Mesopredator and competitor release can lead to population increases of invasive house mice (Mus musculus) after larger introduced mammals are controlled or eradicated. In New Zealand, mammal-resistant fences have enabled multi-species mammal eradications in order to protect indigenous species. When house mice are the only mammals remaining in these biodiversity sanctuaries, they may reach a high population density, with potential consequences for their indigenous prey. We studied mouse populations in the absence of other mammals for 5 years at mammal-resistant fenced forest sites at Maungatautari, Waikato. We used spatially explicit capture-recapture (SECR) to estimate mouse population density quarterly in two independently fenced sites, with contrasting levels of mouse management that were switched half-way through the study. In the absence of mouse control, mouse population density reached 30-46 ha -1 at one site each year after summer breeding, and 23 ha -1 at the other site. Mouse tracking rates in inked footprint tunnels were positively related to numbers of mice captured in each session, but not significantly to mouse density. The highest mouse densities were similar to estimates in New Zealand forest and alpine ecosystems after mass seeding (masting) events, but lower than estimates in another sanctuary and on some islands lacking larger terrestrial mammals. We suggest that in the absence of competition and predation from other mammals, food limitation may have prevented mouse populations from attaining very high densities in this mainland forest location.