This article summarises current knowledge of marsupial population genetics, and discusses its relevance to the conservation of marsupial species. It has been suggested that there is much lower genetic variation within marsupial populations than in eutherian mammals. This trend is not evident in the electrophoretic data summarised here. However, genetic differentiation between populations, subspecies, and species of marsupials appears to be slightly lower than comparable values for eutherians. Genetic estimates of migration between populations are scarce at present, but show values that are comparable with eutherians. Some studies of marsupial population genetics have used non-electrophoretic characteristics, or have addressed the possibility of selection on the characters analysed. Although few, these studies indicate the suitability of marsupials for such investigations. Recent debate over the theories and applications of conservation genetics has made it clear that more research is required on individual species. Given the record of extinction of marsupials in the last 200 years, it is important to test the applicability of these theories to individual marsupial species. Several examples are discussed emphasising the need for ecological studies that estimate the effective number of reproducing individuals per generation. This figure, called the effective size, is the corner- stone of conservation genetics theory, being an important determinant of both the rate of loss of variation between individuals, and the rate of inbreeding. The effective size of the mainland population of the eastern barred bandicoot, Perameles gunnii, appears to be only about one-tenth of its census number. This result is comparable with estimates made in other vertebrates, and demonstrates that many marsupial species which appear to have an adequate census size on ecological grounds may face genetic problems resulting from small effective size.