Distinctive inter-annual patterns of tree seed production can include spatial synchronicity, periodicity, and high variability among individuals within a population. Synchronicity and high variability are now commonly used to define mast seeding, with ‘strict’ mast seeding further distinguished by annual seed production that is either often large or nil and thus bimodal. Here we test for synchronicity, periodicity, and bimodality using 43 years of annual total and viable seed counts, along four transect lines, sampling an elevation gradient (480–1340 m) in a monospecific New Zealand mountain beech Nothofagus solandri var. cliffortioides forest. We expect most support for periodicity and bimodality at relatively high elevation sites that are most resource limited and known to have the greatest variability in seed production. While there was weak evidence for differences in viable seed counts along the elevation gradient, this was not the case for total seed counts. A significant year-effect on seed counts provided some evidence for synchronization, but the correlations of total or viable seed counts rapidly declined (from 0.96 to 0.58) with increasing elevation difference among transect lines. Although we detected a seven-year periodicity in total and viable seed counts at each elevation, we also detected other period lengths at most elevations. We did not find evidence for our expectation of increased period length and bimodality in relatively unproductive high elevation mountain beech forests because they would take more time to recover from seeding events. As a consequence, if resource limitation is an important driver of mountain beech seed production its influence must be strongly regulated by other factors to determine the distinctive characteristics (periodicity and synchronicity but not bimodality) of inter-annual variation in seed production.