The attractions of gravity are felt by us all, not least by the writers of texts. Yet the theory of gravity is a spare and austere subject, discovered by the few for the use of the many. Newton's great work, published more than three hundred years ago, encompassed all that we knew until superseded in the early years of the 20th century by the general theory of relativity. Einstein's theory appeared ready made without precursors or contributions by others: as complete and perfect an intellectual creation as there has ever been.

The single-mindedness of Einstein's development of a relativistic theory of gravity, the elegant economy of his logic, and the mathematical sophistication of his original exposition, has dictated the way it has been taught and understood ever since. The first break with that tradition was Gravitation, the gigantic work of Charles Misner, Kip Thorne and John Wheeler, which introduced the language of differential forms into the mathematical expression of Einstein's theory.

While this formalism is attractive to mathematicians and efficient for the proving of general theorems, it is less attractive for the formulation of particular problems that require initial data to be entered in some coordinate system, or for numerical computation. An interesting feature of this latest co-authored work by Ignazio Ciufolini and Wheeler, one of the greatest contributors to our understanding of the inner workings of Einstein's theory and its experimental consequences, is its adoption of the tried and tested exposition of Einstein's theory in the language of tensors.

But, this book should not be mistaken for just another retreading of Einstein's path to the field equations, followed by the standard solar system tests of its weak-field limit and Aleksandr Friedmann's cosmological solutions. The novelty of this book is its detailed description of all the work going on in experimental gravitation physics. Little of this can be found in other books and none of it is covered in such enlightening detail anywhere else.

ÌÇÐÄVlog

Ciufolini is a specialist in experimental tests of gravitation theories and brings a detailed and wide-ranging knowledge of the achievements and prospects for experimental gravitation studies to the book. Another notable characteristic of the book is its comprehensive list of references which makes it a valuable resource for studying the development of the subject. Its blind spot is in cosmology. Here the authors have not taken modern developments in astrophysical cosmology on board at all and emphasise purely formal aspects of the subject, for example the Bianchi-type classification of homogeneous cosmologies without giving specific solutions or linking it to astrophysical questions like the isotropy of the universe.

Even the Bianchi-type IX universe is discussed geometrically, ignoring the unusual chaotic behaviour found near its singularities. The inflationary universe is squeezed into a page and a half of brief notes and the behaviour of perturbations to the Friedmann universe is not discussed. The cosmology is crammed into a single chapter of only 17 pages and, although it collects together a number of interesting mathematical results about the global structure of cosmological models, it sits a little uneasily alongside the other material.

ÌÇÐÄVlog

The book begins with an overview of the topics to be covered in future chapters. The authors use simple dimensional estimates to express the influence of different gravitational effects and bring out the similarities and differences with electromagnetism. Readers will want to return to this chapter after covering the detailed accounts in later chapters: it is full of memorable Wheelerisms which establish the principal ideas.

A weak point is the immediate use (without warning or explanation) of a system of units in which the gravitation constant and the velocity of light are equal to unity. This has the unfortunate effect of making the estimate of gravitational effects look dimensionally incorrect to the beginner. Thus, on page 6, the gravitational effect of the earth on a rotating gyro is given as the mass of the earth divided its radius with the mass of the earth as 0.44 cm in these units. It would have been best to retain all constants of physics explicitly in this first chapter and then explain the advantages of choosing different systems of units in which some of them equal unity at the end before carrying on with the authors' economical choice for the more formal chapters that follow.

The whole development of tensors, curvature, the field equations and their simplest solutions are all over in less than 60 pages. This material is standard and can be found in many other textbooks. But there are then two long chapters, spanning more than 170 pages, which describe the tests of general relativity and other gravitation theories in great detail.

All the most up-to-date experimental results are there and the physics behind the distinctive physical consequences of the general theory of relativity expounded with care and illuminated by means of vivid contrasts with the structure of electromagnetism.

ÌÇÐÄVlog

The only major topic I looked for in vain was binary pulsars. Since its Nobel prize-winning discovery by Russell Hulse and Joseph Taylor, the binary pulsar PSR 1913+16 has been the prime astronomical laboratory for precision testing of Einstein's predictions, making it the most accurate predictor of nature's workings conceived by the human mind. It deserved a dedicated discussion in the fullest detail in a book such as this; instead it receives several disappointingly brief notices, usually at the ends of sections about far smaller perihelion precession effects in the solar system.

This is not a text that could be used in isolation by undergraduates because it moves so rapidly in the early stages and it lacks problems and exercises. But as a supplement to a lower-level textbook, as a source for project or essay work, or a taster for students contemplating graduate work in (nonquantum, noncosmological) gravity, it is a goldmine of illuminating analogies, simple derivations and apt illustrations.

John D Barrow is professor of astronomy, University of Sussex.