Australia has lost more native mammal species than any other country in the past two centuries, and this record of loss looks likely to worsen over the next few decades. Small-to medium-sized mammals are declining in both distribution and density across large tracts of northern Australia's tropical savannas, including within protected areas. The most likely causes are a combination of changed fire patterns, the impacts of introduced herbivores and predation by feral cats. Here, in contrast to the prevailing trend across northern Australia, we report the recovery of native mammals in response to a large-scale (>40 000 ha) destocking experiment carried out at Mornington Wildlife Sanctuary in the central Kimberley, north-west Australia. Following the removal of introduced herbivores from 2004, the species richness and abundance of small native rodents and dasyurids increased significantly across all sampled habitats over the next 3 years.We discuss the implications of these results for guiding land management and applied research to help to reduce the impending risk of mammalian extinctions in northern Australia.

No abstract

Large predators can have profound impacts on community composition. Not only do they directly affect prey abundance, they also indirectly affect prey abundance through their direct effects on smaller predators. In Australia, dingoes fill the role of a large predator and, in southern Australia, have clear impacts on introduced foxes. Their effect on introduced cats, however, is less clear. Here we present data from multiple sites across northern Australia (where foxes are absent), which reveal a negative correlation between cat and dingo activity.This relationship could arise because cats avoid areas where dingoes are active, or because cats are less abundant in areas with high dingo densities, or a combination of both. At a subset of our study sites, we experimentally reduced dingo (but not cat) abundance by poison baiting. This resulted in a 55% drop in dingo activity within 4 weeks of baiting, but without a compensatory increase in cat activity. This suggests the negative correlation between cat and dingo activity is not a simple consequence of cats reactively avoiding areas with higher dingo traffic, but rather, that there are fewer cats in areas where dingoes are more active.This study is a rare demonstration of the potential for dingoes to affect the behaviour and potentially the population size of feral cats, and therefore reduce the impact of feral cats on vulnerable native prey species.

Cover provides shelter, food, nesting opportunities and protection from predators. The behavioural response of small mammals to reduced cover has been well documented. However, very little is known about the effect of cover on community and population dynamics. Australian small mammals generally inhabit extremely dynamic ecosystems, where cover and food supplies are greatly affected by fire. Species are described as early or late seral specialists, generally returning to a disturbed area once their habitat requirements are met. Habitat requirements have loosely been interpreted as cover and food supply, however, these factors are not mutually exclusive and few studies have attempted to determine the driving factors behind small mammal succession. In this study, we manipulated specific aspects of cover in the eucalypt forests of Fraser Island and show that the behaviour and population dynamics of small mammals were greatly affected. A reduction of cover from grass-trees ( Xanthorrhoea johnsonii ) did not affect small mammal species composition, however, the abundance and size structure distribution of the dominant species ( Rattus fuscipes ) decreased. Patch use by rodents also decreased after cover was reduced. Rattus fuscipes must trade-off remaining in an environment with increased risk of predation, or disperse to an area with greater cover but increased competition. Juveniles dominated ( > 60%) populations of R. fuscipes after cover was reduced, however, size distributions of control sites were relatively more even ( < 25% juvenile). While adult R. fuscipes are either killed by predators or disperse to other areas, juveniles that remained or immigrated to an area of reduced cover gained a selective advantage over those in control sites, because reduced competition with adults increased body condition of juvenile R. fuscipes .