Modelling With Ordinary Differential

Equations: A Comprehensive Approach

(Chapman & Hall/Crc Numerical
Analysis and Scientific Computing)
Alfio Borzì
Visit to download the full and correct content document:
https://textbookfull.com/product/modelling-with-ordinary-differential-equations-a-comp
rehensive-approach-chapman-hall-crc-numerical-analysis-and-scientific-computing-alf
io-borzi/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Modelling with Ordinary Differential Equations Dreyer

https://textbookfull.com/product/modelling-with-ordinary-
differential-equations-dreyer/

Ordinary Differential Equations and Boundary Value

Problems Volume I Advanced Ordinary Differential
Equations 1st Edition John R Graef

https://textbookfull.com/product/ordinary-differential-equations-
and-boundary-value-problems-volume-i-advanced-ordinary-
differential-equations-1st-edition-john-r-graef/

Ordinary Differential Equations Principles and

Applications Cambridge IISc Series 1st Edition A. K.
Nandakumaran

https://textbookfull.com/product/ordinary-differential-equations-
principles-and-applications-cambridge-iisc-series-1st-edition-a-
k-nandakumaran/

Lectures, Problems And Solutions For Ordinary

Differential Equations: Second Edition Yuefan Deng

https://textbookfull.com/product/lectures-problems-and-solutions-
for-ordinary-differential-equations-second-edition-yuefan-deng/
From Ordinary to Partial Differential Equations 1st
Edition Giampiero Esposito (Auth.)

https://textbookfull.com/product/from-ordinary-to-partial-
differential-equations-1st-edition-giampiero-esposito-auth/

Differential Equations A Modern Approach with Wavelets

Textbooks in Mathematics 1st Edition Steven Krantz

https://textbookfull.com/product/differential-equations-a-modern-
approach-with-wavelets-textbooks-in-mathematics-1st-edition-
steven-krantz/

Ordinary differential equations : an introduction to

the fundamentals Second Edition. Edition Howell

https://textbookfull.com/product/ordinary-differential-equations-
an-introduction-to-the-fundamentals-second-edition-edition-
howell/

Methods of Mathematical Modelling: Fractional

Differential Equations 1st Edition Harendra Singh

https://textbookfull.com/product/methods-of-mathematical-
modelling-fractional-differential-equations-1st-edition-harendra-
singh/

Handbook of Ordinary Differential Equations: Exact

Solutions, Methods, and Problems: Third Edition Andrei
D. Polyanin

https://textbookfull.com/product/handbook-of-ordinary-
differential-equations-exact-solutions-methods-and-problems-
third-edition-andrei-d-polyanin/
 Modelling with
Ordinary Differential
Equations
A Comprehensive Approach
Numerical Analysis and Scienfic Compung Series
Series Editors:
Frederic Magoules, Choi-Hong Lai

About the Series

This series, comprising of a diverse collec!on of textbooks, references, and handbooks,
brings together a wide range of topics across numerical analysis and scien!fic compu!ng.
The books contained in this series will appeal to an academic audience, both in
mathema!cs and computer science, and naturally find applica!ons in engineering and the
physical sciences.

Iterave Spli!ng Methods for Differenal Equaons

Juergen Geiser

Handbook of Sinc Numerical Methods

Frank Stenger

Computaonal Methods for Numerical Analysis with R

James P Howard, II

Numerical Techniques for Direct and Large-Eddy Simulaons

Xi Jiang, Choi-Hong Lai

Decomposion Methods for Differenal Equaons

Theory and Applica!ons
Juergen Geiser

Mathemacal Objects in C++

Computa!onal Tools in A Unified Object-Oriented Approach
Yair Shapira

Computaonal Fluid Dynamics

Frederic Magoules

Mathemacs at the Meridian

The History of Mathema!cs at Greenwich
Raymond Gerard Flood, Tony Mann, Mary Croarken

Modelling with Ordinary Differenal Equaons

A Comprehensive Approach
Alfio Borzì

For more informa!on about this series please visit: h#ps://www.crcpress.com/Chapman--HallCRC-

Numerical-Analysis-and-Scien!fic-Compu!ng-Series/book-series/CHNUANSCCOM
 Modelling with
Ordinary Differential
Equations
A Comprehensive Approach

Alfio Borzì
University of Würzburg
First edition published 2020
by CRC Press
6000 Broken Sound Parkway NW, Suite 300, Boca Raton, FL 33487-2742

and by CRC Press

2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN

c 2020 by Taylor & Francis Group, LLC

CRC Press is an imprint of Taylor & Francis Group, LLC

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

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

For permission to photocopy or use material electronically from this work, access
www.copyright.com or contact the Copyright Clearance Center, Inc. (CCC), 222 Rose-
wood Drive, Danvers, MA 01923, 978-750-8400. For works that are not available on CCC
please contact [email protected]

Trademark notice: Product or corporate names may be trademarks or registered trade-

marks, and are used only for identification and explanation without intent to infringe.

ISBN: 9780815392613 (hbk)

ISBN: 9781351190398 (ebk)

Visit the eResources: https://www.crcpress.com/9780815392613

 To my beloved nymph

Mila
(Liudmila Shertsinger)

... quam nunc tibi facta renarro, in latices mutor. sed

enim cognoscit amatas amnis aquas positoque viri,
quod sumpserat, ore vertitur in proprias, ut se mihi
misceat, undas.

(Metamorphosen, Publius Ovidius Naso (P. Ovidii Nasonis) 635, Liber V. )

much quicker than my story could be told, my body

was dissolved to flowing streams. But still the River
recognised the waves, and for the love of me
transformed his shape from human features to his
proper streams, that so his waters might encompass
mine.

(translation by Brookes More)

Contents

Preface xi

Author xv

1 Introduction 1

1.1 Ordinary differential equations . . . . . . . . . . . . . . . . . 1

1.2 The modelling process . . . . . . . . . . . . . . . . . . . . . . 4

2 Elementary solution methods for simple ODEs 13

2.1 Simple ODEs . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.1.1 Simple ODE of 1. type, y 0 = f (x) . . . . . . . . . . . . 13
2.1.2 Simple ODE of 2. type, y 0 = f (y) . . . . . . . . . . . . 14
2.1.3 Simple ODE of 3. type, y 0 = f (x) g(y) . . . . . . . . . 14
2.1.4 Simple ODE of 4. type, y 0 = f (ax + by + d) . . . . . . 15
2.1.5 Simple ODE of 5. type, y 0 = f ( xy ) . . . . . . . . . . . 15
2.2 Linear ODEs . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.3 Method of variation of the constants . . . . . . . . . . . . . . 16
2.4 Bernoulli’s differential equation . . . . . . . . . . . . . . . . . 18
2.5 Riccati’s differential equation . . . . . . . . . . . . . . . . . . 19
2.6 Exact differential equations . . . . . . . . . . . . . . . . . . . 20

3 Theory of ordinary differential equations 25

3.1 The Cauchy problem and existence of solutions . . . . . . . . 25

3.2 Euler’s method . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.3 Uniqueness of solutions . . . . . . . . . . . . . . . . . . . . . 30
3.4 The Carathéodory theorem . . . . . . . . . . . . . . . . . . . 34

4 Systems of ordinary differential equations 39

4.1 Systems of first-order ODEs . . . . . . . . . . . . . . . . . . 39

4.2 Dependence of solutions on the initial conditions . . . . . . . 41
4.3 Systems of linear ODEs . . . . . . . . . . . . . . . . . . . . . 48
4.4 Systems of linear homogeneous ODEs . . . . . . . . . . . . . 49
4.5 The d’Alembert reduction method . . . . . . . . . . . . . . . 52

vii
viii Contents

4.6 Nonhomogeneous linear systems . . . . . . . . . . . . . . . . 55

4.7 Linear systems with constant coefficients . . . . . . . . . . . 57
4.8 The exponential matrix . . . . . . . . . . . . . . . . . . . . . 65
4.9 Linear systems with periodic coefficients . . . . . . . . . . . 71

5 Ordinary differential equations of order n 77

5.1 Ordinary differential equations of order n in normal form . . 77

5.2 Linear differential equations of order n . . . . . . . . . . . . 78
5.3 The reduction method of d’Alembert . . . . . . . . . . . . . 80
5.4 Linear ODEs of order n with constant coefficients . . . . . . 81
5.5 Nonhomogeneous ODEs of order n . . . . . . . . . . . . . . . 83
5.6 Oscillatory solutions . . . . . . . . . . . . . . . . . . . . . . . 85

6 Stability of ODE systems 89

6.1 Local stability of ODE systems . . . . . . . . . . . . . . . . . 89

6.2 Stability of linear ODE systems . . . . . . . . . . . . . . . . 94
6.3 Stability of nonlinear ODE systems . . . . . . . . . . . . . . 96
6.4 Remarks on the stability of periodic ODE problems . . . . . 97
6.5 Autonomous systems in the plane . . . . . . . . . . . . . . . 99
6.6 The Lyapunov method . . . . . . . . . . . . . . . . . . . . . 106
6.7 Limit points and limit cycles . . . . . . . . . . . . . . . . . . 109
6.8 Population dynamics . . . . . . . . . . . . . . . . . . . . . . 114
6.9 The Lorenz model . . . . . . . . . . . . . . . . . . . . . . . . 125
6.10 Synchronisation . . . . . . . . . . . . . . . . . . . . . . . . . 130

7 Boundary and eigenvalue problems 139

7.1 Linear boundary-value problems . . . . . . . . . . . . . . . . 139

7.2 Sturm-Liouville eigenvalue problems . . . . . . . . . . . . . . 142

8 Numerical solution of ODE problems 149

8.1 One-step methods . . . . . . . . . . . . . . . . . . . . . . . . 149

8.2 Motion in special relativity . . . . . . . . . . . . . . . . . . . 166
8.3 The Kepler problem . . . . . . . . . . . . . . . . . . . . . . . 174
8.4 Approximation of Sturm-Liouville problems . . . . . . . . . . 179
8.5 The shape of a drop on a flat surface . . . . . . . . . . . . . 185

9 ODEs and the calculus of variations 189

9.1 Existence of a minimum . . . . . . . . . . . . . . . . . . . . . 189

9.2 Optimality conditions . . . . . . . . . . . . . . . . . . . . . . 194
9.2.1 First-order optimality conditions . . . . . . . . . . . . 197
9.2.2 Second-order optimality conditions . . . . . . . . . . . 198
 Contents ix

9.3 The Euler-Lagrange equations . . . . . . . . . . . . . . . . . 201

9.3.1 Direct and indirect numerical methods . . . . . . . . . 206
9.3.2 Unilateral constraints . . . . . . . . . . . . . . . . . . 207
9.3.3 Free boundaries . . . . . . . . . . . . . . . . . . . . . . 208
9.3.4 Equality constraints . . . . . . . . . . . . . . . . . . . 209
9.4 The Legendre condition . . . . . . . . . . . . . . . . . . . . . 212
9.5 The Weierstrass-Erdmann conditions . . . . . . . . . . . . . 216
9.6 Optimality conditions in Hamiltonian form . . . . . . . . . . 220

10 Optimal control of ODE models 225

10.1 Formulation of ODE optimal control problems . . . . . . . . 225

10.2 Existence of optimal controls . . . . . . . . . . . . . . . . . . 227
10.3 Optimality conditions . . . . . . . . . . . . . . . . . . . . . . 232
10.4 Optimality conditions in Hamiltonian form . . . . . . . . . . 239
10.5 The Pontryagin’s maximum principle . . . . . . . . . . . . . 243
10.6 Numerical solution of ODE optimal control problems . . . . 258
10.7 A class of bilinear optimal control problems . . . . . . . . . . 263
10.8 Linear-quadratic feedback-control problems . . . . . . . . . . 270

11 Inverse problems with ODE models 275

11.1 Inverse problems with linear models . . . . . . . . . . . . . . 275

11.2 Tikhonov regularisation . . . . . . . . . . . . . . . . . . . . . 279
11.3 Inverse problems with nonlinear models . . . . . . . . . . . . 283
11.4 Parameter identification with a tumor growth model . . . . . 284

12 Differential games 289

12.1 Finite-dimensional game problems . . . . . . . . . . . . . . . 289

12.2 Infinite-dimensional differential games . . . . . . . . . . . . . 295
12.3 Linear-quadratic differential Nash games . . . . . . . . . . . 298
12.4 Pursuit-evasion games . . . . . . . . . . . . . . . . . . . . . . 304

13 Stochastic differential equations 309

13.1 Random variables and stochastic processes . . . . . . . . . . 309

13.2 Stochastic differential equations . . . . . . . . . . . . . . . . 314
13.3 The Euler-Maruyama method . . . . . . . . . . . . . . . . . 321
13.4 Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
13.5 Piecewise deterministic processes . . . . . . . . . . . . . . . . 327

14 Neural networks and ODE problems 331

14.1 The perceptron and a learning scheme . . . . . . . . . . . . . 331

14.2 Approximation properties of neural networks . . . . . . . . . 339
x Contents

14.3 The neural network solution of ODE problems . . . . . . . . 345

14.4 Parameter identification with neural networks . . . . . . . . 350
14.5 Deep neural networks . . . . . . . . . . . . . . . . . . . . . . 354

Appendix Results of analysis 357

A.1 Some function spaces . . . . . . . . . . . . . . . . . . . . . . 357

A.1.1 Spaces of continuous functions . . . . . . . . . . . . . 357
A.1.2 Spaces of integrable functions . . . . . . . . . . . . . . 359
A.1.3 Sobolev spaces . . . . . . . . . . . . . . . . . . . . . . 359
A.2 The Arzelà-Ascoli theorem . . . . . . . . . . . . . . . . . . . 362
A.3 The Gronwall inequality . . . . . . . . . . . . . . . . . . . . 364
A.4 The implicit function theorem . . . . . . . . . . . . . . . . . 364
A.5 The Lebesgue dominated convergence theorem . . . . . . . . 366

Bibliography 369

Index 381
Preface

I understand modelling with ordinary differential equations (ODEs) in a broad

sense, and I have always considered this topic as the perfect framework to
introduce students to many modern fields of applied mathematics. I also think
that modelling with ODEs is a very rich research field and all these reason have
motivated me to write this book. It contains much of the teaching material
and resources that I have collected in the course of some years of lecturing on
these equations and related fields.
The book addresses different topics ranging from the general theory of
ODEs, stability properties of solutions and their computation by numerical
methods, to specialised topics with ODE models such as calculus of variation,
optimal control and inverse problems, differential games, stochastic differential
equations, and neural networks. Therefore this book has a very broad scope,
while by no means being exhaustive in any of the topics mentioned above,
but with the purpose to provide students and interested scientists with a first
(sometime intense) reading that illustrates the many facets of modelling with
ordinary differential equations.
In my view the set of topics presented in this book results from our general
attempt to model many real-world systems with ODE equations, and this
effort goes much beyond the mere definition a set of equations whose solution
qualitatively reproduces some observed behaviour or configuration. In fact,
assuming to have such a candidate model, we face the problem of calibrating
it to fit results of measurements on the real system, and once the model
becomes quantitatively adequate, we may use it to make accurate predictions
on future evolution and attainable configurations, and also to design a control
strategy to achieve a desired goal. This reasoning should explain the structure
and the content of the book that I would like to illustrate in the following.
A short introduction defines what an ODE is, and gives some examples
that can be understood just by having knowledge of real analysis. This is part
of my attempt, throughout the book, to accompanying abstract statements
with examples. Also in the introduction, I illustrate some principles behind
the modelling process, and how to investigate the range of validity of a model.
In this introductory chapter, I use the concepts of derivative and integral
whose construction was the prelude to the formulation of ordinary differential
equations. The former is the essential component in the formulation of an
ODE, the latter is the essential tool for its solution. This fact is illustrated in

xi
xii Preface

Chapter 2, where I present some classes of ODE problems, as the Bernoulli

and Riccati differential equations, which can be solved analytically.
However, already in the earlier investigation of ODE problems, scientists
were confronted with models for which the direct construction of solutions by
integration methods appeared not viable, thus posing the question whether a
solution may exist and how to construct it. This is a central issue in the devel-
opment of the ODE theory, and it is the focus of the Chapters 3, 4, and 5. In
Chapter 3, I present proofs of existence and uniqueness of solutions starting
from the theory of Peano and concluding with that of Carathéodory. At this
point, I decided to present the proofs of the corresponding theorems based on
Euler’s method and the Arzelà-Ascoli theorem, and to avoid any explicit use of
fix-point theorems, because in my opinion, also supported by discussions with
my students, with the Euler’s approach we get a better insight in the structure
of ODE initial-value problems and in the strategy of the proofs. Furthermore,
by following this approach, we get a deeper understanding of numerical meth-
ods for solving ODEs. Next, in Chapters 4 and 5, I illustrate systems of ODEs
and differential equations with higher-order derivatives. The purpose is also to
discuss the extension of the theoretical framework of Chapter 3 to these cases,
considering different classes of problems, and to explicitly construct solutions
in special cases. Chapters 4 and 5 are concluded with a discussion on models
with periodic coefficients and problems with oscillatory solutions, respectively.
Assuming to have a valid model of a real system, we would like to establish
the configuration, resp. behaviour, of this system forever. This is the main
motivation for the stability theory presented in Chapter 6. This theory plays
an essential role in the modelling process, and I try to illustrate this fact by
discussing the development of models for population dynamics, the Lorenz
model, and the phenomenon of synchronisation.
In some sense, the first part of this book is completed with Chapters 7 and
8. In the former, I present a short survey of results concerning boundary- and
eigenvalue problems with ODEs. In Chapter 8, I discuss two classes of numer-
ical methods to solve ODE problems, that is, one-step methods for Cauchy
problems and the finite-volume approximation of boundary- and eigenvalue
problems. These methods have large applicability and are supported by well-
developed theoretical tools that I illustrate in this chapter. With these meth-
ods, we can solve different models of large interest. Further, I present the
derivation of a relativistic mechanical model and discuss a one-step method
that is covariant under Lorentz transformation. Thereafter, I discuss the nu-
merical solution of the Kepler problem in classical and relativistic mechanics.
I conclude this chapter illustrating the numerical approximation of Sturm-
Liouville problems and the modelling of a drop of liquid over a solid flat
surface.
Chapter 9 is devoted to the so-called calculus of variation, which is a
fundamental topic in the theoretical development and application of ordinary
differential equations. I tried to make this chapter self-consistent also with
the help of additional material in the Appendix. Chapter 9 and the following
 Preface xiii

ones may be considered more advanced, because they require some additional
knowledge of elements of functional analysis that I have tried to convey in
simple terms.
With the calculus of variation, one can very well illustrate the connec-
tion between infinite-dimensional optimisation and ODEs, and it represents
a central topic in theoretical mechanics. The calculus of variation is also fun-
damental, and it was so also historically, in the formulation and analysis of
optimal control problems governed by ODEs. This is the topic of Chapter 10,
where I discuss both the Lagrange’s and Pontryagin’s frameworks to charac-
terise optimal controls.
Clearly, calculus of variation and optimal control theory are two impor-
tant tools in the modelling with ODEs. The first allows to derive an ODE
model from a “first principle” that is formulated in terms of a functional to
be minimised; as the principle of least action that is of central use in physics.
Optimal control theory is of paramount importance in applications; for exam-
ple, in space flight navigation, control of nano-systems, automotive systems,
robotics, etc.
The tools provided in Chapters 9 and 10 are also instrumental to access the
issue of inverse problems with ODEs. Indeed, whatever ODE model we have, it
becomes useful once we can identify the values of the parameters and functions
entering in it, possibly based on measurements of the real phenomenon that
the ODE should represent. For this purpose, in Chapter 11, I illustrate the
formulation of inverse problems and the challenge posed by their solution,
which in some cases can be obtained based on the optimal control framework.
This chapter is concluded with an application of a parameter identification of
a tumor-growth model.
Chapter 12 is devoted to differential games, which are closely related to
optimal control problems with multiple objectives and controls. Differential
games are essential in the modelling of problems in economics and, more
in general, of problems of conflict and cooperation. This chapter mainly fo-
cuses on the so-called Nash equilibrium concept that I first discuss in a finite-
dimensional framework, and later in the ODE framework.
In Chapter 13, I illustrate the important further development of ODEs in
the realm of stochastic processes. In this chapter, I attempt to discuss this
development step-by-step starting from the definition of a random variable to
arrive at the derivation of a stochastic differential equation including drift, dif-
fusion, and jumps. This chapter is concluded with an introduction to piecewise
deterministic processes.
Chapter 14 is devoted to neural networks and their use to solve ODEs
and related inverse problems. Neural networks represent a class of solution
strategies for different problems ranging from approximation and classification
to solution of differential models.
I think that all concepts and methods mentioned above are better un-
derstood if they are also realised in a numerical setting. For this reason, in
many chapters, I present results of numerical experiments. The corresponding
xiv Preface

codes are available https://www.crcpress.com/9780815392613 (Additional

Resources).
I hope that I have, at least partially, succeeded in my goal to present a
large variety of topics related to modelling with ODEs together and in a unified
manner. Indeed all these topics are otherwise very well treated in dedicated
textbooks. In fact, in my endeavour to reach this goal, I have relied mostly on
very valuable classical references, and in some cases I have mentioned recent
results. However, the ODEs’ world is very large and continuously growing,
and this book can only be a starting point to explore this wonderful field of
mathematics.
Although some of the topics considered in this book can be considered
advanced, the methods and problems are presented with enough details such
that the book may serve as a text book for undergraduate and graduate stu-
dents, and as an introduction for researcher in sciences and engineering that
intend to work with ODE models.
I would like to thank many colleagues and students who have encour-
aged my teaching and research work, and contributed through discussion,
remarks, suggestions, and all that, to improve my insight in applied mathe-
matics. In particular at this point, I would like to acknowledge the contin-
ued support of Mario Annunziato, Jan Bartsch, Ugo Boscain, Tim Breiten-
bach, Francesca Calà-Campana, Kurt Chudej, Gabriele Ciaramella, Andrei V.
Dmitruk, Matthias Ehrhardt, Bino Fonte, Andrei V. Fursikov, Omar Ghattas,
Lars Grüne, Abdou Habbal, Bernadette Hahn, Nadja Henning, Kees Ooster-
lee, Hans Josef Pesch, Georg Propst, Souvik Roy, Richard Schmähl, Volker
Schulz, Georg Stadler, Fredi Tröltzsch, Marco Verani, and Greg von Winckel
(in alphabetical order).
Heartiest thanks to Georg Propst and to Mario Annunziato and Marco
Verani for reading early drafts of this book. I also would like to thank the
anonymous Referees who helped improve with their comments the final version
of this book. Further, I would like to gratefully acknowledge the help of Petra
Markert-Autsch for copyediting some chapters of this work, and the support
of my daughter Zoe in the design of the cover of this book.
For the fortunate inception of this book, I am very grateful to Choi-Hong
Lai who encouraged me to submit a book to CRC Press. I would like to thank
very much Sarfraz Khan, Callum Fraser, and Mansi Kabra from the Editorial
team and Kari Budyk, Production Editor of CRC Press/Taylor & Francis
Group for their kind and very professional assistance in publishing this work.
I owe my thanks also to Narayani Govindarajan, Project Manager and her
team at Nova Techset for their support on this project.

Alfio Borzì
Würzburg, 2019
Author

Alfio Borzì, born 1965 in Catania (Italy), is the professor and chair of
Scientific Computing at the Institute for Mathematics of the University of
Würzburg, Germany. He studied Mathematics and Physics in Catania and
Trieste where he received his Ph.D. in Mathematics from Scuola Internazionale
Superiore di Studi Avanzati (SISSA).
He served as Research Officer at the University of Oxford (UK) and as
assistant professor at the University of Graz (Austria) where he completed
his Habilitation and was appointed as Associate Professor. Since 2011 he has
been Professor of Scientific Computing at University of Würzburg.
Alfio Borzì is the author of three mathematics books and numerous articles
in journals. The main topics of his research and teaching activities are mod-
elling and numerical analysis, optimal control theory and scientific computing.
He is a member of the editorial board for the SIAM Journal on Scientific Com-
puting and for SIAM Review.

xv
Chapter 1
Introduction

1.1 Ordinary differential equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 The modelling process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

In this chapter, ordinary differential equations are defined and illustrated by

means of examples, also with the purpose to introduce the notation and the
basic terminology used throughout this book.
In the second part of this introduction, the main principles behind the
modelling process are discussed focusing on a simple model of population
growth and on Newton’s model of gravitation dynamics.

1.1 Ordinary differential equations

An ordinary differential equation (ODE) is an equation relating a function
of one independent variable to some of its derivatives with respect to this
variable.
Ordinary differential equations represent an important field of mathemat-
ics and its story begins in the seventeenth century and still constitutes a very
active and broad field of research and application in the sciences and tech-
nology. The concept of a differential equation was established with the works
of Gottfried Wilhelm Leibniz and Isaac Newton, of the brothers Jakob I and
Johann Bernoulli, Daniel Bernoulli, and of Leonhard Euler, Giuseppe Luigi
Lagrangia (Joseph-Louis Lagrange), and Pierre-Simon Laplace, among others.
These mathematicians also started the development of a general theory for
ODEs along with numerous applications in geometry, mechanics, and optimi-
sation. This remarkable mathematical development continued in the 19th cen-
tury with the works of Augustin-Louis Cauchy and Giuseppe Peano, Charles
Émile Picard, Henri Poincarè, Vito Volterra, Constantin Carathéodory, etc.,
who greatly contributed to the foundation of the modern theory and method-
ology of ordinary differential equations.
Before giving some introductory examples of ODEs, let us assume that the
independent variable represents time t, then it is quite common to denote the
derivative of function with respect to this variable by a point on the top of this
function, and multiple dots would indicate higher-order derivatives. Thus, if

1
2 Modelling with Ordinary Differential Equations
2
y is a function of time t, we denote ẏ = dy
dt , ÿ = dt2 , and so on. However, this
d y

notation is not unique: derivatives can be denoted by multiple apex as follows

d2 y
y 0 = dy
dt , y = dt2 , etc., and when higher-order
00
derivatives are involved the
(n)
following notation is commonly used, y = dtn . The independent variable
dn y

may represent other quantities as well, e.g., a space coordinate, and this we
usually prefer to denote with x instead of t, and the same notation concerning
derivatives is used. In general, we use t and x interchangeably. Notice that we
consider y, t, and x to be real valued. We do not discuss complex differential
equations, nor differential equations on general differential manifolds; for these
topics, the interested reader may consider the classical references [88, 89].
An ordinary differential equation is said to be of order n if this is the order
of the highest-order derivative appearing in the equation. A general form for
an ODE of order n is as follows:

F (x, y, y 0 , y 00 , ..., y (n−1) , y (n) ) = 0. (1.1)

A solution to (1.1), in an interval I ⊂ R of the independent variable x, is a

n-times differentiable function y = y(x) that satisfies (1.1) in the interval I.
In many cases, it is possible to write (1.1) in the so-called normal form
where the highest-derivative appears explicitly in the equation as follows:

y (n) = f (x, y, y 0 , ..., y (n−1) ). (1.2)

In particular, an n-th order ODE is linear, if it can be written in the form

a0 (x) y (n) + a1 (x) y (n−1) + . . . + an (x) y = g(x), (1.3)

where the functions ak (x) are called the coefficient functions. A nonlinear
ODE is one where F or f above are not linear with respect to one of their
arguments involving the function y or any of its derivatives.
As we shall discuss later in detail, a n-th order ODE may admit a general
parametrised solution with n real parameters (c1 , ..., cn ). For example y(x) =
2
c1 e−x satisfies y 0 = −2xy for every real value of c1 and for every x ∈ R. On
2
the other hand, if the value of c1 = c is fixed, then y(x) = c e−x represents a
particular solution of the differential equation y = −2xy. For this reason, in
0

the following we write y(x; c1 , ..., cn ) to denote a general solution (if it exists)
to a n-th order ODE as (1.1), (1.2), or (1.3) in an interval I . If a solution
exists on all the real line, I = R, then it is called a global solution. A solution
of an n-th order ODE that cannot be obtained from an n-parameter family
of solutions is called a singular solution.
In the case that a general parametrised solution exists, then one can select
a particular solution by choosing the values of the parameters and this can
be done in an arbitrary way. In particular, one can determine these values by
requiring that the function y and its derivatives up to order (n − 1) at a given
point x0 ∈ I take given values as follows:
(n−1)
y(x0 ) = y0 , y 0 (x0 ) = y00 , . . . , y (n−1) (x0 ) = y0 . (1.4)
 Introduction 3

Finding a solution to a n-th order ODE that satisfies the so-called initial
conditions (1.4) defines a Cauchy or initial-value problem.
It is also possible to select a solution by specifying values of the function
y and of some of its derivatives at two or more distinct points of the interval
I. This is called a boundary-value problem.
Notice that, if a solution exists, it might be explicitly given by a function
of x, say y = φ(x). However, this is not always possible. In fact, a solution
can be obtained also in an implicit form through a relation R(x, y) = 0.

Example 1.1 Consider the following second-order ODE y 00 = 3y 0 − 2y with

initial conditions y(0) = 0 e y 0 (0) = 1. One can verify that the particular
solution is given by y(x) = −ex + e2x .
√
Example 1.2 Consider the following first-order ODE y 0 = 2 y with initial
condition y(0) = 0. We find two particular solutions: y(x) = 0 and y(x) = x2 ,
x ≥ 0.

Example 1.3 Consider the following first-order ODE xy 0 − y = 0 with initial

condition y(0) = 1. This Cauchy problem has no solution. On the other hand,
with the initial condition y(0) = 0, it admits an infinite number of solutions,
y(x) = c x, c ≥ 0.

Example 1.4 Consider the following second-order ODE y 00 + y = 0 with

boundary conditions y(0) = 0 and y(π) = 0. This boundary-value problem
has infinite many solutions. On the other hand, with the choice y(0) = 0 and
y(π) = 1, it has no solution.

Example 1.5 The function y(x) = x + 1/x is a solution to the ODE x2 y 00 +

xy 0 − y = 0 in the intervals (−∞, 0) and (0, ∞). It is not a solution in the
interval (−∞, ∞), since it is not defined at x = 0.

Example 1.6 Consider the nonlinear ODE (y 0 )2 + xy 0 − y = 0. This equation

has a one-parameter family of solutions given by y(x) = c x + c2 . However, it
also admits the singular solution y(x) = −x2 /4.

Example 1.7 Consider the ODE y 0 + x/y = 0. This equation has a one-
parameter family of solutions given in implicit form by the relation y 2 + x2 =
c2 .

Notice that in all examples above, the ODE is given in normal form or
can be easily put in this form. However, we may have the following ODE in
implicit form

F (x, y, y 0 ) = 0. (1.5)

In this case, the question arises if, in correspondence to (x, y), it is possible
to uniquely determine y 0 .
4 Modelling with Ordinary Differential Equations

To discuss this issue, consider the continuous function F (x, y, p) in a three-

dimensional domain D and assume that this function is zero at (x, y, p). The
question is if there exists a neighborhood U of (x, y) and a continuous function
f (x, y) such that p = f (x, y) and F (x, y, f (x, y)) = 0 in U . If so, then (x, y) is
a regular point; otherwise it is a singular point. Clearly, if (x, y) is a regular
point, then F (x, y, y 0 ) = 0 defines an ODE that is equivalent to y 0 = f (x, y)
in U . By the implicit function theorem, a sufficient condition for (x, y) to be
a regular point is that the Jacobian ∂F ∂p (x, y, p) =: Fp (x, y, p) be non-singular
at this point.

Example 1.8 Consider the implicit ODE y 02 = 4x2 . It corresponds to

F (x, y, p) = p2 − 4x2 that gives p = ±2x. Furthermore, Fp (x, y, p) = 2p is
non-zero (invertible) if and only if p 6= 0. Therefore, x = 0 is a singular point,
whereas for x 6= 0 we have a regular point. At each regular point, the implicit
ODE gives rise to two distinct linear ODEs, y 0 = 2x and y 0 = −2x, with
distinct solutions y(x; c) = x2 + c and y(x; c) = −x2 + c, respectively.

In the examples above, we are considering scalar ODE problems with y

real valued. In this setting, we are used to interpret the derivative y 0 at x as
the value of the slope of the tangent line to the function y that passes through
the point (x, y) in the Cartesian (x, y) coordinate space. This fact lets us
give a geometric interpretation of the differential equation y 0 = f (x, y). It
specifies the (tangent) direction field of a solution passing at the point (x, y)
in the domain where f is defined. A convenient way to depict part of this
direction field is by plotting it as the (tangent) vectors on points of a (x, y)
domain where the equation is considered. For simplicity, we also refer to this
representation as the direction field itself. This direction field is very useful
for a qualitative understanding of the behaviour of solutions to y 0 = f (x, y),
that is, their dependence on the choice of initial conditions, their asymptotic
behaviour, etc. For illustration, we plot in Figure 1.1 the direction field of
y 0 = x − sin(y) in the domain D = [−4, 4] × [−4, 4].

1.2 The modelling process

A mathematical model of a real system is a representation of this system by
means of mathematical structures made of variables and relationships between
these variables. The purpose of a model is to reproduce, to some degree of
accuracy, some features of an observed phenomenon that is associated to the
system. Therefore, part of the modelling process is also a criterion for testing
the accuracy of the model, and this involves data obtained by measurements
on the system. Thus, we can identify three steps in the modelling process:
 Introduction 5
5

-1

-2

-3

-4

-5
-4 -3 -2 -1 0 1 2 3 4 5

FIGURE 1.1: The direction field of y 0 = x − sin(y).

• characterisation of the system by selection of observed variables and

their measurements;
• mathematical formulation of the relationship between these variables;
and
• comparison of the output (prediction) of the model with additional avail-
able data.

As far as a model meets its purpose it is adequate. However, it is usually

not unique, and additional criteria, including mathematical aesthetics, can be
considered to narrow our choice among possible models.
To illustrate the modelling process mentioned above, we start with a very
simple problem of population dynamics of a species of animals in an isolated
environment; see [46] for more details. We observe this system and realise that
the number of members of the population changes with time, and we decide
to model this phenomenon by defining a mathematical relationship between
the number of members in the population and the time when these members
are counted.
Let x0 be the instant of time when we count N0 members. We denote
with N the function that represents the number of members at time x. In
particular, N (x0 ) = N0 . Suppose that we re-count the population after one
year starting from x0 . Thus, we have the time increment ∆x = 1 year. Now,
let N1 be the new population size after one year since x0 . Hence, we can
compute a growth factor r = N N1
0
in one year. Therefore, we can set a model
of growth of the population given by

N (xn ) = rN (xn−1 ) , n = 1, 2, . . . .

Hence, N (xn ) = rn N0 , where xn = x0 + n∆x.

6 Modelling with Ordinary Differential Equations

With this model, we can make a prediction on the future size of the pop-
ulation. In particular, we can predict the time when the population reaches
the size N2 > N1 as follows. Let N2 = N (xm ) and compute
 
N2 1
m = log ,
N0 log r

which gives xm = x0 + m∆x.

Notice that this model results from observation related to one year interval
(an arbitrary choice), while it would be desirable that the model is independent
of this choice. To circumvent this problem, we assume a linear relationship of
the growth factor r with the time increment ∆x as follows:

r = 1 + p ∆x,

where p is a constant to be determined. Further, we assume that the following

holds:
N (x + ∆x) = (1 + p ∆x) N (x).
This relation can be re-written as follows:
N (x + ∆x) − N (x)
= p N (x).
∆x
Taking the limit ∆x → 0, we obtain

N 0 (x) = p N (x), N (x0 ) = N0 , (1.6)

that is, a differential problem, where p defines the rate of growth. By integra-
tion of this continuous model, we obtain

N (x) = N0 ep(x−x0 ) , (1.7)

which differs from the previous discrete model. In fact, assuming that r =
1 + p ∆x is valid, then the parameter p in the two models takes different
values. In the first case, we have
N1
p= −1 per year.
N0
In the second case, we obtain
 
N1
p = log per year.
N0

Our model of population growth (1.7) appears to be specific of the popula-

tion considered: it contains N0 , and x0 , x refer to 1 year unit of time. However,
we have not specified the species of the population, and we expect that a sim-
ilar model results for different species and with different units of time. In fact,
 Introduction 7

we expect to find an ‘universal’ (platonic) model that has general applica-

bility. This universal model can be obtained by scaling. The scaling strategy
requires to identify quantities that are characteristic of the system, and to use
them as units of measure. In the model (1.6), we can take the characteristic
quantities N̄ = N0 , and x̄ = 1 year. Then, we define
x − x0 N
t= and y= .
x̄ N̄
Thus, we obtain
1 x̄ 0
y(t) = N (x̄ t + x0 ), y 0 (t) = N (x).
N̄ N̄
Therefore, we get
1 0
y (t) = p y(t)
x̄
1
In this model, all parameters can be removed choosing p = x̄ . This choice
leads to the following universal model for population growth

y 0 (t) = y(t), y(0) = 1.

The solution is obtained by integration, y(t) = et , and this result can be used
for any specific case of population growth with the same behaviour as follows:
 
x − x0
N (x) = N̄ y(t) = N0 y = N0 ep(x−x0 ) .
x̄

Another (related) way to arrive at the same universal model is by non-

dimensionalisation. To illustrate this procedure, we denote with [q] the di-
mension/units of the quantity q. Specifically, if q = x represents time, then
[x] = T (e.g., seconds); if q = ` represents length, then [`] = L (e.g., metres);
if q = m represents mass, then [m] = M (e.g., kilograms), etc.
For our population model (1.6), we have [x] = T , [N ] = S (S means size),
[N 0 ] = TS , [p] = T1 . With this preparation, we can perform a dimensional
analysis with the purpose of finding a relation between the parameters of our
model, that is, x̄, N0 , and p. In particular, we assume that there is a relation
between these parameters such that we can express some of them in terms of
the remaining ones. Specifically, we consider the following relation

x̄ = pα N0β α, β ∈ Z, (1.8)

where α and β ought to be determined.

This is possible by considering the dimensions of the variables involved.

α In
fact, the relation above requires that [T ] = [p]α [N0 ]β , that is, T = T1 Sβ ,
which gives α = −1, β = 0. Hence, p = x̄1 , as we have already obtained by the
scaling approach. Notice that, with the choice p = x̄1 , the resulting model is
dimensionless.
8 Modelling with Ordinary Differential Equations

The scaling strategy and the dimensional analysis are powerful tools to
choose appropriate units of measure depending on the application of the prob-
lem. This fact can be illustrated considering the Newton model of a stone of
mass m thrown vertically from the surface of our planet with velocity υ0 ; see
[46] for further details.
Our model is given by Newton’s second law of dynamics m a = F , where
a denotes acceleration of the stone of mass m subject to a force F , and by
Newton’s gravitational law
ME m
F = −G ,
(r + R)2
where F represents the gravitational force between the stone at distance r
from the ground and our planet having mass ME and radius R; G is the
gravitational constant.
Now, let us define the acceleration constant g = G M
R2 , and consider m a =
E

F with the gravitational force. We obtain

g R2
z 00 (x) = −
2 , (1.9)
z(x) + R
where z denotes the vertical position of the stone from the ground at time x.
We assume that the stone is thrown vertically from the ground at time x = 0
and with velocity υ0 . Thus, we have z(0) = 0 and z 0 (0) = υ0 . Notice that in
(1.9) the masses do not appear explicitly, and the characteristic quantities are
given by
metre metre
g ≈ 9.8 , R ≈ 6.371 106 metre, υ0 = 10 . (1.10)
sec2 sec
These parameters have the following dimensions:
L L
[g] = , [R] = L, [υ0 ] = .
T2 T
Now, as in the previous model, we look for a characteristic quantity that
involves as few dimensions as possible. Consider

π = υ0α g β Rγ .

We have
 α  β
L L
[π] = Lγ = Lα+β+γ T −α−2β .
T T2
Therefore, π is dimensionless if α + β + γ = 0 and −α − 2β = 0. Thus, taking

β
α = −2β and γ = β, we obtain π = gυR 2 . With the calculation that follows,
0
υ02
we shall see that the dimensionless parameter ε = gR represents a universal
constant of our problem.
 Introduction 9

Next, we look for π with the dimension of length. In this case, α+β +γ = 1
and −α − 2β = 0. In this way, we obtain the characteristic length

` = υ0−2β g β R1+β = R ε−β .

Another possibility is to have π with time dimension. In this case α+β+γ =

0 and −α − 2β = 1. Hence, the following characteristic time is obtained
R −β
χ = υ0−1−2β g β R1+β = ε .
υ0
Notice that, in all cases, the value of β must be chosen. In particular, if we
take β = 0, then ` = R and χ = R/υ0 , which means that we are choosing a
very large unit of length and of time. Therefore, this choice may be appropriate
or not depending on the data of the problem. This issue is clarified with the
following discussion, where we use scaling to obtain universal versions of (1.9)
with the initial conditions z(0) = 0 and z 0 (0) = υ0 .
Let z̄ represents our unit length, and x̄ be our unit of time. We introduce
the new variable y such that
x
z(x) = z̄ y(τ ), τ= .
x̄
Replacing this function in (1.9), we obtain
z̄ 1
y 00 (τ ) = −  2 , (1.11)
x̄2 g

z̄
R y(τ ) + 1

with the initial conditions

x̄
y(0) = 0, y 0 (0) = υ0 .
z̄
In (1.11), the following three terms appear:
z̄ z̄ x̄
, , υ0 ,
x̄2 g R z̄
which depend on the parameters x̄ and z̄. For the purpose of simplifying (1.11),
we can choose these parameters to make two of the three terms above equal
to 1. We have the following three choices.
q
(a) Let x̄2z̄ g = 1 and R
z̄
= 1. Hence, we obtain z̄ = R and x̄ = R
g , and
q 2 √
z̄ υ0 =
x̄ √υ0 = Rg =
υ0
ε. We obtain the following dimensionless
Rg
model:
1 √
y 00 (τ ) = − , y(0) = 0, y 0 (0) = ε.
(y(τ ) + 1)2
10 Modelling with Ordinary Differential Equations

(b) Let z̄
R = 1 and z̄ υ0 = 1. Thus,
x̄
we have z̄ = R and x̄ = υ0 .
R
Further, we
υ02
obtain z̄
x̄2 g = R g = ε, and the following model results

1
ε y 00 (τ ) = − , y(0) = 0, y 0 (0) = 1.
(y(τ ) + 1)2

υ02
(c) Let x̄2 g = 1 and z̄ υ0 = 1. Hence,
z̄ x̄
we have x̄ = υ0
g and z̄ = g , and we
υ2
obtain R z̄
= R0g = ε. The resulting model is given by

1
y 00 (τ ) = − , y(0) = 0, y 0 (0) = 1.
( y(τ ) + 1)2

Notice that the models in (a), (b), and (c) are all dimensionless. They differ in
the scaling choice and all include ε. However, it is the value of this parameter
that determines which model is adequate.
In fact, consider the data of our problem given in (1.10), and notice that
in this case ε ≈ 10−6 , which appears very small compared to the values of the
data. For this reason, it seems appropriate to neglect it in our three model
variants (a), (b), and (c).
If we set ε = 0 in model (a), we have y(0) = 0, y 0 (0) = 0, y 00 (0) < 0,
which results in a negative value of y, which is not possible. Therefore this
model is not appropriate for our purpose. In fact, a characteristic length of
z̄ = 107 metre, and a characteristic time of x̄ = 103 sec, are too big to describe
the phenomenon of a stone thrown with a velocity of υ0 = 10 metre/sec. The
same remark holds for the model (b), where setting ε = 0 results in a problem
that is not well posed.
On the other hand, if we set ε = 0 in model (c), then we obtain the
following differential problem:

y 00 (τ ) = −1, y(0) = 0, y 0 (0) = 1.

The solution to this problem is obtained by integration: y(τ ) = − 21 τ 2 + τ .

Transforming back to the original variables, we have the well-known result
z(x) = υ0 x − 21 g x2 . Therefore model c) is the appropriate model to describe
our problem with the data given in (1.10). In fact, in this setting we have
z̄ ∼ 10 m and x̄ ∼ 1 s.
In this section, we have illustrated two techniques of how to reformulate
a given model in a way such that it has an universal structure. However,
this is only one step in the development of a model. For example, we may
know that the model (1.9) is adequate, but we do not know the value of g.
In this case, we can perform experiments, like throwing stones vertically and
taking measurements of z at different x, and aim at determining g from this
measurements. This is a parameter identification problem. Now, suppose that,
instead of throwing a stone, we launch a rocket with a steering mechanism.
This means that in addition to the gravitational force, we have the forces
 Introduction 11

exercised by the propulsion engine and by the steering thrusters. In this case,
the modelling work includes the determination of the steering action in order
to follow a given trajectory. This is a control problem.
It is the purpose of this book to illustrate these different aspects of mod-
elling with ordinary differential equations.
Another random document with
no related content on Scribd:
effectually silenced their remarks. He shortly afterwards went out,
and returned with a bowl, containing six quarts of new milk, with
which a little flour and some honey had been mixed. He told me he
was determined to see me eat the whole of it before he left the
apartment. I partook sparingly of it at first, but in the night tasted it
several times, which refreshed and invigorated me.
3d.—Finding myself considerably better this morning, I ordered a
pillow to be laid across my horse’s saddle, that I might travel for a
short distance on his back. Entered a miserable village named
Zunko, and slept there for the night.
4th.—Arose at sun-rise, and after a fatiguing journey reached
Roma at two in the afternoon. This town is built on an eminence, and
commands a delightful view of the country for miles round.
Proceeded to the house of the chief, and slept there. He made me
the accustomed presents of provisions, &c. and received in return a
pair of scissors, a hundred needles, and a paper of cloves. Several
Fellata girls came to me this evening, of a bright copper colour, and
extremely beautiful, with delicate and graceful forms. With a curiosity
so natural to their sex, they were all eager to catch a glance of the
“little Christian,” having seen the “great Christian,” as they termed my
master, before. In the course of the afternoon and evening they
brought me a little milk and butter, for which I recompensed them
with a few beads. Here, I am happy to say, the dysentery quite left
me, and my health wonderfully improved.
5th.—Left Roma for Bogell early in the morning, and arrived there
at half-past four in the afternoon. Found myself so well to-day that I
was enabled to ride on horseback the whole of the journey. As usual,
went to the residence of the chief, who gave me a little corn for the
horses only. In the night slept profoundly, and next morning was
completely recovered.
6th.—At six in the morning left for Zulamee, where we arrived at
two in the afternoon. The chief was overjoyed to see me, and
ordered four armed men to guard the beasts and property. A band of
robbers having infested the town for some time, and committed
numerous depredations, he was fearful some of my goods or cattle
would be stolen, and desired me to fire a gun morning and evening,
that the brigands might not think I was unprepared to meet them.
They did not, however, molest me; but the king of Kashna was not
quite so fortunate: he had a fine horse stolen from him yesterday,
which he never afterwards recovered. Rested here two days, in
order to recruit the spirits of my followers, and refresh the animals
under their charge.
9th.—At six in the morning, having got every thing in readiness,
left Zulamee. Both men and beasts seemed much invigorated with
the rest they had enjoyed, and at two in the afternoon reached
Gundumowah, a small but neat Fellata village. The chief sent me a
little milk.
10th.—Early in the morning started for Sansanee. The country
traversed was thickly wooded, and the path lay for three hours
through a large bush, which, having recently been visited by a horde
of elephants, the prints of whose feet were very perceptible,
rendered travelling extremely unpleasant, and even dangerous.
Reached Sansanee at one at noon. The site on which the town is
built not being long cleared, none of the houses were quite finished.
On our arrival, the chief had an open shed, occupied by fifteen
calves, cleaned out for our reception. In the evening, putting the
goods in the centre, I ordered the men to lie around them, whilst I
placed myself near the most valuable articles. Not deeming them
sufficiently secure, my sleep was rather disturbed; and awaking
about ten o’clock, I found my camel had strayed from outside the
hut, and being unwilling to arouse my drowsy companions, went
myself in search of him. On my return, to my infinite surprise and
alarm, discovered Pascoe had decamped, taking with him a valuable
gun, two pistols, a cutlass, six sovereigns, nineteen dollars, ten large
and ten small knives, and several other articles, which he had
contrived to take from the boxes in which they had been placed. To
deceive me, the artful old villain had put a pillow into a sack, which
he had laid along on his own mat. On the discovery, I immediately
made an alarm, and sent to the chief for twelve horsemen to go in
pursuit of him.
 11th.—About three o’clock in the afternoon, as I was standing in
my shed, I perceived a party of horsemen coming towards me in full
gallop. On coming within a few yards of me, they suddenly checked
their horses, and flourishing their spears over their heads,
exclaimed, “Nasarah, acqui de moogoo!” (Christian, we have the
rogue!) They informed me that a little before day-break in the
morning they heard the report of a gun, and going towards the place
whence the sound seemed to proceed, saw Pascoe perched on the
top of a high tree, and the stolen goods lying at the root of it. They
threatened to shoot him with their poisoned arrows unless he
immediately came down. This had the desired effect. He hastily
descended, and delivered himself into their hands. One of the
soldiers took the trembling scoundrel behind him on his horse, when
the whole party immediately clapped spurs to their horses, and made
all the haste they could to the village. I asked Pascoe what could
have induced him to leave me in so disgraceful a manner. He replied
that his countrymen (the Gooburites) were at war with the Fellatas,
who would cut off his head on arriving at Soccatoo. The chief coming
up at the instant, cried out, “A blessing, a blessing; you have taken
the thief, let me take off his head!” This was Pascoe’s third offence;
and I ordered him to be heavily ironed and pinioned in the town
dungeon.
12th.—Pascoe having expressed a wish to see me, I sent for him.
He begged most piteously to be forgiven, and holding up his naked
arms, which were swollen to thrice their natural size, implored me at
least not to have him pinioned again. This I freely consented to, but
he was not freed from his irons till the morning we left the town.
13th.—This day 500 camels laden with salt, obtained from the
borders of the Great Desert, arrived at the town. They were
preceded by a party of twenty Tuarick salt-merchants, whose
appearance was grand and imposing. They entered at full trot, riding
on handsome camels, some of them red and white, and others black
and white. All the party were dressed exactly alike. They wore black
cotton tobes and trousers, and white caps with black turbans, which
hid every part of the face but the nose and eyes. In their right hand
they held a long and highly-polished spear, whilst the left was
occupied in holding their shields, and retaining the reins of their
camels. The shields were made of white leather, with a piece of
silver in the centre. As they passed me, their spears glittering in the
sun, and their whole bearing bold and warlike, they had a novel and
singular effect, which delighted me. They stopped suddenly before
the residence of the chief, and all of them exclaiming, “Choir!” each
of the camels dropped on its knees, as if by instinct, whilst their
riders dismounted to pay their respects. They came in a body to see
me just after, and notwithstanding their apparent respectability, felt
not the least repugnance to beg money in a most importunate
manner. One of them, in the hope of obtaining some, described
himself as “God’s own slave.” I refused to accede to his request,
observing, that God always loved his servants, and made them
prosperous and happy, and could not believe what he had told me.
Becoming at length very troublesome, I was under the necessity of
turning him out: as he went away, he muttered something I did not
understand, and said, I was the first person that ever refused to give
him money. Like thousands of others, these merchants were very
inquisitive, and amongst other questions asked whether any of my
countrymen had tails like monkeys? I assured them none of them
had that elegant appendage, but they would not believe me. After
remaining an hour, they went to look after their wives and children,
who were on the camels on the road, and not yet arrived.
14th.—The goods being properly secured, I took my gun this
morning and shot enough of pigeons for dinner. The Tuarick men,
women, and children, surrounded me in great numbers on hearing
the report of my gun, and were amazed to see the birds falling dead
at my feet from the tops of the high trees. They examined them with
great attention, declaring I was a beautiful man, too good for a Kafer,
and ought to be a worshipper of the true faith.
16th.—In the afternoon an escort, consisting of fifty armed
horsemen sent by Bello, arrived, to conduct me to Soccatoo. They
brought with them my master’s two camels, to carry the goods,
which, as I afterwards learnt, the Gadado had borrowed under false
pretences. A messenger belonging to the sultan of Kano brought a
letter from my master, unknown to the escort, and was on his way to
Kano with it. The letter informs me of the total ignorance of my
master of my having left Kano, and his expectations of seeing me
there in a week or two. It struck me forcibly at the time that Bello
wished to get me in his power, in order to put us out of the way, to
become the sole possessor of the sheik of Bornou’s presents, &c. As
soon as I had read the letter, I asked the Gadado’s brother, who
commanded the escort, if it was the sultan’s intention to murder us
on my arrival at Soccatoo, as my master knew nothing of the
transaction. He answered, “Fear nothing, the king will not hurt you;
as he has never seen but one Christian, he wishes to view a
second.”
17th and 18th.—The Gadado’s brother came to me several times
for money to buy goora nuts. I at length told him the money in my
possession belonged to my father, who would not permit me to part
with any, unless I had previously obtained his consent. This answer
by no means pleased him: but on my telling him he should be
handsomely rewarded on our safe arrival at Soccatoo, he became a
little appeased.
19th.—Liberated old Pascoe in the morning, who had behaved
peaceably since his imprisonment, and seemed truly sorry for the
offence he had committed; and at two in the afternoon went out of
the town, but was obliged to leave a bullock behind, being lamed and
unable to proceed. Travelled till eleven o’clock the next night, when
we arrived much fatigued at Magaria. The poor camels and horses
could hardly stand, and suffered dreadfully from thirst, not having
drank during this long journey.
21st and 22d.—The horses and camels not being sufficiently
recovered to continue the journey, staid at Magaria both these days.
Resided at a house belonging to the Gadado, who supplied me with
abundance of provisions. Received a message from him to come to
see him, with my gun, in order to show the head men of the town the
manner in which birds were shot in my country. I soon gratified their
curiosity by firing at a small bird, at a distance of fifty yards. The
whole of them testified the greatest astonishment on taking it up, and
would not for a long time believe it was really dead.
 23d.—To my great joy I entered the gates of Soccatoo in the
afternoon, about two o’clock, after a tedious and wearisome journey
of nearly a month. Not having seen my master for three months, I
hastened to his house; but not finding him at home, I went to the
Gadado’s, where, I was told, he had gone. My master was in earnest
conversation with the Gadado and an old Arab, and was much
surprised when informed of the reason of my leaving Kano: he spoke
with warmth of the artful and unhandsome conduct of the sultan; and
after this act of duplicity on the part of Bello, to the hour of his death,
I never observed him to smile. My master had been ill of dysentery
before my arrival, but was then much better.

RESIDENCE AT SOCCATOO — MY MASTER’S DEATH — BURIAL.

On the 13th December, the day after my arrival, Sultan Bello sent
for my master and myself to repair to his residence. As soon as we
entered, he began to make inquiries of the nature of the presents I
had with me, and was extremely desirous to know if I had left any
with Hadji Ben Sallah for the sheik of Bornou. I replied, I had not.
“Are you sure you have not?” said he. I again answered with
firmness in the negative. The sultan then demanded the king of
England’s letters to the sheik of Bornou, which my master reluctantly
produced; but refused to accede to the sultan’s request to open and
read them, observing that, when his king discovered, on his return to
his country, he had so unfaithfully broken his trust, he would
immediately be beheaded. The sultan himself took the letters, and
waving his hand for us to withdraw, we left the apartment. We had
not been in our hut more than a couple of hours when the Gadado,
his brother, Hadji Ben Sallah, and several of the principal inhabitants
of Soccatoo, entered, and demanded, in the name of Bello, the
presents intended for the sheik of Bornou, together with all the arms
and ammunition we did not want ourselves. My master became
deeply agitated when he had heard their errand, and rising up from
his couch, exclaimed with much energy and bitterness—“There is no
faith in any of you; you are an unjust people; you are worse than
highway robbers.” They cautioned him to be more guarded in his
expressions, or it might cost him his head. “If I lose my head,”
rejoined my master in the same determined manner, “I lose it for
speaking for the just rights of my country only.” Ben Sallah and
others entreated him to moderate his anger, or it might indeed be
fatal to him. I also implored him to accede to the demands of the
sultan, how unjust and tyrannic soever they might be, observing, that
two debilitated white men stood no chance in holding out against the
united force of so many Fellatas, who only waited the command of
their sovereign to assassinate them. After much entreaty, and not
without considerable reluctance, my master desired that they should
be given to them; and said to the Gadado, on his leaving the hut,
“Tell your sovereign I never wish to see him again; my business with
him is now at an end.” A short time after Mallam Mudey returned with
a message from the sultan, acquainting my master of his intention of
writing to the king of England in explanation of his conduct. He
desired Mallam Mudey to tell Bello that the king of England would
not even look at a letter from him, after the treatment his subjects
had received.
I took an opportunity one day of acquainting my master of
Pascoe’s villany, who immediately dismissed him, without paying his
wages. The old man went to a native lawyer to obtain advice in what
manner he was to act; but instead of holding out the hope of
obtaining the sum due to him, the learned Fellata expressed his
astonishment that Abdullah had not cut off his head. Pascoe then
turned snuff-merchant; but having given too extensive credit, soon
became a bankrupt, and was finally obliged to cut wood from the
adjoining country, and sell it in the market at Soccatoo, by which
means he contrived to procure a precarious subsistence.
One day the Gadado came and begged me to lend him my camel
to go to the war against the sheik of Bornou: knowing it would be
impolitic to refuse, I let him have it. The animal was returned to me,
about six weeks after, in a most shameful state, reduced to a mere
skeleton, and having two immense holes in its back. I wished the
Gadado to give me another in its stead, but he would not listen to
me.
My master and myself enjoyed pretty good health for some time
after my arrival at Soccatoo, and amused ourselves with going a
shooting almost every day. At one shot my master brought down
thirteen wild ducks, about two miles to the north-east of the city, ten
of which were secured. We remained at Soccatoo much longer than
was intended, believing the sultan would consent to our proceeding
to Bornou when the war had become somewhat abated; but this he
never granted.
On the 12th of March, 1827, I was greatly alarmed on finding my
dear master attacked with dysentery. He had been complaining a
day or two previously of a burning heat in his stomach,
unaccompanied, however, by any other kind of pain. From the
moment he was taken ill he perspired freely, and big drops of sweat
were continually rolling over every part of his body, which weakened
him exceedingly. It being the fast of Rhamadan, I could get no one,
not even our own servants, to render me the least assistance. I
washed the clothes, which was an arduous employment, and obliged
to be done eight or nine times each day, lit and kept in the fire, and
prepared the victuals myself; and the intermediate time was
occupied in fanning my poor master, which was also a tedious
employment. Finding myself unable to pay proper attention to his
wants in these various avocations, I sent to Mallam Mudey, on the
13th, entreating him to send me a female slave to perform the
operation of fanning. On her arrival I gave her a few beads, and she
immediately began her work with spirit; but she soon relaxed in her
exertions, and becoming tired, ran away, on pretence of going out for
a minute, and never returned. Alla Sellakee, a young man my master
had purchased on the road from Kano to take care of the camels,
and whom he had invariably treated with his usual kindness, and
given him his freedom, no sooner was made acquainted with his
master’s illness than he became careless and idle, and instead of
leading the camels to the rich pasturage in the vicinity of Soccatoo,
let them stray wherever they pleased, whilst he himself either
loitered about the city, or mixed with the most degraded people in it:
by this means the camels became quite lean; and being informed of
the reason, I told my master, who instantly discharged him from his
service.
 My master grew weaker daily, and the weather was insufferably
hot, the thermometer being, in the coolest place, 107 at twelve in the
morning, and 109 at three in the afternoon. At his own suggestion I
made a couch for him outside the hut, in the shade, and placed a
mat for myself by its side. For five successive days I took him in my
arms from his bed in the hut to the couch outside, and back again at
sunset, after which time he was too much debilitated to be lifted from
the bed on which he lay. He attempted to write once, and but once,
during his illness; but before paper and ink could be brought him, he
had sunk back on his pillow, completely exhausted by his ineffectual
attempt to sit up in his bed. Fancying by various symptoms he had
been poisoned, I asked him one day whether he thought that, in any
of his visits to the Arabs or Tuaricks, any poisonous ingredients had
been put into the camel’s milk they had given him, of which he was
particularly fond. He replied, “No, my dear boy; no such thing has
been done, I assure you. Do you remember,” he continued, “that
when on a shooting excursion at Magaria, in the early part of
February, after walking the whole of the day, exposed to the
scorching rays of the sun, I was fatigued, and lay down under the
branches of a tree for some time? The earth was soft and wet, and
from that hour to the present I have not been free from cold: this has
brought on my present disorder, from which, I believe, I shall never
recover.”
For twenty days my poor master remained in a low and distressed
state. He told me he felt no pain; but this was spoken only to comfort
me, for he saw I was dispirited. His sufferings must have been acute.
During this time he was gradually, but perceptibly, declining; his
body, from being robust and vigorous, became weak and emaciated,
and indeed was little better than a skeleton. I was the only person,
with one exception, he saw in his sickness. Abderachman, an Arab
from Fezzan, came to him one day, and wished to pray with him,
after the manner of his countrymen, but was desired to leave the
apartment instantly. His sleep was uniformly short and disturbed, and
troubled with frightful dreams. In them he frequently reproached the
Arabs aloud with much bitterness; but being an utter stranger to the
language, I did not understand the tenor of his remarks. I read to him
daily some portions of the New Testament, and the ninety-fifth
Psalm, which he was never weary of listening to, and on Sundays
added the church service, to which he invariably paid the
profoundest attention. The constant agitation of mind and exertions
of body I had myself undergone for so long a time, never having in a
single instance slept out of my clothes, weakened me exceedingly,
and a fever came on not long before my master’s death, which hung
upon me for fifteen days, and ultimately brought me to the very verge
of the grave. Finding myself unequal to pay that attention to my
master’s wants which his situation so particularly required, I solicited
and obtained his consent to have old Pascoe once more to assist
me. On entering the hut, he fell on his knees, and prayed to be
forgiven, promising to be faithful to my master’s service. Master
immediately pardoned him, and said he would forget all that had
passed, if he conducted himself well: by this means the washing and
all the drudgery was taken from my shoulders, and I was enabled to
devote all my time and attention to my master’s person. I fanned him
for hours together, and this seemed to cool the burning heat of his
body, of which he repeatedly complained. Almost the whole of his
conversation turned upon his country and friends, but I never heard
him regret his leaving them; indeed he was patient and resigned to
the last, and a murmur of disappointment never escaped his lips.
On the 1st of April, he became considerably worse, and though
evidently in want of repose, his sleep became more and more
disturbed. He swallowed eight drops of laudanum, four times a day,
for three days; but finding it did him not the least benefit, he
discontinued taking it altogether: this, with the exception of two
papers of Seidlitz powders and four ounces of Epsom salts, was the
only medicine he had during his illness. On the 9th, Maddie, a native
of Bornou, whom master had retained in his service, brought him
about twelve ounces of green bark from the butter tree, and said it
would do him much good. Notwithstanding all my remonstrances,
master immediately ordered a decoction of it to be prepared,
observing, “No man will injure me.” Accordingly Maddie himself
boiled two basins-full, the whole of which he drank in less than an
hour. Next morning he was much altered for the worse, and regretted
his not having followed my advice. About twelve o’clock of the same
day, he said, “Richard, I shall shortly be no more; I feel myself
dying.” Almost choked with grief, I replied, “God forbid, my dear
master: you will live many years yet.” “Don’t be so much affected, my
dear boy, I entreat you,” said he: “it is the will of the Almighty; it
cannot be helped. Take care of my journal and papers after my
death; and when you arrive in London, go immediately to my agents,
send for my uncle, who will accompany you to the Colonial Office,
and let him see you deposit them safely into the hands of the
secretary. After I am buried, apply to Bello, and borrow money to
purchase camels and provisions for your journey over the desert,
and go in the train of the Arab merchants to Fezzan. On your arrival
there, should your money be exhausted, send a messenger to Mr.
Warrington, our consul at Tripoli, and wait till he returns with a
remittance. On reaching Tripoli, that gentleman will advance what
money you may require, and send you to England the first
opportunity. Do not lumber yourself with my books; leave them
behind, as well as the barometer, boxes, and sticks, and indeed
every heavy article you can conveniently part with; give them to
Malam Mudey, who will take care of them. The wages I agreed to
give you my agents will pay, as well as the sum government allowed
me for a servant; you will of course receive it, as Columbus has
never served me. Remark what towns or villages you pass through;
pay attention to whatever the chiefs may say to you, and put it on
paper. The little money I have, and all my clothes, I leave you: sell
the latter, and put what you may receive for them into your pocket;
and if, on your journey, you should be obliged to expend it,
government will repay you on your return.” I said, as well as my
agitation would permit me, “If it be the will of God to take you, you
may rely on my faithfully performing, as far as I am able, all that you
have desired; but I trust the Almighty will spare you, and you will yet
live to see your country.” “I thought I should at one time, Richard,”
continued he; “but all is now over; I shall not be long for this world:
but God’s will be done.” He then took my hand betwixt his, and
looking me full in the face, while a tear stood glistening in his eye,
said, in a low but deeply affecting tone, “My dear Richard, if you had
not been with me, I should have died long ago; I can only thank you,
with my latest breath, for your kindness and attachment to me, and if
I could have lived to return with you, you should have been placed
beyond the reach of want; but God will reward you.” This
conversation occupied nearly two hours, in the course of which my
master fainted several times, and was distressed beyond measure.
The same evening he fell into a slumber, from which he awoke in
much perturbation, and said he had heard with much distinctness the
tolling of an English funeral bell: I entreated him to be composed,
and observed that sick people frequently fancy they hear and see
things which can possibly have no existence. He made no reply.
About six o’clock in the morning of the 11th, on asking how he did,
my master answered he was much better, and requested me to
shave him. He had not sufficient strength to lift his head from the
pillow; and after finishing one side of the face, I was obliged to turn
his head, in order to shave the other. As soon as it was done, he
desired me to fetch him a looking-glass which hung on the other side
of the hut. On seeing himself in it, he observed that he looked quite
as ill at Bornou, on his former journey; and as he had borne his
disorder so long a time, he might yet recover. On the following day
he still fancied himself getting better. I began to flatter myself, also,
that he was considerably improved. He eat a bit of hashed guinea-
fowl in the day, which he had not done before since his illness,
deriving his sole sustenance from a little fowl-soup and milk and
water. On the morning of the 13th, however, being awake, I was
much alarmed by a peculiar rattling noise, proceeding from my
master’s throat, and his breathing was loud and difficult; at the same
instant he called out “Richard!” in a low and hurried tone. I was
immediately at his side, and was astonished at seeing him sitting
upright in his bed, and staring wildly around. I held him in my arms,
and placing his head gently on my left shoulder, gazed a moment on
his pale and altered features: some indistinct expressions quivered
on his lips; he strove, but ineffectually, to give them utterance, and
expired without a struggle or a sigh. When I found my poor master
so very ill, I called out with all my strength, “O God, my master is
dying!” which brought Pascoe and Mudey into the apartment. Shortly
after the breath had left his body, I desired Pascoe to fetch some
water, with which I washed the corpse. I then got Pascoe and Mudey
to assist me in taking it outside of the hut, laid it on a clean mat, and
wrapped it in a sheet and blanket. Leaving it in this state two hours, I
put a large clean mat over the whole, and sent a messenger to
Sultan Bello, to acquaint him of the mournful event, and ask his
permission to bury the body after the manner of my own country, and
also to know in what particular place his remains were to be interred.
The messenger soon returned with the sultan’s consent to the former
part of my request; and about 12 o’clock at noon of the same day a
person came into my hut, accompanied by four slaves, sent by Bello
to dig the grave. I was desired to follow them with the corpse.
Accordingly I saddled my camel, and putting the body on its back,
and throwing an union-jack over it, I bade them proceed. Travelling
at a slow pace, we halted at Jungavie, a small village, built on a
rising ground, about five miles to the south-east of Soccatoo. The
body was then taken from the camel’s back, and placed in a shed,
whilst the slaves were digging the grave; which being quickly done, it
was conveyed close to it. I then opened a prayer-book, and, amid
showers of tears, read the funeral service over the remains of my
valued master. Not a single person listened to this peculiarly
distressing ceremony, the slaves being at some distance, quarrelling
and making a most indecent noise the whole of the time it lasted.
This being done, the union-jack was taken off, and the body was
slowly lowered into the earth, and I wept bitterly as I gazed for the
last time upon all that remained of my generous and intrepid master.
The pit was speedily filled, and I returned to the village about thirty
yards to the east of the grave, and giving the most respectable
inhabitants, both male and female, a few trifling presents, entreated
them to let no one disturb its sacred contents. I also gave them
2,000 cowries to build a house, four feet high, over the spot, which
they promised to do. I then returned, disconsolate and oppressed, to
my solitary habitation, and leaning my head on my hand, could not
help being deeply affected with my lonesome and dangerous
situation; a hundred and fifteen days’ journey from the sea-coast,
surrounded by a selfish and cruel race of strangers, my only friend
and protector mouldering in his grave, and myself suffering
dreadfully from fever. I felt, indeed, as if I stood alone in the world,
and earnestly wished I had been laid by the side of my dear master:
all the trying evils I had endured never affected me half so much as
the bitter reflections of that distressing period. After a sleepless
night, I went alone to the grave, and found that nothing had been
done, nor did there seem the least inclination on the part of the
inhabitants of the village to perform their agreement. Knowing it
would be useless to remonstrate with them, I hired two slaves at
Soccatoo the next day, who went immediately to work, and the
house over the grave was finished on the 15th.
One instance, out of many, of the kindness and affection with
which my departed master uniformly treated me, occurred at Jenna,
on our journey into the interior. I was dangerously ill with fever in that
place, when he generously gave up his own bed to me, and slept
himself on my mat, watched over me with parental assiduity and
tenderness, and ministered to all my wants. No one can express the
joy he felt on my recovery: and who, possessing a spark of gratitude,
could help returning it but by the most inviolable attachment and
devoted zeal? It was his sympathy for me in all my sufferings that
had so powerful a claim on my feelings and affections, and taught
me to be grateful to him in hours of darkness and distress, when
pecuniary recompense was entirely out of the question.
The great sufferings, both mental and bodily, I had undergone at
the death and burial of my master, and the constant agitation in
which I was kept, occasioned a rapid increase in my disorder; and on
the 16th I could with difficulty crawl round my hut, and was obliged to
lay myself on my mat, from which I had not strength to arise till the
27th; old Pascoe, during that period, being very kind and attentive to
me. The Arabs in the city visited me daily, and did all in their power
to raise my spirits; telling me not to be disheartened at the death of
my father, and that no injury would happen to me. But I plainly saw
these visits of condolence did not proceed from a charitable spirit;
they came more for the purpose of obtaining presents than any thing
else: but I did not give them any. The sultan also sent messengers to
inquire after my health nearly as often. The weather was dreadfully
warm; and I was obliged to have a tub of water close at my side, into
which I frequently plunged my hands and arms, and occasionally
sprinkled my head and body. This much refreshed me; indeed it was
the only means by which I was enabled to obtain a little sleep. I had
given up all hopes of life, when on the 26th I found my health
improve in a wonderful manner; the next day I was able to sit up on
my mat. In the course of this day (27th) the Gadado, Malem Moodie,
and Sidi Sheik, came with a commission from the sultan to search
my boxes, as he had been informed they were filled with gold and
silver; but, to their great amazement, found I had not sufficient
money to defray my expenses to the sea-coast. They, however, took
an inventory of all my articles, and carried it to Bello. The gold watch
intended for him, and the private watches of Captains Clapperton
and Pearce, I had taken the precaution to conceal about my person.
In a short time the Gadado and his companions returned with a
message from the sultan, commanding me to deliver to them the
following articles, viz. a rifle-gun, double-barrelled ditto, two bags of
ball, a canister of powder, a bag of flints, a ream and a half of paper,
and six gilt chains, for which he promised to give me whatever I
might ask. I consequently charged him 245,000 cowries, which I was
to receive from Hadji Hat Sallah, at Kano; and an order was given
me to receive this sum, and what more I might require in my journey
over the Great Desert. A letter was also sent by me to Hadji Hat
Sallah.
On the 28th I made Ben Gumso a present of four yards of blue,
and the same quantity of scarlet damask, an unwritten journal-book,
two pairs of scissors, and two knives: with these articles I
endeavoured to get into the good graces of this old Arab. By a
singular piece of good fortune he had just begun to exercise a
powerful influence over the mind and opinions of the sultan. Bello, in
an excursion into the Gooberite country, had come to an
engagement with a large party of the natives; and in the midst of the
fight was shot in the neck with a poisoned arrow, which turned the
tide of victory in favour of his enemies. On his return to Soccatoo,
Ben Gumso wrote a charm on a bit of wood, which was washed off
into a calabash of water, and drank immediately by the sultan; who,
shortly after recovering, attributed it solely to the virtues of this
charm, and he was advanced accordingly. I begged Ben Gumso to
use his influence with the sultan to obtain leave for my departure
from his capital, and make the best of my way homewards. He
accordingly represented to Bello the impolicy and injustice of
detaining any longer a subject of the king of England; advised him to
allow me to quit Soccatoo as soon as possible; and insinuated that, if
I were to die in his dominions, a report would be circulated and
believed that he had murdered both my father and me, by which he
would get a bad name. The sultan approved of these weighty
arguments of Ben Gumso; and word was sent me, almost
immediately, to appear before him. After a little preliminary
conversation, Bello asked me which route I should prefer. Although
my master had advised me to proceed with the Arabs to Fezzan, just
before his death, I much feared that the papers intrusted to my care
would be stolen, and myself murdered, by that wily and treacherous
race, whose behaviour to my master, from the time of his arrival in
Houssa, I very much disliked; and would rather cast myself, unarmed
and unprotected, upon the good faith of the natives, than go with
them. Under these impressions I answered the sultan that, as I
wished to get to England in as short a time as possible, the route to
Kubbi, through Boussa, was most likely to answer that end. “It is
impossible,” he continued, “to travel that way: the rainy season is
commenced; the rivers are overflowed; the country is inundated; and
you will not be able to reach the sea-coast in safety. It will be much
better for you to go over the desert; and, to facilitate your progress, I
will write Hat Sallah to get a trustworthy person to accompany you;
he will also furnish you with camels and provisions, and advance you
what money it is likely you will want.” I only replied, “Very well,
sultan.” He then asked if Abdullah had forgiven Pascoe in his book,
for the roguery he had committed. I said, “He had not been able to
write during the whole of his sickness; and therefore nothing was
stated about the matter.”—“If Abdullah has not pardoned him in his
book,” rejoined the sultan, “your king will certainly cut off his head on
his arrival in England.” I assured him Pascoe would not be punished
by any one, if his future conduct was good; but this the sultan was in
no haste to believe; and observed, “I cannot suffer him to go with
you; he shall stay here to clean and repair my guns!” this latter
consideration having evidently more influence with the sultan than
Pascoe’s safety, which he cared nothing at all about. I then besought
Bello to permit him to accompany me as far as Kano, as an
interpreter; to which he rather reluctantly consented, on condition
that I should procure him a horse to return, and pay him wages, on
my arrival at Kano, to the amount of 15,000 cowries, which of course
I agreed to do; and finding the sultan had nothing more to say, I
bowed profoundly and retired. I never saw him again.

FROM SOCCATOO TO DUNRORA.

On the evening of the 3d of May, a messenger came from the

sultan, and told me to get every thing in readiness to depart on the
following morning, with a promise of a camel and provisions, which I
never received. Early next day, therefore, I left Soccatoo, where I
had suffered so much, heartily tired of the place and its inhabitants;
and, accompanied with a messenger from the kind old Gadado,
Pascoe and Mudey, with our three camels and two horses,
proceeded to a flat, five miles to the east of Magaria, where we
arrived in the afternoon, and rested for the night under the branches
of a large tree, near to a small lake. Mosquitoes were numerous and
troublesome, and consequently could not sleep till morning, when a
refreshing breeze springing up, it drove them away. At this flat we
joined a party of above 4,000 people, consisting of Tuarick salt-
merchants returning to Kilgris, pilgrims on their way to Mecca, Goora
merchants returning to Kano and Nyffé, &c. &c.; all travelling in
company, for mutual protection, with an immense number of camels,
horses, and bullocks. The merchants invariably meet at Kashna,
where they disperse for their different destinations. In the same train
was the king of Jacoba, with fifty slaves, which he had driven to
Soccatoo, as a present to the sultan, who, having learnt the dreadful
losses he had sustained in men and cattle in his wars with the sheik
of Bornou, and the number of his villages which had been plundered
and burnt by the soldiers of the sheik, would not accept of them, and
desired the king of Jacoba to re-conduct them to his own dominions.
At eleven o’clock in the morning of the 4th of May, a signal to
prepare to depart was made with the horns and drums of the party,
which made a loud and most discordant noise; and, about an hour
after, the whole body was in motion. We travelled in great haste till
three o’clock in the afternoon of the 5th, when Boussa Jack, the
horse on which I rode, and which was made a present of by the king
of Boussa to my late master, became much fatigued, and began to
lag. The weather was at this time intolerably hot, and the dust was
rolling in thick clouds in every direction, entering my eyes and
nostrils, and penetrating into the very pores of the skin. I felt nearly
suffocated, and was faint and exhausted. Finding I was unable to
proceed, I ordered Pascoe to overtake the camels, his horse being
fresh and vigorous, and bring me some water. I then dismounted,
and sat under a tree by the road-side, whose branches afforded but
an indifferent shelter against the scorching rays of an African sun,
and holding the bridle of my poor horse in my hand, I implored the
hundreds of Fellatas and Tuaricks who were passing to sell me a
drop of water; but the cold-hearted wretches refused my earnest
request, observing one to another, “He is a Kafir; let him die.” At
length a young Fellata, from Footatoora, accidentally seeing me,
came to the spot, exclaiming, “Nasarah, Nasarah, triffi manora!”
(Christian, Christian, go on!) I answered, “I am faint and sick for want
of water; no one will give me any; and I am so weary that I cannot
proceed.” On hearing which the young man kindly gave me a small
calabash full; part of which I drank, and with the remainder washed
the nostrils of Boussa Jack, and sprinkled a little into his mouth. The
people, who observed the Fellata performing this generous action,
upbraided him in strong language for giving water to the Christian;
but he, showing them a double-barrelled gun, remarked that he had
obtained it of my countrymen, who were all good men, and would do
no harm. This somewhat appeased them. On examining the gun
shortly afterwards, I found it to be of English manufacture, with
“Arnold, maker, London,” on its lock. I, as well as the horse, was
greatly refreshed with the small quantity of water I had taken; but
soon becoming again weak and dispirited, I was almost in as bad a
state as on the former occasion; my legs were swollen prodigiously,
and I felt the most acute pains in every part of my body. At length I
perceived Pascoe, whom I had sent for water three or four hours
previously, comfortably seated under a tree, and seeming to be
enjoying himself much with Mudey,—the camels feeding at a short
distance. I had half an inclination to shoot the heartless old
scoundrel, knowing as he did how keen my sufferings must have
been. Reflecting, however, that the safety of my papers, and even
my own life, was placed in some measure in his hands, I restrained
myself, and merely asked why he did not return with the water, on
which he answered, very composedly, “I was tired.”
The young Fellata who had so generously saved my life came to
me on the 7th, and informed me that the whole of the slaves of the
king of Jacoba being missing, a party of horsemen had been sent in
quest of them, and had just returned with the dreadful account of
having seen thirty-five of their dead bodies exposed on the road; the
remaining fifteen could not be found, but were strongly suspected of
having met a similar fate. These unfortunate creatures had to carry
heavy burdens on their heads the day before; and being unable to
keep up with the rapid pace of the camels, were necessarily obliged
to be left behind, and thus miserably perished of thirst. I
congratulated myself on my own good fortune in having so narrowly
escaped so horrid a death; and thanked the Almighty for having so
providentially rescued me. On leaving me, I gave the Fellata a pair of
scissors, and twenty flints for his gun, which pleased him highly. On
our road to Kano, the king of Jacoba became very sociable with me,
and was my constant companion. He pressed me very much to visit
his country, where he would do all in his power to make my stay
agreeable. He told me that his neighbours, the Yamyam people, who
had assisted him in his war against the sheik of Bornou, were
surrounded, with some of his own people, on a plain near Jacoba, by
the sheik’s soldiers, who made a dreadful slaughter of them. The
fight lasted a whole day, when the Yamyams and people of Jacoba
were entirely routed; he himself narrowly escaping being taken
prisoner. The morning after, the surviving Yamyams repaired to the
field of action, and bearing off a great number of the dead bodies of
their enemies, made a fire, roasted, and ate them!
Soon after our arrival at Markee, on the 19th, the spouse of the
chief, a finely formed and intelligent looking woman, aged about 22
years, came to me with tears in her eyes, and implored me to give
her some magarie, as she had no child. I accordingly presented her
with a couple of tea-spoonfuls of oil of cinnamon, and ordered her to
put two drops of it into a pint of cow’s milk, which should be taken
three times a day till the whole was consumed. I told her that on my
return to Markee I had not the least doubt I should have the pleasure
of seeing her the happy mother of a numerous progeny. On her
husband being made acquainted with the circumstance, he came
and thanked me in the heartiest manner for my kindness, gave me
abundance of fresh milk, fowls, rice, &c.; telling me at the same time,
that on my return from England he would give me a large sum of
money, which, no doubt, he will do. Bidding them adieu the next
morning, they said to me, “Christian! God send you safely to your
country, and may you speedily come back to Markee.”
On my arrival at Kano on the 25th of May, I delivered Sultan
Bello’s order and letter to Hadji Hat Sallah; but he declared, after
perusing the letter, he would have nothing to do with it: he had nearly
lost his head on my father’s account at Soccatoo, and therefore
positively refused to lend me a single cowrie. He had no objection to
give, in goods and a slave, the amount of my demands on Bello, but
refused to let me have any money. I accordingly received from him a
strong female slave, a quantity of unwrought silk, and scarlet caps
and beads. I sold my weakened and diseased camels for 15,000
cowries each, and discharged their keeper, Maddie. Not having
sufficient money to purchase fresh camels, provisions, and presents
for the chiefs on my way to Fezzan, I was necessitated to take a
different route; and for that purpose purchased a horse and two
asses. Bello had not forgotten to mention in his letter for Hadji Hat
Sallah to send back Pascoe to Soccatoo, on his arrival at Kano. I
begged of him, however, to let me take him as far as Coulfo, in
Nyffé, which he at first refused; but giving him two yards of scarlet
damask, two pairs of scissors, two dollars, a large knife, and two
yards of blue damask, the old Arab seemed to relax in his
determination; and calling Pascoe to him, gave him leave to
accompany me to Coulfo, but cautioned him to return the moment he
arrived, on pain of having Jerrub (the devil) sent after him. The old
man came to me just after in great trepidation, every member of his
body trembling like an aspen leaf, and gave me to understand, as
well as he was able, the conversation he had had with Hadji Hat
Sallah, and in consequence thereof, his great reluctance to quit
Kano with me. I did all I could to laugh him out of his superstitious
fears, but it utterly failed of the desired effect; the name of Jerrub
being enough to put him in a fever. I had given the slave that Hadji