Keith Devlin is one of the best popular mathematics writers around, and this is one of his best works. The six chapters cover number theory, set theory, calculus, group theory and topology; but to state it baldly like this is to miss the main value of this seductively illustrated book. Devlin titles his chapters innocuously--"Shape", or "Position"--and the initial discussion, couched generally in English, not mathematics, is so clear that a math-phobic can understand it. By the end of each chapter a great deal of fascinating mathematics has been described, and in some cases the formal basis is sketched--but the emphasis is always on narration, and a lay reader who doesn't even want to understand mathematics can still read this and get a sense of the dramatic history of mathematics. And of the dramatis personae, too; one nice feature is the large number of good pictures of mathematicians, including several more recent figures such as Ribet and Thurston.
Devlin states at the end that he decided to exclude many areas of mathematics in order to focus more effectively on what he did cover. As a result there is little or no coverage of chaos theory, game theory, catastrophe theory, or a long list of other topics. The fact is there will always be holes in a book this size--mathematics has expanded so much in the last hundred years that even a book ten times this size could barely survey it. The decision to focus was a good one, and the subjects chosen are good: the truly exciting stories are here: Archimedes, Fermat, Gauss, Galois, Riemann, Wiles, and many more.
The illustrations deserve an extra comment. I've already mentioned the pictures of mathematicians. There are good diagrams, of the quality you'd expect from Scientific American. There are also plenty of pictures of the sort you see in every maths book of this kind--Escher tessellations, Kepler's nested Platonic solids, a Durer perspective drawing. But there are several more that I've never seen (and I've read a lot of these books). Two notable pictures: a cardboard model of an aperiodic tiling of space, by John Conway; and a picture of a set of tiles at a Dutch high school, designed by Escher: I'm an Escher fan and have never seen these before.
Potential purchasers should note, by the way, that this book was reworked into Devlin's "Language of Mathematics". In Devlin's words (not from either book): "The Language of Mathematics is a restructuring of Science of Patterns that omits most of the color illustrations (a minus) but has two new chapters covering topics not in Science of Patterns (a plus). If you want lots of color, go for patterns; Language of Mathematics covers more ground."
This is a fine book. Strongly recommended.
