Evolutionary Patterns in Some Putative Australian Species in the Ant Genus Rhytidoponera

Genic and morphological variation were compared for 17 putative Rhytidoponera species and a species of the related genus Heteroponera, by use of an allozyme data set and one based on morphometric, surface sculpture, and pilosity characters. Each data set was considered in three versions: the raw data, principal factor scores normalized to the appropriate eigen vectors, and these scores range-coded. The agreement between these data sets, and similar sets derived from published vertebrate studies, was gauged by means of correlation coefficients between distance matrices based on them, calculated by a jack-knife procedure. In all cases, the raw allozyme data sets gave the highest correlation with the morphological sets, but none of the treatments of the morphological data was clearly superior in this regard to the others. For the ant data, congruence between the two types of data was also examined by comparing the branching orders of dendrograms (Wagner and REML), by a new test employing distributions based on the differences between randomly generated branching orders and a reference dendrogram. According to this test, the morphological dendrogram based on range-coded principal-factor scores was significantly more similar to that derived from the raw allozyme data than were those based on the other two treatments of the data. Differences in chromosome number do not correlate well with genic and morphological ones, which indicates that the speed of karyotype change in this genus has been highly variable. Some OTUS showed duplicate-locus expression for IDH, and clustering in the allozyme- based dendrograms occurred on the basis of IDH duplicate-locus expression pattern. The two 'victoriae' populations studied cluster closely with the metallica group on the morphology-based dendrograms, in agreement with conventional views that 'victoriae' is very close to 'metallica', but diverge markedly when allozymes are considered, which indicates that the morphological resemblance is probably due to convergence. The large genetic distance between these 'victoriae' populations indicates the likely presence of sibling species. R. 'tasrnaniensis' populations, in contrast, cluster strongly with 'metallica' in both morphology- and allozyme-based dendrograms. The marked divergence of scabra from other large species in the allozyme- based dendrograms indicates that its large body size has been derived independently.