Download the full version of the ebook at ebookname.

com

R Programming for Bioinformatics 1st Edition

Robert Gentleman

https://ebookname.com/product/r-programming-for-
bioinformatics-1st-edition-robert-gentleman/

OR CLICK BUTTON

DOWNLOAD EBOOK

Download more ebook instantly today at https://ebookname.com

Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

Bioinformatics for Vaccinology 1st Edition Darren R.

Flower

https://ebookname.com/product/bioinformatics-for-vaccinology-1st-
edition-darren-r-flower/

ebookname.com

Bioinformatics for Geneticists 1st Edition Michael R.

Barnes

https://ebookname.com/product/bioinformatics-for-geneticists-1st-
edition-michael-r-barnes/

ebookname.com

Python Programming for Biology Bioinformatics and Beyond

1st Edition Tim J. Stevens

https://ebookname.com/product/python-programming-for-biology-
bioinformatics-and-beyond-1st-edition-tim-j-stevens/

ebookname.com

A History of Political Thought Plato to Marx Second

Edition Subrata Mukherjee

https://ebookname.com/product/a-history-of-political-thought-plato-to-
marx-second-edition-subrata-mukherjee/

ebookname.com
Frommer s Brazil 5th Edition Alexandra De Vries

https://ebookname.com/product/frommer-s-brazil-5th-edition-alexandra-
de-vries/

ebookname.com

Thyroid Hormone Receptors Methods and Protocols 1st

Edition Aria Baniahmad

https://ebookname.com/product/thyroid-hormone-receptors-methods-and-
protocols-1st-edition-aria-baniahmad/

ebookname.com

Electronic Excitations in Organic Based Nanostructures

Volume 31 1st Edition G. Franco Bassani

https://ebookname.com/product/electronic-excitations-in-organic-based-
nanostructures-volume-31-1st-edition-g-franco-bassani/

ebookname.com

The Blackwell Companion to Sociology of Religion Blackwell

Companions to Religion 1st Edition Richard K. Fenn

https://ebookname.com/product/the-blackwell-companion-to-sociology-of-
religion-blackwell-companions-to-religion-1st-edition-richard-k-fenn/

ebookname.com

Palliative Care in Amyotrophic Lateral Sclerosis From

Diagnosis to Bereavement 3rd Edition David Oliver

https://ebookname.com/product/palliative-care-in-amyotrophic-lateral-
sclerosis-from-diagnosis-to-bereavement-3rd-edition-david-oliver/

ebookname.com
Chinese Nuclear Proliferation How Global Politics Is
Transforming China s Weapons Buildup and Modernization 1st
Edition Susan Turner Haynes
https://ebookname.com/product/chinese-nuclear-proliferation-how-
global-politics-is-transforming-china-s-weapons-buildup-and-
modernization-1st-edition-susan-turner-haynes/
ebookname.com
R Programming for
Bioinformatics
Chapman & Hall/CRC
Computer Science and Data Analysis Series

The interface between the computer and statistical sciences is increasing, as each discipline
seeks to harness the power and resources of the other. This series aims to foster the integration
between the computer sciences and statistical, numerical, and probabilistic methods by
publishing a broad range of reference works, textbooks, and handbooks.

SERIES EDITORS
David Blei, Princeton University
David Madigan, Rutgers University
Marina Meila, University of Washington
Fionn Murtagh, Royal Holloway, University of London

Proposals for the series should be sent directly to one of the series editors above, or submitted to:

Chapman & Hall/CRC

4th Floor, Albert House
1-4 Singer Street
London EC2A 4BQ
UK

Published Titles

Bayesian Artificial Intelligence Design and Modeling for Computer

Kevin B. Korb and Ann E. Nicholson Experiments
Computational Statistics Handbook with Kai-Tai Fang, Runze Li, and
MATLAB®, Second Edition Agus Sudjianto
Wendy L. Martinez and Angel R. Martinez Introduction to Machine Learning
Pattern Recognition Algorithms for and Bioinformatics
Data Mining Sushmita Mitra, Sujay Datta,
Sankar K. Pal and Pabitra Mitra Theodore Perkins, and George Michailidis

Exploratory Data Analysis with MATLAB® R Graphics

Wendy L. Martinez and Angel R. Martinez Paul Murrell

Clustering for Data Mining: A Data R Programming for Bioinformatics

Recovery Approach Robert Gentleman
Boris Mirkin Semisupervised Learning for
Correspondence Analysis and Data Computational Linguistics
Coding with Java and R Steven Abney
Fionn Murtagh Statistical Computing with R
Maria L. Rizzo
R Programming for
Bioinformatics

Robert Gentleman
Fred Hutchinson Cancer Research Center
Seattle, Washington, U.S.A.
Chapman & Hall/CRC
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2009 by Taylor & Francis Group, LLC
Chapman & Hall/CRC is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S. Government works

Printed in the United States of America on acid-free paper
10 9 8 7 6 5 4 3 2 1

International Standard Book Number-13: 978-1-4200-6367-7 (Hardcover)

This book contains information obtained from authentic and highly regarded sources Reason-
able 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

Gentleman, Robert, 1959-

R programming for bioinformatics / Robert Gentleman.
p. cm. -- (Chapman & Hall/CRC computer science and data analysis series)
Bibliographical references (p. ) and index.
ISBN 978-1-4200-6367-7
1. Bioinformatics. 2. R (Computer program language) I. Title. II. Series.

QH324.2.G46 2008
572.80285’5133--dc22 2008011352

Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
To Tanja, Sophie and Katja
Contents

1 Introducing R 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 A note on the text . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . 4

2 R Language Fundamentals 5
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1.1 A brief introduction to R . . . . . . . . . . . . . . . . 5
2.1.2 Attributes . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1.3 A very brief introduction to OOP in R . . . . . . . . . 7
2.1.4 Some special values . . . . . . . . . . . . . . . . . . . 8
2.1.5 Types of objects . . . . . . . . . . . . . . . . . . . . . 9
2.1.6 Sequence generating and vector subsetting . . . . . . . 11
2.1.7 Types of functions . . . . . . . . . . . . . . . . . . . . 12
2.2 Data structures . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2.1 Atomic vectors . . . . . . . . . . . . . . . . . . . . . . 12
2.2.2 Numerical computing . . . . . . . . . . . . . . . . . . 15
2.2.3 Factors . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.2.4 Lists, environments and data frames . . . . . . . . . . 18
2.3 Managing your R session . . . . . . . . . . . . . . . . . . . . . 22
2.3.1 Finding out more about an object . . . . . . . . . . . 24
2.4 Language basics . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.4.1 Operators . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.5 Subscripting and subsetting . . . . . . . . . . . . . . . . . . . 28
2.5.1 Vector and matrix subsetting . . . . . . . . . . . . . . 29
2.6 Vectorized computations . . . . . . . . . . . . . . . . . . . . . 36
2.6.1 The recycling rule . . . . . . . . . . . . . . . . . . . . 37
2.7 Replacement functions . . . . . . . . . . . . . . . . . . . . . . 38
2.8 Functional programming . . . . . . . . . . . . . . . . . . . . . 39
2.9 Writing functions . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.10 Flow control . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.10.1 Conditionals . . . . . . . . . . . . . . . . . . . . . . . 44
2.11 Exception handling . . . . . . . . . . . . . . . . . . . . . . . . 45
2.12 Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.12.1 Standard evaluation . . . . . . . . . . . . . . . . . . . 51
2.12.2 Non-standard evaluation . . . . . . . . . . . . . . . . . 52

vii
viii

2.12.3 Function evaluation . . . . . . . . . . . . . . . . . . . 53

2.12.4 Indirect function invocation . . . . . . . . . . . . . . . 54
2.12.5 Evaluation on exit . . . . . . . . . . . . . . . . . . . . 54
2.12.6 Other topics . . . . . . . . . . . . . . . . . . . . . . . 55
2.12.7 Name spaces . . . . . . . . . . . . . . . . . . . . . . . 57
2.13 Lexical scope . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
2.13.1 Likelihoods . . . . . . . . . . . . . . . . . . . . . . . . 61
2.13.2 Function optimization . . . . . . . . . . . . . . . . . . 62
2.14 Graphics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

3 Object-Oriented Programming in R 67
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.2 The basics of OOP . . . . . . . . . . . . . . . . . . . . . . . . 68
3.2.1 Inheritance . . . . . . . . . . . . . . . . . . . . . . . . 69
3.2.2 Dispatch . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.2.3 Abstract data types . . . . . . . . . . . . . . . . . . . 72
3.2.4 Self-describing data . . . . . . . . . . . . . . . . . . . 73
3.3 S3 OOP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.3.1 Implicit classes . . . . . . . . . . . . . . . . . . . . . . 76
3.3.2 Expression data example . . . . . . . . . . . . . . . . 77
3.3.3 S3 generic functions and methods . . . . . . . . . . . . 78
3.3.4 Details of dispatch . . . . . . . . . . . . . . . . . . . . 81
3.3.5 Group generics . . . . . . . . . . . . . . . . . . . . . . 83
3.3.6 S3 replacement methods . . . . . . . . . . . . . . . . . 83
3.4 S4 OOP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.4.1 Classes . . . . . . . . . . . . . . . . . . . . . . . . . . 85
3.4.2 Types of classes . . . . . . . . . . . . . . . . . . . . . . 98
3.4.3 Attributes . . . . . . . . . . . . . . . . . . . . . . . . . 98
3.4.4 Class unions . . . . . . . . . . . . . . . . . . . . . . . 99
3.4.5 Accessor functions . . . . . . . . . . . . . . . . . . . . 100
3.4.6 Using S3 classes with S4 classes . . . . . . . . . . . . . 100
3.4.7 S4 generic functions and methods . . . . . . . . . . . . 101
3.4.8 The syntax of method declaration . . . . . . . . . . . 105
3.4.9 The semantics of method invocation . . . . . . . . . . 106
3.4.10 Replacement methods . . . . . . . . . . . . . . . . . . 107
3.4.11 Finding methods . . . . . . . . . . . . . . . . . . . . . 107
3.4.12 Advanced topics . . . . . . . . . . . . . . . . . . . . . 108
3.5 Using classes and methods in packages . . . . . . . . . . . . . 110
3.6 Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . 110
3.6.1 Finding documentation . . . . . . . . . . . . . . . . . 110
3.6.2 Writing documentation . . . . . . . . . . . . . . . . . 111
3.7 Debugging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
3.8 Managing S3 and S4 together . . . . . . . . . . . . . . . . . . 112
3.8.1 Getting and setting the class attribute . . . . . . . . 113
3.8.2 Mixing S3 and S4 methods . . . . . . . . . . . . . . . 114
 ix

3.9 Navigating the class and method hierarchy . . . . . . . . . . 115

4 Input and Output in R 119

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
4.2 Basic file handling . . . . . . . . . . . . . . . . . . . . . . . . 120
4.2.1 Viewing files . . . . . . . . . . . . . . . . . . . . . . . 124
4.2.2 File manipulation . . . . . . . . . . . . . . . . . . . . . 125
4.2.3 Working with R’s binary format . . . . . . . . . . . . 129
4.3 Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
4.3.1 Text connections . . . . . . . . . . . . . . . . . . . . . 131
4.3.2 Interprocess communications . . . . . . . . . . . . . . 133
4.3.3 Seek . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
4.4 File input and output . . . . . . . . . . . . . . . . . . . . . . 137
4.4.1 Reading rectangular data . . . . . . . . . . . . . . . . 138
4.4.2 Writing data . . . . . . . . . . . . . . . . . . . . . . . 139
4.4.3 Debian Control Format (DCF) . . . . . . . . . . . . . 140
4.4.4 FASTA Format . . . . . . . . . . . . . . . . . . . . . . 141
4.5 Source and sink: capturing R output . . . . . . . . . . . . . . 142
4.6 Tools for accessing files on the Internet . . . . . . . . . . . . . 143

5 Working with Character Data 145

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
5.2 Builtin capabilities . . . . . . . . . . . . . . . . . . . . . . . . 146
5.2.1 Modifying text . . . . . . . . . . . . . . . . . . . . . . 151
5.2.2 Sorting and comparing . . . . . . . . . . . . . . . . . . 152
5.2.3 Matching a set of alternatives . . . . . . . . . . . . . . 153
5.2.4 Formatting text and numbers . . . . . . . . . . . . . . 155
5.2.5 Special characters and escaping . . . . . . . . . . . . . 155
5.2.6 Parsing and deparsing . . . . . . . . . . . . . . . . . . 158
5.2.7 Plotting with text . . . . . . . . . . . . . . . . . . . . 159
5.2.8 Locale and font encoding . . . . . . . . . . . . . . . . 159
5.3 Regular expressions . . . . . . . . . . . . . . . . . . . . . . . . 159
5.3.1 Regular expression basics . . . . . . . . . . . . . . . . 160
5.3.2 Matching . . . . . . . . . . . . . . . . . . . . . . . . . 166
5.3.3 Using regular expressions . . . . . . . . . . . . . . . . 167
5.3.4 Globbing and regular expressions . . . . . . . . . . . . 169
5.4 Prefixes, suffixes and substrings . . . . . . . . . . . . . . . . . 169
5.5 Biological sequences . . . . . . . . . . . . . . . . . . . . . . . 171
5.5.1 Encoding genomes . . . . . . . . . . . . . . . . . . . . 172
5.6 Matching patterns . . . . . . . . . . . . . . . . . . . . . . . . 173
5.6.1 Matching single query sequences . . . . . . . . . . . . 174
5.6.2 Matching many query sequences . . . . . . . . . . . . 175
5.6.3 Palindromes and paired matches . . . . . . . . . . . . 177
5.6.4 Alignments . . . . . . . . . . . . . . . . . . . . . . . . 179
x

6 Foreign Language Interfaces 183

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
6.1.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . 184
6.1.2 The C programming language . . . . . . . . . . . . . . 185
6.2 Calling C and FORTRAN from R . . . . . . . . . . . . . . . 185
6.2.1 .C and .Fortran . . . . . . . . . . . . . . . . . . . . . 186
6.2.2 Using .Call and .External . . . . . . . . . . . . . . . 187
6.3 Writing C code to interface with R . . . . . . . . . . . . . . . 188
6.3.1 Registering routines . . . . . . . . . . . . . . . . . . . 188
6.3.2 Dealing with special values . . . . . . . . . . . . . . . 189
6.3.3 Single precision . . . . . . . . . . . . . . . . . . . . . . 191
6.3.4 Matrices and arrays . . . . . . . . . . . . . . . . . . . 191
6.3.5 Allowing interrupts . . . . . . . . . . . . . . . . . . . . 193
6.3.6 Error handling . . . . . . . . . . . . . . . . . . . . . . 193
6.3.7 R internals . . . . . . . . . . . . . . . . . . . . . . . . 193
6.3.8 S4 OOP in C . . . . . . . . . . . . . . . . . . . . . . . 197
6.3.9 Calling R from C . . . . . . . . . . . . . . . . . . . . . 198
6.4 Using the R API . . . . . . . . . . . . . . . . . . . . . . . . . 198
6.4.1 Header files . . . . . . . . . . . . . . . . . . . . . . . . 198
6.4.2 Sorting . . . . . . . . . . . . . . . . . . . . . . . . . . 199
6.4.3 Random numbers . . . . . . . . . . . . . . . . . . . . . 199
6.5 Loading libraries . . . . . . . . . . . . . . . . . . . . . . . . . 202
6.5.1 Inspecting DLLs . . . . . . . . . . . . . . . . . . . . . 203
6.6 Advanced topics . . . . . . . . . . . . . . . . . . . . . . . . . 204
6.6.1 External references and finalizers . . . . . . . . . . . 204
6.6.2 Evaluating R expressions from C . . . . . . . . . . . . 206
6.7 Other languages . . . . . . . . . . . . . . . . . . . . . . . . . 209

7 R Packages 211
7.1 Package basics . . . . . . . . . . . . . . . . . . . . . . . . . . 212
7.1.1 The search path . . . . . . . . . . . . . . . . . . . . . 212
7.1.2 Package information . . . . . . . . . . . . . . . . . . . 213
7.1.3 Data and demos . . . . . . . . . . . . . . . . . . . . . 215
7.1.4 Vignettes . . . . . . . . . . . . . . . . . . . . . . . . . 215
7.2 Package management . . . . . . . . . . . . . . . . . . . . . . . 216
7.2.1 biocViews . . . . . . . . . . . . . . . . . . . . . . . . 218
7.2.2 Managing libraries . . . . . . . . . . . . . . . . . . . . 219
7.3 Package authoring . . . . . . . . . . . . . . . . . . . . . . . . 219
7.3.1 The DESCRIPTION file . . . . . . . . . . . . . . . . . . 220
7.3.2 R code . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
7.3.3 Documentation . . . . . . . . . . . . . . . . . . . . . . 221
7.3.4 Name spaces . . . . . . . . . . . . . . . . . . . . . . . 224
7.3.5 Finding out about name spaces . . . . . . . . . . . . . 226
7.4 Initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
7.4.1 Event hooks . . . . . . . . . . . . . . . . . . . . . . . . 227
 xi

8 Data Technologies 229

8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
8.1.1 A brief description of GO . . . . . . . . . . . . . . . . 229
8.2 Using R for data manipulation . . . . . . . . . . . . . . . . . 230
8.2.1 Aggregation and creating tables . . . . . . . . . . . . . 230
8.2.2 Apply functions . . . . . . . . . . . . . . . . . . . . . . 232
8.2.3 Efficient apply-like functions . . . . . . . . . . . . . . 234
8.2.4 Combining and reshaping rectangular data . . . . . . 234
8.3 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
8.4 Database technologies . . . . . . . . . . . . . . . . . . . . . . 238
8.4.1 DBI . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
8.4.2 SQLite . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
8.4.3 Using AnnotationDbi . . . . . . . . . . . . . . . . . 243
8.5 XML . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
8.5.1 Simple XPath . . . . . . . . . . . . . . . . . . . . . . . 256
8.5.2 The XML package . . . . . . . . . . . . . . . . . . . . 257
8.5.3 Handlers . . . . . . . . . . . . . . . . . . . . . . . . . . 257
8.5.4 Example data . . . . . . . . . . . . . . . . . . . . . . . 258
8.5.5 DOM parsing . . . . . . . . . . . . . . . . . . . . . . . 258
8.5.6 XML event parsing . . . . . . . . . . . . . . . . . . . . 261
8.5.7 Parsing HTML . . . . . . . . . . . . . . . . . . . . . . 263
8.6 Bioinformatic resources on the WWW . . . . . . . . . . . . . 264
8.6.1 PubMed . . . . . . . . . . . . . . . . . . . . . . . . . . 265
8.6.2 NCBI . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
8.6.3 biomaRt . . . . . . . . . . . . . . . . . . . . . . . . . . 266
8.6.4 Getting data from GEO . . . . . . . . . . . . . . . . . 270
8.6.5 KEGG . . . . . . . . . . . . . . . . . . . . . . . . . . . 272

9 Debugging and Profiling 273

9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
9.2 The browser function . . . . . . . . . . . . . . . . . . . . . . . 274
9.2.1 A sample browser session . . . . . . . . . . . . . . . . 275
9.3 Debugging in R . . . . . . . . . . . . . . . . . . . . . . . . . . 276
9.3.1 Runtime debugging . . . . . . . . . . . . . . . . . . . . 277
9.3.2 Warnings and other exceptions . . . . . . . . . . . . . 278
9.3.3 Interactive debugging . . . . . . . . . . . . . . . . . . 279
9.3.4 The debug and undebug functions . . . . . . . . . . . . 281
9.3.5 The trace function . . . . . . . . . . . . . . . . . . . . 285
9.4 Debugging C and other foreign code . . . . . . . . . . . . . . 289
9.5 Profiling R code . . . . . . . . . . . . . . . . . . . . . . . . . 290
9.5.1 Timings . . . . . . . . . . . . . . . . . . . . . . . . . . 292
9.6 Managing memory . . . . . . . . . . . . . . . . . . . . . . . . 293
9.6.1 Memory profiling . . . . . . . . . . . . . . . . . . . . . 294
9.6.2 Profiling memory allocation . . . . . . . . . . . . . . . 295
9.6.3 Tracking a single object . . . . . . . . . . . . . . . . . 298
xii

References 301
Chapter 1
Introducing R

1.1 Introduction
The purpose of this monograph is to provide a reference for scientists and
programmers working on problems in bioinformatics and computational bi-
ology. It may also appeal to programmers who want to improve their pro-
gramming skills or programmers who have been working in bioinformatics
and computational biology but are familiar with languages other than R. A
reasonable level of programming skill is presumed as is some familiarity with
some of the basic tasks that need to be carried out in bioinformatics. We
concentrate on programming tools and there is no discussion of either graph-
ics or of the multitude of software for fitting models or carrying out machine
learning. Reasonable coverage of these topics would result in a much longer
monograph and to some extent they are orthogonal to our purpose.
Bioinformatics blossomed as a scientific discipline in the 1990s when a num-
ber of technological innovations appeared that revolutionized biology. Sud-
denly, data on the complete genomic sequence of many different organisms
were available, microarrays could measure the abundance of tens of thou-
sands of mRNA species, and other arrays and technologies made it possible to
study protein interactions and many other cellular processes at the molecular
level. Basically, biology moved from a small data discipline to one with large
complex data sets, virtually overnight.
Faced with these sudden challenges, scientific programmers grabbed what-
ever tools were available and made use of them to help address some of the
many problems. Perl was perhaps the most widely used and it remains a
dominant player to this date. Other popular programming languages such as
Java and Python are also used.
R is an implementation of the S language (Becker et al., 1988; Chambers
and Hastie, 1992; Chambers, 1998). S has been a favorite tool for statisticians
and data analysts since the early 1980s when John Chambers and colleagues
started to release versions of it from Bell Labs. It is now becoming one of
the most widely used software tools for bioinformatics. This is mainly due
to its flexibility and data handling and modeling capabilities. Some of these
have been exposed through the Bioconductor Project (Gentleman et al., 2004)
but many users simply find it a useful tool for doing analyses. However, our

1
2 R Programming for Bioinformatics

experience is that it is easy to write inefficient programs, and often the basic
programming idioms are missed or ignored.
In Chapter 2 we discuss the general properties of the R language and some
of the unique aspects of programming in it. In Chapter 3 we discuss object-
oriented programming in R. The paradigm is quite different and may take
some getting used to, but like all object-oriented systems, mastering these
topics is essential to writing good maintainable software. Then Chapter 4 dis-
cusses methods for getting data in and out, for interacting with databases and
includes a discussion of XML, SOAP and other data mark-up and web-services
languages and tools. Chapter 5 discusses different aspects of string handling
and manipulations, including many of the standard sequence similarity tools
that play a prominent role in computational biology. In Chapter 6 we con-
sider interacting with foreign languages, primarily on C, but we also consider
FORTRAN, Perl and Python. In Chapter 7 we describe how to write your
own software packages that can be used locally or distributed more broadly.
Finally we finish with Chapter 9, which discusses debugging and profiling of
R code.
R comes with a substantial amount of documentation. Specifically there
are five manuals: An Introduction to R, The R Language Definition, R In-
stallation and Administration, Writing R Extensions, and R Data Import
and Export. We will draw on material in these manuals throughout this
monograph, and readers who want more detail or alternative examples should
consult them. We will rely most on the Writing R Extensions Manual, which
we abbreviate to R Extensions. R News is a good source of information on R
packages and on aspects of the language written at an accessible level. Read-
ers are encouraged to browse the back issues for more information on topics
that are just touched on in this volume. Venables and Ripley (2000) is another
reference for programming in the S language, as is Chambers (2008).

1.2 Motivation
There are many good reasons to prefer R to other languages for scientific
computation. The existence of a substantial collection of good statistical
algorithms, access to high-quality numerical routines, and integrated data
visualization tools are perhaps the most obvious ones. But as we have been
trying to show through the Bioconductor Project (www.bioconductor.org),
there are many more.
Reproducibility is an essential part of any scientific investigation, but to
date very little attention has been paid to this topic. Our efforts are R-based
(Gentleman, 2005) and make use of the Sweave system (Leisch, 2002). Indeed,
as we discuss later, this entire book has been written so that every example
 Introducing R 3

is reproducible on the reader’s machine. The ability to integrate text and

software into a single document greatly facilitates the writing of scientific
papers and helps to ensure that all figures, tables and facts are based on the
same data, and are essentially reproducible by the reader.
A second strong motivation for using R is its ability to interoperate with
many other languages. Algorithms that have been written in another language
seldom need to be reimplemented for use in R. Typically one need merely write
a small amount of interface code and the routines can be accessed from within
R (this is described in Chapter 6). This approach also helps to ensure maximal
code reuse.
And finally, R supports the creation and use of self-describing data struc-
tures. In the Bioconductor Project we have relied heavily on this capability in
our design and use of the ExpressionSet class. This data structure is designed
to hold the output of a microarray experiment as well as detailed informa-
tion on the experimental design, other covariates that are available for the
samples that were run and links to information on the genes that correspond
to the spots on the array. While this has been successful in that context,
we have reused this data structure with similar benefits for other data types
such as those that arise in proteomic studies (the PROcess package) and flow
cytometry (the flowCore package).

1.3 A note on the text

This monograph was written using the Sweave system (Leisch, 2002), which
is a tool that allows authors to integrate text (using LATEX) and computer code
for the R language. Hence, all examples are reproducible by the reader and
readers can obtain complete source code (but not the text) on a per-chapter
basis from the web site for this monograph. There are a number of exercises
given and solutions for some of them are available in the online supplements.
The examples themselves are often shown integrated into the text of the
chapters. Not all code is displayed; in many cases preliminary computations,
loading of libraries and other mundane tasks are not displayed in the text
version; they are included in the R code for the chapters. Any example that
relies on simulation or the use of a random number generator will have a call
to the set.seed function as a preliminary command. The sole reason for this
is to ensure reproducibility of the output on the user’s machine.
In cases where the code is intended to signal an error, the call is enclosed in
either a call to try or more often in a call to tryCatch. This is done because any
error signaled by R interrupts the Sweave process and causes typesetting to
fail. Details on the behavior of try and tryCatch can be found in Section 2.11.
Markup is used to distinguish some entities. For example, functions are
4 R Programming for Bioinformatics

marked up like mean, R packages using Biobase, function arguments, myarg,

R classes with ExpressionSet, and R objects using x. When R prints a value
that corresponds to a vector, some indexing information is provided. In the
example below, we print a vector of integers from 1 to 30. The first thing
printed is [1], which indicates that the first value on that line is the first
value in the vector, and on the second printed line the [18] indicates that the
first value in that line corresponds to the 18th element of the vector. The grey
background is used for all code examples that were processed by the Sweave
system.

> 1:30

[1] 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
[18] 18 19 20 21 22 23 24 25 26 27 28 29 30

It is essential that the reader follow along and experiment with some of the
examples given, so two basic strategies are advised. First, make use of the help
system, either by constructs such as help("[") or the shorthand, equivalent,
?"[". Many special functions and symbols need to be quoted. All help pages
should have examples and these can be run using the function example, e.g.,
example("[")). The second basic strategy is to investigate the code itself,
and for this purpose get is most useful; for example, try get("mode") and see
if you can better understand how it works.

1.4 Acknowledgments
Many people have contributed, both directly and indirectly, to the cre-
ation of this book. Both the R and Bioconductor development teams have
contributed substantially to my understanding, and many members of those
projects have provided examples, clarified misunderstandings and provided a
rich environment in which to discuss relevant issues. Members of my research
group have contributed to many aspects; in particular, J. Gentry, S. DebRoy,
H. Pagés, M. Morgan, N. Li, T.-Y. Liu, M. Carlson, P. Aboyoun, D. Sarkar,
F. Hahne and S. Falcon have contributed many ideas, examples and helped
clarify issues. Wolfgang Huber and Vincent Carey made extensive comments
and recommendations. All errors remain my own and I will attempt to remedy
those that are found and reported, in a timely fashion.
Chapter 2
R Language Fundamentals

2.1 Introduction
In this chapter we introduce the basic language data types and discuss
their capabilities and structures. Then topics such as flow-control, iteration,
subsetting and exception handling will be presented. R directly supports two
different object-oriented programming (OOP) paradigms, which are discussed
in detail in Chapter 3. Many operations in R are vectorized, and understand-
ing and using vectorization is an essential component of becoming a proficient
programmer.
The R language was primarily designed as a language for data manipula-
tion, modeling and visualization, and many of the data structures reflect this
view. However, R is itself a full-fledged programming language, with its own
idioms – much like any other programming language. In some ways R can be
considered as a functional programming language, although it is not purely
functional. R supports a form of lexical scope that provides a useful paradigm
for encapsulating computations.
R is an implementation of the S language (Becker et al., 1988; Chambers and
Hastie, 1992; Chambers, 1998). There is another commercial implementation
available from Insightful Corporation, called S-PLUS. The two implementa-
tions are quite similar, and much of the material covered here can be used in
either. However, there are many R-specific extensions that are used in this
monograph and users of R are our intended audience.

2.1.1 A brief introduction to R

We presume a reasonable familiarity with R but there are a few points that
will help to clarify some of the discussion. When R is started, a workspace is
created and that workspace is where the user creates and manipulates vari-
ables. This workspace is an environment, and an environment is a set of
bindings of names, or symbols, to values. The top-level workspace can be
accessed through its name, which is .GlobalEnv.
Assignment of value to a variable is generally done with either the = (equals)
character, or a special symbol that is the concatenation of less than and mi-
nus, <-. Assignment creates a binding between a symbol and a value, in a

5
6 R Programming for Bioinformatics

particular environment. Removal of bindings is done with the function rm. In

the next code chunk, we create a symbol x and assign to it the value 10. We
then create a second symbol and assign the same value as x has.

> x = 10
> y = x

The value associated with y is a copy of the value associated with x, and
changes to x do not affect y.
The semantics of rm(x) are that the association between x and its value
is broken and the symbol x is removed from the environment, but nothing is
done to the value that x referred to. If this value can be accessed in other
ways, it will remain available. We provide an example in Section 2.2.4.3.
Valid variable names, sometimes referred to as syntactic names, are any
sequence of letters, digits, the period and the underscore, but they cannot
begin with a digit or the underscore. If they begin with a period, the second
character cannot be a digit. Variable names that violate these rules must be
quoted (see the Quotes manual page) and the preferred quote is the backtick.

> _foo = 10
> "10:10" = 20
> ls()

[1] "10:10" "Rvers" "_foo" "basename"

[5] "biocUrls" "repos" "x" "y"

2.1.2 Attributes
Attributes can be attached to any R object except NULL and they are used
quite extensively. Attributes are stored, by name, in a list. All attributes can
be retrieved using attributes, or any particular attribute can be accessed or
modified using the attr function. Attributes can be used by programmers to
attach any sort of information they want to any R object. R uses attributes
for many things; the S3 class system is based largely on attributes, dimensions
of arrays, and names on vectors, to name but a few.
In the code below, we attach an attribute to x and then show how the
printing of x changes to reflect the fact that it has an attribute.

> x = 1:10
> attr(x, "foo") = 11
> x
 R Language Fundamentals 7

[1] 1 2 3 4 5 6 7 8 9 10
attr(,"foo")
[1] 11

2.1.3 A very brief introduction to OOP in R

In order to fully explain some of the concepts in this chapter, the reader will
need a little familiarity with the basic ideas in object oriented-programming
(OOP), as they are implemented in R. A more comprehensive treatment of
these topics is given in Chapter 3. There are two components: one is a class
system that is used to define the class of different objects, and the second
is the notion of a generic function with methods. R has two OOP systems:
one is referred to as S3, and it mainly supports generic functions; the other is
referred to as S4, and it has support for classes as well as generic functions,
although these are somewhat different from the S3 variants. We will only
discuss S3 here.
In S3, the class system is very lax, and one creates an object (typically called
an instance) from a class by attaching a class attribute to any R object. As a
result, no checking is done, or can easily be done, to ensure common structure
of different instances of the same class. A generic function is essentially a
dispatching mechanism, and in S3 the dispatch is handled by concatenating
the name of the generic function with that of the class. An example of a
generic function is mean.

> mean

function (x, ...)

UseMethod("mean")
<environment: namespace:base>

The general form of a generic function, as seen in the example above, is for
a single expression, which is the call to UseMethod, which is the mechanism
that helps to dispatch to the appropriate method. We can see all the defined
methods for this function using the methods command.

> methods("mean")

[1] mean.Date mean.POSIXct mean.POSIXlt

[4] mean.data.frame mean.default mean.difftime
8 R Programming for Bioinformatics

And see that they all begin with the name mean, then a period. When
the function mean is called, R looks at the first argument and determines
whether or not that argument has a class attribute. If it does, then R looks
for a function whose name starts with mean. and then has the name of the
class. If one exists, then that method is used; and if one does not exist, then
mean.default is used.

2.1.4 Some special values

There are a number of special variables and values in the language and
before embarking on data structures we will introduce these. The value NULL
is the null object. It has length zero and disappears when concatentated with
any other object. It is the default value for the elements of a list.

> length(NULL)

[1] 0

> c(1, NULL)

[1] 1

> list("a", NULL)

[[1]]
[1] "a"

[[2]]
NULL

Since R has its roots in data analysis, the appropriate handling of missing
data items is important. There are special missing data values for all atomic
types and these are commonly referred to by the symbol NA. And similarly
there are special functions for identifying these values, such as is.na, and many
modeling routines have special methods for dealing with missing values. It is
worth emphasizing that there is a distinct missing value (NA) for each basic
type and these can be accessed through constants such as NA_integer_.

> typeof(NA)

[1] "logical"

> as.character(NA)
 R Language Fundamentals 9

[1] NA

> as.integer(NA)

[1] NA

> typeof(as.integer(NA))

[1] "integer"

Note that the character string formed by concatenating the characters N

and A is not a missing value.

> is.na("NA")

[1] FALSE

The appropriate representation of values such as infinity and not a number

(NaN) is provided. There are accompanying functions, is.finite, is.infinite
and is.nan, that can be used to determine whether a particular value is one of
these special values. All mathematics functions should deal with these values
appropriately, according to the ANSI/IEEE 754 floating-point standard.

> y = 1/0
> y

[1] Inf

> -y

[1] -Inf

> y - y

[1] NaN

2.1.5 Types of objects

An important data structure in R is the vector. Vectors are ordered collec-
tions of objects, where all elements are of the same type. Vectors can be of
any length (including zero), up to some maximum allowable, which is deter-
mined by the storage capabilities of the machine being used. Vectors typically
10 R Programming for Bioinformatics

represent a form of contiguous storage (character vectors are an exception).

R has six basic vector types: logical, integer, real, complex, string (or charac-
ter) and raw. The type of a vector can be queried by using one of the three
functions mode, storage.mode or typeof.

> typeof(y)

[1] "double"

> typeof(is.na)

[1] "builtin"

> typeof(mean)

[1] "closure"

> mode(NA)

[1] "logical"

> storage.mode(letters)

[1] "character"

There are also a number of predicate functions that can be used to test
whether a value corresponds to one of the basic vector types. The code chunk
below demonstrates the use of several of the predicate functions available.

> is.integer(y)

[1] FALSE

> is.character(y)

[1] FALSE

> is.double(y)

[1] TRUE

> is.numeric(y)

[1] TRUE
 R Language Fundamentals 11
Exercise 2.1
What does the typeof is.na mean? Why is it different from that of mean?

2.1.6 Sequence generating and vector subsetting

It is also helpful to discuss a couple of functions operators before beginning
the general discussion, as they will help make the exposition easier to follow.
Some of these, such as the subsetting operator [, we will return to later for a
more complete treatment.
The colon, :, indicates a sequence of values, from the number that is to its
left, to the number on the right, in steps of 1. We will also need to make some
use of the subset operator, [. This operator takes a subset of the vector it is
applied to according to the arguments inside the square brackets.

> 1:3

[1] 1 2 3

> 1.3:3.2

[1] 1.3 2.3

> 6:3

[1] 6 5 4 3

> x = 11:20
> x[4:5]

[1] 14 15

These are just ordinary functions, and one can invoke them as if they are.
The usual infix notation, with the : between the lower and upper bounds on
the sequence, may lead one to believe that this is not an ordinary function.
But that is not true, and one can also invoke this function using a somewhat
more standard notation, ":"(2,4)". Quotes are needed around the colon to
ensure it is not interpreted in an infix context by the parser.

Exercise 2.2
Find help for the colon operator; what does it do? What is the type of its
return value? Use the predicate testing functions to determine the storage
mode of the expressions 1:3 and 1.3:4.2.
12 R Programming for Bioinformatics

2.1.7 Types of functions

This section is slightly more detailed and can be skipped. In R there are ba-
sically three types of functions: builtins, specials and closures. Users can only
create closures (unless they want to modify the internals of R), and these are
the easiest functions to understand since they are written in R. The other two
types of functions are interfaces that attempt to pass the calculations down to
internal (typically C) routines for efficiency reasons. The main difference be-
tween the two types of internal functions is whether or not they evaluate their
arguments; specials do not. More details on the internals of R are available
in the R Language Definition (R Development Core Team, 2007b).

2.2 Data structures

2.2.1 Atomic vectors
Atomic vectors are the most basic of all data structures. An atomic vector
contains some number of values of the same type; that number could be zero.
Atomic vectors can contain integers, doubles, logicals or character strings.
Both complex numbers and raw (pure bytes) have atomic representations (see
the R documentation for more details on these two types). Character vectors
in the S language are vectors of character strings, not vectors of characters.
For example, the string "super" would be represented as a character vector
of length one, not of length five (for more details on character handling in R,
see Chapter 5). A dim attribute can be added to an atomic vector to create a
matrix or an array.

> x = c(1, 2, 3, 4)
> x

[1] 1 2 3 4

> dim(x) = c(2, 2)

> x

[,1] [,2]
[1,] 1 3
[2,] 2 4

> typeof(x)

[1] "double"
Random documents with unrelated
content Scribd suggests to you:
better off.”
“I’ve made the break now and again,” Bart explained. “But they
always turn up. Our family line-up is fashioned after that fabled joint
snake. You can disrupt the critter but the pieces crawl back together
again and all stand united.”
“If there’s any more midnight visits made at my cabin,” said Carver;
“there’ll be one middle joint absent from the next family reunion.”
“I take it you’re referring to Noll,” said Bart. “If you’ll only accept my
earnest advice you’ll decoy Noll off to some quiet spot and snap a
cap at him. I promise it won’t upset me a bit.”
On the third day out from the ranch Carver rode with Nate Younger
along a low ridge studded with a straggling stand of black-jack
timber. The old man’s face was stern and set as he viewed the
procession filing for two miles along the open bottoms below them.
A dozen round-up crews made up the picture, for this was a
coöperative move by all the outfits ranging in the Strip, the great final
combing of stock from the unowned lands.
Far up the valley, a mere speck in the distance, the Half Diamond H
wagon led the way while the others trailed at intervals. Two hundred
riders, the personnel probably including the most efficient body of
cowhands in the world, straggled up the bottoms in irregular
formation. The extra horses, if combined into one cavayado would
number over two thousand head. A group of riders hovered near the
last wagon, it having encountered difficulties in making the crossing
of the Cimarron, resuming their way as the quick-sands relinquished
their sucking hold upon the wheels and the floundering horses
snaked the lumbering vehicle out upon the solid shore. A band of
twenty Cherokees flanked the cavalcade and dashed from one outfit
to the next, begging food from each wagon boss in turn. Midway of
the procession a detachment of cavalry rode in double file while the
officer in command conferred with the man in charge of that
particular wagon. As Carver watched they dropped back abreast of
the next in line and he knew the message delivered to each one in
turn by the soldiery,—the instructions to make a thorough sweep and
clear every head of stock from the Cherokee Strip.
The Indians, having gathered contributions sufficient for the moment,
including a steer which was pointed out to them by the owner of the
brand worn by the animal, hazed this moveable meat supply to the
crest of an adjacent knoll and there dropped it with an
accompaniment of rifle shots. Younger waved a hand toward the
scene spread out before him.
“That’s the way I saw the Old West first,” he said. “The picture is
mighty near identical; the wagons rolling along just like that, only
drawn up in more tight formation; the cavayado trailing under guard,
holding all the extra horses of the settlers; maybe a band of
marauding reds clustered off to one side like them that are hacking
up that steer; sometimes a little escort of troopers helping us at bad
crossings where the Kiowas and Comanches was most liable to
jump us while a part of the train was bogged down in the sand. The
wagons was more likely dragged by bulls than horses then, and
buffalo was scattered round the landscape in place of range cows,
but on the whole the picture tallies close enough.” The old man
turned his gaze away. “That’s the way we was first ushered into the
Old West, son. Maybe it’s fitting that we’re being similarly ushered
out of the last bit that’s left for us.”
They rode on in silence and regained the head of the line. The
various wagons made camp at intervals sufficient to permit the
remudas of different outfits to be held on good grass at widely
separate points to prevent the possibility of their mixing. On this
occasion the men rode from one night camp to the next to renew old
friendships, fraternizing with the hands who rode for rival brands.
Another crew of similar magnitude had assembled at another point in
the Strip and during this same hour these men too were mingling
from one outfit to the next. Perhaps among the entire three hundred
odd gathered at these two points there was not one man who fully
realized that this meeting was to be the last of its sort; not one who
could even partly vision the circumstances of the next.
Never again in history were these men to gather as a whole on the
open range. This night was the last. Many would meet in the future;
others would never meet again. Some would be neighbors for a
lifetime and it was slated that the trails of others should cross in far
places. Perhaps it is well that it is not given to man to look far into
the future. This last occasion was not marred by any thought that the
summons for the next gathering would not go forth for more than a
quarter of a century. There were many present who would heed that
plea which would one day be issued for all the old-time peelers and
bronc fighters of the Cherokee lands to assemble for a final rally.
They would not then travel across the open range with chuck
wagons and saddle horse. Some would be carried in luxuriously
appointed coaches that roared along steel rails; others in glittering
vehicles that purred swiftly along fenced and well-kept highways;
some would arrive in strange craft that swept across the skies above
thriving western cities situated on spots now widely known as ideal
cow-camp sites. A few indeed, but very few, would come in
buckboards or ride in on horses, their ropes coiled on ancient
saddles; and it would be these latter ones who would then appear
strange and out of place. But no such glimpse of future actualities
troubled the men as they sought friends who worked with other
wagons. There was a general disposition to scoff at the notion that
there would be no more cows ranged on the Strip. Even if it were
opened for entry it would be long before there were sufficient settlers
to take up any great percentage of the range. The settlement of any
country was a slow and tedious process. In any event there were
long years of life in the open—the only sort of existence which they
could endure with satisfaction—stretching forth ahead of them; so
why concern themselves over vague possibilities of the future? That
was the general attitude of them all, excepting old Nate and his
contemporaries, men who, like himself, were being ushered out of
their domain as they had been ushered in a generation past. Their
day was passing and they knew it.
Throughout the following day various wagons turned aside to the
right or left, branching away toward some far spot allotted to them,
there to begin the first actual work. In the late afternoon the Half
Diamond H wagon made its stand on a creek that flowed to the
Cimarron from the low watershed between that stream and the North
Fork of the Canadian. The cook’s summons brought the men
tumbling from their bed rolls an hour before dawn. The night hawk
hazed the remuda into a corral fashioned by a single rope stretched
between stakes sledged solidly into the sod, and after breakfasting
the men entered in pairs, each to rope a circle horse of his own
particular string. In the first light of day Younger led off up a ridge to
the main divide flanking the creek to the left and turned upstream
along it. Other reps had joined the wagon and there were now nearly
twenty riders following where he led. At the head of each draw he
detailed one or two men to work it. When half of the crew had been
assigned to cover certain stretches Younger dropped again to the
bottoms, mounted to the opposite divide and moved downstream in
a similar fashion until even with the wagon, working the last draw
himself.
The riders combed the scrub-oak side hills and the gulches, shoving
all stock before them to the bottoms and heading them upstream.
The first riders to finish their details were stationed across the valley
to halt the cows brought in by others. The chuck wagon had
lumbered on up the creek to the point from which the next circle
would be thrown. The night hawk had gone off duty with sunrise but
the wrangler held the remuda in a rope corral. While a part of the
men held the herd the others repaired to this enclosure and caught
fresh horses, those who were to engage in the next gathering swing
putting their ropes on circle mounts, while those detailed to bring up
the day herd caught trained cow horses belonging to their individual
strings.
In a breeding-ranch country the herd would have been worked on
the spot, calves roped and ironed with the brand worn by their
mothers, and only the beef steers cut into a day herd, the she-stuff
and all stock younger than two-year-olds being allowed to scatter
once more on the range. But there were no calves to brand, no she-
stock on the range, for of late the cowmen of the Strip had come to
follow one set rule in accord with the transition of the cow business,
forming an intermediate link between the old-time cattle kings of the
open range and the modern feeders of the corn belt. For beef
raising, instead of a one-outfit business from start to finish, had come
to be a business of progression induced by the necessities of later-
day conditions. Big breeding ranches were now mainly confined to
the vast wastes of Texas and the Southwest and to similar stretches
in the ranges of the Northwest.
The breeding ranches of Texas and New Mexico now gathered their
steers as two-year-olds and sold them to the intermediate beef-
brands operating in the Strip, the short-grass plains of Western
Kansas and the Sandhill country of Nebraska. Here they were
ranged on grass till they had turned four-year-olds, then resold to the
feeders of Missouri, Iowa and Illinois, who corn-fed and fattened and
finished them for market. Except for one breeding ranch confined to
a great fenced-in pasture, there was only beef stuff in the whole
expanse of the Strip, which rendered the round-up a comparatively
simple affair. This last event in particular was simplified by the orders
which had just gone forth from governmental sources, and every
head of stock gathered in each circle was held in the day herd.
The rope corral was dismantled, ropes and stakes loaded on the bed
wagon which promptly headed up country, trailed by the wrangler
with the remuda, and Carver led all the hands except those detailed
with the day herd up the bottoms toward the new stand of the cook
wagon. It was but ten o’clock when they dropped from their horses
and fell ravenously upon a hot meal which the cook had already
prepared, for while the cowhand’s day begins an hour before dawn
his nooning comes at ten and his knock-off time is seldom later than
five P. M.
The second circle of the day was completed in the late afternoon.
The hands feasted to repletion and lolled about for an hour, buzzing
angrily over a new rumor which had just reached camp. The men
spread their bed rolls on the ground and retired with the setting sun.
Carver dropped instantly asleep but contrary to his usual custom he
waked within an hour and sleep would not come to him as he tossed
restlessly in his blankets. The turmoil of the round-up, the hoarse
bawls emanating from the throats of five hundred steers, the shrill
yelps of riders, the stifling dust of daytime activities; all these had
been superseded by the night sounds of the cow camp in the open.
A cool breeze stole across the range which now seemed
mysteriously hushed. Occasionally some night horse on picket or
tied to the stake ropes shifted uneasily and stamped a restless foot.
The night hawk held the cavayado on good grass somewhere down
the bottoms and his voice drifted faintly to Carver as he sang to while
away the lonely hours. The night guards on duty with the herd were
likewise singing to soothe their charges on the bed ground a few
hundred yards above the wagon, and fragmentary snatches of their
melodies floated down to Carver’s ears as he blinked sleeplessly up
at the stars. He remained awake till the hour came to stand his turn
on second guard and he rolled out, mounted his night horse and
rode with several others to relieve the weary riders who had stood
the first shift of the night after a fourteen-hour day in the saddle.
As Carver circled the bed ground his thoughts were still concerned
with the text of the rumor so recently set afloat. It was said that not
only cows, but men would be ordered from the unowned lands; that
every foot of fence must be removed from the range and brand
owners forced to abandon home ranches. Bart Lassiter joined him.
“Well, what do you think of our latest bit of news?” he asked. “Think
they’ll go through with it?”
“It don’t seem reasonable that they’d put over any such drastic
measure,” Carver said. “They might. It will be hard on the old man if
they do.”
A figure rode toward him in the moonlight and the old man in
question joined him as Lassiter departed. Nate too had been restless
and had found himself unable to sleep. As Carver had reflected that
such a move would inflict an undeserved hardship upon his
employer, so Nate was wondering as to what effect it would have
upon his hands, for in common with all cowmen of his type, Younger
was proud of the accomplishments of his riders.
Every brand owner would stand back of the men who rode for him;
every rider evidenced a similar devotion to the owner’s interests,—a
loyalty to the brand for which he worked. Perhaps in all history there
has never been another calling which has inspired the same
allegiance throughout its entire personnel. A man must be proficient
in many lines to qualify as a cowhand. First of all he must be a
horseman capable of mastering any horse on the range and of
training his mounts to perform the various and intricate duties
required of them; a roper of parts, able to front-foot a calf or to rope
and hog-tie a mighty range bull with equal facility; sufficiently skilled
in blacksmithing to shoe his own horses; for these and many other
acquirements, working at them sixteen hours a day, he was paid a
lesser sum than any unskilled laborer received for ten hours of far
less gruelling work. It was the wild free life, not the pay, which held
him to his chosen calling. The driving spring rains which soaked his
bed roll as he slept on wet ground in the open; the shrivelling heat of
summer and the shrieking blasts of winter blizzards; the congenial
companionship of round-up days and the long lonely vigils at isolated
winter line camps; all these he chose in preference to the softer life
and greater pay of other less strenuous pursuits.
“What will all the boys be doing in another season?” Younger asked.
“Where’ll they all go when there’s no more range work for them to
do?”
“Texas maybe,” Carver predicted. “Or New Mexico.”
“Both those countries are coming to be overrun with nesters,” Nate
returned. “The big brands are getting their range cut up right now.
They’ve been forced to reduce the size of their outfits in proportion to
the decrease in their range. There’s more cowhands down there now
than there are jobs to go around.”
“Then maybe the Northwest range country,” Carver suggested.
“The surplus bronc peelers of Texas and New Mexico have been
drifting up there for the last ten years,” Nate stated. “They’re a drug
on the market right now, cowhands are. And they’re irrigating that
Northwest country rapid and cutting up the range. Once they settle
the Strip, all the boys down there will have to go into other lines.
That’s sure.”
The herd was worked and reworked almost daily as cows wearing
brands that ranged in different parts of the Strip were culled out and
turned over to some wagon crew whose ultimate destination lay in
that direction. All along a two-hundred-mile front more than a score
of wagons were operating in unison. Owners ranging south of the
Strip sent parties up to trail-herd back any of their stock that had
wandered to these parts. These men brought with them little
bunches of Half Diamond H cows and others that had drifted from
the Strip to southern ranges. Some came from beyond the Canadian
and at least one little assortment had been combed from the distant
Washita. Younger, in common with other large owners of his
neighborhood, maintained drift fences and line camps to prevent the
drifting of his stock from the home range. Even with these
precautions there was a certain annual leakage, but the percentage
of Half Diamond H cows gathered south of the Cimarron was small.
Day after day as the round-up progressed the men threshed out the
fate of the unowned lands. It constituted the sole topic of discussion
whenever two riders met on the circle or paused to converse as they
stood their turn on night guard. It filled that brief period of general
indolence in which they indulged each evening before taking to their
beds with the setting sun. Carver, perhaps to a greater extent than
any of them, had anticipated certain transitions. He had correctly
interpreted the presence of those white-topped wagons camped
along the line and knew what they portended, yet even now he found
it impossible to give credence to such drastic changes as were
predicted by old Nate and others of his kind. He sought for an
analogous example and found it in the settling process which
Kansas had been undergoing for a period of forty years; yet
throughout the whole western half of that State ranches of five to fifty
thousand acres were the rule. In view of this circumstance he could
not quite conceive of the vast expanse of the unowned lands being
cut up into quarter sections in the space of a few short years. It
would all take time. He advanced this idea to Younger on a day
some three weeks out from the ranch.
“All this talk about men being ordered out of the Strip,” he said. “How
are they going about that? I’ve seen the squatter outfits rolling up to
the line and making camp. But we’ve had similar demonstrations
before now; that year the boomers fired the grass for one; and
nothing came of it. They were ordered out. Even if they let ’em in it
will take years to settle up the Strip.”
Younger nodded abstractly. Since that event had cost him a
thousand head of cows it was but natural that the incident was still
fresh in his mind. A few years past a swarm of squatters had invaded
the unowned lands in the face of all regulations. When the cowmen
had sought to expel them after they had refused to obey the
government’s order to move out, the boomers had scattered and
fired the parched fall grass and stock had died by thousands
throughout the burned areas. The negro cavalry had been sent in to
enforce the regulations and were thereafter stationed at Caldwell,
patrolling the line and turning back all insistent settlers who would
enter. Now the Negro troopers had been withdrawn and a second
invasion seemed imminent.
“I know, son; but this time things will be reserved,” Nate prophesied.
“Then they ruled the squatters out and stood by us. This time they’ve
ruled us out instead.”
“They’ll open it for entry,” Carver agreed. “It’s come to that and it’s
likely we can’t postpone it. But this notion that the whole of the
Cherokee country will be settled up solid in a few years’ time seems
overdrawn.”
“A few months’ time, boy,” the old man corrected. “More likely a few
weeks will do the trick.”
Carver’s thoughts reverted to a similar prediction made by Hinman,
“It’ll be one hair-raising, mad stampede,” old Joe had said. But
Carver still dissented.
“It’ll take nearly fifty thousand families to file on every quarter section
in the Strip,” he said. “They’ll come eventually. I know that. But
where will that many come from in a few weeks’ time?”
“Son, they’ll come from every odd corner of the country,” Younger
stated. “They’ll swarm in and settle down in clouds like blackbirds in
a cane field. She’ll be the damnedest, wildest scramble a man will
ever live to witness. I’m telling you.”
 V
The stockyards had been the scene of feverish activity for weeks.
The loading pens were crowded to capacity and throughout every
hour of the day and night there sounded the bawls of thirsty cattle
and the shrill yelps of cowhands as they urged unwilling steers
through the loading chutes. Long trainloads of cows rolled out of
Caldwell in swift succession and loading was resumed as soon as
empty cattle cars could be obtained. An antiquated switch engine
wheezed noisily as it shunted cars along the switches and spotted
them at the chutes. Day by day the congestion increased. The
quarantine belt swarmed with stock, as some two hundred thousand
head had been gathered from the Cherokee lands for shipment. In
addition to these the regular run of summer business from the south
continued as the trail herds from Texas and New Mexico came
plodding up to add to the congestion.
Money flowed back into Caldwell in steady streams as trainloads of
cattle were converted into cash on the Kansas City and Chicago
markets. Many owners, having been deprived of their range by the
stringent orders, found themselves unable to reinvest in cows the
funds received from recent shipments. In their restlessness many of
these turned to the green tables for relaxation and there were stud
games where hundreds and often thousands were wagered on the
turn of every card. All the cowhands of the Strip were banked up in
the quarantine belt, holding the cows of their employers on grass
until such time as they could be cleared and shipped. In their leisure
hours they swarmed the streets of Caldwell. Added to these were the
trail-herd crews from the whole Southwest, among them many
Mexican peelers with their tremendous hats, silver-mounted saddles
and three-inch silver rowels.
Four troops of cavalry were camped along the line and troopers
mingled with the crowds. Caldwell, the last of the old-time cow
towns, had now entered upon her last wild fling. It was now definitely
known that in three months’ time the Cherokee Strip would be
thrown open for settlement and the homeless from all corners of the
country were already beginning to assemble. For weeks on end
there was not a room available in town and men spread their
campbeds in vacant lots. Eating places were crowded to capacity
and new restaurants were being opened up in frame shacks or even
in tents wherever vacant sites were available. As always, where
business is rushing and money freely flowing, there were symptoms
of a boom. It was openly predicted that the settling of the country to
the southward would throw Caldwell into the enviable position of the
one logical metropolis of the whole Southwest.
Cowmen cursed the troopers, seeing in them the visible symbol of
that authority which had excluded them from their rightful domain.
The unowned lands were thoroughly patrolled and detachments of
cavalry were camped at strategic points throughout the Strip. It was
this latter circumstance which had upset Carver’s calculations. He
had planned with Bart Lassiter to hold a bunch of six hundred of
Younger’s three-year-old steers on the forbidden range for a period
of one year, receiving a substantial proportion of the increased price
which they would bring as four-year-olds. Both Carver and Nate
Younger had seen the futility of the attempt. Others had entertained
similar ideas but had abandoned them as events moved swiftly past
the farthest bounds of their previous comprehensions and rendered
their hopes untenable.
Carver, once assured that his plans for the immediate future must be
relinquished, cast about for some substitute occupation which might
prove equally remunerative. He rode away from Younger after their
mutual decision, spinning his lone coin into the air and catching it as
his horse jogged slowly across the range.
“It appears as if it’s going to be real difficult to provide you with all the
company I’d counted on,” he said. “Time is skipping right along and
here you are—occupying my pocket all by yourself without even one
mate to jingle up against. Only last week I had it all mapped out to
gather in several thousand of your sort to keep you company. But
that plan’s flown out the window and here I am without one idea to
work on.”
He turned along the south line fence of the Half Diamond H leases.
“Little lonely dollar, you must mount up to a million,” he asserted.
“But we’ve got to insert our wedge somewheres right soon and start
to mounting.”
His eye traveled along the fence line to where it disappeared in the
distance, and suddenly he turned and rode back to where the outfit
was camped and sought out the boss.
“About those fences being ordered down,” he said. “What
arrangements have you made?”
“Not any,” Nate admitted. “What with gathering eight thousand head
of steers and shipping ’em I haven’t taken time off to worry over
fences. We’ll have the last steer headed north in a few days now.
Then I’ll see about scrapping fences—or let the squatters tear ’em
down when they come in to roost.”
“It won’t leave you short-handed now if Bart Lassiter and I lay off,”
Carver suggested. “You lend me a team and wagon from the home
place and we’ll snatch out those fences for what material there is in
them.”
“The fence is yours,” Younger agreed. “Provided the other half-
owners of any stretches are agreeable. Go get it.”
Lassiter assented instantly when Carver stated the proposition.
“I always did feel suffocated in a fence country,” he announced. “I
was always so much opposed to seeing every fence go up that I
figure it will be a real entertaining pastime to help tear ’em down.”
This spirit of optimism lasted during the two days required to hunt up
other part owners of certain stretches and get their endorsement of
the plan, his enthusiasm lasting through the first few days of actual
work. They were out before sunrise and knocked off after dark,
pulling posts, coiling wire and freighting the materials to the Half
Diamond H home ranch. His interest lagged but he did not openly
rebel until after two thirds of the fence had been salvaged. Carver
roused him one morning for breakfast and Bart blinked sleepily at the
smoky lantern that lighted the sod hut in which they had stayed
overnight.
“We’ve got enough wire piled up to enclose the State of Texas,” he
stated. “There’s thirty miles of three-wire fence we’ve collected if
there’s a foot. That’s twenty-nine miles more than both of us will ever
need. Let’s leave the rest of her set.”
“But we contracted to scrap the whole of it,” Carver dissented.
“Another week will see us through.”
“A week!” Lassiter moaned. “I just can’t face it, honest. I’ve reformed.
I hope I hang if I ever extract another staple.”
“A week’s not such a long stretch,” Carver urged.
“Donald, I’ll break down and cry if you lead me up to just one other
measly fence post,” Lassiter announced. “You take my half and let
me off. I’ve got to amble over to Crowfoot’s and draw my spring
wages. Then, too, I’d ought to collect Molly and get her settled
somewhere in Caldwell. She’s all alone over on Turkey Creek.”
“I’ll pay you thirty dollars for what time you’ve put in—sometime
when I’ve got it—and take over your wire and finish the job myself,”
Carver at last conceded. “You can locate Molly in my little plant in
Caldwell; only mark me now! There’ll be no more balancing of
Cherokee rations conducted on the premises. I’ll remonstrate with
you at some length if I catch you at it again.”
Carver worked on alone and at the end of another ten days he
viewed with satisfaction the numerous coils of fence wire and the
great stack of posts neatly corded behind the deserted buildings of
the Half Diamond H.
“At present that assortment is only wood and iron,” he said. “But it’s
a real imposing pile nevertheless, and I can likely convert it into
dollars when the squatters come romping in.”
When he rode into Caldwell he was amazed at the swift transitions.
The incoming transients had trebled the population in the last two
months. Being unprepared for this sweeping change he was all the
more prepared to lend a willing ear to the prediction that Caldwell
was to become the metropolis of the whole Southwest. There was a
conversational boom in progress and Carver, looking upon the
crowded, teeming streets, the numerous tent houses everywhere in
evidence and the new frame shacks in the process of construction
through the town, divined the possibility of actual boom days just
ahead. He rode out to his little frame cabin to visit with Molly Lassiter
whom he had seen but three times in as many months. He found
neither Bart nor Molly at home but the door was unlocked and he
entered.
The two rooms of the bare little shack had been transformed. Two
worn Navajo rugs were spread on the pine-board floor and soft
curtain materials were draped across the windows.
“She’s made it all homelike,” Carver said. “Just with a touch here and
there. What couldn’t she do with things to work with and a real house
to operate on? We’ll give her one some day if only she’ll agree.” He
drew forth the lucky dollar and consulted it. “Let’s you and me hatch
out a new idea,” he invited. “We can’t be loafing on the job.”
While the idea was hatching he sat peering abstractedly through the
doorway, rousing from his reverie only when he found his gaze
riveted on the girl as she turned into the pathway leading to the
house. Molly halted suddenly when within a few feet of the door, as
she saw him sitting just inside it.
“I hadn’t expected you this soon,” she said.
“Bart told me the fence job would keep you another month at least.
Did you decide not to finish it?”
“It’s salvaged to the last strand of wire,” he returned. “I speeded up
some so as to have it over with.”
“I’m sorry Bart quit,” she said. “You see he won’t stick at anything.”
“Don’t know as I blame him,” said Carver. “The last few days I’ve
developed a downright aversion to the sight of fence wire myself.
Glad to see me?”
“Yes,” she said. “I’ll be out of here and established in a room of my
own so that you can have your house by to-morrow, Don. I’ve been
waiting for the present occupants to vacate.”
“You stay right on here,” he insisted. “I won’t be needing it.”
“Thanks, Don, but I can’t do that,” she said. “I have to stop floating
and find some nook of my own. I can’t follow Bart around any longer.
For three years now we’ve drifted from one spot to the next;
sometimes in line camps; more frequently in some rooming house in
any town where we happened to be, always knowing that wherever it
was it wouldn’t be home for long. I didn’t mind at first, for I was trying
to keep Bart away from Milt and Noll; but they always turn up again
and he follows them off. I’d love even a sod house if only I could call
it my own and know I wouldn’t have to move out on an hour’s notice.
I’m sick of gypsying. I want to feel settled—feel that I’m attached!”
He reached over and rested a hand on her shoulder.
“I know, Honey,” he said. “So do I. That’s exactly my own frame of
mind. The best way all round is for you and me to get attached and
settle. Won’t you?”
She felt that he had failed to grasp the fact of what a sense of
permanency would mean to her after the nomadic existence she had
followed for the past few years.
“Listen, Molly,” he said, divining something of her thoughts. “It’s not
the way a man says a thing but the way he means it that really
counts. And I was meaning that a lot.”
“But you don’t even know to-day what you’re going to do to-morrow,”
she said. “It would be only exchanging one state of gypsying for
another. Don’t you see that?”
He did, at least, see that the moment was not right and he settled
back into his chair and twisted a cigarette.
“You always lean to the dark side of things,” he accused. “Most
complaints I’ve heard about family strife was occasioned because
menfolks generally were so occupied with business that they didn’t
spend much time at home. Now with me not having any special
business it would leave me free to put in most of my time around the
house. There’s that advantage.”
“Yes,” she laughed. “There’s that. Sometimes, Don, I almost wish
you really were a settled sort of a soul; but that time will never be.”
Carver crooned softly:

“Oh, I’m a rolling rambler,”

Said the speeding tumbleweed.
“The prairies are my race track,
The wild wind is my steed.

“I never cease my roaming;

I’m always hard to catch,
But the pumpkin stays forever
In the same old garden patch.

“But I’d rather be a wild, wild weed

Than a sluggish yellow squash:”

“And I’d so much rather be a pumpkin than the wildest of all wild
weeds,” she said. “There’s only that little difference between the two
of us.”
“Tell me,” he urged, “what sort of a quiet home life do you pine for
most? Does your preference run to a cottage in town or stray off
towards a dwelling in the country?”
“The country,” she returned. “Somewhere on a farm where I could
watch things grow.”
“That’s my choice too,” he confessed. “Whatever business I settle on
will have to be at the source of things. Like you said, I want to watch
things grow—calves or crops, it don’t much matter which. I’ll start
casting about for a farm right off.”
After leaving her he mingled with the swarming crowds on the main
street. The conversational boom was in full swing and he heard it
discussed on all sides. There were but few who dissented from the
general prediction that an era of great prosperity lay ahead for
Caldwell. Carver put in three active hours, then sought out Nate
Younger to draw his back wages for the spring work, a sum totalling
a trifle less than two hundred dollars.
He found Younger in his room at the hotel in conference with Joe
Hinman. The two old cowmen had pooled resources and formed the
Plains Land and Cattle Company, Younger having purchased
grasslands adjoining Hinman’s holdings. They planned to make the
new concern a beef ranch straight through instead of a breeding
ranch as now operated by Hinman.
“We’ll be the biggest outfit in this end of the State,” Hinman was
predicting, as Carver thrust his head through the door. “Come in,
son, and set on the bed. The Plains Land and Cattle Company is
going to be the biggest of the lot.”
“I’m counting some on organizing a similar concern myself,” said
Carver. “Maybe a trifle smaller than yours just at first; and in order to
make the start I’ve got to borrow somewhat. I’m owning a nine-
hundred-dollar equity in that bunch of calves we made the deal for
last spring. How about your lending me eight hundred against it?”
“But that would leave you owing me thirteen hundred on the bunch,”
Hinman objected. “And right now those calves wouldn’t fetch that
price on the market.”
“Set the date for maturity of the loan far enough ahead so they’ll
grow into it,” Carver suggested. “Before it comes due they’ll have
advanced way beyond that figure. Then if I don’t pay up you can
close me out at a profit.”
“Now ain’t that a fact!” Hinman exclaimed admiringly. “There was a
time, Buddy, when I marvelled at your ability to shed a season’s
wages overnight. It does look now as if you might also learn me a
few tricks on the reverse side of things. You’ve got a business mind.”
He produced a check book and a stub of pencil.
“How long do you want this loan to run?” Carver asked.
“According to your own figures the longer it runs the more I stand to
make,” said Hinman. “So I don’t know as it makes much difference. It
does appear as if you’d let me in on a pretty good thing—so set the
date yourself.”
“One year from to-day,” Carver decided.
“What do you aim to do with all this money?” Hinman inquired.
“Setting out to break the bank in the Gilded Eagle?”
“I’ve purchased a building,” Carver proclaimed.
“You’ve which?” said Hinman. “What building?”
“Pirie’s place; down in the next block,” Carver informed. “It’s got a
grocery business on the ground floor and the grocer’s wife rents
room upstairs.”
He extended a contract and Hinman perused it, observing that
Carver had agreed to purchase at three thousand dollars, paying six
hundred down and a like amount each year.
“I’d rented my little shack,” Carver explained. “Only to find that there
wasn’t a room for rent in town; not one! It was either buy a place of
my own or set up.”
“It’ll save you considerable room rent,” Hinman agreed, “you being in
town easy three nights out of the year. But what’s the final object?”
“Each season those calves will be worth more and I can borrow
enough additional against them to meet the payments,” Carver
pointed out. “Meantime the grocer pays me thirty dollars rent money
every month, which gives me a steady income to live off till such time
as I can turn the building at a profit and buy a tract of land to run
those calves on.”
“I didn’t know your ambitions run toward owning land,” said Hinman.
“But now since I’ve come into so much surplus fence wire,” Carver
explained, “it looks like the only economical thing to do is to acquire
a piece of land to set inside it.”
“Son, you’ve mapped out a self-operating business,” Nate Younger
congratulated. “All you have to do now is to stand back and watch it
ripen. Meantime why don’t you read up on Belgian hares?”
He handed over the sum due for back wages and Carver studied the
two checks reflectively.
“This surplus now,” he said. “I was figuring to put into horses. They’ll
almost give you horses nowadays just to come and drive them off. If
you don’t mind my throwing a few head up on your range, I’ll buy up
a little bunch and pay you fifty cents a head for pasture fees,
agreeing to get ’em off your grass November first.”
“We’d better let him put ’em on, Joe,” Nate agreed. “It’s that much
more security for that loan.”
Even under favorable circumstances the horse market was poor and
now with all those recently combed from the Strip as a surplus,
horses could be purchased at one’s own price. For a week Carver
rode early and late. The average run of Indian ponies were selling for
less than five dollars a head but it was not this class of horse flesh
which Carver sought. He selected young mares and geldings,
ranging from eleven to twelve hundred pounds in weight, which
would serve for light work stock, and eventually he drove fifty head
well toward the northern extremity of Hinman’s range. They had cost
him an average of ten dollars apiece and he had paid cash for half of
them, issuing verbal promises to pay for the rest. He rode back into
Caldwell with something over a hundred dollars in his pocket.
The equipment of all the deserted ranches in the unowned lands was
banked up in Caldwell. From the Coldstream Pool Carver purchased
ten sets of harness at fifteen dollars a set and three heavy wagons at
forty dollars each, paying his last hundred down and his personal
note for the balance.
Hinman witnessed this last transaction.
“Considering the size of your original stake you’ve stretched it to
cover considerable territory in the last few months,” he said.
“It’s only my surplus I’m spreading out so thin,” Carver explained.
“My capital is still intact.” He exhibited his silver dollar. “My one rule
of life is never to impair my principal.”
“Fine,” Hinman encouraged. “That’s conservative business. I was
satisfied you’d play it slow and safe.”
“Now if you’ll do me just one more little kindness I’ll be grateful;”
Carver said. “You and Nate engage Freel in conversation up on the
corner where he’s standing and inside of five minutes I’ll saunter up
and direct the course of the interview.”
“I’d like to hear it,” Hinman said. “We’ll detain him.”
Carver joined them before the appointed space of time had elapsed.
“Freel, I’ve been feeling real contrite about resisting arrest a few
weeks back,” Carver said. “I’ve decided to surrender and stand trial.”
The deputy marshal glanced apprehensively at the two old cowmen.
“Oh—that,” he said. “Why, I’ve let that matter drop. That’s all closed.”
“And it was real accommodating of you to close it,” Carver returned,
“but I can’t stand by and see you get in trouble on my account.
Orders are orders, and you had yours. That’s the reason I wrote this
letter to Art Webb.” He tendered an unsealed letter to the deputy.
Webb was Freel’s chief, the head United States marshal of the
district. “Webb is a good friend of mine and I’m demanding that he
inform me just why he sent an order down here to you to pick me up.
That will put you in the clear for not rearresting me since that night I
escaped.”
Carver turned to his two friends.
“You’ve both known Webb for years,” he said. “Did you write him like
I asked you?”
“It clear slipped my mind,” Hinman apologized. “I’ll get it off this
evening.”
“Mine goes on the same mail,” Nate concurred. “We’ll sift this thing
right to the bottom layer and clear Freel of any possible blame.”
“Freel will be on my side himself if it comes to a showdown,” Carver
asserted. “He’ll be the first to testify that I’d been away from home for
a solid month prior to the time that charge was lodged. Some one’s
tried to deal me from the bottom, and between the four of us we’ll
discover who it is.”
Freel laughed and slapped Carver on the shoulder.