Professional Documents
Culture Documents
Ordinary Differential Equations by Vladimir I. Arnold: SIAM Review January 2007
Ordinary Differential Equations by Vladimir I. Arnold: SIAM Review January 2007
net/publication/259815160
CITATION READS
1 9,510
1 author:
Carmen Chicone
University of Missouri
173 PUBLICATIONS 3,258 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Carmen Chicone on 10 December 2018.
Book Reviews
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
Being just a humble mathematician, I have not the faintest idea whether mathematics
is the music of the heavenly spheres. Yet, it is fairly safe to bet that, had heavenly
spheres been studied by standard tools of science, had they been measured, weighed,
had their chemical composition been analyzed, had inquiry been made of their ulti-
mate structure, of their mechanics and dynamics, of their science, it would have been
formulated in the language of mathematics.
We all take for granted this universal effectiveness of mathematics, yet is it im-
portant to pause once in a while and ask ourselves what exactly renders mathematics
into such a universal tool. The first, obvious, answer is its rigor. Mathematics places
you into a straitjacket of exact thinking, forces you to define precisely and to infer
rigorously from your assumptions. It leaves at its best no room for woolliness, hand
waving, and “proof by authority” that are besetting other areas of scholarship.
Yet, rigor is just part of the answer. Another, perhaps less evident, secret behind
the surprising effectiveness of mathematics is its ability to express universal relation-
ships. At an elementary level this is clear, since the most basic universal relationships
are enumerative: we count objects and compare them. However, at a more advanced
level, we wish to take heed not just of the number of our “objects” but of their in-
ternal relationship and interconnectedness. At a phenomenological level numbers are
unequal to this task and we need to use graphs: collections of vertices and edges.
The most useful graphs are trees, where a single path of edges extends between
every two vertices. By singling out one vertex as a root, we establish partial ordering:
vertex a precedes vertex b if it features earlier along the (unique!) path extending
from the root. (In particular, the root “precedes” every other vertex.) A useful model
is a family tree, but such rooted trees extend the length and width of applications
of mathematics: from linguistics and genetics all the way to quantum field theory
(Feynman diagrams!).
It is perhaps not surprising that computer scientists (and, indeed, computer pro-
grammers) are the main beneficiaries of mathematical rigor: unlike protein-based
life-forms, computers simply will not settle for less. You can sometimes talk your way
out of a hole with your students or colleagues—certainly not recommended but this
is the real world!—but not with computers. Moreover, the importance of effective im-
plementation of ideas is clearly emphasized in an algorithmic setting, and this, again,
is best undertaken within the confines of formal reasoning and a language which is
rigorously descriptive of the inner structure of the underlying problem. Which brings
us to the slim, yet important, “fascicle” on the generation of trees.
Publishers are invited to send books for review to Book Reviews Editor, SIAM, 3600 University
City Science Center, Philadelphia, PA 19104-2688.
327
328 BOOK REVIEWS
composed of trees) that occur in other areas of mathematical and scientific research
and which reveal, often by means of a shuffle product, Hopf algebra, and dendriform
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
A Course in Mathematical Biology: Quan- topics that are important in the formula-
titative Modeling with Mathematical and tion and analysis of mathematical models
Computational Methods. By Gerda de Vries, in biology. Six chapters are devoted to
Thomas Hillen, Mark Lewis, Johannes Müller, difference equations, ordinary differential
and Birgitt Schönfisch. SIAM, Philadelphia, 2006. equations, partial differential equations,
$69.00. xii+309 pp., softcover. ISBN 0-89871- stochastic processes, cellular automata,
612-8. and parameter estimation (Chapters 2–7).
Mathematical and statistical methods are
A variety of mathematical and statistical introduced through some classical and well-
topics are covered in A Course in Mathe- chosen biological models such as models for
matical Biology: Quantitative Modeling with population genetics, host-parasitoid, SIR
Mathematical and Computational Methods, epidemic, spruce budworm, Fisher’s equa-
330 BOOK REVIEWS
tion, forest succession, polymerase chain of difference and differential equations and
reaction, and simple birth and death pro- the Routh–Hurwitz criteria for higher order
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
using these tools and techniques to study are increasingly used. This book is about
realistic biological problems. A course de- such supply chain problems and software.
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
signed with this goal in mind would pro- The comment on the back jacket accu-
vide excellent training for undergraduate rately describes readers who can clearly
students in interdisciplinary research. benefit from this book: “The authors
address readers involved in optimization
projects in which SAP and, particularly,
REFERENCES
SAP APO are implemented in companies.
These are the project designers, project
[1] N. F. Britton, Essential Mathematical Bi-
ology, Springer Undergrad. Math. Ser., leaders, the IT personnel inside the com-
Springer-Verlag, London, 2003. panies, but also operations research practi-
[2] S. P. Ellner and J. Guckenheimer, Dy- tioners, supply chain management consul-
namic Models in Biology, Princeton tants, and decision makers in the area of tool
University Press, Princeton, NJ, 2006. selection for optimization tasks.” However,
[3] J. A. Lubina and S. A. Levin, The spread the authors, expert consultants in supply
of a reinvading species: Range expan- chain issues but also accomplished astro-
sion in the California sea otter, Am. physicists in academia, are in a position
Nat., 131 (1988), pp. 526–543. to provide relevance as well as precise lan-
[4] D. D. Mooney and R. J. Swift, A Course guage, mathematical perspective, and clar-
in Mathematical Modeling, Mathemat- ity. This bridge between the academic and
ical Association of America, Washing-
commercial can benefit both users of pack-
ton, D.C., 1999.
ages for supply chain planning and opera-
[5] E. K. Yeargers, R. W. Shonkwiler, and
tions and academically oriented members of
J. V. Herod, An Introduction to the
Mathematics of Biology with Compu- the mathematical optimization community.
ter Algebra Models, Birkhäuser Boston, The term “real optimization” in the title
Boston, MA, 1996. can refer to two things. First, it can refer
to the process of finding an actual optimum
LINDA J. S. ALLEN of an objective function, as contrasted with
Texas Tech University merely producing a production or schedul-
ing plan that satisfies requirements. Sec-
ond, it can refer to the fact that the case
Real Optimization with SAP APO. By Josef studies involve real companies with imple-
Kallrath and Thomas I. Maindl. Springer-Verlag, mentations tailored to and used in actual
Berlin, 2006. $79.95. xxvi+321 pp., hardcover. supply chain planning and operation.
ISBN 3-540-22561-7. SAP is a well-known corporation that
provides general software and consulting for
Producing and bringing products to market data acquisition and management, as well as
underlie the lifestyle of those in developed software to compute feasible solutions and
economies. It is easy for the lay person to optimal solutions to supply chain problems.
overlook the complexities inherent in these External companies and consulting services
processes. Such complexities include, but also supply interfaces to SAP products.
are not limited to, adjusting activities to en- “SAP APO” stands for “SAP Advanced
sure that demand is met while not exhaust- Planning and Optimizer.” Most of the book
ing raw materials, labor, or transportation focuses on optimization and supply chain is-
possibilities, and, if demand can be sat- sues in the context of SAP products, which
isfied, minimizing a cost or maximizing a is reasonable in view of those products’
profit. Adding to the complexities are mul- widespread use. Readers with a more gen-
tiple or conflicting product-specific goals, eral interest in applied optimization or sup-
such as minimizing the changes over time ply chain problems can view the book’s use
in use of labor on assembly lines. Advanced of SAP APO as one large case study.
software systems that integrate data acqui- The book is organized into four parts:
sition, user interfaces, and mathematical an extensive introduction, detailed case
ideas and algorithms for constraint satis- studies, a summary section, and three ap-
faction and local and global optimization pendices.
332 BOOK REVIEWS
Part I: The Introduction. Chapter 1 gives fixed model, running a global optimization,
an introduction to the business perspective then using the result to dictate the business
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
of supply chain data acquisition and organi- plan, but is an iterative process:
zation, including standardized terminology 1. posing an initial model;
for supply chain and planning systems and
a brief introduction to the components and 2. doing an optimization to compute an
structure of SAP APO. initial plan;
Chapter 2 contains definitions of terms 3. adjusting objective function weights
related to data models and optimization, and adding previously unidentified
and types of optimization problems (mixed constraints to define business priori-
integer linear programs, nonlinear pro- ties and make the solution correspond
grams, etc.). Problem representation in more closely to reality;
the data (the supply chain model as in
4. repeating steps 1, 2, and 3.
SAP APO) and the mathematical model
(in terms of variables, indices, objective, Indeed, an important early part of the pro-
and constraints) are contrasted. The au- cess is prioritizing different business ob-
thors note that SAP allows definition of jectives, which leads to different objective
a model in supply chain terminology, but functions. Heuristics such as “minimize as-
the automatic generation of objective and sembly line idle time” may be more practical
constraints is proprietary. to use than overall goals such as “maxi-
Chapters 3 and 4, containing numer- mize profit,” and various overall goals may
ous examples and screen shots, illustrate conflict. Often more than one type of opti-
setting up and dealing with the model in mization is involved, with an initial approx-
SAP APO. For instance, the system can be imation given, say, with a continuous linear
instructed to treat mixed integer models program solution, and subsequent refine-
as purely continuous models, etc. These ments done with mixed integer technology.
are practical ways of conquering combina- The book provides details.
torial complexity. The result is a feasible The specific case studies, in Chapters 5
(although not necessarily globally optimal) through 8, are from semiconductor man-
solution obtained within a predictable com- ufacturing, consumer products, the auto-
puting time; the point is that this solution motive and chemical industries, and the
is both practical to obtain and valuable in process industry. Chapter 9, written by the
actual supply chain planning and operation. external authors Rémi Lequette of ILOG
A detailed derivation of a mathematical op- corporation and Axel Hecker of Mathesis
timization problem from an example supply GmbH, describes models and implemen-
chain model (although hidden from the user tations tailored to specific situations, for
in SAP APO) is also presented. which the intrinsic capabilities of SAP APO
do not suffice.
Part II: Detailed Case Studies. Chapters 5
through 9 contain case studies describing Parts III and IV: Concluding Considera-
specific consulting (or application develop- tions and Appendices. The 20-page conclu-
ment) experiences, in candid detail. This sion contains a summary of what users can
includes interaction with the business client expect from SAP APO, general thoughts
to determine goals and specifics of the appli- on design and implementation of business
cation, thoughts on where and how human supply chain software, and thoughts on the
interaction with the software system best relationship of business software to ideas
occurs, how well the packaged components and techniques from the mathematical op-
in SAP APO fit the application, and when timization community. It also contains gen-
customized external solutions are more ap- eral thoughts on what users can and cannot
propriate. expect from SAP APO, how such users can
These case studies illustrate the impor- increase their benefits by doing “real opti-
tance of user interfaces and user interaction. mization” (as opposed to data management
That is, deciding on an actual plan does not and finding a feasible plan), and how SAP
only involve a single pass of formulating a might improve its “APO” product.
BOOK REVIEWS 333
The first appendix, “The Hitchhiker’s to his students. In the meantime about
Guide to SAP APO,” gives a brief introduc- 20 years have passed and a second edition
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
tion to the SAP APO components and SAP was necessary to take into account new ad-
APO’s hierarchical planning philosophy. vances in theory and applications, in par-
The second appendix, 20 pages long, con- ticular providing the basic mathematical
tains a succinct overview of “Mathematical background for stochastic modeling in fi-
Foundations of Optimization.” The topics nance.
include linear programming (with sketches The reviewer was pleased to notice that
of the primal simplex algorithm and asso- this second edition retains the positive fea-
ciated techniques, the dual simplex algo- tures of the first edition with regard to
rithm, and interior point methods), mixed style and mathematical presentation. Ide-
integer linear programming, and multicri- ally, readers of this book should be famil-
teria optimization and goal programming. iar with real analysis and the Lebesgue
(Mixed integer linear programming rep- measure. This would considerably facili-
resents the predominant type of underly- tate the understanding of Chapters 1 to
ing mathematical optimization problem en- 3, which are devoted to the foundations
countered in the book.) of probability theory. In Chapter 1, spe-
The book includes a glossary, 104 refer- cial emphasis is placed on the fundamen-
ences dealing with both mathematical and tal law of probability (better known as the
business issues, biographies of the authors, law of the subconscious statistician) and
a seven-page index, and a list of acronyms key properties of characteristic functions,
and abbreviations. whereas Chapters 2 and 3 give sophisti-
cated treatments of basic concepts such as
Concluding Recommendations. I recom- independence and conditional expectations
mend this book for personnel in corpo- in conjunction with the laws of large num-
rations and governmental entities contem- bers, the Glivenko–Cantelli theorem, ran-
plating improving supply chain operations dom walks and queuing models, martin-
with SAP software. The book may also gales, and the theorem of Radon–Nikodým.
be of more general interest to operations Chapter 4 provides an excellent treatment
researchers and mathematical optimization of probability distributions, characteristic
professionals, especially those interested in functions and their mutual relationships, as
mixed integer linear programs, those want- well as applications to the convergence of
ing to keep their academic activities relevant sums of random variables. Chapter 5 is the
to real-world concerns, and those wishing to “heart of the analytical theory of proba-
gain experience in commercial applications. bility.” From a unifying point of view, it
deals with other topics such as infinite di-
R. BAKER KEARFOTT
visibility and Spitzer’s identity, as well as
University of Louisiana at Lafayette
central and other limit theorems. The er-
ror estimation for central limit theorems
is, unfortunately, restricted to the results
Probability Theory with Applications. of Berry (1941) and Esseen (1945). Other
Second Edition. By M. M. Rao and Randall J. improved results due to Bikelis, Nagaev,
Swift. Springer-Verlag, New York, 2006. $89.95. and in particular Michel (1981) and Paditz
xvii+527 pp., hardcover. ISBN 0-387-27730-7. (1989) are not even mentioned. It is a gen-
eral shortcoming of the book that in crucial
This book is a thoroughly revised and up- cases references to the state of the art are
dated version of the first edition written by missing, particularly in Chapters 6 to 8
M. M. Rao. It expresses the in many ways dealing with martingales, more limit theo-
original views of the authors on how the sub- rems, and some basic concepts of stochastic
ject could be presented in graduate courses processes. Nevertheless, these chapters may
for students of mathematical sciences. The very well serve as a first reading and can
first edition was fairly successful and the re- be strongly recommended to students and
viewer used it for preparing his own lectures other readers who need to upgrade their
on probability theory and recommended it knowledge of the mathematical theory of
334 BOOK REVIEWS
probability to be able to follow modern ap- els, these models are the Poisson’s equation
plications. of contact research: if a method can’t solve
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
Compared to the first edition, the lists problems with these models, it probably
of exercises stated after each chapter have can’t solve any contact problem. The book
been considerably enhanced and adapted to covers constitutive equations for both nor-
changes of content. For more sophisticated mal contact and tangential contact, includ-
exercises, the authors give rather detailed ing friction laws for metal forming, rubber
instructions on how to solve them. Gener- and polymers, concrete, and several others.
ally, lecturers will appreciate this book as a It discusses the computational challenges
valuable source for supplementing their own of incorporating these models into contact
collections of student-tailored problems. As algorithms and demonstrates that a great
with the first edition, the reviewer will defi- deal of research remains to be done.
nitely make use of this book for developing The penalty method and Lagrange multi-
and revising his own lectures and encourage pliers are the two methods most commonly
serious students to use it for accompanying used in contact algorithms. Wriggers covers
reading. these methods plus direct constraint elim-
ination, Nitsche’s method, perturbed and
FRANK BEICHELT
augmented Lagrange methods, the barrier
University of Witwatersrand
method, and the cross-constraint method.
Johannesburg
An especially nice feature of the book is
that the discrete equations are derived and
written out explicitly for each method for a
simple model one-dimensional problem, and
Computational Contact Mechanics. Sec- their accuracy is compared. After devel-
ond Edition. By Peter Wriggers. Springer- oping discretizations for node-to-node con-
Verlag, New York, 2006. $99.00. xii+518 pp., tact and for nonmatching meshes assuming
hardcover. ISBN 3-540-32608-1. small deformations, almost 100 pages are
devoted to large deformation contact dis-
Most books on finite element methods fo- cretizations. Recent research on mortar dis-
cus primarily on element formulations and cretizations and smooth contact discretiza-
treat only the simplest boundary condi- tions is presented.
tions. This is unfortunate given the im- The typical treatments of thermome-
portance of contact mechanics in many chanical contact are naive, ignoring the
engineering problems and the mathemat- interaction between the mechanical contact
ical challenge of imposing inequality con- and the heat conduction across the con-
straints accurately. Computational Contact tact interface. Wriggers demonstrates with
Mechanics, by Peter Wriggers, is an excel- a simple one-dimensional problem that a
lent book that fills this important gap. seemingly straightforward approach may
While contact problems are intrinsically not be well posed. The requirements for a
nonlinear because of the contact inequal- well-posed problem and realistic heat con-
ity, the emphasis is on contact methods for duction models for contact interfaces are
problems that have nonlinear, large defor- presented.
mations, nonlinear material response, and Additional topics that are covered thor-
coupled thermomechanical problems. The oughly include global solution algorithms,
book is especially strong in four areas: (1) contact search and detection strategies,
its coverage of the mechanics of contact; beam contact, critical point analysis with
(2) the comprehensive treatment of differ- contact constraints, and adaptive finite el-
ent methods for imposing inequality con- ement methods with contact.
straints; (3) formulations of large deforma- The book is not simply a monograph re-
tions; and (4) thermomechanical contact. viewing the author’s distinguished research
Most computational contact research fo- career, but a true textbook covering the
cuses on frictionless contact or contact with research of the major players in computa-
Coulomb friction. Compared to the numeri- tional contact mechanics over the last 25
cal difficulties associated with realistic mod- years and offers an extensive list of refer-
BOOK REVIEWS 335
ences. While Peter Wriggers is at the Uni- examples and exercises. The new edition is
versity of Hanover, the book is not a trans- highly recommended as a general reference
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
lation and the writing is very clear. If I were for the essential theory of ordinary differ-
given the opportunity to teach an advanced ential equations and as a textbook for an
special topics course on finite element meth- introductory course for serious undergrad-
ods, I wouldn’t hesitate to use this book. uate or graduate students.
If it has a shortcoming as a textbook, it There is a standard criticism of Professor
is that it does not have homework prob- Arnold’s books: They are beautiful treat-
lems. Although it is self-contained in the ments of their subjects that are appreciated
sense that there are chapters that develop fi- by experts, but too many details are omit-
nite deformation continuum mechanics and ted for students to learn the mathematics
nonlinear finite element methods from the required to prove the statements that he
basics, the pace is too fast to be a first so effortlessly justifies. While this point of
introduction to these topics. I recommend view has some merit in pedagogy, its logical
having a background of at least one year at extreme (the mathematical cookbook at the
the graduate level in continuum mechanics introductory level and the advanced presen-
and finite element methods before tackling tation of pure mathematics devoid of un-
this book. For a researcher who has studied derlying ideas and connections to the other
these areas before and may be a little rusty, sciences) is strongly opposed by Arnold.
the text provides an excellent review. Indeed, there is a rebuttal in the preface:
This book is not written in the theorem-
The completion of the formation of
proof style of math textbooks, but in the
analysis as an independent scientific
constructive style found in science and engi-
discipline is connected with the name of
neering textbooks. As with shell elements,
Leibniz (1646–1716). Another of Leib-
the majority of accurate, robust contact
niz’ grand achievements is the broad
methods are not amenable to formal proofs
publicizing of analysis (his first publi-
of convergence, and this book therefore re-
cation is an article in 1684) and the
flects that fact. Those looking for a more
development of its algorithms to com-
formal style of presentation would probably
plete automatization: he thus discov-
prefer the 1988 SIAM book Contact Prob-
ered a method of teaching how to use
lems in Elasticity: A Study of Variational
analysis (and teaching analysis itself)
Inequalities and Finite Element Methods by
to people who do not understand it
Kikuchi and Oden, or Glowinski, Lions, and
at all—a development that has to be
Trémolières’ Numerical Analysis of Varia-
resisted even today.
tional Inequalities (North-Holland, 1981).
Arnold’s book is certainly meant to teach
DAVID J. BENSON the subject of differential equations to those
University of California who truly wish to understand it.
San Diego The first edition’s first chapter, “Ba-
sic Concepts,” has six sections (a total of
47 pages) with titles “Phase Spaces and
Ordinary Differential Equations. By Phase Flows,” “Vector Fields on the Line,”
Vladimir I. Arnold. Springer-Verlag, Berlin, 2006. “Phase Flows on the Line,” “Vector Fields
$64.95. iv+334 pp., softcover. ISBN 3-540- and Phase Flows in the Plane,” “Nonau-
34563-9. tonomous Equations,” and “The Tangent
Space.” The material in the first chapter
Professor Arnold has expanded his clas- of this edition has expanded to 75 pages
sic book to include new material on ex- with section titles “Phase Spaces,” “Vec-
ponential growth, predator-prey, the pen- tor Fields on the Line,” “Linear Equa-
dulum, impulse response, symmetry groups tions,” “Phase Flows,” “The Action of
and group actions, perturbation and bifur- Diffeomorphisms on Vector Fields and Di-
cation, Sturm–Liouville theory, and first- rection Fields,” and “Symmetries.”
order partial differential equations. In ad- The new material includes a discussion
dition, he has included a wealth of new of some popular elementary topics found
336 BOOK REVIEWS
predator-prey. These concepts are seam- One Degree of Freedom,” together with
lessly woven into the original narrative that a new section called “First-Order Linear
treated elementary mechanics: free fall, and Quasi-linear Partial Differential Equa-
small oscillations, the pendulum, the in- tions.” In this section, the essentials of the
verted pendulum, and the spherical pendu- theory of characteristics (in the simplest
lum. Direct product planar systems, that case) are presented with the underlying
is, systems of the form theme best expressed by a quote from its
introductory paragraph:
ẋ = f (x),
ẏ = g(y), The essence of the connection between
a partial differential equation and a
are introduced and used to illustrate the characteristic equation is that the mo-
typical phase portraits of planar rest points. tion of a solid medium can be described
This leads to a discussion of separable using both the ordinary differential
equations, and culminates in a proof of equations of motion of its particles and
the periodicity of the physical orbits of the the partial differential equations for a
Lotka–Volterra system. The reader is soon field.
introduced to limit cycles, structural sta-
bility, and continuations (of rest points and The last chapters, “Linear Systems,”
periodic orbits) via the implicit function “Proofs of the Main Theorems,” and “Dif-
theorem. ferential Equations on Manifolds,” remain
A simple and elegant treatment of the mostly unrevised with the notable addition
(Dirac) δ-function is given along with an of a section on Sturm’s theory: the inter-
introduction to Green’s function for inho- lacing of zeros, the comparison theorem,
mogeneous linear equations. and the Sturm–Liouville boundary value
The actions of diffeomorphisms on vec- problem.
tor fields and flows (changes of coordinates The appendix, “Sample Examination
for differential equations) is discussed in Problems,” has been expanded to include
depth. This sets the stage for the intro- a sample examination from Moscow State
duction of symmetry groups—Lie groups University and many new supplementary
as they first appeared—and their applica- problems.
tion to the solution of homogeneous and Every serious student of ordinary differ-
quasi-homogeneous differential equations. ential equations should read this book. In
The treatment here exemplifies the differ- the U.S. system, it is an excellent text for an
ence between the Arnold style and the cook- introductory graduate course. While most
book style of most elementary (and widely of the presentation is not out of the reach
used) texts on ordinary differential equa- of undergraduates, it will be unlikely to
tions. The cookbooks give the change of replace the mathematics cookbook in the
variables required to transform a homoge- standard sophomore introduction to ordi-
neous equation into a linear equation (which nary differential equations. Also, most un-
can be solved explicitly), but Arnold reveals dergraduate students do not have time in
the origin of this transformation as a natu- their schedules to take a second course on or-
ral result of an underlying symmetry. dinary differential equations. On the other
Chapter 2, “Basic Theorems,” has been hand, the U.S. system regularly produces
expanded from 47 to 63 pages. It con- students who overcome their sophomore
tains the sections “Rectification Theorems” introduction to differential equations by
(which includes the important theorems on “complete automatization” (perhaps with
existence, uniqueness, and smooth depen- some insight provided by Professor Arnold)
dence on initial data and parameters), “Ap- and eventually gain a complete understand-
plications to Equations of Higher Order ing of this wonderful and useful subject.
than First,” “The Phase Curves of an Au- CARMEN CHICONE
tonomous System,” “The Derivative in the University of Missouri
BOOK REVIEWS 337
beci and Jean Cousteix. Springer-Verlag, New linear stability (Orr–Sommerfeld equation)
York, 2005. $159.00. xvi+502 pp., hardcover. as well as through application of eigenvalue
ISBN 3-540-24459-X. methods such as the en method. Where ap-
plicable, empirical correlations for the on-
Boundary layer theory has a long his- set and end of the transition regime are
tory; its mathematical formulation dates also given that can be profitably used in
back to Ludwig Prandtl (1904) and is fa- integral as well as finite-difference solutions
mously known to every postgraduate stu- to the entire laminar, transitional through
dent and researcher in fluid dynamics as the to turbulent development of a wall bound-
boundary layer approximations of the three- ary layer. Chapters 6, 7, and 12 deal with
dimensional Navier–Stokes (NS) equations. turbulent flows where approaches to turbu-
The history is indeed very rich in a variety lence modeling through zero-equation (or
of problems of practical relevance for which, mixing length) models as well as through
in the days before modern CFD, solutions models based on transport equations (e.g.,
had to be found, the main ones being the one-equation models based on νt -equation
fluid dynamic drag of a body and its coun- and two-equation models based on k- and
terpart the fluid dynamic lift of a body. The k-ω variables) have been listed. These are
latter, of course, is of particular interest to used to model isotropic turbulent viscosity
aeronautical engineers. µt in the Stokes-like relation
Although several books and treatises over
∂ui ∂uj 2
the years have provided fruitful descrip- − ρ u i u j = µt + − δij k,
∂xj ∂xi 3
tions of this theory to various depths, the
book under review provides depths previ- where coplanarity between stress and strain
ously unmatched in a single volume (albeit is indicated. While this approach is quite
while dealing with flows of interest only to satisfactory in most long-and-thin bound-
aeronautical engineers). It is this aspect ary layer flows, there are more complex
that should immediately attract a potential situations where transport equations for
buyer of this book. −ρui uj must be solved directly. Such stress
The book begins invitingly with case equation models have also been listed. Mod-
studies of (a) skin-friction drag reduction els of increasing complexity are required to
through shaping of airfoils with the aim capture effects of surface curvature, rapidly
of delaying the onset of transition and varying pressure gradients and suction, and
through boundary layer suction, and (b) other effects where the viscous inner layer of
prediction of the maximum lift-coefficient of a turbulent boundary layer is influenced by
multi-element wings. This telling introduc- events in the outer layers. In this respect,
tion not only does justice to the authors’ description of the great near-universality
long-standing contributions to aerodynam- of the inner and outer layers of a turbu-
ics research but also creates curiosity in a lent boundary layer that was established by
reader about the mathematical nuances of (say) the late 1950s through phenomeno-
the developments in this theory that are to logical arguments is a most welcome fea-
follow in the rest of the book. ture. Another feature that will be partic-
In this 13-chapter book, Chapters 2 to ularly welcomed by students and teachers
9 are devoted to incompressible flows and is the large number of end-of-chapter prob-
Chapters 10 to 13 to compressible flows. lems whose solutions will be an enriching
Laminar, transitional, and turbulent flows experience. To aid solutions, a CD-ROM
are considered in two and three dimensions. containing computer programs is provided.
Both the wall boundary layers as well as free Chapters 9 and 13 are devoted to interac-
shear layers are considered. These chapters tive boundary layer theory (IBLT). Conven-
comprehensively illustrate how, when, and tionally, the boundary layer equations for a
why similarity, local nonsimilarity, integral, wall flow are solved for a prescribed pressure
and finite-difference (Keller’s box method gradient that is evaluated from inviscid flow
is preferred) methods are to be employed theory. In real viscous flows, however, the
338 BOOK REVIEWS
true pressure gradient deviates from that admirably on both counts. Although the
derived from inviscid flow theory. This is methods presented are often motivated by
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
parametric multiplicity problems for the this is the way to teach mathematics. Use-
Lienard and the Rayleigh equations (the ful resource for writing lectures? Certainly.
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
Ambrosetti–Prodi problem) and further ex- Good way to read and learn about analysis?
tensions of them. Resonance and nonres- You decide.
onance problems are studied in Chapter DONALD ESTEP
7, along with Landesman–Lazer conditions Colorado State University
and results based on the Fredholm alterna-
tive. The existence of positive solutions to
the Dirichlet problem
Triangulations and Applications. By Oyvind
Hjelle and Morten Daehlen. Springer-Verlag,
u + f (t, u) = 0, u(a) = 0 = u(b)
Berlin, 2006. $64.95. xi+234 pp., hardcover.
ISBN 3-540-33260-X.
is considered in Chapter 8. Various types of
singular problems are considered in Chapter
9, and in Chapter 10 the lower and upper Let me start with what I like least about the
solution method is applied to singular per- book: the title. I find it fairly misleading: in
turbations. the area of visualization (the book appears
In addition to the very well presented in a series called Mathematics and Visu-
introduction, the authors include an ex- alization), the term triangulation typically
tensive list of bibliographical notes to each refers to a 3D mesh. The book, however,
of the chapters and also provide an ap- is entirely restricted to triangulations of 2D
pendix which includes basic statements on planar domains.
degree theory, variational methods, spectral For those domains, however, it does pro-
results, and the maximum/antimaximum vide a detailed and adequate treatment. A
principles. There are more than 300 refer- brief overview of the contents:
ences to the literature as well. • basic facts and algorithms about tri-
This is a valuable book because of the angles and triangulations;
exhaustive treatment of the lower and up- • graphs and data structures;
per solution method in its many forms, as • Delaunay triangulations and Voronoi
well as the historical details and manifold diagrams: theory and algorithms;
developments in the theory since its incep-
• data dependent triangulations;
tion. Many illustrative examples are also
presented in which lower and upper solu- • constrained triangulations;
tions are constructed. This should be a • refinement and mesh generation;
very useful book for researchers in the area • piecewise linear least squares approx-
of BVPs and applications. imation;
LYNN ERBE • introduction to the companion soft-
University of Nebraska–Lincoln ware.
Triangulations are well understood in the
field of CG (computational geometry) [1]—
Metric Spaces. By Satish Shirali and Harkri- one might ask what another text could add.
shan Vasudeva. Springer-Verlag, London, 2006. This book is definitely not a rehash of CG
$44.95. viii+222 pp., softcover. ISBN 1-85233- concepts. CG is mainly concerned with the
922-5. order of complexity of geometric algorithms
or uniqueness of solutions. These concepts
This book introduces the fundamentals of are covered briefly here as well, but the
analysis in metric spaces. It’s written in main approach is totally different, and in
a very spare theorem-proof-example style; a sense is complementary to the CG view
has illustrative examples and exercises; of triangulations. The overall flavor of the
spends little time on discussion, develop- book reflects the use of triangulations for
ment of intuition, or substantial applica- the description of planar objects and as the
tions; begins by stating that the abstract domain for bivariate functions.
postulational method has a vital role in Triangulated planar objects are meshes;
modern mathematics; implicitly assumes important topics such as mesh editing or
340 BOOK REVIEWS
One of the main strengths of the book is Cambridge University Press, Cambridge, UK,
the coverage of bivariate functions over tri- 2005. $65.00. xii+219 pp., hardcover. ISBN
angulations. These functions are piecewise 0-521-78124-8.
linear over the triangulation and are used
to approximate bivariate functions as well Hydrodynamic turbulence is often referred
as discrete data. This problem gives rise to to as the last open problem of classical me-
a kind of triangulation which is not con- chanics. Despite nearly a century of efforts
sidered in CG: the data dependent triangu- by the best minds, many questions remain
lation. For this, triangles are not optimal open, and considerable progress has only
with respect to a Delaunay criterion but been achieved for the special case of homo-
rather they are constructed such that they geneous and isotropic turbulence in three
“line up” optimally with a given function. dimensions [15, 16, 4]. To a lesser extent,
The chapter on least squares approxi- there has also been some progress toward
mations is excellent. It shows how to re- understanding homogeneous and isotropic
alistically approximate huge data sets by turbulence in two dimensions [11, 20, 5].
a manageable piecewise linear surface. A On the theoretical front there has been very
mix of interpolation and approximation is limited success in understanding inhomoge-
also covered which is important in several neous turbulence, compressible turbulence,
applications. and geophysical turbulence.
The book is meant as a text for a graduate Our intuitive understanding of turbu-
class on triangulations. Such a class would lence goes back to Richardson [18], Kol-
fit into the area of scientific computing or mogorov [9, 8], and Batchelor [1], who con-
numerical analysis. A course text should jectured that if energy is injected into the
have problems and exercises, and indeed system by forcing at large scales, then for
there are about eight problems at the end sufficiently large Reynolds number, hydro-
of each chapter, with a good mix of theoret- dynamic instability results in the transfer
ical problems and programming exercises. of energy to smaller scales, and for suffi-
For the programming part, students (as ciently small scales this energy is dissipated
well as other readers) do not have to start by viscosity. In between the forcing length
from scratch: a complete software pack- scale 0 , where the energy comes in, and the
age is available from a companion web site. dissipation length scale η, where the energy
There one can find descriptions of the ba- comes out, there is a range of scales, where
sic data structures as well as algorithms the energy is cascaded by local nonlinear
for constrained and unconstrained Delau- interactions from wavenumber to wavenum-
nay triangulations, a triangulation editor, ber down the range. This region is called
and various query tools. This reviewer did the inertial range, and it was conjectured
not try it out, but knowing that it is based by Kolmogorov that it is governed by uni-
on highly successful work by the Norwegian versal statistical principles. The conjecture
SINTEF organization, there is little doubt was motivated by the notion that at length
that it is a good product. scales in the inertial range, the system for-
gets how the energy gets there, and the
only dynamics in the inertial range is the
REFERENCE
transfer of energy to smaller scales. This
conjecture leads to predictions about the
[1] M. de Berg, M. van Krefeld, M. Over- energy spectrum E(k) of turbulence as well
mars, and O. Schwarzkopf, Compu- as the scaling properties of the statistical
tational Geometry: Theory and Ap- moments of velocity differences.
plications, 2nd ed., Springer-Verlag, Kolmogorov’s idea remained a conjecture
Berlin, 2000.
until 1962, when it was confirmed exper-
imentally by measurements of the veloc-
GERALD FARIN ity of deep oceanic currents [6]. It was in
Arizona State University that same year that Kolmogorov [10] and
BOOK REVIEWS 341
dictions, and further work over the next 3 discusses the traditional LES models. An-
decades showed that this was indeed true other chapter of general interest is Chap-
[19]. So, one of the open questions that ter 4, which discusses spectral LES models
has received a lot of theoretical attention such as EDQNM and RNG. These models
is, Why is the energy spectrum E(k) of are often used, not only for numerical simu-
turbulence in the inertial range in such sig- lations, but also for theoretical arguments,
nificant agreement with the predictions of as in, for example, [14]. There’s also an in-
Kolmogorov’s 1941 theory (K41), and what teresting discussion of the famous “bump”
is the origin of the deviations from that in the inertial range energy spectrum, for
theory? which a variety of explanations have been
Now, one can argue that we already have suggested [2, 12]. The remaining chapters
a pretty good theory of hydrodynamic tur- discuss various more modern LES models
bulence: the Navier–Stokes equations! All and a wide range of engineering and indus-
the physics of the Kolmogorov energy cas- trial applications of these models. Chap-
cade can come from the numerical solu- ter 5 discusses models for inhomogeneous
tion of the Navier–Stokes equations! The turbulence, Chapter 6 discusses structure
problem is that a numerical solution of the function models, and Chapter 7 discusses
Navier–Stokes equations by direct numeri- models for compressible turbulence. Com-
cal simulation is not possible for Reynolds pressible turbulence is of unique interest
numbers large enough to be of any practi- due to its applications in aerodynamics and
cal interest. To the best of my knowledge, aerospace engineering. Finally, Chapter 8
the state of the art in direct numerical sim- discusses geophysical fluid dynamics, begin-
ulations is 40963 resolution [7], performed ning with a very nice conceptual introduc-
by one of the largest supercomputers in tion, and emphasizes the modeling of storm
the world, the Earth Simulator in Japan. formation.
Furthermore, for geophysical fluid dynam- For graduate students, I should note that
ics and engineering applications one has to the book does not give a comprehensive
go beyond the Navier–Stokes equations and introduction to the theory of turbulence,
consider more accurate models that include Chapter 1 notwithstanding. One should
temperature and density. A practical al- also study other textbooks, starting from
ternative to direct numerical simulation is Frisch [3] and Lesieur [13], to gain a deeper
large eddy simulation (LES), the topic of appreciation of what is understood and
the book under review. The essential idea what is not understood in this very excit-
is that in practical situations we are inter- ing research area, before specializing into
ested in the effect of turbulence on fluid numerics. The book also doesn’t discuss
motions at large scales, so we go ahead and the numerical implementation of the cor-
model the nonlinear dynamics at smaller responding simulation codes. On the other
scales so that it is not necessary to resolve hand, the objective of the book, as stated in
them numerically. the introduction, is to present in detail a va-
The book begins in the preface with a riety of LES models, and to give the reader
delightful historical introduction by Jim Ri- a thorough understanding of turbulence dy-
ley. The first chapter gives a very concise namics through numerical results obtained
overview of the Kolmogorov theory as well by these models. The authors have done an
as a conceptual introduction to LES in gen- excellent job in achieving these objectives.
eral. The second chapter gives a very inter- Furthermore, they have included material,
esting discussion of vortex dynamics which such as in Chapter 7, that hasn’t been re-
will be of interest to a wider audience of viewed previously in the literature, as well
turbulence researchers. Emphasis is given as material of general interest to the com-
to criteria that can be used to characterize munity, such as in Chapters 2 and 4. With
coherent structures in turbulence. In ad- its impressive bibliography, it will be an
dition to researchers of three-dimensional invaluable resource for students who want
turbulence, this topic is also relevant to the to study the literature of this field, as well
342 BOOK REVIEWS
as researchers who are currently working in [13] M. Lesieur, Turbulence in fluids, Kluwer
LES. Academic, Dordrecht, The Nether-
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
lands, 1990.
[14] V. S. L’vov, A. Pomyalov, and I. Pro-
REFERENCES caccia, Quasi-Gaussian statistics of
hydrodynamic turbulence in 3/4 +
[1] G. Batchelor, Kolmogorov’s theory of dimensions, Phys. Rev. Lett., 89
locally isotropic turbulence, Proc. (2002), article 064501.
Camb. Phil. Soc., 43 (1947), pp. 533–
[15] V. S. L’vov and I. Procaccia, Hydro-
559.
dynamic turbulence: A 19th century
[2] G. Falkovich, Bottleneck phenomenon in problem with a challenge for the 21st
developed turbulence, Phys. Fluids, 6 century, in Turbulence Modeling and
(1994), pp. 1411–1414. Vortex Dynamics (Istanbul, 1996),
[3] U. Frisch, Turbulence: The legacy of O. Boratav, A. Eden, and A. Erzan,
A. N. Kolmogorov, Cambridge Uni- eds., Lecture Notes in Phys. 491,
versity Press, Cambridge, UK, 1995. Springer-Verlag, Berlin, 1997, pp. 1–
[4] M. Giles, Anomalous scaling in homoge- 16.
neous isotropic turbulence, J. Phys. [16] V. L’vov and I. Procaccia, Analytic cal-
A, 34 (2001), pp. 4389–4435. culation of the anomalous exponents
[5] E. Gkioulekas and K. K. Tung, Re- in turbulence: Using the fusion rules
cent developments in understand- to flush out a small parameter, Phys.
ing two-dimensional turbulence and Rev. E, 62 (2000), pp. 8037–8057.
the Nastrom-Gage spectrum, J. Low
[17] A. Oboukhov, Some specific features of
Temp. Phys., 145 (2006), pp. 25–57.
the atmospheric turbulence, J. Fluid.
[6] H. Grant, R. Stewart, and A. Moil- Mech., 13 (1962), pp. 77–81.
liet, Turbulence spectra from a tidal
channel, J. Fluid. Mech., 12 (1962), [18] L. Richardson, Weather prediction by
pp. 241–263. numerical processes, Cambridge Uni-
versity Press, Cambridge, UK, 1922.
[7] Y. Kaneda, T. Ishihara, M. Yokokawa,
K. Itakura, and A. Uno, Energy dis- [19] K. Sreenivasan and R. Antonia, The
sipation rate and energy spectrum in phenomenology of small-scale turbu-
high resolution direct numerical simu- lence, Ann. Rev. Fluid Mech., 29
lations of turbulence in a periodic box, (1997), pp. 435–472.
Phys. Fluids, 15 (2003), pp. L21–L24. [20] P. Tabeling, Two dimensional turbu-
[8] A. Kolmogorov, Dissipation of energy in lence: A physicist approach, Phys.
the locally isotropic turbulence, Dokl. Rep., 362 (2002), pp. 1–62.
Akad. Nauk. SSSR, 32 (1941), pp. 16–
18 (in Russian). English translation ELEFTHERIOS GKIOULEKAS
in Proc. Roy. Soc. London Ser. A, 434 University of Central Florida
(1991), pp. 15–17.
[9] A. Kolmogorov, The local structure of
turbulence in incompressible viscous
fluid for very large Reynolds num- A Guide to MATLAB for Beginners and
bers, Dokl. Akad. Nauk. SSSR, 30 Experienced Users. Second Edition. By
(1941), pp. 301–305 (in Russian). En- Brian R. Hunt, Ronald L. Lipsman, and Jonathan M.
glish translation in Proc. Roy. Soc. Rosenberg. Cambridge University Press, New
London Ser. A, 434 (1991), pp. 9–13.
York, 2006. $50.00. xv+309 pp., softcover.
[10] A. Kolmogorov, A refinement of pre-
ISBN 0-521-61565-8.
vious hypotheses concerning the local
structure of turbulence in a viscous
incompressible fluid at high Reynolds This is a readable introduction to using
number, J. Fluid. Mech., 13 (1962), MATLAB 7 (and the immediate prede-
pp. 82–95. cessor versions of MATLAB). Much of the
[11] R. Kraichnan and D. Montgomery, book is devoted to describing how to use the
Two dimensional turbulence, Rep. various computational facilities provided by
Prog. Phys., 43 (1980), pp. 547–619.
MATLAB and only a small proportion to
[12] S. Kurien, M. Taylor, and T. Mat-
the language itself and to its mathematical
sumoto, Cascade time-scales of en-
ergy and helicity in homogeneous, iso- software library. The authors assume from
tropic turbulence, Phys. Rev. E, 69 the start that the reader has access to MAT-
(2004), article 066313. LAB, Simulink, and the Symbolic Toolbox,
BOOK REVIEWS 343
which is essentially the content of the MAT- to set up and solve finite difference equa-
LAB and Simulink Student Version. There tions for the heat equation with variable
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
are also references to and examples of the conductivity and, alternatively, how to use
use of some of MATLAB’s other toolboxes, the built-in MATLAB function pdepe to
but these are less likely to be available in solve the same problem.
the classroom. As promised in the book’s
title, it switches frequently between intro-
REFERENCE
ductory and advanced material, sometimes
without warning. [1] D. J. Higham and N. J. Higham, MAT-
While reading this book I compared it for LAB Guide, 2nd ed., SIAM, Philadel-
content with another well-known textbook/ phia, 2005.
manual on MATLAB 7 by Higham and
IAN GLADWELL
Higham (H & H) [1]. Of course, both
Southern Methodist University
texts are already, marginally, out of date as
MATLAB is now at version R2006a, and
the next release, R2006b, will be available
before this review is published. Sturm–Liouville Theory: Past and Pre-
These two books could hardly be more sent. Edited by Werner O. Amrein, Andreas M.
different. Indeed, they could be viewed as Hinz, and David B. Pearson. Birkhäuser, Basel,
complementary. H & H’s text is more nar- 2005. $89.95. xx+335 pp., hardcover. ISBN
rowly focused, is somewhat longer, and has 3-7643-7066-1.
a much deeper coverage of the MATLAB
language and its mathematical software li- When I first saw this book on display at the
brary. In the book under review (H, L & Joint Mathematics Meeting, I immediately
R), the symbolic approach to solving prob- added it to my mental list of “must reads”
lems often takes precedence over numerical even though I did not stop to look at the con-
approaches, whereas in H & H discussion of tents. So when the opportunity to review
the symbolic toolbox is restricted to one (of it arose, I jumped at the chance since now
the 22) chapters. In H, L & R, a chapter is “must” would become “have-to.” I must
devoted to Simulink and there is discussion confess a little disappointment when I first
of it elsewhere in the text, whereas H & H do opened the book. I was expecting a mono-
not appear to mention it. H, L & R include graph detailing the history, significance, and
one chapter and some earlier discussion of intertwining of the Sturm–Liouville story in
publish and related new commands intro- one continuous narrative, but instead found
duced in MATLAB 7, and another chapter a collection of essays on various aspects of
on the construction of GUIs; H & H mention the problem. My disappointment was ex-
some of these topics briefly but essentially tremely short lived, however, for what we
only to say that they are outside the scope have here is a collection of important pa-
of their discussion. pers showing the impact that Sturm, and
Features of the new book include three Sturm–Liouville theory in particular, has
practice sets with complete answers, a long, had on the direction of mathematical re-
and mainly straightforward, chapter on ap- search in the last 150 plus years. It is the
plications and modeling, a chapter on trou- fundamental papers by Sturm [10, 11, 12]
bleshooting and debugging, and a glossary and his joint paper with Liouville [13] that
of MATLAB commands. The practice sets form the basis for what we refer to today
give a flavor of the content and level of as Sturm–Liouville theory.
the book: Algebra and Arithmetic; Cal- In its simplest form, the basic Sturm–
culus, Graphics, and Linear Algebra; and Liouville problem consists of the differential
Developing your MATLAB Skills. The ap- equation
plications and modeling chapter digs a lit- (1) (py ) + λqy = 0
tle deeper, including showing how to use
both Simulink and the MATLAB built-in together with some boundary conditions,
function ode45 to solve a simple ordinary for example, the Dirichlet
differential equation. It is also shown how (2) y(0) = 0 and y(L) = 0,
344 BOOK REVIEWS
(3) (py )(0) = 0 and (py )(L) = 0, ground-breaking results, and gives sufficient
or some other type of boundary condition. references to survey monographs so that in-
The problem then is to find values of the terested readers can explore some of these
parameter λ for which the problem (1) and ideas in detail on their own.2
(2) (or (1) and (3)) has nontrivial solutions. Simon discusses Sturm theory for differ-
In commemoration of the 200th year of ence equations in his paper “Sturm Oscilla-
the birth of Sturm, a conference took place tion and Comparison Theorems,” and as he
at the University of Geneva (Switzerland) points out, it has long been believed that
from September 15 to 19, 2003. The sixty Sturm obtained his famous comparison re-
participants represented sixteen different sults for difference equations as well as for
countries. The first part of the conference differential equations. In fact, some believe
was devoted to Sturm’s contributions to that Sturm may have discovered these re-
mathematics and physics; the second part sults for difference equations before doing
focused on Sturm–Liouville theory and in- so for differential equations (see Reid [9]).
cluded general presentations of a survey The first place that the difference equation
nature as well as some on recent research in results appear in print is usually attributed
the field. The book under review is based to Bôcher [2, 3] in 1898.
on the second part of the conference and While Everitt in “Charles Sturm and the
contains six survey papers followed by six Development of Sturm–Liouville Theory in
articles devoted more to topics of current the Years 1900 to 1950” issues a disclaimer
research interest. A separate volume of that his paper not be “counted as a history
Sturm’s complete works is being prepared of Sturm–Liouville theory for the period
for publication [8]. 1900 to 1950” due to a lack of sufficient
In “Sturm’s 1836 Oscillation Results, references to many mathematicians in his
Evolution of the Theory,” Hinton gives a bibliography, what he presents is in fact an
detailed history and discussion of the Sturm eloquent description of the development of
comparison theorem and the origin of os- Sturm–Liouville theory in that time period
cillation theory for the second order linear and its influence on the direction and devel-
equation opment of modern analysis. In particular,
he documents the work of Weyl, Dixon,
(4) (py ) + qy = 0, Stone, and Titchmarsh during this period.
In each case, Everitt details concisely the
including some of the basic results of Fite,
significant contributions that each of these
Wintner, and Leighton. The Sturm com-
important figures made to the advancement
parison theorem simply states that if (4) is
of Sturm–Liouville theory beginning with
oscillatory,1
the classic paper of Weyl [15], who intro-
Q≥q and p ≥ P, duced the famous limit-point/limit-circle
classification of solutions.3
then the equation Weidmann (“Spectral Theory of Sturm–
(P y ) + Qy = 0 Liouville Operators. Approximation by
results on further directions in research. theory for linear equations, for example, is the
While this survey at times seems to pro- book by Swanson [14].
ceed at breath-taking speed, touching on 3 For a discussion of the limit-point/limit-
differential equations in the complex do- circle problem, the Weyl alternative, the exten-
main, higher order equations, systems, and sion of the limit-point/limit-circle problem to
nonlinear equations, and the relationship be-
1 Recall that for a second order linear equa- tween the limit-point/limit-circle problem and
tion, if one solution oscillates, then all solutions other asymptotic properties of solutions, see the
are oscillatory. monograph by Bartušek, Došlá, and Graef [1].
BOOK REVIEWS 345
of Sturm–Liouville operators with particu- is on what are called “nodal sets” as they
lar attention to maximal and minimal oper- relate to Sturm’s oscillation results for lin-
ators, the Weyl alternative, and deficiency ear problems. Although it is not mentioned
indices. Of special interest here is the ap- here, in recent years the basic oscillation
proximation of the spectral properties of and comparison results of Sturm have been
singular problems by using regular prob- extended in a variety of directions to non-
lems and appropriate limiting arguments. linear problems, for example, to nonlinear
In “Spectral Theory of Sturm–Liouville differential equations, nonlinear difference
Operators on Infinite Intervals: A Review of equations, equation with delays and neutral
Recent Developments,” Last gives a survey terms, and even to higher order nonlinear
of recent work on spectral theory of Sturm– differential and difference equations, where
Liouville operators on infinite intervals over the complexity of the notation due to the
the last thirty years in both the continuous generality of the equations tends to obscure
and the discrete cases. The discussion of the the results (see, for example, [5]). Most
location of the spectrum and their spectral recently, the comparison results have been
measures is especially of interest here. extended to nonlinear dynamic equations
Gilbert examines the connection between on time scales.
asymptotic behavior of solutions of singular The relationship between the bound-
Sturm–Liouville operators and the spectral ary conditions and the spectra of Sturm–
properties of corresponding self-adjoint op- Liouville operators is discussed by del Río
erators in her paper “Asymptotic Methods in “Boundary Conditions and Spectra of
in the Spectral Analysis of Sturm–Liouville Sturm–Liouville Operators.” This relates
Operators.” Her approach makes use of the back to the work of Weyl [15], who showed
notion of subordinacy of solutions. the stability of the essential spectrum with
The paper by Bennewitz and Everitt, respect to changes in the boundary condi-
“The Titchmarsh–Weyl Eigenfunction Ex- tions.
pansion Theorem for Sturm–Liouville Dif- “Uniqueness of the Matrix Sturm–
ferential Equations,” gives the histori- Liouville Equation Given a Part of the Mon-
cal background of the Titchmarsh–Weyl odromy Matrix, and Borg Type Results” by
eigenfunction expansion results for Sturm– Malamud is a survey of recent results on in-
Liouville differential equations. They give verse spectral theory based on the approach
a new version of the expansion theorem used by Borg [4]. This is an important and
that includes both the continuous and the expanding area of research in which infor-
discrete cases of the spectrum. They make mation about the spectrum is used to obtain
extensive use of the Titchmarsh–Weyl m- the Sturm–Liouville equation itself. The in-
coefficient. terest in this article is in matrix equations
The paper “Sturm’s Theorems on Zero rather than the scalar case.
Sets in Nonlinear Parabolic Equations” by The last paper in this volume, “A Cata-
Galaktionov and Harwin is particularly in- logue of Sturm–Liouville Differential Equa-
teresting from the point of view that many tions” by Everitt, is unique in that it is
people who work on ordinary differential a collection of more than fifty examples
equations may be unaware of Sturm’s re- of Sturm–Liouville equations along with a
sults [12] on the zeros of solutions of partial classification of their endpoints and infor-
differential equations. As the authors point mation about the location of their eigenval-
out, this fact was essentially ignored un- ues and the spectra of their corresponding
til a paper by Pólya [7] appeared in 1933. differential operators. What a wealth of in-
The authors of this contribution, in par- formation in an easily accessible format. In
ticular, discuss the situation for parabolic many cases their connection to special func-
equations and survey results from the last tions is also indicated. Even better, this
quarter century. article begins with a very nicely written
Chen, in “A Survey of Nonlinear Sturm– concise summary of Sturm–Liouville the-
Liouville Equations,” discusses nonlinear ory. This in itself is worth the price of the
346 BOOK REVIEWS
monograph. In cases where the equation [11] C. Sturm, Mémoire sur les Équations
is linked to a physical application, this is différentielles linéaires du second or-
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
noted as well. Known references to numer- dre, J. Math. Pures Appl., 1 (1836),
pp. 106–186.
ical approximations of the eigenvalues are
[12] C. Sturm, Mémoire sur une classe
also given.
d’Équations à différences partielles,
As a final note, the elegant introduction J. Math. Pures Appl., 1 (1836), pp.
by David Pearson gives some background 373–444.
on Sturm himself, his collaboration with [13] C. Sturm and J. Liouville, Extrait d’un
Liouville, and the beautiful connection be- Mémoire sur le développement des
tween Sturm–Liouville theory and quantum fonctions en séries dont les différents
mechanics. It is highly worthwhile reading termes sont assujettis à satisfaire
all on its own. à une méme équation différentielle
linéaire, contenant un paramètre
This volume has something for every-
variable, J. Math. Pures Appl., 2
one who has any interest in the impact of (1837), pp. 220–223.
Charles François Sturm’s fundamental work [14] C. A. Swanson, Comparison and Os-
on current mathematical research. cillation Theory of Linear Differen-
tial Equations, Academic Press, New
York, 1968.
REFERENCES [15] H. Weyl, Über gewöhnliche Differential-
gleichungen mit Singularitäten und
[1] M. Bartušek, Z. Došlá, and J. R.
die zugehörige Entwicklung willkür-
Graef, The Nonlinear Limit-Point/
licher Funktionen, Math. Ann., 68
Limit-Circle Problem, Birkhäuser
(1910), pp. 220–269.
Boston, Boston, MA, 2004.
[2] M. Bôcher, The theorems of oscillation JOHN R. GRAEF
of Sturm and Klein (First Paper),
University of Tennessee at Chattanooga
Bull. Amer. Math. Soc., 4 (1898), pp.
295–313.
[3] M. Bôcher, The theorems of oscillation
of Sturm and Klein (Second Paper), Nonlinear Problems of Elasticity. Second
Bull. Amer. Math. Soc., 4 (1898), pp. Edition. By Stuart S. Antman. Springer-Verlag,
365–376. New York, 2005. $89.95. xviii+831 pp., hard-
[4] G. Borg, Eine Umkehrung der Sturm- cover. ISBN 0-387-20880-1.
Liouvilleschen Eigenwertaufgabe,
Acta Math., 78 (1946), pp. 1–96.
Nonlinear (finite) elasticity is a central
[5] J. R. Graef, A. Miciano-Cariño, and
C. Qian, A general comparison re- paradigm of continuum mechanics. Cor-
sult for higher order nonlinear differ- rectly formulated problems of the subject,
ence equations with deviating argu- incorporating the exact geometry of de-
ments, J. Difference Equations Appl., formation and nonlinear material response,
8 (2002), pp. 1033–1052. are typically difficult; there are many open
[6] Z. Nehari, Characteristic values associ- problems. The subject has immediate ap-
ated with a class of nonlinear second-
plicability to flexible structures and rubber-
order differential equations, Acta
Math., 105 (1961), pp. 141–175. like solids, and it forms a foundation for
[7] G. Pólya, Qualitatives über Wärmeaus- theories of more exotic solid materials un-
gleich, Z. Angew. Math. Mech., 13 dergoing finite deformations. In spite of its
(1933), pp. 125–128. classical roots (dating back to Euler and
[8] J.-C. Pont, ed., The Collected Works of Cauchy, for example), the nonlinear theory
Charles François Sturm, in prepara- was not finalized until the second half of the
tion. twentieth century—a time when nonlinear
[9] W. T. Reid, A historical note on Sturmian analysis also emerged as a mature mathe-
theory, J. Differential Equations, 20 matical subject. The influence of elasticity
(1976), pp. 316–320.
on analysis is well known [1, 2]. The ap-
[10] C. Sturm, Extrait d’un Mémoire sur
plication of modern methods of analysis to
l’intégration d’un système d’équa-
tions différentielles linéaires, Bull. nonlinear elasticity is an ongoing, fertile en-
Sci. Math. Férussac, 12 (1829), pp. deavor. This brings us to the book under
313–322. review.
BOOK REVIEWS 347
overwhelming and can give one the sense of results. Issues with diverse chemical condi-
drinking from a fire hydrant. tions and bottlenecks in the crystallization
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
This is where knowledge discovery plays and analysis process are also elucidated.
an important role and motivates the timely Chapter 6 is a discussion of the approaches
publication of the book by Jurisica and and tools used to integrate diverse forms
Wigle. Knowledge discovery, as the authors of data from public databases as well as
define it, is the process of finding useful pat- the significant issues faced and the benefits
terns in data with an emphasis on the end gained in that process. The book is con-
product of the process, which is knowledge. cluded in Chapter 7 with an overview of
Naturally, this iterative process involves the systems biology and a description of how
management of available knowledge in a knowledge discovery is involved in advanc-
way that optimizes sharing and collabora- ing various areas of research in this field.
tion. However, when attempting to merge The authors recognize the importance of
information it is difficult to determine what understanding how experimental and theo-
information is useful and what should be retical techniques interact in the knowledge
included. It is also difficult to plan ahead discovery process; and therefore they try
because it is impossible to predict what to strike a balance between these areas
new methods and discoveries will be devel- in the book. Researchers involved in the
oped and lead to new data representation knowledge discovery process come from a
requirements years from now. In Chapters variety of different backgrounds including
1 and 2, the authors provide an overview of biology, physics, mathematics, computer
the knowledge discovery process with sev- science, and engineering. Therefore, in or-
eral specific examples that emphasize the der to coordinate everyone’s efforts, it is
usefulness of approaches that have been em- necessary that researchers from all these
ployed while pointing out limitations that areas understand how their results and tools
need to be addressed. By doing so, they will be used by others, and that they conse-
clearly express the need for defining stan- quently optimize for sharing and collabora-
dards for experimental data representation tion. Throughout the text, impressive ex-
that are both comprehensive and flexible, amples from the literature and the authors’
and they motivate the discussion that fol- personal research experiences highlight this
lows in subsequent chapters. key concept and point out where limitations
The authors discuss several influential ex- and major challenges still exist.
perimental and theoretical techniques em- The book was written by eight contribut-
ployed in bioinformatics, proteomics, and ing authors with varying degrees of expe-
systems biology today. Chapter 3 provides rience, from Ph.D. candidates to faculty
an introduction to mass spectrometry (MS) members. As a result, there is a degree
and an overview of the state-of-the-art MS- of disconnect between chapters in terms of
based experimental procedures. Some of the writing styles as well as in the extent to
data processing tools and search algorithms which each topic is addressed. That crit-
used to manage and extract information icism aside, this is a timely publication
from global proteomic studies are also de- and, for the most part, a successful effort
scribed in Chapter 3. Chapters 4 and 5 go to depict many hot areas of research. I
into much more depth than the other chap- would highly recommend the book to new
ters and make up the majority of the content researchers, as it describes the limitations
of the book. Chapter 4 is an in-depth ex- of tools and techniques being used today
planation of the graph theory analysis of and draws attention to many open prob-
protein-protein interactions (PPIs). It ex- lems. I would also recommend the book to
pounds the terminology, types of network more experienced researchers who have a
models, topological features, motifs, algo- primarily theoretical or experimental bent,
rithms, and public data sets used to infer as attaining the goal of unifying existing
and study PPIs. Chapter 5 is a discussion of knowledge will require a concerted effort
HTP protein crystallization approaches and from everyone involved.
the automated tools used to analyze and
WILLIAM J. HEUETT
extract information from the experimental
University of Colorado
BOOK REVIEWS 349
sity Press, Cambridge, UK, 2006. $55.00. xii+ jections, etc., are used in differential the-
233 pp., softcover. ISBN 0-521-68738-1. ories, and composition of these leads ex-
plicitly to the idea of a category; but
Forty years ago it became apparent that various relevant categories arise, depending
certain categorical methods, devised for on which objects are admitted and which
the study of algebraic and analytic spaces, functorial constructions on them are pos-
can also be applied to smooth (C-infinity) sible. The variety of categories also arises
spaces. The ramifications of this obser- from the fact that useful levels of smooth-
vation for continuum physics are still be- ness (polynomial, analytic, or an intermedi-
ing worked out, but major advances were ate level such as the one generated by bump
recorded in the 1991 Springer book Models functions) all give rise to categories whose
for Smooth Infinitesimal Analysis by Ieke general properties and geometric uses are
Moerdijk and Gonzalo Reyes; undergrad- somewhat similar, so that a common non-
uate textbooks on differentiable analysis trivial theory is possible. We refer here to
that explicitly make use of the resulting the categories described by that theory as
insights were prepared by René Lavend- “smooth toposes.”
homme (1996) and by John Bell (1998). Besides the usual cartesian products
These three books, as well as the research (used by Galileo for the description of mo-
papers by Marta Bunge, Eduardo Dubuc, tion), two of the striking features of smooth
Felipe Gago, and others, descend from the toposes are the representability (described
first book to treat these developments, Syn- more fully two paragraphs below) of tan-
thetic Differential Geometry (Cambridge, gent bundles and of function spaces. For
1981) by Anders Kock; the volume under 300 years mathematical scientists such as
review is a revised and augmented version Bernoulli, Euler, Lie, and Cartan have used
of that earlier book. these functors and even their representabil-
Several mathematicians have proposed ity, despite the fact that such representabil-
improvements of the category of finite- ity was inconsistent with the formal founda-
dimensional smooth manifolds. Notably, tions that were available. The issue in all the
K.T. Chen and Alfred Frölicher both con- classical cases is to broaden the definition
tributed interesting proposals to volume of the spaces sufficiently to permit the de-
1174 of the Springer Lecture Notes in Math- sired representabilities, thereby streamlin-
ematics, which records a 1982 meeting at ing constructions and computations with-
Buffalo. Frölicher’s admirably simple cat- out obscuring the spaces and mappings of
egory is the basis for the book The Con- interest.
venient Setting of Global Analysis (AMS, By Yoneda’s lemma, representing objects
1997) by Andreas Kriegl and Peter Michor. are unique, in particular the object T , which
However, that category does not permit the represents the tangent bundle of any space
“representability” of tangent vectors and X as the function space X T ; T is first-order
differential forms. Chen’s category, on the infinitesimal in the sense that the kernel of
other hand, was explicitly designed to allow the bundle projection RT → R for the ring
representability of differential forms, but R of reals consists of quantities of square 0.
does not include representability of tangent Elementary considerations show that the
vectors, nor the representability of smooth natural properties of T are “inconsistent
function spaces. Remarkably, all these rep- with classical logic,” not in any mystical
resentabilities and more do follow from the intuitive way, but just in the sense that
“topos” formulation expounded in the book the lattice of subspaces is not Boolean. It
under review. The practical significance of should be clear from studying this book that
representability is that many constructions it was the development of the geometry that
of a naively set-theoretic character are not led to “logical” considerations such as non-
mere formulas, but correspond to actual Boolean subspace lattices, double-negation
geometric objects, indeed objects with an of points, explicit distinctions between the-
automatic smooth structure. ories and models, etc.; researchers in this
350 BOOK REVIEWS
field did not set out to impose a precon- by the property that T is an ATOM (“amaz-
ceived version of logic onto the subject. ingly tiny objectified motion”), the stronger
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
Zauderer’s text sets itself apart. On first the end of each of the first ten chapters
appearance, because of its tome-like struc- there is a section on Maple methods where
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
ture (930 pages in this new third edition), the author introduces a variety of Maple
it looks more like an intimidating reference procedures to deal with the problems intro-
book, or handbook, than a textbook. But duced in that chapter. The Maple programs
a deeper examination reveals some impor- are available on a web site. Some scientists
tant topics and approaches that are not will be distressed because Maple, rather
common in most PDE books at this level. than MATLAB, is the computer algebra
(See SIAM Review [1] for a comparative system of choice. It is my impression that
review of several PDE texts up to the year many engineers prefer MATLAB for scien-
2000.) tific computation; most of my colleagues use
The book begins with a chapter on ran- MATLAB for this purpose, leaving Maple
dom walks and PDEs. In this context of as an instructional package or as a resource
probability, one visits diffusion, Brown- for symbolic computation. But I agree with
ian motion, the Fokker–Planck equation, the author in stating that it is unwieldy to
Laplace’s equation, Green’s functions, first do both—or even to include Mathematica.
order PDEs, and other hyperbolic equa- Another large improvement over the pre-
tions associated with transition probabil- vious edition is that exercises are now at the
ities. This is the only text I know that end of each section, rather than amassed at
begins this way, and this reviewer finds this the end of chapters. Unlike the second edi-
approach highly satisfying. Therefore, stu- tion, which has the solution to selected exer-
dents are immediately exposed to key ideas cises in the back of the text, the third edition
in probability that lead to the derivation of directs the reader to a web site for solutions.
the standard equations we study in PDEs. The exercises are one of the strengths of this
Later in the text one finds the common book, and there are a number of new exer-
derivations of the various equations, for ex- cises.
ample, the heat equation based on an energy The other material in the text follows the
balance law. presentation in the second edition. To give a
The second major feature of the text is flavor of the basic content, here are the chap-
the coverage of perturbation and asymp- ter topics: random walks and PDEs, first
totic methods for PDEs. This is the most order PDEs, classification and characteris-
distinctive part of the text and it spans tics, initial and boundary value problems
about 170 pages (Chapters 9 and 10). in bounded domains, integral transforms,
The chapter on perturbation methods in- integral relations, Green’s functions, vari-
cludes material on regular and singular per- ational methods and maximum principles,
turbation, boundary layers, and parabolic perturbation and asymptotic methods, and
approximation. The chapter on asymp- numerical methods. It sounds like a list
totic methods includes equations with large of everything a budding engineer, scientist,
parameters, the eikonal and transport equa- or applied mathematician needs to know.
tions, scattering, the propagation of singu- Indeed, it nearly is!
larities and discontinuities, and the asymp- Because of the text’s size, and the fact
totic simplification of PDEs. Here one that our own course is a one-semester, three
will find dissipative wave equations, Burg- credit hour course, I have not adopted Za-
ers’ equation, the Navier–Stokes equations, uderer as a textbook. However, I have of-
KdV, and more. The important techniques ten used it as a reference, particularly for
in these two chapters do not appear in any asymptotics, and as a source of exercises,
other textbook, to my knowledge, with such of which there are many.
accessibility. If you can get beyond the massive struc-
The chief difference between the second ture of the book and you don’t mind sifting
and the new third editions is the inclusion out a selection of topics for your own course,
of topics on the numerical solution of PDEs. then Zauderer would be an excellent choice
Two new, final chapters deal with finite dif- for an upper division or graduate course in
ference methods and finite element meth- PDEs for engineers, scientists, and applied
ods. Complementary to these chapters, at mathematics students. With some careful
352 BOOK REVIEWS
choices, one might be able to use the book in (with most of the pulp still present) that
the standard “Fourier series and boundary is also enjoyable to read. The title comes
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
value problems” course taught to under- from Stillwell’s dual goals of first arguing
graduates at most universities and many his belief that mathematics treads a narrow
smaller colleges. However, the exposition is path between the trivial and the impossi-
at a more mature level than that in some ble, and then illuminating with numerous
of the other standard undergraduate texts; examples just how very narrow that path
so the instructor should be sure of the au- can be. There are minor quibbles with ter-
dience. In summary, this third edition is a minology that one could make (easy enough
noteworthy revision of a popular PDE book, to do with just about any book, I think),
and it should be highly useful to students such as on p. 6 calling 9 + 16 = 25 an equa-
and instructors alike as either a text or a tion rather than an identity. And I remain
reference volume. not quite convinced that Hamilton’s discov-
ery of quaternions is quite as revolutionary
as mathematicians are prone to argue (but
REFERENCE
I’m an engineer, so take that into account!).
[1] J. D. Logan, Featured Review: PDE Those are pretty lame complaints—I admit
Books, Present and Future, SIAM it!—and they are about the “worst” I could
Rev., 42 (2000), pp. 515–522. find.
Yearning for the Impossible succeeds, in
DAVID LOGAN every topic treated, in bringing a fresh eye to
University of Nebraska questions even mathematicians might think
Lincoln have been mined in the past to boring ex-
haustion. Stillwell, however, shows there is
still a lot of gold to be found, if one only
Yearning for the Impossible: The Surpris- thinks about things in a new way. In the
ing Truths of Mathematics. By John Still- discussion of Chapter 1 on irrational num-
well. A K Peters, Wellesley, MA, 2006. $29.95. bers, for example, you’ll be amazed at how
xiv+230 pp., hardcover. ISBN 1-56881-254-X. much Stillwell extracts from beautiful argu-
ments about the square root of two. I think
All too often, when browsing in the math his presentation of the nontermination of
section at my local Barnes & Noble super- the Euclidean algorithm as equivalent to
store, I come across books with provocative irrationality is, as a physicist might put it,
titles that suggest a good read awaits me in “sweet.” And the derivation of the volume
an important branch of mathematics. Alas, of a tetrahedron, in Chapter 4, is just plain
upon opening them up, what I find instead pretty (it provides a “practical” application
is a potpourri of glib pseudohistory, curious of the trick for summing geometric series).
quotes culled from other, earlier books of In that and other chapters he tackles
the same ilk, and, worst of all, a book writ- imaginary numbers, the geometry of how
ten all too obviously by a journalist with things look to the human eye, the infinitesi-
little if any professional training in mathe- mal, the geometry of flat, curved, and peri-
matics. Don’t get me wrong—those authors odic spaces, the fourth dimension, and the
are often good writers, but it soon becomes concept of infinity. In every one of these
clear that they really don’t “understand” treatments Stillwell brings new, unortho-
mathematics. Such books are like orange dox insights to his writing that will stimu-
juice that has had all the pulp strained late readers (from high schoolers to emeri-
out—it all goes down easily enough, sure, tus professors) to think about old topics in
but most of what makes it worth drinking new, nonstale ways. The arguments in the
in the first place has been removed. book are not always easy to follow on first
In his new book, Yearning for the Impos- reading, but are logically correct and, ulti-
sible, Professor Stillwell has achieved what mately, convincing. Yearning for the Impos-
many might well have come to believe to be sible will be a treat for teachers, too, who are
nearly impossible in mathematical exposi- looking for new ways to bring stimulating,
tion for the masses, good nutritional stuff fresh examples into their courses. The book
BOOK REVIEWS 353
is a great read (the publisher, A K Peters, mathematics to the already prominent al-
has done a first-rate job of packaging the gebraic geometry of Castelnouvo, Enriques,
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
book, with an index that is actually useful and Severi. One should not forget either
for finding things) and shows that Professor Levi-Civita, who had correspondence with
Stillwell’s award of the MAA’s 2005 Chau- Einstein that in April 1915 the latter called
venet Prize for expository writing was well outstandingly important for him.
deserved indeed. Of course, in August 1914 World War I
PAUL J. NAHIN broke out. This actually ended the “golden
University of New Hampshire age” of Italian mathematics. After all,
World War I was a turning point for all of
European (and American) society. Noth-
ing remained as it was heretofore. Italian
Italian Mathematics Between the Two mathematicians, by and large, were in fa-
World Wars. By Angelo Guerraggio and vor of Italian intervention on the side of
Pietro Nastasi. Birkhäuser-Verlag, Basel, 2005. the Allies and against the Central Powers
$119.00. x+299 pp., hardcover. ISBN 3-7643- though this did not happen until May 24,
6555-2. 1915. Vito Volterra (1860–1940) was per-
haps the most committed mathematician
This book appears in the Historical Stud- urging Italian intervention on the side of
ies Science Networks series of Birkhäuser- England, France, and Russia. This is clear
Verlag. Since Poincaré in 1908 said that from his correspondence at the time. This is
the Circolo Matématico de Palermo was the perhaps the place to mention that although
most important world mathematical orga- all the letters and commentary in Italian are
nization, the contents of this book should translated in the text, none of the consider-
have considerable interest, even if Italian able number in French are, and while this
mathematics actually came in third place in posed no difficulty to the reviewer, it may
European mathematics in the cited period, well be one for some prospective readers.
after those of Germany and France. Also, Castelnuovo (1865–1952) and Enriques
the works of the Italian school of algebraic (1871–1946) also had the same persuasion
geometry were prominent in this period. as Volterra, though somewhat more muted.
In addition, the names of several Italian Their collaboration, which would culminate
mathematicians are permanently inscribed in a complete classification of algebraic sur-
in textbooks, e.g., Volterra, Betti, Tonelli, faces in 1914, began in 1892. Also, Enriques
Fubini, Vivanti, Vitali, Enriques, and many married Castelnuovo’s sister.
others. Chapter III, titled “Volterra’s Leader-
This book begins (as all good books ship,” is the first truly post–World War
should) “before the beginning” and its first I chapter. Postwar there was the well-
two chapters are devoted to this. The known exclusion under French leadership
“Risorgimento generation” (Betti, Brioschi, of members of the former Central Powers
Casorati, Beltrami, Cremona) had already from international scientific collaboration.
established a name for mathematics done on Volterra was the leader of Italian mathe-
the Italian peninsula by 1870 (when Italian matics during this period until his refusal
unification was completed and Rome be- to take the fascist oath in 1931, though he
came the capital of the newly unified Italy). did not win all “battles,” even within his
Their successors (e.g., Veronese, Corrado own faculty.
Segre, and others) brought algebraic ge- In 1921 a chair became open in the math-
ometry, “italienische Geometrie,” to full ematics faculty in Rome when an analyst
flower. At the turn of the twentieth cen- died. Volterra wanted an analyst to be suc-
tury, though, Volterra was the most promi- ceeded by an analyst. In 1918 Levi-Civita
nent Italian mathematician, in analysis and (1873–1941) had become a faculty mem-
mathematical physics, and mathematicians ber at Rome, after earlier attempts to woo
like Fubini and Vitali had grown beside him. him from Padua had failed. Levi-Civita
These men left their stamp on real analy- was a committed pacifist and socialist and
sis, and were a counterweight in Italian did not agree with Volterra. Among the
354 BOOK REVIEWS
prime candidates were Leonida Tonelli, En- scant attention given to the Congress in this
riques, and Francesco Severi (1879–1961). book. A countermanifesto to Gentile’s was
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
Severi (who had been Enriques’ assistant issued by Benedetto Croce. Croce, who had
earlier in Bologna—the book under review been a friend and collaborator of Gentile,
says he had studied with Enriques, but had earlier broken with him, and was pre-
to this reviewer that is unclear), with the vailed upon to issue this countermanifesto.
support of his friend Levi-Civita, was suc- This was signed by a number of significant
cessful. Enriques also arrived shortly there- mathematicians including Volterra, Castel-
after in Rome, when Castelnuovo volun- nuovo, Levi-Civita, and Severi, though not
tarily resigned the chair that he had held Enriques. Croce became a representative of
for many years. At this time, Severi and the antifascist forces in Italy, though he was
Levi-Civita were both socialists—and the always skeptical of democracy and “mass-
Socialist Party had been opposed to entry man.”
into the war. Levi-Civita’s support of Severi It is Severi who became the leader of Ital-
was based on his belief that he was the best ian mathematics. The former socialist and
mathematician, from all aspects, who was interventionist (like Mussolini) became a
available. For future developments, it may fascist. Severi and Enriques, who had been
be noted here that Levi-Civita, Volterra, coworkers in the first decade of the twenti-
Castelnuovo, and Enriques were all Jew- eth century, developed a very quarrelsome
ish, but Severi was not. During the 1920s relationship that was only formally made up
Volterra was instrumental in the founding in 1929. Though the authors do not men-
of both the Unione Matematica Italiana tion it, Severi was apparently well known
(UMI) and the Consiglio Nazionale delle for his quarrelsomeness. Severi as a politi-
Ricerche (CNR). These would later become cian and mathematician became (at first
something of rivals, though initially the through Gentile’s support—and his appar-
CNR supported the UMI. ent influence on Gentile) “one man alone in
Of course, any discussion of Italy between the lead.” After the war he was threatened
the wars must deal with the rise of fascism. with lynching and became a Catholic who
On October 27, 1922, the Fascist Party an- suggested that mathematics and physics
nounced the “march on Rome”; the King proved God’s existence. Throughout the
refused to sign the previously agreed upon fascist period he was the master of math-
declaration of martial law; on October 29, ematics (though apparently he did not fol-
Mussolini was charged with forming a new low Gentile in support of the “Republic of
government. The leading fascist figure in Salo”).
education was the Sicilian philosopher Gio- One figure not yet mentioned who de-
vanni Gentile. The authors trace the early serves note is Mauro Picone (1885–1977),
mathematical experience of fascism through who considered himself “a black shirt since
him; in fact early on (in 1911) Benedetto the very beginning.” In 1927 he founded
Croce and Gentile had had a philosophical the Istituto Nazionale per le Applicazioni
dispute with Enriques, who was also some- del Calcolo (INAC) that would become the
thing of a philosopher, which, according to envy of some German mathematicians (like
the authors, had a negative effect on math- Gustav Doetsch). While it seems Picone
ematicians wanting active participation in believed in applications, other foreign ob-
the cultural life of Italy. Gentile did not servers (like Ludwig Prandtl) did not feel
join the Fascist Party until May 1923 and that his institute did “applied mathemat-
became the “Philosopher of Fascism.” ics.” In 1929 it became part of the CNR,
In 1925 Gentile issued a fascist manifesto; whose president was Guglielmo Marconi.
the only mathematicians to support it were The authors identify Picone as an applied
Corrado Gini and Salvatore Pincherle. Gini, mathematician, but it is perhaps necessary
of course, is still mentioned in statistics to indicate that this may be disputed.
books for “Gini’s coefficient,” and Pincherle The authors follow Italian mathematics
would be a major figure in organizing the through the insistence on the fascist oath in
International Congress of Bologna in 1928. 1931 and the racial laws of 1938. While Ital-
In fact, the reviewer is surprised by the ian Jews had lived as normal citizens at least
BOOK REVIEWS 355
since 1870 (and often earlier), until this was casional “é” and “con” make me think that
canceled by the racial laws, it is undeniable the book was originally written in Italian,
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
that not only in the distant past (“ghetto” and then “translated” by the authors. Most
is, after all, an Italian word) but also in often these errors are confusing. The lack
the first decades of the twentieth century of care which this book presents extends to
anti-Semitic feelings occasionally revealed the pictures. The same picture of Leonida
themselves in the correspondence of Ital- Tonelli appears on p. 64 and p. 162. Severi’s
ian mathematicians. 1938 was the year for picture (the same one) appears on p. 62 and
anti-Semitic action in other countries as p. 103 and these are both clearly cropped
well, most prominently Kristallnacht. Ital- from the picture of him and Picone (and
ian anti-Semitism did not go as deeply into unnamed others) on p. 261. This careless-
the culture as it did in Germany, Russia, ness begins on the page before the preface
or Poland, but it was real. In any case, it (!) where the cover illustration is identified
was never biological. Severi and Picone were as being of L. E. J. Brouwer (subject of
not the only fascist mathematicians, Enrico another book in the series) when actually
Bomplani was also one, and he in fact wrote it is of three Italian mathematicians: Levi-
a history of Italian mathematics omitting Civita, Enriques, and ? (I do not know).
the names, among others, of Segre, Castel- What Birkhäuser should do (if they wish
nuovo, Volterra, Enriques, and Levi-Civita. to attract English-language readers) is with-
There were other fascist mathematicians as draw the book from circulation, have the
well. French passages translated (by an English
The authors’ book closes with the onset speaker!), have the whole book gone over
of World War II, of which there are brief (by an English-speaking copy editor), and
mentions. then reissue it.
Actually it is not surprising that the 21 S. L. SEGAL
years between the wars is too short a period University of Rochester
for consideration (which is why Volterra so
dominates the early part of the book and
Severi and Enriques its second half). This
book provides much interesting information Perturbations of Positive Semigroups
not otherwise available in English, and the with Applications. By Jacek Banasiak and Luisa
authors have previously written on cognate Arlotti. Springer-Verlag, London, 2006. $89.95.
topics. xiv+438 pp., hardcover. ISBN 1-85233-993-4.
Unfortunately, the book cannot be rec-
ommended to English readers. The reason In 1798, the British economist Thomas
for this is not just the occasional plodding Malthus suggested the linear differential
style of the authors, but the numerous er- equation x = αx as a simple model for
rors (not all minor) in English that litter the growth of populations. Since such a
the book and are sometimes confusing. A model only shows exponential growth or
good proofreader or copy editor would have decay which (in the case of growth) can
caught these many errors, which seem even hardly continue for too long, in 1845 Pierre-
worse in the translations from Italian than Francois Verhulst suggested the so-called
in the original text. Of course, presum- logistic equation x = αx(1 − K x
), which
ably to save money, Birkhäuser (or was incorporates the idea that there is an envi-
it Springer, or perhaps Bertelsmann?) ap- ronmental limit to growth, the carrying ca-
parently did not employ a proofreader or pacity K. Both equations are special cases
copy editor of any sort. Sometimes these of Bernoulli’s equation x = αx − γxθ with
errors are humorous, for example, on p. 57 θ > 1, which is named after Jacob Bernoulli
there appears the phrase “more than 40% of and was solved by Leibniz in 1696, and
[French] students in Mathematics or in Sci- there are many other biologically mean-
ences were killed or blessed [sic] in France ingful nonlinear modifications of x = αx
during the war.” The French word for [30, Chap. 5]. Nevertheless, these two basic
wounded is “blessé” but the authors are here equations are indicative of the two main
citing an English-language author! The oc- different paths dynamical systems theory
356 BOOK REVIEWS
would later take: (linear) operator semi- of (linear) functional analysis is entered,
groups and nonlinear semiflows (dynamical whose powerful methods are responsible for
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
systems). They are also indicative of this re- the special path linear semigroups (evo-
view being written by somebody interested lution systems) have taken as compared
in life science applications. to general semiflows. It is in this realm
A semiflow, whether linear or nonlinear, that the book by Banasiak and Arlotti,
describes the development of a system in from now on called [BA], resides. [BA] also
time according to the following rule: lives in a world without seasons as it re-
Consider three points in time, r < s < t. stricts itself to autonomous linear semi-
If the system starts at x at time r and flows where Φ(t + s, s) = Φ(t, 0) =: S(t).
reaches a state y at the time s and takes y This is not so restrictive as it may ap-
as a starting point at time s and evolves to pear at first glance, as there are procedures
a state z at the time t, then z is the same which transform nonautonomous semiflows
state as the one the system reaches when into autonomous semiflows on a larger state
evolving directly from time r to time t. space. The easiest such transformation is
More formally, if Φ(t, r, x) denotes the Ψ(t, (s, x)) = (t + s, Φ(t + s, s, x)), which
state the system will reach at time t when introduces an autonomous semiflow on the
it starts at state x at time r < t, we have state space I × X. This transformation is
the functional relationship not useful in a linear context as Ψ is not
linear in (s, x) even if the Φ(t, s) are lin-
(1)
ear maps. Even for nonlinear semiflows,
Φ(t, r, x) = Φ(t, s, Φ(s, r, x)), r < s < t, this transformation is of restricted value,
Φ(r, r, x) = x. namely, mainly for so-called asymptoti-
cally autonomous semiflows [22]. Another
To be even more precise, we take a set of
nonlinear transformation leads to skew-
possible states X and a nontrivial interval
product semiflows, where the state space
I. It is then convenient to define the (not
is very complicated as it contains semi-
necessarily everywhere defined) operators
flows, Ψ(t)(x, Φ) = (Φ(t, 0, x), Φt ), where
Φ(t, s)(x) = Φ(t, s, x) for t, s ∈ I, t ≥ s,
Φt is the shifted semiflow Φt (s, r, x) =
and x ∈ X. The semiflow property (1) can
Φ(t + s, t + r, x). In spite of the unwieldy
then conveniently be written as
state space, skew-product semiflows have
(2) been proved useful in nonlinear dynamical
Φ(t, s) ◦ Φ(s, r) = Φ(t, r) r ≤ s ≤ t, systems theory [12]. Again, linearity is not
Φ(s, s) = I r, s, t ∈ I. preserved even if the maps Φ(t, s) are linear.
This is accomplished by so-called (Howland)
Here the first equation includes the equal-
evolution semigroups (see [5] and the refer-
ity of the domains of definition, where for
ences therein), where the new state space
two operators F and G the domain of their
is Lp (I, X), 1 ≤ p < ∞, and
composition is
D(G ◦ F ) = {x ∈ D(F ); F (x) ∈ D(G)}. [S(t)f ](r)
It is still useful to keep the formulation (1) Φ(t, t − r)f (t − r) if t, t − r ∈ I,
= ,
in mind because typically X is a topological 0 otherwise
space. The most fruitful continuity assump- t ≥ 0, r ∈ I.
tion is that Φ is continuous from ∆ × X to
X, where ∆ = {(t, r); r, t ∈ I, r ≤ t} and Before we describe what is specifically done
∆ × X carries the product topology. In in [BA], let us summarize what is not cov-
the Malthus world of unrestricted growth, ered (on purpose because the authors want
X usually is a Banach space and the op- to follow their own interests and not re-
erators Φ(t, s) are everywhere defined on peat topics that are easily available in the
X and linear. The continuity of Φ then literature): spectral theory, large-time be-
equivalently translates to Φ(t, s) being a havior (see the references in [BA, p. xii]
strongly continuous family of bounded lin- and [7, 36]), dual semigroups [4, 6, 7, 8, 33],
ear operators. This means that the realm integrated semigroups [2, 29, 32], analytic
BOOK REVIEWS 357
semigroups [1, 7], applications to Markov (iii) For every x ∈ D(A)+ , there ex-
processes [11, 28, 15], and applications to ists some x∗ ∈ X ∗ , x∗ ≥ 0, with
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
functional differential equations [8]. [BA] x, x∗ = x and (A + B)x, x∗ ≤
has an impressive list of 172 references; I 0.
have tried to make the list of references be-
Then there exists an extension à of A+B,
low as complementary as possible and, in
D(Ã) ⊇ D(A), which generates a positive
order to limit its length, have only repeated
C0 -semigroup T of contractions. The resol-
references in [BA] if necessary to keep this
vent of à satisfies the strong series expan-
review self-contained.
sion,
The focus of [BA] is positive semigroups.
This reflects that the relevant states of most
∞
biological and many physical problems are (λ − Ã)−1 x = (λ − A)−1 (B(λ − A)−1 )n x,
positive. More precisely, the state space X n=0
contains a closed positive convex cone X+ . (3)
x ∈ X, λ > 0.
A semigroup S = {S(t); t ≥ 0} is called
positive if S(t)X+ ⊆ X+ . From a nonlin-
The positivity of the perturbed semi-
ear point of view, a positive semigroup is
group, T , is not explicitly stated in [BA,
a monotone semiflow. Monotone semiflows
Thm. 5.2], and the subsequent material
have a powerful theory of their own [13, 27]
in [BA, sec. 5.1] makes the reader wonder
which, however, has not too much bearing
whether additional assumptions are needed
on positive semigroups because linearity al-
for positivity. However, by (3), Ã is re-
lows the derivation of even farther-reaching
solvent positive, which implies that T is
results [34]. More precisely, as correctly
positive by the exponential formula [BA,
described in the title, [BA] concentrates
(3.22)]. Actually T is the smallest positive
on the perturbation of positive semigroups
semigroup whose generator extends A + B
and its applications to birth and death
[BA, Prop. 5.7].
processes, fragmentation processes (pure or
Condition (iii) is central in applying
with growth or decay), and to kinetic the-
[BA, Thm. 5.2]. An important class of
ory. The book can be divided roughly into
Kantorovich–Banach spaces is that of ab-
three parts. Chapters 2, 3, and 4 describe
stract L spaces, i.e., Banach lattices where
(mostly known) topics from functional anal-
the norm is additive on the positive cone,
ysis and C0 -semigroup theory that will be
x + y = x + y if x, y ∈ X+ . Then
needed for the remaining parts. The central
there exists a unique positive φ ∈ X ∗ such
Chapters 5 and 6 generalize what the au-
that
thors call Kato’s perturbation theorem to
abstract L spaces and develop the theme of
(4) φx = x, φ = x, x ∈ X+ .
honest and dishonest semigroups. The third
part, actually more than half of the book, The following corollary holds (cf. [31]).
consists of applications. Corollary 2. Let X be an abstract L
One of the highlights of the central Chap- space. Let A be the generator of a posi-
ters 5 and 6 is Theorem 5.2, which I give tive C0 -semigroup S of contractions and let
in slightly modified form. In the follow- B : D(A) → X be A-bounded and satisfy
ing, “spr” denotes the spectral radius of a the following assumptions:
bounded linear operator.
Theorem 1. Let X be a Kantorovich– (i) B is positive.
Banach space (i.e., a Banach lattice with a (iii) φ(A + B)x ≤ 0 for all x ∈ D(A)+ .
fully regular cone). Let A be the generator
of a positive C0 -semigroup S of contractions Then there exists an extension à of A+B,
and let B : D(A) → X be A-bounded and D(Ã) ⊇ D(A), which generates a positive
satisfy the following assumptions: C0 -semigroup, T , of contractions and satis-
fies (3).
(i) B is positive: Bx ∈ X+ for all
It is easy to see that condition (iii) is also
x ∈ D(A)+ = X+ ∩ D(A).
necessary. Corollary 2 is stated and proved
(ii) lim sup spr(B(λ − A)−1 ) ≤ 1. in [BA, Cor. 5.17] for the important class
λ→∞
358 BOOK REVIEWS
of abstract L spaces X = L1 (Ω, µ). In this while (ii) =⇒ (iii) (proof of [BA, Thm. 6.8])
case x, φ = Ω xdµ. While every abstract and (iii) =⇒ (i) follow from (3).
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
L space is isomorphic to some L1 (Ω, µ) [25, In [BA, sec. 6.2], Theorem 3 is extended
Thm. II.8.5], this does not help in verifying to strictly substochastic semigroups intro-
(iii) . Another important example of an ab- ducing the concepts of honest and dishonest
stract L space is M(Ω), the Banach space semigroups. While these concepts are cen-
of (signed) measures µ of finite variation on tral to [BA], their discussion may lead too
some σ-algebra of subsets of Ω [15]. In this far for this review. Stochastic semigroups
case µ, φ = µ(Ω). An application of Corol- are an example of honest semigroups, while
lary 2 to M(Ω) can be found in [31, sec. 5]. the substochastic semigroup T in Theorem 3
Returning to Theorem 1, we cannot ex- is a dishonest semigroup if T is not stochas-
pect à = A + B unless lim supλ→∞ spr tic in spite of φ(A + B) = 0.
B(λ − A)−1 < 1, for, in an abstract L Notice that condition (iii) in Theorem 3 is
space, A + B is resolvent positive if and the only one which is stated in terms of the
only if lim supλ→∞ spr B(λ − A)−1 < 1 [35] original ingredients A and B. Differently
[BA, Thm. 5.10]. This raises the possibil- from what one might expect, (iii) does not
ity that at least à = A + B if lim supλ→∞ seem to play any major role in discussing
spr B(λ − A)−1 = 1. As it turns out, this the honesty of the semigroups which arise
does not necessarily hold even in the sce- in the many applications that are discussed
nario of Corollary 2. In order to show that in Chapters 7 to 11. Rather the authors use
this question is not only of abstract inter- extension techniques which are developed in
est, let us introduce some more terminology section 6.3 and rely on the special structure
[BA, sec. 6.1]. A positive C0 -semigroup of of X = L1 (Ω, µ). The applications cover
contractions is called a substochastic semi- some 240 pages and present a dazzling mul-
group. A positive C0 -semigroup T is called titude of problems: birth and death pro-
a stochastic semigroup if T (t)x = x cesses (though not the general Kolmogorov
for all x ∈ X+ . If X = M(Ω), a positive differential equations; see [14, 31, 18] and
semigroup is stochastic if and only if it pre- the references therein), pure fragmentation
serves the property of being a probability processes, fragmentation with growth and
measure, µ ≥ 0, µ(Ω) = 1, obviously an im- decay, kinetic theory (Maxwell–Bolzmann
portant feature for stochastic processes. A and semiconductor equations), as well as
stochastic semigroup on an abstract L space elastic and inelastic collision models. It is
is called strictly stochastic if φ(A + B) = 0. worth mentioning that birth and death pro-
Notice that strict stochasticity is stronger cesses are not the only possible applications
than being substochastic and not stochas- in mathematical biology. Nonlinear versions
tic. As it turns out, Ã = A + B is equiv- of the Kolmogorov differential equations
alent to the stochasticity of the semigroup have been suggested to model prion prolif-
generated by Ã. eration (which supposedly plays a role in
Theorem 3. Let X be an abstract L transmitted spongiform encephalopathies
space. Let A be the generator of a pos- like mad cow disease) [20, 23] as well as
itive semigroup S of contractions and let spatially implicit metapopulation models
B : D(A) → X be A-bounded and positive. and host-macroparasite diseases [17, 19].
Assume that φ(A + B) = 0. Then the fol- Fragmentation models with growth also ap-
lowing are equivalent for the extension à of ply to size-structured cell division models
A + B and the substochastic semigroup T [3, 9, 17, 21, 36] and again to prion prolif-
generated by à according to Corollary 2: eration models [10, 16, 24, 26].
(i) Ã is the closure of A + B. Obviously [BA] should be in the hand of
(ii) T is a stochastic semigroup. every researcher who is interested in any
of these applications. It also seems to be
(iii) (B(λ − A)−1 )n → 0 strongly as
the only book available that deals with rel-
n → ∞.
atively bounded positive perturbations of
Indeed, (i) =⇒ (ii) follows from the def- positive semigroups and with general ques-
inition of the closure and standard semi- tions of their stochasticity and honesty.
group theory (proof of [BA, Thm. 6.13]),
BOOK REVIEWS 359
[1] H. Amann, Linear and Quasilinear [15] T. Lant and H. R. Thieme, Markov tran-
Parabolic Problems. Volume I. Ab- sition functions and semigroups of
stract Linear Theory, Birkhäuser, measures, Semigroup Forum, http://
Basel, 1995. dx.doi.org/10.1007/s00233-006-0636-
[2] W. Arendt, C. J. K. Batty, M. Hieber, z.
and F. Neubrander, Vector-valued [16] P. Laurençot and C. Walker, Well-
Laplace Transforms and Cauchy posedness for a model of prion pro-
Problems, Birkhäuser, Basel, 2001. liferation dynamics, J. Evol. Equ., to
[3] O. Arino, A survey of structured cell appear.
population dynamics, Acta Biotheo- [17] P. Magal and S. Ruan, eds., Struc-
retica, 43 (1995), pp. 3–25. tured Population Models in Biology
[4] P. L. Butzer and H. Berens, Semi- and Epidemiology, Springer, to ap-
groups of Operators and Approxi- pear.
mation, Springer, Berlin, Heidelberg, [18] M. Martcheva, H. R. Thieme, and T.
1967. Dhirasakdanon, Kolmogorov’s dif-
[5] C. Chicone and Y. Latushkin, Evolution ferential equations and positive semi-
Semigroups in Dynamical Systems groups on first moment sequence
and Differential Equations, AMS, spaces, J. Math. Biol., 53 (2006), pp.
Providence, RI, 1999. 642–671.
[6] Ph. Clément, O. Diekmann, M. Gyl- [19] M. Martcheva and H. R. Thieme, In-
lenberg, H. J. A. M. Heijmans, finite ODE systems modeling size-
and H. R. Thieme, A Hille-Yosida structured metapopulations, macro-
theorem for a class of weakly contin- parasitic diseases, and prion pro-
uous semigroups, Semigroup Forum, liferation, in Structured Population
38 (1989), pp. 157–177. Models in Biology and Epidemiology,
Springer, to appear.
[7] Ph. Clément, H. J. A. M. Heijmans,
[20] J. Masel, V. A. A. Jansen, and M. A.
S. Angenent, C. J. van Duijn, and
Nowak, Quantifying the kinetic pa-
B. de Pagter, One-Parameter Semi-
rameters of prion replication, Bio-
groups, North–Holland, Amsterdam,
phys. Chemistry, 77 (1999), pp. 139–
1987.
152.
[8] O. Diekmann, S. A. van Gils, S. M.
[21] J. A. J. Metz and O. Diekmann, The
Verduyn Lunel, and H.-O. Wal-
Dynamics of Physiologically Struc-
ther, Delay Equations: Functional-,
tured Populations, Lecture Notes in
Complex-, and Nonlinear Analysis,
Biomath. 68, Springer, Berlin, Hei-
Springer, New York, 1995.
delberg, 1986.
[9] O. Diekmann, H. J. A. M. Heijmans,
[22] K. Mischaikow, H. L. Smith, and
and H. R. Thieme, On the stability
H. R. Thieme, Asymptotically au-
of the cell size distribution, J. Math.
tonomous semiflows: Chain recur-
Biology, 19 (1984), pp. 227–248.
rence and Lyapunov functions, Trans.
[10] H. Engler, J. Prüss, and G. F. Webb, Amer. Math. Soc., 347 (1995), pp.
Analysis of a model for the dynamics 1669–1685.
of prions II, J. Math. Anal. Appl., [23] M. A. Nowak, D. C. Krakauer, A.
324 (2006), pp. 98–117. Klug, and R. M. May, Prion in-
[11] W. Feller, An Introduction to Probabil- fection dynamics, Integrative Biol., 1
ity Theory and Its Applications, Vol. (1998), pp. 3–15.
II, Wiley, New York, 1965. [24] J. Prüss, L. Pujo-Menjouet, G. F.
[12] J. K. Hale, Asymptotic Behavior of Dis- Webb, and R. Zacher, Analysis of
sipative Systems, AMS, Providence, a model for the dynamics of prions,
RI, 1988. Discrete Contin. Dyn. Syst. Ser. B, 6
[13] M. W. Hirsch and H. L. Smith, Mono- (2006), pp. 215–225.
tone dynamical systems, in Hand- [25] H. H. Schaefer, Banach Lattices and
book of Differential Equations, Vol. 2, Positive Operators, Springer, Berlin,
A. Cañada, P. Drábek, and A. Fonda, Heidelberg, 1974.
eds., Elsevier, Amsterdam, 2005, pp. [26] G. Simonett and C. Walker, On the
239–357. solvability of a mathematical model
[14] T. Kato, On the semi-groups generated by for prion proliferation, J. Math. Anal.
Kolmogoroff ’s differential equations, Appl., 324 (2006), pp. 580–603.
360 BOOK REVIEWS
[27] H. L. Smith, Monotone Dynamical Sys- book is in the long tradition of study of
tems: An Introduction to the Theory extreme value theory in the Netherlands,
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
of which (49) are connected with Chapters life spans of people born in the Netherlands
1–3. The bibliography lists 105 papers and during the period (1877–1881) and still alive
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
books, 24 of which are authored or coau- on January 1, 1971 (with 10391 points).
thored by de Haan. (This should be com- These case studies are developed further
pared with the approximately 370 entries and used to illustrate the theory in Chap-
in [1] and 646 in [5].) ters 3 and 4 in a total of about 14 pages.
Two main themes are given full treatment Unfortunately the data sets are apparently
here for the first time in textbook form. not publicly available, so this precludes nu-
The first of these is the progression from merical verification of the results obtained
univariate extreme value theory to multi- or comparison with other methods. While
variate extreme value theory, and further to this material makes a start at delivering on
function-valued extreme value theory. This the authors’ stated goal of giving a “thor-
is a valuable feature of this book, since it is ough account of the basic theory of extreme
an area in which considerable research and values. . . statistical. . . and applied,” I was
progress have been made over the past 25 left somewhat disappointed that more had
years. Multivariate extreme value theory is not been done in this direction. There is no
also treated in Chapters 8 and 9 of the re- treatment of profile likelihood methods for
cent book [1], which also makes an effort to finding confidence bounds (see, e.g., Chap-
cover both the probabilistic and statistical ters 5 and 7 of [1]), essentially no treatment
aspects of the theory. It is not covered in of data dependent choices of k for estimation
[5], but it is a topic that is of current in- of the extreme value index (see, e.g., [4], sec-
terest in connection with a wide range of tion 4.7 of [1], and [7] for a different method
applications, including environmental mon- involving kernel smoothing), no treatment
itoring and financial statistics; see, e.g., [8] of extreme value theory in the context of
and [9]. time series models (see, e.g., Chapter 10 of
The second theme is the systematic use [1] or Chapter 7 of [5]), and no discussion
of the theory of “extended regular varia- of regression analysis of extreme values (see
tion,” due in large part to the work of Chapter 7 of [1]). Although a reader or stu-
the first author, and the associated theory dent will get some taste of the statistical
of “second order regular variation.” The theory of extreme values from this book, for
mathematics of these classes of functions a thorough treatment of statistical methods
is developed in Appendix B and used sys- for extreme values and more complete treat-
tematically throughout the text. In this ment of applications, the reader will need
connection it is somewhat curious that the turn to other books (for example, to [1] and
authors do not reference the 1987 book [2], [5]) and the still growing research literature.
but this might be partly explained by the This book is a “must have” for experts
fact that [6], which also appeared in 1987, in extreme value theory, and would be an
was also not referenced in [2]. Both [2] and excellent choice for a special topics course
[6] are included in the comprehensive bibli- on the probabilistic and theoretical sides of
ography of [5]. extreme value theory, especially a course
The authors do make an effort to treat emphasizing the multivariate aspects of ex-
the statistical and applications sides of ex- tremes. For thorough treatments of the sta-
treme value theory. They devote Chapters 3 tistical aspects and applications the reader
and 4 to statistical aspects of univariate ex- will need to look elsewhere.
treme value theory, and Chapters 7 and 8 to
statistical aspects of multivariate extreme
REFERENCES
value theory. Moreover, they treat three
sets of data as “case studies”: the first in-
[1] J. Beirlant, Y. Goegebeur, J. Segers,
volves sea levels at Delfzijl on the northeast
and J. Teugels, Statistics of Ex-
coast of the Netherlands (with a univari- tremes, Wiley Series in Probability
ate data set consisting of 1873 points); the and Statistics, John Wiley, Chiches-
second involves the S&P total return index ter, UK, 2004.
over the period January 1, 1980 to May [2] N. H. Bingham, C. M. Goldie, and
14, 2002 (a univariate data set consisting of J. L. Teugels, Regular Variation,
5835 points); while the third concerns the Encyclopedia Math. Appl. 27, Cam-
362 BOOK REVIEWS
bridge University Press, Cambridge, which has been observed to take many in-
UK, 1987. tricate forms. The richness of its dynamics,
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
[3] L. de Haan, Fighting the arch-enemy with patterns, and instabilities has made RBC a
mathematics, Statist. Neerlandica, 44 paradigm for pattern formation. It has been
(1990), pp. 45–68. an important testbed for studies of stability,
[4] G. Draisma, L. de Haan, L. Peng, and bifurcations, and weakly nonlinear analysis.
T. T. Pereira, A bootstrap-based
Large numbers of research articles and sev-
method to achieve optimality in esti-
mating the extreme-value index, Ex- eral books provide excellent introductions
tremes, 2 (1999), pp. 367–404. to the nonlinear dynamics of such systems,
[5] P. Embrechts, C. Klüppelberg, and for example, [1].
T. Mikosch, Modelling Extremal The current book, however, is not an
Events, Appl. Math. (N.Y.) 33, introduction; rather, it is a systematic
Springer-Verlag, Berlin, 1997. overview of the dynamics and instabilities
[6] J. L. Geluk and L. de Haan, Regu- that occur in more realistic and inevitably
lar Variation, Extensions and Taube- more complicated forms of RBC. The first
rian Theorems, CWI Tract 40, Stich- chapter is a brisk fifteen pages that gives
ting Mathematisch Centrum Cen-
the educated reader a good cross-section of
trum voor Wiskunde en Informatica,
Amsterdam, The Netherlands, 1987. references to previous works in the course of
[7] P. Groeneboom, H. P. Lopuhaä, and
introducing the governing equations. The
P. P. de Wolf, Kernel-type estima- beginning of the hierarchy of models ex-
tors for the extreme value index, Ann. tending RBC is Bénard convection, where
Statist., 31 (2003), pp. 1956–1995. the top of the fluid layer is a moving bound-
[8] N. D. Le and J. V. Zidek, Statistical ary to be determined as part of the solution.
Analysis of Environmental Space- If stresses created by temperature-induced
Time Processes, Springer Ser. Sta- variations in the surface tension along this
tist., Springer-Verlag, New York,
interface are included, then the system
2006.
is called the Bénard–Marangoni problem.
[9] Y. Malevergne and D. Sornette,
Classic one-layer models neglect influences
Extreme Financial Risks, Springer-
Verlag, Berlin, 2006. of gas above the fluid layer on the dynamics
[10] Wikipedia, Hurricane Katrina, http://en.
of the interface. In contrast, in two-layer
wikipedia.org/wiki/HurricaneKatrina and multilayer models the motion of in-
(2006). terfaces separating immiscible layers of flu-
ids is determined by balancing forces from
JON A. WELLNER the adjacent layers. One-layer models with
University of Washington their competing Marangoni (thermocapil-
lary) and Rayleigh (buoyancy) instability
mechanisms are described on pages 4–10.
Interfacial Convection in Multilayer Sys- Page 10 and the following 230 pages give
tems. By A. Nepomnyashchy, I. Simanovskii, and the authors’ intensive examination of insta-
J. C. Legros. Springer-Verlag, New York, 2006. bilities in multilayer convection problems.
$79.95. xii+306 pp., hardcover. ISBN 0-387- The opening sections in Chapter 2 begin
22194-8. with a linear analysis of Rayleigh instabili-
ties in the basic two-layer model. Successive
This monograph presents the current state sections incorporate more effects and cor-
of the art in the understanding of a class of responding instabilities into the model. In
problems in the study of Rayleigh–Bénard the simplest cases, where the interface is
convection (RBC). RBC has been the focus assumed fixed, stability curves are given
of extensive research in physics and math- for many different parameter ranges and
ematics for at least the past five decades. descriptions of the notable instabilities (in-
In its simplest form, RBC describes the be- cluding oscillatory instabilities and anticon-
havior of a uniform layer of fluid which is vection) are accompanied by figures illus-
heated from below. Instabilities stemming trating the streamlines and isotherms in
from the thermal expansion of the fluid each situation. When Marangoni effects,
can drive convective motion in the fluid, deformable interfaces, and surfactants are
BOOK REVIEWS 363
included, full results have not been worked for all researchers seeking to understand
out in all cases, but the authors provide a the nature and complexities of this class of
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
work [2], but it is clear that it represents a E sup Xt = sup E sup Xt | F ⊆ T, F finite .
significant advance with more depth of anal- t∈T t∈F
ysis and a greatly extended set of models The main theme of the book under review
given in a systematic presentation. While is a delicate but powerful procedure created
this book is part of Springer’s Monographs by the author in order to simplify the proof
in Mathematics series, I believe it will be that he obtained in [7] of a long-standing
warmly welcomed by physicists and engi- conjecture of Fernique [4].
neers. It is a rich and thorough guide to Assume that (Xt )t∈T is a centered Gaus-
the diverse and complex physical behaviors sian process, and let d : T × T → R
that occur in RBC problems. These prob- be a distance on T defined by d(s, t) =
lems are of interest to physicists and applied (E((Xs − Xt )2 ))1/2 for every s ∈ T and
mathematicians alike, though readers may t ∈ T.
be hard-pressed to try to directly extend A very old problem that has its origin
the results to other problems as most of the in Kolmogorov’s work is to find a strictly
analysis is done very specifically for RBC. positive function Γ(T, d) of the metric space
This book will serve as a valuable reference (T, d) (that is, a quantity that depends only
364 BOOK REVIEWS
on (T, d)) which has the property that there sequence of partitions of T , and using these
exists a strictly positive universal constant partitions to define Γ(T, d) (for the inequal-
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
under review. Actually, I agreed to review This biography of Karl Pearson, a key fig-
the work, even though my research interests ure in the development of modern statis-
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
are very far from the results discussed in tics, is much more about the person than
the monograph, exactly because I want to about his statistical accomplishments. Of
emphasize the fact that the book is of in- its ten chapters, only Chapters 8 and 9
terest to a larger audience, and not only to contain more than brief mention of his sta-
people who publish in the areas presented tistical work, and a reader seeking details
in the work. regarding his statistical contributions will
As we all know, there is a perception find very few even in the chapters covering
in the mathematical community (especially his career as a statistician. This reflects the
among the people who have written at least author’s intent. The author convinced me
one book) that one cannot find flawless that Pearson’s life was presented as Pear-
monographs in mathematics. In the work son saw it: a unified story, a life lived whole
under review, the author refers repeatedly that couldn’t be separated into “student
to an appendix that is assumed to be part of philosophy and history,” “poet,” “pro-
of the book, and the bad news is that there fessor of statistics,” “husband and father,”
is no appendix in the volume. However, etc. By the end of the book, I felt that I
the good news is that the appendix can knew this complex person well. In the next
be downloaded at http://www.math.ohio- few paragraphs, I summarize what I learned
state.edu/˜talagran/appendix.pdf. about Pearson from the book to give read-
ers a sense of his complexity; references are
to pages in the book. This summary is no
REFERENCES
substitute for reading the book itself!
[1] M. Ajtai, J. Komlós, and G. Tusnády,
Pearson was born in 1857 to an ambi-
On optimal matchings, Combinatorica, tious lawyer father and a sensitive mother.
4 (1984), pp. 259–264. The book contains little about his child-
[2] J. Bourgain, Bounded orthogonal sets hood; he was “plagued by delicate health,”
and the Λp -problem, Acta Math., 162 “earnest and highly sensitive” (p. 20). As a
(1989), pp. 227–246. teenager he was tutored at Cambridge, with
[3] R. M. Dudley, The sizes of compact sub- an emphasis on mathematics, but entered
sets of Hilbert space and continuity of King’s College (Cambridge) where human-
Gaussian processes, J. Funct. Anal., 1 istic studies were prominent and mathemat-
(1967), pp. 290–330. ical offerings limited. He read poetry and
[4] X. Fernique, Regularité des trajectoires philosophy and was impressed by Goethe,
des fonctions aléatoires gaussiennes, Rousseau, Spinoza, and Hegel. However, he
Lecture Notes in Math. 480, Springer,
Berlin, 1975, pp. 1–96.
also studied mathematics and physics, scor-
ing third in the Mathematical Tripos, a key
[5] B. Klartag and S. Mendelson, Empiri-
cal processes and random projections, exam. He was “almost despondent about
J. Funct. Anal., 225 (2005), pp. 229– the result” (p. 41), though his tutors and
245. others assured him that the top three were
[6] M. Ossiander, A central limit theorem un- close, and he would have scored higher if
der metric entropy with L2 bracketing, he had not wisely decided to pursue a more
Ann. Probab., 15 (1987), pp. 897–919. general course of study instead of focusing
[7] M. Talagrand, Regularity of Gaussian only on the Tripos. At this point in his life,
processes, Acta Math., 159 (1987), pp. he was conflicted about philosophy, reli-
99–149. gion (moving from Christianity towards ag-
nosticism), and career (law? mathematical
RADU ZAHAROPOL physics?). He spent a year studying in Ger-
Mathematical Reviews many in hopes of resolving these conflicts.
His efforts after returning from Germany,
still in his twenties, focused on philosophy,
Karl Pearson: The Scientific Life in a Sta- literature, history, and politics. Although a
tistical Age. By Theodore M. Porter. Princeton restrained person, he embraced enthusiastic
University Press, Princeton, NJ, 2004. $35.00. fervor (channeled by reason) in his intellec-
viii+342 pp., hardcover. ISBN 0-691-11445-5.
366 BOOK REVIEWS
tual life. At age 23, he wrote a fictionalized ories both physical and metaphysical. In
autobiography, a romantic novel called The 1892, at age 35, he published a book, The
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
New Werther. Pearson’s poetic bent, love of Grammar of Science, examining the foun-
nature, and respect for Goethe as not just a dations of science. Perceptions and con-
writer but a moral teacher are all evident in ceptions were the basic elements of sci-
this novel. But critics were not impressed. ence. Experience provided perceptions, but
Pearson became a socialist, embracing so- the human mind produced the phenomena
cialism based on science and reading Marx and laws of nature through which these
with admiration. Marx declined an offer perceptions were explained. The scientific
from Pearson to translate Das Kapital into method involved the classification of facts
English. Pearson was attracted to Catholi- and search for generalizations, not only in
cism during this period but remained reli- physics but in any investigation, e.g., of
giously conflicted. He wrote “a passion play social or economic questions. He soon re-
for the nineteenth century” (p. 69; like his alized that probability and mathematical
novel, a critical failure) but also studied for statistics provided a model of rigorous rea-
the bar exam (though he soon decided he soning that could be applied to this search
hated law). His research interests during for generalizations, and he committed him-
this period were divided between German self to the program of quantitative research
history and mathematical physics, and he that would occupy him for the rest of his
was under consideration for a position as life. I invite you to read about that part of
lecturer in German literature and history his life in the book that is the subject of
at Cambridge when he was offered, and ac- this review.
cepted, a permanent appointment as pro- What I have failed to capture in the
fessor of applied mathematics at University above summary is captured very well in
College London in 1884. However, he con- Porter’s book. Pearson’s temperament and
tinued to publish historical essays and gave self-image determined the course of his pro-
extension lectures in cultural history. He fessional and personal life and made it a
also “lectured to workingmen’s associations unified story. It was a story left to Porter
and to such groups as the Sunday Lec- to tell. Pearson wrote Galton’s life in three
ture Society” (p. 109). He knew and cor- volumes but never wrote his own. Although
responded with George Bernard Shaw and he kept journals, wrote many letters, and
other Fabian socialists and admired, though even wrote an autobiographical novel at a
never met, William Morris. However, “he young age, he never wrote an autobiogra-
was too strong-willed and idiosyncratic to phy. Pearson “insisted on the integrity and
join anyone else’s movement” (p. 109). coherence of his career and vision” (p. 314)
Pearson’s “interest in women’s issues and and hoped to be remembered for it. Yet
curiosity about women’s feelings and per- he was the founder of an approach that
ceptions” led him to take the initiative in provided a routine, impersonal method for
establishing a “Men and Women’s Club” arriving at scientific conclusions. So it is
in 1885. The club was to be “devoted to not surprising that a fear he expressed to-
research and frank discussion on issues con- ward the end of his life has been realized.
cerning women and relations between the His name is associated not with a person
sexes” (p. 125). By 1889, when the club but with a correlation coefficient and other
dissolved, Pearson was better known for particular statistics and statistical meth-
his writings on the history and status of ods. Hopefully Porter’s book changes that
women, which he considered his most im- for his readers.
portant area of research, than for his math- Perhaps I enjoyed the book particularly
ematics. In spite of his attempt to make because I also studied philosophy before I
the club a place for purely intellectual dis- studied statistics but was always interested
cussions between men and women, he fell in science. As was, I believe, the case for
in love with Maria Sharpe, secretary of the Pearson, one of the appeals of a career in
club, and married her in 1890. statistics for me was that, as a statistician,
However, he continued to apply his math- one can be involved in contributing to the
ematical skills to physics, developing the- solutions of diverse scientific problems. My
BOOK REVIEWS 367
philosophical, religious, and political back- reading the story. There is a certain amount
ground made Porter’s book easier reading of jumping back and forth between different
Downloaded 06/18/17 to 161.130.252.114. Redistribution subject to SIAM license or copyright; see http://www.siam.org/journals/ojsa.php
for me than it may be for some who are parts of Pearson’s life in many chapters, but
less familiar with his historical references. I did not find that problematic. It was just
However, the book contains many footnotes part of Porter’s method of telling the story.
and a long bibliography, so the reader can
JUDITH ZEH
find background information while or after
University of Washington