Artificial cavities (nest boxes, carved and drilled cavities) may be used to restore habitat quality that may otherwise take many decades to recover naturally. We investigated whether artificial cavities could support populations of two species of Australian cavity-dependent mammals, the brush-tailed phascogale (Phascogale tapoatafa, 200 g) and the inland sugar glider (Petaurus notatus, 110 g). We censused animals twice per year over 6 years in 164 nest boxes arranged in 74 clusters through a 600-ha forest block. We addressed four key questions: (1) were nest boxes used by many individuals of the target species, (2) was use sustained over multiple years, (3) were nest boxes used for breeding, (4) what proportion of the local population might be supported by the nest boxes? We detected up to 58 adult phascogales and 232 adult sugar gliders. Use was sustained over time with a mean probability of occupancy per cluster per year of 0.66 in the phascogale and 0.89 in the sugar glider. Differences in life history in these species led to differences in how breeding was assessed. We detected breeding in 98% of female phascogales and at 68% of sites with sugar gliders. We estimated the total number of adults of each species that might occupy our study area based on their known spatial requirements. Our nest boxes potentially supported 31–63% of the adult female phascogales and 50–100% of the adult sugar gliders that could occur in our study area. Such population support could be highly beneficial to manage populations of cavity-dependent mammals.

This study was conducted at Diamond Hill, in a section of Bendigo Regional Park near Bendigo, Victoria, Australia.

Nest boxes (n=164) were arranged in 74 clusters (i.e. sites), that were spaced 100–300 m apart. A cluster contained 1–4 boxes spaced 5–50 m apart. The spacing was partly a consequence of the history of nest box installation in the study area. The clusters were distributed along the major management trails giving broad coverage through the forest. The boxes were of various designs due to the history of installation and construction by different individuals. One-third (0.33) were made of PVC tubing while the remainder were made of ply (0.18), hardwood (0.25) or pine (0.25). Boxes had entrance holes of 3.5–4.5 cm diameter. They were hung on a single large nail at a height of 3–4 m and positioned on the east side of trees to minimise solar radiation in the afternoon.

We timed our nest box inspections to encompass different periods of the reproductive cycles of the brush-tailed phascogale and the inland sugar glider. Nest boxes were inspected twice per year (usually May and October) for 6 years from 2018 to 2023. The phascogale has a semelparous life history with complete male die-off occurring in June in Victoria. Consequently, only adult female phascogales with young are present in October. We chose not to tag any animals, so our inspections consisted mostly of counting individuals within the boxes from a 3 m long ladder. We hand-captured phascogales and sexed them from 2019 onwards to clarify the number of adult males present in May.

We conducted camera-trapping between January 2023 and February 2024. The aim was to investigate the reliability of the assumption that the abundance of our target species was influenced by the nest boxes. We established 10 replicate sites of three different types for monitoring: sites ‘near’ a previously occupied nest box, sites at least 300 m from a nest box site (i.e. ‘far’), and sites within another forest area (Sedgwick State Forest) without any nest boxes. The 10 nest box sites were randomly chosen from 29 used as maternal nests by phascogales in 2022 and selected to provide broad coverage through the study area. All but two were located at least 500 m apart. A camera was installed on a tree within a 10 m radius of the target nest box. The 10 sites ‘far’ from a nest box site were in a randomly chosen direction from a site. A spacing of 300 m was chosen to place the survey site beyond the 4–13 ha area (radius of ~200 m) used intensively by a breeding female phascogale. Sedgwick Forest (approximately 1300 ha in area) is located 4.6 km southeast of Diamond Hill. Sites were randomly located along management trails throughout the forest block. To ensure our monitoring of sites without nest boxes was not biased by a complete absence of tree cavities we searched to locate a tree with a cavity of sufficient size for a phascogale (a single entrance of 2.5–6 cm diameter; at least 10 cm deep) at each of the ‘far’ and Sedgwick (‘no nest box’) sites. Cameras were installed on a tree within a 10 m radius of these trees. Cameras were mounted on metal stakes and directed at a tree 2–3 m away on which a vented canister (PVC tube, 5 x 15 cm) containing bait (peanut butter, oats, maple syrup and fish oil) was installed at 1.5 m high. Cameras were serviced (SD cards replaced; bait renewed) approximately every 2 months. Cameras (Swift Enduro and Swift 3C wide-angle cameras) were set to take 5 images when triggered with a 1 min quiet period between triggers.