Get the full ebook with Bonus Features for a Better Reading Experience on ebookluna.

com

Parallel programming: concepts and practice 1st

Edition - eBook PDF

https://ebookluna.com/download/parallel-programming-
concepts-and-practice-ebook-pdf/

OR CLICK HERE

DOWLOAD NOW

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

 Instant digital products (PDF, ePub, MOBI) ready for you
Download now and discover formats that fit your needs...

An Introduction to Parallel Programming 2. Edition Pacheco

- eBook PDF

https://ebookluna.com/download/an-introduction-to-parallel-
programming-ebook-pdf/

ebookluna.com

(eBook PDF) Parallel Computer Organization and Design

https://ebookluna.com/product/ebook-pdf-parallel-computer-
organization-and-design/

ebookluna.com

(eBook PDF) Parallel Computer Organization and Design

https://ebookluna.com/product/ebook-pdf-parallel-computer-
organization-and-design-2/

ebookluna.com

Data Science: Concepts and Practice 2nd Edition- eBook PDF

https://ebookluna.com/download/data-science-concepts-and-practice-
ebook-pdf/

ebookluna.com
Problem Solving and Python Programming 1st edition - eBook
PDF

https://ebookluna.com/download/problem-solving-and-python-programming-
ebook-pdf/

ebookluna.com

(eBook PDF) Community and Human Services Concepts for

Practice

https://ebookluna.com/product/ebook-pdf-community-and-human-services-
concepts-for-practice/

ebookluna.com

(eBook PDF) Introduction to Leadership: Concepts and

Practice 4th Edition

https://ebookluna.com/product/ebook-pdf-introduction-to-leadership-
concepts-and-practice-4th-edition/

ebookluna.com

Introduction to Leadership: Concepts and Practice 4th

Edition (eBook PDF)

https://ebookluna.com/product/introduction-to-leadership-concepts-and-
practice-4th-edition-ebook-pdf/

ebookluna.com

(eBook PDF) Professional Nursing Practice: Concepts and

Perspectives 7th Edition

https://ebookluna.com/product/ebook-pdf-professional-nursing-practice-
concepts-and-perspectives-7th-edition/

ebookluna.com
Parallel Programming
Parallel Programming
Concepts and Practice
Bertil Schmidt
Institut für Informatik
Staudingerweg 9
55128 Mainz
Germany

Jorge González-Domínguez
Computer Architecture Group
University of A Coruña
Edificio área científica (Office 3.08), Campus de Elviña
15071, A Coruña
Spain

Christian Hundt
Institut für Informatik
Staudingerweg 9
55128 Mainz
Germany

Moritz Schlarb
Data Center
Johannes Gutenberg-University Mainz
Germany
Anselm-Franz-von-Bentzel-Weg 12
55128 Mainz
Germany
Morgan Kaufmann is an imprint of Elsevier
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States
Copyright © 2018 Elsevier Inc. All rights reserved.
No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including
photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on
how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as
the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted
herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes
in research methods, professional practices, or medical treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information,
methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety
and the safety of others, including parties for whom they have a professional responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or
damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods,
products, instructions, or ideas contained in the material herein.

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

ISBN: 978-0-12-849890-3

For information on all Morgan Kaufmann publications

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

Publisher: Katey Birtcher

Acquisition Editor: Steve Merken
Developmental Editor: Nate McFadden
Production Project Manager: Sreejith Viswanathan
Designer: Christian J. Bilbow
Typeset by VTeX
Preface

Parallelism abounds. Nowadays, any modern CPU contains at least two cores, whereas some CPUs
feature more than 50 processing units. An even higher degree of parallelism is available on larger sys-
tems containing multiple CPUs such as server nodes, clusters, and supercomputers. Thus, the ability
to program these types of systems efficiently and effectively is an essential aspiration for scientists,
engineers, and programmers. The subject of this book is a comprehensive introduction to the area of
parallel programming that addresses this need. Our book teaches practical parallel programming for
shared memory and distributed memory architectures based on the C++11 threading API, Open Mul-
tiprocessing (OpenMP), Compute Unified Device Architecture (CUDA), Message Passing Interface
(MPI), and Unified Parallel C++ (UPC++), as well as necessary theoretical background. We have in-
cluded a large number of programming examples based on the recent C++11 and C++14 dialects of
the C++ programming language.
This book targets participants of “Parallel Programming” or “High Performance Computing”
courses which are taught at most universities at senior undergraduate level or graduate level in com-
puter science or computer engineering. Moreover, it serves as suitable literature for undergraduates in
other disciplines with a computer science minor or professionals from related fields such as research
scientists, data analysts, or R&D engineers. Prerequisites for being able to understand the contents
of our book include some experience with writing sequential code in C/C++ and basic mathematical
knowledge.
In good tradition with the historic symbiosis of High Performance Computing and natural science,
we introduce parallel concepts based on real-life applications ranging from basic linear algebra rou-
tines over machine learning algorithms and physical simulations but also traditional algorithms from
computer science. The writing of correct yet efficient code is a key skill for every programmer. Hence,
we focus on the actual implementation and performance evaluation of algorithms. Nevertheless, the
theoretical properties of algorithms are discussed in depth, too. Each chapter features a collection of
additional programming exercises that can be solved within a web framework that is distributed with
this book. The System for Automated Code Evaluation (SAUCE) provides a web-based testing en-
vironment for the submission of solutions and their subsequent evaluation in a classroom setting: the
only prerequisite is an HTML5 compatible web browser allowing for the embedding of interactive
programming exercise in lectures. SAUCE is distributed as docker image and can be downloaded at
https://parallelprogrammingbook.org
This website serves as hub for related content such as installation instructions, a list of errata, and
supplementary material (such as lecture slides and solutions to selected exercises for instructors).
If you are a student or professional that aims to learn a certain programming technique, we advise to
initially read the first three chapters on the fundamentals of parallel programming, theoretical models,
and hardware architectures. Subsequently, you can dive into one of the introductory chapters on C++11
Multithreading, OpenMP, CUDA, or MPI which are mostly self-contained. The chapters on Advanced
C++11 Multithreading, Advanced CUDA, and UPC++ build upon the techniques of their preceding
chapter and thus should not be read in isolation.
ix
x Preface

If you are a lecturer, we propose a curriculum consisting of 14 lectures mainly covering applications
from the introductory chapters. You could start with a lecture discussing the fundamentals from the
first chapter including parallel summation using a hypercube and its analysis, the definition of basic
measures such as speedup, parallelization efficiency and cost, and a discussion of ranking metrics. The
second lecture could cover an introduction to PRAM, network topologies, weak and strong scaling.
You can spend more time on PRAM if you aim to later discuss CUDA in more detail or emphasize
hardware architectures if you focus on CPUs. Two to three lectures could be spent on teaching the
basics of the C++11 threading API, CUDA, and MPI, respectively. OpenMP can be discussed within
a span of one to two lectures. The remaining lectures can be used to either discuss the content in the
advanced chapters on multithreading, CUDA, or the PGAS-based UPC++ language.
An alternative approach is splitting the content into two courses with a focus on pair-programming
within the lecture. You could start with a course on CPU-based parallel programming covering selected
topics from the first three chapters. Hence, C++11 threads, OpenMP, and MPI could be taught in full
detail. The second course would focus on advanced parallel approaches covering extensive CUDA
programming in combination with (CUDA-aware) MPI and/or the PGAS-based UPC++.
We wish you a great time with the book. Be creative and investigate the code! Finally, we would be
happy to hear any feedback from you so that we could improve any of our provided material.
Acknowledgments

This book would not have been possible without the contributions of many people.
Initially, we would like to thank the anonymous and few non-anonymous reviewers who com-
mented on our book proposal and the final draft: Eduardo Cesar Galobardes, Ahmad Al-Khasawneh,
and Mohammad Olaimat.
Moreover, we would like to thank our colleagues who thoroughly peer-reviewed the chapters and
provided essential feedback: André Müller for his valuable advise on C++ programming, Robin Kobus
for being a tough code reviewer, Felix Kallenborn for his steady proofreading sessions, Daniel Jünger
for constantly complaining about the CUDA chapter, as well as Stefan Endler and Elmar Schömer for
their suggestions.
Additionally, we would like to thank the staff of Morgan Kaufman and Elsevier who coordinated
the making of this book. In particular we would like to mention Nate McFadden.
Finally, we would like to thank our spouses and children for their ongoing support and patience
during the countless hours we could not spend with them.

xi
 CHAPTER

INTRODUCTION

Abstract
1
In the recent past, teaching and learning of parallel programming has become increasingly important
due to the ubiquity of parallel processors in portable devices, workstations, and compute clusters. Stag-
nating single-threaded performance of modern CPUs requires future computer scientists and engineers
to write highly parallelized code in order to fully utilize the compute capabilities of current hardware
architectures. The design of parallel algorithms, however, can be challenging especially for inexpe-
rienced students due to common pitfalls such as race conditions when concurrently accessing shared
resources, defective communication patterns causing deadlocks, or the non-trivial task of efficiently
scaling an application over the whole number of available compute units. Hence, acquiring parallel
programming skills is nowadays an important part of many undergraduate and graduate curricula.
More importantly, education of concurrent concepts is not limited to the field of High Performance
Computing (HPC). The emergence of deep learning and big data lectures requires teachers and stu-
dents to adopt HPC as an integral part of their knowledge domain. An understanding of basic concepts
is indispensable for acquiring a deep understanding of fundamental parallelization techniques.
The goal of this chapter is to provide an overview of introductory concepts and terminologies in parallel
computing. We start with learning about speedup, efficiency, cost, scalability, and the computation-to-
communication ratio by analyzing a simple yet instructive example for summing up numbers using a
varying number of processors. We get to know about the two most important parallel architectures:
distributed memory systems and shared memory systems. Designing efficient parallel programs re-
quires a lot of experience and we will study a number of typical considerations for this process such
as problem partitioning strategies, communication patterns, synchronization, and load balancing. We
end this chapter with learning about current and past supercomputers and their historical and upcoming
architectural trends.

Keywords
Parallelism, Speedup, Parallelization, Efficiency, Scalability, Reduction, Computation-to-communica-
tion ratio, Distributed memory, Shared memory, Partitioning, Communication, Synchronization, Load
balancing, Task parallelism, Prefix sum, Deep learning, Top500

CONTENTS
1.1 Motivational Example and Its Analysis ............................................................................ 2
The General Case and the Computation-to-Communication Ratio..................................... 8
1.2 Parallelism Basics .................................................................................................... 10
Distributed Memory Systems................................................................................ 10
Shared Memory Systems..................................................................................... 11

Parallel Programming. DOI: 10.1016/B978-0-12-849890-3.00001-0

Copyright © 2018 Elsevier Inc. All rights reserved.
1
2 CHAPTER 1 INTRODUCTION

Considerations When Designing Parallel Programs ...................................................... 13

1.3 HPC Trends and Rankings ........................................................................................... 16
1.4 Additional Exercises.................................................................................................. 18

1.1 MOTIVATIONAL EXAMPLE AND ITS ANALYSIS

In this section we learn about some basic concepts and terminologies. They are important for analyzing
parallel algorithms or programs in order to understand their behavior. We use a simple example for
summing up numbers using an increasing number of processors in order to explain and apply the
following concepts:

• Speedup. You have designed a parallel algorithm or written a parallel code. Now you want to
know how much faster it is than your sequential approach; i.e., you want to know the speedup.
The speedup (S) is usually measured or calculated for almost every parallel code or algorithm and
is simply defined as the quotient of the time taken using a single processor (T (1)) over the time
measured using p processors (T (p)) (see Eq. (1.1)).
T (1)
S= (1.1)
T (p)
• Efficiency and cost. The best speedup you can usually expect is a linear speedup; i.e., the maximal
speedup you can achieve with p processors or cores is p (although there are exceptions to this,
which are referred to as super-linear speedups). Thus, you want to relate the speedup to the number
of utilized processors or cores. The Efficiency E measures exactly that by dividing S by P (see
Eq. (1.2)); i.e., linear speedup would then be expressed by a value close to 100%. The cost C is
similar but relates the runtime T (p) (instead of the speedup) to the number of utilized processors
(or cores) by multiplying T (p) and p (see Eq. (1.3)).
S T (1)
E= = (1.2)
p T (p) × p
C = T (p) × p (1.3)

• Scalability. Often we do not only want to measure the efficiency for one particular number of pro-
cessors or cores but for a varying number; e.g. P = 1, 2, 4, 8, 16, 32, 64, 128, etc. This is called
scalability analysis and indicates the behavior of a parallel program when the number of processors
increases. Besides varying the number of processors, the input data size is another parameter that
you might want to vary when executing your code. Thus, there are two types of scalability: strong
scalability and weak scalability. In the case of strong scalability we measure efficiencies for a vary-
ing number of processors and keep the input data size fixed. In contrast, weak scalability shows the
behavior of our parallel code for varying both the number of processors and the input data size; i.e.
when doubling the number of processors we also double the input data size.
• Computation-to-communication ratio. This is an important metric influencing the achievable
scalability of a parallel implementation. It can be defined as the time spent calculating divided by
the time spent communicating messages between processors. A higher ratio often leads to improved
speedups and efficiencies.
 1.1 MOTIVATIONAL EXAMPLE AND ITS ANALYSIS 3

The example we now want to look at is a simple summation; i.e., given an array A of n numbers we
want to compute n−1
i=0 A[i]. We parallelize this problem using an array of processing elements (PEs).
We make the following (not necessarily realistic) assumptions:

• Computation. Each PE can add two numbers stored in its local memory in one time unit.
• Communication. A PE can send data from its local memory to the local memory of any other PE
in three time units (independent of the size of the data).
• Input and output. At the beginning of the program the whole input array A is stored in PE #0. At
the end the result should be gathered in PE #0.
• Synchronization. All PEs operate in lock-step manner; i.e. they can either compute, communicate,
or be idle. Thus, it is not possible to overlap computation and communication on this architecture.

Speedup is relative. Therefore, we need to establish the runtime of a sequential program first. The
sequential program simply uses a single processor (e.g. PE #0) and adds the n numbers using n − 1
additions in n − 1 time units; i.e. T (1, n) = n − 1. In the following we illustrate our parallel algorithm
for varying p, where p denotes the number of utilized PEs. We further assume that n is a power of 2;
i.e., n = 2k for a positive integer k.

• p = 2. PE #0 sends half of its array to PE #1 (takes three time units). Both PEs then compute the
sum of their respective n/2 numbers (takes time n/2 − 1). PE #1 sends its partial sum back to PE
#0 (takes time 3). PE #0 adds the two partial sums (takes time 1). The overall required runtime is
T (2, n) = 3 + n/2 − 1 + 3 + 1. Fig. 1.1 illustrates the computation for n = 1024 = 210 , which has
a runtime of T (2, 1024) = 3 + 511 + 3 + 1 = 518. This is significantly faster than the sequential
runtime. We can calculate the speedup for this case as T (1, 1024)/T (2, 1024) = 1023/518 = 1.975.
This is very close to the optimum of 2 and corresponds to an efficiency of 98.75% (calculated
dividing the speedup by the number of utilized PEs; i.e. 1.975/2).
• p = 4. PE #0 sends half of the input data to PE #1 (takes time 3). Afterwards PE #0 and PE #1
each send a quarter of the input data to PE #2 and PE #3 respectively (takes time 3). All four PEs
then compute the sum of their respective n/4 numbers in parallel (takes time n/4 − 1). PE #2 and
PE #3 send their partial sums to PE #0 and PE #1, respectively (takes time 3). PE #0 and PE #1
add their respective partial sums (takes time 1). PE #1 then sends its partial sum to PE #0 (takes
time 3). Finally, PE #0 adds the two partial sums (takes time 1). The overall required runtime is
T (4, n) = 3 + 3 + n/4 − 1 + 3 + 1 + 3 + 1. Fig. 1.2 illustrates the computation for n = 1024 = 210 ,
which has a runtime of T (4, 1024) = 3 + 3 + 255 + 3 + 1 + 3 + 1 = 269. We can again calculate the
speedup for this case as T (1, 1024)/T (4, 1024) = 1023/269 = 3.803 resulting in an efficiency of
95.07%. Even though this value is also close to 100%, it is slightly reduced in comparison to p = 2.
The reduction is caused by the additional communication overhead required for the larger number
of processors.
• p = 8. PE #0 sends half of its array to PE #1 (takes time 3). PE #0 and PE #1 then each send a
quarter of the input data to PE #2 and PE #3 (takes time 3). Afterwards, PE #0, PE #1, PE #2, and
PE #3 each send a 1/8 of the input data to PE #5, PE #6, PE #7, and PE #8 (takes again time 3).
Fig. 1.3 illustrates the three initial data distribution steps for n = 1024 = 210 . All eight PEs then
compute the sum of their respective n/8 numbers (takes time n/8 − 1). PE #5, PE #6, PE #7, and
PE #8 send their partial sums to PE #0, PE #1, PE #2, and PE #3, respectively (takes time 3).
4 CHAPTER 1 INTRODUCTION

FIGURE 1.1
Summation of n = 1024 numbers on p = 2 PEs: (A) initially PE #0 stores the whole input data locally; (B) PE #0
sends half of the input to PE #1 (takes time 3); (C) Each PE sums up its 512 numbers (takes time 511);
(D) PE #1 sends its partial sum back to PE #0 (takes time 3); (E) To finalize the computation, PE #0 adds the
two partial sums (takes time 1). Thus, the total runtime is T (2, 1024) = 3 + 511 + 3 + 1 = 518.

Subsequently, PE #0, PE #1, PE #2, and PE #3 add their respective partial sums (takes time 1). PE
#2 and PE #3 then send their partial sums to PE #0 and PE #1, respectively (takes time 3). PE #0
and PE #1 add their respective partial sums (takes time 1). PE #1 then sends its partial sum to PE #0
(takes time 3). Finally, PE #0 adds the two partial sums (takes time 1). The overall required runtime
is T (8, n) = 3 + 3 + 3 + n/8 − 1 + 3 + 1 + 3 + 1 + 3 + 1. The computation for n = 1024 = 210
thus has a runtime of T (8, 1024) = 3 + 3 + 3 + 127 + 3 + 1 + 3 + 1 + 3 + 1 = 148. The speedup
for this case is T (1, 1024)/T (8, 1024) = 1023/148 = 6.91 resulting in an efficiency of 86%. The
decreasing efficiency is again caused by the additional communication overhead required for the
larger number of processors.

We are now able to analyze the runtime of our parallel summation algorithm in a more general way
using p = 2q PEs and n = 2k input numbers:

• Data distribution time: 3 × q.

• Computing local sums: n/p − 1 = 2k−q − 1.
• Collecting partial results: 3 × q.
• Adding partial results: q.
 1.1 MOTIVATIONAL EXAMPLE AND ITS ANALYSIS 5

FIGURE 1.2
Summation of n = 1024 numbers on p = 4 PEs: (A) initially PE #0 stores the whole input in its local memory;
(B) PE #0 sends half of its input to PE #1 (takes time 3); (C) PE #0 and PE #1 send half of their data to PE #2
and PE #3 (takes time 3); (D) Each PE adds its 256 numbers (takes time 255); (E) PE #2 and PE #3 send their
partial sums to PE #0 and PE #1, respectively (takes time 3). Subsequently, PE #0 and PE #1 add their
respective partial sums (takes time 1); (F) PE #1 sends its partial sum to PE #0 (takes time 3), which then
finalizes the computation by adding them (takes time 1). Thus, the total runtime is
T (4, 1024) = 3 + 3 + 511 + 3 + 1 + 3 + 1 = 269.
6 CHAPTER 1 INTRODUCTION

FIGURE 1.3
The three initial data distribution steps for n = 1024 and p = 8: (A) Initially PE #0 stores the whole input in its
local memory and sends half of its input to PE #1; (B) PE #0 and PE #1 send half of their (remaining) data to
PE #2 and PE #3; (C) PE #0, PE #1, PE #2, and PE #3 each send half of their (remaining) input data to PE #5,
PE #6, PE #7, and PE #8.

Thus, we get the following formula for the runtime:

T (p, n) = T (2q , 2k ) = 3q + 2k−q − 1 + 3q + q = 2k−q − 1 + 7q. (1.4)

Fig. 1.4 shows the runtime, speedup, cost, and efficiency of our parallel algorithm for n = 1024
and p ranging from 1 to 512. This type of runtime analysis (where the input size is kept constant
and the number of PEs is scaled) is called strong scalability analysis. We can see that the efficiency
 1.1 MOTIVATIONAL EXAMPLE AND ITS ANALYSIS 7

FIGURE 1.4
Strong scalability analysis: runtime, speedup, cost, and efficiency of our parallel summation algorithm for
adding n = 1024 numbers on a varying number of PEs (ranging from 1 to 512).

is high for a small number of PEs (i.e. p  n), but is low for a large number of PEs (i.e. p ≈ n).
This behavior can also be deduced from Eq. (1.4): for the case p  n, holds 2k−q  7q (i.e., the
term for computation time dominates), while it holds 2k−q  7q for the case p ≈ n (i.e., the term
for communication time dominates). Thus, we can conclude that our algorithm is not strongly scal-
able.
Now, we want to change our analysis a bit by not only increasing the number of PEs but additionally
increasing the input data size at the same time. This is known as weak scalability analysis. Fig. 1.5
shows the speedup and efficiency of our algorithm for n ranging from 1024 to 524,288 and p ranging
from 1 to 512. We can see that the efficiency is kept high (close to 100%) even for a large number of
PEs. This behavior can again be deduced from Eq. (1.4): since both n and p are scaled at the same
rate, the term relating to the computation time is constant for varying number of PEs (i.e. 2k−q = 1024
in Fig. 1.5), while the term for the communication time (7q = 7 × log(p)) only grows at a logarithmic
rate. Thus, we can conclude that our algorithm is weakly scalable.
The terms weak and strong scalability are also related to two well-known laws in parallel comput-
ing: Amdahl’s law and Gustafsson’s law, which we will discuss in more detail in Chapter 2.
Other documents randomly have
different content
“lightning expresses,” and to stop them at every little wayside station. The
engines were fed upon wood, and it was a common thing for trains to stop
their intolerable jolting for full twenty minutes to take fresh supplies of
wood and water.
There was immeasurably more of weariness then, in a journey from
Richmond or Cincinnati or Buffalo to New York, than would be tolerated
now in a trip across the continent. As a consequence few people travelled
except for short distances and a journey which we now think nothing of
making comfortably in a single night, was then a matter of grave
consequence, to be undertaken only after much deliberation and with much
of preparation. New York seemed more distant to the dweller in the West or
South than Hong Kong and Yokohama do in our time, and the number of
people who had journeyed beyond the borders of our own country was so
small that those who had done so were regarded as persons of interestingly
adventurous experience.
Quite necessarily all parts of the country were markedly provincial in
speech, manner, habits and even in dress. New England had a nasal dialect
of its own, so firmly rooted in use that it has required two or three
generations of exacting Yankee school marms to eradicate it from the
speech even of the educated class. New York state had another, and the
Southerner was known everywhere by a speech which “bewrayed” him.
And as it was with speech, so also was it with manners, customs, ideas.
Prejudice was everywhere rampant, opinion intolerant, and usage merciless
in its narrow illiberality. Only in what was then the West—the region
between the Alleghenies and the Mississippi—was there anything of
cosmopolitan liberality and tolerance. They were found in that region
because the population of the West had been drawn from all parts of the
country. Attrition of mind with mind, and the mingling of men of various
origin had there in a large degree worn off the angles of provincial prejudice
and bred a liberality of mind elsewhere uncommon in our country.
Edmonia and Dorothy were to make their formidable journey to New
York in easy stages. They remained for several days with friends in
Richmond, while completing those preliminary preparations which were
necessary before setting out for the national capital. They were to stay in
Washington for a fortnight, in Baltimore for three weeks, in Philadelphia for
a week or two, and in New York for nearly two months before sailing for
Europe in May.
The time was a very troubled one, on the subject of slavery. Not only
was it true that if the owner of a slave took the negro with him into any one
of the free states he by that act legally set him free, but it was also true that
the most devoted and loyal servant thus taken from the South into a
Northern state was subjected to every form of persuasive solicitation to
claim and assert his freedom. It was nevertheless the custom of Southern
men, and still more of Southern women, to take with them on their travels
one or more of their personal servants, trusting to their loyalty alone for
continued allegiance. For the attitude of such personal servants in Virginia
at that time was rather that of proud and voluntary allegiance to loved
masters and mistresses who belonged to them as a cherished possession,
than that of men and women held in unwilling bondage.
Accordingly it was arranged that Edmonia’s maid, Dinah—or Diana as
she had come to call herself since hearing her mistress read a “history
pome” aloud—should accompany the two young women as their joint
servitor.
As soon as this arrangement was announced at Branton, Diana began
what Polydore called “a puttin’ on of airs.” In plainer phrase she began to
snub Polydore mercilessly, whereas she had recently been so gracious in her
demeanor towards him as to give him what he called “extinct
discouragement.”
After it was settled that she was to accompany “Miss Mony an’ Miss
Dorothy” to “de Norf” and to “Yurrop”—as she wrote to all her friends who
were fortunate enough to know how to “read writin’,” there was, as
Polydore declared, “no livin’ in de house wid her.” She sailed about the
place like a frigate, delivering her shots to the right and left—most of them
aimed at Polydore, with casual and contemptuous attention, now and then,
to the other house servants.
“I ’clar’ to gracious,” said Elsie, one of the housemaids, “ef Diana ain’t a
puttin’ on of jes’ as many airs as ef she’d been all over a’ready, an’ she ain’t
never been out of dis county yit.”
“Wonder ef she’ll look at folks when she gits back,” said Fred, the cadet
of the dining room, who was being trained under Polydore’s tutelage to
keep his nails clean and to offer dishes to guests at their left hands.
 “Don’ you be in too big a hurry to fin’ out dat, you nigga,” rejoined
Polydore, the loyalty of whose love for Diana would brook no criticism of
her on the part of an underling. “You’se got enough to attend to in gittin’ yer
manners into shape. Diana’s a superior pusson, an’ you ain’t got no ’casion
to criticise her. You jes’ take what yer gits an’ be thankful like Lazarus wuz
when de rich man dropped water outer his hand on his tongue.”
Polydore’s biblical erudition seems to have been a trifle at fault at this
point. But at any rate his simile had its intended effect upon the young
darkey, who, slipping a surreptitious beaten biscuit into his pocket, retreated
to the distant kitchen to devour it.
At that moment Diana entered the dining room with the air of a Duchess,
and, with unwonted sweetness, said:
“Please, Polydore, bring me de tea things. De ladies is faint.”
Polydore, anxious that Diana’s gentle mood should endure, made all
haste to bring what she desired. He made too much haste, unluckily, for in
his hurry he managed to spill a little hot water from a pitcher he was
carrying on a tray, and some drops of it fell upon the sleeve of Diana’s
daintily laundered cambric gown.
The stately bronze colored namesake of the ancient goddess rose in
offended dignity, and looked long at the offender before addressing him.
Then she witheringly put the question:
“Whar’s your manners dis mawnin’, Polydore? Jes’ spose I was Miss
Mony now; would you go sloppin’ things over her dat way?”
Even a worm will turn, we are told, and Polydore was prouder than a
worm. For once he lost his self-control so far as to say in reply:
“But you ain’t Miss Mony, dough you seems to think you is. I’se tired o’
yer highty tighty airs. Git de tea things for yerse’f!” With that Polydore left
the dining room, and Diana, curiously enough, made no reference to the
incident when next she encountered him, but was all smiles and sweetness
instead.
 XXVIII
THE ADVANCING SHADOW

N O sooner were Dorothy and Edmonia gone than Arthur turned again to
affairs. It was a troubled uneasy time in Virginia, a time of sore
apprehension and dread. The “irrepressible conflict” over slavery had
that year taken on new and more threatening features than ever before.
There was now a strong political party at the North the one important
article of whose creed was hostility to the further extension of slavery into
the territories. It was a strictly sectional party in its composition, having no
existence anywhere at the South. It was influential in Congress, and in 1856
it had strongly supported a candidate of its own for president. By the
beginning of 1860 its strength had been greatly increased and circumstances
rendered probable its success in electing a president that year, for the
hopeless division of the Democratic party, which occurred later in the year,
was already clearly foreshadowed, an event which in fact resulted in the
nomination of three rival candidates against Mr. Lincoln and made his
election certain in spite of a heavy popular majority against him.
Had this been all, Virginia would not have been greatly disturbed by the
political situation and prospect. But during the preceding autumn the
Virginians had been filled with apprehension for the safety of their homes
and families by John Brown’s attempt, at Harper’s Ferry, to create a negro
insurrection, the one catastrophe always most dreaded by them. That raid,
quickly suppressed as it was, wrought a revolution in Virginian feeling and
sentiment. The Virginians argued from it, and from the approval given to it
in some parts of the North, that Northern sentiment was rapidly ripening
into readiness for any measures, however violent they might be, for the
extinction of slavery and the destruction of the autonomy of the Southern
States.
They found it difficult under the circumstances to believe the Republican
party’s disclaimer of all purpose or power to interfere with the institution in
the states. They were convinced that only opportunity was now wanting to
make the Southern States the victims of an aggressive war, with a servile
insurrection as a horrible feature of it. They cherished a warm loyalty to that
Union which Virginia had done so much to create, but they began seriously
to fear the time when there would be no peace or safety for their state or
even for their wives and children within the Union. They were filled with
resentment, too, of what they regarded as a wanton and unlawful purpose to
interfere with their private concerns, and to force the country into disunion
and civil war.
There were hot heads among them, of course, who were ready to
welcome such results; but these were very few. The great body of Virginia’s
people loved the Union, and even to the end—a year later—their strongest
efforts were put forth to persuade both sides to policies of peace.
But in the meantime a marked change came over the Virginian mind
with respect to slavery. Many who had always regarded the institution as an
inherited evil to be got rid of as soon as that might be safely accomplished,
modified or reversed their view when called upon to stand always upon the
defensive against what they deemed an unjust judgment of themselves.
Arthur Brent did not share this change of view, but he shared in the
feelings of resentment which had given it birth. In common with other
Virginians he felt that this was a matter belonging exclusively to the
individual states, and still more strongly he felt that the existing political
situation and the methods of it gravely menaced the Union in ways which
were exceedingly difficult for Southern men who loved the Union to meet.
He saw with regret the great change that was coming over public and
private sentiment in Virginia—sentiment which had been so strongly
favorable to the peaceable extinction of slavery, that John Letcher—a
lifelong advocate of emancipation as Virginia’s true policy—had been
elected Governor the year before upon that as the only issue of a state
campaign.
But Arthur was still bent upon carrying out his purpose of emancipating
himself and, incidentally his slaves. And the threatening aspect of political
affairs strengthened his determination at any rate to rid both his own estate
and Dorothy’s of debt.
“When that is done, we shall be safe, no matter what happens,” he told
himself.
To that end he had already done much. In spite of his preöccupation with
the fever epidemic he had found time during the autumn to institute many
economies in the management of both plantations. He had shipped and sold
the large surplus crops of apples and sweet potatoes—a thing wholly
unprecedented in that part of Virginia, where no products of the soil except
tobacco and wheat were ever turned to money account. He was laughed at
for what his neighbors characterized as “Yankee farming,” but both his
conscience and his bank account were comforted by the results. In the same
way, having a large surplus of corn that year, he had fattened nearly double
the usual number of hogs, and was now preparing to sell so much of the
bacon as he did not need for plantation uses. In these and other ways he
managed to diminish the Wyanoke debt by more than a third and that of
Pocahontas by nearly one-half, during his first year as a planter.
“If they don’t quit laughing at me,” he said to Archer Bannister one day,
“I’ll sell milk and butter and even eggs next summer. I may conclude to do
that anyhow. Those are undignified crops, perhaps, but I’m not sure that
they could not be made more profitable than wheat and tobacco.”
“Be careful, Arthur,” answered his friend. “It isn’t safe to make planting
too profitable. It is apt to lead to unkindly remark.”
“How so? Isn’t planting a business, like any other?”
“A business, yes, but not like any other. It has a certain dignity to
maintain. But I wasn’t thinking of that. I was thinking of Robert Copeland.”
“I wish you’d tell me about him, Archer. What has he done? I observe
that everybody seems to shun him—or at least nobody seems quite willing
to recognize him as a man in our class, though they tell me his family is
fairly good, and personally he seems agreeable. Nobody says anything to
his discredit, and yet I observe a general shoulder shrugging whenever his
name is mentioned.”
“He makes too many hogsheads of tobacco to the hand,” answered
Archer smiling but speaking with emphasis and slowly.
“Is he cruel to his negroes?”
“Yes, and no. He is always good natured with them and kindly in his
fashion, but he works them much too hard. He doesn’t drive them
particularly. Indeed I never heard of his striking one of them. But he has
invented a system of money rewards and the like, by which he keeps them
perpetually racing with each other in their work. They badly overtax
themselves, and the community regards the matter with marked disfavor. In
the matter of family he isn’t in our class at all, but his father was much
respected. He was even a magistrate for some years before his death. But
the son has shut himself out of all social position by over working his
negroes, and the fact that he does it in ways that are ingenious and not
brutal doesn’t alter the fact, at least not greatly. Of course, if he did it in
brutal ways he would be driven out of the county. As it is he is only shut out
of society. I was jesting when I warned you of danger of that sort. But if
you are not careful in your application of ‘practical’ methods down here,
you’ll get a reputation for money loving, and that wouldn’t be pleasant.”
Arthur stoutly maintained his right and his duty to market all that the two
plantations produced beyond their own needs, especially so long as there
were debts upon them. Till these should be discharged, he contended, he
had no moral right to let products go to waste which could be turned into
money. Archer admitted the justice of his view, but laughingly added:
“It isn’t our way, down here, and we are so conservative that it is never
quite prudent to transgress our traditions. At the same time I wish we could
all rid our estates from debt before the great trouble comes. For it is surely
coming and God only knows what the upshot of it all will be. Don’t quote
me as saying that, please. It isn’t fashionable with us to be pessimistic, or to
doubt either the righteousness or the ultimate triumph of our cause. But
nobody can really foresee the outcome of our present troubles, and
whatever it may be, the men who are out of debt when it comes—if there
are any—will be better equipped to meet fate with a calm mind than the rest
of us.”
 XXIX
THE CORRESPONDENCE OF DOROTHY

F ROM the very beginning of her travels Dorothy revealed herself to

Arthur in rapidly succeeding letters which were—at the first, at least—
as frank as had been her talks with him during their long morning rides
together. If in the end her utterance became more guarded, with a touch of
reserve in it, and with a growing tendency to write rather of other things
than of her own emotions, Arthur saw in the fact only an evidence of that
increasing maturity of womanhood which he had intended her to gain. For
Arthur Brent studied Dorothy’s letters as scrutinizingly as if they had been
lessons in biology. Or, more accurately speaking, he studied Dorothy herself
in her letters, in that way.
From Richmond she wrote half rejoicings, half lamentations over the
long separation she must endure from him and from all else that had
hitherto constituted her life. Arthur observed that while she mentioned as a
troublesome thing the necessity of having still another gown made before
leaving for the North, she told him nothing whatever about the gown itself.
“Gowns,” he reflected, “are merely necessary adjuncts of life to Dorothy
—as yet. She’ll think of them differently after a while.”
From Washington she wrote delightedly of things seen, for although the
glory of the national capital had not come upon it in 1860, there was even
then abundant interest there for a country damsel.
From Baltimore she wrote:
“I’ve been wicked, and I like it. Edmonia took me to Kunkel and
Moxley’s Front Street Theatre last night to hear an opera, and I haven’t yet
wakened from the delicious dream. I think I never shall. I think I never
want to. Yet I am very sorry that I went, for I shall want to go again and
again and again, and you know that I mustn’t listen to such music as that. It
will make me bad, and really and truly I’d rather die than be bad. I don’t
understand it at all. Why is it wrong for me to hear great music when it isn’t
wrong for Edmonia? And Edmonia has heard the greatest music there is, in
New York and Paris and Vienna and Naples, and it hasn’t hurt her in the
least. I wish you would tell me why I am so different, won’t you, Cousin
Arthur?”
From New York she wrote of music again and many times, for she had
accepted Arthur’s assurance that there was no harm but only good for her in
listening to music; and in obedience to his injunction she went twice each
week to the opera, where Edmonia’s friends, whose guest she was, had a
box of their own. Presently came from her a pleading letter, asking if she
might not herself learn to play a little upon the violin, and availing herself
of Arthur’s more than ready permission, she toiled ceaselessly at the
instrument until after a month or so Edmonia reported that the girl’s music
master was raving about the extraordinary gifts she was manifesting.
“I am a trifle worried about it, Arthur,” Edmonia added. “Perhaps her
father was right to forbid all this. Music is not merely a delight to her—it is
a passion, an intoxication, almost a madness. She is very fond of dancing
too, but I think dancing is to her little more than a physical participation in
the music.
“And she mightily enjoys society. Still more does society enjoy her. Her
simplicity, her directness and her perfect truthfulness, are qualities not very
common, you know, in society, in New York or anywhere else. People are
delighted with her, and, without knowing it, she is the reigning attraction in
every drawing room. Ah me! She will have to know it presently, for I
foresee that ‘the young Virginia belle’ as they all call her, will have many
suitors for her hand before we sail—two weeks hence.
“She astonishes society in many ways. She is so perfectly well always,
for one thing. She is never tired and never has a headache. Most astonishing
of all, to the weary butterflies of fashion, she gets up early in the morning
and takes long rides on horseback before breakfast.
“In certain companies—the sedater sort—she is reckoned a brilliant
conversationalist. That is because she reads and thinks, as not many girls of
her age ever do. In more frivolous society she talks very little and is perhaps
a rather difficult person for the average young man to talk to. That also is
because she reads and thinks.
“On the whole, Arthur, Dorothy is developing altogether to my
satisfaction, but I am troubled about the music. Dr. South had a reason, of
which you know nothing, for fearing music in her case, as a dangerous
intoxication. Perhaps we ought not to disobey his instructions on the
subject. Won’t you think the matter over, Arthur, and advise me?”
To this request Arthur’s reply came promptly. It was an oracular
deliverance, such as he was accustomed to give only when absolutely sure
of his judgment.
“Have no fear as to the music,” he wrote. “It is not an intoxication to
Dorothy, as your report shows that indulgence in it is not followed by
reaction and lassitude. That is the sure test of intoxication. For the rest
Dorothy has a right to cultivate and make the most of the divine gifts she
possesses. Every human being has that right, and it is a cruel wrong to
forbid its exercise in any case. I am delighted to see from your letters and
hers that she has not permitted her interest in music to impair her interest in
other things. She tells me she has been reading a book on ‘The Origin of
Species’ by Charles Darwin. I have heard of it but haven’t seen it yet, as it
was published in England only a few months ago and had not been
reprinted here when I last wrote to New York for some books. So please ask
Dorothy to send me her copy as soon as she has finished it, and tell her
please not to rub out the marginal notes she tells me she has been making in
it. They will be helpfully suggestive to me in my reading, and, as
expressions of her uninfluenced opinion, I shall value them even more than
the text of the book itself. She tells me she thinks the book will work a
revolution in science, giving us a new foundation to build upon. I sincerely
hope so. We have long been in need of new foundations. But, pardon me,
you are not interested in scientific studies and I will write to Dorothy
herself about all that.”
At this point in her reading Edmonia laid the letter in her lap and left it
there for a considerable time before taking it up again. She was thinking, a
trifle sadly, perhaps, but not gloomily or with pain.
“He is right,” she reflected. “I sympathize with him in his high purposes
and I share the general admiration of his character and genius. But I do not
share his real enthusiasms as Dorothy does. I have none of that love for
scientific truth for its own sake, which is an essential part of his being. I
have none of his self-sacrificing earnestness, none of that divine discontent
which is the mainspring of all his acts and all his thinking. It is greatly
better as Fate has ordered it. I am no fit life partner for him. Had he married
me I should have made him happy in a way, perhaps, but it would have
been at cost of his deterioration. It is better as it is—immeasurably better,—
and I must school myself to think of it in that way. If I am worthy even of
the friendship that he so generously gives me, I shall learn to rejoice that he
gives the love to Dorothy instead, and not to me. I must learn to think of his
good and the fit working out of his life as the worthiest thing I can strive
for. And I am learning this lesson. It is a little hard at first, but I shall master
it.”
A few days later Dorothy, in one of her unconsciously self-revealing
letters, wrote:
“I am sending you the Darwin book, with all my crude little notes in the
margins. I have sealed it up quite securely and paid full letter postage on it,
because it would be unjust to the government to send a book with writing in
it, at book postage rates. Besides, I don’t want anybody but you to read the
notes. Edmonia asked me to let her see the book before sending it, but I told
her I couldn’t because I should die of shame if anybody should read my
presumptuous comments on so great a book. For it is great, really and truly
great. It is the greatest explanation of nature that anybody ever yet offered.
At least that is the way it impresses me. Edmonia asked me why I was so
chary of letting her see notes that I was entirely willing for you to see, and
at first I couldn’t explain it even to myself, for, of course, I love Edmonia
better than anybody else in the world, and I have no secrets from her. I told
her I would have to think the matter over before I could explain it, and she
said: ‘Think it out, child, and when you find out the explanation you may
tell me about it or not, just as you please.’ She kindly laughed it off, but it
troubled me a good deal. I couldn’t understand why it was that I couldn’t
bear to let her see the notes, while I rather wanted you to read them. I found
it all out at last, and explained it to her, and she seemed satisfied. It’s
because you know so much. You are my Master, and you always know how
to allow for your pupil’s wrong thinking even while you set me right.
Besides, somehow I am never ashamed of my ignorance when only you
know of it. Edmonia said that was quite natural, and that I was entirely right
not to show the scribbled book to her. So I don’t think I hurt her feelings, do
you?
“Now, I want to tell you about another thought of mine which may
puzzle you—or it may make you laugh as it does other people. There’s a
woman here—a very bright woman but not, to my taste, a lovely one—who
is very learned in a superficial way. She knows everything that is current in
science, art, literature, and fashion, though she seems to me deficient in
thoroughness. She ‘has the patter of it all at her tongue’s end,’ as they say
here, but I don’t think she knows much behind the patter. A wise editor
whom I met at dinner a few days ago, described her as ‘a person who holds
herself qualified to discuss and decide anything in heaven or earth from the
standpoint of the cyclopædia and her own inner consciousness.’ She writes
for one of the newspapers, though I didn’t know it when she talked with me
about Darwin. I told her I thought of Darwin’s book as a great poem. You
would have understood me, if I had said that to you, wouldn’t you? You
know I always think of the grass, and the trees and the flowers, and the
birds and the butterflies, and all the rest, as nature’s poems, and this book
seems to me a great epic which dominates and includes, and interprets them
all, just as Homer and Milton and Virgil and Tasso and especially
Shakespeare, dominate all the little twitterings of all the other poets.
Anyhow it seems to me that a book which tells us how all things came
about, is a poem and a very great one. I said so to this woman, and next day
I saw it all printed in the newspaper for which she writes. I shouldn’t have
minded that as she didn’t tell my name, but I thought she seemed to laugh
and jeer at my thought. She said something witty about trying to turn
Darwin into dactyls and substitute dithyrambics for dogmatics in the
writings of Sir Isaac Newton. Somehow it all sounded bright and witty as
one read it in the newspaper. But is that the way in which a serious thought
ought to be treated? Do the newspapers, when they thus flippantly deal with
serious things, really minister to human advancement? Do they not rather
retard it by making jests of things that are not jests? I have come to know a
good many newspaper writers since I have been here, and I am convinced
that they have no real seriousness in their work, no controlling conscience.
‘The newspaper’ said one of the greatest of them to me not long ago, ‘is a
mirror of today. It doesn’t bother itself much with tomorrow or yesterday.’ I
asked him why it should not reflect today accurately, instead of distorting if
with smartness. ‘Oh,’ he answered, ‘we can’t stop to consider such things.
We must do that which will please and attract the reader, regardless of
everything else. Dulness is the only thing we must avoid as we shun the
pestilence, and erudition and profundity are always dull.’
“ ‘But doesn’t the newspaper assume to teach men and women?’ I asked.
‘And is it not in conscience bound to teach them the truth and not
falsehood?’
 “ ‘Oh, yes, that’s the theory,’ he answered. ‘But how can we live up to it?
Take this matter of Darwinism, for example. We can’t afford to employ
great scientific men to discuss it seriously in our columns. And if we did,
only a few would read what they wrote about it. We have bright fellows on
our editorial staffs who know how to make it interesting by playing with it,
and for our purpose that is much better than any amount of learning.’
“I have been thinking this thing over and I have stopped the reading of
newspapers. For I find that they deal in the same way with everything else
—except politics, perhaps, and, of course, I know nothing of politics. I read
a criticism of a concert the other day in which a singer was—well, never
mind the details. The man that wrote that criticism didn’t hear the concert at
all, as he confessed to me. He was attending another theatre at the time. Yet
he assumed to criticise a singer to her detriment, utterly ignoring the fact
that she has her living to make by singing and that his criticism might
seriously affect her prospects. He laughed the matter off, and when I
seemed disturbed about it he said: ‘For your sake, Miss South, I’ll make
amends. She sings again tomorrow, and while I shall not be able to hear her,
I’ll give her such a laudation as shall warm the cockles of her heart and
make her manager mightily glad. You’ll forgive me, then, for censuring her
yesterday, I’m sure.’ I’m afraid I misbehaved myself then. I told him I
shouldn’t read his article, that I hated lies and shams and false pretences,
and that I didn’t consider his articles worth reading because they had no
truth or honesty behind them. It was dreadfully rude, I know, and yet I’m
not sorry for it. For it seemed to make an impression on him. He told me
that he only needed some such influence as mine to give him a conscience
in his work, and he actually asked me to marry him! Think of the absurdity
of it! I told him I wasn’t thinking of marrying anybody—that I was barely
seventeen, that—oh, well, I dismissed the poor fellow as gently as I could.”
But while the proposal of marriage by the newspaper man, and several
other such solicitations which followed it, struck Dorothy at first as
absurdities, they wrought a marked change in her mental attitude. Two at
least of these proposals were inspired by higher considerations than those of
the plantation which Dorothy represented, and were pressed with fervor and
tenderness by men quite worthy to aspire to Dorothy’s hand. These were
men of substance and character, in whose minds the fascination which the
Virginia girl unwittingly exercised over everybody with whom she came
into contact—men and women alike—had quickly ripened into a strong and
enduring passion. Dorothy suffered much in rejecting such suits as theirs,
but she learned something of herself in the process. She for the first time
realized that she was a woman and that she had actually entered upon that
career of womanhood which had before seemed so far away in the future
that thoughts of it had never before caused her to blush and tremble as they
did now.
These things set her thinking, and in her thinking she half realized her
own state of mind. She began dimly to understand the change that had come
over her attitude and feeling towards Arthur Brent. She would not let herself
believe that she loved him as a woman loves but one man while she lives;
but she admitted to herself that she might come to love him in that way if he
should ever ask her to do so with the tenderness and manifest sincerity
which these others had shown. But of that she permitted herself to entertain
no hope and even no thought. His letters to her, indeed, seemed to put that
possibility out of the question. For at this time Arthur held himself under
severe restraint. He was determined that he should not in any remotest way
take advantage of his position with respect to Dorothy, or use his influence
over her as a means of winning her. He knew now his own condition of
mind and soul in all its fulness. He was conscious now that the light of his
life lay in the hope of some day winning Dorothy’s love and making her all
and altogether his own. But he was more than ever determined, as he
formulated the thought in his own mind, to give Dorothy a chance, to take
no advantage of her, to leave her free to make choice for herself. It was his
fixed determination, should she come back heart whole from this journey, to
woo her with all the fervor of his soul; but the more determined he became
in this resolution, the more resolutely did he guard his written words against
the possibility that they might reveal aught of this to her. “If she ever comes
to love me as my wife,” he resolved, “it shall be only after she has had full
opportunity to make another choice.”
Accordingly his letters to her continued to concern themselves with
intellectual and other external things. He wrote her half a ream of comment
upon Darwin’s book, taking up for discussion every marginal note she had
made concerning it. But that part of his letter was as coldly intellectual as
any of their horseback conversations had been. In all the intimate parts of
that and his other letters, he wrote only as one might to a sympathetic
friend, as he might have written to Edmonia, for example. He even took
half unconscious pains to emphasize the fatherly character of his relations
with her, lest they assume some other aspect to her apprehension.
On her side Dorothy began now to write outside of herself, as it were.
She described to him all her new gowns and bonnets, laughing at the
confusion of mind in which a study of such details must involve him. In her
childlike loyalty she told him of the wooings that so distressed her, but she
did so quite as she might have written to him of the loves of Juliet and
Ophelia and the Lady of Lyons. For the rest she wrote objectively now, in
the main, and speculatively concerning certain of those social problems in
which she knew him to be profoundly interested, and which she was
somewhat studying now, because of the interest they had for him.
The word “slumming” had not been invented at that time by the
insolence that does the thing it means. But Dorothy, chiefly under the
guidance of her friends among the newspaper men, went to see how the
abjectly poor of a great city lived, and she wrote long letters of comment to
Arthur in which she told him how great and distressing the revelation was,
and how she honored his desire to do something for the amelioration of
these people’s lives. “Your aspiration is indeed a noble one,” she wrote in
one of her letters; “the life you proposed to yourself, and from which you
were diverted by your inheritance of a plantation, is the very greatest, the
very noblest that any man could lead. I once thought you were doing even
better in the care you are taking of the negroes at Wyanoke and Pocahontas,
and in your efforts ultimately to set them free. But that was when I did not
know. I know now, in part at least, and I understand your feeling in the
matter as I never could have done had I not seen for myself.
“People here sometimes say things to me that hurt. But I am ready with
my answer now. One woman—very intellectual, but a cat—asked me
yesterday how I could bear to hold negroes in slavery, and to buy fine
gowns with the proceeds of their toil. I told her frankly that I didn’t like it,
but that I couldn’t help it, and in reply to her singularly ignorant inquiries as
to why I didn’t end the wrong or at least my participation in it, I explained
some difficulties to her that she had never taken the trouble to ask about. I
told her how hard you were working to discharge the debts of your estate in
order that you might send your negroes to the west to be free, and that you
might yourself return to New York to do what you could for the
immeasurably worse slaves here. She caught at my phrase and challenged
it. I told her what I meant, and as it happened to be in a company of highly
intellectual people, I suppose I ought not to have talked so much, but
somehow they seemed to want to hear. I said:
“ ‘In Virginia I always visit every sick person on the plantation every
day. We send for a doctor in every case, and we women sit up night after
night to nurse every one that needs it. We provide proper food for the sick
and the convalescent from our own tables. We take care of the old and
decrepit, and of all the children. From birth to death they know that they
will be abundantly provided for. What poor family around the Five Points
has any such assurance? Who provides doctors and medicine and dainties
for them when they are ill? Who cares for their children? Who assures
them, in childhood and in old age, of as abundant a supply of food and
clothing, and as good a roof, as we give to the negroes? I go every morning,
as I just now said, to see every sick or afflicted negro on my own plantation
and on that of my guardian. How often have you gone to the region of the
Five Points to minister to those who are ill and suffering and perhaps
starving there?’
“ ‘Oh, that is all cared for by the charitable organizations’ she said, ‘and
by the city missionaries.’
“ ‘Is it?’ I answered. ‘I do not find it so. I have emptied my purse a
dozen times in an effort to get a doctor for a very ill person here, and to buy
the medicines he prescribed, and to provide food for starving ones. And
then, next day I have found that the sick have died because the well did not
know how to cook the food I had provided, or how to follow the doctor’s
directions in the giving of medicine. I tell you these poor people are
immeasurably worse off than any negro slave at the South is, or ever was.
So far as I can learn there is no working population in the world that gets
half so much of comfort and care and reward of every sort for its labor, as
the negroes of Virginia get.’
“Then the woman broke out. She said: ‘You are dressed in a superb
satin’—it was at a social function—‘and every dollar of its cost was earned
by a negro slave on your plantation.’ I answered, ‘You are equally well
dressed. Will you tell me who earned the money that paid for your satin
gown?’ Then, Cousin Arthur, I lost my temper and my manners. I told her
that while we in Virginia profited by the labor of our negroes, we gave
them, as the reward of their labor, every desire of their hearts and, besides
that, an assurance of support in absolute comfort for their old age, and for
their children; while the laboring class in New York, from whose labor she
profited, and whose toil purchased her gown, had nobody to care for them
in infancy or old age, in poverty and illness and suffering. ‘It is all wrong
on both sides,’ I said. ‘The toilers ought to have the full fruits of their toil in
both cases. The luxury of the rich is a robbery of the poor always and
everywhere. There ought not to be any such thing anywhere. The woman
who made your underclothing was robbed when you bought it at the price
you did. You wronged and defrauded the silk spinners and weavers and the
sewing women when you bought your gown. Worse than that; you have
among you men who have accumulated great fortunes in manufactures and
commerce. How did they do it? Was it not in commerce by paying the
producers for their products less than they were worth? Was it not in
manufactures by paying men and women and children less than they have
earned? Was not the great Astor estate based upon a shrewd robbery of the
Indian trappers and hunters? And has it not been swelled to its present
proportions by the growth of a city to whose growth the Astors have never
contributed a single dollar? Isn’t the whole thing a wrong and a robbery?
Isn’t the “Song of the Shirt” a reflection of truth? Isn’t there slavery in New
York as actually as in Virginia, and isn’t it infinitely more cruel?’
“Then the woman shifted her ground. ‘But at least our laborers are free,’
she said. ‘Are they?’ I answered. ‘Are they free to determine for whom they
will work or at what wages? Cannot their masters, who are their employers,
discharge them at will, when they get old or feeble or otherwise
incompetent, and leave them to starve? No master of a Virginia plantation
can do that. His neighbors would actually lynch him should he turn a
decrepit old negro out to die or even should he deny to him the abundant
food and clothing and housing that he gives to the able-bodied negroes who
make crops. And,’ I added, for I was excited, ‘this cruelty is not confined to
what are ordinarily called the laboring classes. I know a man of unusual
intellectual capacity, who has worked for years to build up the fortunes of
his employers. He has had what is regarded as a very high salary. But being
a man of generous mind he has spent his money freely in educating the ten
or a dozen sons and daughters of his less fortunate brother. He is growing
old now. He has earned for his master, a thousand dollars for every dollar of
salary that he ever received just as all his fellow workers in the business
have done. But he is growing old now, and under the strain of night and day
work, he has acquired the habit of drinking too much. He hasn’t a thousand
dollars in the world as his reward for helping to make this other man, his
master, absurdly, iniquitously rich. Yet in his age and infirmity, the other
man, luxuriating in his palatial summer home which is only one of the
many palaces that other men’s toil late into the night has provided for him,
decides that the old servitor is no longer worth his salary, and decrees his
discharge. Is there anything so cruel as that in negro slavery? Is that man
half so well off as my negro mammy, who has a house of her own and all
the food and clothes she wants at the age of eighty, and who could have the
service of a dozen negro attendants for the mere asking?’
“Now, Cousin Arthur, please don’t misunderstand me. Even what I have
seen at the Five Points doesn’t tempt me to believe in slavery. I want of all
things to see that exterminated. But, really and truly, I find an immeasurably
worse slavery here in New York than I ever saw in Virginia, and I want to
see it all abolished together, not merely the best and kindliest and most
humane part of it. I want to see the time when every human being who
works shall enjoy the full results of his work; when no man shall be any
other man’s master; when no man shall grow rich by pocketing the proceeds
of any other man’s genius or industry. I said all this to that woman, and she
replied: ‘You are obviously a pestilent socialist. You are as bad as Fourier
and Albert Brisbane and Horace Greeley.’ That was very rude of her, seeing
that Mr. Greeley was present, but I’ve noticed that the people who most
highly pride themselves on good manners are often rude and inconsiderate
of others to a degree that would not be tolerated in Virginia society—except
perhaps from Mr. Madison Peyton. By the way, Mr. Jefferson Peyton was
present, and to my regret he said some things in defence of slavery which I
could not at all approve. Mr. Greeley interrupted him to say something like
this:
“ ‘The dear young lady is quite right. We have a horrible slavery right
here in New York, and we ought to make war on that as earnestly as we do
on African slavery at the South. I’m trying to do it in the Try-bune’—that’s
the way he pronounces the name of his paper—‘and I’m going to keep on
trying.’
“That encouraged me, for I find myself more and more disposed to
respect Mr. Greeley as I come to know him better. I don’t always agree with
him. I sometimes think he is onesided in his views, and of course he is
enormously self-conceited. But he is, at any rate, an honest man, and a
brave and sincere one. He isn’t afraid to say what he thinks, any more than
you are, and I like him for that. Another great editor whom I have met
frequently, seems to me equally courageous, but far less conscientious. He
is inclined to take what he calls ‘the newspaper view’ of things,—by which
he means the view that appeals to the multitude for the moment, without
much regard for any fixed principle. Socially he is a much more agreeable
man than Mr. Greeley, but I don’t think him so trustworthy. Mr. Greeley
impresses me as a man who may be enormously wrong-headed, under the
influence of his prejudiced misconceptions, but who, wrong-headed or
right-headed, will never consciously wrong others. If he had been born the
master of a Virginia plantation he would have dealt with his negroes in the
same spirit in which he has insisted upon giving to his fellow workers on
the Tribune a share in the profits of their joint work. Mr. Greeley is odd, but
I like him better than any editor I have met.”
So the girl went on, writing objectively and instinctively avoiding the
subjective. But she did not always write so seriously. She had “caught the
patter” of society and she often filled pages with a sparkling, piquant
flippancy, which had for Arthur a meaning all its own.
 XXX
AT SEA

T HE voyage to Europe in 1860 was a much more serious undertaking

than the like voyage is in our later time. It occupied a fortnight or three
weeks, for one thing, where now a week is ample time for the passage.
The steamers were small and uncomfortable—the very largest of them
being only half the size of the very smallest now regarded as fit for
passenger service. There was no promenade deck or hurricane deck then,
above the main deck, which was open to the sky throughout its length and
breadth, except for the interruption of one small deck house, covering the
companion way, a ventilator pipe here and there, and perhaps a chicken
coop to furnish emaciated and sea sick fowls for the table d’hôte. There was
no ice machine on board, and no distilling apparatus for the production of
fresh water. As a consequence, after two days out, the warm water which
passengers must drink began to taste of the ancient wood of the water tanks;
at the end of a week it became sickeningly foul; and before the end of the
voyage it became so utterly undrinkable that the most aggressive teetotaler
among the passengers was compelled to order wine for his dinner and to
abstain from coffee at breakfast.
The passenger who did not grow seasick in those days was a rare
exception to an otherwise universal rule, while, in our time, when the
promenade deck is forty or fifty feet above the waves, and the passengers
are abundantly supplied both with palatable food and with wholesome
water, only those suffer with mal de mer who are bilious when they go on
board, or who are beset by a senseless apprehension of the sea.
The passenger lists were small, too, even allowing for the diminutive
size of the ships. One person crossed the ocean then where perhaps a
hundred cross in our time.
There were perhaps twenty passengers in the cabin of the ship in which
Dorothy sailed. By the second day out only two of the ship’s company
appeared at meals or at all regularly took the air on deck. Dorothy was one
of these two. The other she herself introduced, as it were, to Arthur, in a