ = (x(t), y(t)) is a curve such that κg (t) = 0,
then
x′ (t)y ′′ (t) − x′′ (t)y ′ (t) = 0.
We need to find solutions to this differential equation or determine
how solutions are related. Since X ⃗ is regular, X
⃗ ′ ̸= ⃗0 for all t ∈ I.
′
Let I1 be an interval where y (t) ̸= 0. Over I1 , we have
x′ y ′′ − x′′ y ′ d x′ x′
= = 0 =⇒ = C,
(y ′ )2 dt y ′ y′
where C is a constant. Thus, x′ = Cy ′ for all t. Integrating with
respect to t we deduce x(t) = Cy(t)+D. Similarly, over an interval I2
where x′ (t) ̸= 0, we deduce that y(t) = Ax(t) + B for some constants
A and B. Since I can be covered with intervals where x′ (t) ̸= 0 or
⃗ is a piecewise linear curve.
y ′ (t) ̸= 0, we deduce that the locus of X
⃗
However, since X is regular, it has no corners, and hence its locus is
a line segment. □
Our formula for curvature given in (1.12) arose from calculating

X⃗ ′′ as the derivative of the expression X ⃗ ′ = s′ T⃗ , which involved
d ⃗ ⃗
finding an expression for dt T (t). Taking the third derivative of X(t)
and using the decomposition in (1.11), we get
⃗ ′′′ = s′′′ T⃗ + (3s′′ s′ κg + (s′ )2 κ′g )U
X ⃗ + (s′ )2 κg U
⃗ ′.
We now need an expression for the derivative U ⃗ ′ (t). By definition,

for all t, the set {T⃗ , U
⃗ } forms an orthonormal basis and hence
⃗ ′ = (U
U ⃗ ′ · T⃗ )T⃗ + (U
⃗ ′·U
⃗ )U
⃗.
⃗ (t) is a unit vector function, we have U

However, since U ⃗ ·U
⃗ ′ = 0 for
⃗ · T⃗ = 0 for all t, we deduce that
all t. Furthermore, since U
⃗ ′ · T⃗ = −U
U ⃗ · T⃗ ′ = −s′ κg .
Consequently,
⃗ ′ (t) = −s′ (t)κg (t)T⃗ (t).
U
Since {T⃗ , U
⃗ } forms an orthonormal basis of R2 for all t, every
⃗
higher derivative of X(t), if it exists, can be expressed as a linear
combination of T⃗ and U ⃗ . The components of X ⃗ (n) (t) in terms of
T⃗ and U ⃗ involve sums of products of derivatives of s(t) and κg (t).
Furthermore, if X ⃗ is parametrized by arc length, then the coefficients
⃗ (n)
of X (s) only involve sums of powers of derivatives of κg (s).
We note that if a curve X ⃗ : [−a, a] → R2 is reparametrized by
⃗
X(−t), then the modified curvature function is −κg (t). Thus, the sign
of the curvature depends on what one might call the “orientation”
of the curve. Excluding this technicality, curvature has a physical
interpretation that one can eyeball on particular curves. If a curve
is almost a straight line, then the curvature is close to 0, but if a
regular curve bends tightly along a certain section, then the curvature
is high (in absolute value). Of particular interest are points where the
curvature reaches a local extremum.
Definition 1.3.8 Let C be a regular curve parametrized by X ⃗ :I→

2
R with curvature function κg (t). A vertex of the curve C is a point
⃗ 0 ) where the curvature function κg (t) attains an extremum.
P = X(t
By the First Derivative Test, extrema of κg (t) occur at t = t0 ,

if κ′g (t) changes sign through t0 . This obviously must occur where
κ′g (t0 ) = 0 but the converse is not necessarily true.
Problems
In Problems 1.3.1 through 1.3.7, calculate the curvature function of the

given curve.
1. The curve parametrized by X(t)⃗ = (tm , tn ) for t ≥ 0.
⃗
2. The ellipse: X(t) = (a cos t, b sin t) for t ∈ [0, 2π].
⃗
3. The cycloid: X(t) = (t − sin t, 1 − cos t) for t ∈ R.
⃗
4. The astroid: X(t) = (cos3 t, sin3 t) for t ∈ [0, 2π].
⃗
5. The curve parametrized by X(t) = (ecos t , esin t ) for t ∈ [0, 2π].
⃗
6. The linear spiral: X(t) = (t cos t, t sin t) for t ∈ R≥0 .
⃗
7. The flower curve: X(t) = (sin(nt) cos t, sin(nt) sin t) for t ∈ [0, 2π].
Another random document with
no related content on Scribd:
Lincoln enviously how he had got hold of these things. He answered
laconically that he had “liberated” them.
Captain Blyth had good news for me the next morning. Two
armored vehicles had left Munich for the mine. I gave the message
to George when he called just before noon.
“They’re a little late,” he said. “Thanks to the fine co-operation of
the 11th Armored Division, I am being taken care of from this end of
the line. I’ll try to catch your fellows in Salzburg and tell them to go
back where they came from. I am sending you a letter by the next
convoy. It’s about a repository which ought to be evacuated right
away. I can’t give you any of the details over the phone without
violating security regulations. As soon as Posey gets back, you
ought to go to work on it. After that I want you to help me here.”
After George had hung up, I asked Lincoln if he had any idea what
repository George had in mind. Lincoln said it might be the
monastery at Hohenfurth. It was in Czechoslovakia, just over the
border from Austria. While we were discussing this possibility, Craig
Smyth walked in.
As soon as he had recovered from the surprise of finding me at
Captain Posey’s desk, he explained the reason for his visit.
Something had to be done right away about the building he was
setting up as a collecting point. He had been promised a twenty-four-
hour guard. He had been promised a barrier of barbed wire. So far,
Third Army had failed to provide either one. A lot of valuable stuff
had already been delivered to the building, and George was sending
in more. He couldn’t wait any longer.
“Let’s have a talk with Colonel Hamilton,” I said. As we walked
down to the Assistant Chief of Staff’s office, Craig told me that the
buildings he had requisitioned were the ones Bancel La Farge had
suggested when we saw him at Wiesbaden a month ago.
“Can’t this matter wait until Captain Posey returns?” the colonel
asked.
“I am afraid it can’t, sir,” said Craig. “As you know, the two
buildings were the headquarters of the Nazi party. The Nazis meant
to destroy them before Munich fell. Having failed to do so, I think it
quite possible that they may still attempt it. Both buildings are
honeycombed with underground passageways. Only this morning we
located the exit of one of them. It was half a block from the building.
We hadn’t known of its existence before. The works of art stored in
the building at present are worth millions of dollars. In the
circumstances, I am not willing to accept the responsibility for what
may happen to them. I must have guards or a barrier at once.”
The colonel reached for the telephone and gave orders that a
cordon of guards was to be placed on the buildings immediately. He
made a second call, this time about the barbed wire. When he had
finished he told Craig that the guard detail would report that
afternoon; the barbed wire would be strung around the building the
next morning. We thanked the colonel and returned to Posey’s office.
We found two officers who had just come in from Dachau. They
were waiting to see someone connected with Property Control. They
had brought with them a flour sack filled with gold wedding rings; a
large carton stuffed with gold teeth, bridgework, crowns and braces
(in children’s sizes); a sack containing gold coins (for the most part
Russian) and American greenbacks. As we looked at these
mementos of the concentration camp, I thought of the atrocity film I
had seen at Versailles and wondered how anyone could believe that
those pictures had been an exaggeration.
I went back with Craig to his office at the Königsplatz. The damage
to Munich was worse than I had realized. The great Deutsches
Museum by the river was a hollow-eyed specter, but sufficiently
intact to house DPs. Aside from its twin towers, little was left of the
Frauenkirche. The buildings lining the Brienner-Strasse had been
blasted and burned. Along the short block leading from the Carolinen
Platz to the Königsplatz, the destruction was total: on the left stood
the jagged remnants of the little villa Hitler had given D’Annunzio; on
the right was a heap of rubble which had been the Braun Haus. But
practically untouched were the two Ehrentempel—the memorials to
the “martyrs” of the 1923 beer-hall “putsch.” The colonnades were
draped with the same kind of green fishnet that had been used to
camouflage the Haus der Deutschen Kunst. The classic façades of
the museums on either side of the square—the Glyptothek and the
Neue Staatsgalerie—were intact. The buildings themselves were a
shambles.
I followed Craig up the broad flight of steps to the entrance of the
Verwaltungsbau, or Administration Building. It was three stories high,
built of stone, and occupied almost an entire block. True to the Nazi
boast, it looked as though it had been built to “last a thousand
years.” And it was so plain and massive that I didn’t see how it could
change much in that time. There was nothing here that could “grow
old gracefully.” The interior matched the exterior. There were two
great central courts with marble stairs leading to the floor above.
Although the building had not been bombed, it had suffered
severely from concussion. Craig said that when he had moved in two
weeks ago the skylights over the courts had been open to the sky.
On rainy days one could practically go boating on the first floor.
There had been no glass in the windows. Now they had been
boarded up or filled in with a translucent material as a substitute. All
of the doors had been out of line and would not lock. But the repairs
were already well under way and, according to Craig, the place
would be shipshape in another month or six weeks. He said that his
colleague, Hamilton Coulter, a former New York architect now a
naval lieutenant, was directing the work and doing a magnificent job.
Even under normal conditions it would have been a staggering task.
With glass and lumber at a premium, to say nothing of the scarcity of
skilled labor, a less resourceful man would have given up in despair.
Vermeer’s Portrait of the Artist in His Studio in the Alt Aussee
mine was purchased by Hitler for his proposed museum. It has
been returned to Vienna.
One of the picture storage rooms in the mine, constructed with

wooden partitions and racks. The temperature was constant,
averaging 40° Fahrenheit in summer, 47° in winter.
Panel from the Louvain altarpiece, Feast of the
Passover, by Bouts, was stored at Alt Aussee.
Hitler acquired the Czernin Vermeer for an alleged
price of 1,400,000 Reichsmarks.
Craig was rapidly building up a staff of German scholars and

museum technicians to assist him in the administration of the
establishment. It would soon rival a large American museum in
complexity and scope. Storage rooms on the ground floor had been
made weatherproof. Paintings and sculpture were already pouring in
from the mine at Alt Aussee—six truckloads at a time. In accordance
with standard museum practice, Craig had set up an efficient
accessioning system. As each object came in, it was identified,
marked and listed for future reference. Quarters had been set aside
for a photographer. Racks were being built for pictures. A two-storied
record room was being converted into a library.
Craig had also requisitioned the “twin” of this colossal building—
the Führerbau, where Hitler had had his own offices. This was only a
block away on the same street and also faced the Königsplatz. It
was connected with the Verwaltungsbau by underground
passageways. It was in the Führerbau that the Munich Pact of 1938
—the pact that was to have guaranteed “peace in our time”—had
been signed. Craig showed me the table at which Mr. Chamberlain
had signed that document. Craig was using it now for a conference
table.
Repairs were being concentrated on the Administration Building,
since its “twin” was being held in reserve for later use. At the
moment, however, a few of the rooms were occupied by a small
guard detail. The truck drivers and armed guards who came each
week with the convoys from the mine were also billeted there.
Just as Craig and I were finishing our inspection of the Führerbau,
a convoy of six trucks, escorted by two half-tracks, pulled into the
parking space behind the building. The convoy leader had a letter for
me. It was the one George Stout had mentioned on the telephone.
Lincoln was right. The repository George had in mind was the
monastery at Hohenfurth. In his letter he stressed the fact that the
evacuation should be undertaken at once. He suggested that I try to
persuade Posey to send Lincoln along to help me.
Craig had a comfortable billet in the Kopernikus-Strasse, a four-
room flat on the fourth floor of a modern apartment building. The
back windows looked onto a garden. Over the tops of the poplar
trees beyond, one could see the roof of the Prinz Regenten Theater
where, back in the twenties, I had seen my first complete
performance of Wagner’s Ring. Craig told me that the theater was
undamaged except for the Speisesaal where, in prewar days, lavish
refreshments were served during intermissions. That one room had
caught a bomb.
I accepted Craig’s invitation to share the apartment with him while
in Munich and made myself at home in the dining room. It had a
couch, and there was a sideboard which I could use as a chest of
drawers. The bathroom was across the hall and he said that the
supply of hot water was inexhaustible. By comparison, the officers’
billets at Third Army Headquarters were tenements.
Ham Coulter had similar quarters on the ground floor. We stopped
there for a drink on our way to supper at the Military Government
Detachment. “Civilized” was the word that best described Ham. He
was a tall, broad-shouldered fellow with sleek black hair, finely-
chiseled features and keen, gray eyes. When he smiled, his mouth
crinkled up at the corners, producing an agreeably sarcastic
expression. Ham poured out the drinks with an elegance the ordinary
German cognac didn’t deserve. They should have been dry martinis.
I liked him at once that first evening, and when I came to know him
better I found him the wittiest and most amiable of companions. He
and Craig were a wonderful combination. They had the greatest
admiration and respect for each other, and during the many months
of their work together there was not the slightest disagreement
between them.
The officers’ dining room that evening was a noisy place. The
clatter of knives and forks and the babble of voices mingled with the
rasping strains of popular American tunes pounded out by a
Bavarian band. As we were about to sit down, the music stopped
abruptly and a second later struck up the current favorite, “My
Dreams Are Getting Better All the Time.” Colonel Charles Keegan,
the commanding officer, had entered the dining hall. I was puzzled
by this until Ham told me that it happened every night. This particular
piece was the colonel’s favorite tune and he had ordered it to be
played whenever he came in to dinner. The colonel was a colorful
character—short and florid, with a shock of white hair. He had
figured prominently in New York politics and would again, it was said.
He had already helped Craig over a couple of rough spots and if
some of his antics amused the MFA&A boys, they seemed to be
genuinely fond of him.
While we were at dinner, three officers came and took their places
at a near-by table. Craig identified one of them as Captain Posey. I
had been told that he was in his middle thirties, but he looked
younger than that. He had a boyish face and I noticed that he
laughed a great deal as he talked with his two companions. When he
wasn’t smiling, there was a stubborn expression about his mouth,
and I remembered that Lincoln had said something about his
“deceptively gentle manner.” On our way out I introduced myself to
him. He was very affable and seemed pleased at my arrival. But he
was tired after the long drive from Frankfurt, so, as soon as I had
arranged to meet him at his office the next day, I joined Ham and
Craig back at the apartment.
I got out to Third Army Headquarters early the following morning
and found Captain Posey already at his desk, going through the
papers which had accumulated during his absence. We discussed
George’s letter at considerable length, and I was disappointed to find
that he did not intend to act on it at once. Somehow it had never
occurred to me that anyone would question a proposal of George’s.
For one thing, Captain Posey said, he couldn’t spare Lincoln; and
for another, there was a very pressing job much nearer Munich
which he wanted me to handle. He then proceeded to tell me of a
small village on the road to Salzburg where I would find a house in
which were stored some eighty cases of paintings and sculpture
from the Budapest Museum. He showed me the place on the map
and explained how I was to go about locating the exact house upon
arrival. I was to see a certain officer at Third Army Headquarters
without delay and make arrangements for trucks. He thought I would
need about five. It shouldn’t take more than half a day to do the job if
all went well. After this preliminary briefing, I was on my own.
My first move was to get hold of the officer about the trucks. How
many did I need? When did I want them and where should he tell
them to report? I said I’d like to have five trucks at the Königsplatz
the following morning at eight-thirty. The officer explained that the
drivers would be French, as Third Army was using a number of
foreign trucking companies to relieve the existing shortage in
transportation.
Later in the day I had a talk with Lincoln about my plans and he
gave me a piece of advice which proved exceedingly valuable—that
I should go myself to the trucking company which was to provide the
vehicles, and personally confirm the arrangements. So, after lunch I
struck out in a jeep for the west side of Munich, a distance of some
seven or eight miles.
I found a lieutenant, who had charge of the outfit, and explained
the purpose of my visit. He had not been notified of the order for five
trucks. There was no telephone communication between his office
and headquarters, so all messages had to come by courier, and the
courier hadn’t come in that day. He was afraid he couldn’t let me
have any trucks before the following afternoon. I insisted that the
matter was urgent and couldn’t wait, and, after much deliberating
and consulting of charts, he relented. I told him a little something
about the expedition for which I wanted the trucks and he showed
real interest. He suggested that I ought to have extremely careful
drivers. I replied that I should indeed, as we would be hauling stuff of
incalculable value.
Thereupon he gave me a harrowing description of the group under
his supervision. All of them had been members of the French
Resistance Movement—ex-terrorists he called them—and they
weren’t afraid of God, man or the Devil. Well, I thought, isn’t that
comforting! “Oh, yes,” he said, “these Frenchies drive like crazy men.
But,” he continued, “one of the fellows has got some sense. I’ll see if
I can get him for you.” He went over to the window that looked out on
a parking ground littered with vehicles of various kinds. Here and
there I saw a mechanic bent over an open hood or sprawled out
beneath a truck. The lieutenant bellowed, “Leclancher, come up here
to my office!”
A few seconds later a wiry, sandy-haired Frenchman of about
forty-five appeared in the doorway. Leclancher understood some
English, for he reacted with alert nods of the head as the lieutenant
gave a brief description of the job ahead, and then turned and asked
me if I spoke French. I told him I did, if he had enough patience. This
struck him as inordinately funny, but I was being quite serious. What
really pleased him was the fact that I was in the Navy. He said that
he had been in the Navy during the first World War. Then and there a
lasting bond was formed, though I didn’t appreciate the value of it at
the time.
Eight-thirty the next morning found me pacing the Königsplatz. Not
a truck in sight. Nine o’clock and still no trucks. At nine thirty, one
truck rolled up. Leclancher leaped out and with profuse apologies
explained that the other four were having carburetor trouble. There
had been water in the gas, too, and that hadn’t helped. For an hour
Leclancher and I idled about, whiling away the time with
conversation of no consequence, other than that it served to limber
up my French. At eleven o’clock Leclancher looked at his watch and
said that it would soon be time for lunch. It was obvious that he
understood the Army’s conception of a day as a brief span of time, in
the course of which one eats three meals. If it is not possible to finish
a given job during the short pauses between those meals, well,
there’s always the next day. I told him to go to his lunch and to come
back as soon as he could round up the other trucks. In the meantime
I would get something to eat near by.
When I returned shortly after eleven thirty—having eaten a K
ration under the portico of the Verwaltungsbau in lieu of a more
formal lunch—my five trucks were lined up ready to go. I appointed
Leclancher “chef de convoi”—a rather high-sounding title for such a
modest caravan—and he assigned positions to the other drivers,
taking the end truck himself. Since my jeep failed to arrive, I climbed
into the lead truck. The driver was an amiable youngster whose
name was Roger Roget. During the next few weeks he was the lead
driver in all of my expeditions, and I took to calling him “Double
Roger,” which I think he never quite understood.
To add to my anxiety over our belated start, a light rain began to
fall as we pulled out of the Königsplatz and turned into the Brienner-
Strasse. We threaded our way cautiously through the slippery streets
choked with military traffic, crossed the bridge over the Isar and
swung into the broad Rosenheimer-Strasse leading to the east.
Once on the Autobahn, Roger speeded up. The speedometer
needle quivered up to thirty-five, forty and finally forty-five miles an
hour. I pointed to it, shaking my head. “We must not exceed thirty-
five, Roger.”
He promptly slowed down, and as we rolled along, I forgot the
worries of the morning. I dozed comfortably. Suddenly we struck an
unexpected hole in the road and I woke up. We were doing fifty. This
time I spoke sharply, reminding Roger that the speed limit was thirty-
five and that we were to stay within it; if we didn’t we’d be arrested,
because the road was well patrolled. With a tolerant grin Roger said,
“Oh, no, we never get arrested. The MPs, they stop us and get very
angry, but—” with a shrug of the shoulders—“we do not understand.
They throw the hands up in the air and say ‘dumb Frenchies’ and we
go ahead.”
“That may work with you,” I said crossly, “but what about me? I’m
not a ‘dumb Frenchy.’”
For the next hour I pretended to doze and at the same time kept
an eye on the speedometer. This worked pretty well. Now and again
I would look up, and each time, Roger would modify his speed.
Presently we came to a bad detour, where a bridge was out. We
had to make a sharp turn to the left, leave the Autobahn and
descend a steep and tortuous side road into a deep ravine. That day
the narrow road was slippery from the rain, so we had to crawl along.
The drop into the valley was a matter of two or three hundred feet
and, as we reached the bottom, we could see the monstrous
wreckage of the bridge hanging drunkenly in mid-air. The ascent was
even more precarious, but our five trucks got through.
We had now left the level country around Munich and were in a
region of rolling hills. Along the horizon, gray clouds half concealed
the distant peaks. Soon the rain stopped and the sun came out. The
mountains changed to misty blue against an even bluer sky. The
road rose sharply, and when we reached the crest, I caught a
glimpse of shimmering water. It was Chiemsee, largest of the
Bavarian lakes.
In another ten minutes the road flattened out again and we came
to the turnoff marked “Prien.” There we left the Autobahn for a
narrow side road which took us across green meadows. Nothing
could have looked more peaceful than this lush, summer
countryside. Reports of SS troops still hiding out in the near-by
forests seemed preposterous in the pastoral tranquillity. Yet only a
few days before, our troops had rounded up a small band of these
die-hards in this neighborhood. The SS men had come down from
the foothills on a foraging expedition and had been captured while
attempting to raid a farmhouse. It was because of just such
incidents, as well as the ever-present fire hazard, that I had been
sent down to remove the museum treasures to a place of safety.
The road was dwindling away to a cow path and I was beginning
to wonder how much farther we could go with our two-and-a-half-ton
trucks, when we came to a small cluster of houses. This was
Grassau. I had been told that a small detachment of troops was
billeted there, so I singled out the largest of the little white houses
grouped around the only crossroads in the village. It had clouded
over and begun to rain again. As I entered the gate and was
crossing the yard, the door of the house was opened by a corporal.
He didn’t seem surprised to see me. Someone at Munich had sent
down word to Prien that I was coming, and the message had
reached him from there. I asked if he knew where the things I had
come for were stored. He motioned to the back of the house and
said there were two rooms full of big packing cases. He explained
that he and one other man had been detailed to live in the house
because of the “stuff” stored there. They had been instructed to keep
an eye on the old man who claimed to be responsible for it. That
would be Dr. Csanky, director of the Budapest Museum, who,
according to my information, would probably raise unqualified hell
when I came to cart away his precious cases. The corporal told me
that the old man and his son occupied rooms on the second floor.
I was relieved to hear that they were not at home. It would make
things much simpler if I could get my trucks loaded and be on my
way before they returned. It was already well after two and I wanted
to start back by five at the latest. I asked rather tentatively about the
chances of getting local talent to help with the loading, and the
corporal promptly offered to corral a gang of PWs who were working
under guard near by.
While he went off to see about that, I marshaled my trucks. There
was enough room to back one truck at a time to the door of the
house. A few minutes later the motley “work party” arrived. There
were eight of them in all and they ranged from a young fellow of
sixteen, wearing a faded German uniform, to a reedy old man of
sixty. By and large, they looked husky enough for the job.
I knew enough not to ask my drivers to help, but knew that the
work would go much faster if they would lend a hand. Leclancher
must have read my thoughts, for he immediately offered his services.
As soon as the other four saw what Leclancher was doing, they
followed suit.
There were eighty-one cases in all. They varied greatly in size,
because some of them contained sculpture and, consequently, were
both bulky and heavy. Others, built for big canvases, were very large
and flat but relatively light. We had to “design” our loads in such a
way as to keep cases of approximately the same type together. This
was necessary for two reasons: first, the cases would ride better that
way, and second, we hadn’t any too much space. As I roughly
figured it, we should be able to get them all in the five trucks, but we
couldn’t afford to be prodigal in our loading.
The work went along smoothly for an hour and we were just
finishing the second truck when I saw two men approaching. They
were Dr. Csanky and his son. This was what I had hoped to avoid.
The doctor was a dapper little fellow with a white mustache and very
black eyes. He was wearing a corduroy jacket and a flowing bow tie.
The artistic effect was topped off by a beret set at a jaunty angle. His
son was a callow string bean with objectionably soulful eyes
magnified by horn-rimmed spectacles. They came over to the truck
and began to jabber and wave their arms. We paid no attention
whatever—just kept on methodically lifting one case after another
out of the storage room.
The dapper doctor got squarely in my path and I had to stop. I
checked his flow of words with a none too civil “Do you speak
English?” That drew a blank, so I asked if he spoke German. No luck
there either. Feeling like an ad for the Berlitz School, I inquired
whether he spoke French. He said “Yes,” but the stream of
Hungarian-French which rolled out from that white mustache was
unintelligible. It was hopeless. At last I simply had to take him by the
shoulders and gently but firmly set him aside. This was the ultimate
indignity, but it worked. At that juncture he and the string bean took
off. I didn’t know what they were up to and I didn’t care, so long as
they left us alone.
Our respite was short-lived. By the time we had the third truck
ready to move away, they were back. And they had come with
reinforcements: two women were with them. One was a rather
handsome dowager who looked out of place in this rural setting. Her
gray hair, piled high, was held in place by a scarlet bandanna, and
she was wearing a shabby dress of green silk. Despite this getup,
there was something rather commanding about her. She introduced
herself as the wife of General Ellenlittay and explained in perfect
English that Dr. Csanky had come to her in great distress. Would I
be so kind as to tell her what was happening so that she could
inform him?
“I am removing these cases on the authority of the Commanding
General of the Third United States Army,” I said. If she were going to
throw generals’ names around, I could produce one too—and a
better one at that.
Her response to my pompous pronouncement was delivered
charmingly and with calculated deflationary effect. “Dear sir, forgive
me if I seemed to question your authority. That was not my intention.
It is quite apparent that you are removing the pictures, but where are
you taking them?”
“I am sorry, but I am not at liberty to say,” I replied. That too
sounded rather lordly, but I consoled myself by recalling that Posey
had admonished me not to answer questions like that.
She relayed this information to Dr. Csanky and the effect was
startling. He covered his face with his hands and I thought he was
going to cry. Finally he pulled himself together and let forth a flood of
unintelligible consonants. His interpreter tackled me again.
“Dr. Csanky is frantic. He says that he is responsible to his
government for the safety of these treasures and since you are
taking them away, there is nothing left for him to do but to blow his
brains out.”
My patience was exhausted. I said savagely, “Tell Dr. Csanky for
me that he can blow his brains out if he chooses, but I think it would
be silly. If you must know, Madame, I am a museum director myself
and you can assure him that no harm is going to come to his
precious pictures.”
I should never have mentioned that I had even so much as been
inside a museum, for, from that moment until we finished loading the
last truck, the little doctor never left my side. I was his “cher
collègue,” and he kept up a steady barrage of questions which the
patient Madame Ellenlittay tried to pass along to me without
interrupting our work.
As we lined the trucks up, preparatory to starting back to Munich,
Dr. Csanky produced several long lists of what the cases contained.
He asked me to sign them. This I refused to do but explained to him
through the general’s wife that if he cared to have the lists translated
from Hungarian and forwarded to Third Army Headquarters, they
could be checked against the contents and eventually returned to
him with a notation to that effect.
Our little convoy rolled out of Grassau at six o’clock, leaving the
group of Hungarians waving forlornly from the corner. Just before we
turned onto the Autobahn, Leclancher signaled from the rear truck
for us to stop. He came panting up to the lead truck with a bottle in
his hand. With a gallant wave of the arm he said that we must drink
to the success of the expedition. It was a bottle of Calvados—fiery
and wonderful. We each took a generous swig and then—with a
rowdy “en voiture!”—we were on our way again. It was a nice
gesture.
The trip back to Munich was uneventful except for the
extraordinary beauty of the long summer evening. The sunset had all
the extravagance of the tropics. The sky blazed with opalescent
clouds. As we drove into Munich, the whole city was suffused with a
coral light which produced a more authentic atmosphere of
Götterdämmerung than the most ingenious stage Merlin could have
contrived.
It was nearly nine when we rolled into the parking area behind the
Führerbau—too late to think about unloading and also, I was afraid,
too late to get any supper. We had pieced out with K rations and
candy bars, but were still hungry. A mess sergeant, lolling on the
steps of the building, reluctantly produced some lukewarm stew.
After we had eaten, I prevailed on one of the building guards to take
my five drivers out to their billets south of town. It was Saturday
night. I told Leclancher we would probably be making a longer trip on
Monday and that I would need ten drivers. He promised to select five
more good men, and we arranged to meet in the square as we had
that morning. But Monday, he promised, they would be on time.
After checking the tarpaulins on my five trucks, I sauntered over to
the Central Collecting Point on the off-chance that Craig might be
working late. I found him looking at some of the pictures which
George had sent in that day from the mine. The German packers
whom Craig had been able to hire from one of the old established
firms in Munich—one which had worked exclusively for the museums
there—had finished unloading the trucks only a couple of hours
earlier. Most of the things in this shipment had been found at the
mine, so now the pictures were stacked according to size in neat
rows about the room. In one of them we found two brilliant portraits
by Mme. Vigée-Lebrun. Labels on the front identified them as the
likenesses of Prince Schuvalov and of the Princess Golowine. Marks
on the back indicated that they were from the Lanckoroncki
Collection in Vienna, one of the most famous art collections in
Europe. Hitler was rumored to have acquired it en bloc—through
forced sale, it was said—for the great museum he planned to build at
Linz. In another stack we came upon a superb Rubens landscape, a
fine portrait by Hals and two sparkling allegorical scenes by Tiepolo.
These had no identifying labels other than the numbers which
referred to lists we didn’t have at the moment. However, Craig said
that the documentation on the pictures, as a whole, was surprisingly
complete. Then we ran into a lot of nineteenth century German
masters—Lenbach, Spitzweg, Thoma and the like. These Hitler had
particularly admired, but they didn’t thrill me. I was getting sleepy
and suggested that we had had enough art for one day. I still had a
report of the day’s doings to write up for Captain Posey before I
could turn in, so we padlocked the room and took off.
Even though the next day was Sunday, it was not a day of rest for
me. The trek to Hohenfurth, scheduled for Monday, involved infinitely
more complicated preliminary arrangements than the easy run of
yesterday. Captain Posey got out maps of the area into which the
convoy would be traveling; gave me the names of specific outfits
from whom I would have to obtain clearance as well as escorts along
the way; and, most important of all, supplied information concerning
the material to be transported. None of it, I learned, was cased. He
thought it consisted mainly of paintings, but there was probably also
some furniture. This wasn’t too definite. Nor did we have a very clear
idea as to the exact number of trucks we would need. I had spoken
for ten on the theory that a larger number would make too
cumbersome a convoy. At least I didn’t want to be responsible for
more at that stage of the game, inexperienced as I was. In the
circumstances, two seasoned packers might, I thought, come in
handy, so I was to see if I could borrow a couple of Craig’s men.
There was the problem of rations for the trip up and back. Posey
procured a big supply of C rations, not so good as the K’s, but they
would do. While at Hohenfurth we would be fed by the American
outfit stationed there. It was a good thing that there was so much to
be arranged, because it kept me from worrying about a lot of things
that could and many that did happen on that amazing expedition.
In the afternoon I went out to make sure of my trucks and on the
way back put in a bid for the two packers. My request wasn’t very
popular, because Craig was shorthanded, but he thought he could
spare two since it was to be only a three-day trip—one day to go up,
one day at Hohenfurth, and then back the third day. Craig gloomily
predicted that I’d never get ten trucks loaded in one day, but I airily
tossed that off with the argument that we had loaded five trucks at
Grassau in less than four hours. “But that stuff was all in cases,” he
said. “You’ll find it slow going with loose pictures.” Of course he was
right, but I didn’t believe it at the time.
(4)
MASTERPIECES IN A MONASTERY
We woke up to faultless weather the following morning, and on the

way down to the Königsplatz with Craig I was in an offensively
optimistic frame of mind. All ten of my trucks were there. This was
more like it—no tiresome mechanical delays. We were all set to go.
Leclancher had even had the foresight to bring along an extra driver,
just in case anything happened to one of the ten. That was a smart
idea and I congratulated him for having thought of it.
It wasn’t till I started distributing the rations that I discovered our
two packers were missing. But that shouldn’t take long to straighten
out. Craig’s office was just across the way. I found them cooling their
heels in the anteroom. They looked as though they had come right
out of an Arthur Rackham illustration—stocky little fellows with
gnarled hands and wizened faces as leathery as the Lederhosen
they were wearing. Each wore a coal-scuttle hat with a jaunty
feather, and each had a bulging bandanna attached to the end of a
stick. There was much bowing and scraping. The hats were doffed
and there was the familiar “Grüss Gott, Herr Kapitän,” when I walked
in.
Craig appeared and explained the difficulty. Until the last minute,
no one had thought to ask whether the men had obtained a Military
Government permit to leave the area—and of course they hadn’t.
With all due respect to the workings of Military Government, I knew
that it would take hours, even days, to obtain the permits. So, what
to do? I asked Craig what would happen if they went without them.
He didn’t know. But if anyone found out about it, there would be
trouble. I didn’t see any reason why anyone should find out about it.
The men would be in my charge and I said I’d assume all
responsibility.

Full Ebook of Differential Geometry of Curves and Surfaces Third Edition Thomas F Banchoff Online PDF All Chapter

Uploaded by

Copyright:

Available Formats

You might also like

Full Ebook of Differential Geometry of Curves and Surfaces Third Edition Thomas F Banchoff Online PDF All Chapter

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

Full Ebook of Differential Geometry of Curves and Surfaces Third Edition Thomas F Banchoff Online PDF All Chapter

Uploaded by

Copyright:

Available Formats

Differential Geometry of Curves and

Surfaces: Third Edition Thomas F.

Topological, Differential and Conformal Geometry of

Tensor Algebra and Analysis for Engineers: With

Parametric Geometry of Curves and Surfaces

Differential Geometry of Plane Curves 96th Edition

Differential Geometry 1st Edition Victor V. Prasolov

Manifolds, Vector Fields, and Differential Forms: An

CRC Standard Curves and Surfaces with Mathematica 3rd

Differential Geometry and General Relativity Volume 1

ISBN: 9781032281094 (hbk)

1 Plane Curves: Local Properties 1

2 Plane Curves: Global Properties 43

3 Curves in Space: Local Properties 65

4 Curves in Space: Global Properties 91

5 Regular Surfaces 111

5.3 Regular Surfaces . . . . . . . . . . . . . . . . . . . . . 128

6 First and Second Fundamental Forms 149

7 Fundamental Equations of Surfaces 219

8 Gauss-Bonnet Theorem; Geodesics 241

9 Curves and Surfaces in n-Dimensional Space 313

A Tensor Notation 335

What Is Differential Geometry?

of a point, or global properties, which refer to properties of the curve

Using This Textbook

is an analogue of the fundamental theorem of calculus. The global

advent of computer graphics has renewed the interest in classical dif-

6.1, 6.3–6.6 A.1

6.2 6.7 7.1–7.2

Note that Appendix A.1 on tensor notation is valuable for back-

Changes from the Second to the Third Edition

 adding more exercises, both introductory and advanced;

 improving the explanations in select areas;

 rewriting the computer demos into Maple, as described earlier.

We would like to thank Dawson Bremner, Jonathan Higgins, and

lists dozens of them, together with descriptions of their contributions.

dents on developing electronic books for differential geometry and

Second Edition Acknowledgments

Thomas F. Banchoff is a geometer and a professor at Brown Uni-

Stephen Lovett is a professor of mathematics at Wheaton College.

Plane Curves: Local Properties

Just as calculus courses introduce real functions of one variable be-

Example 1.1.1 (Lines) Euclid’s first postulate of geometry is that

Figure 1.1: A line in the plane.

and p⃗2 = (x2 , y2 ), the vector

⃗v = p⃗2 − p⃗1 = (x2 − x1 , y2 − y1 )

Example 1.1.2 (Circles) The pair of functions

traces out the unit circle in a counterclockwise manner. Multiplying

Definition 1.1.3 Let ⃗v be a vector in R2 with coordinates ⃗v =

If p and q are two points in Rn with coordinates given by vectors ⃗v

Definition 1.1.4 Let X ⃗ be a vector function from a subset of R into

Definition 1.1.5 Let I be an open interval of R, let a ∈ I, and let

The above definitions mirror the usual definition of a limit of a real

Proposition 1.1.6 Let X ⃗ be a vector function from a subset of R

Corollary 1.1.7 Let I be an open interval of R, let a ∈ I, and con-

Definition 1.1.5 and Corollary 1.1.7 provide the mathematical

Definition 1.1.8 Let I be an interval of R. A parametrized curve (or

It is important to note the distinction in this definition between

adding more exercises, both introductory and advanced;

improving the explanations in select areas;

rewriting the computer demos into Maple, as described earlier.