Abstract: The genus Nothofagus in New Zealand and Australia exhibits strong mast seeding (i.e. highly variable seed crops between years). Seed crop variation is synchronized within and between species over large spatial scales, and results in greatly increased wind pollination efficiency which could provide a selective benefit favouring the maintenance of mast seeding. However, the null hypothesis (that plants simply match their reproductive effort to the variable resources available each year) has not been tested in Nothofagus. Here we use a 33-year dataset on seedfall and wood ring increments for 19 individual Nothofagus truncata trees at Orongorongo, New Zealand, to test for the presence of switching (exaggeration of seedfall variability by diverting resources into, then out of, reproduction). A generalized least squares model explained 40.7% of the variance in standardized ring widths, using six weather variables (absolute minimum temperatures in March (lag 0) April (lag 0 and lag 1), May (lag 0) and rainfall in November and February (lag 0) ) and seedfall. Seedfall had a negative relationship with the current year's ring widths even after controlling for all significant weather variables. This shows that switching is occurring in N. truncata within individuals among years, and therefore that masting in this species is the result of selective forces such as increased wind pollination efficiency. As this result has been demonstrated for very few masting species, we call for this test to be applied more widely.