Are morphology-based cladograms of arthropods really less robustthan those of vertebrates?

Abstract

Received wisdom holds that many groups of arthropods are particularly morphologically homoplastic and difficult to resolve phylogenetically. Here we test this assertion by comparing 90 arthropod cladograms with 300 cladograms of vertebrate taxa. Conventional indices of homoplasy, including the homoplasy index (HI = 1-CI), are influenced by dataset parameters (i.e., the number of taxa and possibly the number of characters). The Homoplasy Excess Ratio is much more suitable for comparing levels of homoplasy between datasets, but requires an intensive permutation procedure and is therefore very rarely reported. Regression models including dataset dimensions, taxonomic level and taxonomic group as predictors reveal that homoplasy in arthropod cladograms is significantly greater than in vertebrate trees. Moreover, measures of tree support including the total support index (ti) are significantly lower for arthropod than vertebrate cladograms, reflecting inferior Bremer support at internal nodes. Initial results suggest that small perturbations in the taxon sample have a greater impact upon the relationships in arthropod trees than those of their vertebrate counterparts. This may partially reflect differences in the adequacy of the available (living and fossil) taxon samples in the two groups. Indeed, when arthropod cladograms are plotted onto stratigraphic range charts of first fossil occurrences, the ghost ranges inferred are often extensive, and stratigraphic congruence (GER, MSM etc.) is frequently poor. This contrasts with the excellent congruence in many vertebrate groups. The difference may partially reflect the particular concentration of important arthropod fossils in a limited number of exceptional sites. Finally, we present preliminary results comparing arthropod trees inferred variously from more and less fossilisable characters. Reassuringly, these data partitions often do not yield significantly different trees, which speaks for the utility of even incomplete fossils for increasing the taxon sample.

Cite this document ^(BETA)