ANALYSIS OF MORPHOLOGICAL GROUPINGS USING ANOPA, A PATTERN RECOGNITION AND MULTIVARIATE STATISTICAL METHOD: A CASE STUDY INVOLVING CENTRARCHID FISHES

Abstract

Morphological relationships within and among taxonomic groups can be very complicated, with anatomical data often supporting two or more incongruent groupings. One possibility is that incongruent character states are taxonomically informative, although in an N-dimensional taxic space. To test the above, morphological relationships of centrarchid fish species were examined using a new pattern recognition, multivariate correlation, and multivariate statistical analysis method (ANOPA). The objective of ANOPA is to identify N-dimensional pattern space correlations among character states, relations that cannot be detected with standard phenetic or phylogenetic approaches. ANOPA provides a solution to an inherent weakness in statistical analysis which occurs in the face of set classification ambiguity, where there is no a priori reason to assign a membership or class identification within multivariate statistical groups. This approach revealed the percoid fish family Centrarchidae to be a statistically significant, cohesive group with complicated internal relationships. Centrarchid taxa are resolved into three major generic aggregates by two and three-dimensional ANOPA, and discrete subgroups were also detected. The complex interrelationships within the Centrarchidae cannot be readily collapsed to a bifurcating tree-structure, explaining the multitude of conflicting phylogenetic hypotheses that have been presented. This is the first robust study of anatomical disparity in teleostean fishes. Applications of ANOPA to the study of morphological gaps, complex taxonomic patterns, and anatomical disparity are discussed.