Experimental tests of the spatiotemporal scale of seed predation in mast-fruiting Dipterocarpaceae

Mast-fruiting Dipterocarpaceae exhibit highly synchronous, interspecific seedfall at irregular, multiyear intervals. To investigate how the temporal pattern of seedfall affects dipterocarp seed and seedling survival, in both a logged and a primary lowland tropical forest, we planted Shorea stenoptera Burck seeds in the last three weeks of a 12-wk synchronous dipterocarp seedfall during a major community mast-fruiting event in West Kalimantan, Indonesia. As a result of commercial timber harvest of dipterocarp individuals eight years before, total dipterocarp seed production in the logged site was only 23% of that in the primary forest. At both sites, an average of 35 kg of seed was sown across a large area (≥1 km2) to examine the spatial pattern of seed destruction. During the period in which "natural" community mast seed was available to predators, 92% and 99% of experimentally sown seed escaped predators in logged and in primary forest, respectively. After regional seed resources were exhausted, nomadic vertebrates (primarily the bearded pig, Sus barbatus) were observed in both forest areas, and all ungerminated seed was destroyed. Seed predators arrived earlier in the logged area, before most experimentally sown seed had germinated, and the logged site experienced greater seed loss to vertebrates than did the primary forest. Because nomadic seed predators were absent during peak fruit fall of naturally occurring communities at both study sites, there was no evidence of local predator satiation. Rather, experimentally sown seed escaped predation because of rapid germination before predator arrival, as opposed to being ignored by satiated predators. Seed escape was more dependent on the late arrival of pigs than on the amount of local seed production. There was no significant spatial autocorrelation of seed predation. All remaining seed at the scale of the experiment (>1 km2) was destroyed by predators. These findings suggest that satiation of nomadic predators occurs at the landscape scale. Postdispersal seed predators caused significantly greater seed destruction in the experimentally sown seed populations than in naturally dispersed, mast-fruiting communities at both sites. In both logged and primary forests, there was significantly greater loss of experimentally sown seed to predation than was found in the entire natural mast-fruiting Shorea community combined (21 spp.). Moreover, a naturally occurring, but late-fruiting, Shorea species also exhibited