Developments in Petrology, Volume 8: Numerical Petrology: Statistical Interpretation of Geochemical Data presents the methods that are likely to be useful to the average petrologist. This book deals with the problems of closed data and singular matrices in multiple discriminant analysis and classification procedure. Organized into 12 chapters, this volume begins with an overview of the petrological data that can be quantified, including both discrete and continuous variables. This text then examines the methods of testing for differences between the means of two populations. Other chapters consider the three methods of evaluating linear trends within such bivatiate plots, namely, the use of the correlation coefficient, linear regression analysis, and either structural or functional relationships. This book discusses as well the propagation of errors in mineral and normative recalculations. The final chapter deals with the use of computers to manage the tremendous amount of information that is available. This book is a valuable resource for petrologists, geochemists, and geologists.