This volume represents a watershed in writing about the mental illness services, in that the authors are proposing a model that brings together a public health concern with the health of populations, with an insistence on remembering that services must also be judged by their effectiveness in dealing with disorders at a patient level. It was all there before they wrote it, but no-one has previously put it together so elegantly. Space and time arefairly obvious dimensions to choose for a model; what they have achieved is to produce a simple model that neatly serves as a framework for comparisons between different services, and illuminates the way in which planning at different levels of the model relates to events at other levels.It is especially ingenious to use the ‘time’ dimension to indicate time with respect to the treatment of illnesses, rather than merely indicating the historical passage of time. This gives the whole model greater power, which the authors have exploited brilliantly. Nothing will ever be quite the same again; it is a book certain to be widely read and quoted.

One point needs further clarification. Models can be predictive, they can be explanatory, they can be heuristic, or they can be descriptive. In Chapter 17 misgivings are expressed from the Nordic countries that the model is not specific, which implies that no predictions follow from it. First and foremost, this is a descriptive model,which helps to ensure that like is compared with like, by carefully disentangling the nine divisions in the proposed matrix.