Multivariate analysis of multi-character individual profiles as a tool in systematics

The biological concept of species, which considers either reproductive isolation or the recognition mating system as criteria for the species definition, suffers several operational difficulties. Reproductive isolation can be seldom demonstrated, especially when hybridization and introgressive phenomena occur, when populations are allopatric, and in non-amphigonically reproducing organisms. In such situations, traditional populational approaches may present problems and the description of taxonomic samples or their variability as averages, variances, or frequencies may be inadeguate. Multivariate ordination techniques may be usefully applied to individual data sets. This method offers several advantages in clarifying systematic and evolutionary relationships among demes, populations or species. First, it makes possible to evaluate the relative location of each individual with respect to various character sets with different properties. Second, it makes possible an identification a posteriori of each individual as > or > a specific cluster. Third, it makes possible to establish relative similarity among taxa within a reference system which is not deformed by cluster algorithms or a priori established character states. Finally, it provides essential indication on the most reliable taxonomic characters to be considered for current systematic work. Two study examples are reported to illustrate the usefulness of this kind of approach in natural situations, where hybridization and introgression occur