Simulating Copulas Stochastic Models Sampling

Algorithms and Applications 2nd Edition Edition

Jan-Frederik Mai pdf download

https://ebookname.com/product/simulating-copulas-stochastic-
models-sampling-algorithms-and-applications-2nd-edition-edition-
jan-frederik-mai/

Get Instant Ebook Downloads – Browse at https://ebookname.com

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

Simulating Copulas Stochastic Models Sampling

Algorithms and Applications 2nd Edition Edition Jan-
Frederik Mai

https://ebookname.com/product/simulating-copulas-stochastic-
models-sampling-algorithms-and-applications-2nd-edition-edition-
jan-frederik-mai/

Handbook of parallel computing models algorithms and

applications 1st Edition Sanguthevar Rajasekaran

https://ebookname.com/product/handbook-of-parallel-computing-
models-algorithms-and-applications-1st-edition-sanguthevar-
rajasekaran/

An Introduction to Continuous Time Stochastic Processes

Theory Models and Applications to Finance Biology and
Medicine 1st Edition Vincenzo Capasso

https://ebookname.com/product/an-introduction-to-continuous-time-
stochastic-processes-theory-models-and-applications-to-finance-
biology-and-medicine-1st-edition-vincenzo-capasso/

Radiation Tolerant Electronics Volume II 1st Edition

Paul Leroux

https://ebookname.com/product/radiation-tolerant-electronics-
volume-ii-1st-edition-paul-leroux/
The Bass Handbook of Leadership Theory Research and
Managerial Applications Bernard M. Bass

https://ebookname.com/product/the-bass-handbook-of-leadership-
theory-research-and-managerial-applications-bernard-m-bass/

Energy Budgets at Risk EBaR A Risk Management Approach

to Energy Purchase and Efficiency Choices Wiley Finance
1st Edition J. Jackson

https://ebookname.com/product/energy-budgets-at-risk-ebar-a-risk-
management-approach-to-energy-purchase-and-efficiency-choices-
wiley-finance-1st-edition-j-jackson/

Evidence Matters Randomized Trials in Education

Research Frederick Mosteller

https://ebookname.com/product/evidence-matters-randomized-trials-
in-education-research-frederick-mosteller/

Don t Panic 5th Edition William H. Bassichis

https://ebookname.com/product/don-t-panic-5th-edition-william-h-
bassichis/

Ace Your Case Consulting Interviews 2nd Edition Wetfeet

https://ebookname.com/product/ace-your-case-consulting-
interviews-2nd-edition-wetfeet/
Digital Signal and Image Processing using MATLAB Volume
1 Fundamentals 2nd Edition Gérard Blanchet

https://ebookname.com/product/digital-signal-and-image-
processing-using-matlab-volume-1-fundamentals-2nd-edition-gerard-
blanchet/
 2nd Edition
SIMULATING
COPULAS
Stochastic Models,
Sampling Algorithms, and Applications

10265hc_9789813149243_tp.indd 1 16/5/17 2:45 PM

 Series in Quantitative Finance ISSN: 1756-1604

Series Editor: Ralf Korn (University of Kaiserslautern, Germany)

Editorial Members: Tang Shanjian (Fudan University, China)

Kwok Yue Kuen (Hong Kong University of Science
and Technology, China)

Published
Vol. 1 An Introduction to Computational Finance
by Ömür Uğur

Vol. 2 Advanced Asset Pricing Theory

by Chenghu Ma

Vol. 3 Option Pricing in Incomplete Markets:

Modeling Based on Geometric Lévy Processes and
Minimal Entropy Martingale Measures
by Yoshio Miyahara

Vol. 4 Simulating Copulas:

Stochastic Models, Sampling Algorithms, and Applications
by Jan-Frederik Mai and Matthias Scherer

Vol. 5 Extreme Financial Risks and Asset Allocation

by Olivier Le Courtois and Christian Walter

Vol. 6 Simulating Copulas:

Stochastic Models, Sampling Algorithms and Applications
(Second Edition)
by Jan-Frederik Mai and Matthias Scherer

Alisha - 10265 - Simulating Copulas.indd 1 16-05-17 3:18:47 PM

 Series in Quantitative Finance – Vol. 6

2nd Edition
SIMULATING
COPULAS
Stochastic Models,
Sampling Algorithms, and Applications

Jan-Frederik Mai
XAIA Investment AG, Germany

Matthias Scherer
Technische Universität München, Germany

with contributions by

Claudia Czado • Elke Korn • Ralf Korn • Jakob Stöber

World Scientific
NEW JERSEY • LONDON • SINGAPORE • BEIJING • SHANGHAI • HONG KONG • TAIPEI • CHENNAI • TOKYO

10265hc_9789813149243_tp.indd 2 16/5/17 2:45 PM

 Published by
World Scientific Publishing Co. Pte. Ltd.
5 Toh Tuck Link, Singapore 596224
USA office: 27 Warren Street, Suite 401-402, Hackensack, NJ 07601
UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE

British Library Cataloguing-in-Publication Data

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

Series in Quantitative Finance — Vol. 6

SIMULATING COPULAS
Stochastic Models, Sampling Algorithms and Applications
Second Edition
Copyright © 2017 by World Scientific Publishing Co. Pte. Ltd.
All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means,
electronic or mechanical, including photocopying, recording or any information storage and retrieval
system now known or to be invented, without written permission from the publisher.

For photocopying of material in this volume, please pay a copying fee through the Copyright Clearance
Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to photocopy
is not required from the publisher.

ISBN 978-981-3149-24-3

Desk Editor: Alisha Nguyen

Typeset by Stallion Press

Email: [email protected]

Printed in Singapore

Alisha - 10265 - Simulating Copulas.indd 2 16-05-17 3:18:47 PM

May 17, 2017 16:28 Simulating Copulas - 9in x 6in b2921-Main page v

To the new generation:

Viola, Lara, Fabian, and Paul.

v
 b2530   International Strategic Relations and China’s National Security: World at the Crossroads

This page intentionally left blank

b2530_FM.indd 6 01-Sep-16 11:03:06 AM

May 17, 2017 16:28 Simulating Copulas - 9in x 6in b2921-Main page vii

Preface

The joint treatment of d ≥ 2 random variables requires vector-valued

stochastic models. In the financial industry, multivariate models are ap-
plied to, e.g., asset allocation problems (portfolio optimization), the pricing
of basket options, risk management, and the modeling of credit portfolios.
In particular, the development during the past years highlighted that the
financial industry is in urgent need of realistic and viable models in large
dimensions. Other fields of application for multivariate stochastic mod-
els include geostatistics, hydrology, insurance mathematics, medicine, and
reliability theory.
Besides specifying the univariate marginals, for multivariate distribu-
tions it is additionally required to appropriately define the dependence
structure among the modeled objects. In most applications, a portfolio per-
spective is significantly more demanding compared to modeling univariate
marginals. One consequence is that analytical solutions for the aforemen-
tioned applications can typically be derived under restrictive assumptions
only. An increasingly popular alternative to accepting unrealistic simplifi-
cations is to solve the model in question by Monte Carlo simulation. This
allows for very general models but requires efficient simulation schemes for
multivariate distributions. This book aims at providing a toolbox for the
simulation of random vectors with a considerable spectrum of dependence
structures.

vii
May 17, 2017 16:28 Simulating Copulas - 9in x 6in b2921-Main page viii

viii Simulating Copulas: Stochastic Models, Sampling Algorithms, and Applications

Why Sampling Copulas?

This book focuses on sampling copulas, i.e. distribution functions on [0, 1]d
with uniform univariate marginals. At first glance, this standardization
to univariate margins seems to be a rather artificial assumption. The
justification for considering copulas instead of more general multivariate
distribution functions is provided by Sklar’s seminal decomposition (see
Sklar (1959) and Section 1.1.2). Heuristically speaking, Sklar’s theorem
allows us to decompose any d-dimensional multivariate distribution func-
tion F into its univariate margins F1 , . . . , Fd and the dependence structure
among them. The latter is described by the copula behind the model, de-
noted C. More precisely, we have F (x1 , . . . , xd ) = C(F1 (x1 ), . . . , Fd (xd ))
for (x1 , . . . , xd ) ∈ Rd . The converse implication also holds, i.e. coupling uni-
variate margins with some copula yields a multivariate distribution. This
observation is especially convenient for the specification of a multivariate
model, since a separate treatment of the dependence structure and uni-
variate margins is usually easier compared to specifying the multivariate
distribution in one step.
Sklar’s decomposition also applies to sampling applications. Assume
that we want to simulate from a multivariate distribution function F with
univariate marginal distribution functions F1 , . . . , Fd and copula C. Given
a sampling scheme for the copula C, the following algorithm generates
a sample from the distribution F by applying the generalized inverses
F1−1 , . . . , Fd−1 (see Lemma 1.4) to the sample of the copula.
Algorithm 0.1 (Sampling Multivariate Distributions)
Let F (x1 , . . . , xd ) = C(F1 (x1 ), . . . , Fd (xd )) be a d-dimensional distribution
function. Let sample C () be a function that returns a sample from C.
Sampling F is then possible via the following scheme:
FUNCTION sample F ()
Set (U1 , . . . , Ud ) := sample C ()


RETURN F1−1 (U1 ), . . . , Fd−1 (Ud )
May 17, 2017 16:28 Simulating Copulas - 9in x 6in b2921-Main page ix

Preface ix

Why Another Book on Copulas?

Our main motivation for writing this book was to summarize the fast-
growing literature on simulation algorithms for copulas. Several results on
new sampling techniques for classical copulas, e.g. the Archimedean and
Marshall–Olkin families, have lately been published. Moreover, new fam-
ilies and construction principles have been discovered; an example is the
pair-copula construction. At the same time, the financial industry has be-
come aware that copula models (beyond a Gaussian dependence structure)
are required to realistically model various aspects of quantitative finance.
This book takes account of this fact by providing a comprehensive toolbox
for financial engineering, and, of course, for other applications as well. All
algorithms are described in pseudo-code. Thus, they can easily be imple-
mented in the user’s preferred programming language. Moreover, we aim at
being comprehensive with respect to sampling schemes for univariate ran-
dom variables as well as with respect to the use of Monte Carlo sampling
engines in general. We purposely included sampling schemes for very basic
copulas, even though this might not be required for an expert in the field.
Another intention is to provide an elementary introduction to copulas from
the perspective of probabilistic representations. Hence, an experienced re-
searcher might skip some parts of the book. But someone who is new to
the field of copulas can use the book as a stand-alone textbook. The book,
however, does not treat statistical estimation of dependence models.
Especially for sampling applications, the dimension of the copula plays
a crucial role. To give an example, the original probabilistic model behind
the so-called d-dimensional Marshall–Olkin copula is based on 2d − 1 ran-
dom variables, i.e. the dimension d enters exponentially. Hence, this book
explicitly focuses on the d-dimensional case and discusses the efficiency of
the provided algorithms with respect to their dimension. Especially in the
field of portfolio credit risk modeling, there are some applications requiring
high-dimensional models with d = 125 or even more.
Copulas can be investigated from two (not necessarily disjoint) perspec-
tives: (1) analytically, i.e. viewing them as d-dimensional functions, and (2)
probabilistically, i.e. viewing them as the dependence structure behind some
random vector. Both perspectives have their distinct advantages.

(1) The analytical perspective aims at deriving statements about cop-

ulas from their functional form. This is especially successful for
analytically tractable families. In this case, it is often possible
May 17, 2017 16:28 Simulating Copulas - 9in x 6in b2921-Main page x

x Simulating Copulas: Stochastic Models, Sampling Algorithms, and Applications

to deduce very general dependence properties from the paramet-

ric form of the respective copula. For high-dimensional sampling
applications, however, this approach is of minor use, since the func-
tional form does not, in general, provide a convenient rule for the
construction of a sampling algorithm.
(2) Investigating copulas from a probabilistic perspective is based on
stochastic representations of the dependence structure. This means
that there is an explicit (and preferably simple) probability space,
on which the random vector associated with a copula is defined.
The immediate advantage is that such probabilistic constructions
provide a recipe for a sampling algorithm. Besides sampling, it is
also possible to investigate the copula from the respective repre-
sentation.
This book pursues a mostly probabilistic treatment. This is especially
suitable for sampling applications.

Acknowledgments

First of all, we would like to thank C. Czado, E. Korn, R. Korn, and

J. Stöber for providing the chapters on pair copulas, univariate sampling al-
gorithms, and Monte Carlo techniques. We also would like to thank several
friends and colleagues for patiently answering questions, reading previous
versions of the manuscript, and pointing us at several improvements. These
are K. Bannör, G. Bernhart, F. Durante, C. Hering, P. Hieber, M. Hofert,
H. Joe, D. Krause, A. Min, A. Reuß, D. Neykova, W. Richter, S. Schenk,
D. Selch, and N. Shenkman. Finally, we would like to thank our editor,
Tasha D’Cruz, for her extremely valuable feedback and professional han-
dling of the manuscript.
Jan-Frederik Mai and Matthias Scherer
May 17, 2017 16:28 Simulating Copulas - 9in x 6in b2921-Main page xi

Preface xi

Preface to the second edition

We were very pleased when ICP/WS approached us in early 2016 with

the proposal of a second edition of this book. For several reasons, we
immediately agreed to provide a revised and extended edition. First of all,
there was the opportunity to correct some typographical errors that we
(and others) noticed in the meanwhile. Second, we negotiated a soft-cover
version that allows the reader to acquire the book at a cheaper price. We
hope the book enjoys an even wider dissemination consequently. Third, and
most important, we felt we made significant progress in the main theme of
the book: the identification of conditionally independent (sub-)families of
multivariate distributions and the design of new parametric copula models.
This progress was supported by the remarkable PhD theses of German
Bernhart and Steffen Schenk.1 This strand of research particularly led to a
deeper understanding of multivariate extreme-value distributions and fatal-
shock models. We have chosen to add these results as a separate, additional
chapter. This preserves the structure of the first edition, so references to
theorems etc. are consistent across the two editions. Furthermore, we have
also monitored the fast growing literature on copulas and included new
results and references related to the simulation of copulas. We would like
to thank A. Hüttner and H. Sloot for providing valuable suggestions for
improvements.
Jan-Frederik Mai and Matthias Scherer, December 2016

1 Gentlemen, many thanks for the fruitful and pleasant collaboration!

 b2530   International Strategic Relations and China’s National Security: World at the Crossroads

This page intentionally left blank

b2530_FM.indd 6 01-Sep-16 11:03:06 AM

May 17, 2017 16:28 Simulating Copulas - 9in x 6in b2921-Main page xiii

Contents

Preface vii

1. Introduction 1
1.1 Copulas . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.1.1 Analytical Properties . . . . . . . . . . . . . . . . 7
1.1.2 Sklar’s Theorem and Survival Copulas . . . . . . 14
1.1.3 General Sampling Methodology in Low
Dimensions . . . . . . . . . . . . . . . . . . . . . . 22
1.1.4 Graphical Visualization . . . . . . . . . . . . . . . 26
1.1.5 Concordance Measures . . . . . . . . . . . . . . . 28
1.1.6 Measures of Extremal Dependence . . . . . . . . . 33
1.2 General Classifications of Copulas . . . . . . . . . . . . . 36
1.2.1 Radial Symmetry . . . . . . . . . . . . . . . . . . 36
1.2.2 Exchangeability . . . . . . . . . . . . . . . . . . . 39
1.2.3 Homogeneous Mixture Models . . . . . . . . . . . 41
1.2.4 Heterogeneous Mixture Models/Hierarchical Models 48
1.2.5 Extreme-Value Copulas . . . . . . . . . . . . . . . 52

2. Archimedean Copulas 57
2.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . 58
2.2 Extendible Archimedean Copulas . . . . . . . . . . . . . . 61
2.2.1 Kimberling’s Result and Bernstein’s Theorem . . 62
2.2.2 Properties of Extendible Archimedean Copulas . . 65
2.2.3 Constructing Multi-Parametric Families . . . . . . 69
2.2.4 Parametric Families . . . . . . . . . . . . . . . . . 69
2.3 Exchangeable Archimedean Copulas . . . . . . . . . . . . 76

xiii
May 17, 2017 16:28 Simulating Copulas - 9in x 6in b2921-Main page xiv

xiv Simulating Copulas: Stochastic Models, Sampling Algorithms, and Applications

2.3.1 Constructing Exchangeable Archimedean Copulas 82

2.3.2 Sampling Exchangeable Archimedean Copulas . . 85
2.3.3 Properties of Exchangeable Archimedean Copulas 87
2.4 Hierarchical (H-Extendible) Archimedean Copulas . . . . 89
2.4.1 Compatibility of Generators . . . . . . . . . . . . 90
2.4.2 Probabilistic Construction and Sampling . . . . . 91
2.4.3 Properties . . . . . . . . . . . . . . . . . . . . . . 93
2.4.4 Examples . . . . . . . . . . . . . . . . . . . . . . . 95
2.5 Other Topics Related to Archimedean Copulas . . . . . . 97
2.5.1 Simulating from the Generator . . . . . . . . . . . 97
2.5.2 Asymmetrizing Archimedean Copulas . . . . . . . 99

3. Marshall–Olkin Copulas 101

3.1 The General Marshall–Olkin Copula . . . . . . . . . . . . 102
3.1.1 Canonical Construction of the MO Distribution . 104
3.1.2 Alternative Construction of the MO Distribution 110
3.1.3 Properties of Marshall–Olkin Copulas . . . . . . . 118
3.2 The Exchangeable Case . . . . . . . . . . . . . . . . . . . 122
3.2.1 Reparameterizing Marshall–Olkin Copulas . . . . 126
3.2.2 The Inverse Pascal Triangle . . . . . . . . . . . . 129
3.2.3 Efficiently Sampling eM O . . . . . . . . . . . . . 131
3.2.4 Hierarchical Extensions . . . . . . . . . . . . . . . 138
3.3 The Extendible Case . . . . . . . . . . . . . . . . . . . . . 140
3.3.1 Precise Formulation and Proof of Theorem 3.1 . . 141
3.3.2 Proof of Theorem 3.2 . . . . . . . . . . . . . . . . 146
3.3.3 Efficient Simulation of Lévy-Frailty Copulas . . . 150
3.3.4 Hierarchical (H-Extendible) Lévy-Frailty Copulas 153

4. Elliptical Copulas 159

4.1 Spherical Distributions . . . . . . . . . . . . . . . . . . . . 161
4.2 Elliptical Distributions . . . . . . . . . . . . . . . . . . . . 166
4.3 Parametric Families of Elliptical Distributions . . . . . . . 170
4.4 Elliptical Copulas . . . . . . . . . . . . . . . . . . . . . . . 174
4.5 Parametric Families of Elliptical Copulas . . . . . . . . . 175
4.6 Sampling Algorithms . . . . . . . . . . . . . . . . . . . . . 179
4.6.1 A Generic Sampling Scheme . . . . . . . . . . . . 179
4.6.2 Sampling Important Parametric Families . . . . . 181
May 17, 2017 16:28 Simulating Copulas - 9in x 6in b2921-Main page xv

Contents xv

5. Pair Copula Constructions 185

5.1 Introduction to Pair Copula Constructions . . . . . . . . . 186
5.2 Copula Construction by Regular Vine Trees . . . . . . . . 191
5.2.1 Regular Vines . . . . . . . . . . . . . . . . . . . . 191
5.2.2 Regular Vine Matrices . . . . . . . . . . . . . . . 196
5.3 Simulation from Regular Vine Distributions . . . . . . . . 203
5.3.1 h-Functions for Bivariate Copulas and Their
Rotated Versions . . . . . . . . . . . . . . . . . . 204
5.3.2 The Sampling Algorithms . . . . . . . . . . . . . . 208
5.4 Dependence Properties . . . . . . . . . . . . . . . . . . . . 218
5.5 Application . . . . . . . . . . . . . . . . . . . . . . . . . . 223
5.5.1 Time Series Model for Each Margin . . . . . . . . 224
5.5.2 Parameter Estimation . . . . . . . . . . . . . . . . 224
5.5.3 Forecasting Value at Risk . . . . . . . . . . . . . . 226
5.5.4 Backtesting Value at Risk . . . . . . . . . . . . . 227
5.5.5 Backtest Results . . . . . . . . . . . . . . . . . . . 228

6. Sampling Univariate Random Variables 231

6.1 General Aspects of Generating Random Variables . . . . . 231
6.2 Generating Uniformly Distributed Random Variables . . . 232
6.2.1 Quality Criteria for RNG . . . . . . . . . . . . . . 233
6.2.2 Common Causes of Trouble . . . . . . . . . . . . 234
6.3 The Inversion Method . . . . . . . . . . . . . . . . . . . . 234
6.4 Generating Exponentially Distributed Random Numbers . 235
6.5 Acceptance-Rejection Method . . . . . . . . . . . . . . . . 235
6.6 Generating Normally Distributed Random Numbers . . . 238
6.6.1 Calculating the Cumulative Normal . . . . . . . . 238
6.6.2 Generating Normally Distributed Random
Numbers via Inversion . . . . . . . . . . . . . . . 238
6.6.3 Generating Normal Random Numbers with Polar
Methods . . . . . . . . . . . . . . . . . . . . . . . 239
6.7 Generating Lognormal Random Numbers . . . . . . . . . 240
6.8 Generating Gamma-Distributed Random Numbers . . . . 240
6.8.1 Generating Gamma-Distributed RNs with β > 1 . 241
6.8.2 Generating Gamma-Distributed RNs with β < 1 . 242
6.8.3 Relations to Other Distributions . . . . . . . . . . 243
6.9 Generating Chi-Square-Distributed RNs . . . . . . . . . . 243
6.10 Generating t-Distributed Random Numbers . . . . . . . . 244
May 17, 2017 16:28 Simulating Copulas - 9in x 6in b2921-Main page xvi

xvi Simulating Copulas: Stochastic Models, Sampling Algorithms, and Applications

6.11 Generating Pareto-Distributed Random Numbers . . . . . 245

6.12 Generating Inverse Gaussian-Distributed Random
Numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
6.13 Generating Stable-Distributed Random Numbers . . . . . 246
6.14 Generating Discretely Distributed Random Numbers . . . 247
6.14.1 Generating Random Numbers with Geometric
and Binomial Distribution . . . . . . . . . . . . . 248
6.14.2 Generating Poisson-Distributed Random
Numbers . . . . . . . . . . . . . . . . . . . . . . . 248

7. The Monte Carlo Method 251

7.1 First Aspects of the Monte Carlo Method . . . . . . . . . 251
7.2 Variance Reduction Methods . . . . . . . . . . . . . . . . 254
7.2.1 Antithetic Variates . . . . . . . . . . . . . . . . . 255
7.2.2 Antithetic Variates for Radially Symmetric
Copulas . . . . . . . . . . . . . . . . . . . . . . . . 257
7.2.3 Control Variates . . . . . . . . . . . . . . . . . . . 258
7.2.4 Approximation via a Simpler Dependence
Structure . . . . . . . . . . . . . . . . . . . . . . . 260
7.2.5 Importance Sampling . . . . . . . . . . . . . . . . 262
7.2.6 Importance Sampling via Increasing the
Dependence . . . . . . . . . . . . . . . . . . . . . 263
7.2.7 Further Comments on Variance Reduction
Methods . . . . . . . . . . . . . . . . . . . . . . . 265

8. Further Copula Families with Known Extendible Subclass 267

8.1 Exogenous Shock Models . . . . . . . . . . . . . . . . . . 268
8.1.1 Extendible Exogenous Shock Models . . . . . . . 271
8.2 Extreme-Value Copulas . . . . . . . . . . . . . . . . . . . 285
8.2.1 Multivariate Distributions with Exponential
Minima . . . . . . . . . . . . . . . . . . . . . . . . 293
8.2.2 Hierarchical (H-extendible) Extreme-Value
Copulas . . . . . . . . . . . . . . . . . . . . . . . . 294

Appendix A Supplemental Material 301

A.1 Validating a Sampling Algorithm . . . . . . . . . . . . . . 301
A.2 Introduction to Lévy Subordinators . . . . . . . . . . . . 302
A.2.1 Compound Poisson Subordinator . . . . . . . . . 306
May 17, 2017 16:28 Simulating Copulas - 9in x 6in b2921-Main page xvii

Contents xvii

A.2.2 Gamma Subordinator . . . . . . . . . . . . . . . . 308

A.2.3 Inverse Gaussian Subordinator . . . . . . . . . . . 309
A.2.4 Stable Subordinator . . . . . . . . . . . . . . . . . 310
A.3 Scale Mixtures of Marshall–Olkin Copulas . . . . . . . . . 311
A.4 Generalizations of Lévy Subordinators . . . . . . . . . . . 315
A.4.1 Additive Subordinators . . . . . . . . . . . . . . . 315
A.4.2 IDT Subordinators . . . . . . . . . . . . . . . . . 316
A.5 Further Reading . . . . . . . . . . . . . . . . . . . . . . . 319

Bibliography 323

Index 335
May 17, 2017 16:28 Simulating Copulas - 9in x 6in b2921-Main page 1

Chapter 1

Introduction

Before we start, let us clarify some notations.

General comment: The dimension of a random vector is typically
denoted by d ≥ 2.
Important sets: N denotes the set of natural numbers {1, 2, . . .}, and
N0 := {0} ∪ N. R denotes the set of real numbers. Moreover, for d ∈ N,
Rd denotes the set of all d-dimensional row vectors with entries in R. For
v := (v1 , . . . , vd ) ∈ Rd , we denote by v
its transpose. For some set A, we
denote by B(A) the corresponding Borel σ-algebra, which is generated by
all open subsets of A. The cardinality of a set A is denoted by |A|. Subsets
and proper subsets are denoted by A ⊂ B and A
B, respectively.
Probability spaces: A probability space is denoted by (Ω, F , P), with
σ-algebra F and probability measure P. The corresponding expectation op-
erator is denoted by E. The variance, covariance, and correlation operators
are written as Var, Cov, Corr, respectively. Random variables (or vectors)
are mostly denoted by the letter X (respectively X := (X1 , . . . , Xd )). As
an exception, we write U := (U1 , . . . , Ud ) for a d-dimensional random vec-
tor with a copula as joint distribution function.1 If two random variables
d d
X1 , X2 are equal in distribution, we write X1 = X2 . Similarly, → denotes
convergence in distribution. Elements of the space Ω, usually denoted by
ω, are almost always omitted as arguments of random variables, i.e. instead
of writing X(ω), we simply write X. Finally, the acronym i.i.d. stands for
“independent and identically distributed”.
Functions: Univariate as well as d-dimensional distribution functions
are denoted by capital letters, mostly F or G. Their corresponding survival
functions are denoted F̄ , Ḡ. As an exception, a copula is denoted by the
letter C; its arguments are denoted (u1 , . . . , ud ) ∈ [0, 1]d. The characteristic
1 The letter U indicates that U1 , . . . , Ud are uniformly distributed on the unit interval.

1
May 17, 2017 16:28 Simulating Copulas - 9in x 6in b2921-Main page 2

2 Simulating Copulas: Stochastic Models, Sampling Algorithms, and Applications

function of a random variable X is denoted by φX (x) := E[exp(i x X)].

The Laplace transform of a non-negative random variable X is denoted
by ϕX (x) := E[exp(−x X)]. Moreover, the nth derivative of a real-valued
function f is abbreviated as f (n) ; for the first derivative we also write f
.
The natural logarithm is denoted log.
Stochastic processes: A stochastic process X : Ω × [0, ∞) → R on
a probability space (Ω, F , P) is denoted by X = {Xt }t≥0 , i.e. we omit the
argument ω ∈ Ω. The time argument t is written as a subindex, i.e. Xt
instead of X(t). This is in order to avoid confusion with deterministic
functions f , whose arguments are written in brackets, i.e. f (x).
Important univariate distributions: Some frequently used proba-
bility distributions are introduced here. Sampling univariate random vari-
ables is discussed in Chapter 6.
(1) U[a, b] denotes the uniform distribution on [a, b] for −∞ < a < b <
∞. Its density is given by f (x) = ½{x∈[a,b]} (b − a)−1 for x ∈ R.
(2) Exp(λ) denotes the exponential distribution with parameter λ > 0,
i.e. with density f (x) = λ exp(−λ x) ½{x>0} for x ∈ R.
(3) N (µ, σ 2 ) denotes the normal distribution with mean µ ∈ R and
variance σ 2 > 0. Its density is given by
2
1 1 (x−µ)
f (x) = √ e− 2 σ2 , x ∈ R.
2πσ 2
(4) LN (µ, σ 2 ) denotes the lognormal distribution. Its density is given
by
2
1 1 (log(x)−µ)
f (x) = √ e− 2 σ2 ½{x>0}, x ∈ R.
x 2πσ 2
(5) Γ(β, η) denotes the Gamma distribution with parameters β, η > 0,
i.e. with density
η β −η x β−1
f (x) = e x ½{x>0} , x ∈ R.
Γ(β)
Note in particular that the exponential law Exp(λ) = Γ(1, λ) is a
special case of the Gamma distribution.
(6) χ2 (ν) denotes the χ2 -distribution (Chi-square) with ν ∈ N degrees
of freedom. The density satisfies
1
f (x) = ν/2 xν/2−1 e−x/2 ½{x>0} , x ∈ R.
2 Γ(ν/2)
From the density we can easily see that χ2 (ν) = Γ(ν/2, 1/2) is a
second important special case of the Gamma distribution. If Z has
May 17, 2017 16:28 Simulating Copulas - 9in x 6in b2921-Main page 3

Introduction 3

a χ2 (ν)-distribution and X1 , . . . , Xν are i.i.d. standard normally

d
distributed random variables, then Z = X12 + . . . + Xν2 .
(7) IG(β, η) denotes the inverse Gaussian distribution with parameters
η, β > 0, i.e. with density
β 1 2 2
f (x) = 3 √ eη β− 2 (β /x+η x) ½{x>0} , x ∈ R.
x2 2 π
(8) S(α, h) denotes the (exponentially tilted) α-stable distribution with
0 < α < 1 and h ≥ 0, as discussed for instance in Barndorff-Nielsen
and Shephard (2001). This distribution is characterized via its
Laplace transform. If X has a S(α, h)-distribution, then
 
E e−xX = e−((x+h) −h ) , x > 0.
α α

√
We write S(α, 0) =: S(α). Note: S(1/2, h) = IG(2−1/2 , 2 h).
(9) Bin(n, p) denotes the binomial distribution with n trials and suc-
cess probability p ∈ (0, 1). The (discrete) density of X with
Bin(n, p)-distribution is
 
n k
P(X = k) = p (1 − p)n−k , k ∈ {0, . . . , n}.
k
(10) P oi(λ) denotes the Poisson distribution with mean λ > 0. The
(discrete) density of X with P oi(λ)-distribution is P(X = k) =
λk exp(−λ)/k! for k ∈ N0 .
(11) t(ν) denotes the Student’s t-distribution with ν ∈ N degrees of
freedom. The density2 is given by


Γ ν+1  x2 −( 2 )
ν+1

f (x) = √ 2 ν 1 + , x ∈ R.
πνΓ( 2 ) ν
(12) t(µ, ν) denotes the (non-central) Student’s t-distribution with non-
centrality parameter µ ∈ R and ν ∈ N degrees of freedom. This
distribution is composed of Z, a normally distributed random vari-
able with unit variance and zero mean, and V , a Chi-square dis-
tributed random variable with ν degrees of freedom (independent
of Z), via (Z + µ)/ V /ν.
(13) P areto(α, x0 ) denotes the Pareto distribution with parameters
α, x0 > 0, i.e. with survival function
 x α
F̄ (x) = 1 − F (x) =
0
½{x≥x0 } + ½{x<x0 } , x ∈ R.
x
2 The motivation for this distribution is the composition of Z, a normally distributed

random variable with unit variance and zero mean, and V , a Chi-square

distributed
random variable with ν degrees of freedom (independent of Z): Z/ V /ν has a t(ν)-
distribution.
May 17, 2017 16:28 Simulating Copulas - 9in x 6in b2921-Main page 4

4 Simulating Copulas: Stochastic Models, Sampling Algorithms, and Applications

(14) Geo(p) denotes the geometric distribution with success probability

p ∈ (0, 1]. The (discrete) density of X with Geo(p)-distribution is
P(X = k) = (1 − p)k−1 p for k ∈ N.
(15) Beta(a, b) denotes the Beta distribution with parameters a, b > 0.
Its density is given by
Γ(a + b) a−1
f (x) = ½{x∈(0,1)} x (1 − x)b−1 .
Γ(a) Γ(b)
Abbreviations for other distributions are introduced when they first appear.
The symbol ∼ means “distributed according to”, e.g. E ∼ Exp(1) means
that E is an exponential random variable with unit mean.

1.1 Copulas

The law of a d-dimensional random vector X := (X1 , . . . , Xd ), defined on a

probability space (Ω, F, P), is usually studied from its distribution function
F (x1 , . . . , xd ) := P(X1 ≤ x1 , . . . , Xd ≤ xd ), x1 , . . . , xd ∈ R.
For i = 1, . . . , d the distribution function Fi of Xi is called the (univariate)
marginal law or margin and can be retrieved from F via
Fi (xi ) := P(Xi ≤ xi ) = F (∞, . . . , ∞, xi , ∞, . . . , ∞), xi ∈ R.
Taking ∞ as an argument of F is used as a shorthand notation for taking the
limits as the arguments x1 , . . . , xi−1 , xi+1 , . . . , xd tend to infinity. It is im-
portant to mention that it is not enough to know the margins F1 , . . . , Fd in
order to determine F . Additionally it is required to know how the marginal
laws are coupled. This is achieved by means of a copula of (X1 , . . . , Xd ).
Generally speaking, knowing the margins and a copula is equivalent to
knowing the distribution. It is now appropriate to give the definition of a
copula.
Definition 1.1 (Copula)

(1) A function C : [0, 1]d → [0, 1] is called a (d-dimensional) copula,

if there is a probability space (Ω, F , P) supporting a random vector
(U1 , . . . , Ud ) such that Uk ∼ U[0, 1] for all k = 1, . . . , d and
C(u1 , . . . , ud ) = P(U1 ≤ u1 , . . . , Ud ≤ ud ), u1 , . . . , ud ∈ [0, 1].
(2) On a probability space (Ω, F, P) let (U1 , . . . , Ud ) be a random vec-
tor on [0, 1]d whose joint distribution function (restricted to [0, 1]d)
May 17, 2017 16:28 Simulating Copulas - 9in x 6in b2921-Main page 5

Introduction 5

is a copula C : [0, 1]d → [0, 1]. For i = 2, . . . , d and indices

1 ≤ j1 < . . . < ji ≤ d the notation Cj1 ,...,ji : [0, 1]i → [0, 1] is
introduced for the joint distribution function of the random subvec-
tor (Uj1 , . . . , Uji ). It is itself a copula and called an i-margin of
C.

For a random vector (U1 , . . . , Ud ) ∈ [0, 1]d on the d-dimensional unit cube
the values of its distribution function on Rd \ [0, 1]d are completely deter-
mined by its values on [0, 1]d . Thus, copulas are defined on [0, 1]d only. A
d-dimensional copula C induces a probability measure dC on the unit cube
[0, 1]d . More clearly, if a random vector (U1 , . . . , Ud ) on [0, 1]d is defined on
the probability space (Ω, F , P) and has distribution function C, then



dC(B) := P (U1 , . . . , Ud ) ∈ B , B ∈ B [0, 1]d .
The measure dC is called the probability measure associated with the copula
C. It is uniquely determined by C. The three simplest examples of copulas
are defined in the following examples.
Example 1.1 (Independence Copula)
The function Π : [0, 1]d → [0, 1], given by
d
Π(u1 , . . . , ud ) := ui , u1 , . . . , ud ∈ [0, 1],
i=1
is called the independence copula. To see that Π actually is a copula,
consider a probability space (Ω, F , P) supporting i.i.d. random variables
U1 , . . . , Ud with U1 ∼ U[0, 1]. The random vector (U1 , . . . , Ud ) then has
U[0, 1]-distributed margins and joint distribution function
d d
P(U1 ≤ u1 , . . . , Ud ≤ ud ) = P(Ui ≤ ui ) = ui
i=1 i=1

= Π(u1 , . . . , ud ), u1 , . . . , ud ∈ [0, 1].

The independence of the components of Π explains the nomenclature.

Example 1.2 (Comonotonicity Copula)

Considering a probability space (Ω, F , P) supporting a single random vari-
able U ∼ U[0, 1], the random vector (U1 , . . . , Ud ) := (U, . . . , U ) ∈ [0, 1]d has
U[0, 1]-distributed margins and joint distribution function
P(U1 ≤ u1 , . . . , Ud ≤ ud ) = P(U ≤ min{u1 , . . . , ud })
= min{u1 , . . . , ud }, u1 , . . . , ud ∈ [0, 1].
May 17, 2017 16:28 Simulating Copulas - 9in x 6in b2921-Main page 6

6 Simulating Copulas: Stochastic Models, Sampling Algorithms, and Applications

Consequently, the function M : [0, 1]d → [0, 1], defined by

M (u1 , . . . , ud ) := min{u1 , . . . , ud }, u1 , . . . , ud ∈ [0, 1],

is a copula called the copula of complete comonotonicity (also called the

upper Fréchet–Hoeffding bound, see Lemma 1.3).

Example 1.3 (Countermonotonicity Copula)

Considering a probability space (Ω, F , P) supporting a single random vari-
able U ∼ U[0, 1], the bivariate random vector (U1 , U2 ) := (U, 1 − U ) ∈ [0, 1]2
has perfectly negatively associated components (i.e. if U1 is large, then U2 is
small, and vice versa). This random vector has U[0, 1]-distributed margins
and joint distribution function

P(U1 ≤ u1 , U2 ≤ u2 ) = P(1 − u2 ≤ U ≤ u1 )
= (u1 + u2 − 1) ½{1−u2 ≤u1 } , u1 , u2 ∈ [0, 1].

Consequently, the function W : [0, 1]2 → [0, 1], defined by

W (u1 , u2 ) := (u1 + u2 − 1) ½{u1 +u2 ≥1} , u1 , u2 ∈ [0, 1],

is a bivariate copula called the copula of complete countermonotonicity

(also called the lower Fréchet–Hoeffding bound, see Lemma 1.3).

When dealing with copulas, especially in larger dimensions, many computa-

tions exhibit a combinatorial character. This is mainly due to the fact that
the dimension d ≥ 2 is a “discrete” number. The following basic fact from
probability calculus will sometimes be useful in this regard and is stated
here for later reference. It shows how to compute the probability of a union
of events in terms of probabilities of intersections.

Lemma 1.1 (Principle of Inclusion and Exclusion)

Let (Ω, F , P) be a probability space. With A1 , . . . , An ∈ F one has3
 n   
P Ai = (−1)|I|+1 P Aj
i=1 ∅=I⊂{1,...,n} j∈I
n  k 
= (−1)k+1 P Aij .
k=1 1≤i1 <...<ik ≤n j=1

Proof. See, e.g., Billingsley (1995, p. 24).

3 For n = 2: P(A1 ∪ A2 ) = P(A1 ) + P(A2 ) − P(A1 ∩ A2 ).
May 17, 2017 16:28 Simulating Copulas - 9in x 6in b2921-Main page 7

Introduction 7

1.1.1 Analytical Properties

Univariate distribution functions can be characterized by analytical prop-
erties. It is well known that a function F : R → [0, 1] is the distribution
function of some (real-valued) random variable X if and only if it is right-
continuous, non-decreasing, limx→−∞ F (x) = 0, and limx→∞ F (x) = 1
(see, e.g., Billingsley (1995, Theorem 12.4, p. 176)). There is an analo-
gous characterization of multivariate distribution functions via analytical
properties, using the notions of groundedness and d-increasingness.

Definition 1.2 (Groundedness and d-Increasingness)

Let C : [0, 1]d → [0, 1] be an arbitrary function.

(1) C is called grounded, if C(u1 , . . . , ud ) = 0 whenever ui = 0 for at

least one component i ∈ {1, . . . , d}.
(2) C is called d-increasing, if for all

u := (u1 , . . . , ud ), v := (v1 , . . . , vd ) ∈ [0, 1]d

with ui < vi for all i = 1, . . . , d one has4

(−1)|{i : wi =ui }| C(w1 , . . . , wd ) ≥ 0.

i=1 {ui ,vi }
(w1 ,...,wd )∈×d

Obviously, a copula C : [0, 1]d → [0, 1] is a grounded function by Defini-

tion 1.1. To see that a d-dimensional copula is also d-increasing, consider
a probability space (Ω, F , P) supporting a random vector (U1 , . . . , Ud ) ∼ C
and let u := (u1 , . . . , ud ), v := (v1 , . . . , vd ) ∈ [0, 1]d with ui < vi for
all i = 1, . . . , d. For each subset ∅ = I ⊂ {1, . . . , d}, define the vector
(I) (I)
w(I) := (w1 , . . . , wd ) by


(I) uk , k ∈ I
wk := , k = 1, . . . , d.
vk , k ∈
/I

4 For
 
d = 2: dC [u1 , u2 ] × [v1 , v2 ] = C(v1 , v2 ) − C(v1 , u2 ) − C(u1 , v2 ) + C(u1 , u2 ) ≥ 0.
 Another Random Document on
Scribd Without Any Related Topics
The Project Gutenberg eBook of The Strange
Adventures of Andrew Battell, of Leigh, in
Angola and the Adjoining Regions
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 Strange Adventures of Andrew Battell, of Leigh, in Angola

and the Adjoining Regions

Author: Andrew Battell

Contributor: active 1591-1649 Anthony Knivet

Editor: Samuel Purchas

Ernest Georg Ravenstein

Release date: November 3, 2012 [eBook #41282]

Most recently updated: October 23, 2024

Language: English

Credits: Produced by Chris Curnow, Turgut Dincer, Joseph Cooper

and
the Online Distributed Proofreading Team at
http://www.pgdp.net

*** START OF THE PROJECT GUTENBERG EBOOK THE STRANGE

ADVENTURES OF ANDREW BATTELL, OF LEIGH, IN ANGOLA AND
THE ADJOINING REGIONS ***
 WORKS ISSUED BY

The Hakluyt Society

THE STRANGE ADVENTURES

OF

ANDREW BATTELL.
SECOND SERIES.
No. VI.

THE

STRANGE ADVENTURES
OF

ANDREW BATTELL
OF LEIGH,

IN ANGOLA AND THE ADJOINING REGIONS.

REPRINTED FROM “PURCHAS HIS PILGRIMES.”

Edited, with Notes and a Concise

HISTORY OF KONGO AND ANGOLA,
 BY

E. G. RAVENSTEIN.
Reproduced, by permission of the
HAKLUYT SOCIETY
from the edition originally published by the Society
in 1901
KRAUS REPRINT LIMITED
Nendeln/Liechtenstein
1967

Printed in Germany
Lessing-Druckerei—Wiesbaden

COUNCIL
OF

THE HAKLUYT SOCIETY.

Sir Clements Markham, K.C.B., F.R.S., Pres. R.G.S., President.

The Right Hon. The Lord Stanley of Alderley, Vice-President.
Rear-Admiral Sir William Wharton, K.C.B., Vice-President.
Commr. B. M. Chambers, R.N.
C. Raymond Beazley, M.A.
Colonel G. Earl Church.
Sir W. Martin Conway.
F. H. H. Guillemard, M.A., M.D.
Edward Heawood, M.A.
Dudley F. A. Hervey, C.M.G.
E. F. Im Thurn, C.B., C.M.G.
J. Scott Keltie, LL.D.
F. W. Lucas.
A. P. Maudslay.
E. J. Payne, M.A.
Howard Saunders.
H. W. Trinder.
Charles Welch, F.S.A.

William Foster, B.A., Honorary Secretary.

CONTENTS.

PAGE
Introduction ix
Bibliography xviii

The Strange Adventures of Andrew Battell of Leigh.

I. Andrew Battell, his voyage to the River of Plate, who
being taken on to the coast of Brazill was sent to
Angola 1
II. His trading on the coast; offer to escape; imprisonment;
exile; escape and new imprisonment; his sending to
Elamba and Bahia das Vaccas; many strange
occurrences 9
III. Discovery of the Gagas: their wars, man-eating; over-
running countries. His trade with them, betraying,
escape to them, and living with them; with many
strange adventures. And also the rites and manner
of life observed by the Iagges, or Gagas, which no
Christian would ever know well but this author 19
IV. His return to the Portugals: invasions of diverse
countries; abuses; flight from them, and living in the
woods divers months; his strange boat and coming
to Loango 36
 V. Of the province of Engoy, and other regions of Loango;
with the customs there observed by the King and
people 42
VI. Of the provinces of Bongo, Calongo, Mayombe,
Manikesocke, Motimbas; of the ape-monster Pongo;
their Hunting, Idolatries, and divers other
observations 52
VII. Of the Zebra and Hippopotamus; the Portuguese Wars in
those parts; the Fishing, Grain, and other things
remarkable 63
On the Religion and the Customs of the Peoples of Angola, Congo,
and Loango, from Purchas His Pilgrimage, 1613 (1617) 71

APPENDICES.
I. Anthony Knivet in Kongo and Angola 89
II. A Sketch of the History of Kongo to the end of the
Seventeenth Century 102
III. A List of the Kings of Kongo 136
IV. A Sketch of the History of Angola to the end of the
Seventeenth Century 139
V. A List of the Governors of Angola 188

Index and Glossary 192

MAPS.
A General Map of Kongo and Angola.
An Enlarged Map of Angola.
 INTRODUCTION.
OUR Englishmen are known to have visited Angola
towards the close of the sixteenth century, namely,
Thomas Turner, Andrew Towres, Anthony Knivet and
Andrew Battell. All four were taken by the Portuguese
out of English privateers in South-American waters, and
spent years of captivity as prisoners of war; happy, no
doubt, in having escaped the fate of many of their less fortunate
companions, who atoned with their lives for the hazardous
proceedings in which they had engaged.

Thomas Turner,1 although he furnished Samuel Purchas with a few

notes on Brazil, never placed on record what happened to him whilst
in Portuguese Africa. Towres was sent to prison at Rio de Janeiro for
the heinous offence of eating meat on a Friday; he attempted an
escape, was retaken, and condemned to spend the rest of his
captivity in Angola. He died at Masanganu, as we learn from Knivet.
Knivet himself has left us an account of his adventures in Angola and
Kongo; but this account contains so many incredible statements that
it was with some hesitation we admitted it into this volume, as by
doing so we might be supposed to vouch for the writer’s veracity.
Andrew Battell, fortunately, has left behind him a fairly circumstantial
record of what he experienced in Kongo and Angola. His narrative
bears the stamp of truth, and has stood the test of time. It is
unique, moreover, as being the earliest record of travels in the
interior of this part of Africa; for, apart from a few letters of Jesuit
missionaries, the references to Kongo or Angola printed up to
Battell’s time, were either confined to the coast, or they were purely
historical or descriptive. Neither F. Pigafetta’s famous Relatione del
Reame di Congo, “drawn out of the writings and discourses of
Duarte Lopez,” and first published at Rome in 1591, nor the almost
equally famous Itinerarium of Jan Huyghen van Linschoten, of which
an English translation appeared as early as 1598, can be classed
among books of travel.2 Samuel Braun, of Basel, who served as
barber-surgeon on board Dutch vessels which traded at Luangu and
on the Kongo, 1611-13, never left the coast.3 Nor did Pieter van der
Broeck, who made three voyages to the Kongo between 1607 and
1612 as supercargo of Dutch vessels, penetrate inland.4 Nay, we are
even able to claim on behalf of Battell that he travelled by routes not
since trodden by European explorers.

Of Andrew Battell’s history we know nothing, except what may be

gathered from his “Adventures,” and an occasional reference to him
by his friend, neighbour, and editor, the Rev. Samuel Purchas. He
seems to have been a native of Leigh, in Essex, at the present day a
mere fishing village by the side of its populous upstart neighbour
Southend, but formerly a place of considerable importance. As early
as the fifteenth century it could boast of its guild of pilots, working in
harmony with a similar guild at Deptford Strond, the men of Leigh
taking charge of inward bound ships, whilst Deptford provided pilots
to the outward bound. Henry VIII incorporated both guilds as the
“Fraternity of the Most Glorious and Indivisible Trinity and of St.
Clement;” and in the venerable church of St. Clement, at Leigh, and
the surrounding churchyard may still be seen monuments erected in
honour of contemporaries of Battell who were Brethren of the Trinity
House; among whom are Robert Salmon (born 1567, died 1661) and
Robert Chester (died 1632). But there is no tombstone in memory of
Andrew Battell; and if a memorial tablet was ever dedicated to him,
it must have been removed when the church was renovated in 1837.
Nor do the registers of the church afford a clue to Battell’s death, for
the earliest of these documents only dates back to the year 1684. At
the present time no person of the name of Battell lives at Leigh.

Samuel Purchas was Vicar of Eastwood, a small village two miles to

the north of Leigh, from 1604 to 1613. Battell returned to Leigh
about 1610, bringing with him a little negro boy, who claimed to
have been kept a captive by a gorilla (see p. 55). Purchas had many
conferences with Battell, and the information obtained in this
manner was incorporated by him in Purchas His Pilgrimage, the first
edition of which was published in 1613,5 and will be found in this
volume, pp. 71-87. Battell’s papers, however, only reached Purchas
after the author’s death, and were first published by him in
Hakluytus Posthumus, or Purchas His Pilgrimes, in 1625.6 There is
reason to fear that Purchas did not perform his duties as editor, as
such duties are understood at the present day. As an instance, we
notice that Battell distinctly told his editor in private conference (see
p. 83) that in his day nothing was known about the origin of the
Jagas, expressly denying that Duarte Lopez could have any
information about it; yet, elsewhere (p. 19), Battell is made
responsible for the statement that they came from Sierra Leone. Nor
is it likely that Battell ever mentioned a lake Aquelunda (p. 74), for
no such lake exists; and Purchas’s authority for its supposed
existence is once more Duarte Lopez or Pigafetta.
Moreover, there is some ground for supposing that Purchas abridged
portions of the MS.; as, for instance, the account of the overland
trading trip to Kongo and Mbata. Perhaps he likewise rearranged
parts of his MS., thus confusing the sequence of events, as will be
seen when we come to inquire into the chronology of Battell’s
travels.

There exists no doubt as to the object with which Abraham Cocke

sailed for the Plate River in 1589. Philip of Spain had acceded to the
throne of Portugal in 1580, and that prosperous little kingdom thus
became involved in the disaster which overtook the Armada, which
sailed out of Lisbon in May, 1588. English skippers therefore felt
justified in preying upon Portuguese trade in Brazil, and intercepting
Spanish vessels on their way home from the Rio de la Plata. We do
not think, however, that we do Abraham Cocke an injustice when we
assume him to have been influenced in his hazardous enterprise
quite as much by the lust of gain as by patriotism.
The determination of the chronology of Battell’s adventures presents
some difficulty, as his narrative contains but a single date, namely,
that of his departure from England on May 7th, 1589. There are,
however, incidental references to events the dates of which are
known; and these enable us to trace his movements with a fair
amount of confidence, thus:—
1. Having left Plymouth in May, 1589, we suppose Battell to have
reached Luandu in June, 1590.
2. His journey up to Masanganu, his detention there for two months,
and return to Luandu, where he “lay eight months in a poor estate”
(p. 7), would carry us to the end of June, 1591.
3. Battell tells us that the Governor, D. João Furtado de Mendonça,
then employed him during two years and a half trading along the
coast. This, however, is quite impossible: for Mendonça only
assumed office in August, 1594; but, as he is the only Governor of
Battell’s day who held office for a longer period than two and a half
years—his term of office extending to 1602—and as Battell is not
likely to have forgotten the name of an employer who gave him his
confidence, we assume that he really did make these trading trips,
but at a subsequent period. Purchas may be responsible for this
transposition.
4. He made a first attempt to escape (in a Dutch vessel), but was
recaptured, and sent to Masanganu, where he spent “six miserable
years,” 1591-96.
5. Second attempt to escape, and detention for three months in
irons at Luandu, up to June, 1596.
6. Campaign in Lamba and Ngazi (see p. 13, note). After a field
service of over three years, Battell was sent back to Luandu,
wounded. This would account for his time up to 1598 or 1599.
7. I am inclined to believe that, owing to the confidence inspired by
his conduct in the field, the Governor now employed him on the
trading ships referred to above.
9. Trading trips to Benguella in 1600 or 1601.
10. Battell joins the Jagas, and spends twenty-one months with
them. Incidentally he mentions that the chief, Kafuche, had been
defeated by the Portuguese seven years before that time (he was
actually defeated in April, 1594).
11. Battell was at Masanganu when João Rodrigues Coutinho was
Governor (Coutinho assumed office in 1602).
12. Battell was present at the building of the presidio of Kambambe
by Manuel Cerveira Pereira in 1604; and stayed there till 1606, when
news was received of the death of Queen Elizabeth, and he was
promised his liberty. The Queen died March 24th, 1603.
13. A journey to Mbamba, Kongo, etc., may have taken up six
months.
14. The Governor having “denied his word,” and a new Governor
being daily expected, Battell secretly left the city, spent six months
on the Dande, and was ultimately landed at Luangu. (The new
Governor expected was only appointed in August, 1607; and his
arrival was actually delayed.)
15. In Luangu, Battell spent two years and a half—say up to 1610.
Great pains have been taken by me with the maps illustrating this
volume; and, if the outcome of my endeavour does not differ in its
broad features from the maps furnished by M. d’Anville, in 1732, to
Labat’s Relation Historique de l’Éthiopie Occidentale, this should
redound to the credit of the great French geographer, but should not
be accounted a proof of lack of industry on my own part. Still, my
maps exhibit an advance in matters of detail, for our knowledge of
the country has increased considerably since the days of d’Anville.
They would have proved still more satisfactory had the Portuguese
thought it worth while to produce a trustworthy map of a colony of
which they had claimed possession during four centuries. It seems
almost incredible that even now many of the routes followed by the
Conquistadores and missionaries of old cannot be laid down upon a
modern map for lack of information. Sonyo, for instance, through
which led the high road followed by soldiers, traders, and
missionaries going up to San Salvador (the present route leaves the
Kongo River at Matadi), is almost a terra incognita. I am almost
ashamed to confess that I have even failed to locate the once-
famous factory of Mpinda; all I can say is, that it cannot have
occupied the site assigned to it on some Portuguese maps.
I need hardly say that modern research lends no support to the
extravagant claims of certain geographers as to the knowledge of
Inner Africa possessed by the Portuguese in the sixteenth century.
Pigafetta’s fantastic map, with its elaborate system of lakes and
rivers, merely proves the utter incapacity of its author to deal with
questions of critical geography. This has long since been recognised.
The map which accompanies Isaac Vossius’s De Nili et aliorum
Fluminum Origine (Hagae Com., 1659) only shows one lake in Inner
Africa, which borders on “Nimeamaie vel Monemugi,” and may
without hesitation be identified with our Nyasa: for the Monemugi
(Muene Muji) is the chief of the Maravi or Zimbas. The “Iages, gens
barbara et inculta,” are placed right in the centre of Africa. The
“Fungeni,” which are shown as neighbours of the “Macoco,” ought to
have been placed to the west of Abyssinia, as they are the Funj, or
Fung, of the Egyptian Sudan. If Ludolfus had carried out his
intention of compiling a map of the whole of Africa (in 1681), these
extravagancies of early map-makers would have been exposed more
fully long since.7
In collecting materials for the maps and for the notes illustrating
Battell’s narrative, I felt bound to consult all accessible literary
sources dealing with the history and geography of Kongo and
Angola. Whilst ploughing my way through this mass of material, it
struck me that a concise history of these African countries, from the
time of their discovery to the end of the seventeenth century, might
form an acceptable appendix to Battell’s Adventures, and at the
same time increase the bulk of the volume dedicated to him to more
respectable proportions. Much material of use for such a purpose
has seen the light since the publication of J. J. Lopes de Lima’s
historical sketches. Yet I am bound to confess that the result of all
this tedious labour is disappointing. I may have been able to rectify
a few dates and facts; but much remains to be done before we can
claim to be in possession of a trustworthy history of that part of
Africa. Possibly my little sketch may rouse a Portuguese into taking
up the work of the late Luciano Cordeiro. Many documents not yet
published should be discoverable in the archives of Portugal, Spain,
and Luandu.8
The spelling of the proper names mentioned by Battell is retained, as
a matter of course; but it is obvious that in the historical appendices
the various ways in which native names are spelt had to be reduced
to a common system. Much might be said in favour of accepting the
Portuguese manner of spelling, but after due consideration I decided
to adopt the system now generally followed (even by a few
Portuguese writers), viz., that all vowels should be sounded as in
Italian, and the consonants as in English, with the only exception
that the letter g should always be hard. I therefore write Sonyo,
instead of Sonho, Sogno, or Sonjo, as the name of that district is
spelt according to the nationality of the writer. In transcribing the
native names I have had the unstinted assistance, among others, of
the Rev. Thomas Lewis, of the Baptist Missionary Society; yet I am
fully aware that the spelling adopted for many names is at least
doubtful, if not absolutely incorrect. This arises quite as much from a
defective hearing on the part of my authorities, as from the
illegibility of many early manuscripts or the carelessness of copyists.
All such doubtful cases are dealt with in the Glossary and Index.

In conclusion, I feel bound to acknowledge with gratitude the kindly

assistance rendered me by Mr. R. E. Dennett, who is spending a life-
time in Luangu; Mr. R. C. Phillips, who is thoroughly acquainted with
the Lower Kongo; the Rev. Thomas Lewis, of the Baptist Missionary
Society; Captain Binger, of the French Foreign Office; and last, not
least, our ever-obliging Secretary, Mr. William Foster.

BIBLIOGRAPHY.
Only the titles of a few books cited merely by the author’s
name, or by abbreviated references, are included in this
list.
How cited:

Alguns Doc.—Alguns documentos do archivo nacional da Torre do Tombo ácerca

das navegações e conquistas Portuguezas. Lisboa (Impr. nac.), 1892.
A Collection of documents, 1416-1554, edited by José Ramos-
Coelho. See Index sub Angola, Kongo, Manicongo.

Paiva Manso.—Historia do Congo, obra posthuma do (Dr. Levy) Visconde de Paiva

Manso. Lisboa (Typ. da Acad.), 1877.
A collection of documents, 1492-1722.

Boletim.—Boletim da Sociedade de Geographia de Lisboa.

The volume for 1883 contains documents now in the Bibliothéque
Nationale (instructions given to B. Dias, 1559; Letters of F.
Garcia Simões, F. Balthasar Barretta, and other Jesuits).
 Memorias do Ultramar, Viagens explorações e conquistas do
Portuguezes. Collecção de Documentos por Luciano Cordeira.
Lisboa (Impr. nac.) 1881.

The following Parts have been published:—

Garcia Mendes.
(a) 1574-1620. Da Mina ao Cabo Negro segundo Garcia Mendes
Castello Branco (the writer of these reports was one of the
companions of Paulo Dias de Novaes).

Rebello de Aragão.
(b) 1593-1631. Terras e Minas Africanas segundo Balthazar Rebello
de Aragão. (He went out to Africa in 1593).

Benguella e seu Sertão.

(c) 1617-1622. Benguella e seu sertão per um Anonymo. (The
author of this account of the conquest of Benguella may
possibly have been Manuel Cerveira Pereira).

Estabelecimentos.
(d) 1607. Estabelecimentos e Resgates Portuguezes na costa
occidental de Africa por um Anonymo.

Escravos e Mimas.
(e) 1516-1619. Escravos e Minas de Africa segundo Diversos.

D. Lopez.—Relatione del Reame di Congo e delle circonvicine contrade tratta dalli

Scritti e ragionamente di Odoardo Lopez, per Filippo Pigafetta. Roma,
1591.
This work has been translated into Latin, German, Dutch, French
and English, but has not hitherto found a competent editor. I
quote the English translation by Mrs. M. Hutchinson, published
at London in 1881.
Duarte Lopez went out to Kongo in 1578; and the bulk of this
volume is based upon information imparted to his editor when
he was in Rome in 1591. Pigafetta has most unwisely
expanded the information thus obtained into a description of
the greater part of Africa.
Cavazzi.—Istorica descrizione de’ tre regni Congo, Matamba, e Angola,
accuratamente compilata, dal P. Gio. Antonio Cavazzi da Montecuccolo.
Bologna, 1687.
Cavazzi, a Capuchin, visited Kongo and Angola twice (1654-67,
1670-??) and died at Genoa in 1693. This bulky folio only
deals with his first visit, and was edited by P. Fortunato
Alamandini, of Bologna. Labat (“Relation historique de
l’Éthiopie,” Paris, 1732) has given a useful version of it in
French, which must, however, be used with some caution. It is
by far the most important work we have at the hand of one of
the early Catholic missionaries. W. D. Cooley’s observation
(“Inner Africa Laid Open,” London, 1852, p. 3), that the works
published up to the time of Cavazzi “would hardly furnish
twenty pages of sound geographical intelligence,” can apply
only to what they say of Inner Africa; whilst Lopez de Lima
(“Ensaios,” p. xi) is hardly justified in calling Cavazzi a
“fabulista,” unless that opprobrious term be confined to what
the friar relates of the miracles wrought by himself and
others.

Dapper.—Nauwkeurige beschrijving der Afrikaansche gewesten van Olf. Dapper.

Amst., 1668.
I quote the German translation (“Beschreibung von Afrika,” Amst.,
1670).
This is a very careful compilation; more especially interesting, as it
contains information on the country collected during the
Dutch occupation (1642-48), not to be found elsewhere.

Cadornega.—Historia das guerras de Angola (Historia General Angolana), por D. A.

de Oliveira Cadornega, in 1680-82.
Cadornega, a native of Villa Viçosa, accompanied D. Pedro Cezar
de Menezes to Angola in 1639, and died at Luandu in 1690.
His work (in three volumes) only exists in MS. in the library of
the Academy of Sciences, Lisbon, and in the Bibliothéque
Nationale, Paris. I have not been able to consult it with the
minuteness which it deserves. A rough copy of a considerable
portion of it is to be found in the British Museum (Add. MS.
15,183, fol. 33). Copious extracts from it are given by Paiva
Manso and D. José de Lacerda (“Exame das Viagens do Dr.
Livingstone,” Lisbon, 1867).
Catalogo.—Catalogo dos Governadores do Reine de Angola (Collecçao de Noticias
para a historia das nações ultramarinas publicada pela Academia real
das Sciencias, tome III, pt. 2). Lisboa, 1826.
This is an anonymous compilation, continued to the year 1784. J.
C. Feo Cardozo, in his “Memorias contendo a biographia do
Vico-Almirante Luiz da Motta Feo e Torres,” Paris, 1825, also
printed this chronological history, and continued it to the year
1825. He has added the map drawn in 1790 by Colonel L. C.
C. Pinheiro Furtado. The “Catalogue” is useful, but it is not
free from very serious errors.

Bentley.—Dictionary and Grammar of the Kongo Language, by the Rev. W. Holman

Bentley. 1887.

Cordeiro da Matta.—Ensaio de Diccionario Kimbundu-Portugueze coordenado par L.

D. Cordeiro da Matta. Lisboa, 1893.

Lopes de Lima, ensaio.—Ensaios sobre a Statistica das possessões Portuguezes (III.

Ensaio sobre a Statistice d’Angola e Benguella), por José Joaquim
Lopes de Lima (Imp. nac.), 1846.
This is a fundamental work. The historical account is contained in
the Introduction and in chap. v.

Lopes de Lima, an. mar.—Descobrimento, posse, e conquista do reino do Congo

pelos Portuguezes no Seculo xvi, por J. J. Lopes de Lima (“Annaes
maritimos e coloniaes,” Lisboa, 1845, pp. 93-108).

Lopes de Lima.—Successos do Reino do Congo, no seculo xvii, pelo J. J. Lopes de

Lima (ibid., pp. 194-99).
 THE STRANGE ADVENTURES
OF

ANDREW BATTELL OF LEIGH IN ESSEX,

SENT BY THE PORTUGALS PRISONER TO ANGOLA,

WHO LIVED THERE, AND IN THE ADJOINING

REGIONS, NEAR EIGHTEEN YEARS.

§ I.

Andrew Battel, his Voyage to the River of Plate, who being taken on
the coast of Brasill, was sent to Angola.

[From the Thames to Cape Palmas.]

N the year 1589, Abraham Cocke9 of Limehouse, began

his voyage toward the River of Plate, with two
pinnaces10 of fifty tons apiece: the one was called the
May-Morning, the other the Dolphin.
We sailed from the river Thames the twentieth of April; and the six
and twentieth of the same month we put into Plimmoth [Plymouth],
where we took in some provision for the voyage. The seventh of May
we put to sea, and with foul weather were beaten back again into
Plimmoth, where we remained certain days, and then proceded on
our voyage: And running along the coast of Spain and Barbary we
put into the road of Sancta Cruz,11 and there set our Light-horse-
man12 together which we carried in two pieces. Abraham Cocke
made great account hereof, thinking that this boat should have
made his voyage. This done, we put to sea, and running along the
coast of Guinea we were becalmed, because we were so near the
coast.

[St. Thomé and the Gulf of Guinea.]

Here our men fell sick of the scurvy, in such sort, that there were
very few sound. And being within three or four degrees of the
equinoctial line we fell with the Cape de las Palmas, where we had
some refreshing, wherewith our men recovered. The people of the
Cape de las Palmas [Cabo das Palmas] made much of us, saying that
they would trade with us; but it was but to betray us, for they are
very treacherous, and were like to have taken our boat, and hurt
some of our men. From this Cape we lay south-west off;13 but the
current and the calms deceived us, so that we were driven down to
the isle of St. Thomé,14 thinking that we had been further off to the
Sea than we were. And being in distress for wood and water, we
went in on the south end between San Tome and the islands das
Rolas,15 where we rode very smooth, and with our light-horse-man
went on shore, thinking to have watered, but we found none in the
island. Here we had great store of plantains and oranges. We found
a village of negroes, which are sent from San Tome, for the
Portugals of San Tome do use, when their slaves be sick or weak, to
send them thither to get their strength again. For the islands are
very fruitful, and though there be no fresh water, yet they maintain
themselves with the wine of the palm-trees. Having refreshed
ourselves with the fruit of this island, we burned the village. And
running on the east side of San Tome we came before the town;16
but we durst not come near, for the castle shot at us, which hath
very good ordnance in it.
Then we lay east and by south toward the main, and in four and
twenty hours we had sight of the Cape de Lopo Gonsalves:17 and
being within three leagues of the said cape we cast about and stood
again toward the island of San Tome, and turned up on the west
side of the island; and coming to a little river, which runneth out of
the mountains, we went on shore with our Light-horse-man, with six
or seven butts to fill with water. But the governor had ambushed one
hundred men of the island; and when we were on shore they came
upon us, and killed one of our men and hurt another: wherefore we
retired to our boat and got aboard.

[Across the Atlantic to the Brazils.]

Then Abraham Cocke determined to fetch the coast of Brasil, and lay
west-south-west into the sea: and being some fifty leagues off, we
fell into a shoal of dolphins,18 which did greatly relieve us, for they
did follow our ship all the way, till we fell [in] with the land, which
was some thirty days. And running along the coast of Brasil till we
came to Ilha Grande,19 which standeth in five [sic] degrees
southward of the line, we put in betwixt the island and the main,
and haled our ships on shore, and washed them, and refreshed
ourselves, and took in fresh water. In this island are no inhabitants,
but it is very fruitful. And being here some twelve days there came
in a little pinnace which was bound to the River of Plate, which came
in to water and to get some refreshments: and presently we went
aboard, and took the Portugal merchant out of the pinnace, which
told Abraham Cocke, that within two months there should two
pinnaces come from the River of Plate, from the town of Buenos
Aires.

[The Rio de la Plata.]

From this town there come every year four or five caravels to
Bahia20 in Brasil, and to Angola in Africa, which bring great store of
treasure, which is transported overland out of Peru into the River of
Plate. There Abraham Cocke, desirous to make his voyage, took
some of the Dolphin’s men into his ship, and sent the Dolphin home
again, which had not as yet made any voyage. This Portugal
merchant carried us to a place in this island, where there was a
banished man,21 which had planted great store of plantains, and told
us that we might, with this fruit, go to the River of Plate: for our
bread and our victuals were almost all spent.
With this hard allowance we departed from this island, and were six-
and-thirty days before we came to the Isle of Lobos Marinos,22
which is in the mouth of the River of Plate. This island is half a mile
long, and hath no fresh water, but doth abound with seals and sea-
morses,23 in such sort that our light-horseman could not get on
shore for them, without we did beat them with our oars: and the
island is covered with them. Upon these seals we lived some thirty
days, lying up and down in the river, and were in great distress of
victuals. Then we determined to run up to Buenos Aires, and with
our light-horseman to take one of the pinnaces that rid at the town.
And, being so high up the river as the town, we had a mighty storm
at south-west,24 which drove us back again, and we were fain to ride
under the Isla Verde25—that is, the green island—which is in the
mouth of the river on the north side.

[A Prisoner of the Portuguese.]

Here we were all discomforted for lack of victuals and gave over the
voyage, and came to the northward again, to the isle of Sant
Sebastian, lying just under the tropic of Capricorn.26 There we went
on shore to catch fish, and some went up into the woods to gather
fruit, for we were all in a manner famished. There was at that time a
canoe fraught with Indians, that came from the town of Spiritu
Sancto.27 These Indians landed on the west side of the island, and
came through the woods and took five of us, and carried us to the
River of Janeiro [Rio de Janeiro]. After this mischance our captain,
Abraham Cocke, went to sea, and was never heard of more.28

[Transported to Angola—A Voyage to the Zaire.]

When we that were taken had remained four months in the River of
Janeiro, I and one Torner29 were sent to Angola in Africa, to the city
of Saint Paul,30 which standeth in nine degrees to the southward of
the equinoctial line. Here I was presently taken out of the ship and
put into prison, and sent up the River Quansa,31 to a town of
garrison, which is 130 miles up the river. And being there two
months the pilot of the governor’s pinnace died: then I was
commanded to carry her down to the city, where I presently fell sick,
and lay eight months in a poor estate, for they hated me because I
was an Englishman. But being recovered of my sickness, Don John
Hurtado de Mendoça,32 who then was governor, commanded me to
go to the river of Congo, called Zaire, in a pinnace, to trade for
elephants’ teeth,33 wheat,34 and oil of the palm-tree. The river
Zaire35 is fifty leagues from the city, to the northward, and is the
greatest river in all that coast. In the mouth of that river is an island,
called the Isle de Calabes, which had at that time a town in it. Here
we laded our pinnace with elephants’ teeth, wheat, and oil of the
palm, and so returned to the city again.

§ II.
His trading on the coast; offer to escape; imprisonment; exile;
escape and new imprisonment; his sending to Elamba and Bahia
das Vaccas; many strange occurrences.

[Trading in Loango.]
When I was sent to Longo [Loango], which is fifteen leagues to the
northward of the River Zaire, and carried all commodities fit for that
country, as long glass beads, and round blue beads, and seed beads,
and looking-glasses, blue and red coarse cloth, and Irish rugs, which
were very rich commodities. Here we sold our cloth at a great rate,
for we had for one yard of cloth three elephants’ teeth, that weighed
120 pounds; and we bought great store of palm-cloth36 and
elephants’ tails.37 So, in little time we laded our pinnace. For this
voyage I was very welcome to the governor, who promised me my
liberty if I would serve him. So I went in his pinnace two years and a
half upon the coast.

[An Attempted Escape.]

Then there came a ship of Holland to the city, the merchant of which
ship promised to carry me away. And, when they were ready to
depart I went secretly on board, but I was betrayed by Portugals
which sailed in the ship, and was fetched on shore by sergeants of
the city and put in prison, and lay with great bolts of iron two
months, thinking that the governor would have put me to death. But
at last I was banished for ever to the Fort of Massangano, to serve
in the conquest of those parts. Here I lived a most miserable life for
the space of six years without any hope to see the sea again.

[A Second Attempt at Escape.]

In this fort there were Egyptians and Moriscoes that were banished
as myself. To one of these Egyptians38 I brake my mind, and told
him that it were better for us to venture our lives for our liberty than
to live in that miserable place. This Egyptian was as willing as
myself, and told me he would procure ten of his consorts to go with
us. So we got three Egyptians and seven Portugals. That night we
got the best canoe that we could find, and went down the river
Cuanza, and being as far down as Mani Cabech,39 which is a little
lord in the province of Elamba [Lamba], we went on shore with our
twelve muskets, powder and shot. Here we sunk our canoe, because
they should not know where we went on shore. We made a little fire
in the wood, and scorched Guinea wheat,40 which we [had] brought
from Massangano, to relieve us, for we had none other food.
As soon as it was night, we took our journey all that night and the
next day, without any water at all. The second night we were not
able to go, and were fain to dig and scrape up roots of trees, and
suck them to maintain life. The third day we met with an old negro
which was travelling to Mani Cabech. We bound his hands behind
him, and made him lead us the way to the Lake of Casansa.41 And,
travelling all that day in this extreme hot country we came to the
Bansa [mbanza], or town, of Mani Casansa, which lyeth within the
land twelve leagues from the city of San Paulo. Here we were forced
to ask water, but they would give us none. Then we determined to
make them flee their houses with our shot; but seeing that we were
desperately bent they called their Lord, Mani Casansa, who gave us
water and fair speeches, desiring us to stay all night, only to betray
us; but we departed presently, and rested that night in (sic) the lake
of Casansa.
The fourth day, at night, we came to the river which is towards the
north,42 and passed it with great danger. For there are such
abundance of crocodiles in this river that no man dare come near
the riverside when it is deep. The fifth day, at night, we came to the
river Dande, and travelled so far to the eastward that we were right
against the Serras, or mountains of Manibangono,43 which is a lord
that warreth against the King of Congo, whither we intended to go.
Here we passed the river, and rested half the night. And being two
leagues from the river we met with negroes, which asked us whither
we travelled. We told them that we were going to Congo. These
negroes said that we were in the wrong way, and that they were
Masicongos,44 and would carry us to Bambe,45 where the Duke of
Bambe lay.
So we went some three miles east, up into the land, till we perceived
that we were in the wrong way, for we travelled by the sun, and
would go no further that way, and turned back again to the
westward; they stood before us with their bows, arrows and darts,
ready to shoot at us. But we, determining to go through them,
discharged six muskets together and killed four, which did amaze
them, and made them to retire. But they followed us four or five
miles, and hurt two of our company with their arrows. The next day
we came within the borders of Bamba, and travelled all that day. At
night we heard the surge of the sea. The seventh day, in the
morning, we saw the captain of the city come after us with
horsemen and great store of negroes. Hereupon our company being
dismayed, seven of our faint-hearted Portugals hid themselves in the
thickets. I, and the four Egyptians, thought to have escaped, but
they followed us so fast that we were fain to go into a little wood. As
soon as the captain had overtaken us he discharged a volley of shot
into the wood, which made us lose one another.

[Surrenders to a Portuguese Captain.]

Thus, being all alone, I bethought myself that if the negroes did take
me in the woods they would kill me: wherefore, thinking to make a
better end among the Portugals and Mulatoes, I came presently out
of the wood with my musket ready charged, making none account of
my life. But the captain, thinking that we had been all twelve
together, called to me and said: “Fellow Soldier, I have the
governor’s pardon; if you will yield yourselves you shall have no
hurt.” I, having my musket ready, answered the captain that I was
an Englishman, and had served six years at Massangono, in great
misery; and came in company with eleven Portugals and Egyptians,
and here am left all alone; and rather than I will be hanged, I will
die amongst you. Then the captain came near unto me and said:
“Deliver thy musket to one of the soldiers; and I protest, as I am a
gentleman and a soldier, to save thy life for thy resolute mind.”
Whereupon I yielded up my musket and myself.
Then the captain commanded all the soldiers and negroes to search
the woods, and to bring them out alive or dead, which was presently
done. Then they carried us to the city of San Paulo, where I and the
three Egyptians lay in prison three months with collars of iron, and
great bolts upon our legs, and hardly escaped.

[A campaign in Lamba.]46
At that time the governor sent four hundred men, that were
banished out of Portugal, up into the country of Elambe. Then I was
with proclamation through the city banished for ever to the wars,
and marched with them to Sowonso,47 which is a lord that obeyeth
the Duke of Bamba; from thence to Samanibansa, and then to
Namba Calamba, which is a great lord, who did resist us. But we
burnt his town, and then he obeyed us, and brought three thousand
warlike negroes to us. From thence [we marched] to Sollancango, a
little lord, that fought very desperately with us, but was forced to
obey; and then to Combrecaianga,48 where we remained two years.
From this place we gave many assaults and brought many lords to
subjection. We were fifteen thousand strong, and marched to the
Outeiro,49 or mountain, of Ingombe. But first we burnt all Ingasia,
which was his country, and then we came to the chief town of
Ingombe, which is half a day’s journey to go up.50
This lord came upon us with more than twenty thousand bows, and
spoilt many of our men. But with our shot we made a great spoil
among them, whereupon he retired up into the mountain, and sent
one of his captains to our general, signifying that the next day he
would obey him. The next day he entered our camp with great
pomp, with drums, petes,51 and Pongoes,52 or waits, and was royally
received; and he gave great presents, and greatly enriched the
general, and them which marched up. Upon the top of the mountain
is a great plain, where he hath his chief town; very fresh, full of
palm-trees, sugar-canes, potatoes, and other roots, and great store
of oranges and lemons. Here is a tree that is called Engeriay,53 that
beareth a fruit as big as a pome-water,54 and hath a stone in it,
present remedy (sic) for the wind colic, which was strange to the
Portugals. Here is a river of fresh water, that springeth out of the
mountains and runneth all along the town. We were here five days,
and then we marched up into the country, and burned and spoiled
for the space of six weeks, and then returned to Engombe again,
with great store of margarite stones,55 which are current money in
that land. Here we pitched our camp a league from this pleasant
mountain, which remained twelve months: but I was shot in my
right leg, and many Portugals and Mulatoes were carried to the city
to be cured.

[A Voyage to Benguella.]
Then the governor sent a fregatte to the southward, with sixty
soldiers, myself being one of the company, and all kinds of
commodities. We turned up to the southward until we came into
twelve degrees. Here we found a fair sandy bay. The people of this
place brought us cows and sheep, wheat56 and beans; but we staid
not there, but came to Bahia das Vaccas: that is, the Bay of Cows,
which the Portugals call Bahia de Torre,57 because it hath a rock like
a tower. Here we rode on the north side of the rock, in a sandy bay,
and bought great store of cows, and sheep—bigger than our English
sheep—and very fine copper. Also, we bought a kind of sweet wood,
called Cacongo,58 which the Portugals esteem much, and great store
of wheat and beans. And having laded our bark we sent her home;
but fifty of us staid on shore, and made a little fort with rafters of
wood, because the people of this place are treacherous, and not to
be trusted. So, in seventeen days we had five hundred head of
cattle; and within ten days the governor sent three ships, and so we
departed to the city.
In this bay may any ship ride without danger, for it is a smooth
coast. Here may any ship that cometh out of the East Indies refresh
themselves. For the Portugals carracks59 now of late come along the
coast, to the city, to water and refresh themselves. These people are
called Endalanbondos,60 and have no government among
themselves, and therefore they are very treacherous, and those that
trade with these people must stand upon their own guard. They are
very simple, and of no courage, for thirty or forty men may go boldly
into the country and fetch down whole herds of cattle. We bought
the cattle for blue glass beads of an inch long, which are called
Mopindes,61 and paid fifteen beads for one cow.
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!

ebookname.com