The role of predators in ship rat and house mouse population eruptions: drivers or passengers?

Four hypotheses regarding the role of predation in the population dynamics of eruptive small mammal communities were tested using the small mammal assemblage found in mixed forests in New Zealand. Large‐scale (750 ha) predator removal was conducted, targeting stoats ( Mustela erminea ). House mouse ( Mus musculus ) and ship rat ( Rattus rattus ) population dynamics during an eruption were compared in areas with and without predator reduction. The success of predator reduction was measured by comparing live‐capture rates of predators on treatment and non‐treatment areas, and by recruitment rates of the threatened northern brown kiwi ( Apteryx australis mantelli ). Overall, predator reduction was successful, although there was a continual low rate of reinvasion. The predictions and results were that 1) Predators can slow but not prevent a population eruption. Supported: Populations of mice and rats erupted to high densities in areas with and without predator reduction, following synchronous southern beech ( Nothofagus spp.) seeding. 2) Predators cannot truncate peak prey population size. Supported: Peak densities of mice and rats were not significantly different between treatment and non‐treatment areas. 3) Predators can hasten the rate of decline in prey populations during the crash phase. Not supported: There was evidence of populations of mice and rats declining slower in areas with predators removed, but none of the trends were significant. 4) Predators can limit low‐phase prey populations. Equivocal: Populations of rats in beech forest, and population of mice and rats in coastline habitats were significantly higher in areas with predators removed, but were not significantly different in tawa‐podocarp forest. Therefore, the role of food in driving the early stages of the mouse and rat eruption was demonstrated, but the role of predation in the decline and low phases is unclear.