Navier Stokes equations and turbulence 1st Edition C. Foias

maggyzteven 7 views 85 slides Feb 27, 2025
Slide 1
Slide 1 of 85
Slide 1
1
Slide 2
2
Slide 3
3
Slide 4
4
Slide 5
5
Slide 6
6
Slide 7
7
Slide 8
8
Slide 9
9
Slide 10
10
Slide 11
11
Slide 12
12
Slide 13
13
Slide 14
14
Slide 15
15
Slide 16
16
Slide 17
17
Slide 18
18
Slide 19
19
Slide 20
20
Slide 21
21
Slide 22
22
Slide 23
23
Slide 24
24
Slide 25
25
Slide 26
26
Slide 27
27
Slide 28
28
Slide 29
29
Slide 30
30
Slide 31
31
Slide 32
32
Slide 33
33
Slide 34
34
Slide 35
35
Slide 36
36
Slide 37
37
Slide 38
38
Slide 39
39
Slide 40
40
Slide 41
41
Slide 42
42
Slide 43
43
Slide 44
44
Slide 45
45
Slide 46
46
Slide 47
47
Slide 48
48
Slide 49
49
Slide 50
50
Slide 51
51
Slide 52
52
Slide 53
53
Slide 54
54
Slide 55
55
Slide 56
56
Slide 57
57
Slide 58
58
Slide 59
59
Slide 60
60
Slide 61
61
Slide 62
62
Slide 63
63
Slide 64
64
Slide 65
65
Slide 66
66
Slide 67
67
Slide 68
68
Slide 69
69
Slide 70
70
Slide 71
71
Slide 72
72
Slide 73
73
Slide 74
74
Slide 75
75
Slide 76
76
Slide 77
77
Slide 78
78
Slide 79
79
Slide 80
80
Slide 81
81
Slide 82
82
Slide 83
83
Slide 84
84
Slide 85
85

About This Presentation

Navier Stokes equations and turbulence 1st Edition C. Foias
Navier Stokes equations and turbulence 1st Edition C. Foias
Navier Stokes equations and turbulence 1st Edition C. Foias


Slide Content

Visit https://ebookultra.com to download the full version and
explore more ebooks
Navier Stokes equations and turbulence 1st
Edition C. Foias
_____ Click the link below to download _____
https://ebookultra.com/download/navier-stokes-
equations-and-turbulence-1st-edition-c-foias/
Explore and download more ebooks at ebookultra.com

Here are some suggested products you might be interested in.
Click the link to download
The Navier Stokes Equations An Elementary Functional
Analytic Approach 1st Edition Hermann Sohr (Auth.)
https://ebookultra.com/download/the-navier-stokes-equations-an-
elementary-functional-analytic-approach-1st-edition-hermann-sohr-auth/
Turbulence in the Atmosphere 1st Edition John C. Wyngaard
https://ebookultra.com/download/turbulence-in-the-atmosphere-1st-
edition-john-c-wyngaard/
Turbulence Structure and Vortex Dynamics 1st Edition J. C.
R. Hunt
https://ebookultra.com/download/turbulence-structure-and-vortex-
dynamics-1st-edition-j-c-r-hunt/
Memoir and Scientific Correspondence of the Late Sir
George Gabriel Stokes Bart 1st Edition George Gabriel
Stokes
https://ebookultra.com/download/memoir-and-scientific-correspondence-
of-the-late-sir-george-gabriel-stokes-bart-1st-edition-george-gabriel-
stokes/

Memoir and Scientific Correspondence of the Late Sir
George Gabriel Stokes Bart 1st Edition George Gabriel
Stokes
https://ebookultra.com/download/memoir-and-scientific-correspondence-
of-the-late-sir-george-gabriel-stokes-bart-1st-edition-george-gabriel-
stokes-2/
Race and Human Rights 1st Edition Curtis Stokes
https://ebookultra.com/download/race-and-human-rights-1st-edition-
curtis-stokes/
Partial Differential Equations 2nd Edition Lawrence C.
Evans
https://ebookultra.com/download/partial-differential-equations-2nd-
edition-lawrence-c-evans/
An Introduction to Ordinary Differential Equations James
C. Robinson
https://ebookultra.com/download/an-introduction-to-ordinary-
differential-equations-james-c-robinson/
Non equilibrium Statistical Mechanics and Turbulence 1st
Edition John Cardy
https://ebookultra.com/download/non-equilibrium-statistical-mechanics-
and-turbulence-1st-edition-john-cardy/

Navier Stokes equations and turbulence 1st Edition C.
Foias Digital Instant Download
Author(s): C. Foias, O. Manley, R. Rosa, R. Temam
ISBN(s): 9780521360326, 0521360323
Edition: 1
File Details: PDF, 1.72 MB
Year: 2001
Language: english

This page intentionally left blank

Navier–Stokes Equations and Turbulence
This book aims to bridge the gap between practicing mathematicians and the practitioners of
turbulence theory. It presents the mathematical theory of turbulence to engineers and physi-
cists as well as the physical theory of turbulence to mathematicians. The book is the result
of many years of research by the authors,who analyze turbulence using Sobolev spaces and
functional analysis. In this way the authors have recovered parts of the conventional theory
of turbulence,deriving rigorously from the Navier–Stokes equations what had been arrived at
earlier by phenomenological arguments.
The mathematical technicalities are kept to a minimum within the book,enabling the dis-
cussion to be understood by a broad audience. Each chapter is accompanied by appendices
that give full details of the mathematical proofs and subtleties. This unique presentation should
ensure a volume of interest to mathematicians,engineers,and physicists.
Ciprian Foias is an Emeritus Professor in the Department of Mathematics at Indiana University
at Bloomington and Professor of Mathematics at Texas A&M University at College Station. He
has held numerous visiting professorships,including those at Virje University (Netherlands),
Israel Institute of Technology,University of California at San Diego,Université Paris-Sud,and
the Collège de France. In 1995,he was awarded the Norbert Wiener prize by the American
Mathematical Society.
Oscar P. Manley works as a consultant and independent researcher on the foundations of tur-
bulent flows. He has acted as head of the U.S. Department of Energy’s Engineering Research
Program and as the Program Manager for the Department of Energy’s research on magnetic
fusion theory. Dr. Manley has held visiting professorships at the Université Paris-Sud and
Indiana University.
Ricardo Rosa is a Professor of Mathematics at the Universidade Federal do Rio de Janeiro. He
has also held positions as Visiting Researcher at the Institute for Scientific Computing and Ap-
plied Mathematics at Indiana University and as Visiting Professor at the Université Paris-Sud
in Orsay,France.
Roger Temam is a Professor of Mathematics at the Université Paris-Sud and Senior Scientist at
the Institute for Scientific Computing and Applied Mathematics at Indiana University. He has
been awarded an Honorary Professorship at Fudan University (Shanghai),the French Acad-
emy of Science’s Grand Prix Alexandre Joannidès,and the Seymour Cray Prize in Numerical
Simulation. Professor Temam has authored or co-authored nine books and published more
than 260 articles in international refereed journals. His current research interests in fluid me-
chanics are in the areas of control of turbulence,boundary layer theory,and geophysical fluid
dynamics.

ENCYCLOPEDIA OF MATHEMATICS AND ITS APPLICATIONS
EDITED BY G.-C. ROTA
Editorial Board
R. Doran,P. Flajolet,M. Ismail,T.-Y. Lam,E. Lutwak
Volume 83
Navier–Stokes Equations and Turbulence
6 H. MincPermanents
19 G. G. Lorentz,K. Jetter,and S. D. RiemenschneiderBirkhoff Interpolation
22 J. R. BastidaField Extensions and Galois Theory
23 J. R. CannonThe One-Dimensional Heat Equation
24 S. WagonThe Banach–Tarski Paradox
25 A. SalomaaComputation and Automata
27 N. H. Bingham,C. M. Goldie,and J. L. TeugelsRegular Variation
28 P. P. Petrushev and V. A. PopovRational Approximation of Real Functions
29 N. White (Ed.)Combinatorial Geometries
30 M. Pohst and H. ZassenhausAlgorithmic Algebraic Number Theory
31 J. Aczel and J. DhombresFunctional Equations in Several Variables
32 M. Kuczma,B. Choczewski,and R. GerIterative Functional Equations
33 R. V. AmbartzumianFactorization Calculus and Geometric Probability
34 G. Gripenberg,S.-O. Londen,and O. StaffansVolterra Integral and Functional Equations
35 G. Gasper and M. RahmanBasic Hypergeometric Series
36 E. TorgersenComparison of Statistical Experiments
38 N. KorneichukExact Constants in Approximation Theory
39 R. Brualdi and H. RyserCombinatorial Matrix Theory
40 N. White (Ed.)Matroid Applications
41 S. SakaiOperator Algebras in Dynamical Systems
42 W. HodgesBasic Model Theory
43 H. Stahl and V. TotikGeneral Orthogonal Polynomials
45 G. Da Prato and J. ZabczykStochastic Equations in Infinite Dimensions
46 A. Björner et al.Oriented Matroids
47 G. Edgar and L. SuchestonStopping Times and Directed Processes
48 C. SimsComputation with Finitely Presented Groups
49 T. PalmerBanach Algebras and the General Theory of *-Algebras
50 F. BorceuHandbook of Categorical Algebra 1
51 F. BorceuHandbook of Categorical Algebra 2
52 F. BorceuHandbook of Categorical Algebra 3
53 V. F. KolchinRandom Graphs
54 A. Katok and B. HasselblattIntroduction to the Modern Theory of Dynamical Systems
55 V. N. SachkovCombinatorial Methods in Discrete Mathematics
56 V. N. SachkovProbabilistic Methods in Discrete Mathematics
57 P. M. CohnSkew Fields
58 R. GardnerGeometric Topography
59 G. A. Baker,Jr. and P. Graves-MorrisPadé Approximants
60 J. KrajicekBounded Arithmetic Propositional Logic and Complexity Theory
61 H. GroemerGeometric Applications of Fourier Series and Spherical Harmonics
62 H. O. FattoriniInfinite Dimensional Optimization and Control Theory
63 A. C. ThompsonMinkowski Geometry
64 R. B. Bapat and T. E. S. RaghavanNonnegative Matrices with Applications
65 K. EngelSperner Theory
66 D. Cvetkovic,P. Rowlinson,and S. SimicEigenspaces of Graphs
67 F. Bergeron,G. Labelle,and P. LerouxCombinatorial Species and Tree-Like Structures
68 R. Goodman and N. R. WallachRepresentations and Invariants of the Classical Groups
69 T. Beth,D. Jungnickel,and H. LenzDesign Theory,vol. 1
70 A. Pietsch and J. WenzelOrthonormal Systems for Banach Space Geometry
71 G. E. Andrews,R. Askey,and R. RoySpecial Functions
72 R. TicciatiQuantum Field Theory for Mathematicians
73 M. SternSemimodular Lattices
74 I. Lasiecka and R. TriggianiControl Theory for Partial Differential Equations I
75 I. Lasiecka and R. TriggianiControl Theory for Partial Differential Equations II
76 A. A. IvanovGeometry of Sporadic Groups 1
77 A. SchinzelPolynomials with Special Regard to Reducibility
78 H. Lenz,T. Beth,and D. JungnickelDesign Theory,vol. 2

encyclopedia of mathematics and its applications
Navier–Stokes Equations and Turbulence
C. FOIAS O. MANLEY
R. ROSA R. TEMAM

         
The Pitt Building, Trumpington Street, Cambridge, United Kingdom
  
The Edinburgh Building, Cambridge CB2 2RU, UK
40 West 20th Street, New York, NY 10011-4211, USA
477 Williamstown Road, Port Melbourne, VIC 3207, Australia
Ruiz de Alarcón 13, 28014 Madrid, Spain
Dock House, The Waterfront, Cape Town 8001, South Africa
http://www.cambridge.org
First published in printed format
ISBN 0-521-36032-3 hardback
ISBN 0-511-03936-0 eBook
Cambridge University Press 2004
2001
(netLibrary)
©

Contents
Preface pageix
Acknowledgments xiv
ChapterIIntroductionandOverviewofTurbulence 1
Introduction 1
1.ViscousFluids.TheNavier–StokesEquations 1
2. Turbulence: Where the Interests of Engineers and Mathematicians
Overlap 5
3. Elements of the Theories of Turbulence of Kolmogorov and
Kraichnan 9
4.FunctionSpaces,FunctionalInequalities,andDimensionalAnalysis 14
Chapter II Elements of the Mathematical Theory of the
Navier–StokesEquations 25
Introduction 25
1.EnergyandEnstrophy 27
2.BoundaryValueProblems 29
3.Helmholtz–LerayDecompositionofVectorFields 36
4.WeakFormulationoftheNavier–StokesEquations 39
5.FunctionSpaces 41
6.TheStokesOperator 49
7.ExistenceandUniquenessofSolutions:TheMainResults 55
8.AnalyticityinTime 62
9.GevreyClassRegularityandtheDecayoftheFourierCoefficients 67
10.FunctionSpacesfortheWhole-SpaceCase 75
11.TheNo-SlipCasewithMovingBoundaries 77
12.DissipationRateofFlows 80
13.NondimensionalEstimatesandtheGrashofNumber 87
AppendixA.MathematicalComplements 90
AppendixB.ProofsofTechnicalResultsinChapterII 102
vii

viii Contents
ChapterIIIFiniteDimensionalityofFlows 115
Introduction 115
1.DeterminingModes 123
2.DeterminingNodes 131
3.AttractorsandTheirFractalDimension 137
4.ApproximateInertialManifolds 150
AppendixA.ProofsofTechnicalResultsinChapterIII 156
Chapter IV Stationary Statistical Solutions of the Navier–Stokes
Equations,TimeAverages,andAttractors 169
Introduction 169
1. Mathematical Framework,Definition of Stationary Statistical
Solutions,andBanachGeneralizedLimits 172
2. Invariant Measures and Stationary Statistical Solutions in
Dimension2 183
3.StationaryStatisticalSolutionsinDimension3 189
4.AttractorsandStationaryStatisticalSolutions 194
5.AverageTransferofEnergyandtheCascadesinTurbulentFlows 198
AppendixA.NewConceptsandResultsUsedinChapterIV 218
AppendixB.ProofsofTechnicalResultsinChapterIV 227
Appendix C. A Mathematical Complement: The Accretivity Property
inDimension3 244
Chapter V Time-Dependent Statistical Solutions of the
Navier–Stokes Equations and Fully Developed
Turbulence 255
Introduction 255
1.Time-DependentStatisticalSolutionsonBoundedDomains 262
2.HomogeneousStatisticalSolutions 271
3.ReynoldsEquationfortheAverageFlow 280
4.Self-SimilarHomogeneousStatisticalSolutions 283
5. Relation with and Application to the Conventional Theory of
Turbulence 295
6.SomeConcludingRemarks 310
AppendixA.ProofsofTechnicalResultsinChapterV 312
References 331
Index 343

Preface
This monograph isan attempt to address the theory of turbulence from the points of
view of several disciplines. The authors are fully aware of the limited achievements
here as compared with the task of understanding turbulence. Even though necessar-
ily limited,the results in this book benefit from many years of work by the authors
and from interdisciplinary exchanges among them and between them and others. We
believe that it can be a useful guide on the long road toward understanding turbulence.
One of the objectives of this book is to let physicists and engineers know about the
existing mathematical tools from which they might benefit. We would also like to
help mathematicians learn what physical turbulence is about so that they can focus
their research on problems of interest to physics and engineering as well as mathe-
matics. We have tried to make the mathematical part accessible to the physicist and
engineer,and the physical part accessible to the mathematician,without sacrificing
rigor in either case. Although the rich intuitionof physicists and engineers has served
well to advance our still incomplete understanding of the mechanics of fluids,the rig-
orous mathematics introduced herein will serve to surmount the limitations of pure
intuition. The work is predicated on the demonstrable fact that some of the abstract
entities emerging from functional analysis of the Navier–Stokes equations represent
real,physical observables: energy,enstrophy,and their decay with respect to time.
Beside this didactic objective,one of our scientific goals – in this book and in its
underlying research – was to see what we can learn about the physical properties of
turbulence using Sobolev spaces and the functional analysis methods that are based
on them. As we subsequently show,these spaces – which seem to be abstract mathe-
matical inventions – are in fact representations of observable physical quantities. In
this way we have recovered several parts of the conventional theory of turbulence,de-
riving rigorously from the Navier–Stokes equations (NSE) what had been arrived at
earlier by phenomenological arguments (Kolmogorov [1941a,b]), but in addition we
derive new results. We have shown that the conventional estimate of the number of de-
grees of freedom in homogeneous,isotropic turbulence (viz.,(Reynolds number)
9/4
)
is at best an upper bound on the number of degrees of freedom needed for numeri-
cal simulations of real flows. We have also provided a rigorous,mathematical way
to avoid the common underlying assumption of the ergodicity of turbulent flows. In
ix

x Preface
fact,we show that (in a suitable sense) time averages of various turbulent flow proper-
ties equal the related ensemble averages with respect to adequate statistical solutions;
we have also found a means for removing the high-wavenumber components of the
flow in such a way as to yield an effective viscosity,while providing a rough upper
bound on the error committed relative to the true solution of the flow equation.
Another task,the second scientific objective of this book,was to make the con-
nection between three of the classical approaches to turbulence: the Navier–Stokes
equations; the dynamical systems approach (following the work and ideas of Lorenz
[1963],Smale [1967],and Ruelle and Takens [1971]); and the conventional statistical
theory of turbulence (following the works and ideas of Kolmogorov [1941a,b, 1962],
Batchelor [1959],Kraichnan [1967],and others – e.g.,Landau and Lifshitz [1971]
and Monin and Yaglom [1975]). Before the researchunderlying the material pre-
sented here,these classical approaches evolved largely independently. In particular,
the conventional theoryof turbulence is based mostly on dimensional phenomeno-
logical arguments that traditionally make little referenceto the NSE (see Tennekes
and Lumley [1972]). However,we believe it is useful and instructive to show that
many known results can be directly derived from the Navier–Stokes equations. We
develop those connections to the widestpossible extent.
The level of mathematical preparation necessary for understanding this material is
an elementary knowledge of partial differential equations and their solutions in terms
of eigenfunction expansions. Terms and concepts beyond that level are presented in
detail as needed. Also included is a brief tutorial on Sobolev spaces and inequalities.
To aid readers unfamiliar with some useful classical inequalities,they are presented
(without proof) in Chapter I along with the tutorial.
Mathematically oriented readers are assumed to be familiar with elementary physics
and continuum mechanics,including such principles as conservation of momentum
and energy and the relationship between stress and strain. For their benefit,Chapter I
contains also a short tutorial on the Kolmogorov (conventional) theory of turbulence.
One of the unresolved difficulties encountered in this monograph is due to limita-
tions in the present stage of the mathematical theory of the NSE; the theory is fairly
complete in the 2-dimensional case but still incomplete in dimension 3. Thus,while
we realize that natural turbulence is usually 3-dimensional,here we sometimes em-
phasize 2-dimensional flows,which are fully within the grasp of modern methods of
functional analysis.
The wordturbulencehas different meanings to different people,which indicates
that turbulence is a complex and multifaceted phenomenon. For mathematicians,
outstanding problems revolve around the Navier–Stokes equations (such as well-
posedness and low-viscosity behavior,especially in the presence of walls or singular
vortices). For physicists,major questions include ergodicity and statistical behavior
as related to statistical mechanics of turbulence. Engineers would like responses to
questions simple to articulate but amazingly difficult to answer: What are the heat
transfer properties of a turbulent flow? What are the forces applied by a fluid to its
boundary (be it a pipe or an airfoil)? To others pursuing the dynamical system ap-
proach,of interest is the large time behavior of the flow. Another ambitious question

Preface xi
for engineers is the control of turbulence (to either reduce or enhance it),which is al-
ready within reach. Finally,a major goal in turbulence research – of interest to all
and toward which progress is constantly made – is trustworthy and reliable compu-
tation of turbulent flows (see e.g. Orszag [1970] and Ferziger,Mehta,and Reynolds
[1977]).
We do not address here any computational aspects,although this problem is very
much present in our thoughts; neither do we address control problems,nor most of the
practical engineering problems (see Schlichting [1960]). After the introductory and
tutorial Chapter I,the core of the book consists of four chapters,Chapters II–V. Each
of them,in addressing a particular topic,could actually be developed into a whole
independent volume. Chapter II summarizes some classical and some more recent
aspects of the mathematical theory of the Navier–Stokes equations – namely,their
formulation and well-posedness. We start by presenting the physical background of
the mathematical theory,introducing kinetic energy and enstrophy,conservation of
kinetic energy,and the Helmholtz–Leray decomposition of vector fields. We present
function spaces,the spaces of finite kinetic energy and finite enstrophy vector func-
tions,as well as some additional related abstract spaces. After recalling the weak
formulation of the NSE,a starting pointof their mathematical theory going back to
the work of Jean Leray in the early 1930s,we recall the main theorems of existence,
uniqueness,and regularity of solutions. Then we describe analyticity properties of
the solutions; first,analyticity in time,which is sometimes related to intermittency (a
question briefly addressed in Sections 6.2 and 6.3 of Chapter V); and second,analytic-
ity in space and time (Gevrey class regularity),which is related in the space-periodic
case to the decline of Fourier coefficients of the solution. Finally,we briefly discuss
the no-slip case with moving boundaries and establish properties of the rate of dissi-
pation of flows.
Chapter III revolves around the idea (hinted at long ago by Landau and Lif-
shitz) that,in the permanent regime,turbulent flows as solutions of NSE are finite-
dimensional. This concept,which in fact follows easily from the Kolmogorov ap-
proach to turbulence,was novel in its time; by now it has been substantiated in many
different ways and extended as well to other equations modeling other physical phe-
nomena. In Chapter III we discuss finite dimensionality of turbulent flows in the
context of determining modes and nodes,showing that such flows are fully deter-
mined by either a finite (sufficiently large) number of modes or a finite number of
observation points (nodes). We discuss also the large time behavior in the context
of attractors and show finite dimensionality of attractors; all these dimensions are
physically relevant and related to the Landau–Lifshitz estimates.We briefly discuss
approximate inertial manifolds,the initial point for multilevel numerical algorithms
under development; in some sense,these algorithms produce in time what multigrid
or wavelet methods produce in space. Chapter IV comes closest to the issue of er-
godicity. We introduce,in space dimensions 2 and 3,stationary statistical solutions
and relate them to the limits of time averages. We consider also the corresponding in-
variant measure and relate it to the attractor that carries it. We then apply these tools
to the study of the cascade processes in turbulent flows.

xii Preface
Finally,in Chapter V,we study the concept of statistical solutions,the evolution
of the probability distribution of the flow,and homogeneous flows. We start by in-
troducing the time-dependent statistical solutions on bounded domains. Then we
introduce the (space-invariant) homogeneous statistical solutions for space-periodic
flows and flows in the whole spaces. The Reynolds averaged equations are intro-
duced,and we then discuss self-similar homogeneous statistical solutions (SSHSS);
we introduce a 2-parameter family of such solutions from which,on the one hand,we
resolve a paradox on SSHSS pointed out by Hopf [1952] and,on the other hand,we
recover and complete some elements of the conventional theory of turbulence. For
instance,we show how the Kolmogorov spectrum follows naturally from NSE and
how the intermittency of turbulent flows is related to the fractal nature (see Novikov
and Stewart [1964] and Mandelbrot [1982]) of energy dissipation in 3-dimensional
flows.
As with all interdisciplinarywork,it is not easy to write a book that is readable
by (and of equal interest to) people with differing perspectives. In order to overcome
this difficulty,we have divided each of the main chapters into two parts: the main
one,in which we hope the language is understandable by all,contains as few math-
ematical technicalities as possible yetstill states the results in a rigorous way. Then,
as needed,a long appendix gives the details of the proofs.
The reader should note that some of the cited original articles underlying this
monograph may treat the same problem in two distinct publications: a more phys-
ically oriented treatment appearing in a physics or mechanics journal as well as a
corresponding “heavy” mathematical treatment presented in a mathematics journal.
That is clearly due to the idiosyncrasies of the two kinds of publications and the need
for different presentation styles when addressing the different audiences.
A few remarks will conclude this Preface. First,the authorsare fully aware that
this book is difficult to read because,owing to the nature of the subject,it assumes
the reader’s familiarity with several distinct areas of knowledge. The three senior au-
thors hope that the younger generation,more accustomed to interdisciplinary work
than their predecessors,will find this work more readily accessible than will their
elders. In that regard,the three senior authors are delighted that their younger col-
league (RR) had agreed to involve himself so deeply inallthe aspects of this book,
and they hope that this bodes well for its future – especially insofar as its interest and
accessibility to the younger generation are concerned.
Second,on the anecdotal side,we recall briefly the genesis of this interdisciplinary
collaboration. For a number of years CF and RT had worked independently on the
analysis of the Navier–Stokes equations; CF had learned thesubject from Jacques-
Louis Lions and Giovanni Prodi; he collaborated with Prodi and started to develop a
rigorous theory of statistical solutions of those equations. RT learned the subject from
Jacques-Louis Lions and Jean Leray,and he also worked on the stochastic solutions
of the NSE. Then CF and RT met in the summer of 1970 at a meeting – organized by
Giovanni Prodi – in Varenna (Italy),and CF visited RT at Orsay (France) in the fall
of 1974. Their collaboration started,addressing such different aspects of the NSE as
analysis,statistical solutions,and the long time behavior (dynamical systems point

Preface xiii
of view). At some point RT suggested that their collaboration would become more
interesting if they could join forces with a physicist.
By chance in the spring of 1980,Jacqueline Mossino,a former student of RT,met
Yvain Trève (OM’s co-worker) at a conference on plasma physics in Tucson (Ari-
zona),and contact was established by letter at a time predating e-mail; eventually
they met face-to-face for the first time at a meeting in Dekalb (Illinois) in 1981. At
that time,OM was working with Trève on finite–mode number approximations of
thermal convection satisfying the first and second laws of thermodynamics and dis-
covered that the qualitative nature of the numerical results depended critically on the
number of modes retained (Trève and Manley [1982]). As they started to interact,
CF,OM,and RT realized immediately the extent of the common ground between the
two communities and perspectives that theyrepresent. That realization was the orig-
inal stimulus for much of the research reported in this volume. More specifically,the
direction of that research was set by the recognition that a simple,physically based
argument (conservation ofenergy and momentum in thermal convection; Trève and
Manley [1981]) yielded a result – a bound on the sufficient number of degrees of free-
dom for this fluid flow – that isessentially equivalent to an elaborate mathematical
exercise in Sobolev spaces (Foias,Manley,Temam,and Trève [1983b]). This collab-
oration has extended through the rest of the 1980s,the 1990s,and beyond.
The youngest author was a graduate studentat Indiana University from 1992 to
1996,and he had many opportunities to be exposed to this research through courses,
informal discussions,and seminar lectures. He enthusiastically agreed to participate
in this book,which has been in process for a number of years,and eventually started
to collaborate on more recent works. As indicated earlier,the three senior authors are
delightedthat RR has joined them in this task,and they see it as a good omen for a
successful transmittal of these results to the next generation.
Beside the prolonged and extended efforts of the four authors,this book has bene-
fitted extensively from the input and influence of many others by occasional collab-
orations,discussions,and other forms of interaction. It is not possible to name them
all,but we want to thank them for their constructive influence on us. Also,we would
like to extend our deepest thanks to those who have co-authored relevant publications
with one or more of the authors of this book,works that are partially or fully reported
in this monograph: Hari Bercovici,Peter Constantin (with whom three of us had an
extended collaboration),Arnaud Debussche,Jean-MichelGhidaglia,David Gottlieb,
Martine Marion,Jean-Claude Saut,George Sell,Denis Serre,Edriss Titi,and Yvain
Trève.
Finally,the authors are very grateful to David Tranah and Alan Harvey of Cam-
bridge University Press for their interest,their encouragement,and their great pa-
tience in waiting for the delivery of the manuscript.
Ciprian Foias Oscar Manley
Ricardo Rosa Roger Temam

Acknowledgments
The research presented in this book has been supported along the years by grants from:
the National Science Foundation NSF-DMS,8802596,9024769,9400615,9705229,
9706903,CDA-9601632; the U.S. Department of Energy,Office of Scientific Com-
puting,DE-AC02-82ER12049,DE-FG02-86ER25020,DE-FG02-92ER25120; the
Air Force Office of Scientific Research AFOSR-88-103; and the Office of Naval Re-
search,N00014-91-J-1140,N00014-96-1-0-425.
This research was also partially supported by: the Research Fund of Indiana Univer-
sity; the Université de Paris-Sud and the Centre Nationalde la Recherche Scientifique
(CNRS) through the Laboratoire d’Analyse Numérique d’Orsay; CNPq (Brasília,
Brazil),FUJB,and FAPERJ (Rio de Janeiro,Brazil); the Institute for Mathematics
and Its Applications,University of Minnesota; and the Center for Nonlinear Studies,
Los Alamos National Laboratory,New Mexico.
OM gratefullyacknowledges the understanding and encouragement of his efforts
shown by his former supervisor,Dr. J. Coleman,at the U.S. Department of Energy.
Furthermore,he wishes to express his deep appreciation to the Laboratoire d’Analyse
Numérique d’Orsay at Université de Paris-Sud,the Department of Mathematics at
Indiana University,and Ecole Central de Lyon,where he has been a visitor on many
inspiring and productive occasions.
xiv

I
Introduction and
Overview of Turbulence
Introduction
In this chapter we first briefly recall, in Section 1, the derivation of the Navier–Stokes
equations (NSE) starting from the basic conservation principles in mechanics: con-
servation of mass and momentum. Section 2 contains some general remarks on turbu-
lence, and it alludes to some developments not presented in the book. For the benefit
of the mathematically oriented reader (and perhaps others), Section 3 provides a fairly
detailed account of the Kolmogorov theory of turbulence, which underlies many parts
of Chapters III–V. For the physics-oriented reader, Section 4 gives an intuitive intro-
duction to the mathematical perspective and the necessary tools. A more rigorous
presentation appearsin the first half of Chapter II and thereafter as needed. Foreach
of the aspects that we develop, the present chapter should prove more useful for the
nonspecialist than for the specialist.
1 Viscous Fluids. The Navier–Stokes Equations
Fluids obey the general laws of continuum mechanics: conservation of mass, energy,
and linear momentum. They can be written as mathematical equations once a repre-
sentation for the state of a fluid is chosen. In the context of mathematics, there are
two classical representations. One is the so-called Lagrangian representation, where
the state of a fluid “particle” at a given time is described with reference to its ini-
tial position. The other representation (adopted throughout this book) is the so-called
Eulerian representation, where at each timetand positionxin space the state – in par-
ticular, the velocityu(x,t)– of the fluid “particle” at that position and time is given.
In the Eulerian representation of the flow, we also represent the densityρ(x,t)as
a function of the positionxand timet.The conservation of mass is expressed by the
continuity equation
∂ρ
∂t
+div(ρu)=0. (1.1)
The conservation of momentum is expressed in terms of the accelerationγand the
Cauchy stress tensorσ:
ργ
i=
3
.
j=1
∂σij
∂xj
+fi,i=1,2,3. (1.2)
1

2 I Introduction and Overview of Turbulence
Hereγ=(γ 1,γ2,γ3)andσ=(σ ij)i,j=1,2,3,componentwise in the 3-dimensional
case. Moreover,f=(f
i,f2,f3)represents volume forces applied to the fluid.
The acceleration vectorγ=γ(x,t)of the fluid at positionxand timetcan be
expressed, using purely kinematic arguments, by the so-called material derivative
γ=
DuDt
=
∂u
∂t
+(u·∇)u, (1.3)
or, componentwise,
γ
i=
∂u
i
∂t
+
3

j=1
uj
∂ui
∂xj
,i=1,2,3.
Inserting this expression into the left-hand side (LHS) of equation (1.2) yields the
termρ(u·∇)u,which is the only nonlinear term in theNavier–Stokes equations;
this term is also called theinertial term.The Navier–Stokes equations are among the
very few equations of mathematical physics for which the nonlinearity arises not from
the physical attributes of the system but rather from the mathematical (kinematical)
aspects of the problem.
Further transformations of the conservation of momentum equation necessitate
additional physical arguments and assumptions. Rheology theory relates the stress
tensor to the velocity field for different materials through the so-called stress–strain
law and other constitutive equations. Assuming the fluid is Newtonian, which is the
case of interest to us, amounts to assuming that the stress–strain law is linear. More
precisely, for Newtonian fluids the stress tensor is expressed in terms of the velocity
field by the formula
σ
ij=µ
+
∂u
i
∂xj
+
∂u
j
∂xi
=
+(λdivu−p)δ
ij, (1.4)
wherep=p(x,t)is the pressure. Here,δ
ijis the Kronecker symbol andµ, λare
constants. The constantµis called the shear viscosity coefficient, and 3λ+2µis the
dilation viscosity coefficient. For thermodynamical reasons,µ>0 and 3λ+2µ≥
0.Inserting the stress–strain law (1.4) into the momentum equation (1.2), we obtain
ρ
+
∂u
∂t
+(u·∇)u
=
=µu+(µ+λ)∇divu−∇p+f. (1.5)
Equations (1.1) and (1.5) govern the motion of compressible Newtonian fluids such
as the air at high speeds (Mach number larger than 0.5). If we also assume that the
fluid is incompressible and homogeneous, then the density is constant in space and
time:ρ(x,t)≡ρ
0.In this case, the continuity equationis reduced to thedivergence-
freecondition:
divu=0. (1.6)
Because the density is constant, we may divide the momentum equation (1.5) byρ
and consider the so-called kinematic viscosityν=µ/ρ
0;we may then replace the
pressurepand the volume forcefby the kinetic pressurep/ρ
0and the mass den-
sity of body forcesf/ρ
0,respectively. In doing so, and taking into consideration the

1 Viscous Fluids. The Navier–Stokes Equations 3
divergence-free condition (1.6), we obtain theNavier–Stokes equations for a viscous,
incompressible, homogeneous flow:
∂u
∂t
−ˆ}u+(u·∇)u+∇p=f, (1.7a)
∇·u=0, (1.7b)
where, for notational simplicity, we represent the divergence ofuby∇·u.For all
pratical purposes, the density has actually been normalized to unity; even so, we may
sometimes replace (1.7a) by (1.5), remembering then that∇·u=0 andρis constant.
For more details on the physical aspects of fluid mechanics, we refer the reader to
the classical books of Batchelor [1988] and Landau and Lifshitz [1971].
It is readily accepted that the Navier–Stokes equations govern the motion of com-
mon fluids such as air or water, so we are faced with the persistent challenging ques-
tion of recovering from (1.7) such complex motions as that of smoke dispersion in
the air and the turbulent flow of a river around a bridge pillar.
The flow of fluids at the microscopic level is governed by phenomena in the realm
of statistical mechanics of fluids. The appropriate statistics is given by the solution
of the Boltzmann equation. That equation represents theevolution of the governing
distribution function, which is dependent on the position and velocity of the particles
colliding with one another as a result of thermal excitation at any finite temperature.
The collisions are described by an integral collision operator. In general, the colli-
sion operator represents simultaneous collisions among many particles, necessitating
the use of a many-particle distribution. As such, it is very complicated and essen-
tially impossible to evaluate precisely. Only in the case of dilute gases can one limit
oneself to considering the evolution of a single-particle distribution and to binary col-
lisions, since many body collisions are highly unlikely. In this idealized situation,
the collision operator can be approximated by first-order and second-order spatial
derivatives. The former is the familiar pressure gradient and the latter is the Laplac-
ian operating on the velocity, multiplied by a constant known as the viscosity. With
that approximation in hand, we can take the appropriate moments of the one-particle
Boltzmann equation and so derive first the conservation of mass equation and sec-
ond the conservation of momentum equation that we recognize as the NSE (when the
incompressibility condition is a valid assumption).
Although such a derivation has been carried out for dilute gases, a corresponding
exercise for liquids remains an open problem. This is because binary collisions play
a relatively minor role in liquids, which are much denser than gases and hence feature
collisions between clusters of particles. However, for practical reasons and lacking a
better option, we use the Navier–Stokes equations with a simple constant viscosity
as a reasonable model for liquid flows.
The origin of viscosity imposes a limit on the domain of validity of the Navier–
Stokes equations. Thus phenomena on a length scale comparable to or smaller than
the collision mean free path in air at atmospheric pressure (say, 10
−3
cm) cannot be
described by a continuum model such as the NSE. Subsequently we will learn about

4 I Introduction and Overview of Turbulence
some natural lengths that characterize the length scale region in which flow energy
dissipation is dominated by viscous phenomena. It will be important then to be sure
that we are still in the regime characterized by a continuum model of the flow. A
similar cautionary remark applies to the amplitude of fluctuations in turbulent flows:
once we are in a regime in which those fluctuations are comparable with thermally (fi-
nite temperature) induced fluctuations, the model based on Navier–Stokes equations
ceases to be relevant.
Nondimensional Form of the Navier–Stokes Equations
It is sometimes convenient, both for physical discussions and mathematical trans-
parency, to consider a nondimensional form of the conservation of momentum equa-
tion. For that purpose we introduce a reference lengthL
∗and a reference timeT ∗for
the flow, andwe set
x=L
∗x
.
,t=T ∗t
.
,p=P ∗p
.
,u=U ∗u
.
,f=
L

T
2

f
.
,
whereP
∗=U
2

andU ∗=L∗/T∗are a reference pressure and a reference velocity,
respectively. By substitution into (1.7) we obtain foru
.
,p
.
,f
.
the same equation but
withνreplaced by Re
−1
,where Re is a nondimensional number called theReynolds
number:
Re=
L
∗U∗
ν
. (1.8)
The value of the Reynolds number depends on the choice of the reference length and
velocity. Usually, if¯(the domain occupied by the fluid) is bounded thenL
∗can be
taken as the diameter of¯or as some other large-scale length related to¯≥such as
the width of a channel. The choice ofU
∗(and hence ofT ∗)depends on the type of
forcing of the flow; it can be related to the forces applied at the boundary of¯or to
a pressure gradient, for example. Various choices ofL
∗andU ∗can be appropriate
for a given flow, leading to various definitions of the Reynolds number, but turbulent
flows result for all appropriate choices when Re is large. How large depends to some
extent on the shape of the domain occupied by the fluid. Once the shape of the do-
main¯is fixed, however, rescalings in length(L
∗)and velocity(U ∗)and changes in
viscosity(ν)affect the equations only through the single parameter Re.
Hence, different experiments may lead to the same nondimensional equations. For
example, multiplying the velocity by 2 and dividing the diameter of the domain by
2leaves the Reynolds number unchanged, so we can pass from one experimentto
another; this is the Reynolds similarity hypothesis constantly used in mechanical en-
gineering. At a given Reynolds number, flows remote from the boundaries of the
domain¯≥irrespective of the latter’s shape, are similar owing to some universal-
ity properties of turbulent flows. Moreover, with flows around blunt bodies (say,
a sphere), as the body’s radius increases and the flow velocity and/or viscosity is
adjusted so as to maintain the Reynolds number constant, the flow throughout the

2 Turbulence: Where the Interests of Engineers and Mathematicians Overlap 5
modified flow domain remains similar. That is what has made possible the design of
aircraft by means of relatively small models tested in moderately sized wind tunnels.
In ChapterIII, instead of the Reynolds number we will use another nondimen-
sional number: the Grashof number (see Section 13 in Chapter II).
A heuristic argument illustrating the significance of the Reynolds number emerges
by comparing the inertial and dissipation terms of the Navier–Stokes equations. The
inertial term(u·∇)uhas dimension
U
2

L∗
,
while the dissipation term has dimension
ν
U

L
2

.
The inertial term dominates when
Re=
L
∗U∗
ν
1.
However, a much more subtle analysis that is valid at each length scale is made for
the Kolmogorov theory of turbulence.
By setting Re=+∞(i.e.,ν=0),we obtain the case of inviscid flows. In this
case, the divergence-free condition is retained but the momentum equation changes,
resulting in the Euler equations for inviscid perfect fluids:
∂u
∂t
+(u·∇)u+∇p=f, (1.9a)
∇·u=0. (1.9b)
Note that some of the difficulties encountered in studying turbulent behavior, a
largely inviscid regime, arise because the transition from Euler’s equations to the
Navier–Stokes equations necessitates a change from a first-order system to a second-
order one in space(∇to}=≥which involves a singular perturbation.
2 Turbulence: Where the Interests of Engineers
and Mathematicians Overlap
Principal substantive questions related to turbulence have been raised since the begin-
ning of the twentieth century, and a large number of empirical and heuristical results
were derived – motivated principally by engineering applications. This includes the
work of Lamb [1957], mostly on addressing idealized inviscid flows; Prandtl [1904],
on eddy viscosity and boundary layers; Taylor [1935, 1937], on viscous flows; and
von Karman[1911, 1912], on the nature of theboundary layer.
At the same time, in mathematics there appears the pioneering work of Jean Leray
[1933, 1934a,b] on the Navier–Stokes equations. Leray speculated that turbulence is

6 I Introduction and Overview of Turbulence
due to the formation of point or “line vortices” on which some component of the ve-
locity becomes infinite.
1
To enable dealing with such a situation, he suggested the
concept of weak, nonclassical solutions to the Navier–Stokes equations (1.7), and
this has become the starting point of the mathematical theory of the Navier–Stokes
equations to this day. We will consider this approach in Chapter II and beyond. It is
noteworthy that, more generally, Leray’s ideas serve also as the starting point for sev-
eral important elements of the modern theory of partial differential equations. Even
today, despite much effort, Jean Leray’s conjecture concerning the appearance of sin-
gularities in 3-dimensional turbulent flows has been neither proved nor disproved.
Let us mention, however, the result of Caffarelli, Kohn, and Nirenberg [1982] (see
also Scheffer [1977]), which considerably extends an earlier result of Leray: Given
the possibility that the singular points are a fractalset (assuming that such a set exists),
the 1-dimensional Hausdorff measure of that set in space and time is 0. Hence the
occurence of smooth line vortices is not possible, explaining our quotation marks
around “line vortices.” Nevertheless, for all physical purposes this powerful mathe-
matical result leaves room for a tremendously complex set of singularities, and so we
remain far from closing the issues raised by Leray’s conjecture.
Before continuing with these historical notes, we remark in passing that engineers
are not directly affected bysuch purely mathematical issues; rather, they want to
calculate or measure certain physical quantities (forces, velocities, pressures, etc.).
Here, however, beside the possible occurrence of singularities, another critical aspect
of turbulence comes to mind: in a turbulent flow, many interesting quantities vary
rapidly in time and cannot be readily measured. In practice, all that can be measured
in laboratory experiments are averages (usually time averages). These averages are
well-defined, reproducible quantities. This leads to the concept of ensemble averages
underlying the conventional theory ofturbulence, and to the concept of statistical so-
lutions of the Navier–Stokes equations (1.7). It leads also to the idea ofergodicity,
which is taken for granted by engineers. Loosely speaking, for all initial experimental
conditions and for all sorts of reasonable ensemble averages, the experiments always
yield the same measured results to within the accuracy of the measurements. We
address here those questions of direct interest to engineers: the need for statistical so-
lutions, the equivalence between ensemble averages and time averages (a question
addressed in Chapter IV), and the so far unchallenged issue of the axiomatic nature
of ergodicity.
We return to our brief overview of some highlights in the history of the studies
of turbulent flows. It is impossible to explore here all the aspects of that history.
Hence, with apologies to all whose important contributions are not mentioned here,
we limit ourselves to those aspects of the history most relevant to the subject of this
monograph.
1
In fact, if such discontinuities occur then another question of physical nature needs to be
raised concerning the validity of the Navier–Stokes equations themselves; indeed, at very
short distances of order 10
−3
cm (the collision mean free path of the particles), the fluid
equations are no longer pertinent.

2 Turbulence: Where the Interests of Engineers and Mathematicians Overlap 7
Turbulent flows have mystified people for ages, as evidenced for example by
Leonardo da Vinci’s sketches of the turbulent wakes downstream of some bridge
columns. Beginning with careful experimental studies of flows under various exper-
imental conditions (Reynolds [1883, 1895]) and with the subsequent formulation of
the Navier–Stokes equations, turbulence became a subject of thorough scientific in-
quiry. For many years, two difficulties held the attention of various investigators. The
first was a technical mathematical obstacle: the presence of the inertial term (a qua-
dratic nonlinearity) precludes a straightforward use of the many available tools of
perturbation methods. The structure of the equations demands that, at any given step
in an approximation scheme, information from the next step is necessary. This had
led to many attempts at formulating the so-called closure schemes, where at some
step in the approximation sequence an assumption about the nature of the subsequent
term is made, thereby terminating that sequence. Such an assumption, usually jus-
tified in terms of intuitive physical arguments, was then used to break the impasse
in the approximation sequence. In principle, closure schemesby and large call for
unprovable assumptions beyond those composing the basis for the Navier–Stokes
equations. Some of the better-known closure schemes may be found in such texts as
Tennekes and Lumley [1972], Leslie [1973], and Lesieur [1997], although further at-
tempts (and controversies) in this area continue. As we shall find in the present work,
the invention of the so-called inertial manifolds in the context of the rigorous theory
of NSE (as well as of other nonlinear partial differential equations) opens the door to
mathematically more soundly based schemes for computationalapproaches, offering
an alternative to the conventional closure schemes.
The second obstacle to progress in the theory of turbulence was largely concep-
tual. Namely, how was it possible for a system described by perfectly deterministic
equations to exhibit behavior that was undeniably statistical in nature? This aspect
of turbulent flows, both from the experimental side and from the nascent theoretical
side, is dealt with at length in the monumental work of Monin and Yaglom [1975].
Hopf [1952], followed by Foias and Prodi [1976] (see also Foias [1972, 1973, 1974]),
studied an extension of Liouville’s theorem that in principle yields the probability
distribution function underlying the Navier–Stokes equation. Many of these efforts
rested on the experimental and theoretical work of Taylor [1935, 1937] and von Kar-
man and Howarth [1938], who clarified,on intuitive grounds, the nature of homoge-
neous isotropic turbulence. The simplifications resulting from the symmetries inher-
ent in this idealized form of turbulence yielded the well-known von Karman–Howarth
ordinary differential equation for the self-similar evolution of the two-point veloc-
ity correlation tensor. This idealization has also yielded Kolmogorov’s theory for
the spectrum of homogeneous isotropic turbulence in three dimensions(Kolmogorov
[1941a,b]) (and later Batchelor’s [1959] and Kraichnan’s [1967] corresponding re-
sults for turbulence in two dimensions), a subject of the next section. All of these
results were obtained without full understanding of the origin of the statistical na-
ture of turbulence. A significant breakthrough occurred in the 1960s and 1970s with
the discovery of stochastic instabilities in seemingly innocuous low-order ordinary

8 I Introduction and Overview of Turbulence
differential equations (Lorenz [1963]) and in some nonlinear difference equations
(Feigenbaum [1980]). Subsequent research (Foias and Prodi [1976], Vishik and Fur-
sikov [1977a,b, 1978], Foias and Temam [1979]) on dynamical systems governed by
nonlinear partial differential equations revealed that such dynamical systems may
reside, in finite-dimensional function spaces, on compact attractors that may be char-
acterized by chaotic behavior.
It is now appropriate to reiterate a point hinted at earlier, namely, the essential
need for careful mathematical analysis when dealing with nonlinear entities such as
the Navier–Stokes equations. While much of our physical intuition serves us well in
the domain of linear phenomena modeled adequately by linear differential and par-
tial differential equations, it can fail us – with potentially disastrous consequences –
in nonlinear domains. A fairly instructive example, outside the realm of thisbook
but worth mentioning here, concerns modeling sonic flow transition as a boundary
value problem rather than (and more correctly) as an initial value problem (Greenberg
and Trève [1960]). Although this may appear to be unnecessarypedantry, it clearly
makes a lot of difference in the context of, say, nuclear reactor safety (Bilicki et al.
[1987]). Unlike the case in linear systems, in nonlinear systems small causes can lead
to very large effects indeed, as well as to qualitative differences. Because nonlinear
equations can have multiple, qualitatively different solutions (different basins of at-
traction), a small change in initial conditions can sometimes lead to radically different
time-asymptotic behavior. An even more dramatic, counterintuitive example is the
previously mentioned possibility of chaotic behavior in what at first sight seem to be
innocent deterministic systems (Lorenz [1963], Feigenbaum [1980], Smale [1967]).
Here is a class of problems in which necessarily limited computer “experiments” can
lead to misleading conclusions about the behavior of a system as a function of the
governing parameters. Only a thorough analysis of the system can reveal its true
nature. Occasionally, such an analysis will reveal, even without detailed numerical
computations, an unphysical aspect of the system (e.g., infinite energy density, de-
creasing entropy, or other pathologies), which is a clear alert to the flawed nature of
the system model.
In this work we concentrate on those aspects of turbulent fluid flows that can be rep-
resented in terms of so-called Sobolev spaces – that is, a class of functions satisfying
the given boundary conditions – andthe given physical constraints, such as diver-
gence-free (incompressible) flow. The various norms (i.e., various integrals of some
seemingly abstract quantities) in these function spaces are in fact readily recognized
as tangible physical quantities that are more or less readily accessible to direct ex-
perimental observation. The relationships among these norms, and the rules for their
manipulations, reveal some aspects of the turbulent flows that justifymany ad hoc
interpretations and inspire insights derived from direct observations of turbulence
while also revealing some hitherto unrealized ones. As such, these mathematical en-
deavors can serve to enlarge our intuitive horizons beyond the limits of linear theories
and models.

3 Elements of the Theories of Turbulence of Kolmogorov and Kraichnan 9
3 Elements of the Theories of Turbulence
of Kolmogorov and Kraichnan
Turbulent flows seem to display self-similar statistical properties at length scales
smaller than the scales at which energy is delivered to the flow. Kolmogorov [1941a,b]
argued that, at these scales, in three dimensions, the fluids display universal statisti-
cal features. Turbulent flow is conventionally visualized as a cascade of large eddies
(large-scale components of the flow) breaking up successively into ever smaller sized
eddies (fine-scale components of the flow; Onsager [1945]). Such a cascade, or flow
of kinetic energy from large to small scales, is taken to occur in a regime at lengths suf-
ficiently large for the effects of viscosity to be inconsequential. The apparent energy
dissipation – that is, the removal of energy from one length scale to a smaller one –
is solely due to the presence of the nonlinear (inertial) term in the Navier–Stokes
equations. The energy dissipation rate˚=νκ
3
0
|∇u(x,t)|
2
is assumed to be con-
stant in spaceand time. A further essential assumption is that the cascade proceeds so
that, at every length scale (or at every corresponding wavenumber), there is an equi-
librium between energy flowing in from above to a given scale and that flowing out
to a lower scale. Such a picture and the associated assumptions imply that, in this
range of length scales (or this range of wavenumbers), the energy density at a given
wavenumber can depend only on the energy dissipation rate˚and the wavenumber
kitself. Then dimensional analysis alone yieldsS(κ)=const.×˚
2/3

5/3
for the
energy density. Such a cascade process cannot continue to arbitrarily small length
scales because, as the norm of the Laplacian operator increases with the decreasing
length scale, eventually the effects of molecular dissipation begin to dominate the
nonlinear inertial term. That length, denoted by
d,is the endpoint of the inertial
range and the beginning of the dissipation range.
Let us determine
d.At each scale (or wavenumberκ=
−1
),we can define by
dimensional analysis, through˚and ,a natural time scaleτand speedu.Indeed,
˚=
2

3
givesτ=(
2
+≈=
1/3
andu= /τ=−˜≈=
1/3
.Now, the dissipation length

dis where the viscous term˘∕ustarts to dominate, on average, the inertial term.
Hence,
˘∕u∼
νu

2

ν
τ
>(u·∇)u∼
u
2



τ
2
.
Therefore,

2
<ντ=ν
·

2
˚

1/3
⇐⇒
4/3
<
·
ν
3
˚

1/3
and

d=
·
ν
3
˚

1/4
. (3.1)
Kolmogorov thus inferred that, in 3-dimensional turbulent flows, the eddies of
length size sensibly smaller than
dare of no dynamical consequence. As we said,

10 I Introduction and Overview of Turbulence
the length das defined by (3.1) is known as the Kolmogorovdissipation length.The
corresponding wavenumber,
κ
d=
1
d
=
·

ν
3

1/4
, (3.2)
is the Kolmogorovdissipation wavenumber.
The inertial range, within which inertial effects dominate, is the range
1< <

d,where 1=L1is the wavelength at which energy is injected in the flow. To each
length in this range we can associate a Reynolds number Re
=u /ν;hence,
Re
3/4

=
·

ν
3

1/4
.
The largest of these Reynolds numbers obtained for =the Kolmogorov macro-scale
lengthL
∗µL1is the Reynolds number Re of the flow. Hence, with (3.1),
Re=
·
L

d

4/3
,orL ∗=Re
3/4
d. (3.3)
This relationship leads naturally to the heuristic estimate of the number of degrees
of freedom in 3-dimensional flows, which is Re
9/4
.As we shall see, this heuristic
estimate is actually an upper bound on the sufficient (but not necessary) number of
degrees of freedom in 3-dimensional turbulent flows.
We now present a somewhat more elaboratederivation (but one that is still divorced
from the Navier–Stokes equations) of the so-called Kolmogorov spectrum.
Let≈denote the average of the energy per unit mass. Then, according to the Kol-
mogorov theory, the length
dat which the turbulent eddies are rapidly annihilated
by the viscosity should be a universal function of≈and the kinematic viscosityν,
namely:

d=f(ν,). (3.4)
In particular,fshould be independent of the choice of units for space and time. Thus,
if we pass fromx,ttox
.
=ξxandt
.
=τtthen we should still have

.
d
=f(ν
.
≥≈
.
). (3.5)
Hereν
.
and≈
.
are not independent ofνand≈≥and dimensional analysis yields

.
d
=ξ d,ν
.
=
ξ
2
τ
ˆ≥ ≈
.
=
ξ
2
τ
3
≈; (3.6)
that is,
$∼−ˆ≥ ≈==f(ξ
2
τ
−1
ν, ξ
2
τ
−3
). (3.7)
With the choices
ξ
2
τ
=
1
ν
and
ξ
2
τ
3
=
1

,(i.e.,τ=−≈ˆ=
1/2
andξ=≈
1/4

3/4
),

3 Elements of the Theories of Turbulence of Kolmogorov and Kraichnan 11
the relation (3.7
∼−ˆ≥≈==
1
ξ
f(1,1)∼
C
ν
3
˚

1/4
.
Following Kolmogorov, one can also argue that the average energy per unit mass,
e
κ,2κ,of the eddies of lengths between /2 and (i.e., between the wavenumbersκ
and 2κ,whereκ=1/ )should enjoy a similar universal property – namely,
e
κ,2κ=g(, κ), (3.8)
provided thatκλκ
d(so that the effect of the viscosity can be neglected) and thatκ
is much larger than the wavenumber at which energy is pumped into the flow. Again
the universality ofgimplies that
ξ
2
τ
−2
g(, κ)=g(ξ
2
τ
−3
≈≥ $
−1
κ);
whence, upon takingξ=κandτ=˚
1/3
κ
2/3
,one obtains
e
κ,2κ=c
˚
2/3
κ
2/3
, (3.9)
wherec=g(1,1),a universal constant.
Consider now the Navier–Stokes equations with periodic boundary conditions.
Thatis, we consider the solutions of equations (1.7) that are periodicin space with
periodLin each direction. Using Fourier series expansions (see Sections 2 and 5 in
Chapter II for details), we can write
u(x)=
Z
k∈Z
3
\{0}
ˆuke
κ1ik·x
,ˆu k·k=0,ˆu −k=
ˆuk. (3.10)
Forkin (3.10),κ=κ
1|k|is the corresponding wavenumber, whereκ 1=2π/L.The
lowest wavenumber isκ
1.The component ofuwith wavenumbers betweenκ
Z
and
κ
ZZ
is
u
κ
Z

ZZ=
Z
κ
Z
≤κ<κ
ZZ
ˆuke
κ1ik·x
. (3.11)
The energy per unit mass and the enstrophy (square of vorticity) per unit mass are,
respectively,
|u
κ
Z

ZZ|
2
=
Z
κ
Z
≤κ<κ
ZZ
|ˆuk|
2
(3.12)
and
ıu
κ
Z

ZZı
2

2
1
Z
κ
Z
≤κ<κ
ZZ
|k|
2
|ˆuκ|
2
. (3.13)
Note thatu
κ1,∞=u.
Physicists and engineers assume that, forL1,the time averages of
|u
κ
Z

ZZ(t)|
2
andκ
2
1
ıuκ
Z

ZZ(t)ı
2
(3.14)

12 I Introduction and Overview of Turbulence
exist (see Chapters IV and V). Note that the first average should bee κ
,

,,,the aver-
age of energy per unit mass of the eddies of linear size between 1/κ
,,
and 1/κ
,
.We
denote these averages by
˙u
κ
,

,,(·)|
2
˝andκ
2
1
˙ıuκ
,

,,(t)ı
2
˝. (3.15)
Moreover, it is also assumed that these values can be viewed (at least whenκ
1is
small) as integrals in the wavenumbers; that is,

κ
,,
κ
,
S(κ) dκand

κ
,,
κ
,
S1(κ) dκ. (3.16)
Comparing (3.12) and (3.13) with (3.15), we see that if (3.16) makes sense then one
is led to the relation
S
1(κ)≡κ
2
S(κ). (3.17)
The functionS(κ) (≥0)is called theenergy spectrumof the turbulent flow pro-
duced by the driving forcefin (1.7). Also, the driving force is assumed to have no
high-wavenumber components:f=f
κ1,¯κ,where¯κis comparable in size withκ 1(the
lowest wavenumber).
So, according to Kolmogorov’s theory, we have (see (3.9


κ
S(χ)dχ∼c

2/3
κ
2/3
, (3.18)
at least as long as
¯κλκλκ
d. (3.19)
Taking the derivative in (3.18) yields
S(κ)−S(2κ)∼
2c
3

2/3
κ
5/3
,
whence
S(κ)∼S(2
m+1
κ)+
2c
3
·
1+
1
2
5/3
+
1
2
10/3
+···+
1
2
5m/3


2/3
κ
5/3
(3.20)
as long as 2
m
κλκ d.For turbulent flows,κ dκ1≈¯κand so we may takem1
in (3.20
S(κ)∼C
,
K

2/3
κ
5/3
, (3.21)
whereC
,
K
=(2/3)c(1−2
−5/3
)
−1
.The form (3.21
theKolmogorov energy spectrumof the turbulent flow. The constantC
,
K
is known as
theKolmogorov constantin energy space (there isa similar relation in which a con-
stantC
Kappears and takes the name of Kolmogorov constant in physical space; see
(5.26 C
,
K
is of the order of unity. The range
ofκin (3.19) for which (3.21 inertial range.
It must be noted that the estimate (3.21) is really a time average, as the ampli-
tude ofS(κ)fluctuates wildly in time. Furthermore, it is only an approximation. In
reality, for a turbulent flow in a bounded domain, intermittency effects in the energy

3 Elements of the Theories of Turbulence of Kolmogorov and Kraichnan 13
dissipation rate≈result in small but measurable corrections to the simple expression
(3.21) (see e.g. Kolmogorov [1962] and Novikov and Stewart [1964]). These cor-
rections depend on the size (say,L
∗)of¯≥thus destroying – to some extent – the
universality of turbulence.
As seen in the preceding, the arguments leading up to (3.21) are clearly divorced
from the NSE itself and are applicable solely to turbulent flows in three dimensions.
For flows in two dimensions (which, as stated earlier, are amenable to deep analy-
sis), we must turn to a phenomenological theory proposed by Batchelor [1959] and
Kraichnan [1967]. That theory is in the spirit of Kolmogorov’s approach but does not
parallel it because the physical situation is quite different; hence we offer a separate
exposition.
Two-dimensional flows are not commonly encountered in nature. Examples that
do come to mind are thin liquid films and (to within some approximation) the at-
mospheric layer on the surface of the Earth – although clearly the significant phe-
nomena (e.g.,weather and climate changes) occur on scales within which the finite
thickness of that layer must be taken explicitly into account. However, some firm
mathematical results derived in the study of 2-dimensional flows appear to carry
over to some 3-dimensional flows, so it is instructive to follow what can be learned
about 2-dimensional flows. Moreover, further advances in functional analysis of the
Navier–Stokes equations in three dimensions may yield the necessary tools for solv-
ing some critical open problems. For now we turn to a summary of Kraichnan’s work
on the phenomenological theory of 2-dimensional turbulence.
We limit ourselves here to fluid flows in the plane, although much of the theory
could carry over to more general 2-dimensional manifolds. The principal physical dif-
ference between 2-dimensional and 3-dimensional flows is that, in the 2-dimensional
case, the vorticity (i.e., the curl of the velocity) has only one component – in the direc-
tion normal to the plane of the flow. This imposes a severe constraint on the kinematics
and the dynamics of the turbulence. For instance, in addition to the conservation of
energy, the flow must conserve enstrophy, that is, the integral of the square of the
vorticity over the flow domain must be constant. The nonlinear interactions may
be viewed in wavenumber (Fourier) space as three-wave interactions. They cannot
simultaneously satisfy the two conservation principles. Hence, the energy cascade
cannot coexist with the enstrophy cascade; they must occur in distinct portions of
the wavenumber domain. As a consequence, in the turbulent regime (large Reynolds
number) for the 2-dimensional case, there are two contiguous ranges: for wavenum-
bers lower than that at which the forcing of the flow is introduced there is an inverse
energy cascade, with small eddies coalescing into larger ones and with the cascade
terminating at a wavenumber determined by the size of the flow domain. The spec-
trum of the energy,S(κ),in that domain is the same as the Kolmogorov spectrum.
Toward the higher wavenumbers, the principal cascading entity is the enstrophy –
that is, the successive breakup of the vortices into ever smaller ones, with the atten-
dant enstrophy dissipation rateηresulting from nonlinear interactions and not being
affected by the molecular viscosity above theKraichnan cutoff length.According to

14 I Introduction and Overview of Turbulence
Kraichnan, the two portions of the inertial range (i.e., the inverse cascade and the
enstrophy cascade) cannot overlap.
Using arguments based on dimensional analysis along the lines followed by Kol-
mogorov in the 3-dimensional regime, we can now determine the energy spectrum
in the enstrophy cascade region as well as the Kraichnan cutoff wavenumber. Note
first that the enstrophy dissipation rate has the dimension of (time
−3
.Assuming
that, in the enstrophy cascade range, the energy spectrum depends only onηand the
wavenumberκ,we then find thatS(κ)=const.×η
2/3

3
.Similarly, it follows from
dimensional considerations that the Kraichnan cutoff wavenumber is given by
κ
η=
C
η
ν
3

1/6
. (3.22)
An extended treatment of turbulent flows in two dimensions is presented in Sec-
tion 5 of Chapter IV.
4 Function Spaces, Functional Inequalities, and Dimensional Analysis
The mathematical theory of the Navier–Stokes equations is based on the use of func-
tion spaces, which are at the heart of the modern theory of partial differential equa-
tions. A formal presentation of the needed tools appears in Chapter II and thereafter
as needed. However, in this section we give an informal introduction and empha-
size that these spaces are not merely inventions of mathematicians; rather, they are
strongly related to the physics of the problem.
The Fundamental Function Spaces
Consider the domain¯occupied by the fluid;¯is a domain ofR
3
,which could
be the whole spaceR
3
in certain idealized cases. The first natural function space
isL
2
−¯=≥the space of square integrable functions on¯;we also haveL
2
−¯=
3
,the
space of square integrable vector fields on.These spaces are endowed with a scalar
product, which we denote by
(u,v)=

¯
u(x)v(x)dx
in the scalar case and, similarly,
(u,v)=

¯
u(x)·v(x)dx
in the vector-valued case. To these scalar products correspond the following norms
2
(mean square norms):
|u|=
C≥
¯
|u(x)|
2
dx

1/2
,|u|=
C≥
¯
|u(x)|
2
dx

1/2
.
2
We will use the same notation|·|for the Euclidean norm inR
2
andR
3
(and evenC
2
and
C
3
),and also for variousL
2
norms. This abuse of notation, for purposes of simplification,
does not lead to any confusion once the context is taken into account.

4 Function Spaces, Functional Inequalities, and Dimensional Analysis 15
The inner product and the corresponding norm are also related by the so-called
Cauchy–Schwarz inequality:
|(u,v)|=





¯
u(x)v(x)dx





C≥
¯
|u(x)|
2
dx

1/2C≥
¯
|v(x)|
2
dx

1/2
=|u||v|
(4.1)
for allu,vinL
2
().Similarly, for anyu,v∈L
2
−¯=
3
,we have
|(u,v)|≤|u||v|. (4.2)
Now, given the velocity vector fielduof the fluid in¯≥
u:x∈¯fiflu(x)∈R
3
,
we see that the square of theL
2
-norm,|u|
2
,is merely twice the kinetic energy of the
flow (assuming that the density of the fluid has been normalized to unity):
e(u)=
1
2

¯
|u(x)|
2
dx=
1
2
|u|
2
.
Without entering into the details of measure theory, we recall thatL
2
−¯=and
L
2
−¯=
3
are Hilbert spaces for these scalar products and norms. Also, forL
2
−¯=(and
the same is true forL
2
−¯=
3
),the following characterizaton holds:
3
u∈L
2
−¯=if and only ifthere exists a sequence of smooth
functionsu
n,compactly supported in¯≥such that|u n|(ore(u n))
remains bounded andu
nis converging tou(in the distribution
sense) asn→∞.
(4.3)
Most of the spaces that we will consider are derived from the spaceL
2
−¯=and from
another space that we will now introduce, the so-called Sobolev spaceH
1
().
In Chapter II we will address the concept of enstrophy of a fluid velocityu=
(u
1,u2,u3),namely,
E(u)=
3
R
i,j=1

¯




∂u
i
∂xj




2
dx.
By comparison with (4.3
What can we say of a sequence of smooth velocity vector fieldsu
n
such thatE(u n)remains bounded?
(4.4)
If¯is bounded,
4
then one can prove thatu ncontains one (or more) subsequence(s
that converge in the distribution sense to some limitu.This vector functionu=
(u
1,u2,u3)is inL
2
−¯=
3
,as are its (distributional) first derivatives
∂u
i
∂xj
,i,j=1,2,3;
3
The functionsu nin this characterization ofL
2
−¯=are defined, say, by some kind of approx-
imation procedure.
4
If¯is not bounded, then we should also require thate(u n)remain bounded as well.

16 I Introduction and Overview of Turbulence
in fact,
E(u)≤sup
n∈N
E(un).
We say that such a functionubelongs to the Sobolev spaceH
1
−¯=
3
(with a similar
definition forH
1
−¯=in the scalar case). In some sense:
The spaceL
2
−¯=
3
consists of all the vector fieldsuwith finite
kinetic energy, and the spaceH
1
−¯=
3
consists of all the vector
fieldsuwith finite enstrophy.
(4.5)
From the mathematical point of view, the Sobolev spacesH
1
−¯=andH
1
−¯=
3
are
Hilbert spaces for the following inner products and norms (see Chapter II):
((u,v))
1=
1
L
2

¯
u(x)v(x)dx+
3
Z
j=1

¯
∂u
∂xj
∂v
∂xj
dx,
((u,v))
1=
1
L
2

¯
u(x)·v(x)dx+
3
Z
i,j=1

¯
∂ui
∂xj
∂vi
∂xj
dx,
ıuı
1=[((u, u))1]
1/2
,ıuı 1=[((u,u)) 1]
1/2
,
whereLis a typical length (e.g., the diameter of).In nondimensional variables,
L=1.As with the spaceL
2
−¯=and any other Hilbert space, we have the Cauchy–
Schwarz inequality, which in this context reads
|((u,v))
1ˇıuı 1ıvı1,|((u,v)) 1ˇıuı 1ıvı1 (4.6)
for allu,vinH
1
−¯=and allu,vinH
1
−¯=
3
.
Most function spaces that we consider are derived from these two physically obvi-
ous spaces, the spaceL
2
−¯=
3
of finite kinetic energy and the spaceH
1
−¯=
3
of finite
enstrophy. For instance, two central spacesVandHappear throughout the book; as-
suming for simplicity that¯is bounded, a mathematically rigorous and physically
intuitive definition of the spacesVandHis as follows:
Vis made up of all the limit points (in the distributional sense)
of all the possible sequences of smooth vector fieldsu
nwhich
are divergence-free, which satisfy the boundary conditions of the
problem, and whose enstrophy remains bounded; that is,E(u
n)≤
const.<∞.
(4.7)
The spaceHis defined in a similar way, replacing the boundedness of the enstro-
phy by the boundedness of the kinetic energy:e(u
n)≤const.<∞.More details
are given in Chapter II.
Functional Inequalities
The functions belonging to the spaceH
1
−¯=(and to other related spaces) satisfy cer-
tain inequalities, which are called Sobolev inequalities in the mathematical literature.

4 Function Spaces, Functional Inequalities, and Dimensional Analysis 17
We make extensive use of these inequalities in the course of this book. Some of them
are proven by interpolation, others by appropriate direct methods (e.g., Gagliardo–
Nirenberg’s, Agmon’s, Ladyzhenskaya’s, and Poincaré’s inequalities). Our objective
here is twofold:
(i) to prove the Ladyzhenskaya and the Poincaré inequalities; and
(ii) to emphasize thephysical invariance by dilatationorby change of scaleof all
such inequalities.
The Ladyzhenskaya inequality may be described as follows. For a smooth, com-
pactly supported scalar functionuinR
2
,we have

R
2
|u(x)|
4
dx≤
C≥
R
2
|u(x)|
2
dx
−·≥
R
2
|∇u(x)|
2
dx

. (4.8)
InR
3
,we have

R
3
|u(x)|
4
dx≤
C≥
R
3
|u(x)|
2
dx

1/2C≥
R
3
|∇u(x)|
2
dx

3/2
. (4.9)
In (4.8
tionsuandgis assumed for the sake of simplicity, since these inequalities are valid
for more general (less smooth) functions.
Remark 4.1Note the difference between space dimensions 2 and 3. It is this very
discrepancy between the two cases that induces many ofthe difficulties in the math-
ematical theory of the Navier–Stokes equations in space dimension 3.
That no dimensional constant appears in the RHS of (4.8
that these inequalities are invariant by dilatation, or homothety(xfiflλx),or (in
physical terms) that both sides of these inequalities have the same dimension:
U
4
L
2
∼U
2
L
2
C
U
2
L
2
L
2

for (4.8
U
4
L
3
∼(U
2
L
3
)
1/2
C
U
2
L
3
L
2

3/2
for (4.9.
Remark 4.2This invariance of the inequality by dilatation (or homogeneity) is com-
mon to many functionalinequalities. However, the lack of a multiplicative constant
in the RHS of (4.8
existence of a multiplicative constant in the right-hand side of the inequalities, which
in some cases may be obtained explicitly. In general, however, such constants can be
taken as of order unity.
We present now the proof of (4.8
(4.8
|g(x)|
2

C≥
R
|g(ξ)|
2


1/2C≥
R
|g
R
(ξ)|
2


1/2
for allx∈R, (4.10)

18 I Introduction and Overview of Turbulence
for any smooth functiongwith compact support inR.Sinceghas compact support,
there existsL>0 sufficiently large such thatg(x)vanishes forxoutside(−L, L).
Then we can write
g(x)
2
=2

x
−L
g(ξ)g
R
(ξ) dξ≤2

x
−L
|g(ξ)||g
R
(ξ)|dξ
and
g(x)
2
=−2

L
x
g(ξ)g
R
(ξ) dξ≤2

L
x
|g(ξ)||g
R
(ξ)|dξ.
Adding both identities, we obtain
g(x)
2


L
−L
|g(ξ)||g
R
(ξ)|dx=

R
|g(ξ)||g
R
(ξ)|dx.
Then, using the Cauchy–Schwarz inequality, we find (4.10).
We now prove the 2-dimensional Ladyzhenskaya inequality (4.8
Agmon inequality (4.10) twice, once in each space direction. We have
u(x
1,x2)
4
=u(x1,x2)
2
u(x1,x2)
2

;·≥
R
|u(ξ1,x2)|
2
dξ1
−·≥
R
|uξ1
(ξ1,x2)|
2
dξ1

C≥
R
|u(x1,ξ2)|
2
dξ2
−·≥
R
|uξ2
(x1,ξ2)|
2
dξ2
−∗
1/2
.
Thus,

R

R
u(x1,x2)
4
dx1dx2≤

R
C≥
R
|u(ξ1,x2)|
2
dξ1

R
|uξ1
(ξ1,x2)|
2
dξ1

1/2
dx2

R
C≥
R
|u(x1,ξ2)|
2
dξ2

R
|uξ2
(x1,ξ2)|
2
dξ2

1/2
dx1.
Using the Cauchy–Schwarz inequality, we obtain

R

R
u(x1,x2)
4
dx1dx2

C≥
R

R
|u(ξ1,x2)|
2
dξ1dx2

1/2C≥
R

R
|uξ1
(ξ1,x2)|
2
dξ1dx2

1/2

C≥
R

R
|u(x1,x2)|
2
dx1dx2
−·≥
R

R
|∇u(x1,x2)|
2
dx1dx2

,
C≥
R

R
|u(x1,ξ2)|
2
dξ2dx1

1/2C≥
R

R
|uξ2
(x1,ξ2)|
2
dξ2dx1

which proves (4.8
one first writesu(x
1,x2,x3)
4
as a product of two squares; then one applies the Agmon
inequality (4.10) with respect to one variable and the 2-dimensional Ladyzhenskaya
inequality (4.8

4 Function Spaces, Functional Inequalities, and Dimensional Analysis 19
With little effort we can now prove another fundamental (often used) inequality:
the Poincaré inequality.
5
We assume thatuis smooth and that it vanishes outside a
bounded set; in fact, it suffices foruto vanish outside a slab, say−L/2<x
1<L/2.
In space dimension 2 (the proof is the same in any space dimension), we write (4.10)
in directionx
1withx 2fixed:
u(x
1,x2)
2

C≥
R
|u(ξ1,x2)|
2
dξ1

1/2C≥
R




∂u∂ξ1
(ξ1,x2)




2
dξ1

1/2
.
Observing that the RHS of this inequality does not depend onx
1,we integrate with
respect tox
1,from−L/2toL/2,and find

R

R
|u(x1,x2)|
2
dx1dx2
≤L

R
C≥
R
|u(ξ1,x2)|
2
dξ1

1/2C≥
R




∂u∂ξ1
(ξ1,x2)




2
dξ1

1/2
dx2
≤(after using the Cauchy–Schwarz inequality and renaming the dummy variable)
≤L
C≥
R

R
|u(x1,x2)|
2
dx1dx2

1/2C≥
R

R




∂u
∂x1
(x1,x2)




2
dx1dx2

1/2
.
This implies thePoincaré inequality
C≥
R

R
|u(x1,x2)|
2
dx1dx2

1/2
≤L
C≥
R

R




∂u
∂x1
(x1,x2)




2
dx1dx2

1/2
,(4.11)
as well as the following form, which is used more often:
C≥
R

R
|u(x1,x2)|
2
dx1dx2

1/2
≤L
C≥
R

R
|∇u(x1,x2)|
2
dx1dx2

1/2
.(4.12)
The Poincaré inequality in higher dimensions can be proved similarly. In three di-
mensions, the form that we will often use reads
C≥
R
3
|u(x)|
2
dx

1/2
≤L
C≥
R
3
|∇u(x)|
2
dx

1/2
. (4.13)
Remark 4.3(cf. Remark 4.2) Note that a coefficientL(with dimension of length)
appears in the RHS of (4.11) and (4.12). This length is given in the assumption onu
(it vanishes outside−L/2<x
1<L/2),and both sides of (4.11) and of (4.12) have
the same dimension, namelyUL.
Another frequently used version of the Poincaré inequality (which we state without
proof) relates to functionsudefined on a bounded domain˛whose average on˛
vanishes. We will use this inequality in space dimensiond=2 or 3, but for any space
dimensiondit reads as follows: There exists a constantc=⊃−¯=such that
5
More precisely, we prove one of the forms of this inequality; another form (valid for space-
periodic functions) is given later in this section.

20 I Introduction and Overview of Turbulence

¯
|u(x1,...,xd)|
2
dx1...dxd≤⊃−¯=

¯
|∇u(x1,...,xd)|
2
dx1...dxd(4.14)
for all functionsusuch that

¯
u(x1,...,xd)dx1...dxd=0. (4.15)
This inequality will be frequently used for space-periodic functions with zero av-
erage on the period.Note that the constant⊃−¯=≥which is not easy to determine,
has dimensionL
2
(square of a length; see Remark 4.3). In the periodic case, we write
this constant more explicitly as⊃−¯==˜⊃−¯=≤
2
,whereLis the smallest period. In
this case, the constant˜⊃−¯=depends only on the “shape” of¯≥in the sense that it is
invariant under dilatation; see Remark 4.2.
More Inequalities
The methods of functional analysis employed throughoutthis volume rely heavily on
the use of some relatively simple and well-known inequalities, as well as on more so-
phisticated ones. For the convenience of the reader, we list here (without proof) all
inequalities in the first category. We shall then also list those in the second category –
namely, the Sobolev inequalities and some of their variants, which extend (4.8
(4.9
Schwarz’s inequality:
ab≤
1
2
C
εa
2
+
b
2
ε

(4.16)
for all real numbersa,band allε>0.
Young’s inequality:
ab≤
1
p
a
p
+
1
p
R
b
p
R
(4.17)
for alla,b >0 and all 1<p<∞,withp
R
=p/(p−1)(i.e., 1/p+1/p
R
=1).Also,
ab≤
ε
p
a
p
+
1
ε
1/( p−1)
p
R
b
p
R
(4.18)
for alla,b,p,p
R
as before and allε>0.
Hölder’s inequality:

¯
u(x)v(x)dx≤
C≥
¯
|u(x)|
p
dx

1/pC≥
¯
|v(x)|
p
R
dx

1/p
R
(4.19)
forall measurable functionsuandvforwhich the right-hand side is finite. Here, 1<
p<∞,p
R
=p/(p−1),and¯is an arbitrary open set inR
d
,d∈N.Also,

¯
u(x)v(x)dx≤sup
x∈¯
|u(x)|

¯
|v(x)|dx. (4.20)
A weaker form of (4.20 uappears in the sequel.
In addition to the inequalities just listed, we use extensively the Poincaré inequalities

4 Function Spaces, Functional Inequalities, and Dimensional Analysis 21
(4.11), (4.12), and (4.14), as well as related forms of this inequality discussed in the
main text.
Lebesgue Spaces
We have recalled already the definition of the spaceL
2
−¯=of square integrable func-
tions on a domain˛inR
d
,d∈N.This is one of the so-called Lebesgue spaces. More
generally, for any 1≤p<∞,we have the Lebesgue spaceL
p
−¯=≥which is the
space of measurable functions whose absolute value to thepth power is integrable:

˛
|u(x)|
p
dx<∞.
This space is endowed with the norm
ıuı
L
p
−¯==
C≥
˛
|u(x)|
p
dx

1/p
. (4.21)
The spaceL
2
−¯=is a particular case of the preceding spaces, and it is the only one
whose norm is associated with an inner product. Because of the frequent use of the
L
2
-norm, we denote it simply by|·|and denote the associated inner product by(·,·).
The limiting casep=∞can also be considered. The spaceL

−¯=is the space
of measurable functions that are uniformly bounded almost everywhere on.More
precisely, a measurable functionu=u(x)on˛belongs toL

−¯=if and only if
there is a numberM≥0 such that|u(x)|≤Mfor almost everyxin˛(i.e., except
on a set of measure zero). The smallest suchMis the norm ofu,which is denoted
ıuı
L

−¯=.
With those spaces in mind, the Hölder inequality (4.19) reads as
(u,v)ˇıuı
L
p
−¯=ıvı
L
p
R
−¯=
(4.22)
for allu∈L
p
−¯=and allv∈L
p
R
−¯=≥wherep
R
=p/(p−1)(i.e., 1/p+1/p
R
=1),
1≤p≤∞.
Higher-Order Sobolev Spaces
A number of technical function spaces will be introduced as needed, especially in
Chapter II. However, at this point we would like to mention the higher-order Sobolev
spacesH
m
−¯=≥which are central in the mathematical theory of partial differential
equations.
For any domain˛inR
d
,which may be bounded or unbounded (possibly˛=R
d
)
and whose boundary;¯may or may not be smooth:
6
for integerm≥1,we denote
byH
m
−¯=the space of square integrable real functionsuon˛whose distributional
derivatives of orders up tomare also square integrable. In mathematical notation,
6
Let us mention that there are some difficulties with nonsmooth domains, a question that is
not addressed in this monograph.

22 I Introduction and Overview of Turbulence
H
m
−¯=={u∈L
2
−¯=;D
α
u∈L
2
−¯=≥[α]≤m},
whereα=(α
1,...,αd)∈N
d
is a multi-integer, [α]=α 1+···+α d,andD
α
is a
short notation for

α1+ ··· +α d

α1
x1···∂
αd
xd
.
We endow this space with the following inner product and norm:
ıuı
H
m
−¯=={((u, u))H
m
−¯=}
1/2
,
((u, u))
H
m
−¯==
m
R
k=0
1
L
2(m−k)
R
[α]=k

¯
D
α
u(x)D
α
v(x)dx. (4.23)
Here,Lis a typical length scale associated with¯;we introduced it in (4.23
the RHS dimensionally homogeneous. In the mathematical context we usually use
nondimensional variables, and thenL=1.The spaceH
m
−¯=is the Sobolev space
of orderm;it is a Hilbert space for the inner product((·,·))
H
m
−¯=and the associated
norm.
Of course, form=1,we recover the spaceH
1
−¯=previously mentioned. Note
that, likeH
1
−¯=(see (4.3
u∈H
m
−¯=if and only ifthere exists a sequence of smooth
functionsu
j,compactly supported in¯ ,such thatıu jıH
m
−¯=
remains bounded andu jis converging tou(in the distribution
sense) asj→∞.
(4.24)
Sobolev spaces based on the Lebesgue spacesL
p
−¯=for any 1≤p≤∞can also
be defined. They are very important in the mathematical theory of partial differential
equations, but they will not be necessary for our purposes in this monograph.
Sobolev Embeddings and Inequalities
An important property of Sobolev spaces are the Sobolev embeddings, based on in-
equalities similar to (4.8
ties. For instance, (4.8
H
1
(R
2
)⊂L
4
(R
2
);
such inequalities are valid more generally for any (smooth but not necessarily
bounded) domain¯⊂R
2
,and imply similarly for such a domain that
H
1
−¯=⊂L
4
().
In the same manner, in space dimension 3, (4.9
H
1
(R
3
)⊂L
4
(R
3
)
and, for any (smooth but not necessarily bounded) domain¯⊂R
3
,

4 Function Spaces, Functional Inequalities, and Dimensional Analysis 23
H
1
−¯=⊂L
4
().
In general, for 1≤m<d/2,wheredis the space dimension,
H
m
−¯=⊂L
q
−¯=≥ (4.25)
where 1/q=1/2−m/d .Form=d/2 (e.g., for the important case ofd=2 and
m=1),the functions inH
m
−¯=belong toL
q
−¯b),for any finiteq(1≤q<∞)and
for any bounded and smooth subdomain˛
bof.The Sobolev embedding (4.25
derived from an inequality similar to (4.8
(dimensionally invariant):
C≥
R
d
|u(x)|
q
dx

1/q
≤⊃−⊥≥ ⊂≥ ¯=
Z
[α]=m
C≥
R
d
|D
α
u(x)|
2
dx

1/2
, (4.26)
where the constant⊃−⊥≥ ⊂≥ ¯=depends onm, d,and the shape of.Note that both
sides of this inequality have the dimension ofUL
d/q
=UL
d/2−m
,and the constant
⊃−⊥≥ ⊂≥ ¯=is a constant that has no dimension.
Remark 4.4Unlike for (4.8 ⊃−⊥≥ ⊂≥ ¯=in (4.26
difficult to obtain; and the proof of (4.26
and (4.9
Adams [1975], Lions and Magenes [1972], and Mazja [1985]; see also a summary in
Temam [1997, Chap. II].
Remark 4.5By interpolation, one can supplement the embedding (4.25
large collection of similar inequalities. Indeed, using the Hölderinequality (4.19), it
is easy to show that
C≥
R
d
|u(x)|
r
dx

1/r

C≥
R
d
|u(x)|
2
dx

d(q−r)
r(q−2)
C≥
R
d
|u(x)|
q
dx

d(r−2)
r(q−2)
(4.27)
when 2≤r≤q.Hence, with (4.26
C≥
R
d
|u(x)|
r
dx

1/r
≤⊃−⊥≥ ⊂≥ ¯=
C≥
R
d
|u(x)|
2
dx

d(q−r)
r(q−2)
C
Z
[α]=m

R
d
|D
α
u(x)|
2
dx

dq(r−2)
2r(q−2)
.(4.28)
Compact Mappings, the Rellich Lemma, and Compact Sobolev Embeddings
We conclude this section with a more technicalnonintuitiveconcept: compact embed-
dings and the Rellich lemma, showing that certain Sobolev embeddings are compact.
The Sobolev embedding (4.25
following inequality holds (cf. (4.26
ıuı
L
q
−¯=≤cıuı H
m
−¯=

24 I Introduction and Overview of Turbulence
for some constantcindependent ofu.This means that the identity operator is contin-
uous fromH
m
−¯=intoL
q
().This holds under the condition that 1/q=1/2−m/d ,
for 1≤m<d/2 or in the critical casem=d/2 as just explained. When the domain
¯is bounded,L
q
−¯=is included inL
p
−¯=for anyp<q.For simplicity, we consider
only the casem=1.Then, if¯is bounded and smooth, we have the embedding
H
1
−¯=⊂L
p
−¯=
for 1/p >1/2−1/d .The Rellich lemma (see e.g. Adams [1975]) asserts that this
embedding is not only continuous but alsocompact.This means that a bounded set
inH
1
−¯=is precompact as a subset ofL
p
().One of the important consequences of
this property is that any sequence{u
n}n∈Nof functions whose norms inH
1
−¯=are
uniformly bounded contains a subsequence{u
nj
}j∈Nthat converges in the norm of
L
p
−¯=to some elementuinH
1
().The usual notationfor compact embedding is
H
1
−¯=⊂⊂L
p
−¯=≥ (4.29)
for 1/p >1/2−1/dand for¯bounded and smooth. This important result is often
invoked in this monograph, especially for proving that the Stokes operator has a
compact inverse, whence we deduce the existence of an orthonormal basis of eigen-
functions of the Stokes operator (see Section 6 in Chapter II).
Alternatively, saying that the embedding (4.29
weakly convergent inH
1
−¯=is strongly convergent inL
p
−¯=for such values ofp,
that is, convergent in theL
p
−¯=norm. See Appendix A.1 in Chapter II for some
explanations on weak convergences.

II
Elements of the Mathematical Theory
of the Navier–Stokes Equations
Introduction
The purpose of this chapter is to recall some elements of the classical mathemati-
cal theory of the Navier–Stokes equations (NSE). We try also to explain the physical
background of this theory for the physics-oriented reader.
As they stand, the Navier–Stokes equations are presumed to embody all of the
physics inherent in the given incompressible, viscous fluid flow. Unfortunately, this
does not automatically guarantee that the solutions to those equations satisfy the
given physics. In fact, it is not even guaranteed a priori that a satisfactory solution
exists. This chapter addresses the means for specifying function spaces – that is, the
ensembles of functions consistent with the physics ofthe situation (such as incom-
pressibility, boundedness of energy and enstrophy, as well as the prescribed boundary
conditions) – that can serve as solutions to the Navier–Stokes equations. An important
point is made that the kinematic pressure,p,is determined uniquely by the velocity
field up to an additive constant. Hence, one cannot specify independently the initial
boundary conditions for the pressure. This observation leads naturally to a represen-
tation of the NSE by an abstract differential equation in a corresponding function
space for the velocity field.
Two types of boundary conditions are considered:no-slip,which are relevant to
flows in domains bounded by solid impermeable walls; andspace-periodicbound-
ary conditions, which serve to study some idealized flows (including homogeneous
flows) far away from real boundaries.
A simplification, without compromising the physics or mathematical rigor, results
from the so-called Helmholtz–Leray decomposition, which is a generalization of the
decomposition of any vector field into a gradient of a potential plus a solenoidal
(divergence-free) component. The generalization consists of taking into account the
accompanying boundary conditions. From the mathematical point ofview, it is pos-
sible that some solutions of the Navier–Stokes equations could be highly irregular,
even though no such irregularities are encountered in nature. However, these solu-
tions when smoothed in some sense could represent physical realities. Toward this
end, one can introduce the so-called weak solutions: essentially, an inner product of
the possibly irregular solutions with some sufficiently regular (nice
operation is somewhat similar to taking moments of the velocity field with respect
25

26 II Elements of the Mathematical Theory of the Navier–Stokes Equations
to some distribution function, yielding observable averages of flow quantities. How-
ever, care must be taken not to push this analogy too far, because weak solutions are
not the same as the statistical solutions.
The next major task undertaken in this chapter is the question of existence and
uniqueness of solutions to the Navier–Stokes equations. In three dimensions, where
all practical fluid flows arise, we find that this is still an open problem, suggest-
ing that the 3-dimensional NSE are not a complete description of fluid flows. This
is not too unexpected, because we know (Schlichting [1979, Chap.III]) that these
equations have been obtained under the natural but unverified hypothesis that the
normal and shearing stresses are strictly linear functions of the rate of deforma-
tion. The departure from linearity may be so small that there are no experimentally
observable consequences.Yet an absence of the corresponding correction in the
equations can be significant enough to reveal mathematical pathologies. It should
be remarked here that other (even small) additional terms serve to regularize the
NSE in three dimensions (Lions [1969]). We do not addressthis problem here. On
the other hand, the corresponding situation in 2-dimensional flows is well in hand,
with existence and uniqueness of both strong and weak solutions proved beyond any
challenge.
Another important issue raised in this chapter addresses the problem that arises be-
cause not all the terms in the Navier–Stokes equations belong to the same function
space. Thus, all the terms linear in the velocity fielduare divergence-free, whereas
the inertial term(u·∇)uis not. In principle, some mathematical operationsinvolving
pairs of quantities (say,aandb)require, in concert with our intuitive physical senses,
thataandbbe of the “same kind” – not “mixing apples and oranges.” Mathemati-
cally speaking, this can be articulated as requiring that, in general,aandbbelong to
the same function space. But sometimes, as in the present case, we must deal with
pairs that arenotin the same function space. In order to make the necessary opera-
tions (e.g., inner products) meaningful, it is necessary to introduce special rules. The
concept of dual spaces, mentioned in the body of this chapter and dealt with at greater
length in Appendix A, serves this purpose.
We address the analyticity of solutions of the equations on the real-time axis. Phys-
ically, this corresponds to the demand that the solutions do not evidence any singular
behavior for all time, which is true for the case of flows in two dimensions. We re-
turn then to the dependence of the solutions on the boundary conditions and study the
special case of flows driven by a moving boundary. The chapter ends with Appen-
dices A and B, addressing some delicate mathematical questions raised in the body
of the chapter.
In short, the mathematical theory presented here serves to ensure that the solutions
of the Navier–Stokes equations, when they exist, are consistent with the physics
underlying the flow of incompressible, viscous flows. It leads to tools presented fur-
ther on, tools that allow us to obtain insight into the nature of turbulent flows without
the burden of elaborate numerical computations. At the same time, such tools are
useful for testing the validity of purely numerical efforts.

1 Energy and Enstrophy 27
We now describe the content of this chapter in words that are mathematically more
technical. In Section 1 we introduce the concepts of kinetic energy and enstrophy; in
Section 2, we present the corresponding boundary value problems to be considered.
We show also that, in the incompressible case, pressure (using the terminology of
meteorology) is a diagnostic variable: this means that, at each instant of time, pres-
sure is a function (functional t=0.Hence,
as we shall see, att=0 the initial field of velocities is prescribed, but not the initial
pressure distribution. In Section 3 we introduce the Helmholtz–Leray decomposi-
tion of vector fields, which is a version of the classical Helmholtz decomposition of
vector fields adapted to the boundary conditions. In Section 4 we give the weak for-
mulation of the NSE in which the pressure term disappears; this is the starting point
of the mathematical theory of these equations as introduced by the French mathema-
tician Jean Leray [1933, 1934a,b]. In Section 5 we introduce the function spaces that
we use, and we explain their physical meaning by relating them to the spaces of finite
kinetic energy and finite enstrophy. Section 6 is devoted to the study of the Stokes
operator, which is associated with the linear part of the Navier–Stokes equations and
is a fundamental tool in the mathematical theory of the NSE.
Section 7 recalls the major existence and uniqueness results for the Navier–Stokes
equations and introduces the concepts of weak (less regular) and strong (more regu-
lar) solutions. In Section 8 we recall the fact that, in dimension 2, the solutions of the
NSE are analytic in time – a property with important physical consequences. Sec-
tion 9 concerns analyticity in space in the 2- and3-dimensional cases; we also show
how this space analyticity implies the exponential decay of the Fourier coefficients
of the solutions. Section 10 introduces the function spaces needed for the Navier–
Stokes equations on the whole space(R
2
orR
3
).Section 11 gives the mathematical
framework needed for the treatment of the no-slip case with moving boundaries. We
introduce apropriate “background flows” to rewrite the equations with homogeneous
boundary conditions and added (linear) terms. In Section 12, we present universal
estimates for the energy dissipation in terms of a characteristic velocity and a charac-
teristic length; in Section 13, we introduce the nondimensional Grashof number and
discuss nondimensional estimates. Finally, in Appendices A and B, we give some
mathematical complements (for the mathematically oriented reader) and prove some
of the results alluded to in this chapter.
1 Energy and Enstrophy
We recall from Chapter I the Navier–Stokes equations for a viscous, incompressible,
homogeneous flow:
∂u
∂t
−ˆ}u+(u·∇)u+∇p=f, (1.1a)
∇·u=0. (1.1b)
With the density normalized toρ=1,the kinetic energy of a fluid with velocity
fieldu=u(x)and occupying a region¯is given by

28 II Elements of the Mathematical Theory of the Navier–Stokes Equations
e(u)=
1
2

¯
|u(x)|
2
dx. (1.2)
Another important quantity is the enstrophy,
E(u)=
d
R
i=1

¯
|∇ui(x)|
2
dx=
d
R
i,j=1

¯




∂u
i
∂xj
(x)




2
dx, (1.3)
whered=2 or 3 depending on whether the flow is 2- or 3-dimensional. As will be
seen shortly, the significance of enstrophy is that it determines the rate of dissipation
of kinetic energy.
If the domain¯is the whole space(R
2
orR
3
)and if the velocityudecays suffi-
ciently rapidly at infinity, then integration by parts using the divergence-free condition
(1.1b) implies the following representation for the enstrophy:
E(u)=

¯
|ω(x)|
2
dx, (1.4)
whereω=curluis the vorticity vector. (In fact,strophycomes from Greek, mean-
ingrotation.) Relation (1.4) is also valid when¯is not the whole space, provided the
boundary terms (resulting from the integration by parts) vanish. This is true in many
interesting cases.
Now assume, for simplicity, that¯=R
3
and thatuandpvanish and decay suffi-
ciently rapidly at infinity. Then, we take the scalar product of the momentum equation
(1.1a) withuand integrate over.The first term yields

∂u
∂t
(x,t)·u(x,t)dx=
1
2
d
dt

¯
|u(x,t)|
2
dx.
The contributions of the inertial term and of the pressure vanish owing to the diver-
gence-free condition:

¯
[(u·∇)u]·udx=
3
R
i=1
3
R
j=1

¯
uj
∂ui
∂xj
uidx=
1
2
3
R
i=1
3
R
j=1

¯
uj
∂(u
2
i
)
∂xj
dx
=−
1
2
3
R
i=1

¯
(∇·u)u
2
i
dx=0(1.5)
and

¯
∇p·udx=
3
R
i=1
∂p
∂xi
uidx=−

¯
p(∇·u)dx=0. (1.6)
Integrating by parts, we see that the viscous term yields
−ν

¯
u·udx=−ν
3
R
i,j=1

¯

2
ui
∂x
2
j
uidx=ν
3
R
i,j=1

¯
C
∂u
i∂xj

2
dx
=νE(u). (1.7)
We are then left with the equation of conservation of energy, which gives the rate of
decay of the kinetic energy:

2 Boundary Value Problems 29
d
dt
e(u)+νE(u)=

˛
f·udx. (1.8)
When there are no volume forces (i.e., whenf=0),the conservation of energy equa-
tion implies the decay of the kinetic energy by viscous effect at the rate−νE(u):
d
dt
e(u)=−νE(u). (1.9)
The calculations performed in (1.5), (1.6), and (1.7) play an essential role and are re-
peatedly used in the mathematical theory of the Navier–Stokes equations, even when
the domain˛is not the whole space.
2 Boundary Value Problems
The Navier–Stokes equations (1.1) are inescapably supplemented with initial and
boundary conditions that depend on the physical problem under consideration.
Throughout this book we mainly consider two distinct types of boundary condi-
tions: theno-slipboundary condition (for bounded domains) and thespace-periodic
boundary condition.
The no-slip boundary condition (flow past a rigid boundary) is one of the few that
correspondto a physically accessible boundary condition. Another physically acces-
sible boundary condition is the open boundary (i.e., an open surface of a flowing
fluid for part or all of the boundary
1
); we will not treat this case specifically, as it can
be handled by methods very similar to the no-slip case. The space-periodic case is
not a physically achievable one, but it is relevant on the physical side as a model for
some flows and is needed in the study of homogeneous turbulence. On the mathe-
matical side, the space-periodic case includes many of the difficulties encountered in
the no-slip case. However, the former is simpler to treat because of the absence of
boundaries (no boundary terms and often no boundary layers). Furthermore, using
the Fourier series as a tool simplifies the analysis and eases visualization of the phys-
ical aspects of the flow.
Flows in unbounded domains, or flows in the whole space(R
2
orR
3
),will not
be emphasized here. This case is often discussed in the literature but is not appro-
priate for our present studies owing, in particular, to the lack of compactness. This
characteristic raises technical difficulties that are related to the feasibility of standard
approximation methods and hence to the computability of quantities such as (1.2) and
(1.3). Very special assumptions about the flows must be made to obtain sensible re-
sults. Thus, we will discuss only the whole-space case (withouta decay condition at
infinity) as an idealization in the context of homogeneous statistical solutions; some
material preliminary to that study appears in Section 10 of this chapter.
1
In the language of partial differential equations, this corresponds to a Neumann-type bound-
ary condition; as we shall see, the no-slip case corresponds to a Dirichlet-type boundary
condition.

Another Random Scribd Document
with Unrelated Content

aikaan, rämeet niukkoine kasvullisuuksineen, joitakin vaivaisia pajuja
ja harvoja poppelia etäisellä niityllä. Vielä etäämpänä näkyi valkosia
pilkkuja, ne olivat kaupungit Marchienne pohjoisessa ja Montsou
etelässä. Lännessä aivan taivaan rannalla siinti kapeana
sinipunasena juovana Vandame-metsä. Talvihämärän vähäisessä
valossa tuntui hänestä, että koko tasanko oli mustunut Voreux'in
naapuruudesta, hieno hiilitomu oli peittänyt puut, tiet ja pellon.
Eniten hämmästytti häntä Scrape-joki, joka oli muutettu
kanavaksi. Yöllä hän ei ollut lainkaan nähnyt sitä. Tämä kanava
virtasi suoraan Voreux'sta Marchienne'en kuten kymmenen
kilometrin pituinen hopeanauha. Sitten hävisi se etäisyyteen kuten
leveä lehtikuja, jonka molemmin puolin oli istutettu tuuheita puita ja
jonka pinnalla liukui punasia purjeveneitä. Aivan kaivoksen luona oli
laivasilta, missä seisoi laivoja, niihin kuletettiin siltaa myöten vaunuja
täynnä hiiliä. Kaivoksen toisella puolen joki teki jyrkän mutkan
kulkien suoperäisen alangon läpi.
Etiennen katse siirtyi joelta työväen kylän punasiin kattoihin. Sen
jälkeen katsoi hän taas Voreux'ta, sen rumia rakennuksia ja
tiilikasoja, jotka poltettiin siinä. Puuaidan takana risteilivät yhtiön
rautatiekiskot, joita käytettiin kaivoksen palvelukseen.
Tällä hetkellä luultavasti laskettiin kaivokseen viimeisen ryhmän
työläisiä. Rautatiekiskoilla vikisivät vaunujen pyörät työläisten
lykätessä niitä. Nyt ei ympäristö tuntunut enää salaperäiseltä, eikä
siinä ollut käsittämättömiä ääniä ja tulia. Päivän valettua himmenivät
koksi- ja sulatus-uunit. Kuului ainoastaan pumpun tasainen huohotus
ikäänkuin ahneen ihmis-syöjän hengitys.
Silloin teki Etienne nopeasti päätöksensä. Ehkä hän luuli
näkevänsä Katarinan kirkkaat silmät kylän portilla. Tai ehkä häneen

tarttui yleinen tyytymättömyys, joka alkoi vallata Voreux'n. Hän ei
ymmärtänyt oikein itsekään, mikä pakotti hänet päättämään, hän
tiesi vain, että hän menee kaivokseen kärsiäkseen ja taistellakseen,
hän ajatteli raivolla niitä ihmisiä, joista Bonnemort oli hänelle
puhunut, tuota luhistunutta epäjumalaa, jolle kymmenet tuhannet
työmiehet antaa lihansa tietämättä edes, mikä epäjumala se on.
TOINEN OSA

I.
Gregoires'ien kartano, Piolaine, sijaitsi kaksi kilometriä itään
Montsou'sta Joisellen tien varrella. Se oli suuri neliskulmainen talo,
joka oli rakennettu seitsemänsata-luvun alussa ilman erikoista tyyliä.
Ennen oli kartanolle kuulunut laajoja maa-aloja, joista nyt oli enää
jälellä noin kolmekymmentä hehtaaria, jota tiheä muuri ympäröi.
Erittäin kuuluisia olivat hedelmäpuutarha ja kyökkitarha
erinomaisista hedelmistään ja vihanneksistaan, joita pidettiin seudun
paraimpina. Vanha lehmuskuja kulki kartanolta aina rautatieasemaan
saakka. Se oli seudun huomattavimpia merkillisyyksiä, sillä siinä oli
vähän puita.
Tänä aamuna oli Gregoire-pariskunta noussut kello kahdeksan
aamulla. Tavallisesti nousivat he vasta kello yhdeksän, sillä he
nauttivat paljosta nukkumisesta, mutta öinen myrsky oli
hermostuttanut heidät.
Gregoiren mennessä ulos katsomaan oliko tuuli tehnyt jotain
vahinkoa, meni hänen vaimonsa mukavissa aamupukimissaan
keittiöön. Hän oli pieni lihavanlainen, joka viidestäkymmenestä
vuodestaan ja harmaasta tukastaan huolimatta oli säilyttänyt
nukenilmeen kasvoissaan.

— Melaine, sanoi hän keittäjättärelle, paistakaappa nyt heti makea
kakku. Taikina on varmaankin jo kystä. Neiti nousee vasta puolen
tunnin kuluttua, niin ehtii kakku hänen aamusuklaatilleen. Se olisi
hauska yllätys hänelle.
Keittäjätär, laiha nainen, joka oli ollut heidän palveluksessaan jo
kolmekymmentä vuotta, naurahti hyväksyvästi:
— Niin, niin, siitä tulee mainio yllätys. Uuni on lämmin ja Honorine
voi auttaa minua.
Honorine oli kahdenkymmenen vanha tyttö, joka jo lapsena oli
otettu taloon ja palveli nyt sisäkkönä. Paitsi näitä kahta naista oli
talossa vain yksi miespalvelija, kuski, joka toimitti kaikki raskaat työt.
Puutarhuri vaimoineen piti huolen hedelmistä, vihanneksista ja
siipikarjasta.
Rouva Gregoire oli jo vuoteessaan keksinyt kakkuyllätyksen ja jäi
katsomaan, miten sitä pannaan uuniin. Heidän keittiönsä oli hyvin
suuri ja siisti, niin että heti näki, kuinka tärkeä merkitys sillä oli.
Hyllyillä kiilsi koko joukko keitto- ja paistinastioita y.m. keittiökaluja.
Ilmassa tunsi miellyttävää ruuan hajua. Kaapit ja laatikot olivat
täynnä kaikellaisia varastoja.
— No tehkää se oikein kauniin näköiseksi, sanoi hän mennessään
ruokasaliin.
Vaikka talossa oli lämmönjohtaja, niin siitä huolimatta paloi
ruokasalin kamiinissa tuli. Sisustus oli yksinkertainen: suuri pöytä,
tuolit, punanen astiakaappi, vain kaksi matalaa nojatuolia puhui
komeuden halusta ja pitkistä tunneista, joita siinä vietettiin kylläisten

päivällisten jälkeen. Gregoire'illa ei ollut tapana siirtyä
vierashuoneeseen, vaan viettivät he iltansa kodikkaasti siinä.
Gregoire palasi juuri ulkoa, hän näytti kuudenkymmenen ikäiseksi
vielä reippaalta, punaposkiselta, hyväntahtoisine kasvoineen, joita
ympäröi valkoset kutrit. Hän oli puettu parkkumipuseroon.
Hän oli puhutellut kuskia ja puutarhuria. Mitään erikoista ei ollut
tapahtunut. Joka aamu teki hän kiertokulun kartanossaan, josta hän
ammensi kaiken omistajan tyydytyksen.
— Entä Cecile? kysyi hän. — Eikö hän aiokaan nousta?
— En ymmärrä, vastasi vaimo. — Mutta minä olin kuulevinani
liikettä hänen huoneestaan.
Pöytä oli katettu, kolme kuppia seisoi valkosella pöytäliinalla.
Honorine lähetettiin katsomaan, miten neidin laita on. Tämä palasi
heti ja nauraen puoliääneen kertoi:
— Oi, jospa herra ja rouva näkisi neitiä!… Hän nukkuu, nukkuu
kuin enkeli… Niin herttaisesti, että nautinnokseen katsoo.
Isä ja äiti liikutettuina vaihtoivat silmäyksen.
— Tuletko katsomaan? sanoi mies hymyillen.
— Tulen, vastasi rouva nousten, minun pikkuraukkani.
He nousivat ylös portaita, — Cecilen makuuhuone oli ainoa komea
huone koko talossa. Sen seinät olivat verhotut vaalean sinisellä
silkillä ja huonekalut kiillotetut valkosiksi. Se oli hemmotellun lapsen
oikku, jonka vanhemmat mielellään täyttivät.

Sängyssä makasi nuori tyttö nojaten poskensa paljaaseen
käsivarteen. Hän ei ollut kaunis, hän oli liian hyvinvoipa ja kehittynyt
kahdeksantoista ikäiseksi, mutta hänellä oli hieno, maidonvalkea iho,
ruskea tukka, pyöreät kasvot sekä pieni pystynenä, joka hävisi
poskien väliin. Peitto oli solahtanut alas ja hän hengitti niin hiljaa,
ettei edes hänen upea rintansa kohonnut.
— Tuo kirottu tuuli varmaankin ei ollut antanut hänen nukkua, —
kuiskasi äiti.
Isä viittasi, että hän olisi hiljaa. Molemmat lähestyivät hiljaa
vuodetta ja kaikella hellyydellä ihailivat kauan odotettua tytärtään.
Hän oli syntynyt myöhään, silloin kuin he jo olivat kadottaneet toivon
saada lapsia. Hän näytti heistä hurmaavalta eikä lainkaan liian
lihavalta, päin vastoin he yhäti pelkäsivät hänen terveytensä
puolesta. Jokin heikko varjo solui hänen kasvojensa yli. He
pelästyivät, että hän herää ja menivät pois varpaillaan.
— Hiljaa! kuiskasi isä ovessa. — Jos hän ei ole nukkunut yöllä, niin
nukkukoon nyt.
— Tietysti, nukkukoon sydänkäpynen niin kauan kuin tahtoo, —
myönsi äiti. — Me voimme odottaa.
He palasivat ruokahuoneeseen ja istuivat nojatuoleihin, palvelijain
lämmittäessä suklaata ja nauraessa neidin hyvälle unelle. Mies luki
lehteä, vaimo kutoi villapeitettä. Ruokasalissa oli hiljaista ja
lämmintä, ei ainoakaan ääni kuulunut sinne.
Gregoiren omaisuuden, joka tuotti heille neljäkymmentä tuhatta
frankia korkoja vuodessa, muodostivat Montsou kaivoksen osakkeet.

He kertoivat mielellään rikkautensa alkuperästä, yhtiön
muodostamisen ajoilta.
Kahdeksantoistaluvun alkupuolella olivat kaikki Lillesta
Valencienneen saakka kiihkon vallassa hakien hiiltä. Ensimäisten
yrittelijäin, jotka myöhemmin muodostivat Anzinin yhtiön,
onnistuttua, tulivat kaikki sekasin päästä. Joka kunnassa alettiin
kaivaa maata, yhtiöitä kasvoi kuin sieniä. Kaikista näistä hulluista
hiilenetsijöistä oli vapaaherra Desrumaut innokkain, sivistynein ja
lujaluontoisin. Neljäkymmentä vuotta taisteli hän väsymättä kaikkia
esteitä vastaan. Ensimäiset etsinnöt epäonnistuivat, uusi kaivos oli
täytynyt monikuukautisen työn jälkeen hyljätä; maanvieremät
hävittivät hänen kaivantojaan, odottamattomat vedentulvat tappoivat
työläisiä, kymmenet tuhannet frankit menivät hukkaan. Vihdoin
onnistui hänen perustaa yhtiö Desrumaux, Fauquenoix ja kumpp.
Montsou kaivantojen kehittämiseksi. Kaivos oli alkanut tuottaa, kun
kaksi uutta yhtiötä, joista toinen kuului kreivi Cougnylle ja toinen
Cornille ja Jenardille, olivat vähällä musertaa hänet kilpailullaan.
Onneksi tuli 25 p:nä elokuuta v. 1760 laadittua näitten kolmen yhtiön
välillä sopimus, joka liitti ne yhteen. Siten oli Montsou kaivantojen
yhtiö muodostunut. Yhteinen omaisuus jaettiin sen ajan rahayksikön
mukaan kahteenkymmeneen neljään sou'hun, jokainen sou jakaantui
kahteentoista osaan — denier'eihin. Kun joka denier vastasi
kymmentä tuhatta frankkia, niin oli koko pääoma lähes kolme
miljoonaa frankia. Desrumaux, joka jo oli lähellä kuolemaa, voitti ja
hänen osakseen tuli kuusi sou'ta ja kolme denier'iä.
Siihen aikaan kuului Piolaine vapaaherralle. Maatilaan kuului kolme
sataa hehtaaria maata ja maatilanhoitajana oli Leon Gregoiren,
Cecilen isän esi-isä Honore Gregoire. Kun Montsou'n sopimus
laadittiin, oli Honore'lla säästössä viisikymmentä tuhatta frankia.

Vavisten ja epäillen antoi hän perää uskoen horjumattomasti
isäntäänsä ja hankki itselleen yhden denierin kymmenellä tuhannella
frankilla. Mutta häntä yhä vaivasi pelko, että hän on ehkä ryöstänyt
lapsensa. Hänen poikansa Eugene sai todellakin hyvin pieniä tuloja.
Sen lisäksi oli hän kyllin tyhmä jättääkseen virkansa ja hävittääkseen
jälelle jääneen isän perinnön, neljäkymmentä tuhatta frankia, jossain
uhkarohkeassa yrityksessä, niin että hän lopulta oli jotenkin
tukalassa tilassa. Mutta vähitellen alkoivat tulot osuudesta kasvaa ja
omaisuutta karttui jo Felicien aikana, joka oli onnellinen
toteuttaakseen isoisänsä toiveen, hän nim. osti takaisin Piolainen,
joka oli paloitettu eri ostajille. Hän sai sen polkuhinnasta sen jälkeen
kuin se oli tullut kansalliseksi omaisuudeksi. Seuraavat vuodet olivat
taas huonoja, täytyi odottaa kunnes vallankumous saapui. Leon
Gregoire sai ensimäisenä nauttia isoisänsä aran yrityksen runsaita
hedelmiä. Tämän pahaset kymmenen tuhatta kasvoivat yhtiön
varojen kasvaessa. Jo v. 1820 tuottivat ne sata prosentia s.o.
kymmenen tuhatta frankia. V. 1844 — kaksikymmentä tuhatta ja v.
1850 — neljäkymmentä. Ja vihdoin kahden vuoden kuluessa oli
voitto-osuus kasvanut uskomattomasti viiteenkymmeneen tuhanteen
frankiin. Denier'in arvo määrättiin Lillen pörssissä miljoonaksi ja oli
siten sadassa vuodessa noussut satakertaiseksi.
Gregoire kehoitettiin myymään osake, silloin kuin sen arvo oli yksi
miljoona, mutta hän kieltäytyi hymyillen tapansa mukaan tyyneesti.
Puoli vuotta sen jälkeen syntyi teollisuuspula ja osuuden arvo laski
kuuteen sataan tuhanteen. Mutta Gregoire hymyili kuten ennenkin
uskoen horjumatta kaivantoihinsa. Kyllä se nousee taas, ennemmin
romahtaa jumala taivaista, mutta ei heidän osakkeensa laske. Tähän
horjumattomaan uskoon liittyi syvä kiitollisuuden tunne arvopaperia
kohtaan, josta heidän perheensä on elänyt jo sata vuotta
tarvitsematta tehdä työtä lainkaan. Se oli ikäänkuin kotipyhyys, jota

he kaikessa itsekkyydessään jumaloivat. Ja kehdosta asti sirotteli se
heille lahjojaan; se tuuditti heitä pehmeillä vuoteilla, se ravitsi heitä
kylläisillä aterioilla. Ja sitä oli jatkunut polvesta polveen. Kuinka siis
voisikaan ruveta epäilemään kohtaloa, sen kautta voisikin ehkä
herättää sen epäsuosiota. Heidän uskonsa johtui osaksi
taikauskoisesta pelosta, että tuo miljoona voisi äkkiä hävitä, jos he
muuttaisivat sen rahaksi ja piilottaisivat pöytälaatikkoonsa. Ei, se
säilyi paremmin tuolla maan alla, josta sadat kivihiilenkaivajat,
sukupolvet nälkäisiä, kalvoivat sen esille heidän hyväkseen aina
vähän kerrallaan heidän jokapäiväiseksi tarpeekseen.
Onni seurasi tätä perhettä kaikessa. Vielä aivan nuorina oli
Gregoire nainut Marchiennen apteekkarin tyttären, joka oli ruma
tyttö ilman myötäjäisiä, mutta Gregoire jumaloi häntä ja tämä
vastasi samalla. Hän antautui talousaskareisiin, ihaili miestään ja oli
kaikessa samaa mieltä kuin hän. Siten olivat he eläneet
neljäkymmentä vuotta molemminpuolisessa rakkaudessa ja
ainaisissa huolehtimisissa toinen toisestaan. Se oli rauhallinen
olemassaolo, jolloin he hiljalleen kuluttivat neljäkymmentä tuhatta
frangia, tekivät säästöjä ja kuluttivat Cecileen. Tyttären myöhäinen
syntyminen oli joksikin ajaksi tuonut häiriön heidän laskuihinsa.
Mutta he täyttivät tähän päivään saakka kaikki Cecilen oikut:
hankkivat vielä toisen hevosen, kahdet ajoneuvot ja tilasivat pukuja
Pariisista. Siitä oli heillä rajaton ilo, vaikka itselleen he eivät sallineet
mitään komeutta. Kaikkinainen meno, joka ei tuottanut mitään,
tuntui heistä järjettömältä.
Äkkiä avautui ovi ja kuului voimakas huudahdus:
— Mitä tämä on? Syöttekö aamiaisen ilman minua?

Cecile oli juuri hypähtänyt vuoteeltaan, heittänyt yllensä valkosen
lämpimän aamupuvun ja hiukan silittänyt hiuksiansa.
— Emmehän toki. Mehän odotimme sinua, näetkö… Lapsiparkani,
sinä et kai saanut nukuttua tuulelta?
Nuori tyttö katsoi hämmästyneenä äitiin.
— Tuuli. Onko yöllä tuullut. Sitä minä en tiedä, olen nukkunut koko
yön aivan sikeästi.
Tämä tuntui heistä niin naurettavalta, että he kaikki kolme alkoivat
nauraa ja palvelijattaret, jotka toivat aamiaisen, nauroivat myös, niin
huvittavalta tuntui, että neiti oli nukkunut kaksitoista tuntia
yhtenään. Kun kakku tuotiin pöytään sai se kaikki vielä paremmalle
tuulelle.
— Kuinka, onko se jo paistettu! — huudahti Cecile. — Sepä oli
yllätys… Ja aivan lämmin. Mahtaa se maistua suklaan kanssa.
He istuivat vihdoinkin pöytään, missä suklaa höyrysi kupissa eikä
kotvan aikaa puhuttu muuta kuin kakusta. Melaine ja Honorine
seisoivat siinä myös kertoen, miten se paistui vakuuttaen, että oli
nautinto paistaa, kun näkee herrasväen syövän sitä sellaisella
nautinnolla.
Koirat alkoivat haukkua ja he luulivat että se oli soitto-opettajatar,
joka maanantaisin ja perjantaisin saapui Marchiennesta. Paitsi häntä
kävi myös kirjallisuuden opettaja. Siihen rajoittui nuoren tytön
opetus. Hän kasvoi ja oppi Piolainessa täydessä tietämättömyydessä,
viskeli kirjoja ikkunasta, jos ne eivät enää häntä huvittaneet.
— Se on herra Deneulin, sanoi Honorine palatessaan.

Hänen perästään astui saliin itse Deneulin, Gregoiren serkku,
suurkokoinen ja suurääninen mies, joka oli säilyttänyt vanhan
ratsuväen upserin vapaan käytöksen. Vaikka hän oli yli
viidenkymmenen, kiiluivat hänen viiksensä ja lyhyeksi ajeltu
tukkansa pikimustina.
— Niin, tässä minä olen, terve… istukaa vaan!
Koko perhe tervehti häntä. Sitten istuivat kaikki jatkaakseen
keskeytettyä juontiaan.
— Onko sinulla jotain sanottavaa minulle, kysyin Gregoire.
— Ei, ei yhtään mitään, — rauhoitti häntä Deneulin. — Ratsastin
vaan aamukävelylläni ja sattuessani ajamaan porttinne ohi, päätin
pistäytyä tervehtimään teitä.
Cecile kysyi, miten hänen tyttärensä Jeanne ja Lucie voivat. He
voivat mainiosti. Edellinen harrastaa maalaustaidetta, kuten
ennenkin ja toinen harjoittaa ääntään aamusta iltaan. Hän kertoi sen
iloisesti nauraen, mutta tuntui kuitenkin, ikäänkuin hän ei olisi oikein
vilpitön, vaan että jokin häiritsisi häntä.
— Onko kaikki hyvin mitä kaivokseen tulee?
— Oh, minä ja toverini olemme saaneet kylliksemme tästä pulasta.
Nyt täytyy meidän maksaa entisistä onnellisista vuosista. Liian monta
tehdasta on perustettu, liian monta rautatietä rakennettu, ja liian
paljon pääomaa pantu liikkeeseen toivossa saada runsaan saaliin.
Mutta nyt ovat rahat siinä kiinni eikä ole edes sen vertaa, että panisi
liikkeeseen… Onneksi ei asiat vielä ole aivan huonosti ja minä
suoriudun kyllä hyvin asiasta.

Hän oli samoin kuin hänen serkkunsa saanut periä yhden
denier'in. Mutta ollen kovin yritteliäs halusi hän hankkia
kuninkaallisen omaisuuden ja myi heti osakkeensa, kun sen arvo oli
noussut miljoonaan. Monta kuukautta oli hän hautonut erästä
suunnitelmaa. Hänen vaimonsa oli perinyt sedältään vain pienen
Vandamen kaivosalueen, missä oli kaksi kaivosta — Jean-Bart ja
Gaston-Marie, mutta ne olivat niin huonossa kunnossa, että tuotanto
tuskin peitti kulut. Hän haaveili Jean-Bartin saattamisesta kuntoon,
aikoen hankkia uuden koneiston ja laajentaa kaivosaukkoa
voidakseen käyttää enemmän työväkeä, mutta Gaston-Marien aikoi
hän jättää vedenpumppaamisen tarkoituksiin. Silloin voisi lapioida
kultaa, vakuutti hän. Suunnitelma oli kyllä oikea, mutta se oli niellyt
koko hänen miljoonansa ja tuo teollisuuspula oli sattunut juuri kun
koitti aika ruveta saamaan voittoja. Sitä paitsi oli hän huono isäntä;
työmiehiä kohtaan oli väliin kiihkeä, väliin hyvä. Vaimonsa kuoleman
jälkeen ryöstettiin häntä mitä julmimmalla tavalla ja tyttärilleen antoi
hän täyden vapauden. Vanhin aikoi ruveta näyttelijättäreksi ja nuorin
lähetti maalauksiaan näyttelyihin, joista ne oli jo kolmasti hyljätty.
Molemmat säilyttivät huolettomuutensa huolimatta uhkaavasta
hävityksestä, joka päinvastoin kehitti heistä erinomaisia
taloudenhoitajattaria.
— Näetkö, Leon, jatkoi Deneulin epävarmalla äänellä, — sinä teit
väärin kun et myynyt osakettasi samaan aikaan kuin minä. Nyt
laskee paperien arvo tuntuvasti ja sinä voit paljon hävittää… Jospa
olisit uskonut minulle rahasi, niin olisitpa nähnyt, mitä meidän
Vandamesta olisi tullut.
Gregoire joi kiirehtimättä suklaansa ja vastasi sitten:

— Ei koskaan. Sinä tiedät hyvin etten tahdo keinotella. Minä vietän
rauhallista elämää, olisi liian järjetöntä vaivata päätä asiahuolilla. Ja
mitä tulee Montsou'hun, niin laskekoon osakkeitten arvo, kyllä meille
riittää? Ei kannata olla liian ahne. Saat nähdä, että Montsou alkaa
pian taas edistyä ja tulee kannattamaan Cecilen lapsia ja lapsen
lapsia.
Deneulin kuunteli häntä hymyillen hämillään.
— Siis, jos minä pyytäisin sinua panemaan satatuhatta minun
yritykseeni, niin sinä kieltäytyisit?
Mutta huomatessaan Gregoirien levottomat kasvot, katui hän että
oli liian pikaisesti iskenyt asiaan ja päätti lykätä lainayrityksensä
tuonnemmaksi.
— No, se oli leikkiä, rauhoitti hän heitä, — laskin vain leikkiä.
Mahdollisesti sinä olet oikeassa; rahat, jotka ansaituttaa toisilla, on
varmin tulolähde, siitä ihmiset eniten lihoo.
Puhe siirtyi toisille aloille. Cecile alkoi taas puhua serkuistaan, sillä
he huvittivat häntä huomattavasti, vaikka hän samalla tunsi olevansa
loukattu heidän taipumuksistaan. Rouva Gregoire lupasi viedä
Cecilen jonakin auringonpaisteisena päivänä rakkaitten lapsukaisten
luo. Gregoire itse ei ottanut osaa keskusteluun, istuen hajamielisenä,
ääneti. Äkkiä lausui hän:
— Sinun asemassasi minä en enempää niskoittelisi, vaan alkaisin
hieroa sovintoa Montsou'n kanssa. He ovat hyvin halukkaita siihen ja
varmasti saisit rahasi takaisin.

Hän tarkoitti vanhaa vihamielistä kilpailua, joka vallitsi Montsou'n
ja Vandamen välillä. Huolimatta viimemainitun vähäpätöisestä
merkityksestä, oli mahtava naapuri raivoissaan nähdessään tämän
viiden neliökilometrin kaivannon keskellä avaria alojaan
kuudessakymmenessä seitsemässä kunnassa. Se oli turhaan
koettanut tappaa sitä ja nyt keinotteli se koettaen saada sen
haltuunsa polkuhinnasta, kun se oli häviön partaalla. Taistelu jatkui
ilman sovintoja, vaikka tirehtöörit ja insinöörit näön vuoksi säilyttivät
kohteliaan ystävällisiä suhteita.
Deneulinin silmät syttyivät.
— Ei koskaan! — huudahti hän. — Niin kauan kuin elän, ei
Montsou saa
Vandamea.
— Olin torstaina Hennebeau'n luona päivällisillä ja näin kyllä,
miten hän koetti päästä suosiooni. Jo viime syksyllä, kun kaikki nuo
herrat tirehtöörit olivat kokoontuneet tänne, koettivat he kaikin
tavoin hyvitellä minua… Kyllä minä tunnen heidät kaikki, nuo
markiisit ja kreivit, kenraalit ja ministerit! Ne ovat pelkkiä maantien
rosvoja, jotka ovat valmiit riistämään viimeisen paidankin, jos joutuu
heidän kynsiinsä.
Hän puhui pitkältä päästyään vauhtiin. Gregoire ei puolustanut
hallintoaan, jonka muodosti v. 1760 sopimuksen mukaan kuusi
valittua tirehtööriä, jotka itsevaltaisesti johtivat yhtiön asioita. Yhden
kuollessa valitsivat jälelle jääneet viisi toisen sijalle rikkaimpien ja
vaikutusvaltaisimpien osakkaitten keskuudesta. Piolainen omistaja
arveli, että nämä herrat rahanahneudessaan toisinaan menivät yli
kaikkien rajojen.

Melaine tuli korjaamaan pöydältä. Pihalla alkoivat koirat taas
haukkua. Honorine aikoi mennä ovelle, mutta Cecile, jonka oli kuuma
ja tukahduttava istua paljon syömisen jälkeen, nousi pöydästä
sanoen:
— Älä mene, se on varmaankin opettajattareni.
Deneulin oli myös noussut. Hän katsoi nuoren tytön jälkeen ja
kysyi hymyillen:
— No, milloinka tulee häät nuoren Negrelin kanssa.
— Se ei ole vielä päätetty, vastasi rouva Gregoire. — Se on vain
ehdotus… täytyy miettiä asiaa.
— Tietysti, vastasi Deneulin viekkaasti hymyillen.
— Luulempa, että täti ja sisaren poika… Minä en voi kuvitella,
miten rouva Hennebeau voisi syleillä Cecileä.
Mutta Gregoire suuttui. Oli sopimatonta siten puhua
kunnioitettavasta
naisesta, joka lisäksi oli nuorta miestä neljätoista vuotta vanhempi.
Häntä ei miellyttänyt, että tehtiin pilaa tuollaisista asioista.
Deneulin nauraen puristi hänen kättään ja meni.
— Ei se ollutkaan hän, sanoi Cecile palatessaan.
— Se on se työläisnainen kaivokselta, kaksine lapsineen,
muistatko, jonka me tapasimme… Saako hän tulla tänne?
Hetkisen olivat he kahden vaiheilla. Ovatko he hyvin likasia? Ei
erittäin, kenkänsä jättävät he portailla. Vanhemmat istuivat jo

mukavasti nojatuoleissaan sulattaen ruokaa. Mutta vilustumisen
pelko ja haluttomuus mennä ulos pakotti heidät tekemään
päätöksensä.
— Laskekaa heidät sisään, Honorine.
Huoneeseen tuli Maheu'n vaimo pienokaisineen, jotka olivat
nälkäiset ja viluiset ja pelästyivät aivan tultuaan tähän lämpimään
huoneeseen, missä tuoksui niin hyvin kakulle.

II.
Ikkunaluukkujen raoista alkoi vähitellen tunkea huoneeseen päivän
kalpeat valonsäteet. Ummehtunut ilma alkoi käydä yhä
raskaammaksi. Kaikki nukkuivat: Lenore ja Henri sylitysten, Alzire
taapäin taivutetuin päin ja isoisä Bonnemort, joka yksin makasi
Sakarian ja Jeanlinin vuoteessa, kuorsasi suu auki. Konttorista ei
myöskään kuulunut hengästystäkään, äiti oli myös vaipunut uneen,
samoin kuin lapsikin kylläisenä ja tyytyväisenä äidin rinnalla.
Käkikello alhaalla löi kuusi. Kylässä alettiin taas lyödä ovia ja
kolista puukengillä — hiililajittelijat kulkivat työhön. Sitten oli taas
kaikki hiljaa kello seitsemään.
Äkkiä kajahti jossain lyönti korvalle ja sitä seuraava parku, josta
Alzire heräsi; Hän arvasi heti, mitä kello oli ja juoksi avojaloin
herättämään äitiä.
— Äiti, äiti! nouse, kello on paljo. Sinunhan piti mennä tänään…
Hiljaa, sinä painat Estellan kuoliaaksi, ja hän veti lapsen esiin, jonka
äiti oli vähällä tukahuttaa allensa.
— Herranen aika! — mutisi Maheun vaimo hieroen silmiään, —
minä olen niin väsynyt, että nukkuisin vaikka koko päivän. Pue

Lenorea ja Henriä, minä otan heidät mukaani, Estelle jää sinun
kanssasi.
Hän peseytyi nopeasti, puki yllensä vanhan sinisen hameen, joka
oli siistimpi, ja harmaan villaröijyn, jonka hän edellisenä päivänä oli
paikannut.
— Mitä me saamme syödäksemme tänään, herra jumala! —
sopersi hän taas.
Hän meni alas, mutta Alzire palasi makuuhuoneeseen ja otti
mukaansa Estellan, joka oli alkanut huutaa. Hän oli tottunut
vaalimaan pientä lasta, kahdeksan vuoden vanhana oli hän jo pieni
nainen, joka osasi rauhoittaa ja ilahuttaa lasta. Hän pani pienokaisen
vuoteeseensa, joka vielä oli lämmin ja tuuditti sen antaen sormensa
imettäväksi. Tuskin oli pienokainen hiljennyt, kun syntyi uusi melu ja
Alziren täytyi välittää Lenoren ja Henrin riitaa. Nämä olivat hyvässä
sovussa ainoastaan silloin kun he nukkuivat. Niin pian kuin tyttö oli
herännyt kestitsi hän veljeään lyönneillään. Poika, joka oli sisartaan
kahta vuotta nuorempi, alistui nöyränä hänen lyönneilleen.
Molemmilla oli liian suuret, ikäänkuin turvonneet päät ja takkuinen
keltanen tukka. Alziren täytyi vetää sisar jaloista ja uhata
selkäsaunalla. Sitten pesi ja puki hän heidät silloin tällöin antaen
heille korvalle. Ikkunaluukkuja ei avattu, ettei Bonnemort häiriytyisi
unessaan. Hän kuorsasi raskaasti lasten melusta huolimatta.
— Minä olen valmis. Joko te siellä pian joudutte! — huusi Maheun
vaimo alhaalta.
Hän avasi luukut, sytytti tulen pannen hiiliä lisää. Hän toivoi, ettei
ukko ollut syönyt kaikkea lientä. Mutta kattila oli kuin nuoltu. Hän
keitti kourallisen makaroneja, joita hän oli kätkenyt varalle. Täytyy

syödä ne vedessä ilman voita. Eikä mitään muutakaan enää ollut
kaapissa, ei leivänmurusta eikä edes luuta, jota voisi kalvaa. Miten
olla, jos Maigrat ei anna velkaa tai jos Gregoiren herrasväki ei anna
hänelle rahaa? Täytyihän saada miehille ja tytölle ruokaa, kun he
tulevat työstä; onnettomuudeksi eivät ihmiset ole vielä oppineet
elämään ilman ruokaa.
— Tuletteko te sieltä viimein, — huusi hän vielä kerran. — Minun
olisi pitänyt jo kauan sitten lähteä.
Kun Alzire ja molemmat pienokaiset tulivat alas, jakoi hän
makaroneja kolmelle pienelle lautaselle. Hänellä itsellään ei muka
ollut nälkä? Vaikka Katarina oli jo edellisenä päivänä keittänyt poroja,
kaasi hän kuumaa vettä pannuun ja joi kaksi suurta kuppia sellaista
kahvia, joka enemmän muistutti likasta vettä kuin juomaa. Mutta
pitihän edes sillä pysyttää voimia.
— Nukkukoon isoisä niin kauan kuin tahtoo, sanoi hän Alzirelle, —
katso sinä Estellen perään; Jos se rupeaa huutamaan, niin tässä on
sokuripala, sulata se lämpimässä vedessä ja syötä sille lusikalla.
Sinähän olet kiltti tyttö etkä syö sitä itse.
— Entä koulu, äiti?
— Niin, koulu, no, se saa jäädä huomiseen. Tänään sinun täytyy
auttaa minua.
— Valmistanko ruokaa, jos sinä viivyt?
— Ruokaa… Ei, odota, kunnes tulen.
Alzire oli henkisesti hyvin kehittynyt, kuten kaikki, joilla on
ruumiillinen vika, ja osasi hyvin valmistaa ruokaa. Mutta hän käsitti,

miten oli asian laita eikä enää sanonut mitään.
Koko kylä oli jo noussut, lapsijoukot lähtivät kouluun kolisten
puukengillään. Kello löi kahdeksan ja Levaquen puolelta alkoi puhe
yltyä. Naisten päivä alkoi. He istuivat kahvipannujensa ympärillä
kädet puuskassa ja väsymättä pieksivät kieltään.
Jonkun kalpeat kasvot painautuivat ikkunaruutuun.
— Kuule, mikä uutinen! — sanoi naisen ääni.
— Ei ole aikaa, toiste, — vastasi Maheun vaimo minun täytyy heti
mennä.
Peläten kiusausta saada juoda kupin kuumaa, työnsi hän lapset
ulos ja lähti itse perästä. Tultuaan kadulle huomasi hän ihmeekseen,
että tuuli oli tyyntynyt. Äkkiä oli tullut suoja, taivas oli tullut
harmaaksi, seinät tahmeiksi, tiet olivat lokaisia ja mustia hiilipölystä,
niin että puukengät upposivat siihen.
Lenore tuli vallattomaksi. Hän tahtoi ottaa lokaa kenkänsä kärjellä
kuin lapiolla ja sai heti läimäyksen äidiltään.
He kulkivat tehdasrakennuksien ohi, kapeita polkuja peltojen yli
oiaistakseen pitkää matkaa ja saapuivat vihdoin maantielle.
Henri oli myös saanut äidiltään korvalle alkaessaan pyörittää
palloja mullasta ja nyt hiljenivät lapset väsyneinä huonosta tiestä ja
matkasta.
Maantie kulki suoraan Marchienneen näyttäen tervatulta nauhalta
keskellä punertavia peltoja. Toisella puolen teki se mutkan kulkien
Montsou'n ohi, joka oli rakennettu mäen rinteelle. Tällaisten teitten

varrelle, joita rakennetaan kaikkialla tehdaskaupungista toiseen,
alkaa ilmestyä rakennus toisensa jälkeen, niin että pian seutu
muuttuu työväen kyläksi. Molemmin puolin tietä oli pieniä tiilitaloja,
jotka olivat maalatut eri värisiksi, jotta olisi edes hiukan vaihtelua
tässä yksitoikkoisessa näköalassa. Siinä oli keltaisia, sinisiä ja mustia
taloja. Viimemainitut luultavasti siksi, että sen värisiksi kumminkin
täytyi ennemmin tai myöhemmin muuttua. Paikottain tapasi
kaksikerroksisen talon, joka kuului jollekin tehtailijalle, ja aikaan sai
edes vähän vaihtelua tässä yksitoikkoisessa talorivissä. Tiilistä
rakennettu kirkko neliskulmaisine tapulineen muistutti myöskin
jotakin sulatusuunia uutta mallia, sillä se oli mustan hiilipölyn
peittämä. Kaikkien noitten sokuri- ja köysitehtaitten sekä myllyjen
ohella vilisi joka askeleella ravintoloita ja oluttupia niin paljon, että
tuhatta taloa kohti oli yli viisisataa kapakkaa.
Kun Maheun vaimo lähestyi yhtiön rakennuksia ja varastohuoneita,
alkoi hän taluttaa lapsia kädestä. He kulkivat nyt Hennebeau'n talon
ohi, se oli rakennettu villan tapaiseksi, tiestä erotti sen aita ja pieni
puisto vaivaisine puineen. Juuri kun he kulkivat ohi pysähtyivät
portilla ajokalut, joista astui kunniamerkeillä koristettu herra ja
nainen turkiksissa, varmaankin vieraita Pariisista, sillä rouva
Hennebeau, joka näkyi eteisessä, huudahti hämmästyksestä.
— No, astukaa toki, kuhnuset? — haukkui Maheun vaimo vetäen
lapsia perässään…
He saapuivat Maigrat'in talon luo ja vaimo kävi yhä
levottomammaksi. Maigrat asui aivan Hennebeau'n vieressä, vain
muuri erotti hänen pienen talonsa tirehtörin komeasta
rakennuksesta. Siinä oli hänen pitkähköinen varastohuoneensa ja
tien puolessa puoti. Maigratilla oli kaupan vähän joka lajia,

ruokavaroja, hedelmiä, olutta ja astioitakin. Entisiin aikoihin oli hän
Voreux'in päällysmiehenä ja kaupitteli viinaa, mutta sitten ollen
johtajien suosiossa laajensi hän liikkeensä, niin että hän vähitellen oli
tappanut kaiken vähittäiskaupan Montsoussa. Hän osti tavaroita
tukuttain ja laajan työväen ostajapiirin tähden kävi hänelle
mahdolliseksi myydä huokeasta, sekä antaa velaksikin. Mutta hän oli
täydelleen riippuvainen yhtiöstä, joka oli rakentanut hänelle talon ja
varastohuoneen.
— Tulin taas teidän luoksenne, herra Maigrat, sanoi vaimo nöyrästi
nähdessään puotimiehen ovella.
Tämä katsoi häneen vastaamatta. Hän oli lihava ja korkea mies,
käytöksessään kohtelias, mutta tyly ja ylpeä siitä, ettei hän koskaan
peruuttanut päätöksiään.
— Ettehän aja minua pois kuten eilen. Täytyyhän meidän saada
jotain syötävää lauvantaihin saakka. Tosin olemme saaneet teiltä
velaksi… kuudenkymmenen frankin edestä kahden vuoden
kuluessa…
Hän puhui vaivoin etsien sanoja. Se oli vanha velka, jota oli
karttunut viimeisen lakon aikana. He olivat monasti luvanneet
maksaa, mutta eivät voineet. Sitä paitsi oli heille hiljakkoin sattunut
vaikeus — oli täytynyt maksaa suutarille kaksikymmentä frankia, sillä
hän oli uhannut kääntyä lain puoleen. Siksi heillä ei ollut nyt mitään.
Maigrat pudisti vain päätään vastaukseksi.
— Vaikkapa vain kaksi leipää, herra Maigrat. Minä en pyydäkään
kahvia… Vain kaksi kolmen naulan leipää päivässä…

— Ei! huusi kauppias vihdoin täyttä kurkkua.
Ovessa näyttäytyi hänen vaimonsa, heikko vaivainen olento, joka
vietti päiväkaudet konttorilaskujen ääressä uskaltamatta nostaa
päätään. Hän katosi yhtä pian, kun sai nähdä onnettoman vaimon
rukoilevan katseen puodin kynnyksellä. Kerrottiin, että hän
nurisematta luovutti aviovuoteensa naisostajille. Jos jonkun
kivihiilenkaivajan täytyi saada maksuaika lykätyksi, niin ei hänen
tarvinnut muuta kuin lähettää kauppiaan luo vaimonsa tai tyttärensä,
samantekevä oliko hän kaunis vai ruma, kunhan hän vain oli
taipuvainen.
Maheun vaimoa kiusasi noitten sameitten silmien loukkaava katse.
Hän suuttui, voisipa tuon vielä ymmärtää, jos hän olisi nuori, kuten
muinoin, mutta nyt, kun hänellä jo on ollut seitsemän lasta. Hän
tarttui lapsiinsa ja syöksyi ulos.
— Siitä ei teille koitu onnea, herra Maigrat, muistakaa se!
Nyt perusti hän koko toivonsa Piolainen herrasväkeen. Jolleivät he
anna viittä frankia, niin täytyy kuolla nälkään. Hän kääntyi
vasemmalle matkalla Joiselle'en. Tienristeyksessä oli tirehtörin
asunto. Se oli oikea palatsi, missä Pariisin herrat, ruhtinaat, kenraalit
ja muut hallituksen jäsenet joka syksy antoivat suuret päivälliset.
Matkalla teki hän laskunsa, miten viisi frankia oli käytettävä: ennen
kaikkea ostaisi hän leipää, sitten kahvia, kapan perunoita,
neljänneksen voita ja ehkä sitten myös vähän sianlihahyytelöä, sillä
isä tarvitsi välttämättä lihaa.
Häntä vastaan tuli pappi Joire Montsousta, hän astui varovasti
kantaen hamettaan, ettei se likaantuisi. Hän oli kaikille ystävällinen,

mutta pysyi syrjässä kaikista asioista säilyttääkseen hyvät välit sekä
isäntiin että työläisiin.
— Terve, herra pastori.
Mutta tämä ei edes pysähtynyt, vaan hymyillen lapsille astui
nopeasti ohi, niin että Maheun vaimo jäi suu auki seisomaan
maantiellä. Hän ei ollut uskonnollinen, mutta jostain syystä hän oli
luullut, että pappi antaisi hänelle jotain.
Täytyi taas kulkea likasta tietä eteenpäin. Oli jälellä vielä kaksi
kilometriä ja lapset olivat jo aivan väsyksissä riippuen hänen
käsivarsillaan. Tien molemmin puolin oli yhä samallaisia autioita
aloja, joitten välillä oli sammaltuneita aitoja ja savuttuneita
tehdasrakennuksia korkeine savutorvineen. Mutta edempänä laajeni
musta tasanko kuin liikkumaton meri ilman ainoatakaan puuta, vain
taivaan rannalla siinti Vandamen metsä.
— Kanna minut äiti.
Ja hän kantoi heidät vuorotellen. Tiellä oli vesilätäkköjä ja hän
kokosi hameensa, jottei saapuisi sinne liian likasena. Kolmasti oli hän
vähällä kaatua liukkaaseen lokaan. Vihdoin saapuivat he portaille,
mutta silloin hyökkäsi heidän kimppuunsa kaksi tavattoman suurta
koiraa haukkuen niin että lapset alkoivat parkua. Kuskin täytyi tulla
avuksi ajamaan ruoskalla koirat pois.
— Riisukaa kengät ja käykää sisälle, sanoi Honorine.
Tultuaan ruokasaliin äiti ja lapset hämmentyivät ja jäivät
äänettöminä oven suuhun. Täällä oli niin lämmin ja herra ja vanha
rouva nojatuolissa katsoivat niin tyystin heihin.

— Cecile, sanoi rouva, — tee nyt tehtäväsi.
Cecilen tehtävänä oli jakaa almuja. Vanhempain mielestä kuului se
hyvään kasvatukseen. Täytyy olla armelias ihmisiä kohtaan ja he
kutsuivat kotiansa jumalan huoneeksi. He ylpeilivät siitä, että he
harjoittavat järjellisesti armeliaisuutta, sillä he alati pelkäsivät että
heitä petetään ja että he siten edistäisivät paheita. Siksi he eivät
koskaan antaneet rahaa. Ei koskaan! ei ainoatakaan sou'ta, sillä
tiettyhän se oli, että niin pian kuin köyhä sai lantin, niin joi hän sen
heti. He jakoivat pääasiallisesti lämpimiä vaatteita sairaille ja lapsille.
— Oi, lapsiraukat! — huudahti Cecile, — kuinka he ovat kalpeat ja
viluiset. Honorine, tuo pian kääre kaapista.
Myöskin palvelijattaret, joitten ei tarvinnut huolehtia
päivällisestään, katsoivat säälien noita onnettomia. Sisäpalvelijatar
meni ylös makuuhuoneeseen, mutta keittäjätär pani kakun jätteet
pöydälle ja jäi siihen.
— Minulla on juuri jälellä kaksi villahametta ja pari huivia, jatkoi
Cecile. — Saatte nähdä kuinka pienokaisten tulee lämmin niissä.
— Kiitos, neiti kulta, sopersi Maheun vaimo, — te olette kaikki niin
hyviä.
Hänen silmänsä täyttyivät kyynelistä. Hän luuli jo olevansa varma
viiden frankin suhteen. Ei vain osannut keksiä, miten pyytäisi sitä,
ellei he itse ehdota. Lapset piiloutuivat äidin hameeseen katsoen
sieltä ahnein silmin kakkua.
— Onko teillä vain kaksi lasta kysyi rouva Gregoire
keskeyttääkseen vaitiolon.

— Oh, hyvä rouva, minulla on koko seitsemän.
Gregoire oli alkanut lukea taas sanomalehteä, mutta tämän
kuullessaan hän melkein hypähti paikaltaan.
— Seitsemän kappaletta! Mutta miksi, herran nimessä?…
— Se on järjetöntä, mutisi vanha rouva.
— Mitä sille voi, — vastasi Maheun vaimo ikäänkuin pyytäen
anteeksi. — Ei tullut tuota ajatelleeksi ja niitä yhä vain tipahteli. Ja
sitten tuottaahan nekin kun kasvavat. Eläisimmehän mekin yhtä
hyvin kuin toisetkin ilman isoisää, joka on tullut kaikille taakaksi.
Lapsista voi ainoastaan kaksi poikaa ja vanhin tytär laskeutua
kaivokseen. Mutta syöttää täytyy kaikkia, pienokaisiakin.
— Oletteko jo kauan työskennelleet kaivoksessa? kysyi rouva
Gregoire.
Maheun vannon kasvot kirkastuivat hymystä.
— Kyllä, kyllä. Minä työskentelin tosin vain, kunnes olin täyttänyt
kaksikymmentä vuotta. Kun sain toisen lapsen, niin sanoi lääkäri,
etten enää nousisi sieltä elävänä, jos laskeutuisin, sillä jotain
ruumiissani oli tullut epäjärjestykseen. Ja olihan kotonakin kyllin
tekemistä naimisissa ollen. Mutta miehen puolelta ovat ne
työskennelleet ikiajoilta saakka, aina siitä saakka kun alettiin
kaivannot.
Gregoire katsoi mietteissään naista ja hänen surkuteltavia lapsiaan
vahankalpeine kasvoineen ja haalistuneine hiuksineen, mikä teki
heidät rumiksi ja verettömiksi, kuten ovat tavallisesti ne, jotka eivät
milloinkaan saa syödä kyllikseen. Syntyi taas hiljaisuus eikä kuulunut

muuta kuin tulen räiskettä uunissa. Ilmassakin huokui porvarillista
onnellisuutta.
— Mutta missä hän viipyy! — huudahti Cecile kärsimättömänä. —
Melaine, mene ja sano hänelle, että käärö on kaapissa alhaalla,
vasemmalla puolen.
Mutta Gregoire lopetti ääneen mietiskelyjään, joita heräsi hänen
mieleensä näitä nälkäisiä ihmisiä nähdessä.
— Maailmassa on kyllä paljon pahaa, mutta, hyvä nainen, täytyy
myöntää, etteivät työläisetkään aina ole järkeviä. Sen sijaan, että he
säästäisivät, kuten talonpojat tekevät, niin kivihiilenkaivajat juovat,
tekevät velkoja, niin että lopuksi heillä ei ole, millä elättäisivät
perhettä.
— Herra on aivan oikeassa, vastasi vaimo, eivät kaikki elä niinkuin
pitäisi. Samaa minäkin sanon aina juopoille, jos alkavat valittaa…
Minun on sentään käynyt hyvin, mieheni ei juo. Luonnollisesti voi
sattua pyhäpäivisin seurassa, että hänkin ottaa vähän liikaa, mutta
siinä kaikki. Joi hänkin ennen naimisiin menoaan kuin porsas, jos
saan luvan sanoa, mutta nyt jumalan kiitos, on kaikki hyvin… Mutta
ei sekään paljoa auta. Sattuu päiviä, kuten tänään esimerkiksi, ettei
löydä ainoatakaan souta, vaikka kääntäisi nurin kaikki laatikot.
Hän tahtoi valmistaa heitä viiden frankin asiaan ja jatkoi
samallaisella hiljaisella äänellä kertoa velastaan, joka alussa oli niin
vähäpätöinen, mutta joka sitten oli kasvanut niin suunnattomasti.
Aluksi maksoivat he saamispäivänä, mutta sattuivat kerran
myöhästymään eivätkä nyt voi sitä peittää. Totta puhuen ei
kivihiilenkaivaja voi tulla toimeen ilman oluttuoppia, täytyyhän hänen
huuhtoa kurkkunsa pölystä. Siitä monet alkavat ja jäävätkin

istumaan kapakkaan. Mahdollisesti kaivajat saavat liian vähän
palkkaa, mutta eihän siihen ole kukaan syynä.
— Minä luulin, sanoi rouva Gregoire, — että yhtiö antaa teille
vapaan asunnon ja lämmön.
Vaimo katsahti kirkkaasti palavia hiiliä kamiinissa.
— Kyllähän me saamme hiiliä, lausui hän, eipä se ole parasta lajia,
mutta palaahan se sentään. Asunnosta maksamme kuusi frankia
kuussa, eihän se tosin ole paljo, mutta toisinaan on hyvin vaikea
saada sitäkään kokoon. Tänään voisi vaikka lyödä minut kuoliaaksi,
mutta ei voisi löytää ainoatakaan souta.
Herra ja rouva Gregoire olivat vaiti istuen mukavasti, heitä alkoi jo
kyllästyttää tuo pitkä kertomus kaikesta tuosta kurjuudesta. Vaimo
pelästyi, että oli ehkä suututtanut heidät ja lisäsi tyynen ja järkevän
naisen tavoin:
— En minä valita. Niinhän sitä on maailmassa, eikä sitä voi
muuttaa, vaan täytyy mukautua. Parasta on täyttää kunnollisesti
velvollisuutensa siinä, mihin jumala on asettanut. Eikö totta, herra?
Tähän myöntyi herra Gregoire täydelleen.
— Jos siten ajattelee, nainen hyvä, niin on yläpuolella
onnettomuuksien.
Honorine ja Melaine toivat vihdoinkin käärön. Cecile avasi sen ja
veti esiin kaksi hametta, liinoja, sukkia ja rukkasiakin. Kaikki sopi
mainiosti ja hän kiirehti palvelijatarta sitomaan kaikki nyyttiin. Hänen
soitto-opettajattarensa oli vihdoin tullut ja hän koetti saada Maheun
vaimon ja lapset pikemmin ulos.

— Me olemme nyt hyvin tukalassa tilassa, — sai vaimo vaivoin
sanottua, — jospa edes saisin viisi frankia.
Sanat tarttuivat hänen kurkkuunsa, kaikki Maheut olivat ylpeitä
eivätkä koskaan kerjänneet. Cecile katsoi pelästyneenä isään, mutta
tämä lausui aivan tyynesti, kuten se, joka on täyttänyt
velvollisuutensa.
— Mahdotonta. Se ei kuulu meidän tapoihimme.
Nuorta tyttöä liikutti äidin toivoton katse ja hän tahtoi edes tehdä
lapsille jotain mieliksi. Nämä tuijottivat yhä vieläkin kakkuun. Hän
leikkasi siitä pari palaa ja antoi niille.
— Tässä on teille.
Mutta sitten otti hän ne takaisin ja pyysi vanhaa sanomalehteä.
— Odottakaa, te voitte jakaa sisartenne ja veljienne kanssa.
Hänen vanhempansa katsoivat häneen liikutettuina hänen
työntäessä perhettä ulos. Lapset, joilla ei ollut leivän muruakaan
kotona, kantoivat kaikella kunnioituksella tuota makeaa kakkua
jäätynein kätösin.
Vaimo meni vetäen lapsia kädestä. Nyt hän ei nähnyt autioita
peltoja, likasta tietä eikä harmaata synkkää taivasta. Mennessään
Montsou'n ohi kääntyi hän päättäväisenä Maigrat'in luo ja pyysi
häneltä niin vakuuttavasti, että sai lopuksi kaksi leipää, kahvia, voita
jopa viisi frankiakin, sillä hän lainasi myös rahaa korkoa vastaan.
Mutta ei kauppias tarvinnutkaan Maheun vaimoa, vaan Katarinaa,
sen hän ymmärsi heti, kun kauppias kehotti lähettämään tytärtä
hakemaan puodista, mitä tarvitaan. No, sittenpähän nähdään.

Katarina kyllä antaa hänelle terveistä nyrkeistään, jos hän koettaisi
lähestyä häntä.

III.
Pienen tiilisen kirkon kello työväen kylässä No. 240 löi yksitoista.
Tässä kirkossa piti pastori Joire sunnuntaisin jumalanpalvelusta.
Kirkon vieressä oli koulun kivitalo. Vaikka ikkunat olivat suljetut,
kuului sieltä lasten tavailemista.
Leveällä tiellä, joka jakoi neljä yksitoikkoista rakennusryhmää
toisistaan, ei näkynyt ainoatakaan sielua. Pienet puistot näyttivät
synkiltä ja surullisilta; paljaine puineen ja likaisine maaperineen.
Kaikista savupiipuista savusi, kaikkialla laitettiin päivällistä, väliin
näkyi joku nainen ovella ja katosi taas. Vesiränneistä tippui vettä
vesisäiliöihin, vaikka sade oli jo lakannut. Nämä punaset rännit, joita
sade huuhtoi, tekivät kylän vähän vaihtelevamman näköiseksi, joka
äkkiä oli kohonnut kesken avaraa lakeutta ja jota reunustivat mustat
tiejuovat.
Palattuaan kylään Maheun vaimo poikkesi erään päällikön vaimon
luo, jolla oli jälellä perunasatoa. Syrjässä oli taloryhmä, jota pieni
poppelilehdikkö ympäröi, aina neljä taloa yhdessä puisto ympärillä.
Yhtiö oli määrännyt nämä rakennukset päälliköille ja vuorivoudeille,
jonka vuoksi työläiset nimittivät tätä kylän nurkkaa "Silkkisukaksi",
mutta omaa kyläänsä kutsuivat he: "maksa velkasi".

— Hohoi, vihdoinpa olemme perillä, — huudahti vaimo Maheu
tultuaan kotiin kädet täynnä kääreitä ja työntäen lapsia, jotka olivat
aivan nääntyneet ja likaiset.
Tulen ääressä huusi Estella täyttä kurkkua Alziren sylissä.
Sokurivesi oli aikoja loppunut ja tyttö koettaessaan saada lasta
tyyntymään oli keksinyt olla antavinaan hänelle rintaa. Toisinaan
tämä kepponen onnistui, mutta tällä kertaa se ei käynyt. Turhaan oli
hän päästänyt röijyn napeista ja painoi lasta laihaan rintaansa, lapsi
kimmastui vielä enemmän imiessään nahkaa, josta ei mitään
lähtenyt.
— Anna se minulle, — sanoi äiti pantuaan ostokset käsistään. —
Hän ei anna meidän puhua sanaakaan.
Niin pian kuin jukura oli imeytynyt hänen rintaansa, voitiin
rauhassa puhua. Kaikki oli järjestyksessä. Pikku emäntä oli siivonnut
huoneessa ja hoitanut tulta. Hiljaisuudessa kuului selvään isoisän
kuorsaaminen yhtä tasaisena ja säännöllisenä kuin ennenkin.
— Kas, kuinka paljon! — sanoi Alzire hymyillen katsoessaan
ostoksia.
— Tahdotko äiti, että minä keittäisin lientä?
Pöytä oli täynnä kääröjä, siinä oli vaatekäärö, kaksi leipää,
perunoita, voita, kahvia, sikuria ja puoli naulaa sianlihahyytelöä.
— Ehtii tuota vielä, sanoi äiti väsyneesti, — täytyy poimia
suolaruohoja ja purjulaukkaa… Minä keitän kyllä miehille, mutta
pane sinä perunat tulelle, niin saamme syödä niitä voin kanssa. Ja
kahvia, älä unohda kahvia!

Äkkiä muisti hän kakkua. Hän katsoi lapsiin, jotka kädet tyhjinä
tepastelivat lattialla. He olivat siis ahmineet sen matkalla. Ja hän
soimasi heitä aika lailla, vaikka Alzire koetti heitä puolustaa.
— Anna heidän olla, äiti. Jos se oli tarkoitettu minulle, niin ei ole
väliä. Heidänhän tuli nälkä kulkiessaan niin pitkälle.
Kello löi kaksitoista. Kuului koulusta palaavien lasten puukenkien
kolinaa. Perunat olivat valmiit ja kahvi keitettynä puoleksi sikurista
tuoksullaan kiihotti ruokahalua. Tehtiin tilaa pöydän toisessa päässä
ja äiti istui siihen. Lapset saivat tyytyä omiin polviinsa. Nuorin poika
katsoi koko ajan ahneesti hyytelöä, joka rasvasen paperin sisällä
kiinnitti hänen huomiotaan.
Äiti joi kahvia pienin kulauksin pitäen lasia molemmin käsin
lämmitelläkseen niitä. Ukko Bonnemort tuli myös. Tavallisesti heräsi
hän myöhemmin niin että aamiainen odotti häntä tulella. Nyt alkoi
hän murista, kun ei ollut lientä. Mutta kun hänelle sanottiin, ettei
aina saa sitä mitä tahtoo, alkoi hän ääneti syödä perunaansa.
Toisinaan nousi hän sylkeäkseen tuhkaan.
— Olin vallan unohtaa, äiti, sanoi Alzire, naapurivaimo kävi täällä…
— Menköön menojaan! keskeytti hänet äiti.
Hän ja Levaquen vaimo salaa eivät pitäneet toinen toisistaan.
Edellisenä päivänä oli Levaquen vaimo valittanut köyhyyttään, jottei
hänen tarvitsisi antaa velkaa. Mutta Levaqueilla olikin nyt juuri hyvät
päivät. Bouteloup oli maksanut kahdesta viikosta edeltäpäin. Mutta
perheelliset harvoin antoivat toisilleen lainaa.

Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade
Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.
Let us accompany you on the journey of exploring knowledge and
personal growth!
ebookultra.com