Many mammal species depend on tree cavities for shelter and for breeding. Some species may use tree cavities as well as other varied shelters but the extent of their dependence on tree cavities is unknown. One such species is the eastern pygmy-possum (Cercartetus nanus) from eastern Australia which typically shelters alone. I addressed four key questions: i) do different age and sex classes differ in their preferences for small (4 cm diameter) and large (10 cm diameter) cavities, ii) does the availability of large cavities influence female breeding status, iii) does the availability of large cavities influence where adult females and males are detected, and iv) does the availability of large cavities influence survival and abundance? I provisioned two 9-ha plots containing few (≤2 per ha) ‘large’ natural cavities with artificial cavities in different years. When plots were provisioned with small and large cavities breeding females (n = 46) showed a pronounced preference for sheltering in the large cavities whilst adult males (n = 67) and subadults (n = 40) preferred the small cavities. When one plot and not the other was provisioned with large cavities adult females were significantly more likely to be reproductive and be detected on the plot with large cavities, but survival was not influenced. Detection of adult males did not vary with large cavity provisioning. The abundance of adult pygmy-possums varied among years, being higher on the plot provisioned with large cavities when the other plot was not provisioned, but equivalent when both were provisioned. The findings of this study are consistent with the hypothesis that breeding females are dependent on a large cavity to raise a litter. Other mammal species around the world are likely to exhibit similar but currently unrecognised shelter site strategies which have implications for their management.

The data set is based on the capture of animals within installed nest boxes in Royal National Park, Sydney, Australia.

There are four sets of data. These relate to four questions addressed in the results of the paper. Q1 describes cavity preference for small or large installed cavities by different age-sex classes; Q2 are the data used in a logistic regression analysis to investigate whether adult females were detected as breeding or non-breeding individuals on one of two study plots in which large cavities (nest boxes) were installed in some years but not others; Q3 are the detection history data for females and males in which one study plot and not the other contained nest boxes with this experiment reversed across the two years; Q4 contains the capture histories of adult females on the two study plots across three years with nest boxes present in two of the three years but the timing differing across the two plots.