DNA hybridization was used to compare representatives of the major groups of marsupials and a eutherian outgroup. Because of the large genetic distances separating marsupial families, trees were calculated from normalized percentages of hybridization; thermal-melting statistics, however, gave identical topologies for the well-supported clades. The most notable results were the association of the only extant microbiotheriid, Dromiciops australis, an American marsupial, with the Australasian Diprotodontia, and of both together with the Dasyuridae. Estimates of the rate of divergence among marsupial genomes suggest that the Dromiciops-Diprotodontia split occurred -=50 million years ago, well after the establishment of the major clades of marsupials but before deep oceanic barriers prohibited dispersal among Australia, Antarctica, and South America. Because Dromiciops is nested within an Australasian group, it seems likely that dispersal from Australia accounts for its present distribution.Dromiciops australis, the monito del monte of southern Chile and adjacent Argentina, had been considered a small-bodied opossum until Reig (1) noted dental and other similarities to members of the Tertiary family Microbiotheriidae and suggested that Dromiciops is a living representative of this otherwise extinct taxon. Subsequent study of skull anatomy (2) confirmed Reig's proposal, and serology (3) indicated the distinctness of Dromiciops vis-a-vis didelphids. Later, Szalay (4) demonstrated similarities between the tarsi of Australasian marsupials and Dromiciops and proposed that Dromiciops represents the sister group of all Australasian species; together they comprise a monophyletic taxon (Australidelphia) apart from other marsupials (Ameridelphia) in his system. Much recent work has supported Szalay's view: Sharman (5) noted details of Dromiciops' chromosomes comparable to those of several Australasian families; Gallardo and Patterson (6) found male sex-chromosome mosaicism like that in the Australian diprotodontian Petauroides volans; and Temple-Smith (7) documented similarities among the spermatozoa of Dromiciops, phalangers, and kangaroos. Reig et al. (8), however, criticized Szalay's interpretation of tarsal anatomy, and argued that Australasian affinity was incompatible with craniodental evolution, suggesting instead that Dromiciops is the terminus of a lineage that evolved independently of Australasian marsupials.As part of a DNA hybridization study of the relationships of opossums (Marsupialia: Didelphidae), we constructed a matrix including representatives of all major marsupial groups and a eutherian, Procyon lotor. Since one of our aims was to root the didelphid tree, more exemplars of Didelphidae (Didelphis, Metachirus, Monodelphis, Caluromys) were included than of any other marsupial group. Because of the limited range of thermal-melting statistics (9), and the fact that a recent study of Dromiciops using At50H (median melting temperature of hybridized plus potentially hybridizable sequences) (10) did not re...