Existence Theory for Generalized

Newtonian Fluids 1st Edition Breit

Visit to download the full and correct content document:
https://textbookfull.com/product/existence-theory-for-generalized-newtonian-fluids-1st
-edition-breit/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Heat Transfer to Non Newtonian Fluids Fundamentals and

Analytical Expressions 1st Edition Aroon Shenoy

https://textbookfull.com/product/heat-transfer-to-non-newtonian-
fluids-fundamentals-and-analytical-expressions-1st-edition-aroon-
shenoy/

Mathematical Theory of Compressible Viscous Fluids

Analysis and Numerics 1st Edition Eduard Feireisl

https://textbookfull.com/product/mathematical-theory-of-
compressible-viscous-fluids-analysis-and-numerics-1st-edition-
eduard-feireisl/

Generalized Intuitionistic Multiplicative Fuzzy

Calculus Theory and Applications Shan Yu

https://textbookfull.com/product/generalized-intuitionistic-
multiplicative-fuzzy-calculus-theory-and-applications-shan-yu/

Generalized Locally Toeplitz Sequences Theory and

Applications Volume II Carlo Garoni

https://textbookfull.com/product/generalized-locally-toeplitz-
sequences-theory-and-applications-volume-ii-carlo-garoni/
Theory and Modeling of Rotating Fluids Convection
Inertial Waves and Precession 1st Edition Keke Zhang

https://textbookfull.com/product/theory-and-modeling-of-rotating-
fluids-convection-inertial-waves-and-precession-1st-edition-keke-
zhang/

Non-cooperative Stochastic Differential Game Theory of

Generalized Markov Jump Linear Systems 1st Edition
Cheng-Ke Zhang

https://textbookfull.com/product/non-cooperative-stochastic-
differential-game-theory-of-generalized-markov-jump-linear-
systems-1st-edition-cheng-ke-zhang/

Necessary Existence Alexander R. Pruss

https://textbookfull.com/product/necessary-existence-alexander-r-
pruss/

Non Newtonian Fluid Mechanics and Complex Flows Angiolo

Farina

https://textbookfull.com/product/non-newtonian-fluid-mechanics-
and-complex-flows-angiolo-farina/

Rheology of Drag Reducing Fluids 1st Edition Aroon

Shenoy

https://textbookfull.com/product/rheology-of-drag-reducing-
fluids-1st-edition-aroon-shenoy/
EXISTENCE
THEORY FOR
GENERALIZED
NEWTONIAN
FLUIDS
Mathematics in Science and
Engineering

EXISTENCE
THEORY FOR
GENERALIZED
NEWTONIAN
FLUIDS

Dominic Breit
Mathematical & Computer Sciences; Mathematics
Heriot-Watt University
Edinburgh, Midlothian, United Kingdom

Series Editor
Goong Chen

AMSTERDAM • BOSTON • HEIDELBERG • LONDON

NEW YORK • OXFORD • PARIS • SAN DIEGO
SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO
Academic Press is an imprint of Elsevier
Academic Press is an imprint of Elsevier
125 London Wall, London EC2Y 5AS, United Kingdom
525 B Street, Suite 1800, San Diego, CA 92101-4495, United States
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom
Copyright © 2017 Elsevier Inc. All rights reserved.
No part of this publication may be reproduced or transmitted in any form or by any means,
electronic or mechanical, including photocopying, recording, or any information storage and
retrieval system, without permission in writing from the publisher. Details on how to seek
permission, further information about the Publisher’s permissions policies and our arrangements
with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency,
can be found at our website: www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright by the
Publisher (other than as may be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and experience
broaden our understanding, changes in research methods, professional practices, or medical
treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in evaluating
and using any information, methods, compounds, or experiments described herein. In using such
information or methods they should be mindful of their own safety and the safety of others,
including parties for whom they have a professional responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors,
assume any liability for any injury and/or damage to persons or property as a matter of products
liability, negligence or otherwise, or from any use or operation of any methods, products,
instructions, or ideas contained in the material herein.

Library of Congress Cataloging-in-Publication Data

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

British Library Cataloguing-in-Publication Data

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

ISBN: 978-0-12-811044-7

For information on all Academic Press publications

visit our website at https://www.elsevier.com

Publisher: Nikki Levy

Acquisition Editor: Graham Nisbet
Editorial Project Manager: Susan Ikeda
Production Project Manager: Paul Prasad Chandramohan
Designer: Maria Ines Cruz
Typeset by VTeX
Dedicated
to
Janine
PREFACE

In continuum mechanics, the motion of an incompressible, homo-

geneous fluid is described by the velocity field and the hydrodynamical
pressure. The time evolution of the fluid is governed by the Navier–
Stokes system of partial differential equations which describes the balance
of mass and momentum. In the classical formulation – which goes back
to C.-L. Navier and G.G. Stokes – the relation between the viscous stress
tensor and the symmetric gradient of the velocity field is linear (i.e. we
have a Newtonian fluid). This system is already quite challenging from a
mathematical point of view and has fascinated many mathematicians. How-
ever, it can only model fluids with a very simple molecular structure such
as water, air and several oils. In order to study more complex fluids one has
to deal with a generalized Navier–Stokes system. In this model for gen-
eralized Newtonian fluids, the viscosity ν is assumed to be a function of
the shear-rate γ̇ . Very popular among rheologists is the power-law model
in which the generalized viscosity function is of power-type ν ∼ γ̇ p−2 with
p > 1. For a specific fluid physicists can identify this power by experiments.
Instead of the Laplacian, as in the classical Navier–Stokes system, the main
part of the mathematical model is a power-type nonlinear second order dif-
ferential operator. So, in addition to the convective term, a second highly
nonlinear term appears. The only framework which is available today is the
concept of weak solutions. These solutions belong to appropriate Sobolev
spaces: Derivatives have to be understood in the sense of distributions and
singularities may occur.
The mathematical observation of power-law fluids began in the late
sixties in the pioneering work of J.-L. Lions and O.A. Ladyzhenskaya. The
cases they could handle were restrictive for real world situations. Never-
theless, it was a breakthrough in the theory of partial differential equations.
Since then, there has been huge progress in the mathematical theory of
generalized Newtonian fluids. The first systematic study was initiated by
the group around J. Nečas in the 1990s. Situations which are realistic for
physical and industrial applications could finally be handled. In the 2000s
the Lipschitz truncation (the approximation of a Sobolev function by a Lip-
schitz continuous one done in such a way that both are equal outside of a
small set whose size can be controlled) was shown to be a powerful tool
in the analysis of generalized Newtonian fluids. A very wide range of non-
Newtonian fluids could finally be included in the mathematical theory. The

ix
x Preface

best known bound p > 65 for three dimensional flows of power-law fluids
was achieved.
A drawback of the Lipschitz truncation is its nonlinear and nonlo-
cal character. In fact, the property of a function to be solenoidal (that
is, divergence-free) is lost by truncating it. So, one has to introduce the
pressure function in the weak formulation which results in additional tech-
nical difficulties. An advanced pressure decomposition via singular integral
operators is necessary in the non-stationary case. An improved version, the
“solenoidal Lipschitz truncation”, was developed only recently. It allows the
existence of weak solutions to the generalized Navier–Stokes system to be
shown without the appearance of the pressure function and therefore highly
simplifies the proofs. Moreover, it allows the Prandtl–Eyring fluid model to
be studied which was out of reach before. In this model the power-growth
is replaced by some logarithmic function: the law ν ∼ ln(1 + γ̇ )/γ̇ was in-
troduced in 1936 based on a molecular theory. This leads to a limit case in
the functional analytical setting of generalized Newtonian fluids. It is not
possible to introduce the pressure in the expected function space. Neither
can the divergence be corrected.
The aim of this book is to present a complete and rigorous mathe-
matical existence theory for generalized Newtonian fluids – for stationary,
non-stationary and stochastic models. The balance laws are formulated in
all situations via a generalized Navier–Stokes system. The proofs are pre-
sented as self-contained as possible and require from the reader only basic
knowledge of nonlinear partial differential equations.
The heart of this book is the construction of the “solenoidal Lipschitz
truncation”. It has numerous applications and is of interest for future re-
search beyond the scope of this monograph. The stationary truncation is
presented in Chapter 3 and the non-stationary version in Chapter 6. Based
on the “solenoidal Lipschitz truncation” the existence of weak solutions to
generalized Navier–Stokes equations is shown. The existence proof itself is
only slightly more complicated than the classical monotone operator theory
and easy to follow.
In Chapter 4 we study the stationary Prandtl–Eyring model in two di-
mensions. Here, several important tools like the Bogovskiı̆ operator and
Korn’s inequality loose some of their continuity properties. Optimal re-
sults, which might allow for a loss of integrability, can be achieved in the
framework of Orlicz spaces. They are flexible enough to study fine prop-
erties of measurable functions which are required in the Prandtl–Eyring
fluid model. We present optimal versions of the Bogovskiı̆ operator, Nečas’
 Preface xi

negative norm theorem and Korn’s inequality in this framework in Chap-

ter 2. The background about Orlicz- and Orlicz–Sobolev spaces is revised
in Chapter 1.
In Chapter 7 we deal with non-stationary flows of power law fluids.
A first step is to approximate the equations by a system whose solutions are
known to exist. In order to pass to the limit in the regularization parame-
ter, one has to apply compactness of the velocity, the “solenoidal Lipschitz
truncation” and arguments from monotone operator theory.
In the last part of the book we study stochastic partial differential equa-
tions in fluid mechanics. Probabilistic models have become more and more
important for applications and earned a strong interest amongst mathemati-
cians. They can, for instance, take into account physical uncertainties and
model turbulence in the fluid motion. We present first existence theorems
for generalized Navier–Stokes equations under random perturbations. The
results and methods build a basis for future research on stochastic partial
differential equations in the analysis of generalized Newtonian fluids. All
probabilistic tools are presented as well. The Chapter can be viewed as an
introduction to stochastic partial differential equations from an analytical
point of view. Thus, the proofs of the main result given in Chapter 10 are
accessible to analysts without prior knowledge in stochastic analysis.
ACKNOWLEDGMENT

This monograph is based on my habilitation thesis [26]. It contains re-

sults from the papers [27,28,30,31,33,34]. Note that the results from [31]
are an improvement on [32] and are not yet contained in [26]. Thanks goes
to J. Frehse, J. Málek and E. Süli for being referees of [26].
Special thanks go to all the people without whom the writing of this
book would have been impossible: to M. Fuchs for introducing me to
mathematical fluid mechanics, to L. Diening for being a constant source of
inspiration, for much valuable advise and his unique view on the Lipschitz
truncation method, to A. Cianchi for sharing his deep knowledge of Or-
licz spaces and to M. Hofmanová for opening the door for me to stochastic
partial differential equations. I would like to extend my sincere appreciation
to all the people with whom I have discussed research related to this mono-
graph, in particular: M. Bildhauer, J. Frehse, S. Schwarzacher. Moreover,
thanks goes to all people who carefully read parts of it: L. Diening, M. Hof-
manová, M.V. Lawson, M. Ottobre, S. Paulus, O. Penrose, B.P. Rynne,
Z.B. Wyatt.

Dominic Breit
Edinburgh
August 2016

xiii
NOTATION

Formula Meaning
C Continuous functions
Cα α -Hölder continuous functions
Ck k-times continuously differentiable functions
C∞ ∞-times continuously differentiable functions
spt(ϕ) Support of ϕ
C0∞ C ∞ -functions with compact support
C0k C k -functions with compact support
D Dual of C0∞
C0∞,div C0∞ -functions with vanishing divergence

Ddiv Dual of C0∞,div
C0∞,⊥ C0∞ -functions with vanishing mean value
X Dual space of X
 · X Norm on X
·, ·X Inner product on X
Ln n-dimensional Lebesgue measure
Hs s-dimensional Hausdorff measure
B Borelian σ -algebra
Lp Lebesgue-space of p-integrable functions
p
Lloc Lebesgue-space of locally p-integrable functions
p
Ldiv L p -functions with vanishing divergence
L⊥p L p -functions with vanishing mean
 
p Dual exponent of p: p = p/ p − 1
L A /A(L ) Orlicz-space generated by A
L⊥A /A(L )⊥ L A -functions with vanishing mean

A Young conjugate of A
u∗ Decreasing rearrangement of u
u∗v Convolution of u and v
Mf Hardy-Littlewood maximal operator applied to f
Br Ball with radius r
κ Br Ball with same center as Br and radius κ r
k Averaged Taylor polynomial of order k
W k ,p Sobolev functions with differentiability k and
integrability p

xv
xvi Notation

Formula Meaning
Wlock,p Sobolev functions with differentiability k and local
integrability p
k ,p
Wdiv W k,p -functions with vanishing divergence
W0k,p W k,p -functions with vanishing trace
W −k,p Dual space of W0k,p
k ,p
Wdiv W k,p -functions with vanishing divergence
W0k,div
,p
W0k,p -functions with vanishing divergence
−k,p
Wdiv Dual space of W0k,div
,p

W 1,A Orlicz-Sobolev space generated by A

W01,A W 1,A -functions with vanishing trace
×d
Rdsym Symmetric d × d matrices
Symmetric tensor product
Ep L p -functions with symmetric gradient in L p
E0p Ep -functions with vanishing trace
E0p,div E0p -functions with vanishing divergence
EA L A -functions with symmetric gradient in L A
E0A EA -functions with vanishing trace
E0A,div E0A -functions with vanishing divergence
BD Functions of bounded deformation
M Bounded signed measures
H A /H A(L) L A -functions with divergence in L A
H0A /H0A(L) H A -functions with vanishing normal trace
Bog Bogovskiı̆-operator
−1 Solution operator to the Laplace equation
−2 Solution operator to the bi-Laplace equation
curl−1 Solution operator to the curl equation
Qr Parabolic cube with radius r
κ Qr Cube with same center as Qr and radius κ r
Mα α -parabolic maximal function
Mασ α -parabolic maximal function with power σ
L p (0, T ; V ) Bochner measurable functions with values in V and
integrability p
C ([0, T ]; V ) Continuous functions with values in V
C α ([0, T ]; V ) α -Hölder continuous functions with values in V
Cw ([0, T ]; V ) Weakly continuous functions with values in V
 Notation xvii

Formula Meaning
W k,p (0, T ; V ) k-times weakly differentiable functions with values in V
and integrability p
(, F , P) Probability space with sample space , σ -algebra F and
probability measure P
(Ft )t≥0 filtration
L p (, F , P; V ) Random variable over (, F , P) with values in V and
moments of order p
M2 Quadratically integrable martingales
M2c Quadratically integrable continuous martingales
M2c,loc Locally in time quadratically integrable continuous
martingales
L2 (H1 , H2 ) Hilbert-Schmidt operators from H1 → H2
 
X t t≥0 Quadratic variation process of (Xt )t≥0
 
X , Y t t≥0 Covariation process of (Xt )t≥0 and (Yt )t≥0
Aq Stokes operator on L q
Aq A-Stokes operator on L q
D Domain of an operator
CHAPTER 1

Preliminaries
Contents
1.1 Lebesgue & Sobolev spaces 3
1.2 Orlicz spaces 8
1.3 Basics on Lipschitz truncation 11
1.4 Existence results for power law fluids 22

Abstract
In this chapter we present some preliminary material which will be needed in order to
study stationary models for generalized Newtonian fluids. We begin with the functional
analytic framework. In particular, we define Lebesgue-, Sobolev- and Orlicz-spaces and
describe their basic properties. After this we present the Lipschitz truncation method
in its classical framework and present two applications. Finally, we discuss some mod-
elling aspects concerning power law fluids and provide a historical overview on the
mathematical theory of weak solutions for stationary flows.

1.1 LEBESGUE & SOBOLEV SPACES

In this section we define various function spaces. For proofs, further details
and references we refer to [5].
Definition 1.1.1 (Classical function spaces). Let G ⊂ Rd be open and
k ∈ N. We define
C (G) := {u : G → R : u is continuous} ,
 
C k (G) := u : G → R : ∇ i u is continuous for i = 0, ..., k ,
 
C ∞ (G) := u : G → R : ∇ i u is continuous for all i ∈ N0 .
Remark 1.1.1. In Definition 1.1.1 if we replace G by the closed set G
we are considering functions whose derivatives are continuous up to the
boundary of G.
Definition 1.1.2. Let G ⊂ Rd be open and α ∈ (0, 1]. We define
   
u(x) − u(y)
C (G) := u : G → R : sup 
α
 <∞ ,
x=y x − yα
 
C α (G) := u : G → R : u ∈ C α (K ) for all K  G
as the set of (locally) α -Hölder continuous functions.

Existence Theory for Generalized Newtonian Fluids.

DOI: http://dx.doi.org/10.1016/B978-0-12-811044-7.00002-1
Copyright © 2017 Elsevier Inc. All rights reserved. 3
4 Existence Theory for Generalized Newtonian Fluids

Remark 1.1.2. In the case α = 1 we obtain the set of Lipschitz continuous

functions.
Definition 1.1.3. Let G ⊂ Rd be open and u : G → R. We define the
support of u by

spt u := {x ∈ G : u(x) = 0}.

By C00 (G) and C0∞ (G) we denote subclasses of C 0 (G) and C ∞ (G) whose
elements satisfy spt u  G.
Definition 1.1.4 (Lebesgue spaces). Let (X, , μ) be a measure space. We
define
 
L (X, , μ) := u : X → R : u is μ-measurable,
p
|u|p dμ < ∞ , 1 ≤ p < ∞,
 X

L ∞ (X, , μ) := u : X → R : u is μ-measurable, inf sup |u(x)| < ∞ .

μ(N )=0 x∈X\N

The elements of L p (X, , μ) are equivalence classes. Two functions u and

v belong to the same equivalence class if u = v μ-almost everywhere in X,
i.e. if
μ {x ∈ X : u(x) = v(x)} = 0.

Remark 1.1.3. • L p (X, , μ) is a Banach space together with the natu-

ral norm
 1
p
u p := u L p (X) := |u| dμ
p
for 1 ≤ p < ∞,
X
u ∞ := u L ∞ (X) := inf sup |u(x)| .
μ(N )=0 x∈X\N

• We set p = p−p 1 for 1 < p < ∞, p = ∞ for p = 1 and p = 1 for p = ∞.

Let 1 ≤ p < ∞ and v ∈ L p (X, , μ). Then the mapping

L p (X, , μ) u → u v dμ
G
belongs to the dual space L p (X, , μ) . On the other hand each element
of L p (X, , μ) can be represented via a function v ∈ L p (X, , μ). In
particular, the mapping

L p (X, , μ) v → L p (X, , μ) u → u v dμ ∈ L p (X, , μ)
X

is an isomorphism.
 Preliminaries 5

• The space L p (X, , μ) is reflexive if and only if 1 < p < ∞.

• If μ(X) < ∞ we have that L q (X, , μ) ⊂ L p (X, , μ) for 1 ≤ p ≤ q ≤ ∞.
In particular, the following holds
u L p (X) ≤ c (d, p, q, μ(X)) u L q (X)

for every u ∈ L q (X, , μ).

• We can define vector- or matrix-valued L p -spaces component-wise.
For simplicity we do not mention this in the notation. However, when
considering functions with values in an infinite dimensional spaces X
we denote this by writing L p (X, , μ; X ).

The most important special case is (X, , μ) = (G, B(Rd ), Ld ) where

G ⊂ Rd , B(Rd ) is the Borel σ -algebra on Rd and Ld the d-dimensional
Lebesgue measure. In this case we abbreviate the notation by setting
L p (G) := L p (G, B(Rd ), Ld ). If 1 ≤ p < ∞ then C0∞ (G) is dense in L p (G). In
particular, for every u ∈ L p (G) there is (um ) ⊂ C0∞ (G) such that um → u in
L p (G).

Definition 1.1.5 (Local Lebesgue spaces). Let G ⊂ Rd be open. We define

p
Lloc (G) := {u : G → R : u ∈ L p (K ) for all K  G} for 1 ≤ p ≤ ∞.

Definition 1.1.6 (Weak derivative). Let G ⊂ Rd be open and u ∈ Lloc 1 (G).

• u is called weakly differentiable with respect to the i-th variable if there

is a function vi ∈ Lloc
1 (G) such that

 
u ∂i ϕ dz = − vi ϕ dz for all ϕ ∈ C0∞ (G).
G G

We call vi the weak derivative of u with respect to the i-th variable.

If weak derivatives with respect to all variables exist, we call u weakly
differentiable and denote by ∇ u = (v1 , ..., vn ) the weak gradient of u.
• Let α ∈ N0d and |α| := α1 + ... + αn . For a smooth function u we denote
|α|
by Dα u := ∂ α1 x∂1 ...∂uαn xn its α -th classical derivative. A function u ∈ Lloc
1 (G)

is called k-times weakly differentiable if for all α ∈ N0d with |α| ≤ k there
is a function vα ∈ Lloc 1 (G) such that

 
u Dα ϕ dz = (−1)|α| vα ϕ dz for all ϕ ∈ C0∞ (G).
G G

We call Dα u = vα the α -th weak derivative.

6 Existence Theory for Generalized Newtonian Fluids

Definition 1.1.7 (Sobolev spaces). For k ∈ N and 1 ≤ p ≤ ∞ we define


W k,p (G) := u : G → R, Dα u ∈ L p (G), for all α ∈ N0d , |α| ≤ k ,
W k,p (G)
W0k,p (G) := C0∞ (G) .

Remark 1.1.4. • W k,p (G) (1 ≤ p ≤ ∞) is a Banach space together with

the norm
 1
p
u k ,p := u W k,p (G) := |D u| dz
α p
for 1 ≤ p < ∞,
|α|≤k G

u k,∞ := u W k,∞ (G) := inf sup |Dα u(x)| .

Ld (N )=0 x∈G\N
|α|≤k

• The space W k,p (G) is reflexive iff 1 < p < ∞.

• A sequence (uk ) ⊂ W k,p (G) (1 < p < ∞) converges weakly to u in
W k,p (G) if
 
Dα uk v dz −→ Dα u v dz, k → ∞,
G G

for all v ∈ L p (G, RN ) (p := p−p 1 ) and for all α ∈ N0d with |α| ≤ k.
• The dual space of W0k,p (G) will be denoted by W −k,p (G).
Theorem 1.1.1 (Smooth approximation). Let G ⊂ Rd be open and bounded
with Lipschitz boundary (i.e. ∂ can be locally parametrized by Lipschitz con-
tinuous functions of d − 1 variables) and 1 ≤ p < ∞. Then C ∞ (G) is dense in
W k,p (G). In particular, for every u ∈ W k,p (G) there is (um ) ⊂ C0∞ (G) such that
um → u in W k,p (G).
Theorem 1.1.2 (Sobolev). Let G ⊂ Rd be open.
a) The embeddings
dp
W01,p (G) → L d−p (G) if p < d,
d
W01,p (G) → C 1− p (G) if p > d, Ld (G) < ∞,
are continuous.
b) Let G ⊂ Rd be open and bounded with Lipschitz-boundary. The embeddings
dp
W 1,p (G) → L d−p (G) if p < d,
1− dp
W 1,p (G) → C (G) if p > d,
are continuous.
 Preliminaries 7

Theorem 1.1.3 (Kondrachov). Let G ⊂ Rd be open and bounded. The em-

bedding
pd
W01,p (G) → L q (G), q< ,
d−p
is compact for all p < d.
In Definition 1.1.7 we have interpreted the boundary values of a
Sobolev function as follows: u = 0 on ∂ G iff u ∈ W01,p (G), where W01,p (G)
for p < ∞ denotes the closure of C0∞ (G) in W 1,p (G). We will develop a
rigorous definition of boundary values and show that it coincides with the
former one. In order to do so we need functions which are integrable over
the boundary of G. For G ⊂ Rd open let L p (∂ G) be equal to the set of all
Hd−1 -measurable functions with
 1
p
u L p (∂) := |u|p dHd−1 < ∞.
∂G

Here Hd−1 denotes the (d − 1)-dimensional Hausdorff measure. For

C 1 -functions we define the operator tr : C 1 (G) → L p (∂ G) by tr u := u|∂ G .
Lemma 1.1.1. Let u ∈ C 1 (G) with G ⊂ Rd open and bounded with Lipschitz–
boundary. Then we have
tr u L p (∂ G) ≤ c (d, p, G) u W 1,p (G) .

For u ∈ W 1,p (G) we consider the approximation sequence um ∈

C ∞ (G) with um − u 1,p → 0. Its existence follows from Theorem 1.1.1.
Lemma 1.1.1 shows that (tr um ) is a Cauchy-sequence in L p (∂ G). We define
its limit (which does not depend on the special choice of the sequence) as
the trace of u. The result is a linear operator tr : W 1,p (G) → L p (∂ G) which
coincides with the classical trace operator on C 1 (G).
Theorem 1.1.4. Let  ⊂ Rd be open and bounded with Lipschitz boundary.
Then we have
 
W01,p (G) = u ∈ W 1,p (G); tr u = 0
for 1 ≤ p < ∞.
All results of this section generalize in a straightforward manner to spaces
of vector-valued functions. In order to keep the notation simple we do not
use target spaces in the notation for our function spaces. It will follow from
the context (and the bold-symbol) when we are dealing with these. In fluid
8 Existence Theory for Generalized Newtonian Fluids

mechanics the velocity field is a function from Rd ⊃ G → Rd . In this setting

we need function spaces of solenoidal (that is, divergence-free) functions.
We will use the following notation for 1 ≤ p < ∞
1,p
Wdiv (G) := {ψ ∈ W 1,p (G) : div ψ = 0},
C0∞,div () := {ψ ∈ C0∞ (G) : div ψ = 0},
p L 2 (G) W 1,p (G)
Ldiv (G) := C0∞,div (G) , W01,div
,p
(G) := C0∞,div (G) .
−1,p 1,p
Finally we write Wdiv (G) for the dual of W0,div (G).

1.2 ORLICZ SPACES

In this section we present some important properties of Orlicz spaces
(see [125] and [5]).
A function A : [0, ∞) → [0, ∞] is called a Young function if it is con-
vex, left-continuous, vanishing at 0, and neither identically equal to 0 nor
to ∞. Thus, with any such function, it is uniquely associated a (non-trivial)
non-decreasing left-continuous function a : [0, ∞) → [0, ∞] such that
 s
A(s) = a(r ) dr for s ≥ 0. (1.2.1)
0

The Young conjugate A of A is the Young function defined by

A(s) = sup{rs − A(r ) : r ≥ 0} for s ≥ 0 .
For A we have the representation formula
 s
A(s) = a−1 (r ) dr for s ≥ 0, (1.2.2)
0
where a−1 denotes the (generalized) left-continuous inverse of a. Moreover,
for every Young function A,
r ≤ A−1 (r )A−1 (r ) ≤ 2r for r≥0 (1.2.3)
as well as
A = A. (1.2.4)
Let A be a Young function of the form (1.2.1). Then the convexity of A
and A(0) = 0 imply
A(λs) ≤ λA(s) for all λ ∈ [0, 1] (1.2.5)
and all s ≥ 0. If λ ≥ 1, then
λA(s) ≤ A(λs) for s ≥ 0. (1.2.6)
 Preliminaries 9

As a consequence, if λ ≥ 1, then
A−1 (λs) ≤ λA−1 (s) for s ≥ 0, (1.2.7)
where A−1 denotes the (generalized) right-continuous inverse of A.
A Young function A is said to satisfy the 2 -condition if there exists a
positive constant K such that
A(2s) ≤ KA(s) for s ≥ 0. (1.2.8)
If (1.2.8) just holds for s ≥ s0 for some s0 > 0, then A is said to satisfy the
2 -condition near infinity. We say that A satisfies the ∇2 -condition [near
infinity] if à satisfies the 2 -condition [near infinity].
A Young function A is said to dominate another Young function B near
infinity if there exist positive constants c and s0 such that
B(s) ≤ A(cs) for s ≥ s0 . (1.2.9)
The functions A and B are called equivalent near infinity if they dominate
each other near infinity.
Let G be a measurable subset of Rd , and let u : G → R be a measur-
able function. Given a Young function A, the Luxemburg norm associated
with A, of the function u is defined as
   |u|  
u LA (G) := inf λ : A dx ≤ 1 .
G λ
The collection of all measurable functions u for which this norm is finite
is called the Orlicz space L A (G). It turns out to be Banach
function space.
The subspace of L A (G) of those functions u such that G u(x) dx = 0 will
be denoted by L⊥A (G). A Hölder-type inequality in Orlicz spaces takes the
form 
u(x)v(x) dx
v LA (G) ≤ sup G ≤ 2 v LA (G) (1.2.10)
u∈L A (G) u LA (G)
for every v ∈ L A (G). If |G| < ∞, then
L A (G) → L B (G) if and only if A dominates B near infinity . (1.2.11)
The decreasing rearrangement u∗ : [0, ∞) → [0, ∞] of a measurable func-
tion u : G → R is the (unique) non-increasing, right-continuous function
which is equimeasurable with u. Thus,
u∗ (s) = sup{t ≥ 0 : |{x ∈ G : |u(x)| > t}| > s} for s ≥ 0.
The equimeasurability of u and u∗ implies that
u L A (G) = u∗ L A (0,|G|) (1.2.12)
for every u ∈ L A (G).
10 Existence Theory for Generalized Newtonian Fluids

The Lebesgue spaces L p (G), corresponding to the choice Ap (t) = tp , if

p ∈ [1, ∞), and A∞ (t) = ∞ · χ(1,∞) (t), if p = ∞, are a basic example of Or-
licz spaces. Other instances of Orlicz spaces are provided by the Zygmund
spaces L p logα L (G), and by the exponential spaces exp L β (G). If either p > 1
and α ∈ R, or p = 1 and α ≥ 0, then L p logα L (G) is the Orlicz space asso-
ciated with a Young function equivalent to tp (log t)α near infinity. Given
β > 0, exp L β (G) denotes the Orlicz space built upon a Young function
β
equivalent to et near infinity.
An important tool will be the following characterization of Hardy type
inequalities in Orlicz spaces [46, Lemma 1].

Lemma 1.2.1. Let A and B be Young functions, and let L ∈ (0, ∞].
(i) There exists a constant C such that
  s 
1 
 f (r ) dr  ≤C f (1.2.13)
s L A (0,L )
0 L B (0,L )

for every f ∈ L A (0, L ) if and only if either L < ∞ and there exist constants c > 0
and t0 ≥ 0 such that
 t
B(s)
t ds ≤ A(ct) for t ≥ t0 , (1.2.14)
t0 s2
or L = ∞ and (1.2.14) holds with t0 = 0. In particular, in the latter case, the
constant C in (1.2.13) depends only on the constant c appearing in (7.0.1).
(ii) There exists a constant C such that
 
 L
dr 
 f (r ) ≤C f (1.2.15)
 r LB (0,L) L A (0,L )
s

for every f ∈ L A (0, L ) if and only if either L < ∞ and there exist constants c > 0
and t0 ≥ 0 such that
 t
A(s)
t ds ≤ B(ct) for t ≥ t0 , (1.2.16)
t0 s2
or L = ∞ and (1.2.16) holds with t0 = 0. In particular, in the latter case, the
constant C in (1.2.15) depends only on the constant c appearing in (1.2.16).

Assume now that G is an open set. The Orlicz–Sobolev space W 1,A (G)
is the set of all functions in L A (G) whose distributional gradient also be-
longs to L A (G). It is a Banach space endowed with the norm
u W 1,A (G) := u L A (G) + ∇u L A (G) .
 Preliminaries 11

We also define the subspace of W 1,A (G) of those functions which vanish
on ∂ G as
W01,A (G)
= {u ∈ W 1,A (G) : the continuation of u by 0 is weakly differentiable}.
In the case where A(t) = tp for some p ≥ 1, and ∂ G is regular enough, such a
definition of W01,A (G) can be shown to reproduce the usual space W01,p (G)
defined as the closure in W 1,p (G) of the space C0∞ (G) of smooth compactly
supported functions in G. In general, the set of smooth bounded functions
is dense in L A (G) only if A satisfies the 2 -condition (just near infinity
when |G| < ∞), and hence, for arbitrary A, our definition of W01,A (G)
yields a space which can be larger than the closure of C0∞ (G) in W01,A (G)
even for smooth domains. On the other hand, if G is a Lipschitz domain,
namely a bounded open set in Rd which is locally the graph of a Lipschitz
function of d − 1 variables, then
W01,A (G) = W 1,A (G) ∩ W01,1 (G),
where W01,1 (G) is defined as usual.
Lemma 1.2.2. Let A be a Young-function satisfying the 2 -condition and
G ⊂ Rd open.
• The following holds
W 1,A (G)
W01,A (G) = C0∞ (G) = W 1,A (G) ∩ W01,1 (G).
• For every u ∈ W 1,A there is a sequence (uk ) ∈ C ∞ (G) ∩ W 1,A (G) such that
uk → u in W 1,A (G).
Lemma 1.2.3. Let A be a Young function satisfying the 2- and the
∇2 -condition. Then L A (G) is reflexive with
L A (G) ∼
= L Ã (G).

1.3 BASICS ON LIPSCHITZ TRUNCATION

The purpose of the Lipschitz truncation technique is to approximate a
Sobolev function u ∈ W 1,p by λ-Lipschitz functions uλ that coincide with u
up to a set of small measure. The functions uλ are constructed nonlinearly
by modifying u on the level set of the Hardy–Littlewood maximal function
of the gradient ∇ u. This idea goes back to Acerbi and Fusco [1–3]. Lips-
chitz truncations are used in various areas of analysis: calculus of variations,
12 Existence Theory for Generalized Newtonian Fluids

in the existence theory of partial differential equations, and in regularity

theory. We refer to [62] for a longer list of references.
The basic idea is to take a function u ∈ W 1,p (Rd ), where p ≥ 1, and
cut values where the maximal function of its gradient is large. The Hardy–
Littlewood maximal operator is defined by

M (v)(x) = sup − |v| dy
B: x∈B B

for v ∈ 1 (Rd )
Lloc which can be extended to vector- (or matrix-)valued
functions by setting M (v) = M (|v|). Basic properties of the maximal op-
erator are summarized in the following lemma (see e.g. [134] and [141,
Lemma 3.2] for d)).
Lemma 1.3.1. a) Let v ∈ Lloc1 (Rd ) and λ > 0. The level-set {x ∈ Rd :

|M (v)(x)| > λ} is open.

b) The strong-type estimate
M (v) L p (Rd ) ≤ cp v L p (Rd ) ∀v ∈ L p (Rd )
holds for all p ∈ (1, ∞].
c) The weak-type estimate
  v pp
Ld {x ∈ Rd : |M (v)(x)| > λ} ≤ cp p ∀v ∈ L p (Rd )
λ
holds for all p ∈ [1, ∞) and all λ > 0.
d) We have the estimate
 
|v|p dx ≤ cp |v|p dx ∀p ∈ [1, ∞)
{M (v)>λ} {|v|>λ/2}

for all λ > 0.

The Lipschitz truncation will be defined via the maximal function of
the gradient. For u ∈ W 1,p (Rd ) the “bad set” is defined by
Oλ := {x ∈ Rd : M (∇ u)(x) > λ}, (1.3.17)
where λ ≥ 0. If we have u ∈ for G =
W 1,p (G) Rd
it has to be extended to
u ∈ W (R ). This can be done in an obvious way if u ∈ W01,p (G) where
1,p d

the extension is zero outside G. In the general case we may apply [5,
Thm. 4.26]. Now, for x, y ∈ Rd \ Oλ we have a.e.
 
|u(x) − u(y)| ≤ c |x − y| M (∇ u)(x) + M (∇ u)(y) ≤ 2c λ|x − y|,
see, e.g., [112]. Hence u is Lipschitz-continuous in Rd \ Oλ with Lipschitz
constant proportional to λ. By a standard extension theorem (see e.g. [71,
 Preliminaries 13

p. 201]) we can extend u (defined in Rd \ Oλ ) to uλ (defined in Rd ) such

that the Lipschitz constant is preserved. (When dealing with this simple
extension it is necessary to cut large values of M (u) as well. We neglect this
for brevity.) This means we have
|∇ uλ | ≤ c λ in Rd . (1.3.18)
Moreover, by construction we have
{x ∈ Rd : u = uλ } ⊂ Oλ .
This and Lemma 1.3.1 c) imply
  c ∇u p
p
L {x ∈ R : u = uλ } ≤ L (Oλ ) ≤
d d d
. (1.3.19)
λp
Combining (1.3.18) and (1.3.19) shows
  
|∇ uλ | dx =
p
|∇ uλ | dx +
p
|∇ uλ |p dx
Rd Rd \Oλ Oλ

 
≤ |∇ u|p dx + c λp Ld Oλ
Rd \Oλ

≤c |∇ u|p dx.
Rd

We obtain the following stability result

∇ uλ p ≤ c ∇ u p. (1.3.20)
The basic properties (1.3.18)–(1.3.20) are already enough to make the Lips-
chitz truncation a powerful tool for numerous applications. We present two
rather classical ones.

Lower semi-continuity in W 1,p .

Let G ⊂ Rd be an open and bounded with Lipschitz boundary. Suppose
further that F : G × Rd×D → [0, ∞) is a continuous function with p-growth
(p > 1), i.e.,
F (x, Q) ≤ c |Q|p + g(x) ∀Q ∈ Rd×D (1.3.21)
with a constant c ≥ 0 and a non-negative function g ∈ L 1 (G). We are inter-
ested in minimizing the functional

GF [w] = F (x, ∇ w) dx
G

defined for functions w : G → RD . An important concept in showing the

existence of minimizers is the lower semi-continuity of GF with respect to
14 Existence Theory for Generalized Newtonian Fluids

an appropriate topology. The functional GF is called W 1,p -weakly lower

semi-continuous if
GF [v] ≤ lim inf GF [vn ] (1.3.22)
n→∞

provided vn v in W 1,p (G) for m → ∞. The functional GF is called

1,∞ ∗
W -weakly lower semi-continuous if
GF [v] ≤ lim inf GF [vn ]
n→∞
∗
provided vn v in for n → ∞.
W 1,∞ (G)
Due to (1.3.21) the right concept for the functional GF is W 1,p -weak
lower semi-continuity. It can be deduced from W 1,∞ -weak∗ lower semi-
continuity by the Lipschitz truncation, see Lemma 1.3.3 below. Using this
idea Acerbi and Fusco [1] showed the W 1,p lower semi-continuity of GF in
the case where F is only quasi-convex. Note that W 1,∞ -weak∗ lower semi-
continuity is a consequence of the definition of quasi-convexity, see [1,
Thm. 2.1]. For brevity we do not discuss the concept of quasi-convexity
and refer instead to the fundamental papers [16] and [115].

Lower integrability for the p-Laplace system.

Consider the system
div(|∇ v|p−2 ∇ v) = div F in G, (1.3.23)
v=0 on ∂ G. (1.3.24)
Here, G is an open set in Rd ,
with d ≥ 2, the exponent p ∈ (1, ∞) and the
×D
function F :  → R d is given. A weak solution to (1.3.23) is a function
v ∈ W01,p (G) such that
 
p−2
|∇ v| ∇ v : ϕ dx = F : ∇ϕ dx
G G

for all ϕ ∈ W01,p (G). Its existence can be shown via standard methods pro-
vided F ∈ L p (G). We are concerned here with the question of how the
regularity of F transfers to v (particularly to |∇ v|p−2 ∇ v). In the linear
case p = 2 this is answered by the classical theory of Calderón and Zyg-
mund [43]. It says that F ∈ L q (G) implies ∇ v ∈ L q (G) for all q ∈ (1, ∞).
Note that the case q < 2, where q is below the duality exponent p , is in-
cluded. In that situation existence of weak solutions is not clear a priori.
There has been a great deal of effort in obtaining a corresponding result for
the nonlinear case p = 2 such that
F ∈ L q (G) ⇒ |∇ v|p−2 ∇ v ∈ L q (G) ∀q ∈ (1, ∞) (1.3.25)
 Preliminaries 15

together with a corresponding estimate. This has been positively answered

in the fundamental paper by Iwaniec [97] provided q ≥ p . An improvement
to q > p − δ for some small δ > 0 has been carried out in [98] by different
methods (for an overview and further references see [113]). We remark that
the case q ∈ (1, p − δ) is still open. Based on the Lipschitz truncation we can
give a relatively easy proof for the estimate in the case q ∈ (p − δ, p ) using
the approach in [141] (see also [40] for a more general setting and [101] for
the parabolic problem), see Lemma 1.3.4 below.
Before we give proofs of these applications we present an important
improvement of the Lipschitz truncation which firstly appeared in [62]. It
concerns the smallness of the level-sets. Similar ideas have been used earlier
for the L ∞ -truncation in [78].

Lemma 1.3.2. Let v ∈ L p (Rd ) with p ∈ (1, ∞). Then there exist j0 ∈ N and a
j +1
sequence λj ∈ R with 22 ≤ λj ≤ 22 −1 such that
j

 
λj Ld {x ∈ Rd : M (v) > λj } ≤ c p 2−j v
p p
p

for all j ≥ j0 where c = cp is the constant in Lemma 1.3.1 b).

Proof. We have
  ∞
M (v) pp = ϑ p−1 χ{|v|>ϑ} dϑ dx
Rd 0
  2m+1
= ϑ p−1 χ{|M (v)|>ϑ} dϑ dx
Rd m∈Z 2m

 p (1.3.26)
≥ 2m χ{|M (v)|>2m+1 } dx
Rd m∈Z

2j+1 −1 
 k p
≥ 2 χ{|M (v)|>2·2k } dx.
j∈N Rd
k=2j

The continuity of M on L p (Rd ), see Lemma 1.3.1 b), implies

2j+1 −1 
 k p
2 χ{|M (v)|>2·2k } dx ≤ c p v pp .
j∈N Rd
k=2j

In particular, for all j ∈ N

2j+1 −1 
 k p
2 χ{|M (v)|>2·2k } dx ≤ c p v pp .
Rd
k=2j
16 Existence Theory for Generalized Newtonian Fluids

Since the sum contains 2j summands, there is at least one index kj such that

 p
2kj χ{|M (v)|>2·2kj } dx ≤ c p v pp 2−j . (1.3.27)
Rd

Define λj := 2kj and we conclude from (1.3.27) that


 p p −j
λj χ{|M (v)|>2 λj } dx ≤ c p v p2 .
Rd
This proves the claim.
Lemma 1.3.2 shows that there is a particular sequence of levels (λj ) such
that
 
p
λj Ld Oλj ≤ κj ∇ v p
p (1.3.28)
with κj → 0 for j → ∞. This improves the estimate (1.3.19) and does not
follow from the original results by Acerbi and Fusco. It allows us to simplify
the original proof of W 1,p -lower semi-continuity from [1].
Lemma 1.3.3. Assume that the functionals GF defined in (1.3.22) are
W 1,∞ -weakly∗ lower semi-continuous for any choice of F satisfying (1.3.21). Then
they are W 1,p -weakly lower semi-continuous.
Proof. Let (vn ) ⊂ W 1,p (G) be a sequence with weak limit v such that
un := vn − v 0 in W 1,p (G).
We take the sequence (λj ) in accordance with Lemma 1.3.2 for the level-
sets of ∇ u. We apply the Lipschitz truncation to the sequence (un ) with
level λ = λj , see the construction after (1.3.17), and obtain for the double
sequence (un,j := un,λj )
∇ un,j ∞ ≤ c λj (1.3.29)
∗ 1,∞
un,j 0 in W (G) (1.3.30)
 
p
λj L d {x ∈ Rd : un = un,j } ≤ κj (1.3.31)
due to (1.3.18)–(1.3.20), where κj → 0 for j → ∞. We obtain

lim inf G [vn ] = lim inf F (·, ∇ v + ∇ un ) dx
m→∞ n→∞ G
 
≥ lim inf F (·, ∇ v + ∇ un ) dx = lim inf F (∇ v + ∇ un,j ) dx
n→∞ G\Oλj n→∞ G\Oλj
 
≥ lim inf F (·, ∇ v + ∇ un,j ) dx − lim sup F (·, ∇ v + ∇ un,j ) dx.
n→∞ G n→∞ O λj

(1.3.32)
 Preliminaries 17

We can use the functional GF̃ with

F̃ (x, Q) = F (x, ∇ v + Q), Q ∈ R d ×D .
The function F̃ has p-growth as required in (1.3.21) (by ∇ v ∈ L p ()) such
that

lim inf F (·, ∇ v + ∇ un,j ) dx = lim inf GF̃ [un,λj ] ≥ GF̃ [0] = GF [v]. (1.3.33)
n→∞ G n→∞

This is a consequence of the W 1,∞ -weak∗ lower semi-continuity of GF̃ .

Finally we have
  
 
F (·, ∇ v + ∇ un,j ) dx ≤ c |∇ v|p + g dx + c |∇ un,j |p dx
O λj O λj O λj

   
≤c |∇ v|p + g dx + c λd Ld Oλj
O λj

such that

lim sup lim sup F (·, ∇ v + ∇ un,j ) dx = 0 (1.3.34)
λ→∞ n→∞ O λj

by (1.3.31) and |∇ v|p + g ∈ L 1 (G). Combining (1.3.32)–(1.3.34) shows that

GF [v] ≤ lim inf GF [vn ],
m→∞
i.e. GF is W 1,p -weakly lower semi-continuous.
We now turn to the proof of the lower integrability for the p-Laplace
system. In addition to the Lipschitz truncation crucial ingredients are the
following integral identities. Let 0 <  < ∞, 0 ≤ δ <  < δ and (X, , μ)
be a measure space. There holds for every μ-measurable function f with
|f | ∈ L 1 (X, , μ)
 ∞  
1
ϑ −1−δ |f |δ dμ dϑ = |f | dμ, (1.3.35)
0 {|f |>ϑ} −δ X
 ∞  
1
ϑ −1−δ |f |δ dμ dϑ = |f | dμ. (1.3.36)
0 {|f |≤ϑ} δ− X

Both equalities are easy consequences of Fubini’s Theorem. As we will

apply (1.3.35) and (1.3.36) several times it is important that all estimates
hold for any λ > 0. So Lemma 1.3.2 is no use.
Lemma 1.3.4. There is a number δ > 0 such that for all q ∈ (p −δ) the following
holds. Let v ∈ W01,q(p−1) (G) be a weak solution to (1.3.23) with F ∈ L q (G). Then
we have
 
 
|∇ v|p−2 ∇ vq dx ≤ c |F|q dx.
G G
18 Existence Theory for Generalized Newtonian Fluids

Proof. Take the solution v to (1.3.23) and use its Lipschitz truncation vλ
as a test-function, see the construction after (1.3.17). Note that the Lips-
chitz truncation can preserve zero boundary values at least in the case of a
Lipschitz boundary, cf. [62, Thm. 3.2]. We obtain
   
p−2 p−2
|∇ v| ∇ v : ∇ vλ dx = − |∇ v| ∇ v : ∇ vλ dx + F : ∇ vλ dx
Rd \Oλ Oλ Rd

which implies by (1.3.18)

  
p−1
|∇ v| dx ≤ c λ
p
|∇ v| dx + |F||∇ vλ | dx. (1.3.37)
Rd \Oλ Oλ Rd

Furthermore, the following holds

  
|∇ v| dx =p
|∇ v| dx + p
|∇ v|p dx
{|∇ v|≤λ} {|∇ v|≤λ}∩Oλ {|∇ v|≤λ}\Oλ
 
≤λ |∇ v|p−1 dx + |∇ v|p dx.
Oλ Rd \O λ

Inserting this into (1.3.37) yields

  
|∇ v|p dx ≤ c λ |∇ v|p−1 dx + |F||∇ vλ | dx.
{|∇ v|≤λ} Oλ Rd

As a consequence of Lemma 1.3.1 d) we deduce

  
|∇ v|p dx ≤ c λ |∇ v|p−1 dx + |F||∇ vλ | dx.
{|∇ v|≤λ} {|∇ v|>λ/2} Rd

After multiplying with λq−1−p and integrating we have

 ∞   ∞ 
q−1−p q−p
λ |∇ v| dx dλ ≤ c
p
λ |∇ v|p−1 dx dλ
0 {|∇ v|≤λ} 0 {|∇ v|>λ}
 ∞ 
q−1−p
+ λ |F||∇ vλ | dx dλ.
0 Rd

Setting χλ := Rd |F||∇ vλ | dx we obtain on account of (1.3.35) and (1.3.36)
   ∞
1 c
|∇ v| dx ≤
q
|∇ v| dx +
q
λq−1−p χ (λ) dλ.
p−q q−p+1 0

If q is close enough to p, say p − q < δ̃ , we have

c 1
<
q−p+1 p−q
and hence
  ∞
|∇ v|q dx ≤ c λq−1−p χ (λ) dλ.
0
Another random document with
no related content on Scribd:
several thousands every year, but in 1902 they sank to the
comparatively small total of 600. But whilst every endeavour is made
to govern the country on lines acceptable to and understood by the
people, there are, of course, some points on which the policy of a
civilized government is necessarily in opposition to very deep-rooted
customs and habits to such a degree as to completely upset the old
basis of social life. In a minor degree this is true of the partial
application of the game laws to natives, but of far greater importance
is our attitude towards the institution of slavery and the slave-trade.
The occupation of the Soudan has been a tremendous blow to
slavery; one of the principal recruiting-grounds for slaves has
practically been closed. A certain amount of slave-raiding goes on
along the Abyssinian frontier. Descents are periodically made by
parties one to two hundred strong, well armed, from the south-
western districts of Abyssinia. They raid the Barun negroes, and
carry off the women and children. The same kind of thing is apt to
happen on the Darfur frontier, and some of the remote tribes in the
same quarter sometimes raid each other with the object of getting
slaves. Some of these, but not many, find their way to Dongola, or
the Ghezireh; others are taken to Tripoli. Special steps have now
been taken by the Anti-Slavery Department of the Egyptian
Government, which now has its headquarters at Khartoum, to put
down this traffic. Two extra English inspectors have been posted,
one at Rosaires, on the Blue Nile, the other at El Obeid. They are to
form small mounted corps of the best Arabs and patrol the disturbed
districts.
Apart from the actual work done, it will be a great thing to enlist
the best of the Arab tribesmen in the Government service, and it is
hoped that they will in time form the nucleus of an effective native
police. These men have been in the past some of the principal
exponents of slave-catching themselves; they ought to be very adept
at their new business. With these exceptions the slave-trade within
the borders of the Soudan has practically disappeared. During the
first year or two captures were occasionally made of small caravans,
but very heavy penalties were imposed. There is still a constant
demand for slaves in Arabia, and once a slave is shipped over the
Red Sea a good profit is assured. But it is too dangerous now for
anyone to try to make a regular livelihood by it. There are still,
however, about twenty cases a year of trials for offences against the
slavery laws, mostly isolated cases of kidnapping a woman or a
child, and probably there is besides a fair proportion of undetected
cases. But the regular trade is pretty well stamped out.
The benefits of the abolition of slave-raiding and kidnapping are
immediate and obvious to everyone. Even the Arab can understand
them, but he finds it very difficult to appreciate our attitude—which,
needless to say, is uncompromising enough—towards slavery as a
domestic institution. It is the one serious complaint which he has
against the new government. His domestic habits and customs have
been completely based on slavery for centuries. Slavery is permitted
and recognised by the Koran. In most cases the slaves themselves
have been treated more like members of the family than as slaves,
and no doubt many of them have had far happier lives than they
would have had in their own villages. Nor is it difficult to point to evils
which have arisen from the emancipation of the slaves. It is a
melancholy fact that many of the towns in the Soudan are crowded
with freed slaves, too lazy to do anything but steal, while the women
have recourse to an even less reputable occupation. It is easier to
break down the social system of centuries than to build up a sounder
fabric in its place. But the thing had absolutely to be done if the
Soudan was to have a real regeneration. Even when the slaves have
been well treated, the demoralization caused by slavery has been
great. The Arabs have all the vices of a slave-owning people. It was
a good time to make an absolutely fresh start. All changes of such
magnitude are bound to produce dislocations. The evils of the
change will die out with the present generation. The good must be
waited for patiently, but it is sure to come.
 CHAPTER XIX
EDUCATION AND THE GORDON COLLEGE

Probably the last thing that a military government might be

expected to take an interest in is education, and yet few educational
establishments are so widely famous as the Gordon College at
Khartoum. As yet it owes its reputation to the fact that it was founded
by one great soldier in memory of another, not to its achievements; it
was born great. But already it has justified its founders; its mere
existence marks an extraordinary contrast between the character of
the present régime in the Soudan and that of any preceding.
Whenever the Soudan is mentioned, the first question asked is,
‘How is the Gordon College getting on?’ and the question cannot be
answered in a word. The actual building is indeed for the present
complete. It is a handsome structure of native red brick, built in the
Moorish style, but retaining the collegiate character. It occupies two
sides of a square, the front facing on the river. In the centre is the
principal entrance, and over it a tower. If the original design is finally
carried out, the whole quadrangle will be completed. Along the inside
runs a cool and airy cloister, with winding stairs leading to the upper
story; the class-rooms are spaciously designed. Its commanding
position at the east end of the town makes it a conspicuous
landmark for many miles round. From no point is this so remarkable
as from the hill of Surgham, which overlooks the battlefield of Kerreri.
Here is summed up much of the past and the future of the Soudan.
On the one hand is the scene of the final overthrow of the forces of
darkness and ignorance by war; on the other the symbols of that
longer contest for the conquest of the Soudan by the peaceful arts of
science and learning.
With the eye of faith it is easy to look forward into the future, and
to imagine the time, generations hence, when the Gordon College
will be a true centre of learning for all these vast territories. Then it
will stand, a completed quadrangle, in the middle of large gardens,
its own territory, as green and cultivated as they are now arid and
dusty. Its halls and class-rooms will be crowded with picked students
from all the provincial centres, not vainly pursuing a dry and vain
scholasticism, as in the other degenerate Universities of the
Mohammedan East, but eagerly following in the paths of living
science, and learning by practical teaching in the laboratory and
workshop to wrest from Nature her secrets, and to absorb the
principles underlying practice in the departments of chemistry and
medicine, mechanics, agriculture, and the arts. Perhaps once more,
in years to come, the culture and science of the Arabs will be as
famous as they were in the great days of Arab dominion.
It is a long way to travel from such stimulating forecasts to the
actual state of learning and education in the Soudan to-day. There
was not much learning under the Egyptians, but at least a certain
amount of theological study went on. Under the Khalifa even that
was sedulously discouraged, and the books were ordered to be
destroyed because he feared that they might tend to discredit the
unorthodox doctrine of Mahdism. Reading and writing were not likely
to flourish under a ruler who, possessing neither of these arts
himself, and entertaining strong suspicions of those who did, was
wont to give drastic expression to his views. As a consequence,
there never was a country more absolutely and wholly illiterate.
Writing is practically an unknown art, and reading hardly less so. It is
perfectly useless to post a Government Proclamation unless a
competent person is stationed by it to read it out to any passer-by. At
the same time, there flourishes the most exaggerated respect for a
written document, which is regarded as a kind of magic book, and
cases have been known in which swindlers have extorted large
sums of money by going round exhibiting a paper professing to be
an order to pay issued by the Government. Obviously, education has
had to be on very humble lines at first, and must continue so for
some time.
Was, then, the Gordon College a too ambitious attempt to
anticipate the future? Is it a mere white elephant, doomed to be a
vain monument of an ill-directed wave of enthusiasm? Such a view is
far from the truth. It would be strange indeed if a project so dear to
the heart of Lord Kitchener was of such a nature. Certainly, it is
impossible to start a complete University right away with a building
and an endowment of some £4,000 a year. Time is of the essence of
the question. It is possible to argue that the money used in the
building might have been more advantageously expended in other
ways. But, apart from the fact that the subscribers doubtless wished
to see some immediate result for their munificence, I am sure that it
was the right policy to build at once.
For the Gordon College, though not yet a University, is much
more than a college. It is the centre of all the new educational and
intellectual influences in the Soudan. Its director is also head of the
Education Department; the activities of both are inseparably
connected. It acts as an extraordinary stimulus upon the authorities
in the direction of education. It would have been so easy and so
natural for a Government so hard beset for money to neglect
education for other objects, apparently more practical and more
immediately pressing. The actual material presence of the college
makes it impossible for its claims to be overlooked. Very likely
without it there would not have been an Education Department at all.
Secondly, but for the existence of the building, the Soudan would
certainly never have obtained such valuable gifts as those of Mr.
Wellcome’s bacteriological laboratory and Sir W. Mather’s complete
technical workshop apparatus, containing all that is necessary for the
establishment and organization of departments for manual training
and technical instruction. Thirdly, the building itself has been already,
and will be to an increasing extent, of the greatest use; and,
moreover, there is still about £100,000 of the original endowment
remaining, the income from which is playing a great part, as will be
shown, in providing the beginnings of education in the Soudan, and
so laying the foundations for the future work of the college itself. Lord
Kitchener was wiser than his critics. Among his many claims to fame,
none is greater than the clearness with which he saw that a sound
educational system is one of the fundamental requirements of the
Soudan, as well as a substantial foundation for our rule.
 In 1901 the Soudan Government spent £1,421 on education, in
addition to the Gordon College endowment; in 1902, £3,577; and in
1903 something over £6,000. With such resources as these, it is
obvious that nothing heroic could be attempted. Only the more
immediate needs could be attended to. Looking to the necessity for
the education of a class of native public servants, it was most
important to establish some sound primary schools, in which the
boys should be given a fair general education in reading, writing, and
arithmetic, besides a certain amount of history and geography and
English. Four of these schools are now in existence; two of them,
those at Halfa and Suakin, were established some eight or nine
years ago by the Egyptian Government, and have only recently been
handed over to the Soudan; they are in an efficient condition. The
most interesting, considering their recent establishment, are those at
Khartoum and Omdurman. That at Omdurman was the first, the
direct offspring of the Gordon College. It now numbers over 200
pupils, and is in a most flourishing condition. There are constant
applications for admission, partly, no doubt, from the reason that
pupils of the school are thought likely to obtain Government
employment, but partly also from a real appreciation of the
advantages of education.
The school course is divided into four years, and the curriculum is
based on the Egyptian one—modified to the extent that no subjects
are taught in English, except English itself. An inspection of the
school made it clear to me that, at any rate, in the subject which I
could understand—the teaching of English—the methods were
thoroughly sound, and the results good. Pupils in their last year are
also taught land-measuring. The reason is the great demand which
comes from every province for land-measurers—a most important
thing in view of the assessments for the land-tax—while no trained
men are available. This part of the work is taught both in the class-
room and practically in the field. I am certain that in a very short time
there will be land-measurers available of very good ability, so
excellent was the quality of some of the work done.
Some, of course, of these boys are the sons of Egyptians in the
Government service, to whom it is a great blessing to be able to get
a good education for their children on the spot. But far the majority of
them—at least 90 per cent.—are genuine Soudanese, some of them
members of good Arab families, whose fathers were prominent in the
service of the Khalifa. Originally it was intended that this school
should be transferred to Khartoum, and housed in the college as
soon as the general exodus took place. But as Omdurman shows no
signs at present of diminishing, and is, indeed, once more
increasing, the school has been kept on, and another started on
similar lines in Khartoum itself. This school is also flourishing, but it
naturally contains a larger proportion of the Egyptian-born pupils;
altogether it has about 120 scholars. This school is now housed in
the Gordon College itself.
Book-learning is not the only channel of instruction employed. I
had the good fortune to be umpire in the first football match between
Khartoum and Omdurman schools, in the mosque square at
Khartoum. It was a hot afternoon, and I felt as though I should get a
sunstroke whilst umpiring; but these boys, all hatless as they were,
played with great energy, and appeared to derive nothing but benefit
from the heat. They played a good game, and it was pleasant to see
that what they lacked in experience they made up in courage and
determination. All the players, of whatever shade of black or brown,
and the shades were very various, showed a spirit which augurs well
for the future. Anybody who can play football with energy in Central
Africa must have good stuff in him. The match was drawn.
In another direction also an encouraging start has been made.
There is at present in the Soudan no skilled native labour;
blacksmiths, tinsmiths, carpenters, and bricklayers, are all in
demand, and have to be imported from outside at great trouble and
expense. In the hope of meeting this demand, an industrial school
has been established at the Dockyard Works, which are now on the
river at Omdurman, but which will eventually be moved to Halfaya.
The head of the Steamboat Department agreed to take in sixty boys
as apprentices. These are divided into two shifts of thirty each, and
they alternately receive a day’s schooling and do a day’s practical
work at their separate trades as carpenters, fitters, or riveters. In
school their time is divided between reading, writing, and arithmetic,
and also drawing. The plan has been found to answer admirably. Not
only does the education they receive improve their intelligence as
workers, but some of the boys have shown such proficiency in
drawing that they are able to copy engineering and building designs
with such accuracy as to be able to relieve the English
superintendents of a good deal of work. Many of these boys are
sons of men employed in unskilled labour at the works, who take the
greatest interest in seeing their sons advance so far beyond
themselves. Many applications for admission have to be refused,
and there is little doubt that when Sir W. Mather’s technical school at
the Gordon College is in full swing, it will fill a great need in the
requirements of the country.
On the whole, the most difficult task which the Director of
Education has to face is that of diffusing the elements of knowledge
among the masses of the people. The great distances to be covered
alone impose a tremendous obstacle. But it is extremely important
that at least a portion of the population should be able to understand
the outlines of the machinery of government as laid down in notices
and proclamations, so as to be able to protect themselves against
the exactions of minor officials and the frauds and deceits practised
on them by wandering rogues. It has wisely been determined to
proceed along the lines laid down by Mohammedan tradition. The
kuttabs, or preliminary schools, are a well-known part of the ordinary
religious organization. They are supposed to give instruction in
reading and writing and the Koran, and there are many of them
scattered over the Soudan, as in other Mohammedan countries.
They are, in fact, a sort of private elementary school, something like
the old dames’ schools once existing in parts of England.
Unfortunately, they are almost entirely useless at present. The
teachers are incredibly ignorant. What little instruction they give is
confined to teaching by rote certain passages from the Koran, the
meaning of which is understood by neither pupil nor teacher. The
buildings are usually filthy to the last degree. The idea is to establish
model kuttabs in different parts of the Soudan, and to make them as
efficient as possible, so as to improve the others by their example.
To quote the words of the Director:
 ‘The process of formation has been in all cases the same. With the help of the
Mudir, a suitable building is put up; then the least incompetent sheikh that can be
procured is installed. After confidence has been established, and the nucleus of a
school formed, he is superseded by a trained teacher from Egypt, who, under the
local supervision of the Mudir and occasional supervision from my office, begins to
reduce chaos to order.’

Progress has not been very rapid. Lack of money and lack of
competent schoolmasters sadly hamper all operations. But kuttabs
are now established, attached as a sort of junior class to the schools
at Khartoum, Omdurman, Halfa, and Suakin. A model kuttab has
been established at Berber, which is reported to be doing well, and
another is being built at Dongola. The like is also being attempted at
Wad Medani, a populous town on the Blue Nile, capital of the
province of Sennar, with about 40,000 inhabitants. Reference has
been made to the lack of trained schoolmasters. Egypt itself feels
this difficulty, and Egypt is at present the only source of supply on
which the Soudan can draw. It was to meet this demand—at least,
so far as the kuttabs are concerned—that a small training college for
native sheikhs was opened in the beginning of 1901 in connection
with the school at Omdurman. At first this interesting experiment was
not very successful. The students, who all belonged to the best Arab
families, were all proud, ignorant, and lazy; and as Arabs they were
inclined to despise the Egyptian schoolmasters, whose task it was to
teach them. But now there is a great improvement. They have
increased in number to about thirty, and only lack of room prevents a
further increase. I watched them doing their own lessons, and also
receiving practical instruction in teaching by taking a class of the
school under the guidance of a master. It was impossible to doubt
the value of the experiments. They were nearly all fine-looking,
intelligent young men, some of them really handsome, with the keen,
clear-cut features that mark the pure-bred Arab. Three of them had
come from distant Kassala, where at present there are no means of
education whatever. The course lasts three years. At the end of it
they are to be examined as to their fitness, and they will then be
drafted off either to teach in their kuttabs or else to some posts in
connection with the native Courts. Whether as schoolmasters or
Cadis, they will be most useful elements in the development of the
Soudan.
In the negro portion of the Soudan, inhabited by the pagan tribes,
the people are so backward in civilization that the question of
education does not at present arise, or, if it does arise, assumes a
totally different aspect. Here is the field for the missionary. Two
missions are already established—one, the American Medical
Mission, on the Sobat; and the other, the Austrian Roman Catholic
Mission at Taufikieh, on the White Nile. Both are doing good work,
and both are to be encouraged and assisted by the Government. In
other parts of the Soudan it must be remembered that we are
dealing with a fanatically Mohammedan population, and any
suspicion that the Government was trying to proselytize would
immediately wreck all schemes of education, and probably be the
signal for grave disorders.
It will be a long time before the schools turn out sufficient pupils to
fill the Government Civil Service, and there does not seem to be any
danger of producing mere ‘babus,’ hanging about and relying on a
certain knowledge of English to procure them a job. The teaching of
English is entirely confined to those boys who are going to make use
of it in the Government Service or in commercial pursuits, where its
knowledge is required. For the ordinary mass of the population
nothing is to be gained by an imperfect knowledge of English. The
authorities are unquestionably right in discouraging such teaching;
the supposed political advantages of it are small, if not entirely
imaginary. The Soudan can never be a real white man’s country; its
rulers must always be speakers of Arabic, and its people will do far
better to employ their time in more useful ways than struggling with a
foreign language.
The Gordon College is the centre of education, but even now its
activities are not confined to mere teaching. It is proposed to form a
collection of books dealing with the Soudan, its peoples, its natural
history, and its various productions, accompanied by specimens to
illustrate them, and some progress has been made. Mr. Wellcome’s
valuable bacteriological research laboratory is in full working order,
and a skilled expert from Scotland has been in charge of it for some
months. He is doing a work of great importance, not only to the
Soudan, but also to the scientific world in general. The Soudan is a
land where strange diseases both of men and animals abound.
There is a wide field for research. The scientific and systematic
examination of these obscure subjects is already bearing fruit, and
cannot fail to ameliorate the conditions of life in these tropical
regions for the European as well as for the native.
Lord Cromer has promised that more shall be done for education
in the future. There is no need for hurry; indeed, it is essential that
the educational system shall be built up slowly with caution and
patience. But in time each province will have its own primary and
technical schools, whose pupils will be selected from the elementary
kuttabs. Afterwards, when all this has been carefully organized, the
provincial schools will in their turn pass on their more promising
students to Khartoum to receive the higher education which will then
be demanded. Then the Gordon College will at last become a real
college. In it the germ of a most hopeful future is contained. If the
work proceeds on the same sound lines as hitherto—which there is
no reason to doubt—great days are coming. The Arab is capable of
a very high degree of civilization, and has a great intelligence, which
has as yet had no chance of development. And in those days, I
doubt not, those who founded the Gordon College, and first kindled
the fire of learning, will be praised by its students as sincerely and as
deservedly as the benefactors of any of our own Western
foundations.
 CHAPTER XX

TRADE AND COMMERCE

Sir Rudolf von Slatin, Inspector-General of the Soudan, who

possesses an unrivalled experience of the country, reports that ‘the
whole situation in the country is very satisfactory. Everywhere I went,
from north to south and from east to west, I found that villages and
cultivation had increased. The population is larger and wealthier;
flocks and herds are more numerous; security prevails, and general
satisfaction is expressed with the present rule.’ Once more, or,
rather, nearly for the first time, life and property are safe. Relieved
from the scourge of war and tyranny, people are everywhere
resuming their old occupations. They have even recovered from the
shock of finding themselves under just and settled government, and
are no longer content merely to exist. New wants are being felt, and
with the advance of material prosperity trade and commerce are
springing up.
There are, however, considerable obstacles to the development of
trade. First of all there is a lack of labour. Partly this is due to lack of
population, but partly also to other causes. Slavery has left its mark,
and many of the Arabs are too proud and too lazy to take part in
manual labour; in laziness, though not in pride, many of the negro
tribes are fully their equals. Secondly, there is the difficulty of
communications in so vast a country, and the lack of transport.
Thirdly, the Soudan is very poor, and capital is wanting. Still, every
year shows an improvement in these respects.
The population is steadily growing, partly by natural increase, and
partly by immigration from neighbouring countries of people who had
fled during the rebellion. Attempts have also been made to assist the
increase by colonization. A number of old soldiers from the
Soudanese battalions, who are enlisted for life, were permitted to
retire, and with their wives and children were established in villages
on the Nile and at Kassala. The villages were organized on a more
or less military basis, with a well-known non-commissioned officer as
chief. Each colonist was allotted two or three acres of good rain or
pasture land, or an acre of Nile foreshore. He was given grain for
sowing, besides a quantity of dhurra sufficient to support him until his
crops grew. Markets were also started. Unfortunately, the
experiments were generally unsuccessful. The colonies at Dongola
and Berber failed altogether. It was found that the black when
released from the strict discipline of the regiment was more anxious
to enjoy doing nothing, after the manner of his ancestors, than to
work, and if he saw a chance of living by begging or stealing he was
apt to leave his cultivation alone, and go off to some town.
At Hellet Abbas, on the White Nile, it was found that if the rains
were good the colonists would prepare the ground and sow the crop,
but if it came to artificial irrigation and shadoof work they soon tired
of this heavier labour and left the crops to wither. Happily, a much
better account comes from Kassala. The Mudir of that province
reported:

‘The colony of blacks established at Kassala continues to thrive, and in every

way justifies its existence. They have a well-laid-out village, and are eager to
cultivate along the Gash, and have also a fair amount of rain crops. Labourers can
nearly always be obtained from amongst them for public works, and there are
some very fair masons who are permanently employed. I wish we had the means
of teaching some of them carpentry and blacksmith’s work as well. They have
acquired a good deal of small stock.
‘Those of Gedaref are not so thrifty, and are lazy; they do not cultivate so much
or so well, but I hope for improvement.’

Transport and communications are indeed vastly better than they

were in Egyptian days. The railway from Halfa to Khartoum makes
an all-important difference. It has already given an immense stimulus
to the export trade; without it there would have been quite another
tale to tell of the last five years. But except in the province of Berber
it does not tap any local resources. Great part of it lies in an
uninhabited desert. Every train leaving Halfa has to carry with it
1,520 cubic feet of water. And there are 200 miles of river between
Halfa and the outer world at Assouan, and then 700 miles more to
the sea. Carriage of goods over such a mileage, with its necessary
transhipments, is a long and costly business, nearly prohibitive for
bulky articles. Coal, for example, is seldom less than £4 a ton at
Khartoum, and often nearer £6. No one who had to build a railway in
the Soudan for commercial purposes only would think of crossing the
desert to Halfa; his first thought would be to connect Khartoum with
the sea-board at Suakin.
There is also the railway from Halfa to Kerma, thirty miles from
Dongola. But this, too, was laid down in haste for military purposes.
It is laid so badly, with such sharp curves and such steep gradients,
owing to the nature of the country, that no heavy trains can run on it,
and with the present rolling-stock an engine has sometimes to make
three or four starts before it can master an ascent. It is worked at a
loss of about £20,000 a year, and it has become a serious problem
whether the money so spent could not be employed much more
advantageously elsewhere, so many are the claims on the Soudan
Exchequer. It would be far better to take up the whole line, and relay
it from Dongola to connect with the main-line at Abu Hamed. For not
only is this a much easier country, but the southern part of the
province, which is the richer, and inhabited by a more industrious
and hard-working population, would be opened up. But at present no
money is forthcoming either for this or for the complete repair of the
existing line. So poor Dongola is in the tantalizing position of having
a railway, and yet not being able to take full benefit of it. It is actually
suffering from a surplus of foodstuffs, and is a year behindhand in its
exports.
The province is famous for its dates; not only is it a great exporter
of the ordinary fruit, but it also produces a golden date, which is said
to be the best in the world, better, even, than the Algerine date, so
well known in Europe. It has also abundance of irrigable land as
good as Upper Egypt, which it very much resembles in general
climatic conditions. As its numerous and interesting antiquities show,
it once supported a very large population; but now a great deal of
land is lying waste, and the population, though increasing very fast,
is still not more than about 100,000. With better communications its
prospects are very good. When Egypt becomes overcrowded, as it
must in time if the present rate of increase is maintained, Dongola
will offer a fair field to Egyptian immigration. The conditions of life are
so similar to those in Egypt that it cannot fail to be the most attractive
part of the Soudan. Even as things are, owing to the partial
advantage afforded by the railway and facilities of transport by boat-
carriage on the Nile, the trade of the province is increasing. Cotton
goods and luxuries like tea, sugar, coffee, and perfumeries, are the
principal imports, and cereals are exported as well as dates. The
people are wealthier, and anxious to buy such goods as cutlery,
crockery, soap, agricultural implements, and hardware, but well-to-do
traders have not as yet exploited the field, as they would do if the
railway difficulty were solved.
Once Khartoum is connected with the sea by railway, the principal
obstacle in the way of trade will have been removed, and,
fortunately, this is no longer a mere vision of hopeful men. Practically
the railway has already been begun. It will strike across the desert to
the Atbara, crossing the mountains near Sinkat, and then run along it
to join the main-line near where it crosses that river. If the railway
does not come to Berber itself, Berber will probably travel up the Nile
to meet it. Arab towns are not very difficult to shift. The whole route
has been carefully surveyed, and a good deal has been spent in
improving the port of Suakin. Materials are being rapidly collected.
This time careful preparations are being made; there will be none of
those kaleidoscopic changes of policy which were so fatal in 1885. It
may confidently be expected that in two or three years’ time there
will be something to show very much more substantial than a tennis-
court at Dover built of much-travelled material, the only result of our
former exertions. For the construction of the line the Government will
be its own contractor. The Soudan Railway Department has a very
capable staff, and they will be able to do the work as efficiently as
any outside contractor, and much more cheaply. It is hoped that as
far as possible native labour may be made use of. It is a hopeful
indication of a change of spirit among the Arabs that the local
sheikhs have agreed to bring their tribesmen to work; the experiment
is worth trying even if it fails. No such thing has ever happened in the
Soudan before.
 Meantime Suakin is looking eagerly forward. Its inhabitants have
naturally suffered by the complete diversion of trade to the Nile
Valley route. Very few of them can even afford to repair their houses,
and the town shows signs of decay. Most of the people are
unemployed, and labour is very cheap. There was formerly a good
deal of trade with India and the Red Sea ports, but most of this has
fallen away, and nearly all the Indian merchants who formerly had
their headquarters there have left. But now that the desert which
shuts off the Soudan from the sea is really to be bridged over, there
will be a great change. The one seaport of these immense territories
cannot fail to be a busy and prosperous place.
Once the new line is completed, the distance from Khartoum to
the sea will be reduced by about two-thirds, to some 450 miles, with
a proportionate reduction in expense of carriage, and it will then be
possible to think of building other feeder lines in various parts of the
Soudan. A branch line to Kassala and on to Gedaref and Gallabat
along the Abyssinian frontier will tap a very rich district and open up
the Abyssinian trade. Possibly in the distant future such a line may
be continued southwards so as to connect with Uganda. In the
recent agreement with Abyssinia, powers have been taken to build in
Abyssinian territory for this purpose. But this is still a long way
outside practical politics. Of far more importance for the immediate
development of the country would be a light railway from Omdurman
or Duem to El Obeid, or across the rich Ghezireh from Duem to Wad
Medani on the Blue Nile, or, again, from near Wad Medani to
Gedaref. Easy communication with the sea will render it possible to
bring the necessary plant into the country at a reasonable cost, and
the experience gained as to material and labour in building the
Suakin line will also be invaluable.
Meanwhile the provincial Governors are doing all they can to
improve the caravan routes and roads. In great part of the Soudan
this is simply a question of increasing the number of wells; metalled
roads are scarcely necessary as yet, outside a few towns. Wheel
traffic is almost non-existent; the camel and the ass are the great
public carriers. It is rather strange that the camel has not been more
used for pulling wheel transport than he has. One camel can pull
three or four times as much as he can carry. All the heavy machinery
of the Nile Valley Gold Mining Company, working in Nubia, has been
transported in this way from the river, a distance of sixty miles.
Rarely a camel may be seen hauling a plough in Egypt. But there are
limits to the use of camels. They cannot breed successfully south of
parallel 13°, and in all the country south of this the serut fly makes it
almost impossible for either them or horses to live at certain seasons
of the year.
Above all, access to the sea will greatly stimulate the use made of
water carriage within the country. In the Nile and its tributaries the
Soudan possesses a system of natural trade-routes unequalled in
Africa for internal commerce. The river traffic, though already
growing, is merely in its infancy. The Government has a considerable
fleet of steam and sailing vessels between Wadi Halfa and Assouan,
and also on the Blue and White Niles. They are also encouraging an
English company which has placed some steamers and steam-
barges on both rivers. On the Blue Nile there is regular steamer
communication with Rosaires, 426 miles from Khartoum, during high
Nile, about six months in the year, and even in low Nile most of the
river is navigable by native boats. On the White Nile, now that the
sudd has been cleared, steamers and native boats can ply the whole
year round up to Gondokoro. In time the Sobat may prove a very
good route for trade with Abyssinia. And the Bahr el Ghazal, with its
network of waters, is in this respect the most fortunate of all the
provinces. Many of its waterways are still blocked by sudd, but every
year the navigation is improving. A serviceable channel is now
available on the Jur River as far as Wau, and large steamers will
eventually be able to get up even much further than this. The other
rivers will in time be opened up. In a country where at present
everything has to be carried on men’s heads, this will be an
extraordinary benefit from every point of view, and if the province
answers at all to its old reputation as one of the most fertile spots of
Africa, it will do very well, in spite of mosquitoes, serut flies, malaria,
guinea worm, and all its other plagues.
As for the lack of capital both for private enterprise and public
works, that, like the lack of population, can only be cured by time.
Overhasty development could only do harm, even if it were possible.
As trade improves and agriculture develops, the people will become
more wealthy, and the two will react upon each other. That large
sums will be invested by private capitalists or firms is not to be
expected for years to come. But even now an Englishman is erecting
flour-mills at Wad Medani, and if his venture succeeds it may be
followed by others. The Government is doing all in its power to
encourage agriculture by small loans for the purchase of seeds,
water-wheels, and cattle to work them. Perhaps some form of State-
supported Agricultural Bank will be established.
The Soudan is an agricultural country; in that direction alone can
real progress be made, and the progress depends mainly upon
irrigation. Between the Atbara and the Blue Nile, and between the
Blue Nile and the White Nile, there is, as has been pointed out, a
great field for irrigation works on a large scale, but the amount of
crops now grown by direct irrigation is very small. Along the Blue
Nile water-wheels are numerous for the first ten miles above
Khartoum. There used to be 3,000 between Khartoum and Berber,
but in 1898 there were no more than seventy; and though the
number has increased, it has yet nothing like reached the old level.
The same is true of the rest of Berber Province and of Dongola.
There is also a certain amount of irrigation on islands in the White
Nile above Khartoum. When the river is falling, large mud flats
appear in the centre of the stream. To these the people transport
their cattle and belongings; they sow their seed in the mud, build
themselves huts, and set up shadoofs to water the growing crops of
wheat, barley, dhurra, and onions. They begin operations towards
the end of January, and in good soil the crops are harvested by May.
What perennial irrigation can do is shown by the gardens at
Khartoum, where lemons, figs, oranges, pomegranates, bananas,
vines, and all kinds of vegetables, grow in profusion all the year
round. But at present nearly all the crops are rain crops. The height
to which water has to be raised out of the Blue Nile is too great for
extended irrigation by mere lift. For the first 150 miles from Khartoum
the banks are 26 to 30 feet over the summer level, and further south
33 to 39 feet, the difference between summer and flood level being
about 23 feet. The soil of the Ghezireh is a rich alluvial deposit
sometimes 150 feet deep. The inhabitants build small dykes across
the general slope of the country so as to prevent the rain running off
too quickly, and sow their seed as soon as the rain has fallen. The
result as described by Sir R. Wingate a year or two ago is as follows:

‘I recently rode from Wad Medani on the Blue Nile to opposite Duem on the
White Nile, eighty miles across a perfectly flat plain sown almost throughout its
entire length with dhurra, which was standing 6 to 8 feet high. As there is only one
crop sown during the short rainy season, and as this is planted and harvested
within a period of sixty to eighty days, it follows that, if a system of irrigation were
possible in the Ghezireh it would become a huge granary capable of supplying, not
only the whole Soudan, but other countries as well.’

But, of course, any schemes for utilizing the waters of the Nile
have always to be considered in reference to the prior claims of
Egypt. It is only to Egypt that the Soudan can look for the money
necessary to carry out great works, and, naturally, Egypt would not
allow irrigation to be developed in the Soudan unless her own needs
were amply safeguarded. Whenever Egypt undertakes the great
works contemplated on the Upper Nile, the Soudan will share in the
benefit. That will not be for a long time to come, and meanwhile the
Soudan has opportunity to develop her resources and her population
so as to be able to seize the chance when it comes. There seems to
be no reason, however, why works on a small scale should not be
undertaken, subject always to the question of expense, to utilize in
the Soudan, by means of basin irrigation on the Blue Nile or the
Atbara, some of the flood-water during July, August and September,
which is only an embarrassment to Egypt now that perennial
irrigation is adopted so universally in that country. It is only in the
preceding summer months, when the Nile is low, that Egypt is forced
to watch the proceedings of her neighbour with such a jealous eye. It
would be rash to prophesy exactly when irrigation works will be
undertaken, or what form they will take; but it is certain that they will
come, and when they do, their effect upon the Soudan will be
immense. The subject is of peculiar interest to England; the lands in
question are capable of producing other crops than cereals, and, as
will presently be shown, for none are they more suitable than for
cotton.
Quite apart from any such speculations upon the future, the
external trade of the Soudan is capable of great expansion under
present conditions. An examination of the returns of imports and
exports during the last years before the rebellion is a very useful
guide to the capacities of the country. The figures are taken from
Colonel Stewart’s report.
There are no returns of imports except for the port of Suakin, but
these embrace practically the whole. Of course, a certain amount
came in by the Nile Valley route, just as now a certain amount comes
in by Suakin instead of by the railway. The position has been
practically reversed, and for purposes of comparison it is fair to take
the imports formerly entering the port of Suakin and those now
carried by the railway. The most important item was cotton goods,
which amounted to about 25,000 to 30,000 kantars annually. Linens
were about 200 to 300 kantars, and woollens 100 to 300 kantars.
There was also a certain amount of silks and silk thread and sewing
cotton. The Indian trade brought in a good deal of grain and tobacco.
In 1881 the import of Indian rice amounted to 20,000 kantars.
Petroleum (6,000 kantars), oil, zinc, copper, and iron, appear in the
list, as well as flour and provisions of various kinds in large
quantities, and candles, boots, and clothes. The peculiar tastes of
the Soudanese in luxuries are reflected in the large imports of
sandal-wood, and scents and perfumery, especially fish-scent (700
tons, or 15,400 kantars, a year). Both men and women are
particularly fond of strong, greasy scents. In all native festivities and
entertainments these scents are a very prominent feature, and a
native marriage can be smelt a very long way off. Last, but not least,
comes a very peculiar item: in one year no less than 11,000 dozen
umbrellas were imported. There is certainly enough sunshine in the
Soudan, and in parts of it rain also, to justify a large number of
umbrellas, but it is difficult to believe that 132,000 umbrellas would
find a ready sale in the Soudan to-day. Perhaps the Khalifa
particularly disliked the umbrella-carrying class.