BOOK REVIEWS 187

and analogues. Chapter 7 consists of further generalisations to the case of Toeplitz

operators associated with a kernel K(s, t), defined with reference to two measure
spaces and a function ^>(x, a) measurable with respect to their product measure.
In the later part of the chapter some previous results are re-examined in the light
of these generalisations—for instance, in § 7.6, there is an ingenious alternative
proof of a theorem of § 5.2. If a certain limiting property holds, the class of
Toeplitz matrices corresponding to a class of real-valued functions is said to be
trace-complete ; the first part of chapter 8 concerns the properties of such classes
of matrices. §§ 8.4, 8.5 concern Toeplitz matrices associated with certain special
orthogonal systems, while the rest of the chapter examines two special kernels
K(a, t).
The first chapter (9) of the Applications section of the book makes use of the
properties of the eigenvalues of Toeplitz forms to deduce several results on analytic
functions regular within the unit circle. Some of the results in the final chapters
(10, 11), which deal with the applications to probability theory and statistics, are
recent, and there is a link between Toeplitz forms and the work of Kolmogorov
and Wiener on stationary processes. Some of the work of Grenander on stochastic
processes is incorporated into these chapters.
This is a stimulating book, designed primarily for the research worker who is
anxious to have available, in a compact form, some of the recent work on the
subject, together with a survey of older results. Its value is enhanced by the
bibliography and by the Appendix which provides (with additional references)
an annotated commentary on the text. An unusual amount of material is com-
pressed into the 228 pages, mainly by the omission of tedious detail. Despite this,
the printing is uncramped and the book remains, both in appearance and content,
a pleasure to read. DENNIS C.

CHURCHUL, B. v., Operational Mathematics, 2nd ed. (McGraw-Hill Book Co., New
York, 1958), 306 pp., 41s.
The general arrangement of the book is unaltered from the first edition published
in 1944 under the title Modern Operational Mathematics in Engineering. There
is, however, an extensive revision of detail, the concept of the delta-function is
introduced, the chapter on elementary applications now contains sections on
electric circuits and servomechanisms, and many more examples for the reader
to work are included. The bibliography extends to twenty-two titles including
references to tables of Laplace and Fourier transforms.
Perhaps the most obvious change is in the title, implying that the book is now
of interest to a wider class of reader than was suggested by the title of the first
edition. The change is justified for this is a sound mathematical treatment of
the theory of the Laplace transform and its application to problems of physical
origin. Great care is taken to state clearly and precisely the conditions under
which results are established. These conditions are usually simple and practical,
the author having resisted the temptation to enshroud the theory in mathematical
sophistication. The treatment should satisfy all with a reasonable outlook whether
they incline to pure or applied mathematics.
Recent developments, applying operator methods to partial differential equations
by the use of integral transforms with kernels of various types, are reflected in the
chapters entitled " Sturm-Liouville Systems " and " Fourier Transforms ". The
first of these has been radically revised providing a much more satisfying account
of second order linear differential systems ; an addition of particular interest in
the other chapter is the account of how the kernel appropriate to a given differential
system may be determined.

https://doi.org/10.1017/S0013091500021751 Published online by Cambridge University Press

 188 BOOK REVIEWS
The impression was gained that this extensive revision of a well established
book was carried out by an author still very much absorbed in his subject.
J. FULTON
KOtMOGOBOv, A. N., AND FOMIN, s. v., Elements of the Theory of Functions and
Functional Analysis, vol. i : Metric and Normed Spaces, translated by LEO F.
BORON (Graylock Press, Bochester, N.Y., 1957), 129 pp., 32s.
This is an excellent introduction to the ideas and methods of functional analysis.
The original version was based on lectures given by the author in Moscow, and the
translator has produced a very readable English text. Although nothing beyond
elementary analysis is presupposed, the principal abstract concepts are illustrated
by an ample variety of examples, and the theory is elegantly applied to problems
of considerable practical interest. The approach is essentially " classical " : all
limits are sequential, and no use is made of any principle of transfinite induction.
This is a weakness from the point of view of the serious student of modern analysis,
but it makes substantial parts of the subject easily accessible to many others.
There are four chapters. The first is concerned with some elementary facts
of abstract set theory (with no mention of the Axiom of Choice). Chapter II, the
longest, is devoted to metric spaces, with a brief reference to general topological
spaces. The main themes here are completeness and compactness. The fixed-point
theorem for contraction mappings in a complete metric space is thoroughly exploited
in a discussion of iterative methods for solving equations, including differential
and integral equations, for which several existence theorems are proved. The
fundamental properties of compact sets in metric spaces are established (the word
" compact " is used in the relative sense, which is now unusual), and there is a
careful account of the relations between compactness and equicontinuity in spaces
of continuous functions on compacta. The chapter ends with a useful discussion
of rectifiable curves. Questions of category are not considered, and there is no
mention of local compactness.
Chapter III, on normed vector spaces, is concerned largely with continuous
linear functionals. Conjugate spaces are determined in some simple cases, and
their are brief discussions of reflexivity and weak convergence. A proof of the
Hahn-Banach extension theorem, and an account of weak* compactness, are
restricted to separable spaces (of necessity, since transfinite induction is not
available). Continuous linear operators on Banach spaces are also considered,
and there is a proof of Banach's theorem on the continuity of inverse operators.
The Banach- Steinhaus theorem is, surprisingly, omitted. An addendum to this
chapter gives some of the main facts about " generalized functions " (distributions).
In Chapter IV, the idea of the spectrum of a linear operator is introduced, and
the main theorems on completely continuous operators are proved, giving the
Fredhohn theory of integral equations.
There is a good index, and the translator has added a bibliography. The
authors promise further volumes, in which they propose to discuss, among other
things, Lebesgue integration and Hilbert-space theory. J. D. WESTON
PONTBYAGIN, L. s., Foundations of Combinatorial Topology (Graylock Press,
Rochester, N.Y., 1952), 99 pp., $3.00.
This is a translation of the first (1947) Russian edition of a book which the
author says is " essentially a semester course in combinatorial topology which I
have given several times at Moscow National University ".
There are three chapters. In Chapter I the Betti (homology) groups are defined
for polyhedra; in Chapter II the topological invariance of the groups is proved ;

https://doi.org/10.1017/S0013091500021751 Published online by Cambridge University Press