Summary1 Two non-mutually exclusive hypotheses have been proposed to explain the evolutionary advantages of mast seeding (the intermittent production of large crops of flowers or seeds by a population of perennial plants). Mast seeding could have evolved as a result of increased pollination efficiency in mast-flowering years and/or as an anti-predator adaptation that increases the survival of seeds by alternately starving seed predators in non-mast years and satiating them in mast years. 2 We investigated annual seed crops to test the relative contributions of pollination efficiency and pre-dispersal predator satiation to mast seeding in Fagus crenata , a tall tree species dominating cool-temperate forests in Japan. Thirteen-year (1990Thirteen-year ( -2002 time series data were collected for five beech forests in south-western Hokkaido. 3 The negative relationship observed between the pollination failure rate and the total seed crop in the current year supports the pollination efficiency hypothesis. The predator satiation hypothesis was also supported by the fact that the predation rate showed a good fit to the ratio of successive total seed crops, suggesting that a numerical response (starving the predator in low seed years) operated in F . crenata . 4 Key-factor analysis revealed that pre-dispersal seed predation had a larger effect on seed production per flower than did pollination efficiency. 5 We used a simulation model to examine how the magnitude of fluctuation in the total seed crop would influence the pollination failure rate, the predation rate and the viable seed rate. The mean levels of fluctuation of total seed crops of F. crenata were just large enough to provide maximum benefits from predator satiation at some sites. 6 Mast seeding in F. crenata thus appears to be determined by selective pressures from its seed predators.

No abstract

To examine the proximate factors causing mast seeding in Fagus crenata Blume in Hokkaido, northern Japan, we analyzed a 13-year time series of seed production in relation to both previous reproduction and weather conditions. In an autocorrelation analysis we observed a significant negative correlation in 1-year time lags for the log-transformed total seed crop. This indicates that internal resource dynamics are important for mast seeding. A strong negative correlation was observed between the total seed crop and minimum temperature from late April to mid-May in the year preceding flowering. The critical minimum temperature from late April to mid-May for total seed crop at all five sites was about 1.0 °C higher than the 22-year (1979–2000) mean of the minimum temperatures, above which very few seeds were produced. These results show that a weather cue triggers the cessation of reproduction in F. crenata. Regression models that included reproduction in the previous year and minimum temperature explained 57.8%–83.1% of the total seed crop at the five study sites. Therefore, resource dynamics and weather cues are clearly involved in mast seeding in F. crenata.

No abstract