In 1976 (incidentally, the year the reviewer was born), Thurston first circulated his now-famous preprint on the classification of surface homeomorphisms [81]. It states that every homeomorphism of a compact surface is homotopic to a homeomorphism in standard form.The generic standard form is a pseudo-Anosov homeomorphism. Such a homeomorphism acts on the surface by preserving two transverse singular measured foliations, multiplying the measure of one by a stretch factor λ > 1 and the other by 1/λ. In other words, away from a finite set of singular points, a pseudo-Anosov homeomorphism is modeled on an Anosov map of the torus, or equivalently, an element of SL(2, Z) with two real eigenvalues.The precise statement of Thurston's theorem is: Every homeomorphism of a compact surface is homotopic to a homeomorphism that either (a) has finite order, (b) is reducible (that is, fixes an essential 1-submanifold), or (c) is pseudoAnosov. Moreover, the last case is exclusive from the first two. It was realized after Thurston's work (see [63]) that Nielsen had assembled all of the tools needed for the classification [66][67][68][69], and so this theorem is sometimes called the Nielsen-Thurston classification theorem.Let Mod(S) be the mapping class group of a surface S, that is, the group of homotopy classes of homeomorphisms of S. Another way to state the classification which suggests an analogy with the Jordan canonical form is: for every element of Mod(S), there is a representative φ and a (possibly empty) φ-invariant 1-submanifold C so the restriction of φ to S − C has two kinds of components, finite order and pseudo-Anosov. Birman, Lubotzky, and McCarthy showed that the reduction system C is canonical [13].Shortly after Thurston's announcement, there was a flurry of activity to understand what he did. The book under review is the product of a year-long seminar at Orsay devoted to Thurston's work. The lectures were based on notes from Thurston's graduate course at Princeton University, handwritten by Michael Handel and William Floyd.In the broadest of strokes, Thurston's proof proceeds as follows. Let S g be a closed, connected, orientable surface of genus g ≥ 2. Fricke showed that the Teichmüller space Teich(S g ), the space of hyperbolic metrics on S g up to isotopy, is an open ball of dimension 6g − 6. The key idea of Thurston is that Teich(S g ) has a compactification that is homeomorphic to a closed ball; the boundary sphere is PMF(S g ), the space of projective classes of measured foliations on S g . Moreover, Mod(S g ) acts on this closed ball. Then, in a most spectacular application of the Brouwer fixed point theorem, Thurston concludes that each element of Mod(S g ) fixes some point of Teich(S g ) ∪ PMF(S g ); by analyzing the various cases for the fixed point, we obtain the classification theorem.