Visit ebookfinal.

com to download the full version and

explore more ebooks or textbooks

An Introduction to Statistical Computing A

Simulation based Approach 1st Edition Jochen Voss

_____ Click the link below to download _____

https://ebookfinal.com/download/an-introduction-to-
statistical-computing-a-simulation-based-approach-1st-
edition-jochen-voss/

Explore and download more ebooks or textbook at ebookfinal.com

Here are some recommended products that we believe you will be
interested in. You can click the link to download.

Program Evaluation An Introduction to an Evidence Based

Approach 6th Edition David Royse

https://ebookfinal.com/download/program-evaluation-an-introduction-to-
an-evidence-based-approach-6th-edition-david-royse/

Introduction to Supercritical Fluids A Spreadsheet based

Approach 1st Edition Richard Smith

https://ebookfinal.com/download/introduction-to-supercritical-fluids-
a-spreadsheet-based-approach-1st-edition-richard-smith/

Location Based Services 1st Edition Jochen Schiller

https://ebookfinal.com/download/location-based-services-1st-edition-
jochen-schiller/

An Introduction to Statistical Signal Processing 1st

Edition Robert M. Gray

https://ebookfinal.com/download/an-introduction-to-statistical-signal-
processing-1st-edition-robert-m-gray/
An Introduction to Numerical Methods A MATLAB Approach
Third Edition Guenther

https://ebookfinal.com/download/an-introduction-to-numerical-methods-
a-matlab-approach-third-edition-guenther/

An invariant approach to statistical analysis of shapes

1st Edition Subhash R. Lele

https://ebookfinal.com/download/an-invariant-approach-to-statistical-
analysis-of-shapes-1st-edition-subhash-r-lele/

An Introduction to Statistical Concepts 3rd Edition Debbie

L. Hahs-Vaughn

https://ebookfinal.com/download/an-introduction-to-statistical-
concepts-3rd-edition-debbie-l-hahs-vaughn/

A Comprehensible Guide to J1939 First Edition Voss

https://ebookfinal.com/download/a-comprehensible-guide-to-j1939-first-
edition-voss/

An introduction to computer simulation methods

Applications to physical systems 3rd Edition Harvey Gould

https://ebookfinal.com/download/an-introduction-to-computer-
simulation-methods-applications-to-physical-systems-3rd-edition-
harvey-gould/
An Introduction to Statistical Computing A Simulation
based Approach 1st Edition Jochen Voss Digital Instant
Download
Author(s): Jochen Voss
ISBN(s): 9781118357729, 1118357728
Edition: 1
File Details: PDF, 3.37 MB
Year: 2013
Language: english
 An Introduction to
Statistical Computing
WILEY SERIES IN COMPUTATIONAL STATISTICS

Consulting Editors:

Paolo Giudici
University of Pavia, Italy

Geof H. Givens
Colorado State University, USA

Bani K. Mallick
Texas A & M University, USA

Wiley Series in Computational Statistics is comprised of practical guides and cutting

edge research books on new developments in computational statistics. It features
quality authors with a strong applications focus. The texts in the series provide
detailed coverage of statistical concepts, methods and case studies in areas at the
interface of statistics, computing, and numerics.
With sound motivation and a wealth of practical examples, the books show in
concrete terms how to select and to use appropriate ranges of statistical comput-
ing techniques in particular fields of study. Readers are assumed to have a basic
understanding of introductory terminology.
The series concentrates on applications of computational methods in statistics to
fields of bioinformatics, genomics, epidemiology, business, engineering, finance and
applied statistics.

Titles in the Series

Biegler, Biros, Ghattas, Heinkenschloss, Keyes, Mallick, Marzouk, Tenorio,
Waanders, Willcox – Large-Scale Inverse Problems and Quantification of Uncertainty
Billard and Diday – Symbolic Data Analysis: Conceptual Statistics and Data Mining
Bolstad – Understanding Computational Bayesian Statistics
Borgelt, Steinbrecher and Kruse – Graphical Models, 2e
Dunne – A Statistical Approach to Neutral Networks for Pattern Recognition
Liang, Liu and Carroll – Advanced Markov Chain Monte Carlo Methods
Ntzoufras – Bayesian Modeling Using WinBUGS
Tufféry – Data Mining and Statistics for Decision Making
 An Introduction to
Statistical Computing
A Simulation-based Approach

Jochen Voss
School of Mathematics, University of Leeds, UK
This edition first published 2014

C 2014 John Wiley & Sons, Ltd

Registered office
John Wiley & Sons, Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ,
United Kingdom
For details of our global editorial offices, for customer services and for information about how to apply
for permission to reuse the copyright material in this book please see our website at www.wiley.com.
The right of the author to be identified as the author of this work has been asserted in accordance with the
Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or
transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise,
except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of
the publisher.
Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may
not be available in electronic books.
Designations used by companies to distinguish their products are often claimed as trademarks. All brand
names and product names used in this book are trade names, service marks, trademarks or registered
trademarks of their respective owners. The publisher is not associated with any product or vendor
mentioned in this book.
Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in
preparing this book, they make no representations or warranties with respect to the accuracy or
completeness of the contents of this book and specifically disclaim any implied warranties of
merchantability or fitness for a particular purpose. It is sold on the understanding that the publisher is not
engaged in rendering professional services and neither the publisher nor the author shall be liable for
damages arising herefrom. If professional advice or other expert assistance is required, the services of a
competent professional should be sought.
Library of Congress Cataloging-in-Publication Data
Voss, Jochen.
An introduction to statistical computing : a simulation-based approach / Jochen Voss. – First edition.
pages cm. – (Wiley series in computational statistics)
Includes bibliographical references and index.
ISBN 978-1-118-35772-9 (hardback)
1. Mathematical statistics–Data processing. I. Title.
QA276.4.V66 2013
519.501 13–dc23
2013019321

A catalogue record for this book is available from the British Library.
ISBN: 978-1-118-35772-9
Typeset in 10/12pt Times by Aptara Inc., New Delhi, India

1 2014
Contents

List of algorithms ix

Preface xi

Nomenclature xiii

1 Random number generation 1

1.1 Pseudo random number generators 2
1.1.1 The linear congruential generator 2
1.1.2 Quality of pseudo random number generators 4
1.1.3 Pseudo random number generators in practice 8
1.2 Discrete distributions 8
1.3 The inverse transform method 11
1.4 Rejection sampling 15
1.4.1 Basic rejection sampling 15
1.4.2 Envelope rejection sampling 18
1.4.3 Conditional distributions 22
1.4.4 Geometric interpretation 26
1.5 Transformation of random variables 30
1.6 Special-purpose methods 36
1.7 Summary and further reading 36
Exercises 37

2 Simulating statistical models 41

2.1 Multivariate normal distributions 41
2.2 Hierarchical models 45
2.3 Markov chains 50
2.3.1 Discrete state space 51
2.3.2 Continuous state space 56
2.4 Poisson processes 58
2.5 Summary and further reading 67
Exercises 67
vi CONTENTS

3 Monte Carlo methods 69

3.1 Studying models via simulation 69
3.2 Monte Carlo estimates 74
3.2.1 Computing Monte Carlo estimates 75
3.2.2 Monte Carlo error 76
3.2.3 Choice of sample size 80
3.2.4 Refined error bounds 82
3.3 Variance reduction methods 84
3.3.1 Importance sampling 84
3.3.2 Antithetic variables 88
3.3.3 Control variates 93
3.4 Applications to statistical inference 96
3.4.1 Point estimators 97
3.4.2 Confidence intervals 100
3.4.3 Hypothesis tests 103
3.5 Summary and further reading 106
Exercises 106

4 Markov Chain Monte Carlo methods 109

4.1 The Metropolis–Hastings method 110
4.1.1 Continuous state space 110
4.1.2 Discrete state space 113
4.1.3 Random walk Metropolis sampling 116
4.1.4 The independence sampler 119
4.1.5 Metropolis–Hastings with different move types 120
4.2 Convergence of Markov Chain Monte Carlo methods 125
4.2.1 Theoretical results 125
4.2.2 Practical considerations 129
4.3 Applications to Bayesian inference 137
4.4 The Gibbs sampler 141
4.4.1 Description of the method 141
4.4.2 Application to parameter estimation 146
4.4.3 Applications to image processing 151
4.5 Reversible Jump Markov Chain Monte Carlo 158
4.5.1 Description of the method 160
4.5.2 Bayesian inference for mixture distributions 171
4.6 Summary and further reading 178
4.6 Exercises 178

5 Beyond Monte Carlo 181

5.1 Approximate Bayesian Computation 181
5.1.1 Basic Approximate Bayesian Computation 182
5.1.2 Approximate Bayesian Computation with regression 188
5.2 Resampling methods 192
 CONTENTS vii

5.2.1 Bootstrap estimates 192

5.2.2 Applications to statistical inference 197
5.3 Summary and further reading 209
Exercises 209

6 Continuous-time models 213

6.1 Time discretisation 213
6.2 Brownian motion 214
6.2.1 Properties 216
6.2.2 Direct simulation 217
6.2.3 Interpolation and Brownian bridges 218
6.3 Geometric Brownian motion 221
6.4 Stochastic differential equations 224
6.4.1 Introduction 224
6.4.2 Stochastic analysis 226
6.4.3 Discretisation schemes 231
6.4.4 Discretisation error 236
6.5 Monte Carlo estimates 243
6.5.1 Basic Monte Carlo 243
6.5.2 Variance reduction methods 247
6.5.3 Multilevel Monte Carlo estimates 250
6.6 Application to option pricing 255
6.7 Summary and further reading 259
Exercises 260

Appendix A Probability reminders 263

A.1 Events and probability 263
A.2 Conditional probability 266
A.3 Expectation 268
A.4 Limit theorems 269
A.5 Further reading 270

Appendix B Programming in R 271

B.1 General advice 271
B.2 R as a Calculator 272
B.2.1 Mathematical operations 273
B.2.2 Variables 273
B.2.3 Data types 275
B.3 Programming principles 282
B.3.1 Don’t repeat yourself! 283
B.3.2 Divide and conquer! 286
B.3.3 Test your code! 290
B.4 Random number generation 292
B.5 Summary and further reading 294
Exercises 294
viii CONTENTS

Appendix C Answers to the exercises 299

C.1 Answers for Chapter 1 299
C.2 Answers for Chapter 2 315
C.3 Answers for Chapter 3 319
C.4 Answers for Chapter 4 328
C.5 Answers for Chapter 5 342
C.6 Answers for Chapter 6 350
C.7 Answers for Appendix B 366

References 375

Index 379
List of algorithms

Random number generation

alg. 1.2 linear congruential generator 2
alg. 1.13 inverse transform method 12
alg. 1.19 basic rejection sampling 15
alg. 1.22 envelope rejection sampling 19
alg. 1.25 rejection sampling for conditional distributions 22

Simulating statistical models

alg. 2.9 mixture distributions 47
alg. 2.11 componentwise simulation 49
alg. 2.22 Markov chains with discrete state space 53
alg. 2.31 Markov chains with continuous state space 58
alg. 2.36 Poisson process 61
alg. 2.41 thinning method for Poisson processes 65

Monte Carlo methods

alg. 3.8 Monte Carlo estimate 75
alg. 3.22 importance sampling 85
alg. 3.26 antithetic variables 89
alg. 3.31 control variates 93

Markov Chain Monte Carlo methods

alg. 4.2 Metropolis–Hastings method for continuous state space 110
alg. 4.4 Metropolis–Hastings method for discrete state space 113
alg. 4.9 random walk Metropolis 117
alg. 4.11 independence sampler 119
alg. 4.12 Metropolis–Hastings method with different move types 121
alg. 4.27 Gibbs sampler 142
alg. 4.31 Gibbs sampler for the Ising model 155
alg. 4.32 Gibbs sampler in image processing 158
alg. 4.36 reversible jump Markov Chain Monte Carlo 165
x LIST OF ALGORITHMS

Beyond Monte Carlo

alg. 5.1 basic Approximate Bayesian Computation 182
alg. 5.6 Approximate Bayesian Computation with regression 191
alg. 5.11 general bootstrap estimate 196
alg. 5.15 bootstrap estimate of the bias 200
alg. 5.18 bootstrap estimate of the standard error 202
alg. 5.20 simple bootstrap confidence interval 205
alg. 5.21 BCa bootstrap confidence interval 207

Continuous-time models
alg. 6.6 Brownian motion 217
alg. 6.12 Euler–Maruyama scheme 232
alg. 6.15 Milstein scheme 235
alg. 6.26 multilevel Monte Carlo estimates 251
alg. 6.29 Euler–Maruyama scheme for the Heston model 256
Preface

This is a book about exploring random systems using computer simulation and thus,
this book combines two different topic areas which have always fascinated me:
the mathematical theory of probability and the art of programming computers. The
method of using computer simulations to study a system is very different from the
more traditional, purely mathematical approach. On the one hand, computer exper-
iments normally can only provide approximate answers to quantitative questions,
but on the other hand, results can be obtained for a much wider class of systems,
including large and complex systems where a purely theoretical approach becomes
difficult.
In this text we will focus on three different types of questions. The first, easiest
question is about the normal behaviour of the system: what is a typical state of the sys-
tem? Such questions can be easily answered using computer experiments: simulating
a few random samples of the system gives examples of typical behaviour. The second
kind of question is about variability: how large are the random fluctuations? This
type of question can be answered statistically by analysing large samples, generated
using repeated computer simulations. A final, more complicated class of questions is
about exceptional behaviour: how small is the probability of the system behaving in
a specified untypical way? Often, advanced methods are required to answer this third
type of question. The purpose of this book is to explain how such questions can be
answered. My hope is that, after reading this book, the reader will not only be able
to confidently use methods from statistical computing for answering such questions,
but also to adjust existing methods to the requirements of a given problem and, for
use in more complex situations, to develop new specialised variants of the existing
methods.
This text originated as a set of handwritten notes which I used for teaching
the ‘Statistical Computing’ module at the University of Leeds, but now is greatly
extended by the addition of many examples and more advanced topics. The material
we managed to cover in the ‘Statistical Computing’ course during one semester is less
than half of what is now the contents of the book! This book is aimed at postgraduate
students and their lecturers; it can be used both for self-study and as the basis of
taught courses. With the inclusion of many examples and exercises, the text should
also be accessible to interested undergraduate students and to mathematically inclined
researchers from areas outside mathematics.
xii PREFACE

Only very few prerequisites are required for this book. On the mathematical side,
the text assumes that the reader is familiar with basic probability, up to and including
the law of large numbers; Appendix A summarises the required results. As a con-
sequence of the decision to require so little mathematical background, some of the
finer mathematical subtleties are not discussed in this book. Results are presented in a
way which makes them easily accessible to readers with limited mathematical back-
ground, but the statements are given in a form which allows the mathematically more
knowledgeable reader to easily add the required detail on his/her own. (For example,
I often use phrases such as ‘every set A ⊆ Rd ’ where full mathematical rigour would
require us to write ‘every measurable set A ⊆ Rd ’.) On the computational side, basic
programming skills are required to make use of the numerical methods introduced
in this book. While the text is written independent of any specific programming
language, the reader will need to choose a language when implementing methods
from this book on a computer. Possible choices of programming language include
Python, Matlab and C/C++. For my own implementations, provided as part of the
solutions to the exercises in Appendix C, I used the R programming language; a short
introduction to programming with R is provided in Appendix B.
Writing this book has been a big adventure for me. When I started this project,
more than a year ago, my aim was to cover enough material so that I could discuss
the topics of multilevel Monte Carlo and reversible jump Markov Chain Monte Carlo
methods. I estimated that 350 pages would be enough to cover this material but it
quickly transpired that I had been much too optimistic: my estimates for the final
page count kept rising and even after several rounds of throwing out side-topics and
generally tightening the text, the book is still stretching this limit! Nevertheless, the
text now covers most of the originally planned topics, including multilevel Monte
Carlo methods near the very end of the book. Due to my travel during the last year,
parts of this book have been written on a laptop in exciting places. For example, the
initial draft of section 1.5 was written on a coach travelling through the beautiful
island of Kyushu, halfway around the world from where I live! All in all, I greatly
enjoyed writing this book and I hope that the result is useful to the reader.
This book contains an accompanying website. Please visit www.wiley.com/
go/statistical_computing

Jochen Voss
Leeds, March 2013
Nomenclature

For reference, the following list summarises some of the notation used throughout
this book.

ø the empty set

N the natural numbers: N = {1, 2, 3, . . .}
N0 the non-negative integers: N = {0, 1, 2, . . .}
Z the integers: Z = {. . . , −2, −1, 0, 1, 2, . . .}
n mod m the remainder of the division of n by m, in the range 0, 1, . . . , m − 1
δkl the Kronecker delta: δkl = 1 if k = l and δkl = 0 otherwise
R the real numbers
x the number x ∈ R ‘rounded up’, that is the smallest integer greater than
or equal to x
(an )n∈N a sequence of (possibly random) numbers: (an )n∈N = (a1 , a2 , . . .)
O(·) the big O notation, introduced in definition
 3.16
 
[a, b] an interval of real numbers: [a, b] = x ∈ R  a ≤ x ≤ b
{a, b} the set containing a and b   
A the complement of a set: A = x  x ∈ / A .
A×B the Cartesian
 product
 of the sets A and B:
A × B = (a, b)  a ∈ A, b ∈ B
1 A (x) the indicator function of the set A: 1 A (x) = 1 if x ∈ A and 0 otherwise
(see section A.3)
U[0, 1] the uniform distribution on the interval [0, 1]
U{−1, 1} the uniform distribution on the two-element set {−1, 1}
Pois(λ) the Poisson distribution with parameter λ
X ∼μ indicates that a random variable X is distributed according to a probability
distribution μ
|S| the number of elements in a finite set S; in section 1.4 also the volume of
a subsets S ⊆ Rd
RS space of vectors where the components are indexed by elements of S (see
section 2.3.2)
R S×S space of matrices where rows and columns are indexed by elements of S
(see section 2.3.2)
1

Random number generation

The topic of this book is the study of statistical models using computer simulations.
Here we use the term ‘statistical models’ to mean any mathematical models which
include a random component. Our interest in this chapter and the next is in simu-
lation of the random component of these models. The basic building block of such
simulations is the ability to generate random numbers on a computer, and this is the
topic of the present chapter. Later, in Chapter 2, we will see how the methods from
Chapter 1 can be combined to simulate more complicated models.
Generation of random numbers, or more general random objects, on a computer
is complicated by the fact that computer programs are inherently deterministic: while
the output of computer program may look random, it is obtained by executing the
steps of some algorithm and thus is totally predictable. For example the output of a
program computing the decimal digits of the number

π = 3.14159265358979323846264338327950288419716939937510 · · ·

(the ratio between the perimeter and diameter of a circle) looks random at first sight,
but of course π is not random at all! The output can only start with the string of digits
given above and running the program twice will give the same output twice.
We will split the problem of generating random numbers into two distinct sub-
problems: first we will study the problem of generating any randomness at all, con-
centrating on the simple case of generating independent random numbers, uniformly
distributed on the interval [0, 1]. This problem and related concerns will be discussed
in Section 1.1. In the following sections, starting with Section 1.2, we will study the
generation of random numbers from different distributions, using the independent,
uniformly distributed random numbers obtained in the previous step as a basis.

An Introduction to Statistical Computing: A Simulation-based Approach, First Edition. Jochen Voss.

© 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd.
2 AN INTRODUCTION TO STATISTICAL COMPUTING

1.1 Pseudo random number generators

There are two fundamentally different classes of methods to generate random
numbers:

(a) True random numbers are generated using some physical phenomenon which
is random. Generating such numbers requires specialised hardware and can
be expensive and slow. Classical examples of this include tossing a coin or
throwing dice. Modern methods utilise quantum effects, thermal noise in
electric circuits, the timing of radioactive decay, etc.
(b) Pseudo random numbers are generated by computer programs. While these
methods are normally fast and resource effective, a challenge with this
approach is that computer programs are inherently deterministic and therefore
cannot produce ‘truly random’ output.

In this text we will only consider pseudo random number generators.

Definition 1.1 A pseudo random number generator (PRNG) is an algorithm which

outputs a sequence of numbers that can be used as a replacement for an independent
and identically distributed (i.i.d.) sequence of ‘true random numbers’.

1.1.1 The linear congruential generator

This section introduces the linear congruential generator (LCG), a simple example of
a PRNG. While this random number generator is no longer of practical importance,
it shares important characteristics with the more complicated generators used in
practice today and we study it here as an accessible example. The LCG is given by
the following algorithm.

Algorithm 1.2 (linear congruential generator)

input:
m > 1 (the modulus)
a ∈ {1, 2, . . . , m − 1} (the multiplier)
c ∈ {0, 1, . . . , m − 1} (the increment)
X 0 ∈ {0, 1, . . . , m − 1} (the seed)
output:
a sequence X 1 , X 2 , X 3 , . . . of pseud random numbers
1: for n = 1, 2, 3, . . . do
2: X n ← (a X n−1 + c) mod m
3: output X n
4: end for
 RANDOM NUMBER GENERATION 3

In the algorithm, ‘mod’ denotes the modulus for integer division, that is the value
n mod m is the remainder of the division of n by m, in the range 0, 1, . . . , m −
1. Thus the sequence generated by algorithm 1.2 consists of integers X n from the
range {0, 1, 2, . . . , m − 1}. The output depends on the parameters m, a, c and on the
seed X 0 . We will see that, if m, a and c are carefully chosen, the resulting sequence
behaves ‘similar’ to a sequence of independent, uniformly distributed random vari-
ables. By choosing different values for the seed X 0 , different sequences of pseudo
random numbers can be obtained.

Example 1.3 For parameters m = 8, a = 5, c = 1 and seed X 0 = 0, algorithm 1.2

gives the following output:

n 5X n−1 + 1 Xn

1 1 1
2 6 6
3 31 7
4 36 4
5 21 5
6 26 2
7 11 3
8 16 0
9 1 1
10 6 6

The output 1, 6, 7, 4, 5, 2, 3, 0, 1, 6, . . . shows no obvious pattern and could be con-

sidered to be a sample of a random sequence.

While the output of the LCG looks random, from the way it is generated it is
clear that the output has several properties which make it different from truly random
sequences. For example, since each new value of X n is computed from X n−1 , once the
generated series reaches a value X n which has been generated before, the output starts
to repeat. In example 1.3 this happens for X 8 = X 0 and we get X 9 = X 1 , X 10 = X 2
and so on. Since X n can take only m different values, the output of a LCG starts
repeating itself after at most m steps; the generated sequence is eventually periodic.
Sometimes the periodicity of a sequence of pseudo random numbers can cause
problems, but on the other hand, if the period length is longer than the amount of
random numbers we use, periodicity cannot affect our result. For this reason, one
needs to carefully choose the parameters m, a and c in order to achieve a long enough
period. In particular m, since it is an upper bound for the period length, needs to be
chosen large. In practice, typical values of m are on the order of m = 232 ≈ 4 · 109
and a and c are then chosen such that the generator actually achieves the maximally
possible period length of m. A criterion for the choice of m, a and c is given in the
following theorem (Knuth, 1981, Section 3.2.1.2).
4 AN INTRODUCTION TO STATISTICAL COMPUTING

Theorem 1.4 The LCG has period m if and only if the following three conditions
are satisfied:

(a) m and c are relatively prime;

(b) a − 1 is divisible by every prime factor of m;
(c) if m is a multiple of 4, then a − 1 is a multiple of 4.

In the situation of the theorem, the period length does not depend on the seed X 0
and usually this parameter is left to be chosen by the user of the PRNG.

Example 1.5 Let m = 232 , a = 1 103 515 245 and c = 12 345. Since the only
prime factor of m is 2 and c is odd, the values m and c are relatively prime and condition
(a) of the theorem is satisfied. Similarly, condition (b) is satisfied, since a − 1 is
even and thus divisible by 2. Finally, since m is a multiple of 4, we have to check
condition (c) but, since a − 1 = 1 103 515 244 = 275 878 811 · 4, this condition also
holds. Therefore the LCG with these parameters m, a and c has period 232 for every
seed X 0 .

1.1.2 Quality of pseudo random number generators

PRNGs used in modern software packages such as R or Matlab are more sophisticated
(and more complicated) than the LCG presented in Section 1.1.1, but they still share
many characteristics of the LCG. We will see that no PRNG can produce a perfect
result, but the random number generators used in practice, for example the Mersenne
Twister algorithm (Matsumoto and Nishimura, 1998), are good enough for most
purposes. In this section we will discuss criteria for the quality of the output of
general PRNGs, and will illustrate these criteria using the LCG as an example.

1.1.2.1 Period length of the output

We have seen that the output of the LCG is eventually periodic, with a period length
of at most m. This property that the output is eventually periodic is shared by all
PRNGs implemented in software. Most PRNGs used in practice have a period length
which is much larger than the amount of random numbers a computer program could
ever use in a reasonable time. For this reason, periodicity of the output is not a big
problem in practical applications of PRNGs. The period length is a measure for the
quality of a PRNG.

1.1.2.2 Distribution of samples

The output of almost all PRNGs is constructed so that it can be used as a replacement
for an i.i.d. sample of uniformly distributed random numbers. Since the output takes
 RANDOM NUMBER GENERATION 5

values in a finite set S = {0, 1, . . . , m − 1}, in the long run, for every set A ⊆ S we
should have

  
# i  1 ≤ i ≤ N , Xi ∈ A #A
≈ , (1.1)
N #S

where #A stands for the number of elements in a finite set A.

Uniformity of the output can be tested using statistical tests like the chi-
squared test or the Kolmogorov–Smirnov test (see e.g. Lehmann and Romano, 2005,
Chapter 14).
One peculiarity when applying statistical tests for the distribution of samples to the
output of a PRNG is that the test may fail in two different ways: The output could either
have the wrong distribution (i.e. not every value appears with the same probability),
or the output could be too regular. For example, the sequence X n = n mod m hits
every value equally often in the long run, but it shows none of the fluctuations which
are typical for a sequence of real random numbers. For this reason, statistical tests
should be performed as two-sided tests when the distribution of the output of a PRNG
is being tested.

Example 1.6 Assume that we have a PRNG with m = 1024 possible output values
and that we perform a chi-squared test for the hypothesis

P (X i ∈ {64 j, 64 j + 1, . . . , 64 j + 63}) = 1/16

for j = 0, 1, . . . , 15.
If we consider a sample X 1 , X 2 , . . . , X N , the test statistic of the chi-squared test
is computed from the observed numbers of samples in each block, given by

  
O j = # i  64 j ≤ X i < 64( j + 1) .

The expected count for block j, assuming that (1.1) holds, is

E j = N · 64/1024 = N /16

for j = 0, 1, . . . , 15 and the test statistic of the corresponding chi-squared test is

15
(O j − E j )2
Q= .
j=0
Ej
6 AN INTRODUCTION TO STATISTICAL COMPUTING

For large sample size N , and under the hypothesis (1.1), the value Q follows a
χ 2 -distribution with 15 degrees of freedom. Some quantiles of this distribution are:

q 6.262 7.261 ··· 24.996 27.488

P(Q ≤ q) 0.025 0.05 ··· 0.95 0.975

Thus, for a one-sided test with significance level 1 − α = 95% we would reject the
hypothesis if Q > 24.996. In contrast, for a two-sided test with significance level
1 − α = 95%, we would reject the hypothesis if either Q < 6.262 or Q > 27.488.
We consider two different test cases: first, if X n = n mod 1024 for n =
1, 2, . . . , N = 106 , we find Q = 0.244368. Since the series is very regular, the value
of Q is very low. The one-sided test would accept this sequence as being uniformly
distributed, whereas the two-sided test would reject the sequence.
Secondly, we consider X n = n mod 1020 for n = 1, 2, . . . , N = 106 . Since this
series never takes the values 1021 to 1023, the distribution is wrong and we expect a
large value of Q. Indeed, for this case we get Q = 232.5864 and thus both versions
of the test reject this sequence.

Random number generators used in practice, and even the LCG for large enough
values of m, pass statistical tests for the distribution of the output samples without
problems.

1.1.2.3 Independence of samples

Another aspect of the quality of PRNGs is the possibility of statistical dependence
between consecutive samples. For example, in the LCG each output sample is a
deterministic function of the previous sample and thus consecutive samples are clearly
dependent. To some extent this problem is shared by all PRNGs.
An easy way to visualise the dependence between pairs of consecutive samples
is a scatter plot of the points (X i , X i+1 ) for i = 1, 2, . . . , N − 1. A selection of such
plots is shown in Figure 1.1. Figure 1.1(a) illustrates what kind of plot one would
expect if X i ∼ U[0, 1] was a true i.i.d. sequence. The remaining panels correspond
to different variants of the LCG. Figure 1.1(b) (using m = 81) clearly illustrates that
each X i can only be followed by exactly one value X i+1 . While the same is true for
Figure 1.1(c) and (d) (using m = 1024 and m = 232 , respectively), the dependence
is much convoluted there and in particular the structure of Figure 1.1(d) is visually
indistinguishable from the structure of Figure 1.1(a).
One method for constructing PRNGs where X i+1 is not a function of X i is to
use a function f (X i ) of the state, instead of the state X i itself, as the output of
the PRNG. Here, f : {0, 1, . . . , m − 1} → {0, 1, . . . , m̃ − 1} is a map where m̃ < m
and where the same number of pre-images is mapped to each output value. Then a
uniform distribution of X i will be mapped to a uniform distribution for f (X i ) but
the output f (X i+1 ) is not a function of the previous output f (X i ). This allows to
construct random number generators with some degree of independence between
consecutive values.
 RANDOM NUMBER GENERATION 7

(a) (b)

1.0
1.0

0.8
0.8

0.6
0.6
Xi+1

Xi+1
0.4
0.4

0.2
0.2

0.0
0.0

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
Xi Xi
(c) (d)

1.0
1.0

0.8
0.8
0.6

0.6
Xi+1

Xi+1
0.4

0.4
0.2

0.2
0.0

0.0

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0

Xi Xi

Figure 1.1 Scatter plots to illustrate the correlation between consecutive outputs
X i and X i+1 of different pseudo random number generators. The random number
generators used are the runif function in R (a), the LCG with m = 81, a = 1 and
c = 8 (b), the LCG with m = 1024, a = 401, c = 101 (c) and finally the LCG with
parameters m = 232 , a = 1 664 525, c = 1 013 904 223 (d). Clearly the output in the
second and third example does not behave like a sequence of independent random
variables.

One way to quantify the independence of the output samples of a PRNG is the
following criterion.

Definition 1.7 A periodic sequence (X n )n∈N with values in a finite set S and
period length P is k-dimensionally equidistributed, if every possible subsequence
x = (x1 , . . . , xk ) ∈ S k of length k occurs equally often in the sequence X , that is if
  
N x = # i  0 ≤ i < P, X i+1 = xi , . . . , X i+k = xk

does not depend on x.

A random number generator is good, if the output is k-dimensionally equidis-

tributed for large values of k.
Discovering Diverse Content Through
Random Scribd Documents
The Project Gutenberg eBook of The rebellion
of the Princess
 This ebook is for the use of anyone anywhere in the United
States and most other parts of the world at no cost and with
almost no restrictions whatsoever. You may copy it, give it away
or re-use it under the terms of the Project Gutenberg License
included with this ebook or online at www.gutenberg.org. If you
are not located in the United States, you will have to check the
laws of the country where you are located before using this
eBook.

Title: The rebellion of the Princess

Author: Mary Imlay Taylor

Release date: May 12, 2024 [eBook #73608]

Language: English

Original publication: United States: McCLure, Phillips & Co, 1903

Credits: D A Alexander, David E. Brown, and the Online

Distributed Proofreading Team at
https://www.pgdp.net (This file was produced from
images generously made available by The Internet
Archive)

*** START OF THE PROJECT GUTENBERG EBOOK THE REBELLION

OF THE PRINCESS ***
THE REBELLION OF THE
PRINCESS
 OTHER BOOKS
BY M. IMLAY TAYLOR

On the Red Staircase

The House of the Wizard
The Cardinal’s Musketeer
An Imperial Lover
A Yankee Volunteer
Anne Scarlett
The Cobbler of Nimes
Little Mistress Goodhope and Other Fairy Tales
The Rebellion of
the Princess
By
M. Imlay Taylor

Mc. Clure, Phillips & Co.

New York
1903
 Copyright, 1903, by

McCLURE, PHILLIPS & CO.

Published, March, 1903, R

 CONTENTS
chapter page
I. The Major-domo’s Whip, 1
II. The Miniature, 14
III. The Boyar Kurakin, 23
IV. The Making of a Friend, 35
V. The Princess Daria, 45
VI. The Dwarf, 56
VII. The Summons, 66
VIII. The Great Czarevna, 73
IX. I Make a Prisoner, 81
X. In the Garden of the Kremlin, 89
XI. The Plot Thickens, 98
XII. Advotia as an Interpreter, 104
XIII. The Tocsin, 110
XIV. A Desperate Climb, 121
XV. The Princess and the Czarevna, 129
XVI. The Painted Gallery, 138
XVII. Crowned with Rue, 147
XVIII. An Hour of Peril, 155
XIX. At Nightfall, 163
XX. The Escape, 174
XXI. The Steward’s Revenge, 182
XXII. A Drunken Orgy, 189
XXIII. A Sprig of Rue, 199
XXIV. Galitsyn, 206
XXV. Michaud’s Repentance, 216
 XXVI. Maluta Buys Two Souls, 224
XXVII. “Is It Thou?” 234
XXVIII. The Hut on the Road, 243
XXIX. A Duel with Swords, 254
XXX. The Prince Voronin, 261
XXXI. Vassalissa, 268
XXXII. The Man with the Purple Scar, 277
XXXIII. I Sow Dissension, 285
XXXIV. A Boyar’s Funeral, 293
XXXV. The Dwarf and I, 301
XXXVI. The Princess, 309
XXXVII. The Woman, 316
THE REBELLION OF THE
PRINCESS
 I: THE MAJOR-DOMO’S WHIP
FROM my post at the window I could look down upon the court-yard
of the palace of the Boyar Kurakin. Although it was early in May, it
was a cold day in Moscow, and the sun shone obliquely into the
yard, cut off as it was by the walls of two houses. The black mud of
winter had not dried off the centre of the court, and there was ice in
the corner by the water-butts, and ice hung, too, on the north side
of the roof, under the eaves, like the ragged beard of the old man of
the north, Moroz Treskun, or the Crackling Frost, as the moujik
names him, while above, around the great chimney, a group of
ravens were huddled together in the sun, preening their plumage
and croaking now and then in a solemn fashion.
The boyar’s house was large, and shaped like a Greek cross, the
kitchen and the servants’ quarters opening on the court, which was
crowded now with the serfs, for the steward of the household was
giving one of the varlets a taste of the whip. The doors and windows
of the kitchen gaped wide, filled with curious spectators; some, I
fancied, half in sympathy with the poor rogue who squealed under
the lash, and others applauding the major-domo, whether from fear
or love I knew not. He was a burly fellow with a red head and a
short, red-bearded, fierce-eyed countenance, and had the serf by
the waistband with one giant hand and with the other he laid the
whip on his bare back, leaving a long welt across the brown flesh
with every cut. The slave howled and writhed, the whip cracked, the
spectators applauded or jeered, as fancy seized them, and then,
quite suddenly, there was a diversion.
The water-butts were in the corner at the steward’s back, and a
dwarf darted out from behind them, quick as a wasp, and cut at the
major-domo’s calves with a leather thong and was back under cover
before the big man could wheel around. And he, thinking that he
had cut his own legs with the long end of his lash, and furious at the
titter of the servants, laid it on the poor serf with redoubled venom
until the blood ran. Meanwhile the dwarf executed a weird dance of
triumph on the ice by the water-butts, mocking the steward in dumb
show, and beating an imaginary victim, his thin cheeks blown out
and his brows knotted, to the delight of his audience, thus furnished
with a double entertainment. He was one of those wretched little
creatures that haunted Moscow, the playthings and spies of the
courtiers, and he was unusually small, even for a dwarf, with a
strange pointed face, white and three-cornered, like a patch of
paper, and with great ears shaped like the leaves of a linden
standing out from his head as if upon stems; it was by these ears
that I always knew him afterwards, even in the crowd of court
midgets. Encouraged by the success and the private applause, the
little wretch darted out again and repeated the performance of
whipping the steward’s legs, while the men and women held their
sides with laughter, because the fat beast danced and swore and
lashed, like one beside himself.
But it was an ill jest for the rogue in his clutches, and, minded to
end their sport, I shouted to him, in Russ, to look behind the water-
butts for the wasp. The fat fool gaped at me in amazement, and the
dwarf, darting from his covert, was running full speed for the kitchen
before he spied him and made after him. But one of the men, willing
to save the little beast, no doubt for the sake of the laugh, tripped
the major-domo as if by accident, and down he went in the mud of
the court-yard, bellowing and splashing like a whale.
I laughed until the tears came into my eyes, and whether he heard
me, or thought I had some hand in it, I know not, but when he got
to his feet, all bedaubed with mud and green slime, he shook his
two great fists at me and shrieked defiance; at which I laughed the
more. His face grew as red as his beard.
“Come down, you dog of a tinsmith!” he shouted, cracking his whip
in the dirt. “Come down, and I’ll take the hide off your back!”
I laughed again; for the life of me, I could not be angry with the
wretch. His burly figure and his impotent rage only aroused my
contempt, and I heeded his threats and his gestures as little as I did
the mirth of the kitchen behind him. I know not how long he would
have continued his pantomime, if it had not been for another attack
of his inveterate enemy. While he was shouting at me I saw the
ravens rise suddenly from the roof with a whir of black wings, and
the dwarf came dancing along on the very verge of the eaves. He
had evidently dropped from the windows of the terem, the women’s
quarters, which there, as usual, occupied a separate upper story of
wood, which overlooked the flat roof of the wing. The little creature
executed a fandango over the steward’s head and then suddenly let
fly a pebble, with such accurate aim that he took the fat man fairly
under his left ear. He was alive to dangers now, however, and,
discovering his foe, started for the kitchen-door with a bound, while
the dwarf, waiting only for him to disappear, came sliding down over
the edge of the roof, and swinging by his long arms he dropped,
with marvellous agility, on the ledge of the window below, and from
there, swinging again, monkey-fashion, on the window of the lower
story, he finally dropped into the yard, amidst a burst of applause
from the serfs. Meanwhile, the major-domo, arriving at last at the
window over the roof, looked out in baffled fury, and seeing me still
at my post, cursed me in Russ and two or three other dialects. “O
meat for dogs!” he bellowed, “’tis through some signalling of yours,
and I’ll pay you for it! I will—by the beard of the Saint Nikolas of
Mojaïsk! May the black god smite you!”
I shrugged my shoulders and left the window in disgust. So he called
me meat for dogs and a tinsmith; that was the cream of the jest! By
Saint Denis, a tinsmith—I, Jéhan de Marle, Marquis de Cernay, an
officer of the household troops of Louis XIV., King of France, and
cousin to the Duc de Richelieu! Yet, after all, the varlet had some
reason for his gibe, for did I not figure in Moscow as the apprentice
of Maître le Bastien, the goldsmith of Paris? Ah, and thereby hangs a
tale!
Twelve months before, my evil star took me to Paris for Easter. I had
been in Normandy, on my estates, and had served in the Palatinate.
Before that I won distinction, under the very eye of the king, at
Ghent and Ypres, and the saying at court was that no service paid
out of his sight, while in it there was such a scramble that Spinola bit
the royal finger—when he saluted the king’s hand—to make his mark
among the herd of sycophants. But, as it happened, the king noticed
me without the bite, though afterwards I paid for the recognition.
It was then scarcely three years after the Peace of Nimequen;
France was on a pinnacle of glory; Strasburg had fallen without a
shot, and Catinet had entered Casale. King Charles the Second of
England had taken his wages with some grumblings, and retired
from the war, and the Prince of Orange had been forced to yield to
the Estates of Holland and conclude the peace; the King of France
held a line of towns from Dunkerque to the Meuse, and Spain was
disarmed. Louis had maintained the war against Europe and was
victorious; “singly against all,” as Louvois said. It was a season of
glory and joy for every Frenchman, and especially for every French
soldier.
But what of it? What if fortune seemed to smile, and the rewards of
courage were within my grasp; what of it? I say. My evil star took
me to Paris, and all the world was at the festival. Mme. de
Montespan, the king’s mistress, was at dagger’s point with Mme. de
Maintenon, who was the governess of her children and the rising
star, as all the world knew; for “the Star of Quanto,” as they called
Mme. de Montespan, was near to setting, though she could still
afford to lose and win again, four millions—in one night—at basset.
There also was M. le Vicomte d’Argenson, taken by his evil star—a
deadly evil one it proved—to Paris and to me. He was cousin to
Mme. de Montespan, and as black-hearted a knave as ever wore a
velvet coat and clean ruffles at court, and that, as I would have you
know, is saying much. Ah, well! monsieur and I were in Paris, and
’twas Easter week, and Mme. de Montbazon gave a ball at the Hôtel
de Montbazon. It was one of the most magnificent fêtes in Paris;
wine flowed in the kennel of the Rue de Bethisi, so they said, and
madame gave a silver lily to each of her guests, while Vatel himself
was superintending in the kitchen. The lily for the young Duchess of
Burgundy was of gold set with pearls and diamonds. The world was
there, great and small, and one little maid from Provence, a
dependent of the Princess de Condé, country-bred and honest, as I
chanced to know, although she had an old hag of a mother who
would have sold her soul to make a fine match for her daughter, and
had even been to that great man, Bontemps, the king’s chief valet,
to inquire about the possibilities of securing a rich husband. But that
is neither here nor there.
It all happened at the very height of the ball, and it was thus I lost
my silver lily. I was on the grand staircase, and at the landing was
M. d’Argenson, with a throng of rufflers, waiting for the king. And, at
the moment, as ill luck would have it, the little maid from Provence,
Mlle. Lamoignon, came up the stairs, her face aglow with pleasure
and looking, as I thought, not unlike a Provençal rose herself. Satan
being in the heart of M. le Vicomte, doubtless it was his prompting
that made the man go out, before us all, to meet the child and try to
kiss her; at which she cried out, resisting with all her might, and the
beaux on the landing laughed. M. d’Argenson, being in liquor and
angered, I take it, by the titter behind him, turned on the girl and
grossly insulted her before us all. I was but two steps above them
and, quick as a flash, I caught monsieur by the shoulders, and
flinging him back against the wall with one hand, with the other I
slapped my glove in his face. D’Argenson was a mixture of bully and
coward, and had his sword out in a trice, and was at me, the others
crying to us that the king was coming. But I caught his rapier and,
breaking it across my knee, flung the fragments over the balustrade
with a gibe, and he, with the face of a fury, cursed me, standing on
the same step, while little mademoiselle cowered under my arm like
a frightened pigeon.
“Monsieur will pay for it—with blood!” screamed M. le Vicomte,
growing purple above his cravat.
“Pish!” I retorted, laughing in his face. “Jéhan de Cernay cares not
for vermin.”
“Coward!” he said, and struck me on the side where mademoiselle
cowered, so that I could not ward off the blow, and it slanted on my
cheek.
Then the devil rose in me; I thrust her away, and catching him about
the waist, flung him headlong on the stairs, just as the ushers in the
lower hall began to shout, “The king, the king!”
M. de Mazarin and M. de Besanvel, my friends, hustled me off out of
sight, and there was pandemonium on the staircase! Mme. de
Montbazon furious and in tears because of the fracas, Mlle.
Lamoignon hysterical, and M. le Vicomte, with a bruised head and a
black eye, shrieking for vengeance. To make a long story short, the
next morning I received monsieur’s cartel at my lodgings, and being
privately warned by M. de Mazarin that the king was angry and I
might look for the provost-marshal, I lost no time in choosing the
hour and the weapons. We fought that day in the Place Royale with
swords. What would you? I was accounted one of the best
swordsmen in France, and I had the advantage of being indifferent.
M. de Besanvel was with me and M. de Palisot with him. So far my
evil star shone propitious and sparkled, for monsieur’s nerves were
unstrung and his head sore.
I remember the scene quite well. The spring was forward; it was
Thursday in Easter week, and the trees were feathery with green
and the violets bloomed. ’Twas afternoon, and long shadows fell
aslant the green turf and the sun was warm. Monsieur, stripped of
coat and waistcoat, confronted me in a white ruffled shirt and
trousers of blue satin, with ruffles of point de Venise, and silk
stockings and red-heeled slippers. I saw his bloodshot eyes and his
purple lips, and we crossed swords, while M. de Besanvel engaged
M. de Palisot. It was not long; I spitted him at the second round—
my famous thrust over the guard—and I saw him die without regret
—vermin!
That was the end of it. We left him in the arms of the surgeon and
M. de Palisot, who got but a scratch from Besanvel, and I rode post-
haste from Paris with his majesty’s provost-marshal at my heels—
and all for a girl I did not know. Saint Denis, such is life!
It seemed that Mme. de Montespan, the handsome she-devil, was
hot for my ruin, and would give the king no rest; so Paris would not
hold me, nor Normandy, nor France. In this dilemma I bethought me
of Maître le Bastien, the goldsmith, then on his way to Moscow,
summoned thither by Prince Basil Galitsyn. Maître le Bastien was my
father’s friend and mine, and one whom I had benefited in more
ways than one; to him therefore I went. Was not a journey to Russia
and, mayhap, an adventure or two, better than a dull exile over
seas? To protect Maître le Bastien from trouble, I travelled under an
assumed name; I had the passports of his apprentice, Raoul,—who
fell ill of the small-pox, the week before we left Paris,—and no one
suspected my disguise unless it was the little varlet, Michaud, who
hated me from the first. Thus out of Paris, and its envy and favour, I
dropped into the northern capital, and found it less interesting than I
had hoped—which shows that a man sees but an inch beyond his
own nose.
I had been in Moscow now nearly a year, and the Czar Feodor was
just dead and the two factions—the Naryshkins and the Miloslavskys
—were quarrelling to the knife over the succession to the throne,
and the quarrel was all the more bitter because it was a family one.
It came about in this way. The Czar Alexis the Débonair married first
a Miloslavsky, by whom he had several children, among them the
Czar Feodor, his successor, but just dead; then Alexis had married a
second wife, the young and beautiful Natalia Naryshkin, who
became the mother of a boy and a girl. At the death of Feodor his
natural successor would have been his own brother, Ivan, but Ivan
was weak-minded and blind, and the Patriarch and the Naryshkins
stirred up the populace to elect Natalia’s boy, Peter, a lad of nine.
But the victory, though apparently easy, was destined to bear black
fruit, for behind Ivan, the idiot, was his clever and daring sister, the
Czarevna Sophia, who wanted the throne herself, and supporting her
was her clever cousin, Ivan Miloslavsky, and Prince Basil Galitsyn,
one of the most enlightened of the young Russian statesmen. And
the balance of power seemed to be for the time with the Streltsi, or
national guard, the only military organization of Russia, and both
parties were intriguing with the soldiers, who, dissatisfied with their
officers, their pay in arrears, and some of their hereditary privileges
threatened by political changes, were ripe for mischief. Trouble
growled deep and loud in the lanes and alleys of Moscow; in the
palaces and the hovels of its three towns were whisperings, and
terror, and intrigue.
But little I cared for all this, and time hung heavy on my hands, for I
had many dull hours, and it was in one of these that I watched the
dwarf torment the steward, and found the scene amusing.
I was still pacing the workshop in an idle mood when Michaud, the
apprentice, found me.
“Monsieur,” he said, with his air of knowing more than he chose to
tell, “two ladies are below, determined to see the master.”
“Of what sort, Michaud?” I asked; “old or young, fair or fat?”
“How can I tell, monsieur,” he replied, with a shrug, “they are
hooded as close as an ugly nun.”
I laughed.
“Maître le Bastien shall not have all the fun,” I said; “let them come
up, Michaud, and not a word to tell them I am not the master
goldsmith.”
He gave me an odd look and went out, and presently I heard his
step again on the stairs, and with it the rustle of skirts and the
sound of soft laughter.
“So!” I said to myself, “the jest is not all on one side.”
 II: THE MINIATURE
MICHAUD opened the door and stood back to admit my visitors,
casting another look of intelligence at me. But the two did not enter
at once; instead, there was much ado, whispers and suppressed
laughter in the hall, one hanging back and one pushing forward,
until my curiosity was alive, and I stood waiting with my eyes on the
door. At last, with another ripple of laughter, they came in; two slight
figures, muffled in the long, straight Russian cloaks, fur-edged, with
conical hoods over their heads, their features as completely
concealed as any nun’s of Port Royal. Determined to play my rôle of
goldsmith to the life, I had hastily picked up a mallet and a bit of
beaten gold, and, with these in my hands, I made a becoming
obeisance. Both the cloaked figures responded, and here at once I
noted a difference between them which no similarity of dress could
disguise: the taller of the two inclined her hooded head with the air
of a queen, the smaller one nodded at me with a suggestion of
infinite good humour. They remained silent,—struck dumb, no doubt,
at their own daring,—and we three stood confronting each other
without a word. It was evident that the pause might be eternal, and
I heard Michaud shuffling his feet outside the door; the rogue was
listening. I had learned to speak Russian fairly well and I called it to
my aid.
“How can I serve you, madame?” I said, awkwardly enough, I
suspect, for the shorter girl tittered, while the taller one silenced her
with a gesture, and addressed me in excellent French.
“You are a goldsmith, monsieur,” she said, in a clear voice, her
accent sweet rather than harsh. “I would have this locket opened.”
As she spoke she held out a gold locket and chain which she had
been hiding under her cloak. A glance told me that it was of great
value, and a rare piece of workmanship, encrusted with precious
jewels, and shaped like a pear. I took it gingerly, knowing no more of
a goldsmith’s trade than an unborn babe, and fairly caught in my
own trap. Whether she saw my awkwardness or not, I could not tell,
but she drew back a little, seeming to examine me with curious
eyes. I suddenly remembered my hands, when I became aware that
both girls were looking at them; my signet was on my right hand,
and their sharp eyes had discovered it, beyond a doubt; but what of
it? They knew nothing of French heraldry—or as little as I knew of
them, and I was more anxious than ever to peep under those hoods.
Meanwhile, in spite of my busy thoughts, I was trying in vain to find
an opening in the trinket. It showed not a crevice, but lay in my
hand, a marvellous golden pear, gleaming with rubies and diamonds
and sapphires, and with a crest that I could not decipher on its lower
end.
“It baffles you, monsieur,” remarked the taller maiden, a trifle coldly.
The perspiration gathered on my brow; what, in the name of the
saints, could I do with it? And I was figuring as a master goldsmith
with the abominable thing lying sealed in my hand. The smaller
nymph began to shake with laughter again under her cloak.
“’Tis magic, Daria,” she said, with the merriest laugh in the world,
her hood slipping back enough to disclose the rosy, roguish face of a
girl of sixteen or seventeen, with a pair of eyes as blue as the sky.
“I will have it open for all that,” retorted her companion imperiously.
“Monsieur, there is a secret spring.”
“Precisely, mademoiselle,” I replied, with a bow, “so secret that ’twill
not confide in a stranger.”
At this both laughed a little, but I saw that mademoiselle the
imperious was growing impatient, and, in desperation, I turned the
locket over and over, and as I did so my eye caught sight of the
Russian Imperial arms on the small end of the pear, where a golden
clasp represented the stem. In twisting the trinket thus in my fingers
I must have pressed a spring, for lo! the pear fell apart and
mademoiselle clapped her hands.
“The problem is solved,” she cried, while both of them craned their
necks to look at the two pieces.
These already riveted my attention; in one side was a lock of hair
and in the other a miniature that no one in Moscow could mistake,
flattered though it was. It was the face of the dead Czar’s sister, her
serene highness Sophia Alexeievna. There was an exclamation,
either of surprise or pleasure, from one of the girls, and as I cast a
covert glance at them I discovered that both hoods had been slightly
displaced, and I saw the features of the taller of the two. Saint
Denis, what a face! Young, beautiful, with the spirit of an empress;
the dark eyes, keen and brilliant, the lips and cheeks deeply
coloured, the brows sharply defined, the forehead like milk. My
glance was so searching and so earnest that mademoiselle looked
up and, encountering it, flashed me a look of such hauteur as I had
never before seen in the eyes of woman, but she disdained to draw
her hood. Meanwhile, the smaller and merrier beauty had given
away to delight at the adventure.
“Take out the portrait, monsieur,” she said; “I have one here to put
in its stead.”
“Nay,” interposed Mlle. Daria. “I will have none of it, Lissa; the jest
has gone too far.”
“Daria, Daria!” cried the other, forgetful of me, “thou art afraid! thou,
Daria Kirilovna!”
“I am not!” cried mademoiselle with defiance, tossing her head; “but
I despise the trick.”
“Oh, sweetheart, thou——” Lissa broke off under a lightning glance
from the dark eyes, for Mlle. Daria had remembered me.
But the merry damsel was not to be silenced; plucking at her
companion’s cloak, she drew her off into the corner and whispered,
and laughed, and entreated, apparently between jest and earnest,
while I pretended to examine the miniature, all the while cudgelling
my brains for a solution of this escapade, so rare was it for girls to
be out on an adventure in Moscow, and girls too, of rank, for no one
could doubt that who looked at them and heard them speak.
Meanwhile Daria had been melting under the persuasion of the fair
manœuvrer, and she came back slowly across the room, permitting
rather than encouraging Lissa, who now took the lead.
“Prithee, monsieur,” she said,—she too, spoke French, though with a
strong accent,—“take out that portrait for us and substitute this.” As
she held out her hand her companion made a sudden motion as if to
snatch the bit of ivory from it, but restrained herself and let Lissa
hand me a miniature.
Then I understood mademoiselle’s hesitation, for the face limned on
the ivory, more or less faithfully, was her own. Suppressing my
surprise, I put it down on a table and began the delicate task of
lifting the other miniature from its setting, and a task it was for my
awkward fingers. With no knowledge of such baubles, and as little
dexterity as a bear, I fully expected to break the picture in pieces,
but, as luck would have it, either the ivory was already loose in its
setting, or I again hit upon some secret spring, and out fell Sophia,
just escaping annihilation by falling on Maître le Bastien’s taffety
cloak that lay on the table. But now was the rub, for I had no notion
of how I should set mademoiselle’s face in the room where Sophia’s
had been, and both girls hung on my movements with breathless
interest. I took up the bit of ivory with a gingerly touch and
cautiously dropped it into the gaping setting, and lo! success beyond
my wildest hopes. It seemed to sink into place, as if by magic, and
Mlle. Lissa clapped her hands with delight.
“Good goldsmith!” she cried, beaming upon me. “What a fair
exchange!”
“Hush, Vassalissa!” commanded Mlle. Daria; “for shame!”
But Lissa would not be suppressed.
“And is it not?” she cried mischievously. “Ah, bah; what a fright!” and
she pointed derisively at Sophia’s portrait. “Come, come, Daria, let
us have our frolic while we may!”
“Exactly so, while we may!” retorted Daria grimly; “but afterwards,
my dear,” and she smiled a little.
“The deluge,” replied Vassalissa, laughing. “Ah, good master
goldsmith, give us the trinket that we may get into the ark.”
But here was the difficulty; I could not fasten the miniature in place,
nor could I for the life of me close the locket. The pear was twain
and like to be so, as far as I could see, to the end of the world, and
Mlle. Daria began to cast suspicious glances at me. I think, for the
second time, she doubted that I was a goldsmith.
“Time presses, monsieur,” she said imperiously; “let us have it, as
speedily as may be.”
I was red in the face and almost out of temper, but I saw no escape.
“Mademoiselle must leave it with me,” I replied as blandly as I could;
“it will take time to secure the portrait and reclasp the locket.”
“Impossible!” said Daria; “we must have it now, monsieur; the
matter is imperative.”
I saw that she was uneasy, and I thought that Vassalissa was a little
alarmed; both girls pressed forward eagerly.
“We must have it!” they protested.
I took the bull by the horns. “Certainly, mademoiselle,” I said with a
bow, “but it will not be completed or fastened,” and I held out the
two pieces of that ill-starred pear with a malicious smile.
They looked at each other and at me for a moment with blank faces,
and then they broke out with irresistible, delicious, rippling laughter.
“What on earth shall we do?” cried Vassalissa; “the deluge and no
ark! Monsieur, we have a fable that when the Evil One, in the form
of a mouse, gnawed a hole in the ark, Uzh, the snake, saved the
ship by thrusting his head into the place. Find us a snake therefore,
good goldsmith, or our ark will surely sink. Mend us the pear, or——”
“Pshaw!” interrupted Mlle. Daria, with an imperious gesture, “what
difference? I care not a straw! Finish it, monsieur, and send it to me
at your leisure.”
“Daria!” sharply ejaculated her smaller companion, suddenly grown
cautious.
And Daria bit her lip and turned crimson.
“Mademoiselle may trust me,” I said, drawing myself up to my full
height, which compelled them both to look up at me.
She gave me a swift, penetrating glance, and her face, by nature
haughty, suddenly relaxed and a smile, like sunshine, shone on it.
“I do, monsieur,” she said, with her queenly air. “You will send the
locket, by a safe hand, to the house of the Prince Voronin, to be
delivered only to me—the Princess Daria.”
Her companion fairly gasped, her blue eyes big with amazement, at
mademoiselle’s daring.
“I will bring it with my own hand,” I said, with a profound bow.
And, as I spoke, there was a sharp knock at the door. Vassalissa
started with a little shriek of nervous excitement, but Daria laughed.
“’Tis old Piotr,” she said.
As she spoke, the door opened and a tall, grey-haired Russian,
wearing the dress of a boyar’s retainer, stood on the threshold.
“We have been here too long, little mistress,” he said in Russ,
respectful, but impatient; “’tis neither safe nor wise.”
“Bear with us, Piotr,” said his mistress graciously; “’tis but a half hour
under a whole moon; may not the children play?”
He shook his head, glancing with evident affection at the tall, girlish
figure.
“Time waits for no man, Daria Kirilovna,” he said gravely, “and the
morning is wiser than the evening.”[A]
“I come, I come!” she retorted, and with a gesture of farewell to me,
she left the room, followed by Lissa, who cast a mischievous smile at
me, and a doubtful glance at the trinket in my hands as she went
out.
 III: THE BOYAR KURAKIN
LEFT alone with the trinket, I forgot it in my meditation on the two
girls, or rather, if the truth be told, on the one—the Princess Daria.
Such beauty, such spirit, such dignity; the combination was rare, and
in a Russian, brought up no doubt under the iron rule of some old
Russian dragon of propriety, it was little short of a miracle. How
came this perfect flower to bloom in a waste of snow? And how
came she and the merry one on this strange expedition? There was
some mischief afoot, but I could not fathom it, cudgel my brains as I
would. They both seemed too young and too artless to be engaged
in any very profound intrigue, and yet the portrait of the czarevna
was an unusual possession to cast lightly and publicly aside; publicly,
I say, because I was a stranger to them and might be, for all they
knew, quite unworthy of trust. And how did they escape the vigilant
watchfulness of a Russian household, where the women were kept
in almost Oriental seclusion? It was true that the Czar Alexis the
Débonair had modified the customs of the court in this respect, by
the freedom he had allowed his young wife, Natalia Naryshkin, the
mother of the newly elected Czar Peter. Yet it was undoubtedly an
escapade for two Russian girls to visit the workshop of a stranger
and a Frenchman, for the nation had no love for the French, and
indeed a deep distrust of all foreigners.
But what of it, after all? I reflected, was it not better to remember
the two pretty faces, the slender hands, the soft voices, the ripple of
merry laughter? Saint Denis! ’twas worth something to have seen
them! And I would see them again unless Jéhan de Cernay had
assumed a coat of quite another colour from the one he had worn in
France. As for Daria, she might well be a princess; she looked it, and
no queen was ever more worthy a crown.
How she had graced even Maître le Bastien’s workshop, and
transformed the old room into an enchanted palace! I looked about
it now with a shrug; since she had left it, it had returned to its usual
aspect, and was a workshop again and no more.
The house that Prince Galitsyn had given to Maître le Bastien stood
in the Kitai-gorod, with the bazaars on one side, humming with life,
like so many beehives, and on the other the palaces of the boyars,
the official nobility of Moscow; and yonder were the golden domes
and minarets of the Kremlin. The house itself was much like the
others in Moscow—built of logs, the interstices stuffed with tow, and
the roof also of wood; it was no marvel that there had been great
fires, leaping from town to town, within the walls, and carrying
terror and destruction with smoke and flame. Underground we had
cellars for storing liquors and ice; and above these, on the ground
floor, were the kitchen, refectory, and offices, while on the second
floor were always the living rooms; the Chamber of the Cross, or
private chapel, being in the centre, and a narrow stair led to the
apartments above, usually set aside for the women, in a separate
story of the house, and called the terem.
It was on the second floor that Maître le Bastien had his workshop,
in a long room that had served as a nursery and playroom for the
children of the Russian family who had previously occupied the
dwelling. The windows faced north, and the room was well lighted
and spacious, but very different from the goldsmith’s famous
workshop on the Pont-au-Change, where all the lovers of his art in
Paris flocked. I have seen Louvois there, and Luxemburg himself,
with his hump and his pale face, and Monseigneur, dull and
pompous, and the little Duchess of Burgundy with Mme. de
Maintenon, then called the widow of Scarron, and the court ladies,
Mme. de Mazarin and Mme. de Richelieu, and hundreds of others,
and sometimes the great king himself. It was Le Bastien who made
the famous bracelet for Mme. de Montespan, and Le Bastien who
designed the great candelabra for the king’s table. It was the silver
vase that he had made for Louis that he was to copy now for Prince
Galitsyn to give, so it was whispered, to the Czarevna Sophia, she
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

ebookfinal.com