190 BOOK REVIEWS

chemical reactor. In Case Study II, they discuss the Hopf bifurcation appearing in the
clamped Hodgkin-Huxley equations that form a system of nonlinear ODEs model-
ling electrical activity in the giant axon of a squid under certain controlled
experimental conditions. In Case Study III, devoted to the buckling of rectangular
plate, the authors show how singularity theory can contribute to the understanding of
certain experimentally observed phenomena,
The prerequisites for reading this book are advanced calculus, linear algebra and
elementary aspects of the theory of ODEs, commutative algebra, group theory, and
functional analysis. However, except for linear algebra and advanced calculus, all
relevant ideas are explained in the text or in the appendices. More difficult, technical
and longer proofs are postponed to the end of each chapter. Most paragraphs are
accompanied by exercises which complete the text.
The authors have tried to write the book in such a way that the reader who is not a
mathematician could understand the essentials of the theory and learn how to apply
it. So I can recommend it as an excellent example of how to connect deep theory with
nontrivial applications.

References
1. Arnold, V. I.: Sinoularity Theory, London Mathematical Society Lecture Notes Series, 53, Cambridge
University Press, Cambridge.
2. Mather, J. N.: Stability ofC ~°mappings, IlL Finitely determined map germs, Publ. Math. IHES 35, 127-
156.
3. Thom, R.: Stabilit~ structurelle et morphog~n~se, W. A. Benjamin, 1972.

Institute of Mathematics, MARTIN (3ADEK

Czechoslovak Academy of Sciences,
662 95 Brno, Czechoslovakia

Philippe G. Ciarlet: Mathematical Elasticity, Volume I: Three-Dimensional Elasticity,

Studies in Mathematics and its Applications 20, North-Holland, Amsterdam, 1988,
xlii + 452 pp.

Mathematical elasticity (also called finite or nonlinear elasticity) has made significant
progress during the last decade. The general approach has changed and now the basic
results are achieved using various distinct branches of mathematics such as matrix
theory, topology, and functional analysis.
Contemporary mathematical elasticity follows the program proclaimed by C.
Truesdell characterized by an axiomatic approach, generality and objectivity. The
material is considered as being as general as possible within the definition of elasticity.
The exact steps in the theory are strictly distinguished from steps where some

Acta Applicandae Mathematicae 18 (1990)

BOOK REVIEWS 191

simplifications were adopted. The simplifications are derived by using general

principles to exclude subjectivity of the author.
The book under review can serve as a thorough introduction to contemporary
research as well as a working textbook at graduate level.
The headings of the first reviewed volume are the following:
Part A. Description of three-dimensional elasticity
1. Geometrical and other preliminaries
2. The equations of equilibrium and the principle of virtual work
3. Elastic materials and their constitutive equations
4. Hyperelasticity
5. The boundary value problems of three-dimensional elasticity
Part B. Mathematical methods in three-dimensional elasticity
6. Existence theory based on the implicit function theorem
7. Existence theory based on the minimization of the energy.
The plan for the second volume subtitled Lower-Dimensional Theories of Plates and
Rods is as follows:
Part C. Two-dimensional plate models
8. Justification of the von K~irmhn equations, and other two-dimensional plate
models, from three-dimensional elasticity
9. Mathematical analysis of the von Khrmhn equations
10. Convergence analysis in linearized plate theory
11. Junctions between plates and three-dimensional elastic structures
12. Folded plates
13. Plates with rapidly varying thickness
Part D. One-dimensional rod models
14. Justification of one-dimensional rod models from three-dimensional elasticity
15. Convergence analysis in linearized rod theory
16. Junctions between plates and rods.
The work deals with static problems. According to the author, the state of present
mathematical elasticity does not allow us to write a 'dynamical' counterpart of the
book.

Part A: Description of Three-Dimensional Elasticity

Deformation of a body is described by a mapping tp, It is a bijection between a domain
f~ c R 3 reference configuration (the volume occupied by the body before it is
deformed) and a deformed configuration ~p(f~)- flu volume occupied by the deformed
body. The matrix function F = V~p,called the deformation gradient, characterizes all
geometrical properties: changes of volumes, surfaces and lengths are respectively
governed by scalar det F, matrix C o f F (3 × 3 matrix of subdeterminants of order 2)
192 B O O K REVIEWS

and the right Cauchy-Green strain tensor C = FTF. The Green-St Venant strain
tensor E = ~ ( C - I) measures the deviation between deformation q~ and rigid
deformation, where C = I. Neglecting the second-order terms in E, we obtain the
small strain tensor e of linear elasticity.
The stress state in the body is described by the Cauchy stress tensor T. In static, it
satisfies well-known equilibrium equations. In the case of a mixed problem for a body
subjected to volume forces f and surface tractions g, the equation admits the form
-divOT ~ = f ~ in f~, T"n ~ = g'P on F'~, (1)

where the superscript ~0 indicates that the quantity is taken in the deformed
configuration. In contrast to linear elasticity, nonlinear elasticity strictly distinguishes
the quantities in reference or deformed configuration. Since the deformation q~ is
unknown, Equation (1) should be transformed into a reference configuration. The
problem is solved by the Piola transform of the stress tensor
T(x) = T~'(x*)CofV~o(x), (2)
which takes into account the effect of a change of volume element dx, area element da,
and normal vector n. Thus, the equilibrium equation admits the following form

-divT=f inf,, Tn=g onF1, (3)

where f, g are related to f~, g~ by the formulas f dx = f ~ dx ~, g d a = g~ da ~. Besides

the first Piola-Kirchhoff stress tensor T, the symmetric second Piola-Kirchhoff stress
tensor E = (Vtp)-1T is introduced.
Constitutive equations give the relation between deformation and stress. The
material is said to be elastic if the stress tensor depends on x and the deformation
gradient F = V~0 only:

T(x) = T(x, F(x)) or Y~(x)= ~,(x, F(x)), (4)

where 7" and ,~ are response functions characterizing the properties of the elastic
material. The response functions are considered as being as general as possible. They
are restricted by the axiom of material frame-indifference. This implies that 1~ depends
only on the strain tensor C, C = FrF:

]~(x, F) = ~,(x, C). (5)

According to the symmetries of the material, the response function ~ can be further
reduced. In the case of isotropic material, we have
~.(x, C) = yo(X, ~c)l + el(x, ~c)C + ~'2(x, ~c)C 2,

where )'i are real-valued functions of the principal invariants tl = tr C, t2 = tr Cof C,

13 = det C. Near the stress-free state, in the case of homogeneous material, we have

~.(C) = ).(tr E)I + 2#E + o(llEII),

where E = ~(C - I) and 2, # are the Lain6 constants.
BOOK REVIEWS 193

An elastic material is called hyperelastic if there exists a stored energy function W

such that

7"(x, F) = ~ff- (x, F). (6)

If, in addition, the applied forces are conservative, then the problem admits the
variational formulation:
Find the admissible deformation ¢: fl ~ R a, det V~O > 0 minimizing the total
energy functional

,(@)=ftaI~g(.,Vf)dx-fnfqgdx-fryq~da. (7,

The axiom of frame indifference yields l~(x, F) = if(x, C). In the case of isotropic
material, it can be further reduced. A natural requirement for stored energy function
W is
14"(', F) -o o% for det F ~ 0 +, (8)
which is not satisfied in linear elasticity. The lack of convexity of function if(x, F) with
respect to F is the root of existence problems.
In the formulation of boundary-value problems, two boundary conditions are
usually considered:

(i) condition of place: ~0 = ~Oo on Fo, and

(ii) condition of traction: 7"(', V~o)n= ~(., V~0) on F1.

In practice, other boundary conditions are also used, i.e. nonlocal boundary
conditions and the unilateral condition with or without friction. Deformation ~o
should satisfy the injectivity condition, thus ensuring noninterpenetration of matter.
The lack of uniqueness of the problems is illustrated by physical examples.

Part B: Mathematical Methods in Three-Dimensional Elasticity

The question of the existence of solutions to boundary-value problems in mathemat-
ical elasticity can be approached in two ways: the first consists of a clever way of
applying the implicit function theorem and the second is based on the minimization of
the energy. We give the basic ideas of both methods.
The first approach starts with an equilibrium equation in the form
-div{(I + Vu)E(E(u))} = f in f~, (9)

where u(x)= tp(x)- x is the displacement vector. Considering the homogeneous

Dirichlet boundary conditions, the problem possesses the trivial solution u -- 0 for
f = 0. The left-hand side of Equation (9) represents a nonlinear operator Au. Under
mild smoothness assumptions on the response function Y,, it maps
VP(~q) = {u ~ W2'~(D), u = 0 on F}
194 BOOK REVIEWS

into LP(f~) for p > 3. The proof uses the fact that the Sobolev space WLP(fl) is the
algebra for p > 3. Since
A'(O)u=f inf,, u=0 onF,
is a boundary-value problem of linearized elasticity having a unique solution, one can
prove that the derivative A'(0) is an isomorphism between spaces VP(fl) and LP(f~).
Using the implicit function theorem in the neighbourhood of the origin in
VP(f~) x LP(f~), we obtain the local existence of the solution for small forces f. For a
given f, we define a loading path f(2) = 2f, 2 ~ [0, 1] and denote the corresponding
solution by ~(2). Using the implicit function theorem under the assumption that
A'(~(2)) is an isomorphism for 2 ~ [0, 1], we can reach the existence of the solution
u = ~(1). The discrete counterpart of the method yields an approximate method. For a
partition 0 = 20 < 21 < ... < 2 N = 1, we define a sequence 0 = u °, u 1. . . . . u N by the
relation
(u" + ~ - u")l(,~" + 1 _ ,l") = (A'(u"))- ~f, n = 0, 1 . . . . . N - 1.

By this successive solving of linear problems, we reach the approximate solution u N to

the problem. The implicit function theorem methods can be also adopted for other
boundary conditions.
The second approach used for hyperelastic materials starts with the variational
formulation (7). Since the integrand ff'(x, F) need not be convex with respect to F and
the set of admissible deformation is nonconvex due to injectivity condition det F > 0,
the existence of a minimizer does not follow by the usual methods. The problem was
solved by John Ball introducing the concept of polyconvexity: a function ff'(x, F) is
called polyconvex, if it can be written in the form
l~(x, F) = W(x, F, Cof F, det F),
where W(x, F, H, fi) is a function convex in variables F, H, ft. This polyconvexity
implies that the energy functional I($) is weakly sequentially lower semi-continuous,
which yields that a weakly converging minimizing sequence Sk converges to the
minimizer $. If, moreover, W is coercive, i.e.
IW(x, F)l >/~{IIFII p + IlCof FII q + (det F)'} + fl
on the set of admissible deformations
= {~be WI'~(~), Cof V~beLq(l)), det V~ke L'(f~),
~b = ~bo on ro, det V~b > 0},
then the minimizing sequence is bounded, contains a weakly converging subsequence,
and the existence of the solution follows. The method can be extended to other
boundary conditions, e.g. to problems with unilateral constrains. A polyconvexity
assumption represents a reasonable restriction of constitutive relations.
The book is written in the elegant, bright style of an experienced pedagog. The well
thought out notation is unified throughout the whole volume. Besides specific
BOOK REVIEWS 195

concepts of elasticity, all nontrivial auxiliary topics of functional analysis and other
branches of mathematics are explained and illustrated by examples. The chapters are
concluded by exercises.
The book is also ideal from the formal point of view, using various typefaces for
emphasis and, framing important formulas and figures. The volume is equipped with a
20-page survey of the main notation, definitions and formulas, a 573-item biblio-
graphy, and a 17-page index.
This nice book can be fully recommended as an introductory textbook to
contemporary nonlinear elasticity as well as a working textbook at graduate level for
courses in pure or applied mathematics or in continuum mechanics.

Technical University, JAN FRANCU

Brno, Czechoslovakia

Tullio Valent: Boundary Value Problems of Finite Elasticity, Local Theorems on

Existence, Uniqueness, and Analytic Dependence on Data, Springer Tracts in
Natural Philosophy, Volume 31 (Edited by C. Truedell), Springer, New York, 1988,
xii + 192 pp.

The monograph deals with three-dimensional finite elasticity. The author presents, in
a systematic form, local theorems on existence, uniqueness, and analytic dependence
on load, which he has recently obtained for various types of boundary-value problems
of elastostatics. The results are obtained by methods based on the clever use of the
inverse function theorem and the implicit function theorem. The method started by
Stoppelli in 1954 has recently obtained significant results.
The book has the following headings:
I. A brief introduction to some general concepts in elasticity;
II. Composition operators in Sobolev and Schauder spaces. Theorems on cont-
inuity, differentiability, and analyticity;
III. Dirichlet and Neumann boundary problems in linearized elastostatics. Exis-
tence, uniqueness, and regularity;
IV. Boundary problems of place in finite elastostatics;
V. Boundary problems of traction in finite elastostatics. An abstract method. The
special case of dead loads;
VI. Boundary problems of pressure type in finite elastostatics.
Thb problem of finite elasticity can be described by the equation
div A(D4~) + F(~) = 0 in t) (1)
with some boundary conditions, where l) is the reference configuration, ~: l)--, R 3 -

Acta ApplicandaeMathematicae 18 (1990)