PERSPECTIVES ON THE APPLICATION OF MULTIVARIATE STATISTICS TO TAXONOMY

Summary A brief outline is given of the principal types of multivariate statistical techniques which have found use in taxonomy. Techniques such as correlation, principal components, canonical correlation, and factor analyses are described for problems dealing with analysis of covariation within a single sample. Techniques such as canonical variate, cluster, multidimensional scaling, and network analyses are described for dealing with analyses of among sample variation. The purpose of this account is to give an intuitive understanding of what the various techniques have to offer to research in taxonomy.