The authors have exploited their wealth of practical experience to create an accessible, interesting and up-to-date survey of the statistical issues in soil sampling. They have targeted researchers involved in field sampling but who are not themselves statisticians. The book title specifies environmental science and management but the content is relevant to researchers from any discipline who have to sample soil or any other medium.

Technically, the first seven chapters demand only basic familiarity with mathematical notation, and random variation. These chapters cover simple random sampling; practical considerations; prediction from spatial classification; sampling from designed experiments and sampling for regression and calibration. Each chapter begins with a brief summary of its key points and the statistical concepts are always clearly explained and well illustrated with real data sets.

Chapter eight introduces unbalanced nested sampling designs and their analyses by Restricted Maximum Likelihood (REML). All the didactic strengths of the previous chapters are retained but the option of grappling with the matrix notation of mixed models is available. Similarly chapter nine includes some more technical material in its graded introduction to the model-based prediction methods of kriging. This chapter finishes with a discussion of some recent research into identifying optimal sampling designs by simulated annealing.

My only quibble is the omission of any reference to specific software packages. Some suggestions from the authors would have rounded off perfectly this excellent monograph, which I regret not being available on my bookshelf years ago.