Determining the potential utility of datasets for phylogeny reconstruction

The utility of morphological and molecular data in phylogenetic inference has been widely debated. Potential utility is intrinsically related to size, so the importance of morphological data in phylogenetic inference is becoming increasingly uncertain as molecular datasets grow rapidly larger. To test the continuing relevance of morphological data, we present an empirical investigation of size in a selection of recent data matrices and contrast the possible methods of measuring relative potential utility. Molecular datasets contained significantly more variable and parsimony-informative characters, and implied a greater number of parsimony-informative character-state changes than morphological ones. The latter is suggested to be the most useful measure of potential utility in phylogeny reconstruction, and shows much greater variability and overlap between molecular and morphological datasets. Morphological datasets often have much greater phylogenetic utility than is apparent from their size, and thus molecular datasets are not always more useful simply because they contain a greater number of aligned characters. We recommend these more accurate measures of utility should be more widely reported in phylogenetic research.