Algorithmics of Nonuniformity: Tools

and Paradigms Micha Hofri

Visit to download the full and correct content document:
https://textbookfull.com/product/algorithmics-of-nonuniformity-tools-and-paradigms-mi
cha-hofri/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Internet of Things Principles and Paradigms 1st Edition

Rajkumar Buyya

https://textbookfull.com/product/internet-of-things-principles-
and-paradigms-1st-edition-rajkumar-buyya/

Coulson and Richardson’s Chemical Engineering, Fourth

Edition: Volume 3A: Chemical and Biochemical Reactors
and Reaction Engineering R. Ravi

https://textbookfull.com/product/coulson-and-richardsons-
chemical-engineering-fourth-edition-volume-3a-chemical-and-
biochemical-reactors-and-reaction-engineering-r-ravi/

Paradigms of international human rights law 1st Edition

Fellmeth

https://textbookfull.com/product/paradigms-of-international-
human-rights-law-1st-edition-fellmeth/

Essays on Supply Chain Management in Emerging Markets

1st Edition Micha Hirschinger (Auth.)

https://textbookfull.com/product/essays-on-supply-chain-
management-in-emerging-markets-1st-edition-micha-hirschinger-
auth/
Essentials of Microservices Architecture Paradigms
Applications and Techniques 1st Edition Chellammal
Surianarayanan (Author)

https://textbookfull.com/product/essentials-of-microservices-
architecture-paradigms-applications-and-techniques-1st-edition-
chellammal-surianarayanan-author/

High Performance Python Practical Performant

Programming for Humans 2nd Edition Micha Gorelick Ian
Ozsvald

https://textbookfull.com/product/high-performance-python-
practical-performant-programming-for-humans-2nd-edition-micha-
gorelick-ian-ozsvald/

Big Data Principles and Paradigms 1st Edition Rajkumar

Buyya

https://textbookfull.com/product/big-data-principles-and-
paradigms-1st-edition-rajkumar-buyya/

Globalisation Ideology and Education Reforms Emerging

Paradigms Joseph Zajda

https://textbookfull.com/product/globalisation-ideology-and-
education-reforms-emerging-paradigms-joseph-zajda/

Paradigms in Green Chemistry and Technology 1st Edition

Angelo Albini

https://textbookfull.com/product/paradigms-in-green-chemistry-
and-technology-1st-edition-angelo-albini/
Algorithmics of
Nonuniformity
Tools and Paradigms
Discrete Mathematics and Its Applications

Series Editors
Miklos Bona
Donald L. Kreher
Douglas West
Patrice Ossona de Mendez

Combinatorics of Compositions and Words

Silvia Heubach and Toufik Mansour

Handbook of Linear Algebra, Second Edition

Leslie Hogben

Combinatorics, Second Edition

Nicholas A. Loehr

Handbook of Discrete and Computational Geometry, Third Edition

C. Toth, Jacob E. Goodman and Joseph O’Rourke

Handbook of Discrete and Combinatorial Mathematics, Second Edition

Kenneth H. Rosen

Crossing Numbers of Graphs

Marcus Schaefer

Graph Searching Games and Probabilistic Methods

Anthony Bonato and Paweł Prałat

Handbook of Geometric Constraint Systems Principles

Meera Sitharam, Audrey St. John, and Jessica Sidman,

Additive Combinatorics
Béla Bajnok

Algorithmics of Nonuniformity: Tools and Paradigms

Micha Hofri and Hosam Mahmoud

For more information about this series, please visit:

https://www.crcpress.com/Discrete-Mathematics-and-Its-Applications/book-series/CHDISMTH
APP?page=1&order=dtitle&size=12&view=list&status=published,forthcoming
Algorithmics of
Nonuniformity
Tools and Paradigms

Micha Hofri
Worcester Polytechnic Institute

Hosam Mahmoud
George Washington University
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742

© 2019 by Taylor & Francis Group, LLC

CRC Press is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S. Government works

Printed on acid-free paper

Version Date: 20180622

International Standard Book Number-13: 978-1-4987-5071-4 (Hardback)

This book contains information obtained from authentic and highly regarded sources. Reasonable
efforts have been made to publish reliable data and information, but the author and publisher cannot
assume responsibility for the validity of all materials or the consequences of their use. The authors and
publishers have attempted to trace the copyright holders of all material reproduced in this publication
and apologize to copyright holders if permission to publish in this form has not been obtained. If any
copyright material has not been acknowledged please write and let us know so we may rectify in any
future reprint.

Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced,
transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or
hereafter invented, including photocopying, microfilming, and recording, or in any information
storage or retrieval system, without written permission from the publishers.

For permission to photocopy or use material electronically from this work, please access
www.copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc.
(CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization
that provides licenses and registration for a variety of users. For organizations that have been granted
a photocopy license by the CCC, a separate system of payment has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and
are used only for identification and explanation without intent to infringe.

Library of Congress Cataloging-in-Publication Data

Names: Hofri, Micha, author. | Mahmoud, Hosam M. (Hosam Mahmoud), 1954-

author.
Title: Algorithmics of nonuniformity : tools and paradigms / Micha Hofri and
Hosam Mahmoud.
Description: Boca Raton : CRC Press, Taylor & Francis Group, 2018. | Includes
bibliographical references and index.
Identifiers: LCCN 2018006956 | ISBN 9781498750714
Subjects: LCSH: Combinatorial analysis. | Computer algorithms. | Data
structures (Computer science) | Probabilities
Classification: LCC QA164 .H6427 2018 | DDC 511/.6--dc23
LC record available at https://lccn.loc.gov/2018006956

Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
To Ellen and Fari
Contents

Preface xiii

List of Symbols and Notation xvii

1 Introduction 1
1.1 Computing machines and models . . . . . . . . . . . . . . . . . . 1
1.1.1 Computer elements . . . . . . . . . . . . . . . . . . . . . . 1
1.1.2 Turing machines . . . . . . . . . . . . . . . . . . . . . . . 3
1.1.3 Pseudocode . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2 Asymptotic notation . . . . . . . . . . . . . . . . . . . . . . . . . 7

2 Counting 13
2.1 Generating functions . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.1.1 Multivariate generating functions and special numbers . . . 17
2.1.2 The principle of inclusion and exclusion . . . . . . . . . . . 20
2.2 Stirling numbers: Combinatorial interpretation . . . . . . . . . . . 24
2.2.1 Stirling numbers of the first kind . . . . . . . . . . . . . . . 25
2.2.2 Stirling numbers of the second kind . . . . . . . . . . . . . 26
2.2.3 Stirling numbers and powers . . . . . . . . . . . . . . . . . 27
2.3 Expansion of generating functions . . . . . . . . . . . . . . . . . . 29
2.3.1 Direct expansion of functions . . . . . . . . . . . . . . . . 29
2.3.2 Lagrange inversion theorem . . . . . . . . . . . . . . . . . 34
2.4 Generating functions in probability . . . . . . . . . . . . . . . . . 36
2.4.1 Convolution of random variables . . . . . . . . . . . . . . . 38
2.5 Generating functions in the solution of recurrences . . . . . . . . . 40

vii
viii Contents

2.6 Notes and sources . . . . . . . . . . . . . . . . . . . . . . . . . . 44

3 Symbolic Calculus 49
3.1 Admissible operations . . . . . . . . . . . . . . . . . . . . . . . . 51
3.1.1 The sum and product rules . . . . . . . . . . . . . . . . . . 51
3.1.2 Labeled combinatorial operations . . . . . . . . . . . . . . 56
3.2 Applications of the symbolic calculus . . . . . . . . . . . . . . . . 60
3.2.1 Compositions of integers . . . . . . . . . . . . . . . . . . . 60
3.2.2 Positional tree counting . . . . . . . . . . . . . . . . . . . 64
3.2.3 Plane tree counting . . . . . . . . . . . . . . . . . . . . . . 65
3.2.4 Rooted oriented trees . . . . . . . . . . . . . . . . . . . . . 66
3.3 Notes and sources . . . . . . . . . . . . . . . . . . . . . . . . . . 69

4 Languages and Their Generating Functions 71

4.1 Regular languages . . . . . . . . . . . . . . . . . . . . . . . . . . 74
4.2 Finite-state automata . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.3 Finite-state automata and regular languages . . . . . . . . . . . . . 80
4.4 Generating functions and their regular languages . . . . . . . . . . 85
4.4.1 Word equations . . . . . . . . . . . . . . . . . . . . . . . . 85
4.4.2 Solutions to word equations . . . . . . . . . . . . . . . . . 86
4.5 Counting regular languages . . . . . . . . . . . . . . . . . . . . . 88
4.5.1 A matricial alternative to word equations . . . . . . . . . . 93
4.5.2 Admissibility considerations . . . . . . . . . . . . . . . . . 97
4.6 Notes and sources . . . . . . . . . . . . . . . . . . . . . . . . . . 102

5 Probability in Algorithmics 105

5.1 Random variables . . . . . . . . . . . . . . . . . . . . . . . . . . 108
5.1.1 Independence of discrete random variables . . . . . . . . . 110
5.1.2 Probability spaces for sequences of random variables arising
in combinatorial objects . . . . . . . . . . . . . . . . . . . 112
5.1.3 Illustration via runs . . . . . . . . . . . . . . . . . . . . . . 115
5.2 Characteristic functions . . . . . . . . . . . . . . . . . . . . . . . 118
Contents ix

5.3 Mixed distributions . . . . . . . . . . . . . . . . . . . . . . . . . . 120

5.4 Inequalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
5.4.1 Boole inequality . . . . . . . . . . . . . . . . . . . . . . . 123
5.4.2 Chebyshev inequality . . . . . . . . . . . . . . . . . . . . . 125
5.4.3 Markov inequality . . . . . . . . . . . . . . . . . . . . . . 126
5.4.4 Gauss inequality . . . . . . . . . . . . . . . . . . . . . . . 127
5.4.5 Schwarz inequality . . . . . . . . . . . . . . . . . . . . . . 129
5.5 Modes of probabilistic convergence . . . . . . . . . . . . . . . . . 131
5.6 Some classic results from probability theory . . . . . . . . . . . . . 138
5.6.1 Weak and strong laws . . . . . . . . . . . . . . . . . . . . 139
5.6.2 Further convergence theorems . . . . . . . . . . . . . . . . 143
5.7 Central limit theorems . . . . . . . . . . . . . . . . . . . . . . . . 144
5.8 Martingales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
5.9 Generating random numbers . . . . . . . . . . . . . . . . . . . . . 150
5.9.1 The probability integral transform . . . . . . . . . . . . . . 152
5.10 Notes and sources . . . . . . . . . . . . . . . . . . . . . . . . . . 154

6 Functional Transforms 161

6.1 Mellin transform . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
6.1.1 Properties of the Mellin transform . . . . . . . . . . . . . . 161
6.1.2 Harmonic sums . . . . . . . . . . . . . . . . . . . . . . . . 167
6.2 Poissonization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
6.2.1 Algebraic depoissonization—uniform distribution . . . . . . 175
6.2.2 Algebraic depoissonization—arbitrary distributions . . . . . 178
6.2.3 Asymptotics of the Poisson transform . . . . . . . . . . . . 181
6.3 Notes and sources . . . . . . . . . . . . . . . . . . . . . . . . . . 182

7 Nonuniform Pólya Urn Schemes 187

7.1 Classic Pólya urns . . . . . . . . . . . . . . . . . . . . . . . . . . 187
7.2 Tenability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
7.3 Pólya urns with ball activity . . . . . . . . . . . . . . . . . . . . . 190
x Contents

7.3.1 Pólya-Eggenberger urn with ball activity . . . . . . . . . . 191

7.3.2 Ehrenfest urn with ball activity . . . . . . . . . . . . . . . . 194
7.3.3 Bagchi-Pal urn schemes with ball activity . . . . . . . . . . 198
7.3.4 Triangular urns with ball activity . . . . . . . . . . . . . . . 209
7.4 A nonuniform Pólya process . . . . . . . . . . . . . . . . . . . . . 214
7.5 Notes and sources . . . . . . . . . . . . . . . . . . . . . . . . . . 222

8 Nonuniform Data Models 227

8.1 Restricted permutations . . . . . . . . . . . . . . . . . . . . . . . 227
8.1.1 The combinatorics of 1-away permutations . . . . . . . . . 229
8.1.2 Properties of 1-away permutations via recurrences . . . . . 235
8.2 Automata for restricted permutations . . . . . . . . . . . . . . . . 239
8.2.1 1-away permutations . . . . . . . . . . . . . . . . . . . . . 239
8.2.2 2-away permutations . . . . . . . . . . . . . . . . . . . . . 241
8.3 Random multisets . . . . . . . . . . . . . . . . . . . . . . . . . . 244
8.3.1 Inversions in random multisets . . . . . . . . . . . . . . . . 246
8.3.2 Multinomially generated multisets . . . . . . . . . . . . . . 257
8.4 Binary search trees . . . . . . . . . . . . . . . . . . . . . . . . . . 259
8.4.1 Optimal binary search trees . . . . . . . . . . . . . . . . . 262
8.4.2 Bounds on the (optimal) access cost . . . . . . . . . . . . . 265
8.4.3 Nearly optimal binary search trees . . . . . . . . . . . . . . 269
8.4.4 Binary search trees—unknown p . . . . . . . . . . . . . . . 273
8.5 Digital trees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
8.5.1 The Bernoulli model . . . . . . . . . . . . . . . . . . . . . 275
8.5.2 Depth of nodes in a trie . . . . . . . . . . . . . . . . . . . . 276
8.5.3 Clades . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
8.6 Notes and sources . . . . . . . . . . . . . . . . . . . . . . . . . . 287

9 Sorting Nonuniform Data 295

9.1 Data comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
9.2 Insertion sort . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
Contents xi

9.2.1 Linear insertion sort . . . . . . . . . . . . . . . . . . . . . 298

9.2.2 Inversions under the uniform random permutation model . . 299
9.2.3 Performance on a slightly perturbed input . . . . . . . . . . 301
9.2.4 Sorting a partially sorted file . . . . . . . . . . . . . . . . . 302
9.2.5 Insertion sort for multisets . . . . . . . . . . . . . . . . . . 305
9.3 Quick sort . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
9.3.1 Three-way partition . . . . . . . . . . . . . . . . . . . . . . 308
9.3.2 Analysis of Quick Sort for random multisets . . . . . . . . 310
9.4 Notes and sources . . . . . . . . . . . . . . . . . . . . . . . . . . 318

10 Recursive Trees 321

10.1 Uniform recursive trees . . . . . . . . . . . . . . . . . . . . . . . 323
10.1.1 Outdegrees in uniform recursive trees . . . . . . . . . . . . 323
10.1.2 Depth of nodes in a uniform recursive tree . . . . . . . . . . 325
10.1.3 Leaves in uniform recursive trees . . . . . . . . . . . . . . 326
10.2 Trees with vertex affinity proportional to age . . . . . . . . . . . . 329
10.2.1 Degree profile in age-affinity random recursive trees . . . . 330
10.2.2 Depth of nodes in an age-affinity random recursive tree . . . 331
10.2.3 Leaves in age-affinity random recursive trees . . . . . . . . 332
10.3 Recursive trees grown under the power of choice . . . . . . . . . . 334
10.3.1 Degree profile of k–minimal-label recursive trees . . . . . . 335
10.3.2 Depth of nodes in k-minimum-label tree models . . . . . . . 338
10.3.3 Maximal-label recursive tree . . . . . . . . . . . . . . . . . 341
10.4 Preferential attachment tree model . . . . . . . . . . . . . . . . . . 342
10.4.1 Leaves in a random PORT . . . . . . . . . . . . . . . . . . 344
10.4.2 Depth of nodes in a random PORT . . . . . . . . . . . . . . 345
10.5 Blocks trees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
10.5.1 Building trees from random tree blocks . . . . . . . . . . . 349
10.5.2 Leaves in a blocks tree . . . . . . . . . . . . . . . . . . . . 351
10.5.3 Depth of nodes in blocks trees . . . . . . . . . . . . . . . . 352
xii Contents

10.5.4 The height of a random blocks tree . . . . . . . . . . . . . . 357

10.6 Hoppe trees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360
10.6.1 The number of species . . . . . . . . . . . . . . . . . . . . 361
10.6.2 Sizes of species populations . . . . . . . . . . . . . . . . . 364
10.7 Notes and sources . . . . . . . . . . . . . . . . . . . . . . . . . . 365

11 Series-Parallel Graphs 369

11.1 Some models of binary series-parallel graphs . . . . . . . . . . . . 371
11.2 Enumerating binary series-parallel graphs . . . . . . . . . . . . . . 374
11.3 The order of binary series-parallel graphs . . . . . . . . . . . . . . 377
11.3.1 The order of factorial binary series-parallel graphs . . . . . 377
11.3.2 The order of Catalan binary series-parallel graphs . . . . . . 379
11.4 Path length in binary series-parallel graphs . . . . . . . . . . . . . 381
11.4.1 Path length under the factorial model . . . . . . . . . . . . 381
11.4.2 Path length under the Catalan model . . . . . . . . . . . . . 385
11.5 A series-parallel graph with unrestricted degrees . . . . . . . . . . 391
11.5.1 Nodes of small outdegree . . . . . . . . . . . . . . . . . . . 392
11.6 Notes and sources . . . . . . . . . . . . . . . . . . . . . . . . . . 399

Bibliography 401

Solutions 418

Index 559
Preface

“What is the use of a book,” thought Alice, “without pictures or conversations?”

–Lewis Carroll

This is a book about the analysis of algorithms, and the data structures that sup-
port them. With all due respect to Alice, we have few pictures in the book, but tried
to keep its tone conversational, to the extent possible. We look at algorithms that
are designed to handle general data—sort any array, find the median of any numer-
ical set, identify patterns in any setting—but their analysis does not usually aspire
to such generality; we want to evaluate “average” performance, “typical” behavior,
or in mathematical terms, the expectations—and when possible, the variances and
ultimately the distributions—of the random variables that describe their operations.
Such analyses are usually conducted under the benign assumptions of uniformity,
which evidently appear natural to us. Very little has been done under nonuniform
assumptions on the input to algorithms. Most existing analyses assume that when
searching, the sought key is equally likely to be in any location; when sorting, that
the input is a random permutation of the content, and so forth. The genesis of this
book was our desire to embark on an adventure to explore the issues that arise when
we leave the comfort zone of uniformity. Indeed, like Alice, we drop into the rabbit
hole, where little looks familiar.
As the reader would infer from the foregoing, we do not deal at all with the analysis
of deterministic performance bounds, the so-called worst-case behavior, leading to
complexity theory. We rather analyze particular data structures and algorithms, with
a focus on aspects of nonuniformity.
Many exercises are presented. We think of them as essential; they convey additional
material complementing the content of the chapters. For this reason, the solutions
aim to be more than mere answers, but to explain, and expand, where it seems ap-
propriate, on the related concepts, and motivate further work by the reader. Thus, the
solutions should be seen as an integral part of the main text. The level of difficulty of
an exercise is expressed on a scale of 1–5, with 1 being easiest and 5 being hardest.
Thus, an exercise at level h1i is a straightforward application of the material, and dif-

xiii
xiv Preface

ficulty level h2i would fall in the category of “doable with some effort and attention.”
An exercise of level h5i is one for which we do not have a satisfactory solution. Do
not assume much precision was aimed for at this ranking (or is indeed possible); we
believe that rating the difficulty of a problem is highly subjective.
While the first half of the book contains much introductory material, there is no
way for it to be entirely self-contained. Our assumptions about the background of
the reader are that it includes exposure to programming, data structures, and basic
algorithms and their analysis at the undergraduate level. We also assume, at the same
level, the elements of discrete mathematics and probability, its basic concepts and
standard distributions. The central place of probability in our discussion led us to
include a more complete development of the topics we need, in Chapter 5.
A reader with a first degree in computer science, mathematics, or engineering has
adequate preparation, and falls within our expected audience. We think the book can
be used as a general reference, or as additional reading in the graduate studies in
these disciplines. Certain chapters may serve as a source for projects pursued by
undergraduate students.
Every chapter ends with a section titled “Notes and sources,” where we tell about
sources for the materials presented in the chapter, and sometimes develop back-
ground material that we use in the chapter. Whenever appropriate, we included sug-
gestions for further study of the topics we discuss.
Most of the mathematical symbols are displayed and described in a list beginning on
page xvii. We have adopted some additional conventions:
• Vectors and matrices are printed using boldface letters.
• Random variables are denoted by capital letters, and their realized values are de-
noted by the same letters in lowercase.
• Sequences and their generating functions are, whenever possible, denoted by the
same letter, such as {an } and a(x); likewise for random variables and their trans-
forms.

The book is structured in chapters, mostly falling into two main categories: funda-
mentals (tools) and applications (paradigms).
• Chapter 1 is a general introduction about some basic notions of computing, associ-
ated jargon, and notation meant to put the entire book in perspective. Chapters 2–6
fall in the fundamentals category. These fundamentals include the following tools:
• Chapter 2: Methods of counting and enumeration—generating functions and spe-
cial numbers.
• Chapter 3: Symbolic calculus, where the focus is on abstract combinatorial tools,
like sequences and compositions.
Preface xv

• Chapter 4: Automata and regular expressions constructed for various types of

combinatorics on words, using the abstraction of recognition of languages.
• Chapter 5: Probability. We take a view often ignored in the study of combina-
torics, focusing on underlying sample spaces and treating the analysis of algo-
rithms beyond averages, leading to discussions of central limit theorems and weak
and strong laws of large numbers via measure-theoretic themes.
• Chapter 6: Functional transforms. This short chapter contains some modern tools
based on mathematical transforms, where one switches the view to a domain in
which the problem structure is more transparent. The desired results in the original
domain are then obtained by an inverse transform.
• Chapter 7 falls between tools and paradigms. Standard probability theory is uti-
lized to outline a theory for Pólya urn models, which are models for populating
urns with colored balls according to rules associated with ball drawing. The focus
is on some urn schemes that deviate from uniformity. Under uniform models, balls
in the urn are equally likely. The models we consider associate a different “activ-
ity” level for balls of each color. These models have proved to be a versatile tool
for various studies in random graphs, particularly those focused on the number of
nodes of small degrees in the structure.
The rest of the chapters use these tools in a variety of algorithmic applications:
• Chapter 8: Data models, such as nonuniform permutations, multisets, and trees.
• Chapter 9: Sorting, where the analysis is conducted under nonuniform data mod-
els.
• Chapter 10: Recursive trees, under various nonuniform evolution models.
• Chapter 11: Series-parallel graphs, a type of data structure, where several of the
techniques displayed in the previous chapters are put to work.

It is important to put in perspective one purpose of our book. Traditionally, algorith-

mics have been approached either via a probabilistic view or an analytic approach.
Our coverage espouses the two approaches and brings them together hand-in-hand to
get the best of both worlds and benefit from the advantage of each approach, when it
genuinely has something better to offer.
Many of the calculations shown in the book are best done with the help of a computer
algebra system, and some are practically impossible otherwise. We urge the reader
to master the use of one, such as MAPLE R , Mathematica R , or MATLAB R . To those
who have not used such a system, our enthusiastic advice would be to take the jump,
and find high-order bliss! Know that the learning curve is both steep and long, but
the advantages such a tool bestows are commensurate with the effort of mastering it.
We are using an unconventional method of referencing tables and diagrams. Both
xvi Preface

components are referenced by the page number on which they appear. As further
aid in orientation, several of the text components—exercises, definitions, theorems,
and the like—are all numbered consecutively within each chapter. For example, the
Lyapunov theorem in Chapter 5 is listed as Theorem 5.79 and the following exercise
is labeled Exercise 5.80, and so forth.

ACKNOWLEDGMENTS
Our thanks go to the editors, Miklós Bóna and Robert Ross, who gave us the op-
portunity to prepare this book. Also, we are indebted to Shashi Kumar of Cenveo
Publisher Services in Noida, India, for his prompt and attentive help with LATEX. The
exposition of the book was improved via the careful scrutiny of the native eyes of Re-
becca Condit and the sound advice of the project editor, Paul Boyd. They provided
excellent editorial advice.
Micha Hofri wishes to acknowledge the main influences on his contributions to this
book: Alan Konheim showed him how mere equations capture the life and complex
evolution of data structures; the late Philippe Flajolet made him see that generating
functions are not only useful, but magic as well.
The debt to his wife, Ellen, who has an answer to any question, and an apt observation
whenever one is called for, cannot be computed: It is too large for our tools.
The second author adds his appreciation and gratitude for Philippe Flajolet. He
also acknowledges the deep influence Robert Smythe left on him. The second au-
thor wishes to acknowledge Worthy Friends for a lifetime of friendship and support
and his family for encouragement. In particular, he thanks Farideh Mahmoud, Mona
Mahmoud, Sherif, Kareem, and Mai Hafez, and Moshira Mahmoud.
We finally acknowledge, gratefully, the enormous help we have had in finding, com-
puting, and presenting much of the results we report in this volume—from software.
The MAPLE computer algebra system has done so much more than mere algebra, and
carried us through some hair-raising calculations. The entire book has been typeset
by the authors using the LATEX document-preparation system, aided by a score or two
of “packages,” created and maintained by the large, international, and industrious
TEX community.
8 Introduction

In this sense, f (n) = 2n + 7 sinn2 ∼ g(n) = 2n.

Sometimes we cannot find such a close approximation, and we need to make do with
a bound. We present a few of those. The most common is the “big Oh” notation:



f (n) ∈ O g(n) , some writers prefer the less precise notation f (n) = O g(n) .

This holds, when a natural number n0 and a positive constant C exist, such that
| f (n)| 6 C |g(n)|, for all n > n0 . The reason we treat the symbol ∈ as more precise
here is that the above relation is not a true equation—more on this below—and ex-
presses the statement “ f (n) belongs to a collection of functions, all of which satisfy
this relation.” The big Oh notation tells us that ultimately (a word we use in the sense
“for large enough n)” f (n) is bounded by some multiple of g(n). An alternative for-
mulation is that f (n) increases no faster than g(n) does. However, this notation says
nothing about the value of C (which may be tiny or enormous), and it leaves open the
√
possibility that g(n) is a poor bound. Thus, it is true that n = n1/2 ∈ O(n2 ), but this
is a bound with little merit. Sometimes, we can do no better. Indeed, several complex
algorithms carry a history of successive improvements of the bounds of their cost
function, and some of these bounds differ very much!
Note: Some scientists use the big Oh notation in the sense of “the exact order of”; we
do not do this, since we have other symbols for this purpose, Θ, which we introduce
below, which provides the exact rate of growth, and ∼, which gives the exact value
of the leading term.
There is another notation that carries a similar but stronger version of the message—
f (n) increases more slowly than g(n)—the “little oh” symbol:


f (n)
f (n) ∈ o g(n) or f (n) = o g(n) , if → 0, as n → ∞. (1.1)
g(n)

The little oh makes a stronger claim than its big brother, and turns out to be somewhat
less useful.
Exercise 1.7 h1i Justify the assertion that the little oh notation makes a stronger claim
than big Oh, by showing that f (n) ∈ o(g(n)) =⇒ f (n) ∈ O(g(n)), but there are
function pairs ( f (n), g(n)) such that f (n) ∈ O(g(n)) and f (n) ∈ / o(g(n)). ♦

For the converse statement, that f (n) is ultimately bounded from below by g(n), or
that it increases at least as fast as g(n), we have the omega (Ω) notation, which also
requires the existence of a natural number n0 and a constant K, such that:



f (n) ∈ Ω g(n) , or f (n) = Ω (g(n) , if | f (n)| > K |g(n)|,
∀n > n0 .
(1.2)
You may well ask: What is the point of this additional notation, since it is clear, that
Another random document with
no related content on Scribd:
of moustache) comme ca. Also, I have thick hair which stands up
(illustration) comme ci. I had my head shaved last November, you
know, and it had a good effect. My letter is thrilling, isn’t it, with all
these martial details? A veritable cross-section of battle, a flower from
No Man’s Land. Well, I am sending some battling picture-postcards
soon which you will find amusing, I think—portraits, in two striking
histrionic poses, of myself. You know now, from my previous letters,
that I am no longer (I thank God!) doing statistics, so gently, but widely
and most firmly, correct the statement that I have a bullet-proof job. I
had one, but succeeded, after two months’ intriguing, in getting rid of it.
It wasn’t a shell-proof job—nothing in a real regiment is. (Such jobs are
in the Ordnance or Supply Departments, back of the lines.) It wasn’t
shell-proof, I say, and if I should be squashed by a shell, wouldn’t you
hate to have it said that I was nobly holding my post in the office, or
bravely manning my typewriter? Now, I’m doing work I love—and work
you may be proud of. None of the drudgery of soldiering, but a double
share of glory and thrills. But it is not so glorious and thrilling as you.
Joyce.

* * * * *

May 24, 1918.

Dear Aline:
I don’t know just what kind of a mental picture you have of my
present life—to be accurate it would need to be rather kaleidoscopic,
for my activities and the spheres of them vary frequently and greatly.
But this afternoon you (or perhaps not you, but readers of Empey and
Private Peat and the rest of that mob of war writers—thank God, let me
Pharisaicly say, that I am not one of them—would do so) would have
been amazed at the martial picture before you, had you seen me. Just
outside the edge of the forest of firs and spruce in which we fourteen
men live is a lovely meadow. There, among the knee-high buttercups,
lay in the May sunshine all afternoon three warriors—myself being
one. Whiles we smoked and gazed at the lovely valley miles below us
—whiles we took turns in reading aloud from—what do you suppose?
The Oxford Book of English Verse! We read Gray’s Elegy, the first
chorus from Atalanta in Calydon, “They told me, Heraclitus,” that witch
poem of William Bell Scott, “Love in the Valley,” “Lake Isle of Innisfree,”
“Keith of Ravelston,” and half a dozen other poems, all of which
brought you most poignantly and beautifully before me. No, they didn’t
bring you before me—you are always before me and with me and in
my heart and brain—but it’s dangerous to write this—it draws so tight
the cords that bind me to you that they cut painfully into my flesh. Well,
we are to be together sometime, inevitably, and soon in terms of
eternity. For we are absolutely one, incomplete apart, and in Heaven is
completeness. How unhappy must lovers be who have not the
gracious gift of faith!
There is to be an Homeric banquet at our house one day—the day
when I exhibit to my comrades the glory of my life—yourself. You will
like them all—Watson (a gifted artist from Richmond, who is now at
work on a fine drawing which must accompany Rouge Bouquet in
Scribner’s), Bob Lee, Titterton (my especial friend), Beck, Mott,
Kerrigan, Levinson—say a prayer for them all, they’re brave men and
good, and splendid company. They are all men of education, and
breeding, and humour, and we have fine times. Dangers shared
together and hardships mutually borne develops in us a sort of
friendship I never knew in civilian life, a friendship clean of jealousy
and gossip and envy and suspicion—a fine hearty roaring mirthful sort
of thing, like an open fire of whole pine-trees in a giant’s castle, or a
truly timed bombardment with eight-inch guns. I don’t know that this
last figure will strike you as being particularly happy, but it would if
you’d had the delight of going to war.
Right here I had to pause to discuss theology with Watson,
Titterton, Levinson and Jongberg. Jongberg is up here for a brief spell;
he is a Swedish-Irishman, and now he is posing for St. Michael for
Watson, using a bayonet for a sword. Levinson (in full uniform,
including belt and helmet) is being model for the soldiers leaving
Rouge Bouquet for Heaven. He is a quaint little French-Jewish-
American, with whom we have a lot of fun. They all trooped into the
room where Titterton and I sit writing to our respective sweethearts, to
ask about the style of St. Michael’s sword and that of his halo. These
questions settled, Watson became enamoured of the idea of angels
saluting, and devised a whole manual of arms for angels—as “Angels,
attention! Wings raised, by the numbers. 1 up! 2 down! Wings flap!
Hey you down there! what’s the matter with you? Don’t you know
enough to keep your hands down when you flap your wings? Awkward
squad for you to-morrow!” Then they went back to their work.
Well, I don’t know how long the Intelligence Section is to stay in
this lovely summer resort of a place, but it’s certainly fine while it lasts.
And I am very fortunate to have such a crowd as this to work with. I
never was with so congenial a company in civilian life. Fr. O’Donnell
came up to see us yesterday—he is stationed not far away. Fr. Duffy
has an assistant—Fr. Kennedy—a Jesuit, thank God! I have not met
him yet. He is a brother of the publishers of Barclay Street. I have a
new stripe—an inverted chevron of bright gold on the left cuff for six
months service. Let you be proud of that too, please, and let my
children be proud of it. That is all the news except that I love you.
Joyce.

* * * * *

Dear Aline:
To-day I was made very happy by receiving two nice little ridiculous
letters from you—one of which considered itself to be long. As to Sr.
Mary Leo—if you happen to be writing to her (you will be, for I told her
to ask you for “Rouge Bouquet”) ask her by all means to tell you how
her grandmother taught her geography. It’s an absolutely delightful
story, and she’s a delightful old saint.
You say Tad will go “in a few months.” I’m glad I’m not Tad. There
are many reasons for this, of course, but I mean I am especially glad
I’m not to go in a few months. I’m glad I’m a Sergeant in the 69th, a
volunteer regiment, the bravest and best regiment in the army. I’m glad
I have a golden service chevron on my left arm—that means six
months service in France, and is better than two bars on my shoulder,
which would mean only three months stay at Plattsburg. If I get an
honorable wound that will unfit me for soldiering but not for civilian
work, that civilian work will not be hack-writing. It may be straight
reporting, or editing, or writing (Evening World at $80 a week, for
instance) a weekly essay-column, or dramatic criticism (this last
appeals to me strongly). Also I’ll do less lecturing and more oratory—
bombastic, thundering, literary, Henry Woodfin Grady oratory! It is
exciting, and lecturing isn’t, especially.
Meanwhile I find myself approving your plan of getting literary
work, although doing it makes me feel like a coloured man approving
his wife’s plan of taking in washing. You relieve my mind greatly by
your willingness to keep your name off the book. I thought I would
acknowledge to you the pleasure I got from “A Wind Rose in the
Night.” It is a holy and dear poem.
Please see that Kenton learns to serve Mass, won’t you? Sorry to
keep teasing you about this, but you never write anything about it.
Your poem “To a Sick Child” is beautiful, and “cookies” is the
making of it. I love you.
Joyce.

* * * * *

May 18, 1918.

Dear Aline:
I didn’t get your letter expressing enthusiastic approval of “Rouge
Bouquet” until I had received two others just mentioning it. I certainly
am glad the poem “got” you, as it got me. The Regiment—and soldiers
outside the Regiment as high as a Brigadier General—are wonderfully
enthusiastic. The Brigadier General had twenty-four copies made to
send his friends. Of course we are especially moved by it because the
event is so close to us. But to appreciate it you should have heard taps
played and at a soldier’s funeral. P. C. ought to know the distinction
between peacemakers and pacifists. I wonder he didn’t include St
Michael in his catalogue of pacifists. We are peace-makers, we
soldiers of the 69th, we are risking our lives to bring back peace to the
simple, generous, gay, pious people of France, whom anyone
(knowing them as I have come to know them in the last six months)
must pity and admire and love. They are an invaded people—and
invaded people always are right. The careful study of certain ghostly
villages (I cannot forget them) has taught me where I would have stood
in the 60’s and that is on the side of the invaded Southern States. God
keep me from ever being ordered to take part in an invasion of a
peaceful land! So long as I am helping drive out the invader, I know I’m
right—and all questions of international politics are of no importance
whatever. Here are nice old ladies, fat babies, jovial humorous men,
and little girls just after making their First Communions. They’ve been
driven out of their pretty sleepy little villages. They want to get back
and mend the shell holes in the roof and go to school and take their
place drinking red wine of an evening according to their tastes and
ages. Well, we men of the 69th are helping to give these people back
their homes—and perhaps to prevent our homes from one day being
taken from us by the same Power—of whom nothing at all worse need
be said than that it is an invader. And St. Patrick and St. Bridgid and
St. Columkill and all the other Saints are with us—they are no more
pacifists than they are Roycrofters! And God keep us from Washington
Square! No Man’s Land is a damn sight healthier place—I have tried
both, and I know what I’m talking about.
And as for you, you young flippertigibbet, you won’t be middle-
aged, you say? Well, we’ll see about that! I’ll probably come back with
one eye, one arm, one leg and a horrible temper, and turn your hair
white in a day. Stop calling any child of ours “Billy.” When you referred
to “Billy” in previous letters I thought you meant some rather
unfortunately named dog or cat. Now I am annoyed to find out you
meant Michael. The habit of so designating him probably started
humorously—please stop it at once.
I have not read “The Tree of Heaven,” to which you refer in a
recent letter. But I have read Carnival—and if when I get back to the
States I find that you really think such books are “fine,” as you say in
your letter, I’ll grab a baby (probably young Michael) and start for
whatever South American Republic or other bellicose land has need of
the services of a ranting, roaring, slashing, dashing young Sergeant,
fresh from the ranks of the quarrelsome Irish, veteran of the Great War.
Joyce.
 * * * * *

May 27, 1918.

Dear Aline:
I enclose three bits of verse by various hands. “A Nun’s Prayer for
a Soldier” is by dear old Sister Mary Leo. The flattering sonnet was
sent to me by Sr. Mary Emerentia, having been copied from a paper
published by the girls of St. Mary’s College, Notre Dame, Indiana. I
destroyed Sr. Emerentia’s letter before I copied down the author’s
name, so I couldn’t write to thank her, unfortunately. “The Green
Estaminet” I want carefully preserved, for it must go in every anthology
made by you or me or anybody we know. I like it better than any war
poem I’ve read since I became a soldier, and I deeply regret that I
didn’t write it. Many and many a jovial evening I have spent in just such
bistrals as the poet describes. Dear old Madame—she always has that
“strange disease,” and she finds that white bread is good for it—and in
exchange for that white bread (from army kitchens) she gives the
delicious soft-crumbed brittle-crusted war-bread, better than any bread
in the States. “The great round songs begin”—they do, indeed! “Bing
Bang Bingen on the Rhine!” “Down in the Heart of the Gas House
District,” “Sidewalks of New York,” and always “Madelon,” in which the
French soldiers in the corner enthusiastically join. Cecil Chesterton told
me he always judged a country by its drinking-places. Well, here are
the merriest, bravest drinking places in the world. If the States go dry,
I’m going to bundle all you young critters over here to live—a
comfortable, humorous, Catholic country. Honestly, the only drawback
to living in this country is the absence of you and your gang. I’d like a
house on the very mountain top I now inhabit, and I know you’d enjoy it
tremendously. I’m perfectly serious about this, and look forward to
bringing you here.
Write oftener, and longer! I love you.
Joyce.
 * * * * *

June 1, 1918.
Dear Aline:
A terrible thing, this war, what with a pine forest to live in, all the
latest novels to read, and bridge every evening. And now I am
preparing for a short week-end in a certain nice small city, an hour’s
walk through the forest and across the mountains. It is a lovely forest
in which to wander of a June day—so deliberately European—a forest
whose tall evergreen trees and smooth brown floor suggest all the folk
tales I ever heard your dear voice beautify for our children’s delight.
Reaching this certain town, I have interesting things to do—to collect
my pay for one thing. To get mail—I hope from you. To go to
confession to Fr. Duffy or Fr. Kennedy, S. J., our new assistant
chaplain, or to some curé with thin hands and long white hair. To drink
a bottle of beer at the Sergeants’ Club—among my brother noncoms,
French, Italian and American, in horizon blue, green and olive drab
respectively. To mail to you this letter and a splendid drawing for
“Rouge Bouquet,” by Emmett Watson, of ours. To dine sumptuously on
steak and fried potatoes and Pinard at a merry bistral on a side-street,
where Madame la patronne will present me with a sprig of lilies of the
valley, which I (imagine a boutonierre with an American uniform!) will
present to the baby of the tailor’s daughter. To attend the movies—two
sous! To sleep between clean sheets in a hotel-bedroom—half a franc!
And to-morrow to be (gold service stripe flashing in the sunlight,
trench-stick jauntily swinging) a part of the many-coloured Sunday
afternoon procession along a certain broad avenue. Then back to my
forest-lodge for supper.
I have written a long topical poem about a hike which I’ll send you
soon.
I love you.
Joyce.
 * * * * *

A. E. F., France.
Dear Aline:
So far, I have told about six people about your rejoicing in the fact
that the possession of a kitten justifies you in purchasing a bird-cage.
An absolutely delightful character-revelation—that is, the revelation of
a delightful character. And the most genuine sample of you I have had
since last November.
I am writing this letter in the end of one of our barracks, which its
partition cuts to make an airy office for the Supply Sergeant of
Headquarters Company. You’d like him very much—a very beautiful
person of about my age, a playwright by profession, educated in Paris.
At the other end of the long table sits a bridge party, of which the
members are the Supply Sergeant (Lemist Esler is his name), Howard
Young (who was a Marine before he became a fighting Irisher),
Sergeant Kenneth Russell (a wealthy and charming business man of
very literary taste, who knows most of the living poets we know and
revels in the Oxford Book of English Verse), and Tom O’Kelly, who has
the most beautiful tenor voice you ever heard. He is a professional
singer from Dublin, with very tightly curled close-cropped hair, a
reddish brown face, and the build of a prize-fighter. In the lulls of the
game he sings snatches of “Digging for Gould” and other nice ballads.
Admirable person.
 “HE LIES BURIED, ON THE RIGHT,
BESIDE LIEUTENANT OLIVER AMES,
AT THE EDGE OF A LITTLE COPSE
KNOWN AS THE WOOD OF THE
BURNED BRIDGE, CLOSE TO THE
PURLING OURCQ”
I am with the Regiment now—the Intelligence Section is back from
its Post, the beautiful place from which I wrote you many letters. Now
we are in very flat country, the flattest I ever saw. Very violent flowers,
chiefly scarlet poppies and small blue things, cornflowers, I believe.
Gloriously hot, with good swimming nearby, and the best sort of
shower-bath, obtainable by merely sitting under the spout of a pump
while an obliging friend turns the handle.
Your letter of May twenty-sixth came after I had received a June
letter. I am delighted to hear of Kenton’s confirmation and of his choice
of a name—St. Stephen is a gallant gentleman and a true friend and
does wonderful things for me.
I’m not very enthusiastic about that group picture of office-workers
in which I appear—I like myself better as I am now, a soldier instead of
a clerk, visiting the office only occasionally—to work on my history. And
I’m ever so much happier now than when the picture was taken.
 Is Kenton serving Mass yet? Please have him do so. And please
send me my Anthology—I believe restrictions no longer keep
packages from soldiers. I love you.
Joyce.
MISCELLANEOUS PIECES
 A BALLAD OF NEW SINS

Prelude

I am sick of the riotous roses of rapture,

Of sibilant serpentine lips,
Of the wine cup
And murder
And all that mid-Victorian stuff.
I will sin large purple sins
American
And new.

I
THE SIN OF SOUTH BEND
Clandestinely, by night,
I will board a train
And be borne, palpitant and eager, over the shuddering rails,
Until
In the chaste stillness of Sunday morning
I am hurled into the blushing station of South Bend, Indiana.
Thence I will speed to the Hotel Mortimer
And, leering with hot lips,
Rubbing with tremendous hand my rough jaws
(For this is the shameful signal,
Immemorial, inevitable),
I will ask the way to the barber shop.
The liveried menials will blench,
But one of them, hardened in crime,
Will feverishly seize my proffered gold,
And lead me, on tiptoe,
To the barber-shop’s secret door.
There, in a dim back room,
Screened, impenetrable,
A barber (artfully disguised in a black coat and no chewing
gum)
Will (now and then peering nervously out beyond the screen,
fearful of spies and policemen)
Give to me, for twenty-five cents,
The forbidden delight
Of a Sunday morning shave.

II
THE WICKEDNESS OF WASHINGTON
What wickedness is more witchingly wonderful
Than the wickedness of Washington,
Where it is heinous to hock
And perilous to pawn?
I will go to Washington
And offer my watch to an usurer.
“Wait here,” he will whisper,
“And give me ten cents!”
Then will a messenger, winged like a swallow,
Carry my watch over meadow and hill
Over the riotous River Potomac
To the Virginian Shore;
He will return with a lavender ticket;
He will return with a five-dollar bill;
He will return (Oh, the logic of law!)
He will return with the watch
And the pawnbroker will lock it away.

III
THE IMMORALITY OF INDIANAPOLIS

I will go to Indianapolis,
And there sin strangely
By crossing the street catty-cornered
Instead of at the crossings,
And I shall be hanged.
 WAR SONGS

I—Water-colour
Pushing my way through the chattering throng of my brown-clad
mates to the rail of the troop-ship, I look at still water, greasy and
opaque. A touch of sunlight makes it splendid with rainbows, a great
prismatic expanse, beautiful, more beautiful than clear water could
be. Broken oars shatter the rainbow, bringing a black, clumsy
rowboat close to our ship’s side. Around the black boat the rainbow
settles. The rower rests his oars and lifts graceful entreating arms.
He wears pale blue overalls. In the stern of his boat is a little girl in a
cardinal cloak. On her head is one of the caps that make the French
sailors look so gay and gentle, a flat, round, blue thing with a red
pompom. She claps her hands when cooks lean through the
portholes and throw loaves of bread to her father.

II—Breakfast
I may breakfast in either of two ways. I may, as I pass a
steaming field-kitchen, hold out by its long handle a shining
aluminum basin. John Wilkert will put into it a big ladleful of rice, and
Leo Maher will pour golden syrup over it. Also, before I leave the line
I shall have three long strips of broiled bacon and two thick slices of
white bread, and a canteen cup full of hot, sweet coffee. The
breakfast room is a meadow or the roadside across from the
barracks. There is good company, hungry and mirthful. And over our
heads noisy battalions of crows maneuver, advancing, retreating,
hoarsely shouting down to us news of what awaits us beyond the
frozen hills.
Or I may go to the House by the Fountain. Pierre’s “permission”
is over, so he will not come in from the stable to smoke my tobacco
and tell me of life and death in the trenches. Grandpere sits by the
fire, now and then blowing it to flame by forcing his scant old breath
upon it through a long hollow tube, and toasting for me a thick slice
of war-bread. Madame superintends the heating of the big iron pot of
this morning’s milk and the three-legged pot of coffee. Now my bowl
—a little precious sugar in the bottom—is filled with hot milk.
Madame deftly pours black coffee into it, and it becomes richly
brown. I break my toast into it and eat eagerly—more eagerly than
does demure little Francine, who sits opposite me, her school books
beside her on the bench. She has large innocent brown eyes like her
father’s. Her hands are so tiny that I am surprised at her dexterity
with the large pewter spoon. I am afraid that if I stare at her I shall
embarrass her and make her spill cafe au lait on her immaculate
pinafore. On the great stone mantel over the fire are a spent
seventy-five millimetre shell curiously engraved with wreaths of
roses, a pink china pig and a brass crucifix. A bugle sounds by the
barracks. I give Madame her half-franc. I take my belt and rifle. “Bon
jour, Monsieur,” says Madame, “A demain!” Some day instead of “a
demain” they will say “au revoir.”
 “TRY A TIN TO-DAY”
“AND NOW,” said John Potts, whirling around on the piano-stool
and throwing his lank forelock back with a jerk of his head, “we will
interpret Grace Mallon’s soul!”
“Oh, wait till she comes, John!” boomed the Rev. Morris Gildell.
“She was at the Karl Marx Forum in our parish-house last night, and
she said positively she’d be down here this afternoon.”
“Yes, do wait!” said Mrs. Anna Watkins Wilbur, placing her fat
hands, covered with rings set with semiprecious stones, over those
of the pianist. “Do wait! It will be wonderful for her to hear it!”
The host of the occasion was Edwin Marmaduke, painter. He
was at present engaged in spooning Arabian incense out of a cocoa-
tin and putting it into a large bronze thurible, which hung before a
plaster replica of Rodin’s “Le Baiser.” The incense caught fire, and
dense clouds of fragrant smoke filled the little studio. Then
Marmaduke turned his pale but agreeable face to his guests.
“No, do it now!” he commanded. “Perhaps you and Arthura will
bring her with it.”
So John Potts struck a few minor chords, and Arthura Lewis
lifted her pale-green mantle in both thin arms, smiled at the low
ceiling, closed her eyes, and danced. She did not really dance. She
merely bent and kicked and gestured, approximately in rhythm with
the music.
She was not really Miss Arthura Lewis, either. Her first name,
which she had discarded as too usual, was Alice and her last name
was Potts, for she was the pianist’s wife. They were known to live
together in the little apartment across Patchin Place from
Marmaduke’s studio, but the fact of their marriage was scrupulously
kept a secret from the other members of their emancipated circle. No
liaison was ever hidden from the world more zealously than was the
regularity and mid-Victorian respectability of this seeming “free
comradeship.”
So John Potts played and Arthura danced, and the theme they
were interpreting was Grace Mallon’s soul. Edwin Marmaduke’s eyes
were turned toward Arthura, but he did not see her. Nor did he see
Grace’s soul. He saw her eyes—very gray and lovely—and her hair,
which was golden-brown and had an indefinable air of mirth about it.
Marmaduke loved Grace better than cigarettes, or incense, or
art; better even than the school of painting of which he was the
acknowledged master, the school which proudly accepted a name
first derisively given it—the Incomprehensiblists. So, while Arthura
twisted and turned and John Potts hammered out discords, he
thought of Grace’s beauty and charm.
To do him justice, he did not think of her great wealth. In fact, he
did not like to think of her wealth, for that wealth came from her
father’s success in a most unesthetic business. Grace was the
daughter of “Try a Tin To-day” Mallon, who was known in the world of
canned goods as the Salmon King.
There was no vulgar hand-clapping after the artists had finished
their interpretation of Grace’s soul; but the Rev. Morris Gildell rattled
his spoon against the sides of his teacup, while Mrs. Anna Watkins
Wilbur sighed ecstatically and clanked her chain of heavy amber
beads. Even in the act of lighting a cigarette, John Potts started
dramatically and pointed toward the open window.
“La voilà!” he exclaimed. “We have summoned her! I hear the
rumble of her chariot-wheels.”
There did indeed come from the cobbles of Patchin Place the
whir and snort of an automobile. Soon the great gong on the wall
clanged viciously, causing Arthura Lewis to shudder and clasp her
long, white hands to her eyes. Edwin Marmaduke sped down the
four flights of stairs that intervened between his studio and the street
door.
His deserted guests looked at one another expectantly.
 “Dear Grace!” said Mrs. Anna Watkins Wilbur. “I hope she came
alone. I can’t stand that friend of hers—what is his name?—that
Watson person.”
“He is a terrible bromid,” said Arthura Lewis, lighting a cigarette.
“He’s so very—so very, shall I say, salmony!”
This was said with humorous intent and received with kindly
laughter.
“But what has Mr. Watson to do with salmon, anyway?” asked
John Potts. “Does he catch them, or what is it?”
“Nothing so exciting as that,” said the Rev. Morris Gildell, with a
great chuckle. “He merely celebrates them. He is the advertising
manager of the Mallon Salmon Company, and his chief claim to
immortality is that he invented the ‘Try a Tin To-day’ slogan. You
know those great pink pictures of ridiculous-looking fish that we see
in public vehicles and on the highways and byways, each one
holding out a can bearing the legend, ‘Try a Tin To-day’? Watson is
responsible for those mutilations of the landscape.”
“Does Watson paint them himself?” asked Mrs. Anna Watkins
Wilbur. “He looks capable of it.”
The Rev. Morris Gildell issued another of those shouts of
laughter which made the women’s clubs believe him masculine and
hearty.
“Excellent!” he cried, putting an approving hand on Mrs. Anna
Watkins Wilbur’s fat, bare shoulder. “Oh, excellent! No, Watson
doesn’t paint them, but he does worse—he inspires them. He
persuades some poor devil of an artist to make these hideous
caricatures of the truth. He is the Mæcenas of the ‘Try a Tin To-day’
school of painting.”
After waiting for the murmur of amusement to die down, the Rev.
Morris Gildell continued:
“Grace brought Watson to the Feminist Conference at our church
last week, and several people asked me what a rare, free spirit like
her could see in such a clod. I suppose she feels that she must go
around with him occasionally because he is so useful to her father.
But I hope she doesn’t think of marrying him!”
“Marrying him!” exclaimed Mrs. Anna Watkins Wilbur, Arthura
Lewis, and John Potts in unison. “I hope not!”
“Why,” said Arthura Lewis, “it would be a tragedy for an
emancipated woman like Grace to marry at all! Not to speak of
marrying such a soulless, brainless animal as that David Watson! I’d
as soon see her—”
But the conversation was interrupted by the opening of the
studio door and the entrance of Grace and their young host.
Edwin Marmaduke seemed somehow to look younger than
before. He had lost a little of his expression of languor and disdain;
and Grace was the very personification of radiant girlhood. In spite of
her knowledge of her own good looks and good clothes, she was
charmingly deferential to these people, whom she considered
intellectuals. Her naively respectful greeting to the Rev. Morris Gildell
seemed for the moment to restore to that drawing-room revolutionist
some strange lost dignity.
Men and women alike greeted the young girl with genuine
friendliness. When her gray motor-cloak had been hung over the
corner of an easel, and she had been seated on one of the few real
chairs the studio boasted, and served with sweet biscuits, a cup of
tea, and a cigarette, then the great event of the afternoon occurred—
the unveiling of Marmaduke’s portrait of her.
The blinds were drawn—for the Incomprehensiblist painting is
best appreciated in semidarkness. The incense-pot received some
new fuel; John Potts played something “very golden” on the piano;
and Edwin Marmaduke reverently drew back the gay Indian scarf
that covered his masterpiece.

II