Download the full version of the ebook now at ebooksecure.

com

Spatiotemporal Random Fields, Second Edition:

Theory and Applications George Christakos -
eBook PDF

https://ebooksecure.com/download/spatiotemporal-
random-fields-second-edition-theory-and-
applications-ebook-pdf/

Explore and download more ebook at https://ebooksecure.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

(eBook PDF) Turbomachinery: Basic Theory and Applications,

Second Edition

http://ebooksecure.com/product/ebook-pdf-turbomachinery-basic-theory-
and-applications-second-edition/

ebooksecure.com

Kappa Distributions. Theory and Applications in Plasmas

1st Edition Edition George Livadiotis (Eds.) - eBook PDF

https://ebooksecure.com/download/kappa-distributions-theory-and-
applications-in-plasmas-ebook-pdf/

ebooksecure.com

Solutions for Biot's Poroelastic Theory in Key Engineering

Fields. Theory and Applications 1st Edition Edition
Yuanqiang Cai And Honglei Sun (Auth.) - eBook PDF
https://ebooksecure.com/download/solutions-for-biots-poroelastic-
theory-in-key-engineering-fields-theory-and-applications-ebook-pdf/

ebooksecure.com

Geophysical potential fields: geological and environmental

applications 1st edition - eBook PDF

https://ebooksecure.com/download/geophysical-potential-fields-
geological-and-environmental-applications-ebook-pdf/

ebooksecure.com
Rowan's Primer of EEG, 2e Second Edition Fields - eBook
PDF

https://ebooksecure.com/download/rowans-primer-of-eeg-2e-ebook-pdf/

ebooksecure.com

Schaum's Outline of Probability, Random Variables, and

Random Processes, Fourth Edition Hwei P. Hsu - eBook PDF

https://ebooksecure.com/download/schaums-outline-of-probability-
random-variables-and-random-processes-fourth-edition-ebook-pdf/

ebooksecure.com

(eBook PDF) Psycholinguistics: Introduction and

Applications, Second Edition

http://ebooksecure.com/product/ebook-pdf-psycholinguistics-
introduction-and-applications-second-edition-2/

ebooksecure.com

(eBook PDF) Psycholinguistics Introduction and

Applications Second Edition

http://ebooksecure.com/product/ebook-pdf-psycholinguistics-
introduction-and-applications-second-edition/

ebooksecure.com

(eBook PDF) Probability, Statistics, and Random Signals by

Charles Boncelet

http://ebooksecure.com/product/ebook-pdf-probability-statistics-and-
random-signals-by-charles-boncelet/

ebooksecure.com
Spatiotemporal Random
Fields
This page intentionally left blank
Spatiotemporal Random
Fields
Theory and Applications

Second Edition

George Christakos
Department of Geography, San Diego State University,
San Diego, California, USA;
Institute of Islands and Coastal Ecosystems, Ocean College,
Zhejiang University, Zhoushan, Zhejiang, China
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

Copyright © 2017 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: 978-0-12-803012-7

For Information on all Elsevier publications visit our

website at https://www.elsevier.com/books-and-journals

Publisher: Candice Janco

Acquisition Editor: Marisa LaFleur
Editorial Project Manager: Marisa LaFleur
Production Project Manager: Paul Prasad Chandramohan
Designer: Greg Harris
Typeset by TNQ Books and Journals
Dedicated
to
Yongmei
This page intentionally left blank
Contents

Preface ................................................................................................................................................. xv

CHAPTER I Space, Time, SpaceeTime, Randomness, and Probability .................. 1

1. Introduction ...................................................................................................................1
2. SpaceeTime Continuum and Kolmogorov Probability Space ....................................4
2.1 SpaceeTime Arguments: Points, Lags, Separations, and Metrics....................... 4
2.2 Transformations and Invariance in SpaceeTime ............................................... 17
2.3 SpaceeTime Interpretations................................................................................ 23
2.4 Functions of SpaceeTime Arguments................................................................ 26
3. Random Variables in SpaceeTime.............................................................................31
3.1 Kolmogorov’s Probability Theory ...................................................................... 31
3.2 Useful Inequalities............................................................................................... 35
3.3 Convergence of Random Variable Sequences .................................................... 37
CHAPTER II Spatiotemporal Random Fields........................................................... 39
1. Introduction .................................................................................................................40
1.1 The SpaceeTime Component ............................................................................. 41
1.2 The Randomness Component ............................................................................. 42
2. Characterization of Scalar Spatiotemporal Random Fields .......................................42
2.1 Probabilistic Structure ......................................................................................... 43
2.2 The Characteristic Function ................................................................................ 50
2.3 Spatiotemporal Variability Functions: Complete (or Full) and Partial .............. 51
2.4 Analysis in the Spectral Domain ........................................................................ 56
2.5 Data-Independent Spatiotemporal Variability Function ..................................... 57
2.6 Some Noticeable Special Cases of the Spatiotemporal Random
Field Theory ........................................................................................................ 59
2.7 SpaceeTime Separability.................................................................................... 60
3. Physical Insight Behind the Random Field Concept .................................................61
3.1 Random Field Realizations ................................................................................. 61
3.2 Probable Versus Actual ....................................................................................... 63
3.3 Probability and the Observation Effect............................................................... 64
3.4 Self-consistency and Physical Fidelity ............................................................... 65
4. Geometry of Spatiotemporal Random Fields.............................................................69
5. Vector Spatiotemporal Random Fields .......................................................................70
6. Complex Spatiotemporal Random Fields...................................................................73
7. Classifications of the Spatiotemporal Random Field Model .....................................73
7.1 First Classification: Discrete Versus Continuous Arguments............................. 74
7.2 Second Classification: Scalar Versus Vector Random Fields and Arguments... 74

vii
viii CONTENTS

7.3 Third Classification: Probability Law Shapes .................................................... 74

7.4 Fourth Classification: SpaceeTime Variability .................................................. 75
7.5 Fifth Classification: Spatiotemporal Random Field Memory Versus
Independence ....................................................................................................... 77
8. Closing Comments......................................................................................................78
8.1 The Methodological Importance of SpaceeTime .............................................. 78
8.2 A Conceptual Meeting Point for Modelers and Experimentalists ..................... 80
8.3 There Is No Assumptionless Modeling .............................................................. 81
CHAPTER III SpaceeTime Metrics ......................................................................... 83
1. Basic Notions ..............................................................................................................83
1.1 Formal and Physical Aspects of SpaceeTime Metric Determination ............... 84
1.2 SpaceeTime Metric Forms ................................................................................. 87
1.3 Derived SpaceeTime Metrics............................................................................. 92
1.4 SpaceeTime Metric Differentials ....................................................................... 93
1.5 Specifying SpaceeTime Relationships in the Covariance Function ................. 95
2. Covariance Differential Formulas.............................................................................100
3. SpaceeTime Metric Determination From Physical Considerations........................106
4. Examples ...................................................................................................................108
5. Concerning the Zeta Coefficients .............................................................................117
6. Closing Comments....................................................................................................118
CHAPTER IV SpaceeTime Correlation Theory ..................................................... 121
1. Focusing on SpaceeTime Variability Functions......................................................121
1.1 Basics of SpaceeTime Correlation Theory...................................................... 122
1.2 Physical Investigations Based on SpaceeTime Correlation Theory................ 123
2. SpaceeTime Variability Functions in Terms of Scalar SpaceeTime Statistics......124
2.1 Locality: One-Point SpaceeTime Variability Functions.................................. 125
2.2 Nonlocality: Omnidirectional Two-Point SpaceeTime Variability
Functions ........................................................................................................... 127
2.3 Nonlocality: Direction-Specific SpaceeTime Variability Functions ............... 133
2.4 Physical Considerations and Assumptions of SpaceeTime Variability
Functions ........................................................................................................... 133
2.5 Formal and Physical Covariance Permissibility ............................................... 137
3. Basic Properties of Covariance Functions................................................................139
4. CrosseSpaceeTime Variability Functions...............................................................141
5. Correlation of Gaussian and Related Spatiotemporal Random Fields ....................144
5.1 General Considerations ..................................................................................... 144
5.2 Gaussian Properties ........................................................................................... 144
 CONTENTS ix

6. Correlation Theory of Complex Spatiotemporal Random Fields ............................146

6.1 Basic Notions .................................................................................................... 146
6.2 Other Types of Complex Covariance Functions............................................... 148
6.3 Gaussian Complex Spatiotemporal Random Fields ......................................... 151
6.4 Complex-Valued Versus Real-Valued Covariance Functions
of SpaceeTime Homostationary Random Fields ............................................. 152
6.5 Some Methodological Considerations .............................................................. 153
CHAPTER V Transformations of Spatiotemporal Random Fields......................... 155
1. Introduction ...............................................................................................................155
2. Fourier Transformation .............................................................................................157
2.1 Characteristic Functions .................................................................................... 157
2.2 Harmonizable Random Fields and Covariance Functions ............................... 158
2.3 Transfer Function and Evolutionary Mean Power............................................ 165
2.4 Fourier Transform of Vector Spatiotemporal Random Fields.......................... 167
3. Space Transformation ...............................................................................................168
3.1 Basic Notions .................................................................................................... 168
3.2 Space Transformation of Spatiotemporal Random Fields................................ 173
3.3 Space Transformation for Spatiotemporal Variability Functions..................... 175
3.4 Space Transformation in the Simulation of Spatiotemporal
Random Fields................................................................................................... 177
3.5 Space Transformation in the Solution of Stochastic Partial Differential
Equation............................................................................................................. 180
4. The Traveling Transformation ..................................................................................180
4.1 Basic Notions .................................................................................................... 181
4.2 Determination of the Traveling Vector ............................................................. 186
4.3 Traveling Transformation in Spatiotemporal Random Field Estimation:
The SpaceeTime Projection Technique ........................................................... 195
5. Closing Comments....................................................................................................201
CHAPTER VI Geometrical Properties of Spatiotemporal Random Fields ............ 203
1. Introduction ...............................................................................................................203
2. Stochastic Convergence ............................................................................................204
3. Stochastic Continuity ................................................................................................208
3.1 Basic Types of Stochastic Continuity ............................................................... 209
3.2 Equivalence, Modification, and Separability .................................................... 212
4. Stochastic Differentiation .........................................................................................216
4.1 Basic Notation and Definitions ......................................................................... 217
4.2 Covariances of Random Field Derivatives ....................................................... 223
4.3 Mean Squarely Differentiability Conditions..................................................... 227
4.4 Almost Surely Differentiability Conditions...................................................... 231
Visit https://ebooksecure.com
now to explore a rich
collection of ebook and enjoy
exciting offers!
x CONTENTS

5. The Central Limit Theorem......................................................................................233

6. Stochastic Integration................................................................................................234
CHAPTER VII Auxiliary Hypotheses of Spatiotemporal Variation ....................... 239
1. Introduction ...............................................................................................................239
1.1 Hypothesis 1: Homostationarity........................................................................ 241
1.2 Hypothesis 2: Isostationarity............................................................................. 243
1.3 Hypothesis 3: Heterogeneity ............................................................................. 245
1.4 Hypothesis 4: Ergodicity................................................................................... 246
1.5 Hypothesis 5: Separability ................................................................................ 248
1.6 Hypothesis 6: Symmetry ................................................................................... 249
1.7 Hypothesis 7: Locational Divergence ............................................................... 249
2. SpaceeTime Homostationarity.................................................................................250
2.1 Omnidirectional Spatiotemporal Variability Functions .................................... 251
2.2 Direction-Specific Spatiotemporal Covariance Function ................................. 255
2.3 Anisotropic Features.......................................................................................... 255
2.4 Spatiotemporal Variogram and Structure Functions: Omnidirectional
and Direction Specific ....................................................................................... 257
3. Spectral Representations of SpaceeTime Homostationarity ...................................260
3.1 Spectral Functions of SpaceeTime Homostationary Random
Fields ................................................................................................................. 261
3.2 Properties of the Spectral Density Function ..................................................... 266
3.3 Partial Spectral Representations........................................................................ 268
3.4 More on Dispersion Relations........................................................................... 272
3.5 Separability Aspects .......................................................................................... 273
4. The Geometry of SpaceeTime Homostationarity ...................................................276
4.1 Differentiation Formulas: Physical and Spectral Domains .............................. 276
4.2 Stochastic Continuity and Differentiability ...................................................... 285
4.3 Spatiotemporal Random Field Integrability ..................................................... 296
5. Spectral Moments and Linear Random Field Transformations ...............................297
CHAPTER VIII Isostationary Scalar Spatiotemporal Random Fields ................... 303
1. Introduction ...............................................................................................................303
1.1 Basic Considerations ......................................................................................... 303
1.2 Power-Law Correlations.................................................................................... 309
1.3 Physical Considerations of Variogram Functions............................................. 313
2. Relationships Between Covariance Derivatives and SpaceeTime
Isostationarity ............................................................................................................314
3. Higher-Order Spatiotemporal Variogram and Structure Functions .........................319
4. Separable Classes of SpaceeTime Isostationary Covariance Models.....................320
5. A Survey of SpaceeTime Covariance Models ........................................................324
 CONTENTS xi

6. Scales of Spatiotemporal Dependence and the Uncertainty Principle ....................329

6.1 Scales for Spatiotemporal Random Fields With Restricted SpaceeTime
Variability .......................................................................................................... 330
6.2 Relationships Between Physical and Spectral Domains: The Uncertainty
Principle............................................................................................................. 334
7. On the Ergodicity Hypotheses of Spatiotemporal Random Fields ..........................336
CHAPTER IX Vector and Multivariate Random Fields.......................................... 347
1. Introduction ...............................................................................................................347
2. Homostationary and Homostationarily Connected CrosseSpatiotemporal
Variability Functions and CrosseSpectral Density Functions.................................349
2.1 Basic Notions and Interpretations ..................................................................... 350
2.2 Geometry of Vector Spatiotemporal Random Fields ....................................... 355
3. Some Special Cases of Covariance Functions .........................................................356
4. Solenoidal and Potential Vector Spatiotemporal Random Fields ............................362
5. Partial Cross-Covariance and Cross-Spectral Functions ..........................................365
6. Higher-Order CrosseSpatiotemporal Variability Functions ....................................366
7. Isostationary Vector Spatiotemporal Random Fields ...............................................369
7.1 Direct (Lag-Based) SpaceeTime Isostationarity.............................................. 369
7.2 Composite Lag-FieldeBased SpaceeTime Isostationarity.............................. 372
7.3 Links With Solenoidal and Potential Spatiotemporal
Random Fields................................................................................................... 378
8. Effective Distances and Periods ...............................................................................381
CHAPTER X Special Classes of Spatiotemporal Random Fields......................... 383
1. Introduction ...............................................................................................................383
2. Frozen Spatiotemporal Random Fields and Taylor’s Hypothesis............................384
2.1 Basic Notions .................................................................................................... 385
2.2 Spectral Domain Analysis................................................................................. 389
2.3 Differential Equation Representations .............................................................. 391
2.4 Extensions of the Frozen Random Field Model............................................... 395
2.5 Integrals of Frozen Spatiotemporal Random Fields......................................... 399
2.6 Vector Frozen Spatiotemporal Random Fields ................................................. 399
3. Plane-Wave Spatiotemporal Random Fields ............................................................400
4. Lognormal Spatiotemporal Random Fields..............................................................402
5. Spherical Spatiotemporal Random Fields ................................................................402
6. Lagrangian Spatiotemporal Random Fields .............................................................407
CHAPTER XI Construction of Spatiotemporal Probability Laws .......................... 409
1. Introduction ...............................................................................................................409
2. Direct Probability Density Model Construction Techniques ...................................411
2.1 The Independency Techniques .......................................................................... 412
xii CONTENTS

2.2 The Spherical Symmetry Technique................................................................. 412

2.3 The Transformation Technique ......................................................................... 413
3. Factora-Based Probability Density Model Construction Techniques ......................414
4. Copula-Based Probability Density Model Construction Techniques.......................418
5. Stochastic Differential EquationeBased Probability Density Model
Construction Techniques...........................................................................................421
5.1 The Transformation of Variables Approach ..................................................... 422
5.2 The Characteristic Function Approach ............................................................. 425
5.3 The Functional Approach.................................................................................. 426
6. Bayesian Maximum EntropyeBased Multivariate Probability Density
Model Construction Techniques ...............................................................................428
7. Methodological and Technical Comments ...............................................................431
CHAPTER XII Spatiotemporal Random Functionals ............................................. 433
1. Continuous Linear Random Functionals in the SpaceeTime Domain ...................433
1.1 Basic Notions .................................................................................................... 433
1.2 Generalized Fourier Transform ......................................................................... 436
1.3 SpaceeTime Characteristic Functionals........................................................... 439
1.4 Functional Derivatives....................................................................................... 441
2. Gaussian Functionals ................................................................................................447
CHAPTER XIII Generalized Spatiotemporal Random Fields ................................ 455
1. Basic Notions ............................................................................................................455
1.1 The Notion of Generalized Spatiotemporal Random Field.............................. 456
1.2 Generalized Spatiotemporal Random Field Properties and Physical
Significance ....................................................................................................... 461
1.3 Homostationary Generalized Spatiotemporal Random Fields ......................... 464
2. Spatiotemporal Random Fields of Orders n/m..........................................................468
2.1 Departure From SpaceeTime Homostationarity .............................................. 468
2.2 SpaceeTime Detrending ................................................................................... 470
2.3 Ordinary Spatiotemporal Random Field-n/m Representations of the
Generalized Spatiotemporal Random Field-n/m ............................................... 474
2.4 Determination of the Operator Qn/m and Its Physical Significance .................. 475
3. The Correlation Structure of Spatiotemporal Random Field-n/m ............................477
3.1 SpaceeTime Functional Statistics .................................................................... 477
3.2 Generalized Spatiotemporal Covariance Functions.......................................... 479
3.3 Generalized Spectral Representations and Permissibility of Generalized
Covariances........................................................................................................ 481
3.4 Generalized Covariance Function Models........................................................ 484
 CONTENTS xiii

4. Discrete Linear Representations of Spatiotemporal Random Fields .......................490

4.1 SpaceeTime Random Increments .................................................................... 490
4.2 SpaceeTime Variogram Analysis ..................................................................... 495
CHAPTER XIV Physical Considerations................................................................ 501
1. Spatiotemporal Variation and Laws of Change........................................................501
2. Empirical Algebraic Equations.................................................................................504
3. Physical Differential Equations ................................................................................506
4. Links Between Stochastic Partial Differential Equation and Generalized
Random Fields ..........................................................................................................512
4.1 Links in Terms of the Random Functional....................................................... 513
4.2 Links in Terms of the Detrending Operator ..................................................... 515
5. Physical Constraints in the Form of Integral Relationships, Domain
Restrictions, and Dispersion Equations ....................................................................518
CHAPTER XV Permissibility in SpaceeTime........................................................ 521
1. Concerning Permissibility.........................................................................................521
2. Bochnerian Analysis .................................................................................................522
2.1 Main Results...................................................................................................... 523
2.2 Limitations of Bochnerian Analysis ................................................................. 525
3. Metric Dependence ...................................................................................................528
4. Formal and Physical Permissibility Conditions for Covariance Functions .............529
4.1 Permissibility Conditions for SpaceeTime Homostationary Covariance
Functions ........................................................................................................... 530
4.2 Permissibility Conditions for SpaceeTime Isostationary Covariance
Functions ........................................................................................................... 532
4.3 Permissibility Conditions for Generalized Spatiotemporal Covariance
Functions ........................................................................................................... 535
4.4 Permissibility Conditions for Spatiotemporal Covariance Matrices ................ 537
5. More Consequences of Permissibility ......................................................................540
CHAPTER XVI Construction of Spatiotemporal Covariance Models ................... 543
1. Introduction.............................................................................................................543
2. Probability Density FunctioneBased and Related Techniques .............................545
2.1 Linking Directly Covariance Models and Probability Density
Functions ......................................................................................................... 545
2.2 Using Polynomial-Exponential Functions ...................................................... 548
2.3 Using Spectral Functions ................................................................................ 550
3. Delta and Related Techniques ................................................................................552
3.1 Basic Decomposition....................................................................................... 552
xiv CONTENTS

3.2 Normalized Angular Spectrum Decomposition.............................................. 554

3.3 Normalized Frequency Spectrum (or Coherency Function)
Decomposition................................................................................................. 555
4. Space Transformation Technique ...........................................................................557
5. Physical Equation Techniques ................................................................................560
5.1 Covariance Construction From Stochastic Partial Differential
Equation Representations ................................................................................ 560
5.2 Covariance Construction From Algebraic Empirical Relationships .............. 570
6. Closed-Form Techniques ........................................................................................572
7. Integral Representation Techniques........................................................................580
8. SpaceeTime Separation Techniques ......................................................................582
9. Dynamic Formation Technique ..............................................................................586
10. Entropic Technique .................................................................................................587
11. Attribute and Argument Transformation Techniques ............................................588
11.1 Attribute Transformation............................................................................... 588
11.2 Argument Transformation ............................................................................. 589
12. Cross-Covariance Model Construction Techniques ...............................................590
13. Revisiting the Role of Physical Constraints...........................................................593
14. Closing Comments..................................................................................................594
Exercises ............................................................................................................................................ 597
References.......................................................................................................................................... 643
Appendix............................................................................................................................................ 653
Index ................................................................................................................................................. 665
Preface

The study of real-world phenomena relies on theories of natural (physical, biological, etc.) sciences
that generally involve mathematical models. These models are usually defined by a set of equations
and supplemented by a set of logical reasoning rules for rigorously translating the quantitative analysis
results into meaningful statements about the phenomenon of interest. Additionally, and very impor-
tantly, the real-world study of a phenomenon is faced with various sources of uncertainty, ontic,
and epistemic (including phenomenal, technical, conceptual, and computational sources related to
quantitative modeling, data selection, and processing). As a result, exact deterministic model solutions
in terms of well-known analytic functions often turn out to be unrealistic and lack any visible means of
meaningful interpretation.
In light of the above considerations, in most real-world applications the mathematical models that
we currently use to describe the attributes of a natural phenomenon are stochastic in nature, i.e., these
attributes and the associated boundary/initial conditions are represented by random fields with argu-
ments in a composite spaceetime domain. In this domain, space represents the order of coexistence,
and time represents the order of successive existence of the attribute. Randomness manifests itself as
an ensemble of possible realizations regarding the attribute distribution, where the likelihood that each
one of these possible realizations occurs is expressed by the corresponding probability law. Thus,
spatiotemporal random model solutions are considerably more flexible and realistic than the determin-
istic single-valued solutions. Attribute distributions are well represented by theoretical probability
laws, and this permits us to calculate various spaceetime properties of these distributions with reason-
able accuracy.
The above considerations are the primary reasons for devoting this book to the spatiotemporal
random field theory and its potential applications in natural sciences. In this context, for any such the-
ory there is first the mathematical problem of analyzing, as far as possible, the stochastic model gov-
erning the relevant attributes together with the available data sets (hard or exact and soft or uncertain,
in general), and of finding as realistic and complete a solution as possible to the problem of interest that
maintains good contact with the real-world phenomenon in conditions of in situ uncertainty. Next
comes the interpretation (mathematical and physical) of the conclusions thus obtained, and their uti-
lization to make informative predictions. It should be pointed out that certain exact models and equa-
tions have played very important roles in the study of natural phenomena. It should also be noticed that
because many models and equations describing real-world phenomena are necessarily complicated
(multiparametered, highly nonlinear, and heterogeneous, whereas potentially critical features of the
phenomena remain unspecified), it is very useful to understand what qualitative features these models
and equations might possess, since they have been proven to offer an invaluable guide about the phe-
nomenon of interest.
Certainly, there are several important issues related to the distinction between theory and imple-
mentation. Concerning the in situ implementation of random field theory and techniques, one should
be aware that, although the fact is not always appreciated, the real-world complexities of the phenom-
ena mentioned above also mean that approximate techniques used as part of implementation could run
into hidden complications that have a tendency to distract attention from more useful issues. The sit-
uation may also be partially the fault of those of us working in this discipline, when we occasionally

xv
xvi PREFACE

propose abstract theories for the sake of greater “generality.” Yet this is not an excuse for the fact that,
as real-world experience shows, in the vast majority of cases the “ineffectiveness” is not a feature of
the theory or the modeling technique used but rather of the practitioners often attempting to use them
in a “black-box” manner. In this framework, exact models and equations that can be compared with
approximate or numerical results are very useful in checking the validity of approximation techniques
used in an application.
In addition to the above reasons for devoting this book to the theory of spatiotemporal random
fields, it should be noted that although much work has been done concerning the subject, it is often
not generally known because of the plethora of disciplines, journals, and mathematical terminologies
and notations in which it has appeared. It is hoped that one beneficial effect of the present effort will be
to save the interested readers from spending their time rediscovering already known results. And I
hope the present attempt to characterize the known results invariantly will help readers to identify
any new findings that may emerge. Throughout the book, for the basic issue regarding fundamental
concepts of probability, statistics, and random variables, I refer the reader to other texts, when
necessary.
Naturally, I begin by introducing the basic notions of the spaceetime continuum (points, lags, met-
rics, and geometrical transformations), mathematical field and related functions, probability, uncer-
tainty, and randomness (Chapter 1). Two chapters (Chapters 2 and 4) are devoted to the
presentation of standard results of the ordinary spatiotemporal random field theory, including much
of the terminology used later in the book. Among these two chapters I have interpolated one chapter
(Chapter 3) on spaceetime metrics. Its position is due to the fact that the spaceetime metric properties
can be used to elucidate the validity of certain random field issues introduced in the preceding two
chapters. This chapter also discusses the classification of spaceetime metrics for scalar and vector
random fields, and a physical lawebased metric determination technique is outlined and applied in
several cases. Intuitively, a natural attribute represented as a spatiotemporal random field is projected
on the physical (real) domain. Yet, there may be constructed other domains on which an attribute could
be projected. Such domains are the spectral domain, the reduced dimensionality, and the traveling do-
mains, which provide equivalent representations of the attribute defined on it. And while one’s intu-
ition may be better adapted to the physical domain, in certain cases it may be more convenient to
work in the alternative domains. So in Chapter 5 we discuss important concepts and methods associ-
ated with these alternative domains. Chapter 6 focuses on spatiotemporal random field geometry (con-
tinuity, differentiability, and integrability). This is one of the subjects that would warrant a book of its
own and, thus, I had to be very selective in the choice and manner of the material presented. Because of
its special physical and mathematical interest, the topic of homostationarity (space homogeneous/time
stationary) was given a chapter of its own (Chapter 7). Similar reasons are valid for devoting Chapter 8
to isostationary (space isotropic/time stationary) random fields. In this chapter a large number of
spatiotemporal variability functions (covariance, variogram, and structure functions of high order)
are presented. Chapter 9 deals with multivariate and vectorial random fields varying in the spaceetime
domain, including their main mathematical features and differences as regards their interpretation
(mathematical and physical). In Chapter 10, I discuss a selected group of spatiotemporal random fields
with special properties of particular interest to applications (this group includes the frozen random
field and its variations, the plane-wave, the lognormal, the spherical, and the Lagrangian random
fields). Chapter 11 focuses on techniques for constructing multivariate probability density functions
that offer a complete characterization of the spatiotemporal random field in stochastic terms. Due
 PREFACE xvii

to their fundamental role in the study of spaceetime heterogeneous random fields, an entire chapter
(Chapter 12) is devoted to the theory of spatiotemporal random functionals. Indeed, the functional
description of randomness naturally involves more complex mathematics, but it has its rewards on
both theoretical and application grounds (e.g., many real-world phenomena and their measurements
need to be expressed in terms of random functionals). Chapter 13 provides a rather detailed account
of the theory of spaceetime heterogeneous (generalized) random fields that is useful in the case of
natural attributes characterized by complex variations and patterns (varying trends, fluctuations of
varying magnitude, coarse-grained measurements, etc.). Interestingly, since the first edition (1992)
of the present book, only certain limited aspects of this theory have been thoroughly discussed in
the literature. Chapter 14 emphasizes the importance of accounting for physical laws, scientific
models, and empirical relationships in the development of a spatiotemporal random field theory.
This valuable core knowledge concerning a phenomenon is usually quantitatively expressed in terms
of stochastic partial differential equations, several of which are reviewed in this chapter. Admittedly, I
only tangentially deal with the solution of these equations and relevant topics. The strongest reason for
excluding the omitted topics is that each would fill another book (I do, of course, give references to the
relevant literature). Chapter 15 presents a series of permissibility criteria for spaceetime covariance
functions (ordinary and generalized) that are widely used in applied stochastics. Certain of these
criteria are necessary and sufficient, whereas some others are only sufficient, but they have the advan-
tage that they refer directly to the covariance function itself. Further, some important practical impli-
cations of permissibility in different kinds of applications are discussed. Chapter 16 presents a rather
large number of techniques for constructing spaceetime covariance models, which can be used in a
variety of scientific applications. Formal and substantive model-building techniques are examined,
each of which has its own merits and limitations. There are many covariance model construction tech-
niques in use and they could not all be discussed in full: my choice of what to present in detail and what
to mention only as a reference simply reflects my personal taste and experience.
The book has benefited by the contributions in the field of my colleagues, collaborators, and stu-
dents during the last few decades. The second edition of the book was written mainly during my leave
of absence year at the Ocean College of Zhejiang University (China). I am grateful for the support of
Zhejiang University and of the CNSF (Grant no. 41671399). I am also grateful to my colleagues at the
Ocean College, particularly Professor Jiaping Wu, who did everything possible to create the right envi-
ronment for writing such a book. Last but not least, this work would not have been completed without
Yongmei’s infinite patience during the long process of writing the book, which is why to Yongmei this
book is dedicated.

George Christakos
This page intentionally left blank
 CHAPTER

SPACE, TIME, SPACEeTIME,

RANDOMNESS, AND
PROBABILITY
I
CHAPTER OUTLINE
1. Introduction .......................................................................................................................................1
2. SpaceeTime Continuum and Kolmogorov Probability Space ..................................................................4
2.1 SpaceeTime Arguments: Points, Lags, Separations, and Metrics...........................................4
2.2 Transformations and Invariance in SpaceeTime.................................................................17
2.3 SpaceeTime Interpretations.............................................................................................23
2.4 Functions of SpaceeTime Arguments ...............................................................................26
3. Random Variables in SpaceeTime.....................................................................................................31
3.1 Kolmogorov’s Probability Theory .......................................................................................31
3.2 Useful Inequalities ..........................................................................................................35
3.3 Convergence of Random Variable Sequences .....................................................................37

1. INTRODUCTION
Due to its importance in almost any scientific discipline, random field theory is an active area of
ongoing research. Significant work has been done, indeed, in the theory of spatial random field, but
much less so in the theory and applications of spatiotemporal random field, where many important
topics still need to be studied and notions to be advanced. On the other hand, many practitioners argue
that random field remains a tough theory to work with, due to the difficulty of the “nondeterministic”
mathematics involved. This kind of mathematics is also known as stochastics, a term that generally
refers to the mathematical representation of phenomena that vary jointly in space and time under
conditions of in situ uncertainty. In a formal sense, deterministic mathematics can be viewed as a
special case of stochastics under the limiting and rare conditions that the phenomenon under study is
known with certainty. To phrase it in more words, stochastics deals with any topic covered by the
deterministic theory of functions, and, in addition, the presence of uncertainty (technically, sometimes
characterized as randomness) makes stochastics a much larger, considerably more complex and surely
more challenging subject than the deterministic theory of functions. Historically, the development of
stochastics can be traced back in the works of some of the world’s greatest scientists, such as Maxwell
(1860), Boltzmann (1868), Gibbs (1902), Einstein (1905), Langevin (1908), Wiener (1930), Heisen-
berg (1930) Khinchin (1934), Kolmogorov (1941), Chandrasekhar (1943), Lévy (1948), Ito (1954),
Gel’fand (1955), von Neumann (1955), Yaglom (1962), and Bohr (1963), among many others.

Spatiotemporal Random Fields. http://dx.doi.org/10.1016/B978-0-12-803012-7.00001-5

Copyright © 2017 Elsevier Inc. All rights reserved.
1
Visit https://ebooksecure.com
now to explore a rich
collection of ebook and enjoy
exciting offers!
2 CHAPTER I SPACE, TIME, SPACEeTIME, RANDOMNESS, AND PROBABILITY

It should be pointed out that random field modeling is at the heart of many theoretical advances in
stochastics. It has led to the development of new mathematical concepts and techniques, and, also, it
has raised several interesting theoretical questions worthy of investigation. Computational random
field modeling, on the other hand, deals with computational and numerical aspects of the systematic
implementation of random field theory in the study of complex real-world phenomena, which covers
almost every scientific and engineering discipline. The term “computational” used here should not
create any confusion with computational mathematics and statistics: while computational mathematics
and statistics are concerned with numbers, computational random fields are concerned with physical
quantities.1
In applied sciences, random field modeling deals with spatiotemporal natural attributes, that is,
real-world attributes that develop simultaneously in space and time, and they are measurable or
observable. These natural attributes occur in nearly all the areas of applied sciences, such as ecology
and environment (e.g., concentrations of pollutants in environmental mediadwater/air/soil/biota),
climate predictions and meteorology (e.g., variations of atmospheric temperature, density, moisture
content, and velocity), hydrology (e.g., water vapor concentrations, soil moisture content, and pre-
cipitation data consisting of long time series at various locations in space), oil reservoir engineering
(e.g., porosities, permeabilities, and fluid saturations during the production phase), environmental
health (e.g., human exposure indicators and doseeeffect associations), and epidemiology (e.g., breast
cancer incidence, and Plague mortality). In all these cases, a central issue of random field modeling is
factual accuracy in the informational statements that describe what was observed and experienced.
For sure, the application of random fields in the study of real-world phenomena is not an un-
constrained theoretical exercise. It rather follows certain methodological criteria that involve the
identification of the bounds of the specific application, the evaluation of the context in making sense of
empirical data, a focus on probative evidence from diverse sources, an openness to inductive insights,
and an in-depth analysis justified by the generation of interpretable results. Induction, interpretation,
and abstraction are not competing objectives in this approach, but mutually reinforcing operations.
Random field modeling is concerned, although to varying extends, about both its internal and
external validity. Internal validity relates to whether the findings or results of the random field modeling
relate to and are caused by the phenomena under investigation, and not by other unaccounted for in-
fluences. On the other hand, external validity is assessed by the extent to which these findings or results
can be generalized, and thus applied to other real-world situations. While internal validity is the primary
concern of random field modeling, external validity is also a very important goal.
In this Chapter, I present the fundamentals regarding the conceptual and quantitative characterization
of spaceetime (or space/time, or spacetime) within which random fields will be defined in subsequent
chapters. Arguably, there are many issues surrounding the use and nature of the notion of “spaceetime”
in scientific modeling, and some of them are even controversial. Yet, “space” and “time” are funda-
mental concepts that were invented by humans in their effort to describe Nature, but the map is not the
territory. The formulation of spaceetime introduced in this chapter has the considerable merit of
maintaining close contact between mathematical description and physical reality. Among the central
goals of this formulation are to direct us toward a correct interpretation of spaceetime, and, to the extend
possible, to help us avoid asking the wrong questions and focus on the insignificant issues.

1
Surely, physical quantities are represented by numbers, but they also possess physical meaning and content (which are
ignored by computational mathematics and statistics), and they are also associated with spaceetime arguments.
 1. INTRODUCTION 3

In this book, random quantities like the random variable, the random field, and the vector random
field will be studied in both the physical (real) and the frequency domains. Notationally, a random
variable is represented by lowercase Latin letters (x, y, etc.), a random field by uppercase Latin letters
(X, Y, etc.), and a vector random field by uppercase bold Latin letters (X, Y, etc.). Lowercase Greek
letters (c, j, etc.) denote random variable or random field values (realizations), and lowercase bold
Greek letters (c, j, etc.) denote vector random field values. The frequency domain counterparts of the
above random quantities will be denoted by inserting the symbol w above them (e.g., X). e The N
denotes the set of natural numbers (f0; 1; 2; .g or f1; 2; .g, depending on the context). The R1 (or R)
and C denote, respectively, the spaces of real and complex numbers. In the latter case, z ¼ c þ ij ˛ C,
pffiffiffiffiffiffiffi
where i ¼ 1, and the c h Re(z) ˛ R1 and j h Im(z) ˛ R1 denote, respectively, the real and
imaginary parts of z. The complex conjugate of z is denoted by z ¼ c  ij ˛ C; also,
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
jzj ¼ jz j ¼ c2 þ j2 ˛R1þ;0 (i.e., the positive part of the real line including zero) is the modulus of z,
z. The symbol T is sometimes used to represent the time domain as a subset of R1 (T 4 R1), which is
in agreement with the physical irreversibility of most real-world phenomena. On the real line R1, I use
the convention for closed, open, and half-open intervals written as [c, j], (c, j), [c, j), and (c, j].
Also, Rn is the Euclidean space of dimension n  1.
Scalar, vector, and matrix notation will be used, noticing that scalars can be seen as tensors of rank
n
zero, vectors have rank Pnone, and matrices have rank two. A vector in R will be denoted as
s ¼ ðs1 ; .; sn Þ or s ¼ i¼1 si εi, where εi, i ¼ 1; .; n, are base vectors along the coordinate di-
rections. The simplest choice of an (orthonormal) basis is, of course, the set of unit length vectors εi,
where the ith component is 1 and all others 0. For any two vectors s and s0 ¼ s01 ; .; s0n , their scalar
P 1
product is defined as s$s0 ¼ ni¼1 si s0i . The length of the vector s is jsj ¼ ðs$sÞ2 , and
1
js  s0 j ¼ ððs  s0 Þ$ðs  s0 ÞÞ2 is the distance between s and s0 in Rn. The spaceetime domain is
denoted as Rnþ1, or as Rn,1, if we want to explicitly distinguish space from time (for the same reason,
we may also denote the spaceetime domain as the Cartesian product Rn  R1 or Rn  T). I.e., in the
case of spaceetime, the domain dimensionality increases to n þ 1 by including the additional term s0
or t representing time. Table 1.1 lists some commonly used symbols. Table 1.2 gives a list of special

Table 1.1 Commonly Used Symbols

Symbol Mathematical Expression
R
Dirac delta dsdðsÞ ¼ 1;
dðss0Þ ¼ 0:

Kronecker delta 1 if i ¼ j;
dij ¼
0 otherwise.
8
< þ1 if ði; j; kÞ ¼ ð1; 2; 3Þ or ð2; 3; 1Þ or ð3; 1; 2Þ;
>
Levi-Chivita
εijk ¼ 1 if ði; j; kÞ ¼ ð3; 2; 1Þ or ð1; 3; 2Þ or ð2; 1; 3Þ;
>
:
0 if i ¼ j or j ¼ k or k ¼ i.

Kronecker/Levi-Chivita εijkεipq ¼ djpdkq  djqdkp

relationship
4 CHAPTER I SPACE, TIME, SPACEeTIME, RANDOMNESS, AND PROBABILITY

Table 1.2 Special Functions and Polynomials

Name Notation

Gamma function. G(n)

Bessel function of 1st kind and vth order. Jv
Modified Bessel function of the 1st kind. In
Modified Bessel function of 2nd kind and vth order. Kv
Gegenbauer polynomials of degrees l, q. Gl,q

functions and polynomials that will be used in the mathematical expression of several results
throughout the book. For the readers’ convenience, the mathematical definitions and basic properties
of these special functions and polynomials are briefly reviewed in the book’s Appendix.
Although many of the theoretical results presented in each chapter of the book (in the form of
propositions and corollaries) will be repeatedly used in subsequent chapters, most proofs and other
details will not, so they will not be discussed. Instead, examples illustrating the most important
application-related aspects of these proofs will be presented when appropriate. A consistent effort is
made throughout the book to keep a balance between abstract mathematical rigor and real-world science.
In many cases, this means that by suppressing certain strict mathematical conditions, a more realistic
representation of the observed phenomenon is achieved, and, also, a richness of new material is produced
(this is, e.g., the case with spaceetime metric). The remaining of this chapter presents a review of the
basic concepts and principles (regarding space, time, field, uncertainty, and probability) around which
the spatiotemporal random field theory will be developed in the following chapters of the book.

2. SPACEeTIME CONTINUUM AND KOLMOGOROV PROBABILITY SPACE

This section presents an overview of the fundamental notions pertaining to the description of the
spaceetime domain shared by the natural phenomena and the mathematical constructions that
represent them. A general point to be stressed is that although on formal mathematics grounds many
aspects of the analysis in the n þ 1-dimensional (spaceetime) domain are similar to those in the
n-dimensional (spatial) domain, essential differences could exist on physical grounds. On the same
grounds, crucial links may emerge between space and time, in which case the deconstruction of the
concept of spaceetime into separate types of space and time may be an unnecessary conceptualization
often leading to unsatisfactory conclusions.
Having said that, I start with the introduction of the different geometrical notions and arguments
that play a central role in the spatiochronological specification of a phenomenon, an attribute or an
event occurring in physical spaceetime, highlighting similarities and differences as they emerge.

2.1 SPACEeTIME ARGUMENTS: POINTS, LAGS, SEPARATIONS, AND METRICS

Generally, a continuum E of spaceetime points is associated with the physical spaceetime domain,
and as such, it is assumed to be equipped with the corresponding metric that specifies “distances”
between points in spaceetime. Operationally, a point in E can be defined in different ways, depending
on the situation and the objectives of the analysis.
 2. SPACEeTIME CONTINUUM AND KOLMOGOROV PROBABILITY SPACE 5

Definition 2.1
A point in E is denoted by a vector p, which can be defined, either as an element of the n þ 1-
dimensional domain,
p ¼ ðs1 ; .; sn ; s0 Þ ˛ Rnþ1 ; (2.1a)
where, in notational terms, s1 ; .; sn denote space coordinates and time is considered jointly with space
using the convention s0 ¼ t; or, as a pair of elements
p ¼ ðs; tÞ ˛ Rn;1 ; (2.1b)
where s ¼ ðs1 ; .; sn Þ ˛ Rn is the spatial location vector considered separately than the time instant
t ˛ R1. As noted earlier, instead of Rn,1 some authors use the Cartesian domain notation Rn  T, i.e., the
time axis T (4R1) is indicated separately from space Rn.
The point vector p plays a key role in physical sciences, not only because it uniquely specifies a point in
the spaceetime domain of interest, but also because of the numerous functions of p (also known as fields,
see Section 2.4) encountered in physical sciences. In Definition 2.1, either the entire real line may be
considered as the time axis (referring, e.g., to situations in which the attribute of interest is reversible), or,
due to physical requirements (i.e., the attribute is irreversible in the real-world), the domain R1þ;f0g
(restricted to the positive part of the real line including zero) may be more appropriate to represent the time
axis. To phrase it in more words, time may be seen as a coordinate (s0) of the vector p in the Rnþ1 domain,
as in Eq. (2.1a). Intuitively, a spaceetime point in Rnþ1 is considered as a fusion of a space point and a
time point (e.g., the “here and now” exists as a unity not specifying the “here” and the “now” separately;
similar is the interpretation of the “there and then”). Alternatively, time may enter the analysis as a distinct
variable via the vectorescalar pair (s, t) of the Rn,1 domain, as in Eq. (2.1b). Both approaches have their
merits and uses, which will be discussed in various parts of the book. Yet, one should be aware of certain
noticeable consequences of the two approaches in applications, as illustrated in the examples below.

Example 2.1
From a modeling viewpoint, in many cases we may treat space and time on essentially the same formal
footing. By setting, e.g., s0 ¼ iat, the wave operator2 can be written as
v2 X X
n n;0
v2 v2
a2  ½$  1  ½$; (2.2)
vt 2
i¼1
vsi
2
i¼1
vs2i

and we may view the solution of Poisson’s equation in two dimensions and that of the wave equation in
one dimension as analogous problems. The physical differences between space and time need to be
carefully taken into consideration, though, since certain formal analogies between space and time may
be deceptive. Indeed, the boundary conditions and the initial conditions may enter the problem in
different ways, even though the governing equation may look symmetric (as in the wave operator
above). Also, when we study the spaceetime variation of an air pollutant, the way pollution con-
centration changes across space (the distribution of spatial locations in which pollution exceeds a
critical threshold) can be essentially different than concentration changes as a function of time (fre-
quency of threshold exceedances at each location).
2
This operator expresses certain fundamental physical laws like the electromagnetic laws.
6 CHAPTER I SPACE, TIME, SPACEeTIME, RANDOMNESS, AND PROBABILITY

Our discussion so far of the issues surrounding the essence and use of the term “spaceetime” lead
to the first postulate.

Postulate 2.1
The vast majority of real-world data are interrelated both in space and time. This spaceetime
connection is ingrained through physical relations and is welcomed in scientific modeling because it
allows the representation of the spaceetime variation of a natural attribute from the limited number of
data usually available.
This postulate is supported by reality, including the fact that spaceetime coupling is known to
remove possibly unphysical divergences from the moments of the corresponding transport processes
(e.g., Shlesinger et al., 1993). Unfortunately, these crucial facts are often ignored in purely technical
treatments of spaceetime phenomena. Indeed, in statistical inferences (i.e., the inductive process of
inferring from a limited sample valid conclusions about the underlying yet unknown population) the
proper assessment of spaceetime correlations is often problematic, since most standard statistics tools of
data analysis and processing have been developed based on the key premise of independent (physical
relation-free) experiments.

Example 2.2
Working in the classical Newtonian conceptual framework, many practitioners find it tempting to
completely separate the space component s from the time component t. Although convenient, indeed,
this approach is often inadequate in real-world studies. A common example is the model-fitting
procedure in which a valid covariance function of space is fitted at any fixed time (or a valid
covariance function of time is fitted at any fixed distance). However, the resulting model is not
necessarily a valid spatiotemporal covariance model, as has been discussed in Ma (2003b). Further-
more, in ocean studies involving underwater acoustics propagation, the travel time t is related to the
horizontal (s1, s2) and vertical (s3) coordinates in R3,1 by means of
 1
s21 þ s22 2 cos q þ s3 sin q  yt ¼ 0; (2.3)
where q is the angle of a ray element in a refracting medium with sound speed y (Lurton, 2010). Hence,
the acoustics of the phenomenon imply that space s and time t are closely linked through the physical
relation of Eq. (2.5).
The perspective suggested by Eq. (2.1b) enables the introduction of alternative expressions of
spaceetime point determination, while still accounting for the spaceetime connection posited in
Postulate 2.1. These expressions are called conditional, and the reason for this will become obvious
below. Consider a spaceetime domain represented by the nodes i (i ¼ 1; .; m) of a grid or lattice. An
obvious expression of spaceetime at each grid node i is
ðsi ; ti Þ; i ¼ 1; .m; (2.4a)
denoting location si ¼ ðsi1 ; .; sin Þ and time ti. Eq. (2.4a) assigns the same subscript to space and time,
and, hence, it allows the consideration of a unique time instant at each node i. Sometimes, the physics
of the situation may require that more than one time instants need to be considered at each node, in
which case it is convenient to represent spaceetime at each grid node by
 
si ; tji ; i ¼ 1; .; m; and j ¼ 1; .; m0 ; (2.4b)
Other documents randomly have
different content
only to catch the fish, their labour being finished as soon as the
boats touch the quay. The Newhaven fishwife’s mode of doing
business is well known. She is always supposed to ask double or
triple what she will take; and, on occasions of bargaining, she is
sure, in allusion to the hazardous nature of the gudeman’s
occupation, to tell her customers that “fish are no fish the day,
they’re just men’s lives.” The style of higgling adopted when dealing
with the fisher-folk, if attempted in other kinds of commerce, gives
rise to the well-known Scottish reproach of “D’ye tak’ me for a
fishwife?” The style of bargain-making carried on by the fishwives
may be illustrated by the following little scene:—
A servant girl having just beckoned to one of them is answered by
the usual interrogatory, “What’s yer wull the day, my bonnie lass?”
and the “mistress” being introduced, the following conversation
takes place:—
“Come awa, mem, an’ see what bonnie fish I hae the day.”
“Have you any haddocks?”
“Ay hae I, mem, an’ as bonnie fish as ever ye clappit yer twa een
on.”
“What’s the price of these four small ones?”
“What’s yer wull, mem?”
“I wish these small ones.”
“What d’ye say, mem? sma’ haddies! they’s no sma’ fish, an
they’re the bonniest I hae in a’ ma creel.”
“Well, never mind, what do you ask for them?”
“Weel, mem, it’s? been awfu’ wather o’ late, an’ the men canna
get fish; ye’ll no grudge me twentypence for thae four?”
“Twentypence!”
“Ay, mem, what for no?”
“They are too dear, I’ll give—”
 “What d’ye say, mem? ower dear! I wish ye kent it: but what’ll ye
gie me for thae four?”
“I’ll give you a sixpence.”
“Ye’ll gie me a what?”
“A sixpence.”
“I daur say ye wull, ma bonny leddy, but ye’ll no get thae four fish
for twa sixpences this day.”
“I’ll not give more.”
“Weel, mem, gude day” (making preparations to go); “I’ll tak’
eighteenpence an’ be dune wi’t.”
“No; I’ll give you twopence each for them.”
And so the chaffering goes on, till ultimately the fishwife will take
tenpence for the lot, and this plan of asking double what will be
taken, which is common with them all and sometimes succeeds with
simple housewives, will be repeated from door to door, till the supply
be exhausted. The mode of doing business with a fishwife is
admirably illustrated in the Antiquary. When Monkbarns bargains for
“the bannock-fluke” and “the cock-padle,” Maggie Mucklebackit asks
four and sixpence, and ends, after a little negotiation and much
finesse, in accepting half-a-crown and a dram; the latter commodity
being worth siller just then, in consequence of the stoppage of the
distilleries.
The fishwives while selling their fish will often say something
quaint to the customer with whom they are dealing. I will give one
instance of this, which, though somewhat ludicrous, is characteristic,
and have no doubt the words were spoken from the poor woman’s
heart. “A fishwife who was crying her “caller cod” in George Street,
Edinburgh, was stopped by a cook at the head of one of the area
stairs. A cod was wanted that day for the dinner of the family, but
the cook and the fishwife could not trade, disagreeing about the
price. The night had been stormy, and instead of the fishwife flying
into a passion, as is their general custom when bargaining for their
fish if opposed in getting their price, the poor woman shed tears,
and said to the cook, ‘Tak’ it or want it; ye may think it dear, but it’s
a’ that’s left to me for a faither o’ four bairns.’”
Notwithstanding, however, their lying and cheating in the streets
during the week when selling their fish, there are no human beings
in Scotland more regular in their attendance at church. To go to their
church on a Sunday, and see the women all sitting with their smooth
glossy hair and snow-white caps, staring with open eyes and mouth
at the minister, as he exhorts them from the pulpit as to what they
should do, one would think them the most innocent and simple
creatures in existence. But offer one of them a penny less than she
feels inclined to take for a haddock, and he is a lucky fellow who
escapes without its tail coming across his whiskers. Of late our
fishwives have been considering themselves of some importance.
When the Queen came first to Edinburgh, she happened to take
notice of them, and every printshop window is now stuck full of
pictures of Newhaven fishwives in their quaint costume of short
petticoats of flaming red and yellow colours.[19]
The sketch of fisher-life in the Antiquary applies as well to the
fisher-folk of to-day as to those of sixty years since. This is
demonstrable at Newhaven; which, though fortunate in having a pier
as a rendezvous for its boats, thus admitting of a vast saving of time
and labour, is yet far behind inland villages in point of sanitary
arrangements. There is in the “town” an everlasting scent of new tar,
and a permanent smell of decaying fish, for the dainty visitors who
go down to the village from Edinburgh to partake of the fish-dinners
for which it is so celebrated. Up the narrow closes, redolent of
“bark,” we see hanging on the outside stairs the paraphernalia of the
fisherman—his “properties,” as an actor would call them; nets,
bladders, lines, and oilskin unmentionables, with dozens of pairs of
those particularly blue stockings that seem to be the universal wear
of both mothers and maidens. On the stair itself sit, if it be
seasonable weather, the wife and daughters, repairing the nets and
baiting the lines—gossiping of course with opposite neighbours, who
are engaged in a precisely similar pursuit; and to-day, as half a
century ago, the fishermen sit beside their hauled-up boats, in their
white canvas trousers and their Guernsey shirts, smoking their short
pipes, while their wives and daughters are so employed, seeming to
have no idea of anything in the shape of labour being a duty of
theirs when ashore. In the flowing gutter which trickles down the
centre of the old village we have the young idea developing itself in
plenty of noise, and adding another layer to the incrustation of dirt
which it seems to be the sole business of these children to collect on
their bodies. These juvenile fisher-folk have already learned from the
mudlarks of the Thames the practice of sporting on the sands before
the hotel windows in the expectation of being rewarded with a few
halfpence. “What’s the use of asking for siller before they’ve gotten
their denner?” we once heard one of these precocious youths say to
another, who was proposing to solicit a bawbee from a party of
strangers.
To see the people of Newhaven, both men and women, one would
be apt to think that their social condition was one of great hardship
and discomfort; but one has only to enter their dwellings in order to
be disabused of this notion, and to be convinced of the reverse of
this, for there are few houses among the working population of
Scotland which can compare with the well-decked and well-
plenished dwellings of these fishermen. Within doors all is neat and
tidy. When at the marriage I have mentioned, I thought the house I
was invited to was the cleanest and the cosiest-looking house I had
ever seen. Never did I see before so many plates and bowls in any
private dwelling; and on all of them, cups and saucers not excepted,
fish, with their fins spread wide out, were painted in glowing
colours; and in their dwellings and domestic arrangements the
Newhaven fishwives are the cleanest women in Scotland, and the
comfort of their husbands when they return from their labours on
the wild and dangerous deep seems to be the fishwife’s chief delight.
I may also mention that none of the young women of Newhaven will
take a husband out of their own community, that they are as rigid in
this matrimonial observance as if they were all Jewesses.[20]
 The following anecdotic illustration of the state of information in
Newhaven sixty years since is highly characteristic:—
A fisherman, named Adam L——, having been reproved pretty
severely for his want of Scripture knowledge, was resolved to baulk
the minister on his next catechetical visitation. The day appointed he
kept out of sight for some time; but at length, getting top-heavy
with some of his companions, he was compelled, after several falls,
in one of which he met with an accident that somewhat disfigured
his countenance, to take shelter in his own cottage. The minister
arrived, and was informed by Jenny, the wife, that her husband was
absent at the fishing. The Doctor then inquired if she had carefully
perused the catechism he had left on his last visit, and being
answered in the affirmative, proceeded to follow up his conversation
with a question or two. “Weel, Jenny,” said the minister, “can ye tell
me the cause o’ Adam’s fall?” By no means versed in the history of
the great progenitor of the human race, and her mind being
exclusively occupied by her own Adam, Janet replied, with some
warmth, “’Deed, sir, it was naething else but drink!” at the same time
calling upon her husband, “Adam, ye may as weel rise, for the
Doctor kens brawly what’s the matter; some clashin’ deevils o’
neebours hae telt him a’ aboot it!”
The remains of many old superstitions are still to be found about
Newhaven. I could easily fill a page or two of this volume with
illustrative anecdotes of sayings and doings that are abhorrent to the
fisher mind. The following are given as the merest sample of the
number that might be collected.
They have several times “gone the round” of the newspapers but
are none the worse for that:—
If an uninitiated greenhorn of a landsman chanced to be on board
of a Newhaven boat, and, in the ignorance and simplicity of his
heart, talked about “salmon,” the whole crew—at least a few years
ago—would start, grasp the nearest iron thowell, and exclaim,
“Cauld iron!” “cauld iron!” in order to avert the calamity which such a
rash use of the appellation was calculated to induce; and the said
uninitiated gentleman would very likely have been addressed in
some such courteous terms as “O ye igrant brute, cud ye no ca’d it
redfish?” Woe to the unfortunate wight—be he Episcopalian or
Presbyterian, Churchman or Dissenter—who being afloat talks about
“the minister:” there is a kind of undefined terror visible on every
countenance if haply this unlucky word is spoken; and I would
advise my readers, should they hereafter have occasion, when
water-borne, to speak of a clergyman, to call him “the man in the
black coat;” the thing will be equally well understood, and can give
offence to none. I warn them, moreover, to be guarded and
circumspect should the idea of a cat or a pig flit across their minds;
and should necessity demand the utterance of their names, let the
one be called “Theebet” and the other “Sandy;” so shall they be
landed on terra firma in safety, and neither their ears nor their
feelings be insulted by piscatory wit. In the same category must be
placed every four-footed beast, from the elephant moving amongst
the jungles of Hindostan to the mouse that burrows under the
cottage hearth-stone. Some quadrupeds, however, are more
“unlucky” than others; dogs are detestable, hogs horrible, and hares
hideous! It would appear that Friday, for certain operations, is the
most unfortunate; for others the most auspicious day in the week.
On that day no sane fisherman would commence a Greenland
voyage, or proceed to the herring-ground, and on no other day of
the week would he be married.
In illustration of the peculiar dread and antipathy of fishermen to
swine, I give the following extract from a volume published by a
schoolmaster, entitled An Historical Account of St. Monance. The
town is divided into two divisions, the one called Nethertown and the
other Overtown—the former being inhabited entirely by fishermen,
and the latter by agriculturists and petty tradesmen:—“The
inhabitants of the Nethertown entertained a most deadly hatred
towards swine, as ominous of evil, insomuch that not one was kept
amongst them; and if their eyes haplessly lighted upon one in any
quarter, they abandoned their mission and fled from it as they would
from a lion, and their occupation was suspended till the ebbing and
flowing of the tide had effectually removed the spell. The same
devils were kept, however, in the Uppertown, frequently affording
much annoyance to their neighbours below, on account of their
casual intrusions, producing much damage by suspension of labour.
At last, becoming quite exasperated, the decision of their oracle was
to go in a body and destroy not the animals (for they dared not hurt
them), but all who bred and fostered such demons, looking on them
with a jealous eye, on account of their traffic. Armed with boat-
hooks, they ascended the hill in formidable procession, and dreadful
had been the consequence had they not been discovered. But the
Uppertown, profiting by previous remonstrance, immediately let
loose their swine, whose grunt and squeak chilled the most heroic
blood of the enemy, who, on beholding them, turned and fled down
the hill with tenfold speed, more exasperated than ever, secreting
themselves till the flux and reflux of the tide had undone the
enchantment.... According to the most authentic tradition, not an
animal of the kind existed in the whole territories of St. Monance for
nearly a century; and, even at the present day, though they are fed
and eaten, the fisher people are extremely averse to looking on
them or speaking of them by that name; but, when necessitated to
mention the animal, it is called ‘the beast’ or ‘the brute’ and, in case
the real name of the animal should accidentally be mentioned, the
spell is undone by a less tedious process—the exclamation of ‘cauld
iron’ by the person affected being perfectly sufficient to counteract
the evil influence. Cauld iron, touched or expressed, is understood to
be the first antidote against enchantment.”
At Fisherrow, a few miles east from Newhaven, there is another
fishing community, who also do business in Edinburgh, and whose
manners and customs are quite as superstitious as those of the folks
I have been describing. “The Fisher-raw wives,” in the pre-railway
times, had a much longer walk with their fish than the Newhaven
women; neither were they held in such esteem, the latter looking
upon themselves as the salt of their profession. Dr. Carlyle of
Inveresk, whose memoirs were recently published, in writing of the
Fisherrow women of his time, says:—“When the boats come in late
to the harbour in the forenoon, so as to leave them no more than
time to reach Edinburgh before dinner, it is not unusual for them to
perform their journey of five miles by relays, three of them being
employed in carrying one basket, and shifting it from one to another
every hundred yards, by which means they have been known to
arrive at the fishmarket in less than three quarters of an hour. It is a
well-known fact, that three of these women went from Dunbar to
Edinburgh, a distance of twenty-seven miles, with each of them a
load of herrings on her back of 200 pounds, in five hours.” Fatiguing
journeys with heavy loads of fish are now saved to the wives of both
villages, as dealers attend the arrival of the boats, and buy up all the
sea produce that is for sale. In former times there used to be great
battles between the men of Newhaven and the men of Fisherrow,
principally about their rights to certain oyster-scalps. The Montagues
and Capulets were not more deadly in their hatreds than these rival
fishermen. Now the oyster-grounds are so well defined that battles
upon that question are never fought.
Fisherrow has long been distinguished for its race of hardy and
industrious fishermen, of whom there are about two hundred in all.
They go to the herring-fishing at Caithness, at North Sunderland, at
Berwick, North Berwick, and Dunbar, and about sixty men go to
Yarmouth, on the east coast of England, a distance of about 300
miles. Ten boats, with a complement of eight men each, go to the
deep-sea white-fishing, and two or three boats to the oyster-
dredging.
The white-fishing of Fisherrow has long been a staple source of
income. At what time a colony of fishermen was established at that
village is unknown. They are most likely coeval with the place itself.
When the Reverend Dr. Carlyle, minister of the parish of Inveresk,
wrote (about 1790) there were forty-nine fishermen and ninety
fishwives, but since that time the numbers of both have of course
much increased.
The system of merchandise followed by the fishwives in the old
days of creel-hawking, and even yet to a considerable extent, was
very simple. Having procured a supply of fish, which having
bestowed in a basket of a form fitted to the back, they used to
trudge off to market under a load which most men would have had
difficulty in carrying, and which would have made even the strongest
stagger. Many of them still proceed to the market, and display their
commodities; but the majority, perhaps, perambulate the streets of
the city, emitting cries which, to some persons, are more loud than
agreeable, and which a stranger would never imagine to have the
most distant connection with fish. Occasionally, too, they may be
seen pulling the door-bell of some house where they are in the habit
of disposing of their merchandise, with the blunt inquiry, “Ony
haddies the day?”[21]
While treating of the peculiarities of these people, I may record
the following characteristic anecdote:—“A clergyman, in whose
parish a pretty large fishing-village is situated, in his visitations
among the families of the fish-carriers found that the majority of
them had never partaken of the sacrament. Interrogating them
regarding the reason of this neglect, they candidly admitted to him
that their trade necessarily led them so much to cheat and tell lies,
that they felt themselves unqualified to join in that religious duty.” It
is but justice, however, to add that, when confidence is reposed in
them, nothing can be more fair and upright than the dealings of the
fisher class; and, as dealers in a commodity of very fluctuating
value, they cannot perhaps be justly blamed for endeavouring to sell
it to the best advantage.
At Prestonpans, and the neighbouring village of Cockenzie, the
modern system, as I may call it, for Scotland, of selling the fish
wholesale, may be seen in daily operation. When the boats arrive at
the boat-shore, the wives of those engaged in the fishing are in
readiness to obtain the fish, and carry them from the boats to the
place of sale. They are at once divided into lots, and put up to
auction, the skipper’s wife acting as the George Robins of the
company, and the price obtained being divided among the crew, who
are also, generally speaking, owners of the boat. Buyers, or their
agents, from Edinburgh, Glasgow, Liverpool, Manchester, etc., are
always ready to purchase, and in a few hours the scaly produce of
the Firth of Forth is being whisked along the railway at the rate of
twenty miles an hour. This system, which is certainly a great
improvement on the old creel-hawking plan, is a faint imitation of
what is done in England, where the owners of fishing-smacks
consign their produce to a wholesale agent at Billingsgate, who sells
it by auction in lots to the retail dealers and costermongers.
Farther along on the Scottish east coast is North Berwick, now a
bathing resort, and a fishing town as well; and farther east still is
Dunbar, the seat of an important herring-fishery—grown from a
fishing village into a country town, in which a mixture of agricultural
and fishing interests gives the place a somewhat heterogeneous
aspect; and between St. Abb’s Head and Berwick-on-Tweed is
situated Eyemouth, a fishing-village pure and simple, with all that
wonderful filth scattered about which is a sanitary peculiarity of such
towns. The population of Eyemouth is in keeping with the outward
appearance of the place. As a whole, they are a rough uncultivated
people, and more drunken in their habits than the fishermen of the
neighbouring villages. Coldingham shore, for instance, is only three
miles distant, and has a population of about one hundred fishermen,
of a very respectable class, sober, well dressed, and “well-to-do.” A
year or two ago an outburst of what is called “revivalism” took place
at Eyemouth, and seemed greatly to affect it. The change produced
for a time was unmistakable. These rude unlettered fishermen
ceased to visit the public-houses, refrained from the use of oaths,
and instead sang psalms and said prayers. But this wave of
revivalism, which passed over other villages besides Eyemouth, has
rolled away back, and in some instances left the people worse than
it found them; and I may perhaps be allowed to cite the fish-tithe
riots as a proof of what I say. These riots, for which the rioters were
tried before the High Court of Justiciary at Edinburgh, and some of
them punished, arose out of a demand by the minister for his tithe
of fish.
 Crossing the Firth of Forth, the cost of Fife, from Burntisland to
“the East Neuk,” will be found studded at intervals with quaint
fishing-villages; and the quaintest among the quaint is Buckhaven.
Buckhaven, or, as it is locally named, Buckhyne, as seen from the
sea, is a picturesque group of houses sown broadcast on a low cliff.
Indeed, most fishing villages seem thrown together without any kind
of plan. The local architects had never thought of building their
villages in rows or streets; as the fisher-folks themselves say, their
houses are “a’ heids and thraws,” that is, set down here and there
without regard to architectural arrangement. The origin of
Buckhaven is rather obscure: it is supposed to have been founded by
the crew of a Brabant vessel, wrecked on that portion of the Fife
coast in the reign of Philip II. The population are, like most of their
class, a peculiar people, living entirely among themselves; and any
stranger settling among them is viewed with such suspicion that
years will often elapse before he is adopted as one of the
community. One of the old Scottish chap-books is devoted to a satire
of the Buckhaven people. These old chap-books are now rare, and to
obtain them involves a considerable amount of trouble. Thirty years
ago the chapmen were still carrying them about in their packs; now
it is pleasing to think they have been superseded by the admirable
cheap periodicals which are so numerous and so easy to purchase.
The title of the chap-book referred to above is, The History of
Buckhaven in Fifeshire, containing the Witty and Entertaining
Exploits of Wise Willie and Witty Eppie, the Ale-wife, with a
description of their College, Coats of Arms, etc. It would be a strong
breach of etiquette to mention the title of this book to any of the
Buckhaven people; it is difficult to understand how they should feel
so sore on the point, as the pamphlet in question is a collection of
very vulgar witticisms tinged with such a dash of obscenity as
prevents their being quoted here. The industrious fishermen of
Buckhaven are moral, sober, and comparatively wealthy. Indeed,
many of the Scottish fisher-folk are what are called “warm” people;
and there are not in our fishing villages such violent alternations of
poverty and prosperity as are to be found in places devoted entirely
to manufacturing industry. There is usually on the average of the
year a steady income, the people seldom suffering from “a hunger
and a burst,” like weavers or other handicraftsmen.
As denoting the prosperous state of the people of Buckhaven, it
may be stated that most of the families there have saved money;
and, indeed, some of them are comparatively wealthy, having a bank
account, as well as considerable capital in boats, nets, and lines.
Fishermen, being much away from home, at the herring-fishery or
out at the deep-sea fishing, have no temptation to spend their
earnings or waste their time in the tavern. Indeed, in some Scottish
fishing villages there is not even a single public-house. The
Buckhaven men delight in their boats, which are mostly “Firth-
built,”—i.e. built at Leith, on the Firth of Forth. Many of the boats
used by the Scottish fishermen are built at that port: they are all
constructed with overlapping planks; and the hull alone of a boat
thirty-eight feet in length will cost a sum of £60. Each boat, before it
can be used for the herring or deep-sea fishery must be equipped
with a set of nets and lines; say, a train of thirty-five nets, at a cost
of £4 each, making a sum of £140; which, with the price of the hull,
makes the cost £200, leaving the masts and sails, as well as inshore
and deep-sea lines and many other etceteras, to be provided for
before the total cost can be summed up. The hundred boats which
belong to the men of Buckhaven consequently represent a
considerable amount of capital. Each boat with its appurtenances
has generally more than one owner; in other words, it is held in
shares. This is rather an advantage than otherwise, as every vessel
requires a crew of four men at any rate, so that each boat is usually
manned by two or three of its owners—a pledge that it will be
looked carefully after and not be exposed to needless danger. With
all the youngsters of a fishing village it is a point of ambition to
obtain a share of a boat as soon as ever they can; so that they save
hard from their allowances as extra hands, in order to attain as early
as possible to the dignity of proprietorship. We look in vain, except
at such wonderful places as Rochdale, to find manufacturing
operatives in a similar financial position to these Buckhaven men; in
fact, our fishermen have been practising the plan of co-operation for
years without knowing it, and without making it known. The co-
operative system seems to prevail among the English fisher-folk as
well. At Filey, on the Yorkshire coast, many of the large fishing yawls
—these vessels average about 40 tons each—are built by little
companies and worked on the sharing principle: so much to the men
who find the bait, and so much to each man who provides a net;
and a few shillings per pound of the weekly earnings of the ship go
to the owners. In France there are various ways of engaging the
boats and conducting the fisheries. There are some men who fish on
their own account, who have their own boat, sail, and nets, etc., and
who find their own bait, whether at the sardine-fishery or when
prosecuting any other branch of the sea fisheries. Of course these
boat-owners hire what assistance they require, and pay for it. There
are other men again who hire a boat and work it on the sharing
plan, each man getting so much, the remainder being left for the
owner. A third class of persons are those who work off their
advances: these are a class of men so poor as to be obliged to pawn
their labour to the boat-owners long before it is required. We can
parallel this at home in the herring-fishery, where the advance of
money to the men has become something very like a curse to all
concerned.
The joint-stock fishing system has been prevalent in Scotland, with
various modifications, for a very long period. Ship-carpenters at one
time used to speculate in the fisheries, and build boats in order to
give fishermen a share in them, and persons who had nets would
lend them out on condition of getting a share in the speculation. The
two or three fishermen chiefly concerned would assume a few
landsmen as assistants. At the end of the season the proceeds of the
fishing were divided; the proprietors of the boat drew each one deal,
every man half a deal, and every net was awarded half a deal. The
landsmen, being counted as boys, only drew a quarter of a deal.
The retired Buckhaven fishermen can give interesting information
about the money value of the fisheries. One, who was a young
fellow five-and-twenty years ago, told me the herring-fishery was a
kind of lottery, but that, on an average of years, each boat would
take annually something like a hundred crans—the produce, in all
cases where the crew were part owners, after deducting a fifth part
or so to keep up the boat, being equally divided. “When I was a
younker, sir,” said this person, “there was lots o’ herrin’, an’ we had a
fine winter fishin’ as well, an’ sprats in plenty. As to white fish, they
were abundant five-an’-twenty year ago. Haddocks now are scarce
to be had; being an inshore fish, they’ve been a’ ta’en, in my
opinion. Line-fishin’ was very profitable from 1830 to 1840. I’ve seen
as many as a hunder thoosand fish o’ ae kind or anither ta’en by the
Buckhyne boats in a week—that is, countin’ baith inshore boats an’
them awa at the Dogger Bank. The lot brocht four hunder pound;
but a’ kinds of fish are now sae scarce that it taks mair than dooble
the labour to mak the same money that was made then.”
In the pre-railway era, most of the fishermen along the east coast
of Fife (at Buckhaven, Cellardyke, St. Monance, and Pittenweem), as
also the fishermen along the south coast (North Bewick, Dunbar,
Eyemouth, and Burnmouth), used to carry their catchings of white
fish to villages up the Firth of Forth, and dispose of them to cadgers
and creel-hawkers, who had the retail trade of Edinburgh and Leith
in their own hands. These persons distributed themselves over the
country in order to dispose of their fish, and some of them would
return with farm-produce in its place. The profits realised from thus
retailing the produce of fishermen belonging to distant villages
enabled those who resided on firths bordering the large towns and
cities quietly to lie on their oars. Railways having given facilities to
the east coast of Fife fishers, as well as those on the opposite coast,
to send their produce to market from their own respective villages,
and a new class of traders having arisen—viz. fishmongers having
retail shops—the creel-hawking trade is now fast declining, and as a
following result so also must be the material wealth of the villages
that were in a great measure dependent upon it. In fact, railways
have quite revolutionised the fish trade. There are a few females,
formerly creel-hawkers, who continue still to act as retailers of fish.
But many of them have taken shops, and others stalls in retail
markets, and attend the wholesale market regularly to purchase
their supplies. These retail dealers in fish do remarkably well; but
those who still continue to hawk about a few haddocks or whitings
when they can be procured find that creel-hawking is but a
precarious trade.
I will now carry the reader with me to a very quaint place indeed,
the scene of Sir Walter Scott’s novel of The Antiquary—Auchmithie;
and then on to Fittie, at Aberdeen—another fishing quarter of great
originality: we will go in the steamer.
Steamboat travelling has been in some degree superseded by the
railway carriage; but to tourists going to Inverness or Thurso the
steamer has its attractions. It is preferable to the railroad when the
time occupied in the journey is not an object. On board a fine
steamboat one has opportunities to study character, and there are
always a few characters on board a coasting steam-vessel. And
going north from Edinburgh the coast is interesting. The steamer
may pass the Anster or Dunbar herring-fleet.

“Up the waters steerin’,

The boats are thick and thrang;
Aboon the Bass they’re bearin’,
They’ll shoot their nets ere lang.

“The morn, like siller glancin’,

They’ll haul them han’ to han’;
Syne doon the water dancin’,
Come hame wi’ sixty cran.”

The passengers can see the Bell Rock lighthouse, and think of the
old legend of the pirate who took away the floating bell that had
been erected by a pious abbot on the Inchcape Rock as a warning to
mariners, and who was promptly punished for his sin by being
shipwrecked on the very rock from which he had carried off the bell.
After leaving Aberdeen, the Buffers of Buchan are among the
wonders of the shore, and the sea soughs at times with mournful
cadence in the great caverns carved out by the waves on the
precipitous coast, or it foams and lashes with majestic fury, seeking
to add to its dominions. All the way, till the Old Man of Wick is
descried, guarding the entrance of Pulteneytown harbour, there are
ruined castles, and ancient spires, and curious towers perched on
high sea-cliffs; or there are frowning hills and screaming sea-birds to
add to the poetry of the scene. And along these storm-washed
coasts there are wonders of nature that show the strong arm of the
water, and mark out works that human ingenuity could never have
achieved. Loch Katrine and the Pass of Glencoe have been the
fashion ever since Sir Walter Scott made Scotland; but there are
other places besides these that are worth visiting.
The supposed scene of Sir Walter Scott’s novel of The Antiquary,
on the coast of Forfarshire, presents a conjunction of scenic and
industrial features which commends it to notice. At Auchmithie,
which is distant a few miles from Arbroath, there is often some
cause for excitement; and a real storm or a real drowning is
something vastly different from the shipwreck in the drama of The
Tempest, or the death of the Colleen Bawn. The beetling cliffs
barricading the sea from the land may be traversed by the tourist to
the music of the everlasting waves, the dashing of which only makes
the deep solitude more solemn; the sea-gull sweeps around with its
shrill cry, and playful whales gambol in the placid waters.
The village of Auchmithie, which is wildly grand and romantic,
stands on the top of the cliffs, and as the road to it is steep a great
amount of labour devolves on the fishermen in carrying down their
lines and nets, and carrying up their produce, etc. One customary
feature observed by strangers on entering Auchmithie is, that when
met by female children they invariably stoop down, making a very
low curtsey, and for this piece of polite condescension they expect
that a few halfpence will be thrown to them. If you pass on without
noticing them they will not ask for anything, but once throw them a
few halfpence and a pocketful will be required to satisfy their
importunities. There are two roads leading to Auchmithie from
Arbroath, one along the sea-coast, the other through the country.
The distance is about 3½ miles in a north-east direction, and the
country road is the best; and approaching the village in that
direction it has a very fair aspect. Two rows of low-built slate-roofed
houses, and a school and chapel, stand a few yards off by
themselves. On the north side of the village is a stately farm-house,
surrounded by trees, and on the south side a Coast-Guard station,
clean, white-washed, and with a flagstaff, giving the whole a regular
and picturesque appearance. Entering the village of Auchmithie from
the west, and walking through to the extreme east end, the
imagination gets staggered to think how any class of men could
have selected such a wild and rugged part of the coast for pursuing
the fishing trade—a trade above all others that requires a safe
harbour where boats can be launched and put to sea at a moment’s
warning if any signals of distress be given. The bight of Auchmithie
is an indentation into rocky cliffs several hundred feet in
perpendicular height. About the middle of the bight there is a steep
ravine or gully with a small stream, and at the bottom of this ravine
there is a small piece of level ground where a fish-curing house is
erected, and where also the fishermen pull up their boats that they
may be safe from easterly gales. There are in all about seventeen
boats’ crews at Auchmithie. Winding roads with steps lead down the
side of the steep brae to the beach. There are a few half-tide rocks
in the bight that may help to break the fury of waves raised by
easterly winds; but there is no harbour or pier for the boats to land
at or receive shelter from, and this the fishermen complain of, as
they have to pay £2 a year for the privilege of each boat. The beach
is steep, and strewed with large pebbles, excellently adapted, they
say, for drying fish upon.
The visitor, in addition to studying the quaint people, may explore
one of the vast caves which only a few years ago were the nightly
refuge of the smuggler. Brandy Cove and Gaylet Pot are worth
inspection, and inspire a mingled feeling of terror and grandeur. The
visitor may also take a look at the “Spindle”—a large detached piece
of the cliffs, shaped something like a corn-stack, or a boy’s top with
the apex uppermost. When the tide is full this rock is surrounded
with water, and appears like an island. Fisher-life may be witnessed
here in all its unvarnished simplicity. Indeed nothing could well be
more primitive than their habits and mode of life. I have seen the
women of Auchmithie “kilt their coats” and rush into the water in
order to aid in shoving off the boats, and on the return of the little
fleet carry the men ashore on their brawny shoulders with the
greatest ease and all the nonchalance imaginable, no matter who
might be looking at them. Their peculiar way of smoking their
haddocks may be taken as a very good example of their other
modes of industry. Instead of splitting the fish after cleaning them,
as the regular curers do, they smoke them in their round shape.
They use a barrel without top or bottom as a substitute for a curing-
house. The barrel being inserted a little distance in the ground, an
old kail-pot or kettle, filled with sawdust, is placed at the bottom,
and the inside is then filled with as many fish as can conveniently be
hung in it. The sawdust is then set fire to, and a piece of canvas
thrown over the top of the barrel: by this means the females of
Auchmithie smoke their haddocks in a round state, and very
excellent they are when the fish are caught in season. The daily
routine of fisher-life at Auchmithie is simple and unvarying; year by
year, and all the year round, it changes only from one branch of the
fishery to another. The season, of course, brings about its joys and
sorrows: sad deaths, which overshadow the village with gloom; or
marriages, when the people may venture to hold some simple fête,
but only to send them back with renewed vigour to their
occupations. Time, as it sweeps over them, only indicates a period
when the deep-sea hand-lines must be laid aside for the herring-
drift, or when the men must take a toilsome journey in search of
bait for their lines. Their scene of labour is on the sea, ever on the
sea; and, trusting themselves on the mighty waters, they pursue
their simple craft with persevering industry, never heeding that they
are scorched by the suns of summer or benumbed by the frosts of
winter. There is, of course, an appropriate season for the capture of
each particular kind of fish. There are days when the men fish
inshore for haddocks; and there are times when, with their frail
vessels, the fishermen sail long distances to procure larger fish in
the deep seas, and when they must remain in their open boats for a
few days and nights. But the El-dorado of all the coast tribe is “the
herring.” This abounding and delightful fish, which can be taken at
one place or another from January to December, yields a six weeks’
fishing in the autumn of the year, to which, as has already been
stated, all the fisher-folk look forward with hope, as a period of
money-making, and which, so far as the young people are
concerned, is generally expected to end, like the third volume of a
love-story, in matrimony.
Taking a jump from Auchmithie, it is desirable to pause a moment
at the small fishing village of Findon, in the parish of Banchory-
Devenick, in Kincardineshire, in order to say a few words about a
branch of industry in connection with the fisheries that is peculiar to
Scotland. Yarmouth is famed for its “bloaters,” a preparation of
herrings slightly smoked, well known over England; and in Scotland,
as has already been mentioned in a previous chapter, there is that
unparagoned dainty, the “Finnan haddock,” the best accompaniment
that can be got to the other substantial components of a Scottish
breakfast. Indeed, the Finnan haddock is celebrated as a breakfast
luxury all over the world, although it is so delicate in its flavour, and
requires such nicety in the cure, that it cannot be enjoyed in
perfection at any great distance from the sea-coast. George IV., who
had certainly, whatever may have been his other virtues, a kingly
genius in the matter of relishes for the palate (does not the world
owe to him the discovery of the exquisite propriety of the sequence
of port wine after cheese?), used to have genuine Finnan haddocks
always on his breakfast-table, selected at Aberdeen and sent express
by coach every day for his Majesty’s use. Great houses of brick have
now been erected at various places on the Moray Firth and
elsewhere; and in these immense quantities of haddocks and other
fish are smoked for the market by means of burning billets of green
wood. Formerly the fisher-folk used to smoke a few haddocks in
their cottages over their peat-fires for family use. I have already
described how the fame of the Finnan haddock arose. The trade
soon grew so large that it required a collection to be made in the
fishing districts in order to get together the requisite quantity; so
that what was once a mere local effort has now become a prominent
branch of the fish trade. But it is seldom that the home-smoked fish
can be obtained, with its delicate flavour of peat-reek. The
manufactured Finnan or yellow haddock, smoked in a huge
warehouse, is more plentiful, of course, but it has lost the old relish.
It is pleasant to see the clean fireside and the clear peat-fire in the
comfortably-furnished cottage, with the children sitting round the
ingle on the long winter evenings, listening to the tales and
traditions of the coast, the fish hanging all over the reeking peats,
acquiring the while that delicate yellow tinge so refreshing to the
eyes of all lovers of a choice dish.
Footdee, or “Fittie” as it is locally called, is a quaint suburb of
Aberdeen, figuring not a little, and always with a kind of comic
quaintness, in the traditions of that northern city, and in the stories
which the inhabitants tell of each other. They tell there of one
Aberdeen man, who, being in London for the first time, and visiting
St. Paul’s, was surprised by his astonishment at its dimensions into
an unusual burst of candour. “My stars!” he said, “this maks a
perfect feel (fool) o’ the kirk o’ Fittie.” Part of the quaint interest thus
attached to this particular suburb by the Aberdonians themselves
arises from its containing a little colony or nest of fisher-folk, of
immemorial antiquity. There are about a hundred families living in
Fittie, or Footdee Square, close to the sea, where the Dee has its
mouth. This community, like all others made up of fishing-folk, is a
peculiar one, and differs of course from those of other working-
people in its neighbourhood. In many things the Footdee people are
like the gipsies. They rarely marry, except with their own class; and
those born in a community of fishers seldom leave it, and very
seldom engage in any other avocation than that of their fathers. The
squares of houses at Footdee are peculiarly constructed. There are
neither doors nor windows in the outside walls, although these look
to all the points of the compass; and none live within the square but
the fishermen and their families, so that they are as completely
isolated and secluded from public gaze as are a regiment of soldiers
within the dead walls of a barrack. The Rev. Mr. Spence, of Free St.
Clement’s, lately completed plans of the entire “toun,” giving the
number and the names of the tenants in every house; and from
these exhaustive plans it appears that the total population of the
two squares was 584—giving about nine inmates for each of these
two-roomed houses. But the case is even worse than this average
indicates. “In the South Square only eight of the houses are
occupied by single families; and in the North Square only three, the
others being occupied by at least two families each—one room
apiece—and four single rooms in the North Square contain two
families each! There are thirty-six married couples and nineteen
widows in the twenty-eight houses; and the number of distinct
families in them is fifty-four.” The Fittie men seem poorer than the
generality of their brethren. They purchase the crazy old boats of
other fishermen, and with these, except in very fine weather, they
dare not venture very far from “the seething harbour-bar;” and the
moment they come home with a quantity of fish the men consider
their labours over, the duty of turning the fish into cash devolving, as
in all other fishing communities, on the women. The young girls, or
“queans,” as they are called in Fittie, carry the fish to market, and
the women sit there and sell them; and it is thought that it is the
officious desire of their wives to be the treasurers of their earnings
that keeps the fishermen from being more enterprising. The women
enslave the men to their will, and keep them chained under petticoat
government. Did the women remain at home in their domestic
sphere, looking after the children and their husbands’ comforts, the
men would then pluck up spirit and exert themselves to make money
in order to keep their families at home comfortable and respectable.
Just now there are many fishermen who will not go to sea as long as
they imagine their wives have got a penny left from the last hawking
excursion. There is no necessity for the females labouring at out-
door work. There are few trades in this country where industrious
men have a better chance to make money than fishermen have,
especially when they are equipped with proper machinery for their
calling. At Arbroath, Auchmithie, and Footdee (Fittie), the fishing
population are at the very bottom of the scale for enterprising habits
and social progress. When the wind is in any way from the eastward,
or in fact blowing hard from any direction, the fishermen at these
places are very chary about going to sea unless dire necessity urges
them.
The people of “Fittie” are progressing in morals and civilisation.
One of the local journalists who took the trouble to visit the place
lately, in order to describe truthfully what he saw, says:—“They have
the reputation of being a very peculiar people, and so in many
respects they are; but they have also the reputation of being a
dirtily-inclined and degraded people, and this we can certify from
personal inspection they are not. We have visited both squares, and
found the interior of the houses as clean, sweet, and wholesome as
could well be desired. Their white-washed walls and ceiling, their
well-rubbed furniture, clean bedding, and freshly-sanded floors,
present a picture of tidiness such as is seldom to be met with among
classes of the population reckoned higher in the social scale. And
this external order is only the index of a still more important change
in the habits and character of our fisher-toun, the population of
which, all who know it agree in testifying, has within the past few
years undergone a remarkable change for the better in a moral point
of view. Especially is this noticed in the care of their children, whose
education might, in some cases, bring a tinge of shame to the cheek
of well-to-do town’s folks. Go down to the fisher squares, and lay
hold of some little fellow hardly able to waddle about without
assistance in his thick made-down moleskins, and you will find he
has the Shorter Catechism at his tongue-end. Ask any employer of
labour in the neighbourhood of the shore where he gets his best
apprentices, and he will tell you that for industry and integrity he
finds no lads who surpass those from the fisher squares. Inquire
about the families of the fishermen who have lost their lives while
following their perilous occupation, and you will find that they have
been divided among other families in the square, and treated by the
heads of these families as affectionately as if they had been their
own.”
 As regards the constant intermarrying of the fisher class, and the
working habits of their women, I have read an Italian fable to the
following effect:—“A man of distinction, in rambling one day through
a fishing-village, accosted one of the fishermen with the remark that
he wondered greatly that men of his line of life should chiefly
confine themselves, in their matrimonial connections, to women of
their own caste, and not take them from other classes of society,
where a greater security would be obtained for their wives keeping a
house properly, and rearing a family more in accordance with the
refinement and courtesies of life. To this the fisherman replied, that
to him, and men of his laborious profession, such wives as they
usually took were as indispensable to their vocation as their boat
and nets. Their wives took their fish to market, obtained bait for
their lines, mended their nets, and performed a thousand different
and necessary things, which husbands could not do for themselves,
and which women taken from any other of the labouring classes of
society would be unable to do. ‘The labour and drudgery of our
wives,’ continued he, ‘is a necessary part of our peculiar craft, and
cannot by any means be dispensed with, without entailing
irreparable injury upon our social interests.’ Moral.—This is one
among many instances, where the solid and the useful must take
precedence of the showy and the elegant.”
As I have already mentioned, the fishers are intensely
superstitious. No matter where we view them, they are as much
given to signs and omens at Portel near Boulogne as at Portessie
near Banff. For instance, whilst standing or walking they don’t like to
be numbered. Rude boys will sometimes annoy them by shouting—

“Ane, twa, three;

What a lot o’ fisher mannies I see!”

It is also considered very offensive to ask fisher-people, whilst on

their way to their boats, where they are going to-day; and they do
not like to see, considering it unlucky, the impression of a very flat
foot upon the sand; neither, as I have already explained, can they go
to work if on leaving their homes in the morning a pig should cross
their path. This is considered a particularly unlucky omen, and at
once drives them home. Before a storm, it is usually thought, there
is some kind of warning vouchsafed to them; they see, in their
mind’s eye doubtless, a comrade wafted homeward in a sheet of
flame, or the wraith of some one beckons them with solemn gesture
landward, as if saying, “Go not upon the waters.” When an accident
happens from an open boat, and any person is drowned, that boat is
never again used, but is laid up high and dry, and allowed to rot
away—rather a costly superstition. Then, again, some fisher-people
perform a kind of “rite” before going to the herring-fishery, in
drinking to a “white lug”—that is, that when they “pree” or examine
a corner or lug of their nets, they may find it glitter with the silvery
sheen of the fish, a sure sign of a heavy draught.
But the fishermen of other coasts are quite as quaint,
superstitious, and peculiar as those of our own. The residents in the
Faubourg de Pollet of Dieppe are just as much alive to the signs and
tokens of the hour as the dwellers in the square of Fittie, or those
who inhabit the fishing quarter of Boulogne. It is a pity that the
guide-books say so little about these and similar places. The fishing
quarter of Boulogne is not unlike Newhaven: there is the same
“ancient and fish-like smell,” the same kind of women with a very
short petticoat, the only difference being that our Scottish fishwives
wear comfortable shoes and stockings. We can see too the dripping
nets hung up to dry from the windows of the tumble-down-like
houses, and the gamins of Boulogne lounge about the gutter’s side
on the large side stones, or run up and down the long series of steps
just the same as the fisher-folks’ children do at home.
 A FRENCH FISHWOMAN.

It is only, however, by penetrating into the quaint villages situated

on the coasts of Normandy and Brittany, that we can gain a
knowledge of the manners and customs of those persons who are
daily engaged in prosecuting the fisheries. The clergymen of their
districts, as may be supposed, have great power over them, and all
along the French coast the fisher-people have churches of their own,
and they are constantly praying for “luck,” or leaving propitiatory
gifts upon the altars, as well as going pilgrimages in order that their
wishes may be realised. A dream is thought of such great
consequence among these people, that the women will hold a
conference, early in the day, in order to its interpretation. Each little
village has its storied traditions, many of them of great interest, and
some of them very romantic. I can only briefly allude, however, to
one of these little stories. Some of my readers may have heard of
the Bay of the Departed on the coast of Brittany, where, in the dead
hour of night, the boatmen are summoned by some unseen power
to launch their boats and ferry over to a sacred island the souls of
men who had been drowned in the surging waters. The fishermen
tell that, on the occasion of those midnight freights, the boat is so
crowded with invisible passengers as to sink quite low in the water,
and the wails and cries of the shipwrecked are heard as the
melancholy voyage progresses. On their arrival at the Island of Sein,
invisible beings are said to number the invisible passengers, and the
wondering awe-struck crew then return to await the next
supernatural summons to boat over the ghosts to the storied isle,
which was in long back days the chief haunt of the Druidesses in
Brittany. A similar story may be heard at Guildo on the same coast.
Small skiffs, phantom ones it is currently believed, may be seen
when the moon is bright darting out from under the castle cliffs,
manned by phantom figures, ferrying over the treacherous sands the
spirits whose bodies lie engulphed in the neighbourhood. Not one of
the native population, so strong is the dread of the scene, will pass
the spot after nightfall, and strange stories are told of phantom
lights and woeful demons that lure the unsuspecting wayfarer to a
treacherous death.
The Parisian fishwives are clean and buxom women, like their
sisters of Newhaven, and they are quite as celebrated if not so
picturesque in their costume. About a century and a half ago—and I
need not go further back—there were a great number of fishwives in
Paris, there being not less than 4000 oyster-women, who pursued
their business with much dexterity, and were able to cheat their
customers as well, if not better than any modern fishwife. One of
their best tricks was to swallow many of the finest oysters under the
pretence of their not being fresh. Among the Parisian fishwives of
the last century we are able to pick out Madame Picard, who was
famed for her poetical talent, and was personally known to many of
the eminent Frenchmen of the last century. Her poems were
collected and published in a little volume, and ultimately by marriage
this fishwife became a lady, having married a very wealthy silk
merchant. The fishwives of Paris have long been historical: they
have figured prominently in all the great events connected with the
history of that city. A deputation from these market-women,
gorgeously dressed in silk and lace, and bedecked with diamonds
and other precious stones, frequently took part in public affairs.
Mirabeau was a great favourite of the Parisian fishwives; at his death
they attended his funeral and wore mourning for him. These
Poissardes took an active part in the revolution of 1789, and did
deeds of horror and charity that one has a difficulty in reconciling. It
was no uncommon sight, for instance, to see the fishwives carrying
about on poles the heads of obnoxious persons who had been
murdered by the mob.
As I am on the subject of the foreign fisher-folk, I may as well say
a few words more about the quaint eel-breeders of Comacchio, to
whom I have already had occasion to allude. According to M. Coste,
the social life of the people at Comacchio, who are engaged in the
work of eel-culture, is very curious; but I think the industrial phase is
so much mixed up with the social as to render the two inseparable.
The community is in a sense—that is, so far as discipline is
concerned—a military one, and strict laws are laid down for the
conduct of the fishery. A large number of the men live in barracks,
and observe the monkish rule of passive obedience. Each of the
islands of the lagoon may be described as a small farm, having a
chief cultivator, a few servants, a plentiful supply of the necessary
implements of labour, its living-house, and its store for the harvest.
It appears so natural to the people to suppose these stations to be
farms, that they have from the very earliest times described the
various basins as fields, just as if they were composed of earth
instead of water; and of these places there are no less than four
hundred, the most important of them belonging to the state, the rest
being private property. The government of the whole lagoon is
exclusively in the hands of the farmer-general or his representative,
who rents the fisheries from the Pope. There is a large body of men
employed by him, who are divided into brigades, and whose
business lies in the construction of the dykes, and in the
management of the flood-gates during the seeding of the lagoon,
and the organisation of the labyrinths during the fishing-season. This
cultivating brigade numbers about three hundred men; the police
brigade consists of one hundred and twenty persons; and besides
these there is an administrative brigade of one hundred individuals.
A great deal of work has to be done by the persons employed,
whether at the various farms, in the offices, or in the kitchen, for at
Comacchio a large portion of the fish is cooked for the market. Upon
each farm there are about twelve labourers, who live in a barrack
under severe discipline, having all things in common. There is a
master who exercises absolute power in his own domain; he is paid
a salary of four scudi seventy-five baiocchi per month, with two and
a half pounds of fish per day, and during summer-time, when the
fish are scarce, he gets an additional allowance of money. The rate
of wages at this place appears exceedingly small when contrasted
with the payment of English labour. The wages of the learners or
apprentices are exceedingly modest; they are remunerated with the
“sair-won penny-fee” of 26s. per annum, in addition to their food!
But then the poor people of Comacchio—the widow, the orphan, the
aged and the infirm labourer—are all maintained at the expense of
the community.
But it is right to mention also that a greater than a mere salaried
interest in the labours incidental to the working of these fish-farms is
kept up by the greater portion of the employes having a share of or
commission on the produce, which in good years amounts to as
much as twelve Roman ecus for each man. The captain is, of course,
responsible in every way for his farm, both that the labour be
properly carried on, and also for the moral conduct of the men under
his charge, to whom he is bound to set a good example, as well of
neatness in dress as activity in business.
Exiled in the valley which they cultivate, each family finds it
necessary to devote its attention to those domestic offices so
necessary for economy and comfort. The vallanti take in turn, as our
soldiers do, the duty of cooking. They place the fish which they
receive as a part of their wages in a common stock, to which is
added such provision as the messenger may have brought from the
town. When the cook has prepared the repast, they all sit down to
table in one company, from the head man to the most humble
servant; but although they mix thus promiscuously together, military
etiquette is strictly observed—the foreman occupies the place of
honour, having the under-foreman and the secretary by his side,
next come the vallanti, and then the apprentices and cleaners. A
benediction is then pronounced, after which the foreman serves out
to each man his proper modicum of food, taking care to respect
those rights of precedence which have been indicated. Eels, cooked
upon the gridiron, form the staple of the repast, and the dinner is
washed down with a little bosco-eli-esco wine. After dinner is over,
the labourers return to their work. When evening arrives some
remain awake all night, seated in arm-chairs, and others lie down in
hard beds similar to those of the barracks. None of the employes of
the valley are allowed to be absent from duty without a written
permission, and heavy fines are exacted on any occasion of this rule
being infringed. The discipline of each valley is the same, and one
cannot conceive of a more monotonous life than that led by these
humble fishermen, which season after season is ever the same, and
goes on for years in one dull unvarying round. An unexpected tourist
excites quite a commotion among the simple people, and they have
great hopes that as the place becomes known to the outer world
their prison life will ultimately be ameliorated.
The fish season is opened with great solemnity of prayer, and
many of those other ceremonies of the church peculiar to Roman
Catholic communities—one of which is the consecration of the
lagoon. The labyrinths, which have been constructed from hurdles in
each watery field (see plan in “Fish Culture”) are crowded with fish,
so that there is comparatively little trouble in the capture, and the
salter waters of the sea being let in, the migratory instinct of the
animal is excited, so that it becomes an easy prey to the fishermen.
Upon the occasion of taking a great haul of fish in any particular
valley, a gun is fired to announce the glad tidings to the other
islanders, and next day a feast is held to celebrate the capture,
which must, however, be of a certain amount.
The town of Comacchio is chiefly a long street of one-storied
houses, situated on the principal island of the lagoon. There is a
cathedral in the town, but it is entirely destitute of any architectural
character, and there is a tower, from the top of which a good view of
the lagoon and its various islands may be obtained; but in an
industrial point of view the chief feature of the place is the great
kitchen where the cure of the fish is carried on, one of the
peculiarities of Comacchio being that a large portion of the eels are
cooked before being sent to market. The kitchen where the eels are
cooked is a large room containing a number of fireplaces ranged
along one side. These fireplaces are about five feet square, and in
front of each of them are hung six or seven spits on which the eels
are impaled and roasted. The fire is placed on a low grate, and
immediately below the spits is a trough or duct to catch the grease,
that drops from the eels while cooking. Before being roasted the fish
undergo an operation. A workman seated before a block of wood,
with a small hatchet in his hand, seizes the eels one by one and with
great dexterity cuts off the head and tail, which are given to the
poor, divides the body of the eel into several pieces of equal length
according to its size, and throws them into a basket at his side. Each
piece at the same time is slightly notched to facilitate the work of
the next operator, who with equal skill and quickness puts the bits
on the spit. It is only the large eels, however, that are decapitated
and divided, the smaller ones are simply notched and stuck on the
spit. The spits thus filled are next handed to the women in front of
the fire. Two women are necessary for each fireplace: one regulates
the fire; the second looks after the roasting of the eels, which is the
most important part of the labour, carefully shifting the spits from a
higher to a lower position in front of the fire until the fish are
properly done, when the spits are taken off by the woman, who
places them aside for the next operation. This woman also attends
to the grease that collects in the trough below the spits, and puts it
in jars for future use. Besides these fireplaces, there are a number of