"Of the making of new books there is no end," said the Preacher. He (she?) might have added "and of the reprinting of old books there is not enough." All too often great old books become unavailable and students are denied the opportunity to learn from them. This is all the more serious when the books in question are written by the original creators of the topics they discuss.
Here's one. Jürgen Moser's Stable and Random Motions in Dynamical Systems is motivated by the stability problem in celestial mechanics: can we prove that the solar system is stable? In the first chapter, Moser explains the historical roots of the question, makes it precise, and sets up the mathematical questions that the rest of the book will address. The other chapters center on two big theorems: the Kolmogorov-Arnold-Moser (KAM) theorem, dealing with quasi-periodic motions, and the Smale-Birkhoff theorem, connecting dynamical systems with Bernoulli processes. These are, respectively, the "stable" and "random" aspects mentioned in the title.
The book is part of Princeton University Press's Landmarks in Mathematics series. The series consists of paperback reprints of old mathematical classics, ranging from Milnor's Topology from the Differentiable Viewpoint to Cartan and Eilenberg's Homological Algebra. All in all, a worthy endeavor.
Fernando Q. Gouvêa (fqgouvea@colby.edu) is the editor of FOCUS and MAA Online. He teaches both "History of Mathematics" and "Number Theory", among others, at Colby College. He is a number theorist whose main research focus is on p-adic modular forms and Galois representations.