Population structure of the northern bettong, Bettongia tropica and the northern brown bandicoot, Isoodon macrourus

Abstract

Population structure underpins our understanding of speciation and metapopulation stability. On the one hand, high levels of structure allow the maintenance of high levels of genetic diversity, however the existence of such structure restricts the spread of adaptations. While it is possible to describe patterns of population structure, it is more difficult to disentangle the potential causes of such patterns. In this study I examined the population structure of two species, the northern bettong, Bettongia tropica, and the northern brown bandicoot, Isoodon macrourus, across several spatial scales. Bettongia tropica is an endangered species with a restricted distribution and specialised diet. Isoodon macrourus occurs sympatrically with B. tropica, but has a much wider distribution along the east and north coasts of Australia and an omnivorous diet. With the aim of determining the separate influences of habitat barriers, both current and historic, and behavioural barriers on gene flow in B. tropica, I combined broad scale analysis of phylogeography with studies of local-scale population structure and demography, and compared the patterns detected to those observed for the generalist I. macrourus. Bettongia tropica was demographically stable with high survival and low rates of turnover, even during periods after fire. Dispersal, inferred from microsatellite data, generally occurred over short distances with males dispersing further than females. The mating system can be characterised as overlap promiscuity, possibly with low levels of monogamy resulting from exclusive overlap. Both dispersal and mating systems are therefore likely to result in strong isolation by distance patterns. This was evident from the analysis of population structure in continuous habitat where both nuclear (microsatellite) and mitochondrial DNA (control region) markers detected structure. MtDNA structure was an order of magnitude higher than microsatellite structure. While supporting male-biased gene flow this also indicated that males are more likely to undertake infrequent long distance dispersal than females. While structure is expected in a linear habitat in the presence of moderate gene flow, asymmetrical patterns of gene flow suggested that the local model is more complicated than simple linear isolation by distance. Analysis of phylogeography showed two distinct clades, north and south of a known rainforest biogeographic barrier, the Black Mountain Barrier, but with admixture in the central Lamb Range population and a lower level of divergence detected between the two clades than in rainforest species. Analysis of mtDNA control region sequence from Isoodon on the Lamb Range revealed two highly divergent lineages. This prompted a more widespread analysis of the genus which showed that a population of I. o. peninsulae was present on the Lamb Range, occurring sympatrically with I. macrourus. This represented a significant range extension for the sub-species. Isoodon macrourus had high levels of population turnover, particularly after a low intensity fire. Dispersal, inferred from recruitment, appeared to be male-biased following the fire but female-biased before the fire. Structure within continuous habitat, sympatric with B. tropica, was detected with both nuclear (microsatellite) and mtDNA (control region) markers. The level of structure detected was similar for both markers. Such structure could result from limited dispersal, drift or strong linear habitat effects, though this is not expected for a habitat generalist. Little support was found for rainforest contraction effects on I. macrourus from phylogeography. There were two clades, with samples north of the BMB found in only one clade, but both clades occurred throughout the Lamb Range. Comparison of B. tropica and I. macrourus indicated that B. tropica was historically an ecotone specialist and influenced by rainforest contractions, but that I. macrourus, the more generalist forest species, was less affected by rainforest contractions. B. tropica showed strong genetic structure at all scales; however, while I. macrourus did show structure in continuous habitat this was not reflected over a broader scale. Population structure across distances as short as nine km was an unexpected result for both species, particularly I. macrourus. While dispersal within continuous habitat may be effectively similar for the two species for nuclear makers, this appears to be the result of different processes. Structure within B. tropica can be attributed to effects from behaviour within continuous habitat as well as habitat barriers. In comparison, I. macrourus shows short term structure, possibly resulting from fluctuating population dynamics, that may be ameliorated by gene flow over a longer time scale and across a broader array of habitats.