Book reviews

Abstract

To many of us who are students of quantitative and statistical research methodology, the names of Cliff and Ord immediately bring up the concept "spatial autocorrelation." This is certainly a very natural reaction considering the fact that they have, together and separately, published an impressive amount of material on the subject. The latest effort, Spatial Autocorrelation, perhaps summarizes much of their work to date.The name Haggett is without any doubt one of the most familiar names in modern geography. Although better known for his synthesis of research frontiers in geography, Haggett has made substantial contribution to application of statistical methods to regional planning, as well as medical geography. The other two names, Bassett and Davies are less familiar to this reviewer. Bassett has of course collaborated with Haggett previously in at least an economic forecasting study, and has contributed a chapter to the Progress in Geography series.Given some knowledge of the research interests of four of these five British scholars, it becomes readily apparent with what elements Elements of Spatial Structure deals. Of the ten chapters in the book, each individual contribution basically reflects interests and expertise of the individual scholars. Essentially, Elements of Spatial Structure involves the application of certain quantitative methods, collectively known as spatial series, to study elements of the region. Spatial series can be seen as a derivative of time series. Spatial series, however, deal with two dimensions, whereas time series only deal with one. Another difference between the two is time series analysesthings that are naturally ordered in terms of time, but this natural ordering is absent in spatial series.Three major sections, within which there are ten chapters, underline the organization of this book. Two major aspects of regional structure, static and dynamic, are dealt with in the first two sections and antocorrelation and forecasting form the last section. Although, it contains no more than three hundred pages of material, this book deals with a vast number of techniques, methods, and models. Many of these methods are not easily found in books written for similar purposes.The static aspects of regional structure, Part I, involve a central problem in the planning process--how to aggregate discrete geographical subareas into regions. The first chapter is a discussion of such region-building problems as size and best grouping. The mathematical methods used include combinatorial techniques and the theory of graphs. The second chapter is a discussion of regions as ordered series. The central problem, however, is the well-known rank-size relationships. The authors recognize two approaches to the study of growth and formulation of regional mosaics: that of an analysis of political decision making, and that of the derivation of stochastic models to, probabilistically, treat the cumulative effect of all the events involved. It is the second