(eBook PDF) Turbulence: An

Introduction for Scientists and

Engineers 2nd Edition
Visit to download the full and correct content document:
https://ebooksecure.com/download/ebook-pdf-turbulence-an-introduction-for-scientist
s-and-engineers-2nd-edition/
PREFACE TO THE SECOND EDITION

R oughly 10 years have passed since the publication of the first edition of this book, and
perhaps it is timely to reflect on the ambitions and structure of the text in the light of
recent advances in the subject. In the first edition three recurring themes were: (i) a pref-
erence for working in real space rather than Fourier space, wherever possible; (ii) the importance
of giving equal weight to the large and small scales in the discussion of homogeneous turbulence;
and (iii) a note of caution as to what may be achieved through the use of direct numerical simu-
lations (DNS) of turbulence. As to the first two themes, developments in the intervening years
have, by and large, tended to reinforce my particular preferences. However, I have been proved
overly pessimistic on the third point; the relentless rise in computing power has allowed DNS to
become the preferred means of attack for the theoretician interested in the fundamental structure
of turbulence, though of course DNS remains of limited value to the engineer or applied scientist
interested in practical problems. In any event, the text has been modified to reflect this growing
importance of DNS in fundamental studies.
The motivation for a second edition has been the desire to simplify the text where possible,
while updating those sections where the subject has moved on. As a consequence, significant parts
of Sections 5.1.3, 5.3.6, 6.3, 6.6, 9.2, 9.5, and Chapter 10 have been rewritten, and a new section,
6.2.10, has been introduced on passive scalar mixing. Nevertheless the spirit of the book remains,
I hope, unchanged; a desire to bridge the gap between the elementary coverage of turbulence
found in undergraduate texts and the advanced (if often daunting) discussion in monographs on
the subject. Above all, I have tried at all times to combine the maximum of physical insight with
the minimum of mathematical detail.
Once again it is a pleasure to acknowledge the help of friends and colleagues, particularly
the many stimulating discussions I continue to have with Julian Hunt, Yukio Kaneda, Per-Åge
Krogstad, and many colleagues in Cambridge. Oxford University Press were, as always, a pleasure
to work with.

P. A. Davidson
Cambridge, 2014

preface | vii
PREFACE TO THE FIRST EDITION

T
urbulence is all around us. The air flowing in and out of our lungs is turbulent, as is the
natural convection in the room in which you sit. Glance outside; the wind which gusts
down the street is turbulent, and it is turbulence which disperses the pollutants that belch
from the rear of motor cars, saving us from asphyxiation. Turbulence controls the drag on cars,
aeroplanes, and bridges, and it dictates the weather through its influence on large-scale atmos-
pheric and oceanic flows. The liquid core of the earth is turbulent, and it is this turbulence which
maintains the terrestrial magnetic field against the natural forces of decay. Even solar flares are
a manifestation of turbulence, since they are triggered by vigorous motion on the surface of the
sun. It is hard not to be intrigued by a subject which pervades so many aspects of our lives.
Yet curiosity can so readily give way to despair when the budding enthusiast embarks on ser-
ious study. The mathematical description of turbulence is complex and forbidding, reflecting the
profound difficulties inherent in describing three-dimensional, chaotic processes.
This is a textbook and not a research monograph. Our principal aim is to bridge the gap be-
tween the elementary, heuristic accounts of turbulence to be found in undergraduate texts, and
the more rigorous, if daunting, accounts given in the many excellent monographs on the sub-
ject. Throughout we seek to combine the maximum of physical insight with the minimum of
mathematical detail.
Turbulence holds a unique place in the field of classical mechanics. Despite the fact that the
governing equations have been known since 1845, there is still surprisingly little we can predict
with relative certainty. The situation is reminiscent of the state of electromagnetism before it was
transformed by Faraday and Maxwell. A myriad tentative theories have been assembled, often
centred around particular experiments, but there is not much in the way of a coherent theoretical
framework.1 The subject tends to consist of an uneasy mix of semi-empirical laws and deter-
ministic but highly simplified cartoons, bolstered by the occasional rigorous theoretical result.
Of course, such a situation tends to encourage the formation of distinct camps, each with its own
doctrines and beliefs. Engineers, mathematicians, and physicists tend to view turbulence in rather
different ways, and even within each discipline there are many disparate groups. Occasionally re-
ligious wars break out between the different camps. Some groups emphasize the role of coherent
vortices, while others downplay the importance of such structures and advocate the use of purely

1
One difference between turbulence and nineteenth-century electromagnetism is that the latter was eventually
refined into a coherent theory, whereas it is unlikely that this will ever occur in turbulence.

viii | preface
statistical methods of attack. Some believe in the formalism of fractals or chaos theory, others do
not. Some follow the suggestion of von Neumann and try to unlock the mysteries of turbulence
through massive computer simulations, others believe that this is not possible. Many engineers
promote the use of semi-empirical models of turbulence; most mathematicians find that this is
not to their taste. The debate is often vigorous and exciting and has exercised some of the finest
twentieth-century minds, such as L.D. Landau and G.I. Taylor. Any would-be author embarking
on a turbulence book must carefully pick his way through this minefield, resigned to the fact that
not everyone will be content with the outcome. But this is no excuse for not trying; turbulence is
of immense importance in physics and engineering, and despite the enormous difficulties of the
subject, significant advances have been made.
Roughly speaking, texts on turbulence fall into one of two categories. There are those which
focus on the turbulence itself and address such questions as, where does turbulence come from?
What are its universal features? To what extent is it deterministic? On the other hand, we have
texts whose primary concern is the influence of turbulence on practical processes, such as drag,
mixing, heat transfer, and combustion. Here the main objective is to parameterize the influence
of turbulence on these processes. The word modelling appears frequently in such texts. Applied
mathematicians and physicists tend to be concerned with the former category, while engineers
are necessarily interested in the latter. Both are important, challenging subjects.
On balance, this text leans slightly towards the first of these categories. The intention is to
provide some insight into the physics of turbulence and to introduce the mathematical appar-
atus which is commonly used to dissect turbulent phenomena. Practical applications, alas, take a
back seat. Evidently such a strategy will not be to everyone’s taste. Nevertheless, it seems natural
when confronted with such a difficult subject, whose pioneers adopted both rigorous and heuris-
tic means of attack, to step back from the practical applications and try and describe, as simply as
possible, those aspects of the subject which are now thought to be reasonably well understood.
Our choice of material has been guided by the observation that the history of turbulence has,
on occasions, been one of heroic initiatives which promised much yet delivered little. So we have
applied the filter of time and chosen to emphasize those theories, both rigorous and heuristic,
which look like they might be a permanent feature of the turbulence landscape. There is little
attempt to document the latest controversies, or those findings whose significance is still un-
clear. We begin, in Chapters 1 to 5, with a fairly traditional introduction to the subject. The
topics covered include: the origins of turbulence, boundary layers, the log-law for heat and mo-
mentum, free-shear flows, turbulent heat transfer, grid turbulence, Richardson’s energy cascade,
Kolmogorov’s theory of the small scales, turbulent diffusion, the closure problem, simple closure
models, and so on. Mathematics is kept to a minimum and we presuppose only an elementary
knowledge of fluid mechanics and statistics. (Those statistical ideas which are required are intro-
duced as and when they are needed in the text.) Chapters 1 to 5 may be appropriate as background
material for an advanced undergraduate or introductory postgraduate course on turbulence.
Next, in Chapters 6 to 8, we tackle the somewhat refined, yet fundamental, problem of homo-
geneous turbulence. That is, we imagine a fluid which is vigorously stirred and then left to itself.
What can we say about the evolution of such a complex system? Our discussion of homogeneous
turbulence differs from that given in most texts in that we work mostly in real space (rather than
Fourier space) and we pay as much attention to the behaviour of the large, energy-containing
eddies as we do to the small-scale structures.
Perhaps it is worth explaining why we have taken an unconventional approach to homoge-
neous turbulence, starting with our slight reluctance to embrace Fourier space. The Fourier trans-
form is conventionally used in turbulence because it makes certain mathematical manipulations

preface | ix
easier and because it provides a simple (though crude) means of differentiating between large
and small-scale processes. However, it is important to bear in mind that the introduction of the
Fourier transform produces no new information; it simply represents a transfer of information
from real space to Fourier space. Moreover, there are other ways of differentiating between large
and small scales, methods which do not involve the complexities of Fourier space. Given that
turbulence consists of eddies (blobs of vorticity) and not waves, it is natural to ask why we must
invoke the Fourier transform at all. Consider, for example, grid turbulence. We might picture this
as an evolving vorticity field in which vorticity is stripped off the bars of the grid and then mixed
to form a seething tangle of vortex tubes and sheets. It is hard to picture a Fourier mode being
stripped off the bars of the grid! It is the view of this author that, by and large, it is preferable
to work in real space, where the relationship between mathematical representation and physical
reality is, perhaps, a little clearer.
The second distinguishing feature of Chapters 6–8 is that equal emphasis is given to both large
and small scales. This is a deliberate attempt to redress the current bias towards small scales in
monographs on homogeneous turbulence. Of course, it is easy to see how such an imbalance
developed. The spectacular success of Kolmogorov’s theory of the small eddies has spurred a vast
literature devoted to verifying (or picking holes in) this theory. Certainly it cannot be denied that
Kolmogorov’s laws represent one of the milestones of turbulence theory. However, there have
been other success stories too. In particular, the work of Landau, Batchelor, and Saffman on the
large-scale structure of homogeneous turbulence stands out as a shinning example of what can
be achieved through careful, physically motivated analysis. So perhaps it is time to redress the
balance, and it is with this in mind that we devote part of Chapter 6 to the dynamics of the large-
scale eddies. Chapters 6 to 8 may be suitable as background material for an advanced postgraduate
course on turbulence, or can act as a reference source for professional researchers.
The final section of the book, Chapters 9 and 10, covers certain special topics rarely discussed
in introductory texts. The motivation here is the observation that many geophysical and astro-
physical flows are dominated by the effects of body forces, such as buoyancy, Coriolis, and Lorentz
forces. Moreover, certain large-scale flows are approximately two dimensional and this has led to
a concerted investigation of two-dimensional turbulence over the last few years. We touch on the
influence of body forces in Chapter 9 and two-dimensional turbulence in Chapter 10.
There is no royal route to turbulence. Our understanding of it is limited and what little we
do know is achieved through detailed and difficult calculation. Nevertheless, it is hoped that this
book provides an introduction which is not too arduous and which allows the reader to retain at
least some of that initial sense of enthusiasm and wonder.
It is a pleasure to acknowledge the assistance of many friends and colleagues. Alan Bailey,
Kate Graham, and Teresa Cronin all helped in the preparation of the manuscript, Jean Delery
of ONERA supplied copies of Henri Werle’s beautiful photographs, while the drawing of the
cigarette plume and the copy of Leonardo’s sketch are the work of Fiona Davidson. I am grateful
to Julian Hunt, Marcel Lesieur, Keith Moffatt, and Tim Nickels for many interesting discussions
on turbulence. In addition, several useful suggestions were made by Ferit Boysan, Jack Herring,
Jon Morrison, Mike Proctor, Mark Saville, Christos Vassilicos, and John Young. Finally, I would
like to thank Stephen Davidson who painstakingly read the entire manuscript, exposing the many
inconsistencies in the original text.

P. A. Davidson
Cambridge, 2003

x | preface
CONTENTS

Part 1 The classical picture of turbulence

1. The ubiquitous nature of turbulence 3
1.1 The experiments of Taylor and Bénard 4
1.2 Flow over a cylinder 7
1.3 Reynolds’ experiment 8
1.4 Common themes 9
1.5 The ubiquitous nature of turbulence 13
1.6 Different scales in a turbulent flow: a glimpse at the energy cascade of
Kolmogorov and Richardson 18
1.7 The closure problem of turbulence 27
1.8 Is there a ‘theory of turbulence’? 29
1.9 The interaction of theory, computation, and experiment 30
2. The equations of fluid mechanics 34
2.1 The Navier–Stokes equation 35
2.1.1 Newton’s second law applied to a fluid 35
2.1.2 The convective derivative 38
2.1.3 Integral versions of the momentum equation 40
2.1.4 The rate of dissipation of energy in a viscous fluid 41
2.2 Relating pressure to velocity 43
2.3 Vorticity dynamics 44
2.3.1 Vorticity and angular momentum 44
2.3.2 The vorticity equation 48
2.3.3 Kelvin’s theorem 52
2.3.4 Tracking vorticity distributions 55
2.4 A definition of turbulence 57
3. The origins and nature of turbulence 61
3.1 The nature of chaos 62
3.1.1 From non-linearity to chaos 63
3.1.2 More on bifurcations 66
3.1.3 The arrow of time 69

contents | xi
 3.2 Some elementary properties of freely evolving turbulence 71
3.2.1 Various stages of development 73
3.2.2 The rate of destruction of energy in fully developed turbulence 77
3.2.3 How much does the turbulence remember? 81
3.2.4 The need for a statistical approach and different methods of taking averages 85
3.2.5 Velocity correlations, structure functions, and the energy spectrum 88
3.2.6 Is the asymptotic state universal? Kolmogorov’s theory 94
3.2.7 The probability distribution of the velocity field 97
4. Turbulent shear flows and simple closure models 105
4.1 The exchange of energy between the mean flow and the turbulence 107
4.1.1 Reynolds stresses and the closure problem of turbulence 108
4.1.2 The eddy viscosity theories of Boussinesq and Prandtl 111
4.1.3 The transfer of energy from the mean flow to the turbulence 114
4.1.4 A glimpse at the k-ε model 119
4.2 Wall-bounded shear flows and the log-law of the wall 122
4.2.1 Turbulent flow in a channel and the log-law of the wall 122
4.2.2 Inactive motion—a problem for the log-law? 127
4.2.3 Turbulence profiles in channel flow 129
4.2.4 The log-law for a rough wall 131
4.2.5 The structure of a turbulent boundary layer 131
4.2.6 Coherent structures 134
4.2.7 Spectra and structure functions near the wall 139
4.3 Free shear flows 142
4.3.1 Planar jets and wakes 142
4.3.2 The round jet 148
4.4 Homogeneous shear flow 152
4.4.1 The governing equations 152
4.4.2 The asymptotic state 156
4.5 Heat transfer in wall-bounded shear flows—the log-law revisited 157
4.5.1 Turbulent heat transfer near a surface and the log-law for temperature 157
4.5.2 The effect of stratification on the log-law—the atmospheric boundary layer 163
4.6 More on one-point closure models 169
4.6.1 A second look at the k-ε model 169
4.6.2 The Reynolds stress model 176
4.6.3 Large eddy simulation: a rival for one-point closures? 180
5. The phenomenology of Taylor, Richardson, and Kolmogorov 188
5.1 Richardson revisited 191
5.1.1 Time and length scales in turbulence 191
5.1.2 The energy cascade pictured as the stretching of turbulent eddies 195
5.1.3 The dynamic properties of turbulent eddies: linear and angular impulse 202
5.2 Kolmogorov revisited 211
5.2.1 Dynamics of the small scales 211
5.2.2 Turbulence-induced fluctuations of a passive scalar 221
5.3 The intensification of vorticity and the stretching of material lines 228
5.3.1 Enstrophy production and the skewness factor 228

xii | contents
 5.3.2 Sheets or tubes? 231
5.3.3 Examples of concentrated vortex sheets and tubes 234
5.3.4 Are there singularities in the vorticity field? 236
5.3.5 The stretching of material line elements 240
5.3.6 The interplay of the strain and vorticity fields 243
5.4 Turbulent diffusion by continuous movements 252
5.4.1 Taylor diffusion of a single particle 254
5.4.2 Richardson’s law for the relative diffusion of two particles 257
5.4.3 The influence of mean shear on turbulent dispersion 260
5.5 Why turbulence is never Gaussian 263
5.5.1 The experimental evidence and its interpretation 263
5.5.2 A glimpse at closure schemes which assume near-Gaussian statistics 267
5.6 Closure 268

Part 2 Freely decaying, homogeneous turbulence

6. Isotropic turbulence (in real space) 275
6.1 Introduction: exploring isotropic turbulence in real space 275
6.1.1 Deterministic cartoons versus statistical phenomenology 276
6.1.2 The strengths and weaknesses of Fourier space 281
6.1.3 An overview of this chapter 283
6.2 The governing equations of isotropic turbulence 293
6.2.1 Some kinematics: velocity correlation functions and structure functions 293
6.2.2 More kinematics: the simplifications of isotropy and the vorticity
correlation function 300
6.2.3 A summary of the kinematic relationships 304
6.2.4 Dynamics at last: the Karman–Howarth equation 308
6.2.5 Kolmogorov’s four-fifths law 310
6.2.6 The skewness factor and enstrophy production (reprise) 312
6.2.7 The dynamical equation for the third-order correlations
and the problem of closure 313
6.2.8 Closure of the dynamical equations in the equilibrium range 314
6.2.9 Quasi-normal-type closure schemes (part 1) 316
6.2.10 Passive scalar mixing in isotropic turbulence and Yaglom’s
four-thirds law 318
6.3 The dynamics of the large scales 320
6.3.1 The classical view: Loitsyansky’s integral and Kolmogorov’s decay laws 322
6.3.2 Landau’s angular momentum 323
6.3.3 Batchelor’s pressure forces 328
6.3.4 Saffman’s spectrum 333
6.3.5 A consistent theory of the large scales in Batchelor turbulence 342
6.3.6 A summary of the dynamics of the large scales 346
6.4 The characteristic signature of eddies of different shape 348
6.4.1 Townsend’s model eddy and its relatives 349
6.4.2 Turbulence composed of Townsend’s model eddies of different sizes 353
6.4.3 Other model eddies 356

contents | xiii
 6.5 Intermittency in the inertial-range eddies 357
6.5.1 A problem for Kolmogorov’s theory? 358
6.5.2 The log-normal model of intermittency 360
6.5.3 The β̂ model of intermittency 363
6.6 Measuring the distribution of energy and enstrophy across the different
eddy sizes 366
6.6.1 A real-space function which represents, approximately, the variation
of energy with scale 366
6.6.2 Relating energy distributions in real and Fourier space 374
6.6.3 Cascade dynamics in real space 379
7. The role of numerical simulations 394
7.1 What is DNS or LES? 394
7.1.1 Direct numerical simulations (DNS) 394
7.1.2 Large eddy simulations (LES) 398
7.2 On the dangers of periodicity 404
7.3 Structure in chaos 406
7.3.1 Tubes, sheets, and cascades 407
7.3.2 On the taxonomy of worms and clusters of worms 412
8. Isotropic turbulence (in spectral space) 419
8.1 Kinematics in spectral space 420
8.1.1 The Fourier transform and its properties 421
8.1.2 The Fourier transform as a filter 425
8.1.3 The autocorrelation function and power spectrum 427
8.1.4 The transform of the correlation tensor and the three-dimensional
energy spectrum 431
8.1.5 One-dimensional energy spectra in three-dimensional turbulence 434
8.1.6 Relating the energy spectrum to the second-order structure function 438
8.1.7 A footnote: singularities in the spectrum due to anisotropy 440
8.1.8 Another footnote: the transform of the velocity field 441
8.1.9 Definitely the last footnote: what do E(k) and E1 (k) really represent? 442
8.2 Dynamics in spectral space 445
8.2.1 An evolution equation for E(k) 445
8.2.2 Closure in spectral space 448
8.2.3 Quasi-normal type closure schemes (part 2) 454

Part 3 Special topics

9. The influence of rotation, stratification, and magnetic fields on turbulence 467
9.1 The importance of body forces in geophysics and astrophysics 467
9.2 The influence of rapid rotation and stable stratification 470
9.2.1 The Coriolis force 470
9.2.2 The Taylor–Proudman theorem 473
9.2.3 Properties of inertial waves 474
9.2.4 Turbulence in rapidly rotating systems 479
9.2.5 From rotation to stratification (or from cigars to pancakes) 485

xiv | contents
 9.3 The influence of magnetic fields I: the MHD equations 488
9.3.1 The interaction of moving conductors and magnetic
fields: a qualitative overview 489
9.3.2 From Maxwell’s equations to the governing equations of MHD 494
9.3.3 Simplifying features of low magnetic Reynolds number MHD 498
9.3.4 Simple properties of high magnetic Reynolds number MHD 499
9.4 The influence of magnetic fields II: MHD turbulence 503
9.4.1 The growth of anisotropy in MHD turbulence 505
9.4.2 The evolution of eddies at low magnetic Reynolds number 507
9.4.3 The Landau invariant for homogeneous MHD turbulence 513
9.4.4 Decay laws at low magnetic Reynolds number 514
9.4.5 Turbulence at high magnetic Reynolds number 516
9.4.6 The shaping of eddies by combined Coriolis and Lorentz forces 520
9.5 Turbulence in the core of the earth 522
9.5.1 An introduction to planetary dynamo theory 523
9.5.2 Numerical simulations of the geodynamo 529
9.5.3 Various cartoons of the geodynamo 532
9.5.4 An α2 model of the geodynamo based on inertial wave packets 538
9.6 Turbulence near the surface of the sun 553
10. Two-dimensional turbulence 559
10.1 The classical picture of 2D turbulence: Batchelor’s self-similar
spectrum, the inverse energy cascade, and the E(k)~k–3 enstrophy
flux law 560
10.1.1 What is two-dimensional turbulence? 560
10.1.2 What does the turbulence remember? 565
10.1.3 Batchelor’s self-similar spectrum 565
10.1.4 The inverse energy cascade of Batchelor and Kraichnan 567
10.1.5 Different scales in 2D turbulence 570
10.1.6 The shape of the energy spectrum and the k–3 law 571
10.1.7 Problems with the k–3 law 574
10.2 Coherent vortices: a problem for the classical theories 577
10.2.1 The evidence 577
10.2.2 The significance 579
10.3 The governing equations in statistical form 581
10.3.1 Correlation functions, structure functions, and the energy spectrum 582
10.3.2 The two-dimensional Karman–Howarth equation 586
10.3.3 Four consequences of the Karman–Howarth equation 586
10.3.4 The two-dimensional Karman–Howarth equation in spectral space 588
10.4 Variational principles for predicting the final state in confined domains 591
10.4.1 Minimum enstrophy 592
10.4.2 Maximum entropy 594
10.5 Quasi-two-dimensional turbulence: bridging the gap with reality 594
10.5.1 The governing equations for shallow-water, rapidly rotating flow 595
10.5.2 The Karman–Howarth equation for shallow-water, rapidly
rotating turbulence 597

contents | xv
Epilogue 601
Appendix 1: Vector identities and an introduction to tensor notation 603
A1.1 Vector identities and theorems 603
A1.2 An introduction to tensor notation 606
Appendix 2: The properties of isolated vortices: invariants, far-field properties,
and long-range interactions 611
A2.1 The far-field velocity induced by an isolated eddy 611
A2.2 The pressure distribution in the far field 613
A2.3 Integral invariants of an isolated eddy: linear and angular impulse 614
A2.4 Long-range interactions between eddies 617
Appendix 3: Hankel transforms and hypergeometric functions 620
A3.1 Hankel transforms 620
A3.2 Hypergeometric functions 621
Appendix 4: The kinematics of homogeneous, axisymmetric turbulence 623

Subject index 625

xvi | contents
PA R T 1

The classical picture of turbulence

Cross section through a turbulent jet carrying a passive scalar. The colours indicate the sca-
lar concentration. Note that the edge of the jet is highly convoluted. (See the discussion in
Section 4.3.1.)
Image courtesy of P.E. Dimotakis, California Institute of Technology.
CHAPTER 1

The ubiquitous nature

of turbulence
Vether it’s worth goin’ through so much, to learn so little, as the charity-boy said ven he got to the
end of the alphabet, is a matter o’ taste.
Charles Dickens, The Pickwick Papers

T
he study of turbulence is not easy, requiring a firm grasp of applied mathematics and
considerable physical insight into the dynamics of fluids. Worse still, even after the various
theoretical hypotheses have been absorbed, there are relatively few situations in which we
can make definite predictions!
For example, perhaps the simplest (and oldest) problem in turbulence concerns the decay of
energy in a cloud of turbulence. That is, we stir up a fluid and then leave it to itself. The turbulence
decays because of viscous dissipation and a natural question to ask is, how fast does the kinetic
energy decline? Theoretical physicists and applied mathematicians have been trying to answer
this question for over half a century and still they cannot agree. At times, one is tempted to side
with Weller the elder in The Pickwick Papers.
Nevertheless, there are certain predictions that can be made based on a variety of physical
arguments. This is important because turbulent motion is the natural state of most fluids. As you
read this book the air flowing up and down your larynx is turbulent, as is the natural convection
in the room in which you sit. Glance outside: if you are lucky you will see leaves rustling in the
turbulent wind. If you are unlucky you will see pollutants belching out from the rear of motor cars,
and it is turbulent convection in the street that disperses the pollutants, saving the pedestrians
from an unfortunate fate.
Engineers need to know how to calculate the aerodynamic drag on planes, cars, and buildings.
In all three cases the flow will certainly be turbulent. At a larger scale, motion in the oceans and
in the atmosphere is turbulent, so weather forecasters and oceanographers study turbulence. Tur-
bulence is also important in geophysics, since it is turbulent convection in the core of the Earth
that maintains the Earth’s magnetic field despite the natural forces of decay. Even astrophysicists

Turbulence: An Introduction for Scientists and Engineers. Second Edition. P. A. Davidson.

© P. A. Davidson 2015. Published in 2015 by Oxford University Press.
study turbulence, since it controls phenomena such as solar flares, the 22-year solar cycle, and the
rate of flow of mass through accretion discs onto the surfaces of young and dying stars.
It is the ubiquitous nature of turbulence, from the eruption of solar flares to the rustling of
leaves, that makes the subject both important and intriguing. The purpose of this chapter is to
give some indication as to just how widespread turbulence really is, and to introduce some of its
elementary properties. We start by describing some simple laboratory experiments.

1.1 The experiments of Taylor and Bénard

It is an empirical observation that the motion of a very viscous or slow-moving fluid tends to be
smooth and regular. We call this laminar flow. However, if the fluid viscosity is not too high, or
the characteristic speed is moderate to large, then the movement of the fluid becomes irregular
and chaotic, i.e. turbulent. The transition from laminar to turbulent motion is nicely illustrated
in a number of simple experiments, of which the most famous are probably those of Reynolds,
Bénard, and Taylor.
In 1923 Taylor described a remarkably simple, yet thought-provoking, experiment. Suppose
that we have two concentric cylinders and that the annular gap between the cylinders is filled
with a liquid, say water. The inner cylinder is made to rotate while the outer one remains sta-
tionary. At low rotation rates the fluid within the gap does what you would expect: it also rotates,
being dragged around by the inner cylinder. At higher rotation rates, however, something unex-
pected happens. At a certain critical speed, toroidal vortices suddenly appear, superimposed on
the primary circular motion. (This is shown schematically in Figure 1.1(a).) These axisymmetric
structures are called Taylor vortices, for an obvious reason, and they arise because of an instability
of the basic rotary flow. The net motion of a fluid particle is now helical, confined to a toroidal
surface.
The reason for the sudden appearance of Taylor vortices is related to the centrifugal force. This
tends to drive the rotating fluid radially outward. Below the critical speed this force is balanced by
a radial pressure gradient and there is no radial motion. However, the centrifugal force is greatest

(a) z (b) (c)

r

R d

Taylor vortices Wavy Taylor vortices Turbulent Taylor vortices

Figure 1.1 Flow between concentric cylinders. As the rotation rate of the inner cylinder increases,
the flow becomes progressively more complex until eventually turbulence sets in.

4 | the ubiquitous nature of turbulence

at points where the rotation rate is highest (near the inner cylinder) and so there is always a
tendency for fluid at the inner surface to move outward, displacing the outer layers of fluid. This
tendency is held in check by pressure and viscous forces. At the critical rotation rate, however, the
viscous forces are no longer able to suppress radial disturbances and the flow becomes unstable
to the slightest perturbation. Of course, all of the rapidly rotating fluid cannot move uniformly
outward because the outer fluid is in the way. Thus the flow breaks up into bands (or cells) as
shown in Figure 1.1(a). This is known as the Rayleigh instability.1
Now suppose we increase the rotation rate a little more, say by 25%. The Taylor vortices them-
selves become unstable and so-called wavy Taylor vortices appear. These have the appearance
of non-axisymmetric toroidal vortices which migrate around the inner cylinder (Figure 1.1(b)).
Note that, although complex, this flow is still laminar (non-chaotic).
Suppose we now increase the speed of the inner cylinder yet further. More complex, unsteady
structures start to emerge (so-called modulated wavy Taylor vortices) until eventually, when
the speed of the inner cylinder is sufficiently high, the flow becomes fully turbulent. In this
final state the time-averaged flow pattern resembles that of the steady Taylor vortices shown in
Figure 1.1(a), although the cells are a little larger. However, superimposed on this mean flow we
find a chaotic component of motion, so that individual fluid particles are no longer confined to
toroidal surfaces. Rather, as they are swept around by the mean flow, there is a constant jostling
for position, almost as if the particles are in a state of Brownian motion (Figure 1.1(c)). A typical
measurement of, say, uz (t) would look something like that shown in Figure 1.2. (Here ū is the
time-averaged value of uz .)
In fact, it is not just the rotation rate, , that determines which regime prevails. The kinematic
viscosity of the fluid, v, the annular gap, d, the radius of the inner cylinder, R, and the length of
the apparatus, L, are also important. We can construct three independent dimensionless groups
from , v, d, R, and L, say,
2 d3 R
Ta = , d/R , L/R.
v2
The first of these groups, Ta, is called the Taylor number. (Different authors use slightly different
definitions of Ta.) When the apparatus is very long, L  R, and the gap very narrow, d  R, it
happens to be the value of Ta, and only Ta, which determines the onset of Taylor vortices. In fact,
the critical value of Ta turns out to be 1.70 × 103 and the axial wavenumber of the vortices is
k = 2π/λ = 3.12/d, λ being the wavelength.
Let us now consider a quite different experiment, often called Bénard convection or Rayleigh–
Bénard convection. Suppose that a fluid is held between two large, flat, parallel plates, as shown

Figure 1.2 Variation of the axial component of velocity with

u time at some typical location in the annulus. When the flow
is turbulent there is a mean component of motion plus a
t random component.

1
Rayleigh identified the instability mechanism and produced a stability criterion for inviscid, rotating flows.
Taylor later extended the theory to viscous flows of the type shown in Figure 1.1(a).

the experiments of taylor and b énard | 5

 (a) (b) (c)
COLD (T = T0)

No motion d
g

HOT (T = T0 + ΔT)
Low ΔT Bénard cells Turbulent convection
(High ΔT)

Figure 1.3 Fluid is held between two flat, parallel plates. The lower plate is heated. At low values
of T the fluid is quiescent. As T is increased natural convection sets in, first in the form of
regular convection cells and then in the form of turbulent flow.

in Figure 1.3(a). The lower plate is maintained at temperature T = T0 + T and the upper plate
at temperature T 0 . At low values of T the fluid remains stagnant and heat passes from one plate
to the other by molecular conduction. Of course, there is an upward buoyancy force which is
greatest near the lower plate and tends to drive the fluid upward. However, at low values of T
this force is exactly balanced by a vertical pressure gradient. We now slowly increase T. At a
critical value of T the fluid suddenly starts to convect as shown in Figure 1.3(b).
The flow consists of hot rising fluid and cold falling regions. This takes the form of regular
cells, called Bénard cells, which are reminiscent of Taylor vortices.2 The rising fluid near the top
of the layer loses its heat to the upper plate by thermal conduction, whereupon it starts to fall.
As this cold fluid approaches the lower plate it starts to heat up and sooner or later it reverses
direction and starts to rise again. Thus we have a cycle in which potential energy is continuously
released as light fluid rises and dense fluid falls. Under steady conditions the rate of working of
the buoyancy force is exactly balanced by viscous dissipation within the fluid.
The transition from the quiescent state to an array of steady convection cells is, of course,
triggered by an instability associated with the buoyancy force. If we perturb the quiescent state,
allowing hot fluid to rise and cold fluid to fall, then potential energy will be released and, if the
viscous forces are not too excessive, the fluid will accelerate, giving rise to an instability. On the
other hand, if the viscous forces are large then viscous dissipation can destroy all of the potential
energy released by the perturbation. In this case the quiescent configuration is stable. Thus we
expect the quiescent state to be stable if T is small and v large, and unstable if T is large and v
is small.
It is possible to set up a variational problem where one looks for the form of perturbation that
maximizes the rate of release of potential energy and minimizes the viscous dissipation. On the
assumption that the observed convection cells have just such a shape this yields the critical values
of T and v at which instability first sets in. It turns out that the instability criterion is

gβTd3
Ra = ≥ 1.70 × 103 ,
vα

2
In plan view the convection cells can take a variety of forms, depending on the value of (T)/(T)CRIT and on
the shape of the container. Two-dimensional rolls and hexagons are both common.

6 | the ubiquitous nature of turbulence

where Ra (a dimensionless parameter) is known as the Rayleigh number, β is the expansion co-
efficient of the fluid, d is the height of the gap, and α the thermal diffusivity. The wavenumber
(2π/wavelength) of the convection cells is k = 3.12/d.
Of course, we have seen these numbers before. The critical value of the Taylor number for
narrow annuli is
2 d3 R
Ta = = 1.70 × 103
v2
and the wavenumber of the Taylor cells at the onset of instability is k = 3.12/d. At first sight this
coincidence seems remarkable. However, it turns out that there is an analogy between axisym-
metric flow with swirl (in the narrow-gap approximation) and natural convection. The analogue
of the buoyancy force is the centrifugal force and the angular momentum, = ruθ , in a rotating
axisymmetric flow is convected and diffused just like the temperature, T, in Bénard’s experiment.3
In fact, Rayleigh first derived his famous criterion for the instability of a rotating, inviscid fluid
through a consideration of the analogy between the buoyancy and centrifugal forces.
Now suppose Ra is slowly increased above its critical value. Then there is a point at which the
Bénard cells themselves become unstable, and more complex, unsteady structures start to appear.
Eventually, when Ra is large enough, we reach a state of turbulent convection.
It seems that the general picture that emerges from both Taylor’s and Bénard’s experiments is
the following. For high viscosities the basic configuration is stable. As the viscosity is reduced we
soon reach a point where the basic equilibrium is unstable to infinitesimal perturbations and the
system bifurcates (changes) to a more complex state, consisting of a steady laminar flow in the
form of regular cells. As the viscosity is reduced even further the new flow itself becomes unstable
and we arrive at a more complex motion. Subsequent reductions in viscosity give rise to pro-
gressively more complex flows until eventually, for sufficiently small v, the flow is fully turbulent.
At this point the flow field consists of a mean (time-averaged) component plus a random, chaotic
motion. This sequence of steps is called the transition to turbulence.

1.2 Flow over a cylinder

This kind of behaviour, in which a fluid passes through a sequence of flow regimes of increasing
complexity as v is reduced, is also seen in external flows. Consider, for example, flow past a cy-
linder, as shown in Figure 1.4. An inverse measure of v is given by the (dimensionless) Reynolds
number, Re = ud/v, where u is the upstream speed of the fluid and d the diameter of the cylin-
der. At high values of v we have a steady, symmetric flow pattern. As Re approaches unity the
upstream–downstream symmetry is lost and in the range Re ~ 5–40 we find steady vortices at-
tached to the rear of the cylinder. When Re reaches a value of around 40 an instability is observed
in the form of an oscillation of the wake, and by the time we reach Re ~ 100 the vortices start to
peel off from the rear of the cylinder in a regular, periodic manner. The flow is still laminar but
we have now lost the top–bottom symmetry. This is, of course, the famous Karman vortex street.
The laminar Karman street persists up to Re ~ 200 at which point three-dimensional instabilities
develop. By Re ~ 400 low levels of turbulence start to appear within the vortices. Nevertheless,

3
Actually it turns out that there is a 0.7% difference in the critical values of Ta and Ra, so the analogy is very
nearly, but not exactly, perfect.

flow over a cylinder | 7

 (a)

Re « 1

(b)
Re ~ 10

(c)

Re ~ 100

(d)

Re ~ 104

(e)
Figure 1.4 Flow behind a cylinder:
(a) Re < 1; (b) 5 < Re < 40; Re ~ 106
(c) 100 < Re < 200; (d) Re ~ 104 ;
and (e) Re ~ 106 .

the periodicity of the vortex shedding remains quite robust. Eventually, for high values of Re,
say 105 , the turbulence spreads out of the vortices and we obtain a fully turbulent wake. Within
this wake, however, one can still detect coherent vortex shedding. Thus we have the same sort of
pattern of behaviour as seen in Taylor’s experiment. As v is reduced the flow becomes more and
more complex until, eventually, turbulence sets in.
This example is important since it illustrates the key role played by the Reynolds number
in determining the state of a flow. In general, the Reynolds number, Re = u /v, represents the
ratio of inertial to viscous forces in a fluid, provided that is chosen appropriately.4 Thus, when
Re is large, the viscous forces, and hence viscous dissipation, are small. Such flows are prone to
instabilities and turbulence, as evidenced by our example of flow over a cylinder.

1.3 Reynolds’ experiment

The transition from laminar to turbulent flow, and the important role played by Re in this tran-
sition, was first pointed out by Reynolds in 1883. Reynolds was concerned with flow along a
straight, smooth pipe which, despite having a particularly simple geometry, turns out to be rather
more subtle and complex than either Taylor’s or Bénard’s experiments.
In his now famous paper, Reynolds clearly distinguished between the two possible flow re-
gimes (laminar and turbulent) and argued that the parameter controlling the transition from one

4
The inertial forces are of order u2 /lu where lu is a length scale typical of the streamline pattern, say the curvature
of the streamlines, while the viscous forces are of order vu/lv2 , where lv is a length typical of cross-stream gradients
in velocity (see Chapter 2).

8 | the ubiquitous nature of turbulence

Another random document with
no related content on Scribd:
reconcile their habits, views, and interests with those of the South
and West. The latter are beginning to rule with a rod of iron.”
Pickering knew that the Federalist majority in Massachusetts was
none too great. The election in May, four months later, showed a
Federalist vote of 30,000 against a Republican minority of 24,000,
while in the Legislature Harrison Gray Otis was chosen Speaker by
129 votes to 103. Pickering knew also that his colleague, Senator
Adams, was watching his movements with increasing ill-will, which
Pickering lost no chance to exasperate. Nothing could be more
certain than that at the first suggestion of disunion Senator Adams
and the moderate Federalists would attack the Essex Junto with the
bitterness of long-suppressed hatred; and if they could not command
fourteen votes in the Legislature and three thousand in the State, a
great change must have occurred since the year before, when they
elected Adams to the Senate for the long term over Pickering’s head.
Pickering concealed his doings from his colleague; but Tracy was not
so cautious. Adams learned the secret from Tracy; and the two
senators from Massachusetts drew farther and farther apart, in spite
of the impeachments, which tended to force them together.
The Essex Junto, which sent Pickering to Washington, and to
which he appealed for support, read his letter with evident
astonishment. George Cabot, Chief-Justice Parsons, Fisher Ames,
and Stephen Higginson, who were the leaders consulted,[104]
agreed that the scheme was impracticable; and Cabot, as gently as
possible, put their common decision into words.
“All the evils you describe,” he said,[105] “and many more, are to
be apprehended; but I greatly fear that a separation would be no
remedy, because the source of them is in the political theories of our
country and in ourselves. A separation at some period not very remote
may probably take place,—the first impression of it is even now
favorably received by many; but I cannot flatter myself with the
expectation of essential good to proceed from it while we retain
maxims and principles which all experience, and I may add reason
too, pronounce to be impracticable and absurd. Even in New England,
where there is among the body of the people more wisdom and virtue
than in any other part of the United States, we are full of errors which
no reasoning could eradicate if there were a Lycurgus in every village.
 We are democratic altogether; and I hold democracy in its natural
operation to be the government of the worst.
“There is no energy in the Federal party, and there could be none
manifested without great hazard of losing the State government.
Some of our best men in high stations are kept in office because they
forbear to exert any influence, and not because they possess right
principles. They are permitted to have power if they will not use it.... I
incline to the opinion that the essential alterations which may in future
be made to amend our form of government will be the consequences
only of great suffering or the immediate effects of violence. If we
should be made to feel a very great calamity from the abuse of power
by the National Administration, we might do almost anything; but it
would be idle to talk to the deaf, to warn the people of distant evils. By
this time you will suppose I am willing to do nothing but submit to fate.
I would not be so understood. I am convinced we cannot do what is
wished; but we can do much, if we work with Nature (or the course of
things), and not against her. A separation is now impracticable,
because we do not feel the necessity or utility of it. The same
separation then will be unavoidable when our loyalty to the Union is
generally perceived to be the instrument of debasement and
impoverishment. If it is prematurely attempted, those few only will
promote it who discern what is hidden from the multitude.”

Cabot’s letter, more clearly than any writing of Alexander

Hamilton himself, expressed the philosophy and marked the tactics
of their school. Neither Cabot nor Hamilton was a lively writer, and
the dust which has gathered deep on their doctrines dulls whatever
brilliancy they once possessed; but this letter showed why Cabot
was considered the wisest head in his party, to whose rebuke even
Hamilton was forced to bow. For patient and willing students who
have groped in search of the idea which, used by Hamilton and
Jefferson, caused bitterer feeling and roused deeper terrors than civil
war itself, Cabot’s long and perhaps pedantic letter on the policy of
disunion was full of meaning. “We shall go the way of all
governments wholly popular,—from bad to worse,—until the evils, no
longer tolerable, shall generate their own remedies.” Democracy
must end in a crisis, experience and reason pronounced it
impracticable and absurd, Nature would in due time vindicate her
own laws; and when the inevitable chaos should come, then
conservative statesmanship could set society on a sound footing by
limiting the suffrage to those citizens who might hold in their own
right two thousand dollars value in land. Meanwhile disunion would
be useless, and the attempt to bring it about would break up the
Federalist party. “A war with Great Britain manifestly provoked by our
rulers” was the only chance which Cabot foresaw of bringing the
people of New England to a dissolution of the Union.
Pickering was not so intelligent as Cabot, Parsons, and Ames;
his temper was harsher than theirs; he was impatient of control, and
never forgot or wholly forgave those who forced him to follow
another course than the one he chose. Cabot’s letter showed a
sense of these traits; for though it was in the nature of a command or
entreaty to cease discussing disunion, if the Federalist party in
Massachusetts were to be saved, it was couched in gentle language,
and without affecting a tone of advice suggested ideas which ought
to guide Federalists in Congress. Pickering was to wait for the crisis.
Inaction was easy; and even though the crisis should be delayed five
or ten years,—a case hardly to be supposed,—no step could be
taken without a blunder before the public should be ready for it. With
this simple and sound principle to guide them, conservatives could
not go wrong. Cabot there left the matter.
Such gentleness toward a man of Pickering’s temper was a
mistake, which helped to cost the life of one whom conservatives
regarded as their future leader in the crisis. Pickering was restive
under the sense that his friends preferred other counsellors; whereas
his experience and high offices, to say nothing of his ability, entitled
him, as he thought, to greater weight in the party than Hamilton,
Cabot, or Rufus King. Backed by Tracy, Griswold, and other men of
standing, Pickering felt able to cope with opposition. His rough sense
and democratic instincts warned him that the fine-drawn political
theories of George Cabot and Theophilus Parsons might end in
impotence. He could see no reason why Massachusetts, once
corrupted, might not wallow in democratic iniquities with as much
pleasure as New York or Pennsylvania; and all that was worth saving
might be lost before her democracy would consent to eat the husks
of repentance and ask forgiveness from the wise and good. Cabot
wanted to wait a few months or years until democracy should work
out its own fate; and whenever the public should yearn for repose,
America would find her Pitt and Bonaparte combined in the political
grasp and military genius of Alexander Hamilton. Pickering, as a
practical politician, felt that if democracy were suffered to pull down
the hierarchy of New England, neither disunion nor foreign war, nor
“a very great calamity” of any kind, could with certainty restore what
had once been destroyed.
Cabot’s argument shook none of Pickering’s convictions; but the
practical difficulty on which the home Junto relied was fatal unless
some way of removing it could be invented. During the month of
February, 1804, when the impeachment panic was at its height in
Congress, Pickering, Tracy, and Plumer received letter after letter
from New England, all telling the same story. The eminent Judge
Tapping Reeve, of Connecticut, wrote to Tracy:[106] “I have seen
many of our friends; and all that I have seen and most that I have
heard from believe that we must separate, and that this is the most
favorable moment.” He had heard only one objection,—that the
country was not prepared; but this objection, which meant that the
disunionists were a minority, was echoed from all New England. The
conspirators dared not openly discuss the project. “There are few
among my acquaintance,” wrote Pickering’s nephew, Theodore
Lyman,[107] “with whom I could on that subject freely converse; there
may be more ready than I am aware of.” Plumer found a great
majority of the New Hampshire Federalists decidedly opposed.
Roger Griswold, toward the end of the session, summed up the
result in his letter to Oliver Wolcott:—
“We have endeavored during this session to rouse our friends in
New England to make some bold exertions in that quarter. They
generally tell us that they are sensible of the danger, that the Northern
States must unite; but they think the time has not yet arrived.
Prudence is undoubtedly necessary; but when it degenerates into
procrastination it becomes fatal. Whilst we are waiting for the time to
arrive in New England, it is certain the democracy is making daily
inroads upon us, and our means of resistance are lessening every
day. Yet it appears impossible to induce our friends to make any
decisive exertions. Under these circumstances I have been induced to
look to New York.”
 The representatives of the wise and good looked at politics with
eyes which saw no farther than those of the most profligate
democrat into the morality of the game. Pickering enjoyed hearing
himself called “honest Tim Pickering,” as though he were willing to
imply a tinge of dishonesty in others, even in the Puritan society of
Wenham and Salem. Griswold was to the end of his life a highly
respected citizen of Connecticut, and died while governor of the
State. That both these worthy men should conspire to break up the
Union implied to their minds no dishonesty, because they both held
that the Republican majority had by its illegal measures already
destroyed the Constitution which they had sworn to support; but
although such casuistry might excuse in their own consciences the
act of conspiracy, neither this reasoning nor any other consistent
with self-respect warranted their next step. Griswold’s remark that
the procrastination of New England had led him to look to New York
was not quite candid; his plan had from the first depended on New
York. Pickering had written to Cabot at the outset, “She must be
made the centre of the confederacy.” New York seemed, more than
New England, unfit to be made the centre of a Northern confederacy,
because there the Federalist party was a relatively small minority. If
Massachusetts and Connecticut showed fatal apathy, in New York
actual repulsion existed; the extreme Federalists had no following.
To bring New York to the Federalism of Pickering and Griswold, the
Federalist party needed to recover power under a leader willing to do
its work. The idea implied a bargain and an intrigue on terms such as
in the Middle Ages the Devil was believed to impose upon the
ambitious and reckless. Pickering and Griswold could win their game
only by bartering their souls; they must invoke the Mephistopheles of
politics, Aaron Burr.
To this they had made up their minds from the beginning. Burr’s
four years of office were drawing to a close. The Virginians had paid
him the price he asked for replacing them in power; and had it been
Shylock’s pound of flesh, they could not have looked with greater
care to see that Burr should get neither more nor less, even in the
estimation of a hair, than the exact price they had covenanted to pay.
In another year the debt would be discharged, and the Virginians
would be free. Burr had not a chance of regaining a commanding
place among Republicans, for he was bankrupt in private and public
character. In New York the Clintons never ceased their attacks, with
the evident wish to drive him from the party. Cheetham, after
publishing in 1802 two heavy pamphlets, a “Narrative” and a “View,”
attempted in 1803 to crush him under the weight of a still heavier
volume, containing “Nine Letters on the Subject of Aaron Burr’s
Political Defection.” Nov. 16, 1803, the “Albany Register” at length
followed Cheetham’s lead; and nearly all the other democratic
newspapers followed the “Register,” abandoning Burr as a man who
no longer deserved confidence.
Till near the close of 1803 the Vice-President held his peace. The
first sign that he meant energetic retaliation was given by an
anonymous pamphlet,[108] which won the rare double triumph of
political and literary success, in which ability and ill temper seemed
to have equal shares. The unexpected appearance of “Aristides”
startled New York. This attack recalled the scandal which Alexander
Hamilton had created four years before by his pamphlet against his
own President. “Aristides” wrote with even more bitterness than
Hamilton, and the ferocity of his assault on the personal and political
characters of the Republican leaders made the invectives of
Hamilton and Cheetham somewhat tame; but the scandal in each
case was due not so much to personalities of abuse as to breaches
of confidence. “Aristides” furnished to the enemies of the Clintons
and Livingstons an arsenal of poisoned weapons; but what was
more to the purpose, his defence of Burr was strong. That it came
directly from the Vice-President was clear; but the pamphlet showed
more literary ability than Burr claimed, and the world was at a loss to
discover who could be held responsible for its severities. Cheetham
tried in vain to pierce the incognito. Not till long afterward was
“Aristides” acknowledged by Burr’s most intimate friend, William
Peter Van Ness.
An attempt to separate what was just from what was undeserved
in Van Ness’s reproaches of the Clintons and Livingstons would be
useless. The Clintons and Livingstons, however unprincipled they
might be, could say that they were more respectable than Burr; but
though this were true so far as social standing was concerned, they
could not easily show that as a politician the Vice-President was
worse than his neighbors. The New England Federalists knew well
that Burr was not to be trusted, but they did not think much worse of
him than they thought of De Witt Clinton, or John Armstrong, or
Edward Livingston, at this moment removed from office by Jefferson
for failing to account for thirty thousand dollars due to the United
States Treasury. As a politician Burr had played fast and loose with
all parties; but so had most of his enemies. Seeing that he was about
to try another cast of the dice, all the political gamblers gathered
round to help or hurt his further fortunes; and Van Ness might fairly
have said that in the matter of principle or political morality, none of
them could show clean hands.
Although Vice-President until March, 1805, Burr announced that
he meant to offer himself as a candidate for the post of governor of
New York in April, 1804. At the same time Governor Clinton privately
gave warning of his own retirement. De Witt Clinton was annoyed at
his uncle’s conduct, and tried to prevent the withdrawal by again
calling Jefferson to his aid and alarming him with fear of Burr.
“A certain gentleman was to leave this place yesterday morning,”
wrote De Witt to the President.[109] “He has been very active in
procuring information as to his probable success for governor at the
next election. This, I believe, is his intention at present, although it is
certain that if the present Governor will consent to be a candidate, he
will prevail by an immense majority.... Perhaps a letter from you may
be of singular service.”
Jefferson declined to interfere, putting his refusal on the ground
of Burr’s candidacy.
“I should think it indeed a serious misfortune,” was his reply,[110]
“should a change in the administration of your government be
hazarded before its present principles be well established through all
its parts; yet on reflection you will be sensible that the delicacy of my
situation, considering who may be competitors, forbids my
intermeddling even so far as to write the letter you suggest. I can
therefore only brood in silence over my secret wishes.”
No real confidence ever existed between Jefferson and the
Clintons. A few days after these letters were written, “Aristides”
betrayed the secret that Governor Clinton, in the spring of 1800,
declared Jefferson to be “an accommodating trimmer, who would
change with times and bend to circumstances for the purposes of
personal promotion.” This revelation by “Aristides,” supported by the
names of persons who heard the remark, forced Governor Clinton
into an awkward denial of the charge, and led to an exchange of
letters[111] and to professions of confidence between him and
Jefferson; but time showed that neither the Governor nor his nephew
loved the Virginians more than they were loved by Burr.
The threads of intrigue drew together, as they were apt to do
before a general election. The last week in January came. Three
days before Senator Pickering wrote his conspiracy letter to George
Cabot, a letter which implied co-operation with Burr in making him
governor of New York, Burr asked for a private interview with
Jefferson, and formally offered him the choice between friendship or
enmity. The President thought the conversation so curious that he
made a note of it.
“He began,” said Jefferson,[112] “by recapitulating summarily that
he had come to New York a stranger, some years ago; that he found
the country in possession of two rich families,—the Livingstons and
Clintons; ... that since, those great families had become hostile to him
and had excited the calumnies which I had seen published; that in this
Hamilton had joined, and had even written some of the pieces against
him.... He observed, he believed it would be for the interest of the
Republican cause for him to retire,—that a disadvantageous schism
would otherwise take place; but that were he to retire, it would be said
he shrank from the public sentence, which he would never do; that his
enemies were using my name to destroy him, and something was
necessary from me to prevent and deprive them of that weapon,—
some mark of favor from me which would declare to the world that he
retired with my confidence.”
Jefferson, with many words but with his usual courtesy, intimated
that he could not appoint the Vice-President to an Executive office;
and Burr then united his intrigues with those of Pickering and
Griswold. Thenceforth his chance of retaining power depended on
the New York election; and his success in this election depended on
the Federalists. Before George Cabot had yet written his answer to
Pickering’s questions, Pickering could no longer resist the temptation
to act.
The effect of what passed at Washington was instantly felt at
Albany. Toward the middle of February, about three weeks after
Jefferson had civilly rejected the Vice-President’s advances, Burr’s
friends in the New York legislature announced that they should hold
a caucus February 18, and nominate him as candidate for governor.
The Federalists at once called a preliminary caucus to decide
whether they should support Burr. Alexander Hamilton, who
happened to be engaged in law business at Albany, Feb. 16, 1804,
attended the Federal caucus, and used his influence in favor of the
regular Clinton candidate against Burr’s pretensions. The drift of his
argument was given in an abstract of reasons which he drew up for
the occasion.[113] Unfortunately the strongest of these reasons was
evidently personal; the leadership of Hamilton would not tolerate
rivalry from Burr. Hamilton pointed out that Burr’s elevation by the
Federalists of New York would present him as their leader to the
Federalists of New England, and would assist him to disorganize
New England if so disposed; that there “the ill-opinion of Jefferson,
and jealousy of the ambition of Virginia, is no inconsiderable prop of
good opinions; but these causes are leading to an opinion that a
dismemberment of the Union is expedient. It would probably suit Mr.
Burr’s views to promote this result,—to be the chief of the Northern
portion; and placed at the head of the State of New York, no man
would be more likely to succeed.”
If the Union was to be severed, Hamilton was the intended chief
of the Northern portion; but he wanted no severance that should
leave the germs of the democratic disease. His philosophy was that
of George Cabot, William Pitt, and Talleyrand; he waited for the
whole country to come to its senses and restore sound principles,
that democracy might everywhere die out or be stifled. Burr’s
methods were democratic, and would perpetuate in a Northern
confederacy the vices of the Union; they would break up the
conservative strength without weakening democracy. Within a few
days the danger which Hamilton foresaw came to pass. Burr’s little
band of friends in the Legislature, Feb. 18, 1804, set him in
nomination; and a large majority of Federalists, in defiance of
Hamilton’s entreaties, meant to vote for him.
As the situation became clearer, Hamilton’s personal feeling
became public. While at Albany, February 16, he dined with Judge
Taylor, and at table talked of the political prospect. One of the
company, Dr. Charles D. Cooper, an active partisan, wrote an
account of the conversation to a certain Mr. Brown near Albany:
“General Hamilton and Judge Kent have declared, in substance, that
they looked upon Mr. Burr to be a dangerous man, and one who
ought not to be trusted with the reins of government.” The letter was
printed, and went the rounds of the press. As it roused some
question and dispute, Cooper wrote again: “I could detail to you a
still more despicable opinion which General Hamilton has expressed
of Mr. Burr.” This letter also was printed; the “Albany Register” of
April 24 contained the correspondence.
The news of Burr’s nomination reached Washington at the
moment when Pickering and Tracy received answers to their
disunion scheme; and it served to keep them steady to their plan.
The Federalists, who professed to consider Hamilton their leader,
seldom followed his advice; but on this occasion they set him
somewhat unkindly aside. Too much in awe of Hamilton to say
directly to his face that he must be content with the place of Burr’s
lieutenant, they wrote letters to that effect which were intended for
his eye.
Of all Federalist leaders, moderate and extreme, Rufus King, who
had recently returned from London, stood highest in the confidence
of his party. He was to be the Federalist candidate for Vice-
President; he had mixed in none of the feuds which made Hamilton
obnoxious to many of his former friends; and while King’s manners
were more conciliatory, his opinions were more moderate, than those
of other party leaders. To him Pickering wrote, March 4, 1804, in a
tone of entreaty:—
“I am disgusted with the men who now rule, and with their
measures. At some manifestations of their malignancy I am shocked.
The cowardly wretch at their head, while like a Parisian revolutionary
 monster prating about humanity, would feel an infernal pleasure in the
utter destruction of his opponents.”
After avowing his hopes of disunion, Pickering next touched the
New York election:[114]—
“The Federalists here in general anxiously desire the election of
Mr. Burr to the chair of New York, for they despair of a present
ascendency of the Federalist party. Mr. Burr alone, we think, can
break your democratic phalanx, and we anticipate much good from his
success. Were New York detached, as under his administration it
would be, from the Virginia influence, the whole Union would be
benefited. Jefferson would then be forced to observe some caution
and forbearance in his measures. And if a separation should be
deemed proper, the five New England States, New York, and New
Jersey would naturally be united.”
Rufus King was as cautious as Pickering was indiscreet. He
acknowledged this letter in vague terms of compliment,[115] saying
that Pickering’s views “ought to fix the attention of the real friends of
liberty in this quarter of the Union, and the more so as things seem
to be fast advancing to a crisis.” Even King’s cool head was
possessed with the thought which tormented Hamilton, Cabot,
Ames, Pickering, Griswold, and Tracy,—the crisis which was always
coming, and which, in the midst of peace, plenty, and contentment
such as a tortured world had seldom known, overhung these wise
and virtuous men like the gloom of death.
A week later Roger Griswold followed Pickering’s example by
writing to another of Hamilton’s friends, Oliver Wolcott, who
apparently sent the letter to Hamilton.[116] A Congressional caucus,
February 25, nominated George Clinton as the Republican candidate
for Vice-President by sixty-five votes against forty-one,—Burr’s
friends absenting themselves. This nomination showed some
division between the Northern and Southern democrats; but
Griswold rightly argued that nothing could be done in Congress,—
the formation of a Northern interest must begin at home, and must
find its centre of union in Burr. The arguments for this course were
set forth with entire candor.
 “I have wished to ascertain,” wrote Griswold, “the views of Colonel
Burr in relation to the general government; but having had no intimacy
with him myself, and finding no one on the spot calculated, or indeed
authorized, to require an explanation, I have obtained but little
information. He speaks in the most bitter terms of the Virginia faction,
and of the necessity of a union at the northward to resist it; but what
the ultimate objects are which he would propose, I do not know. It is
apparent that his election is supported in New York on the principle of
resisting Virginia and uniting the North; and it may be presumed that
the support given to him by Federal men would tend to reconcile the
feelings of those democrats who are becoming dissatisfied with their
Southern masters. But it is worthy of great consideration whether the
advantage gained in this manner will not be more than
counterbalanced by fixing on the Northern States a man in whom the
most eminent of our friends will not repose confidence. If Colonel Burr
is elevated in New York to the office of governor by the votes of
Federalism, will he not be considered, and must he not in fact
become, the head of the Northern interest? His ambition will not suffer
him to be second, and his office will give him a claim to the first rank.”
Having proposed this question, Griswold argued it as one in
which the interests of New York must yield to the larger interests
behind, and decided that “unpleasant as the thing may be,” Burr’s
election and consequent leadership of the Federalist party was “the
only hope which at this time presents itself of rallying in defence of
the Northern States.... What else can we do? If we remain inactive,
our ruin is certain. Our friends will make no attempts alone. By
supporting Mr. Burr we gain some support, although it is of a doubtful
nature, and of which, God knows, we have cause enough to be
jealous. In short, I see nothing else left for us.”
Had this been all, though it was a rude blow to Hamilton, it might
have passed as a difference of opinion on a point of party policy; but
Griswold’s object in writing these excuses was to explain that he had
already done more, and had even entered into personal relations
with Colonel Burr in view of a bargain. What this bargain was to be,
Griswold explained:—
“I have engaged to call on the Vice-President as I pass through
New York. The manner in which he gave me the invitation appeared to
indicate a wish to enter upon some explanation. He said he wished
very much to see me, and to converse, but his situation in this place
 did not admit of it, and he begged me to call on him at New York. This
took place yesterday in the library. Indeed, I do not see how he can
avoid a full explanation with Federal men. His prospects must depend
on the union of the Federalists with his friends, and it is certain that his
views must extend much beyond the office of governor of New York.
He has the spirit of ambition and revenge to gratify, and can do but
little with his ‘little band’ alone.”
Even George Cabot deserted Hamilton, and wrote from Boston to
Rufus King a long letter, in the tone of indolent speculation which
irritated restless fighters like Pickering and Griswold:[117]—
“An experiment has been suggested by some of our friends, to
which I object that it is impracticable, and if practicable would be
ineffectual. The thing proposed is obvious and natural; but it would
now be thought too bold, and would be fatal to its advocates as public
men; yet the time may soon come when it will be demanded by the
people of the North and East, and then it will unavoidably take place.”
He explained his favorite thesis,—the last resource of failing
protestants,—that things must be worse before they were better; but
closed by wishing success to Burr. “I should rejoice to see Burr win
the race in your State, but I cannot approve of aid being given him by
any of the leading Federalists.”
Ten days later, March 27, Congress adjourned; and
thenceforward the intrigue centred about Burr and Hamilton in New
York. No sooner did Griswold reach that city, a week afterward, on
his way from Washington to Connecticut, than he kept his
engagement with Burr, and the conversation strengthened him in his
policy.[118] Burr was cautious, but said that in his present canvass
“he must go on democratically to obtain the government; that if he
succeeded, he should administer it in a manner that would be
satisfactory to the Federalists. In respect to the affairs of the nation,
Burr said that the Northern States must be governed by Virginia, or
govern Virginia, and that there was no middle course; that the
democratic members of Congress from the East were in this
sentiment,—some of those from New York, some of the leaders in
Jersey, and likewise in Pennsylvania.” Further than this he would not
go; and Griswold contented himself with such vague allurements.
 On the other hand, Rufus King’s library was the scene of grave
dissensions. There Pickering went, April 8, to urge his scheme of
disunion, and retired on the appearance of his colleague, Senator
Adams, who for the first and last time in his life found himself fighting
the battle of Alexander Hamilton, whom he disliked as decidedly as
Pickering professed to love him. As the older senator left the house
at his colleague’s entrance, King said to Adams:[119] “Colonel
Pickering has been talking to me about a project they have for a
separation of the States and a Northern Confederacy; and he has
also been this day talking of it with General Hamilton. Have you
heard anything of it at Washington?” Adams replied that he had
heard much, but not from Colonel Pickering. “I disapprove entirely of
the project,” said King; “and so, I am happy to tell you, does General
Hamilton.”
The struggle for control between Hamilton and the conspirators
lasted to the eve of the election,—secret, stifled, mysterious; the
intrigue of men afraid to avow their aims, and seeming rather driven
by their own passions than guided by the lofty and unselfish motives
which ought to inspire those whom George Cabot emphatically
called the best! The result was a drawn battle. Hamilton prevented
leading Federalists from open committal of the party, but he could
not prevent the party itself from voting for Burr. The election took
place April 25, 1804; and although Burr succeeded in carrying to the
Federalists a few hundred voters in the city of New York, where his
strength lay, giving him there a majority of about one hundred in a
total vote of less than three thousand, he polled but about twenty-
eight thousand votes in the State against thirty-five thousand for the
Clinton candidate. The Federalists gained nothing by supporting him;
but only a small portion of the party refused him their aid.
The obstinacy of Pickering and Griswold in pressing Burr on the
party forced Hamilton to strain his strength in order to prevent what
he considered his own humiliation. That all Hamilton’s doings were
known to Burr could hardly be doubted. When the election closed, a
new era in Burr’s life began. He was not a vindictive man, but this
was the second time Hamilton had stood in his way and vilified his
character. Burr could have no reason to suppose that Hamilton was
deeply loved; for he knew that four fifths of the Federal party had
adopted his own leadership when pitted against Hamilton’s in the
late election, and he knew too that Pickering, Griswold, and other
leading Federalists had separated from Hamilton in the hope of
making Burr himself the chief of a Northern confederacy. Burr never
cared for the past,—the present and future were his only thought; but
his future in politics depended on his breaking somewhere through
the line of his personal enemies; and Hamilton stood first in his path,
for Hamilton would certainly renew at every critical moment the
tactics which had twice cost Burr his prize.
Pickering and Griswold saw their hopes shattered by the result of
the New York election. They gained at the utmost only an agreement
to hold a private meeting of leading Federalists at Boston in the
following autumn;[120] and as Hamilton was to be present, he
probably intended to take part only in order to stop once for all the
intrigues of these two men. Such an assemblage, under the
combined authority of Cabot, King, and Hamilton, could not have
failed to restore discipline.
Nearly two months passed after the New York election, when, on
the morning of June 18, William P. Van Ness, not yet known as
“Aristides,” appeared in Hamilton’s office. He brought a note from
Vice-President Burr, which enclosed newspaper-cuttings containing
Dr. Cooper’s report of Hamilton’s “despicable” opinion of Burr’s
character. The paragraph, Burr said, had but very recently come to
his knowledge. “You must perceive, sir, the necessity of a prompt
and unqualified acknowledgment or denial of the use of any
expression which would warrant the assertions of Dr. Cooper.”
General Hamilton took two days to consider the subject; and then
replied in what Burr thought an evasive manner, but closed with two
lines of defiance: “I trust on more reflection you will see the matter in
the same light with me; if not, I can only regret the circumstance, and
must abide the consequences.”[121]
These concluding words were the usual form in which men
expressed themselves when they intended to accept a challenge to
a duel. At first sight, no sufficient reason for accepting a challenge
was shown by Hamilton’s letter, which disavowed Dr. Cooper’s report
so far as Burr was warranted in claiming disavowal. Hamilton might
without impropriety have declined to give further satisfaction. In truth,
not the personal but the political quarrel drew him into the field; he
knew that Burr meant to challenge, not the man, but the future
political chief, and that an enemy so bent on rule must be met in the
same spirit. Hamilton fought to maintain his own right to leadership,
so rudely disputed by Burr, Pickering, and Griswold. He devoted
some of his moments before the duel to the task of explaining, in a
formal document, that he fought only to save his political influence.
[122] “The ability to be in future useful, whether in resisting mischief
or effecting good, in those crises of our public affairs which seem
likely to happen, would probably be inseparable from a conformity
with public prejudice in this particular.”
Always the crisis! Yet this crisis which brought Hamilton in July to
the duelling-ground at Weehawken was not the same as that which
Pickering and Griswold had so lately tried to create. Pickering’s
disunion scheme came to a natural end on Burr’s defeat in April. The
legislatures of the three Federalist States had met and done nothing;
all chance of immediate action was lost, and all parties, including
even Pickering and Griswold, had fallen back on their faith in the
“crisis”; but the difference of opinion between Hamilton and the New
Englanders was still well defined. Hamilton thought that disunion,
from a conservative standpoint, was a mistake; nearly all the New
Englanders, on the contrary, looked to ultimate disunion as a
conservative necessity. The last letter which Hamilton wrote, a few
hours before he left his house for the duelling-ground, was a short
and earnest warning against disunion, addressed to Theodore
Sedgwick, one of the sternest Massachusetts Federalists of
Pickering’s class.[123]
“Dismemberment of our empire,” said Hamilton, “will be a clear
sacrifice of great positive advantages, without any counterbalancing
good; administering no relief to our real disease, which is democracy,
—the poison of which, by a subdivision, will only be the more
concentred in each part, and consequently the more virulent.”
 The New Englanders thought this argument unsound, as it
certainly was; for a dissolution of the American Union would have
struck a blow more nearly fatal to democracy throughout the world
than any other “crisis” that man could have compassed. Yet the
argument showed that had Hamilton survived, he would probably
have separated from his New England allies, and at last, like his
friends Rufus King and Oliver Wolcott, would have accepted the
American world as it was.
The tragedy that actually happened was a fitter ending to this
dark chapter than any tamer close could have been. Early on the
morning of July 11, in the brilliant sunlight of a hot summer, the two
men were rowed to the duelling-ground across the river, under the
rocky heights of Weehawken, and were placed by their seconds face
to face. Had Hamilton acted with the energy of conviction, he would
have met Burr in his own spirit; but throughout this affair Hamilton
showed want of will. He allowed himself to be drawn into a duel, but
instead of killing Burr he invited Burr to kill him. In the paper
Hamilton left for his justification, he declared the intention to throw
away his first fire. He did so. Burr’s bullet passed through Hamilton’s
body. The next day he was dead.
As the news spread, it carried a wave of emotion over New
England, and roused everywhere sensations strangely mixed. In
New York the Clinton interest, guided by Cheetham, seized the
moment to destroy Burr’s influence forever. Cheetham affected to
think the duel a murder, procured Burr’s indictment, and drove him
from the State. Charges were invented to support this theory, and
were even accepted as history. In the South and West, on the other
hand, the duel was considered as a simple “affair of honor,” in which
Burr appeared to better advantage than his opponent. In New
England a wail of despair arose. Even the clergy, though shocked
that Hamilton should have offered the evil example of duelling, felt
that they had lost their champion and sword of defence. “In those
crises of our public affairs which seemed likely to happen,”
Hamilton’s genius in council and in the field had been their main
reliance; he was to be their Washington, with more than
Washington’s genius,—their Bonaparte, with Washington’s virtues.
The whole body of Federalists, who had paid little regard to
Hamilton’s wishes in life, went into mourning for his death, and held
funeral services such as had been granted to no man of New
England birth. Orators, ministers, and newspapers exhausted
themselves in execration of Burr. During the whole summer and
autumn, undisturbed by a breath of discord or danger, except such
as their own fears created, they bewailed their loss as the most fatal
blow yet given to the hopes of society.
The death of Hamilton cleared for a time the murky atmosphere
of New York and New England politics. Pickering and Griswold,
Tracy and Plumer, and their associates retired into the background.
Burr disappeared from New York, and left a field for De Witt Clinton
to sacrifice in his turn the public good to private ambition. The bloody
feuds of Burr’s time never again recurred. The death of Hamilton and
the Vice-President’s flight, with their accessories of summer-morning
sunlight on rocky and wooded heights, tranquil river, and distant city,
and behind all, their dark background of moral gloom, double
treason, and political despair, still stand as the most dramatic
moment in the early politics of the Union.
 CHAPTER IX.
President Jefferson was told from day to day of the
communications that passed between Burr and the Connecticut
Federalists. Of all members of the Government, the most active
politician was Gideon Granger, the Postmaster-General, whose
“intimacy with some of those in the secret,” as Jefferson afterward
testified, gave him “opportunities of searching into their
proceedings.”[124] Every day during this period Granger made a
confidential report to the President; and at the President’s request
Granger warned De Witt Clinton of Burr’s intrigues with the
Federalists. What passed in Rufus King’s library and in Burr’s private
room seemed known at once by Granger, and was reported within a
few days to Jefferson, who received the news with his innate
optimism, warranted by experience.[125]
“It will be found in this, as in all other similar cases, that crooked
schemes will end by overwhelming their authors and coadjutors in
disgrace, and that he alone who walks strict and upright, and who in
matters of opinion will be contented that others should be as free as
himself, and acquiesce when his opinion is fairly overruled, will attain
his object in the end.”
If Jefferson and his Virginia friends in 1798, when their own
opinions were overruled, had expressed the idea of acquiescence as
strongly, the nation might perhaps have been saved the necessity of
proving later the truth of his words; but Jefferson could afford to treat
with contempt the coalition between Burr and Pickering, because, as
he wisely said, it had no cohesive force to hold it together, no
common principle on which to rest. When Burr’s defeat in April and
Hamilton’s death in July dissolved the unnatural connection,
Jefferson let the secret die; he wanted no scandal. He stood a little in
awe of the extreme Federalists, whom he called incurables, and was
unwilling to exasperate them without an object.
The Administration had every reason to rejoice that Burr’s
factious influence in the State of New York was at an end; for other
causes of anxiety gave the President more personal annoyance. The