PREDICTING THE INCIDENCE OF MOHUA PREDATION FROM THE SEEDFALL, MOUSE,AND PREDATOR FLUCTUATIONS IN BEECH FORESTS

Predator control will be required to save many mohua (Mohoua ochroceph ala) populations from extinction. However, control may be required onl y in years when stoat (Mustela erminea) densities are high. To manage local stoat populations effectively, a reliable predictor of high risk years is required. We examined whether different levels of beech seed fall and mouse capture rates were related to the levels of mohua preda tion recorded in the Hawdon Valley, Arthur's Pass National Park, and t he Eglinton Valley, Fiordland National Park, between 1989 and 1994. Du ring this period there was only one full beech mast year in each study area during autumn. The full mast seedfall in Hawdon Valley was predo minantly of mountain beech (Nothofagus solandri var. cliffortioides) a nd red beech (N. fusca), and in Eglinton Valley it was predominantly s ilver beech (N. menziesii). During the following summer, mouse and sto at densities, and the predation rate of adult mohua, all increased con siderably. There was very little predation on adult mohua in the summe rs following poor seedfalls when mouse and predator densities remained low. In 1993, a partial mast did not trigger a mouse or stoat irrupti on. We conclude that counts of beech seedfall and indices of mouse den sity are potential predictors of an impending irruption of key predato rs. Winter mouse density appeared to be the most reliable indicator, b ecause neither stoats nor mice respond to seedfall alone. A combinatio n of these indicators could be used as a basis for management decision s on whether to undertake stoat control to protect mohua populations i n the future. However, more information is required on the seedfall th resholds that may trigger sufficient increases in mouse and stoat numb ers and, consequently, bird predation.