EVOLUTIONARY PALEOBIOLOGY AND THE SCIENCE OF FORM (REVISITED)

Gould (1970) published the last major review of trends in morphometric analysis as applied to evolutionary studies. That review focused on a small set of topics that all became major themes in paleobiological research. The past 30 years have also witnessed the development of new morphometric tools and new ideas that guide the interpretation of morphometric data. Although Gould (with R. C. Lewontin) later published a controversial critique of pan-adaptationism, the study of adaptation proceeded down multiple pathways, very much as Gould envisioned. Morphometric analyses have contributed to these approaches, but failed thus far to serve as a basis for combining them into a unified research strategy. Allometry has undergone much conceptual development and has been used to study an expanding range of phenomena. Although the rationale of dimensionality reduction via allometric analysis has now been internalized by most practitioners, its corollary of functional constraint lacks an agreed approach for analysis and demonstration. Progress on this question has come mostly from EvoDevo research programmes. The morphological history of clades as assessed by 'morphospace occupation' has been a popular research program, though there remain lingering conceptual difficulties regarding the (implicit) spatial-domain dependence of empirical results. The distinction (or lack thereof) between empirical and theoretical morphospaces also remains a source of needless confusion. The study of heterochrony has gone full circle from an attempt to bring some systematization to a sprawling topic, to the epitome of all morphological evolutionary hypotheses, and back again. The problem with heterochrony has been that assessment of pattern (which can be inferred from morphometric data) has been mistaken for a knowledge of process (which, in most cases, cannot). The most successful of Gould's themes, though, has been an renewed appreciation of phylogeny's role in morphological comparisons. In addition to these areas, all of which are poised for continued growth during the next 30 years, I predict morphometric studies will also focus strongly on the perennial question of automated object recognition since the algorithms and distributed computer architectures necessary for this task have begun to appear.