Why morphometrics isn't special: coding quantitative data for phylogenetic analysis. 1998. Systematic Biology, 47(3): 508-519

D. L. Swiderski, Zelditch, M. L., and W. L. Fink

One of the main obstacles to using morphometric features in phylogenetic analysis has been the coding of quantitative data. Coding has seemed a problem because many workers treat quantitative data as conceptually distinct from qualitative descriptions of the same traits. A misplaced emphasis on continuity of the measurement scales has led to questionable methods of coding and to questions about the validity of coding in principle. It is more appropriate to focus on evidence of novelty and homology, the hypotheses encoded in the data matrix. In a review of several methods of coding, we found only one method that has this focus. We exemplify this method with comparisons of piranha body shapes using data generated by thin-plate spline analysis, specifically partial warps scores. As with traditional qualitative data, some features provide no evidence of novelty, others provide suggestive but ambiguous evidence of novelty, and others provide evidence as compelling as any we have seen from any data source.