The disparity of the Burgess Shale arthropod fauna and the limits of cladistic analysis: why we must strive to quantify morphospace

Abstract

Three major arguments have been raised against the crucial claim, documented by Whittington and colleagues for the Burgess Shale fauna, and so contrary to traditional views, that disparity of anatomical design reached an early maximum in the history of multicellular life: (1) the presence of many early taxa with low membership and high rank is an artifact of naming; (2) cladistic analysis of Burgess arthropods negates the claim for greater early disparity; and (3) Whittington's argument is a retrospective fallacy based on assigning high rank to differentia only by virtue of their later capacity to define major branches. I show that all these arguments are either false or illogical, and that the claim for increased early disparity is justified: (1) Taxonomic rank is an artifact, but no one has ever based a claim for greater disparity on this false criterion. (2) Cladistics can only deal with branching order, whereas disparity is a phenetic issue. These two legitimate aspects of evolutionary “relationship” are logically distinct. The rooting of a cladogram only illustrates monophyletic ancestry (which no one doubts, as we are not creationists), and cannot measure disparity. (3) The active stabilization of the differentia of Baupläne (for genetic and developmental reasons only dimly understood) provides a powerful rationale for weighting these characters in considerations of disparity; nothing had so stabilized in the Burgess fauna. If these differentia were steadily changing contingencies, rather than actively stabilized features with “deep” architectural status, then the retrospective argument would be justified. Although the three arguments are wrong, the claim for greater early disparity cannot be confidently established until we develop quantitative techniques for the characterization of morphospace and its differential filling through time. This is a dauntingly difficult problem, much harder than cladistic ordering, but not intractable.