DNA-DNA HYBRIDIZATION STUDIES OF CARNIVOROUS MARSUPIALS .2. RELATIONSHIPS AMONG DASYURIDS (MARSUPIALIA, DASYURIDAE)

A suite of comparisons among ten radiolabelled dasyurid species and one outgroup bandicoot was generated using the hydroxyapatite chromatography method of DNA-DNA hybridisation; comparisons were also made with four other dasyurid taxa. Square matrices of DELTA-T(m)s, DELTA-Modes, and DELTA-T50H's were complied and corrected for reciprocity, additivity, and, in the case of DELTA-T(m)'s, normalised percentages of hybridisation. These matrices were analysed using the FITCH algorithm in Felsenstein's PHYLIP (Version 3.1), and all distinct topologies were jackknifed to test for internal consistency. Additionally, uncorrected DELTA-T(m), DELTA-Mode, and DELTA-T50H datasets were bootstrapped and subjected to phylogenetic analysis to assess measurement imprecision. FITCH trees from folded matrices including unlabelled species or those for which heteroduplex comparisons were incomplete were also calculated and jack-knifed, both before and after correction. With the exception of limited measurements to Dasyuroides byrnei and Dasykaluta rosamondae, which showed affinities with Dasyurus spp., the final tree was fully resolved: Sminthopsis crassicaudata and S. murina, together with the more distant Planigale maculata, are the sister-group to all other dasyurids examined, which in turn comprise two clades. One of these includes Dasyurus, Dasyuroides, and Dasykaluta; the other, 'true' Antechinus (A. flavipes, A. stuartii, A. swainsonii) as a sister-group to Antechinus melanurus plus Murexia longicaudata, with Phascogale tapoatafa representing a probable sister-group to all Antechinus with Murexia. DNA-DNA hybridisation provides no support for the genus Satanellus: most of the trees linked Dasyurus albopunctatus with D. maculatus instead of D. hallucatus. Similarly, Antechinus flavipes and A. stuartii appear to be closer to each other than either is to A. swainsonii. The historical biogeographic significance of the adopted phylogeny is considered, and it is concluded that the putative early Miocene separation of Australia and New Guinea was probably too early to account for the independent evolution of the New Guinean clade.