This book avoids repeating or summarising Gleick's "Chaos" and Stewart's "Does God Play Dice?", and introduces an interesting spin on the topic, at least in the early chapters. He introduces the "Burns' Effect" , named after the poem "To a Mouse", and has elsewhere some apposite and interesting quotations, for example from James Clerk Maxwell, shown to anticipate sensitive dependence on initial conditions. The choice of "The Cheat With The Ace of Diamonds" is an interesting illustration and leads to an interesting discussion of different scientific philosophies. Later on, the book attempts to discuss the philosophical and practical implications of the existence of chaos/nonlinear dynamical systems for science. However, I did not end the book with a sufficiently clear and detailed grasp of what the author thought these were. Except a sense that the problems include the difficulty of: model validation, parameter estimation, and interpretation of ensemble forecasts. The author avoids the use of equations, instead sometimes substituting wordy descriptions of the meanings of equations. This is ridiculous. None of the equations so described are beyond school level algebra - anyone interested in this topic is bound to understand the simple equations involved.