Professional Documents
Culture Documents
Full download Acquisition and Processing of Marine Seismic Data 1st Edition Derman Dondurur file pdf all chapter on 2024
Full download Acquisition and Processing of Marine Seismic Data 1st Edition Derman Dondurur file pdf all chapter on 2024
https://ebookmass.com/product/real-time-data-acquisition-in-
human-physiology-real-time-acquisition-processing-and-
interpretation-a-matlab-based-approach-dipali-bansal/
https://ebookmass.com/product/interpreting-subsurface-seismic-
data-rebecca-bell/
https://ebookmass.com/product/intelligent-multi-modal-data-
processing-the-wiley-series-in-intelligent-signal-and-data-
processing-1st-edition-soham-sarkar/
https://ebookmass.com/product/tensors-for-data-processing-1st-
edition-yipeng-liu/
Volume III: Data Storage, Data Processing and Data
Analysis 1st ed. 2021 Edition Volker Liermann (Editor)
https://ebookmass.com/product/volume-iii-data-storage-data-
processing-and-data-analysis-1st-ed-2021-edition-volker-liermann-
editor/
https://ebookmass.com/product/big-data-in-astronomy-scientific-
data-processing-for-advanced-radio-telescopes-1st-edition-linghe-
kong-editor/
https://ebookmass.com/product/annotated-atlas-of-coastal-and-
marine-winds-1st-edition-nazla-bushra/
https://ebookmass.com/product/tensors-for-data-processing-theory-
methods-and-applications-yipeng-liu/
https://ebookmass.com/product/machine-intelligence-big-data-
analytics-and-iot-in-image-processing-ashok-kumar/
ACQUISITION AND PROCESSING
OF MARINE SEISMIC DATA
This page intentionally left blank
ACQUISITION
AND
PROCESSING
OF MARINE
SEISMIC DATA
DERMAN DONDURUR
Elsevier
Radarweg 29, PO Box 211, 1000 AE Amsterdam, Netherlands
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States
# 2018 Elsevier Inc. All rights reserved.
No part of this publication may be reproduced or transmitted in any form or by any means,
electronic or mechanical, including photocopying, recording, or any information storage
and retrieval system, without permission in writing from the publisher. Details on how to
seek permission, further information about the Publisher’s permissions policies and our
arrangements with organizations such as the Copyright Clearance Center and the Copyright
Licensing Agency, can be found at our website: www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright
by the Publisher (other than as may be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and
experience broaden our understanding, changes in research methods, professional
practices, or medical treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in
evaluating and using any information, methods, compounds, or experiments described
herein. In using such information or methods they should be mindful of their own safety
and the safety of others, including parties for whom they have a professional
responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or
editors, assume any liability for any injury and/or damage to persons or property as a
matter of products liability, negligence or otherwise, or from any use or operation of any
methods, products, instructions, or ideas contained in the material herein.
Library of Congress Cataloging-in-Publication Data
A catalog record for this book is available from the Library of Congress
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library
ISBN: 978-0-12-811490-2
1. Introduction 5. Preprocessing
1.1 Underwater Acoustics 7 5.1 Demultiplexing 242
1.2 Marine Acoustic Methods 9 5.2 Data Loading 243
1.3 Fundamentals of Marine 5.3 Single Trace Section 244
Seismics 25 5.4 Geometry Definition 246
5.5 Band-Pass Filter 253
2. Marine Seismic Data Acquisition 5.6 Gain Recovery 272
5.7 Trace Edit 287
2.1 Components of Marine Seismic
5.8 Muting 294
Acquisition 40
5.9 f-k Dip Filters 299
2.2 Air Gun Arrays 62
5.10 Brute Stack 310
2.3 3D Marine Seismic Acquisition 81
2.4 Specific Acquisition Techniques 100
2.5 Data Acquisition Parameters 112 6. Deconvolution
2.6 QC in Data Acquisition 131 6.1 Convolutional Model 316
6.2 Assumptions for Deconvolution 319
3. Noise in Marine Seismics 6.3 Spiking Deconvolution 324
6.4 Predictive Deconvolution 335
3.1 Operational Noise 174
6.5 Determination of Deconvolution Parameters 337
3.2 Bubble Effect of the
6.6 Poststack Deconvolution 349
Air Gun 175
6.7 Maximum Entropy (Burg) Deconvolution 349
3.3 Multiple Reflections 177
6.8 Shaping Filters 350
3.4 Swell Noise 180
6.9 Surface Consistent Deconvolution 356
3.5 Bird Noise 183
6.10 QC in Deconvolution 357
3.6 Inline Waves 185
3.7 Diffractions 193
3.8 Guided Waves 193 7. Suppression of Multiple Reflections
3.9 Seismic Interference 196 7.1 CDP Stack 368
3.10 Other Noise Types 200 7.2 Predictive Deconvolution 372
7.3 Deconvolution in τ-p Domain 382
4. Fundamentals of Data Processing 7.4 Radon Velocity Filter 383
7.5 Wave Equation Multiple Rejection 388
4.1 Autocorrelation 213
7.6 f-k Filtering 391
4.2 Crosscorrelation 215
7.7 Surface-Related Multiple Elimination 394
4.3 Convolution 215
7.8 QC in Multiple Suppression 400
4.4 Fourier Series 216
4.5 1D Fourier Transform 219
4.6 2D Fourier Transform 222 8. CDP Sort and Binning
4.7 z Transform 230 8.1 CDP Geometry 405
4.8 Hilbert Transform 231 8.2 CDP Fold 409
4.9 τ-p Transform (Slant Stack) 232 8.3 Binning in 3D 411
4.10 Sampling Theory 235 8.4 QC in Sort and Binning 415
vii
viii CONTENTS
1
Introduction
O U T L I N E
Seismic reflection exploration today is, obvi- intrusions, where subsalt imaging today has
ously, used extensively by the oil and gas indus- become a challenging phenomenon. Continuous
try. Indeed, the increase in new hydrocarbon development of acquisition and processing tech-
discoveries over the last 50 years almost paral- niques of seismic data, after their first introduc-
lels the advancements in data processing and tion in the early 1920s, has resulted in very high-
acquisition techniques. Following the digital resolution subsurface images, which have
recording of seismic data, incorporating the enabled us to discover much smaller hydrocar-
common midpoint (CMP) acquisition technique, bon traps, and it is now possible to map the
powerful workstations are now used to process target levels in much higher detail.
large amounts of digital data using sophisticated Raw seismic data should be processed using
processing applications. This has led to a consid- several complex data-processing steps in order
erable increase in discovery and production to obtain a subsurface image from 2D or 3D
(Fisher, 1991), since these developments have multichannel seismic datasets. This procedure,
provided many quality subsurface images at rel- consisting of a number of sequential mathemat-
atively higher resolution, especially in complex ical processes, is known as seismic data proces-
geological environments such as areas of salt sing. Data-processing applications commenced
Acquisition and Processing of Marine Seismic Data 1 # 2018 Elsevier Inc. All rights reserved.
https://doi.org/10.1016/B978-0-12-811490-2.00001-3
2 1. INTRODUCTION
following the digital recording of seismic data in exploration and processing (Flinn et al., 1967).
the mid-1960s, resulting in a considerable Today, seismic data is digitally recorded and
increase in the final seismic image quality after processed before interpretation. Several differ-
proper processing. Thenceforward, the basic ent data-processing steps can be applied to the
goal of the processing has not changed and it is seismic data in different domains, such as
still quite simple: increasing the seismic resolu- time-distance, frequency, frequency-distance,
tion, and enhancement of the signal level while etc. The selection of suitable domain for the
suppressing the noise amplitudes – in other application is dependent on which domain
words, increasing the signal-to-noise (S/N) ratio provides the best separation of signal from
of the seismic data. If the noise embedded in the the noise.
data can be separated by one of its specific char- Seismic data processing consists of several
acteristics, such as its frequency band, propaga- processing steps that are consecutively applied
tion velocity and/or direction, its amplitude to the input data. Such a complete procedure
with respect to the primary reflection ampli- composed of different processing steps is
tudes, etc., then it may be possible to remove termed a processing flow. A basic processing
most of the noise components from the data. flow consists of loading the dataset from disk,
Early noise-suppression applications, when applying one or more processing step(s) sequen-
digital recording and processing techniques tially, and then writing the processed data back
were not available, were accomplished by to the disk (Fig. 1.1). The effectiveness of many
grouping geophones to suppress the coherent processing steps strongly depends on the
noise, such as ground roll in land seismic acqui- parameters used in the previous steps. For ins-
sitions. The technological revolution caused by tance, if a band-pass filter is not properly
World War II also resulted in significant devel- applied before the deconvolution, we will not
opments related to seismic data-processing obtain satisfactory results from the deconvolu-
methodology. The practices used in communi- tion step. Therefore, a processor is responsible
cation technology during wartime were later for proper selection of each data-processing
applied to seismic data by the researchers at parameter and testing of the results of each step
the Massachusetts Institute of Technology in to ensure the quality of the output.
the 1950s, which ultimately advanced the meth- Both for 2D and 3D projects, a whole seismic
odology and approaches utilized in seismic reflection study consists of data acquisition,
QC
-1- -2- -3-
Selection of flow and data Selection of data processing Evaluation of the outputs of
processing steps most parameters that are most each processing step,
suitable for the input raw suitable for each processing removal of parameter
seismic data step selection issues
FIG. 1.2 Basic stages of a conventional seismic reflection project, consisting of data acquisition, processing and interpre-
tation, and primary steps of data processing along with the quality control (QC) recurrence in processing.
4 1. INTRODUCTION
processing steps, such as velocity analysis or and is known as quality control in data
prestack migration. processing (QC).
During the processing, some of the processing
An almost conventional seismic data-
steps and their orders of application may differ
processing flow used to process marine seismic
from one dataset to another, depending on the
data is shown in Fig. 1.3. However, since the
selection of the processor. For instance, while a
data-processing flows are dependent on the data
strong tail buoy noise in a marine seismic sur-
and the data processor, any particular noise in
vey may require an application of a suitable
the seismic data may require additional proces-
frequency-wavenumber (f-k) filter, this may not
sing steps and/or application of the steps in a
be necessary for another line. While some data
different order. The data is processed as gathers
processors apply deconvolution to the shot
or ensembles, termed prestack data until it is
gathers, the others prefer an application to CDPs.
stacked. Following the stacking process, the data
A strict quality control (QC) system must be
is called poststack data.
applied to the outputs of each processing step
Because the data-processing flow, steps and
to ensure that the specific processing application
the parameters are determined by the proces-
produced acceptable results. These QC analyses
sing specialists, and hence are operator-
are schematically illustrated in Fig. 1.2:
dependent, the final product of the same raw
• At first, processor should determine the most input dataset processed by different operators
appropriate data processing flow for the may have a different appearance. The output
input dataset, which consists of proper images may have discrepancies by means of
processing steps. their frequency content, S/N ratio, resolution
• Then, the suitable parameters for each and trace-by-trace consistency. The factors
processing step is determined, such as cut-off affecting the quality of the output of the seismic
frequency values for band-pass filtering, data-processing sequence are
deconvolution operator length, etc. These
• S/N ratio of the input dataset
processing steps with initially determined
• environment of the data collected (e.g., land
parameters are then applied to the input data
data almost always has a lower resolution
one at a time.
than marine data)
• After each application, the quality and
• algorithm and software used
acceptability of the outputs are analyzed in
• data-processing flow
different domains and with different tools.
• experience of the processor (especially in
For instance, output of a band-pass filter can
determining the parameters and QC
be analyzed by checking the power spectrum
implications)
of the band-pass filter output, while the
success of a spiking deconvolution can be The data processor should apply a series of
assured by controlling the auto-correlograms parameter tests to the input data for each step
of the deconvolution output. Whenever a to determine the suitable parameters, and the
processing step produces an unacceptable results are carefully analyzed after each test
result, its parameters are updated and the attempt. However, only a specific part of the
step is re-applied to the data with new seismic data is generally extracted for parameter
parameters to adjust the mistakes arising testing, since using the whole seismic line for a
from inappropriate parameter selections. parameter test is time consuming and is not
This recurrence can be done on shots, CDPs, practical. For instance, to determine the cut-off
brute stack sections or a near trace section, frequencies for a band-pass filter, tests can be
1. INTRODUCTION 5
Data loading
Demultiplexing
Band-pass filter
Multiple
suppression
Deconvolution
Band-pass filter
FIG. 1.3 A conventional seismic data-processing flow for marine seismic data.
performed on the very first shots of the line in Fig. 1.4. Here, the dataset used has 500 shots
instead of applying the filter to the whole seis- with 480 recording channels, resulting in a total
mic line. After determination of all parameters of 240,000 traces and 120-fold seismic data. Sam-
for each step, the flow is run for the whole line pling rate and maximum recording time are
using approved parameters, which is known 1 ms and 6 s, respectively. Apart from the hard-
as production processing. ware specifications used to prepare this illustra-
Each of these steps is of a specific run time con- tion, the relative time span for each processing
trolled by data volume, processing hardware, and step is remarkable. In general, the processes per-
the algorithm implemented. As an example, nec- formed on prestack data require much more run
essary rendering times in seconds for different time than those applied to poststack data, because
steps for the same 2D seismic dataset are shown the data volume is significantly reduced after
6 1. INTRODUCTION
FIG. 1.4 Run times in seconds for different processing steps for the same 2D dataset. Note that the vertical axis is set to
logarithmic so that the small rendering times of some particular steps also become evident. AGC, automatic gain control;
NMO, normal moveout correction; TVSW, time variant spectral whitening; TVF, time variant filter; SRME, surface related
multiple elimination; DMO, dip moveout.
stacking. The most costly processes are the pre- • SeisSpace/Promax (#Landmark)
stack migration algorithms. • Geovation, previously Geovecteur (#CGG
Today, several sophisticated commercial soft- Veritas)
ware packages exist in the market. In addition to • GeoThrust (#GeoTomo LLC)
the dedicated private software developers, most • Globe Claritas (#GNS Science)
of the seismic survey companies also develop • RadExPro (#DECO Geophysical)
their own seismic processing software. The most • Vista (#Gedco)
distinguished software suites commercially • Seismic Processing Workshop (SPW)
available today, especially for the petroleum (#Parallel Geoscience)
industry, are:
Apart from the commercial software listed
• Omega (#Western Geco) here, a couple of freeware processing codes
• Echos, previously Disco Focus (#Paradigm) are also available, generally used by academia:
1.1 UNDERWATER ACOUSTICS 7
• Stanford Exploration Project, SEPlib TABLE 1.1 Effects of the Seawater Physical Parameters
(Stanford University) on the Sound Velocity
• Seismic Un*x (SU) (Colorado School
Parameter Effect on the Sound Velocity
of Mines)
• FreeUSP (BP America Inc.) Temperature (T) 2.7 (m/s)/°C
Salinity (S) 1.2 (m/s)/ppt
Pressure or depth (P) 0.017 (m/s)/m
1.1 UNDERWATER ACOUSTICS
T H P
+ =
Depth
Depth
Depth
(A) (B) (C)
FIG. 1.5 Physical zones, or clines, within the water column. (A) Thermocline (T), (B) halocline (H), and (C) pycnocline
(P) layers. M represents the surficial, or mixed, water layer.
rapidly with depth (Fig. 1.5B). It is located below pressure create a layer of minimum sound veloc-
the uniformly saline surface water layer and is ity in the water column. It is termed the Sound
characterized by a strong, vertical salinity gradi- Fixing and Ranging (SOFAR) channel, where
ent. Below the halocline, salinity remains high. the sound waves in the seawater are trapped
Below the mixed layer, there is a horizontal layer and travel for long distances without losing their
within the water column where the density energy significantly. The depth where the mini-
gradient is greatest due to the rapid change in mum sound velocity occurs is the axis of the
temperature or salinity (Fig. 1.5C). This layer is channel. Velocity increases above and below
the pycnocline, where a large density contrast the axis because of the temperature and pressure
is therefore observed between the surficial water increases, respectively. Although the location of
layer and deep oceanic waters, which prevents the SOFAR channel axis varies with the temper-
the formation of vertical currents. Except for ature and water depth, it commonly lies
the arctic zones, where no pycnocline layer between 600 and 1200 m below the sea surface
exists, this layer is quite stable and separates in open oceans.
the surficial layer from deep ocean waters where Sound velocity information in marine
variations in salinity and temperature are acoustic applications is vital, especially in some
very small. specific marine geophysical applications. For
Clines in the seawater extend almost horizon- instance, we definitely need sound velocity in
tally for large distances. The vertical stratifica- seawater in sufficient detail for multibeam
tion at a specific location due to the bathymetric surveys to convert the arrival times
temperature and salinity variations as a function of the signals reflected back from the seafloor
of depth creates “channeling” for sound waves into the water depths. In 3D seismic surveys,
in the water column. This channel is located at the distances between the streamers are main-
a depth where the effects of temperature and tained by acoustic communications among the
1.2 MARINE ACOUSTIC METHODS 9
sensors using the specific instruments in the sea- B ¼ 1:389 1:262E-2T + 7:166E-5T 2
water regularly positioned along the spread. + 2:008E-6T 3 3:21E-8T 4 Þ
Therefore, the sound velocity in the water col- + ð9:4742E-5 1:2583E-5T 6:4928E-8T2
umn must be continuously measured in real
time during 3D surveys, since it may change + 1:0515E-8T 3 2:0142E-10T4 ÞP
with time and location in the survey area. + ð3:9064E-7 + 9:1061E-9T
Variation of the sound velocity in seawater 1:6009E-10T2 + 7:994E-12T3 ÞP2
can be obtained using velocimeters, which + ð1:100E-10 + 6:651E-12T
directly measure the velocity, or the specific sen- 3:391E-13T2 ÞP3
sors termed CTDs (conductivity-transmission-
depth), which measure the physical parameters C ¼ 1:922E-2 4:42E-5T
used to calculate the sound velocity. In addition, + ð7:3637E-5 + 1:7950E-7T ÞP
an expendable bathythermograph (XBT) probe
can be used to measure the temperature of the D ¼ 1:727E-3 7:9836E-6P
upper kilometer of the ocean, and the data is Fig. 1.6 shows a CTD cast from deep waters of
then used to calculate the sound velocity profile. the Black Sea. The thermocline (T) is between
CTD measurements used to determine the con- 15 and 80 m depth, where the seawater temper-
ductivity and temperature as a function of depth ature decreases significantly. There is an
of the ocean are more common in obtaining the approximately 130-m thick halocline (H) layer
velocity. There are several empirical approxima- below the surficial mixed water layer (M). Until
tions to obtain the sound velocity from mea- the bottom of the thermocline, the sound veloc-
sured physical parameters. A more recent ity is predominantly controlled by the tempera-
international standard algorithm has been ture variations in the water column. At greater
developed by Chen and Millero (1977) and later depths, however, the effect of pressure on the
modified by Wong and Zhu (1995). It is also sound velocity value becomes increasingly
known as the UNESCO algorithm today and is dominant, resulting in a linear velocity increase
expressed as since the pressure increases almost linearly with
V ðS, T, PÞ ¼ A + B S + C S3=2 + D S2 (1.1) depth. As a result, velocity is relatively high both
in surficial and deep waters because of the
where velocity (V) is in meters per second, tem- higher temperature in surficial waters and
perature (T) is between 0 and 40°C, salinity (S) is linearly increasing pressure in deeper waters,
between 0 and 40 ppt, and pressure (P) is respectively.
between 0 and 1000 bars, and the coefficients
A, B, C, and D are given by
1.2 MARINE ACOUSTIC METHODS
A ¼ 1402:388 + 5:03830T 5:81090E-2T 2
+ 3:3432E-4T3 1:47797E-6T4 + 3:1419E-9T 5 Þ Marine geophysics studies are performed to
+ 0:153563 + 6:8999E-4T 8:1829E-6T 2 understand the structure and morphology of
+ 1:3632E-7T3 6:1260E-10T 4 ÞP the seafloor and subsurface sediments and
monitor their short- and long-term behaviors,
+ ð3:1260E-5 1:7111E-6T + 2:5986E-8T 2
to safely settle the offshore geo-engineering
2:5353E-10T 3 + 1:0415E-12T 4 ÞP2 structures such as pipelines and platforms,
+ ð9:7729E-9 + 3:8513E-10T and to explore the offshore mineral and energy
2:3654E-12T2 ÞP3 sources. The methodology and equipment used
10 1. INTRODUCTION
FIG. 1.6 Physical properties of seawater and sound velocity obtained from a CTD profile. (A) Calculated sound velocity,
(B) temperature, and (C) salinity as a function of depth. M, T, and H represent mixed water, halocline and thermocline layers,
respectively.
in marine acoustic exploration are different from geo-hazards encountered during shallow marine
those used in onshore surveys. Discrepancies installations.
arise from the purposes of the surveys, the pen- Before drilling an offshore well, there is a
etration depths and resolution differences, the need to define the surficial morphology as well
working principles of the equipment, and the as the subsurface sediments in the area sur-
information obtained. rounding the well location in detail. This opera-
Marine geophysical surveys have been con- tion is often termed a site survey. Since the
ducted since the 1960s utilizing several different resolution of conventional 2D and 3D seismic
acoustic and nonacoustic methods. Among these, data is not sufficient to provide detailed subsur-
gravity and magnetic surveys, seismic methods, face information on the shallow stratigraphy,
heat flow measurements, and other high- high-resolution techniques employing much
resolution acoustic methods such as bathymetric, higher frequency signals are used to map the
side-scan sonar and subbottom profiler surveys geo-hazards, such as slides, excessive seafloor
are the most common techniques. In this section, inclinations, shallow gas and gas hydrates,
the marine geophysical methods employing active faulting, etc. Although single channel or
acoustic signals in different frequency, amplitude multichannel short spread sparker seismic
and signal forms are discussed briefly. Excluding reflection surveys are also used for the shallow
the conventional marine seismics, these methods geo-engineering problems, 2D and 3D conven-
are generally known as high-resolution marine tional multichannel seismic surveys are mostly
geophysical techniques, and are commonly used used for hydrocarbon exploration by dedicated
to solve geo-engineering problems or to map survey companies, as well as by academia for
1.2 MARINE ACOUSTIC METHODS 11
scientific purposes worldwide, and are not 10 and 200 Hz) or spark arrays (between 50 and
included in the site surveys. 800 Hz), while higher-frequency acoustic signals
The marine geophysical methods for high- from approximately 3 to 800 kHz are produced in
resolution site surveys use different sensors to the water column by specific instruments called
produce acoustic signals within a significantly transducers. These instruments convert one type
wide frequency spectrum. Table 1.2 shows gen- of energy (e.g., the electric signal) into another
eral information on the frequency band of the (e.g., the acoustic signal), or vice versa. Trans-
most common marine acoustic systems. The most ducers use piezoelectric crystals made by ceramic
suitable method with an appropriate frequency is material, which vibrate when they are excited by
selected by factors such as required information, an electric pulse. These vibrations are transmitted
resolution and penetration purposes. Relatively into the water column as pressure (P) waves to
low-frequency signals used in seismic surveys generate the acoustic signal. The same transducer
are produced by air guns (generally between is also used to receive the reflected signals, and
Another is so interesting, for its classical turn, and for the names
which it gives to the ‘bishop’ and his crew that I quote it in full[1129].
2. ad honorem Tityri,
festum colant baculi
satrapae et asini.
3. applaudamus Tityro
cum melodis organo,
cum chordis et tympano.
4. veneremur Tityrum,
qui nos propter baculum
invitat ad epulum.
‘Orientis partibus,
adventavit asinus.’
In any case, the oriental example can hardly be responsible for more
than the admission of the feast within the doors of the church. One
cannot doubt that it was essentially an adaptation of a folk-custom
long perfectly well known in the West itself. The question of origin
had already presented itself to the learned writers of the thirteenth
century. William of Auxerre, by a misunderstanding which I shall
hope to explain, traced the Feast of Fools to the Roman Parentalia:
Durandus, and the Paris theologians after him, to the January
Kalends. Certainly Durandus was right. The Kalends, unlike the
more specifically Italian feasts, were co-extensive with the Roman
empire, and were naturally widespread in Gaul. The date
corresponds precisely with that by far the most common for the
Feast of Fools. A singular history indeed, that of the ecclesiastical
celebration of the First of January. Up to the eighth century a fast,
with its mass pro prohibendo ab idolis, it gradually took on a festal
character, and became ultimately the one feast in the year in which
paganism made its most startling and persistent recoil upon
Christianity. The attacks upon the Kalends in the disciplinary
documents form a catena which extends very nearly to the point at
which the notices of the Feast of Fools begin. In each alike the
masking, in mimicry of beasts and probably of beast-gods or
‘demons,’ appears to have been a prominent and highly reprobated
feature. It is true that we hear nothing of a dominus festi at the
Kalends; but much stress must not be laid upon the omission of the
disciplinary writers to record any one point in a custom which after all
they were not describing as anthropologists, and it would certainly be
an exceptional Germano-Keltic folk-feast which had not a dominus.
As a matter of fact, there is no mention of a rex in the accounts of
the pre-Christian Kalends in Italy itself. There was a rex at the
Saturnalia, and this, together with an allusion of Belethus in a quite
different connexion to the libertas Decembrica[1158], has led some
writers to find in the Saturnalia, rather than the Kalends, the origin of
the Feast of Fools[1159]. This is, I venture to think, wrong. The
Saturnalia were over well before December 25: there is no evidence
that they had a vogue outside Italy: the Kalends, like the Saturnalia,
were an occasion at which slaves met their masters upon equal
terms, and I believe that the existence of a Kalends rex, both in Italy
and in Gaul, may be taken for granted.
But the parallel between Kalends and the Feast of Fools cannot
be held to be quite perfect, unless we can trace in the latter feast
that most characteristic of all Kalends customs, the Cervulus. Is it
possible that a representative of the Cervulus is to be found in the
Ass, who, whether introduced from Constantinople or not, gave to
the Feast of Fools one of its popular names? The Feast of Asses
has been the sport of controversialists who had not, and were at no
great pains to have, the full facts before them. I do not propose to
awake once more these ancient angers[1160]. The facts themselves
are briefly these. The ‘Prose of the Ass’ was used at Bourges, at
Sens, and at Beauvais. As to the Bourges feast I have no details. At
Sens, the use of the Prose by Pierre de Corbeil is indeed no proof
that he allowed an ass to appear in the ceremony. But the Prose
would not have much point unless it was at least a survival from a
time when an ass did appear; the feast was known as the asinaria
festa; and even now, three centuries after it was abolished, the Sens
choir-boys still play at being âne archbishop on Innocents’ day[1161].
At Beauvais the heading Conductus quando asinus adducitur in the
thirteenth-century Officium seems to show that there at least the ass
appeared, and even entered the church. The document, also of the
thirteenth century, quoted by the editors of Ducange, certainly brings
him, in the ceremony of January 14, into the church and near the
altar. An imitation of his braying is introduced into the service itself.
At Autun the leading of an ass ad processionem, and the cantilena
super dictum asinum were suppressed in 1411. At Châlons-sur-
Marne in 1570 an ass bore the ‘bishop’ to the theatre at the church
door only. At Prague, on the other hand, towards the end of the
fourteenth century, an ass was led, as at Beauvais, right into the
church. These, with doubtful references to fêtes des ânes at St.
Quentin about 1081, at Béthune in 1474, and at Laon in 1527, and
the Mosburg description of the ‘bishop’ as asinorum dominus, are all
the cases I have found in which an ass has anything to do with the
feast. But they are enough to prove that an ass was an early and
widespread, though not an invariable feature. I may quote here a
curious survival in a ronde from the west of France, said to have
been sung at church doors on January 1[1162]. It is called La Mort de
l’Âne, and begins:
‘Quand le bonhomme s’en va,
Quand le bonhomme s’en va,
Trouvit la tête à son âne,
Que le loup mangit au bois.