A MORPHOSPACE OF PHYLA
Coopting data from Brusca and Brusca, Schram, Nielsen, and others, morphological characters originally coded for phylogenetic purposes have been used in multivariate ordination analyses, allowing the graphical projection of about 30 extinct and extant phyla onto single spaces of lower dimensionality. From various dataset partitionings, including some conducive to explicitly developmental morphospaces, several noteworthy general conclusions emerge:
- morphospaces of phyla are indeed representationally feasible, and there are several, nonexclusive solutions;
- no particular morphospace of phyla supervenes over others. The ideal of a single phylogeny is orthogonal to the theoretical richness of a plurality of morphospatial data analyses;
- interphyla disparities calculated on raw morphospaces are variously and imperfectly, yet nontrivially, approximated by the distances implied by phylum-level phylogenies;
- discordances among different morphospaces of phyla are only partly an effect of the cladistic nature of the data, or of the ordination methods here utilized. More importantly, they appear to be the outcome of the nature of characters, of the logic of presence-absence coding, and of the inclusion of living phyla (qua ghost-taxa subject to a “Pull of the Past”) into any given bodyplan sample of the Cambrian explosion.
Although heuristic, our exploratory analyses strongly suggest that a rigorous morphospace of phyla is tractable. Accordingly, the disparity among phyla can be quantified. Our approach holds promise in inviting further biological and paleobiological research on the Cambrian explosion and on patterns and processes of bodyplan macroevolution.