Discussion

Abstract

R. M. C. Eagar writes: Vasey has defined statistically, by cluster analysis, two groups (subspecies) of Anhaconaut which correspond to already named groups in the Trueman & Weir (1946) system of nomenclature (A. phillipsii and A. tenuis), and has added to the system a third statistically defined subspecies (A. calveri). This is a new achievement. If cluster analysis can be further extended to groups of other non-marine bivalve genera, then the work could pave the way for the ultimate establishment of a trinomial system of nomenclature, as he claims. This is a logical step in classification, however unwelcome it may seem to those who try to identify small, incompletely known and sometimes poorly preserved samples. The trinomial system was considered, but rejected by Trueman & Weir (1946), partly because data then available proved too incomplete in much of the Upper Carboniferous record, and particularly because the second or specific term could not be tablished on any but an arbitrary basis for the division of biospecies. Cluster analysis, based on measurements, appears to offer an objective basis for the division of subspecies and, one can hope, for species. It now therefore seems necessary to look for examples of British assemblages and communities of non-marine bivalves which have been collected, or can be collected, in quantity sufficient to provide ample data for a pioneer study to test fully the application of cluster analysis. Of the six generation characterizing Westphalian A and B (Fig. 6, bottom), Carbonicola and