Many factors influence the discovery, consumption, or removal of seeds from the vicinity of parental plants. During the summer of 2007, we manipulated local seed abundance and seed density at habitat points to examine how these and parent density affected consumption or removal of seeds of eastern leatherwood (Dirca palustris), an understory shrub of temperate forests in Eastern North America. Experiment 1: We selected forested sites with leatherwood shrub densities near either 100/Ha or 1500/Ha. We then collected all leatherwood seeds just before abscission to produce plots with low ambient seed density, or augmented natural seed crops to produce plots with high ambient seed density. In each plot we placed a grid of 6 trays with either 30 seeds/m² or 150 seeds/m²; after 1 month, we counted intact seeds remaining in each tray. We also trapped and marked rodents within each plot. Experiment 2: We selected 24 leatherwood shrubs isolated from conspecifics by at least 10 m. We then produced 24 seed shadows, flagged sites at least 10 m from a leatherwood shrub, and randomly assigned each the canopy area of one of our real shrubs. We placed ambient seeds at 150/m² under the canopy of half the shrubs and at corresponding seed shadows, the remaining shrubs and shadows had no seeds. We placed dishes with either 450 seeds/m² or 150 seeds/m² either under the canopy or one canopy radius beyond the edge of real shrubs or seed shadows.

Results/Conclusions

Experiment 1: Shrub density, ambient seed density, and tray seed density interacted significantly to influence seed removal — only trays with low seed density in areas of low shrub density with seeds removed retained a significant number of intact seeds. Our treatments did not appear to influence local abundance of granivorous rodents (primarily Peromyscus spp.). Experiment 2: Seeds disappeared more rapidly near shrubs than near seed shadows. Addition of ambient seeds generally accelerated seed removal around seed shadows, but had no consistent effect near real shrubs. Neither distance from a shrub or shadow nor dish seed density had consistent effects on seed removal rate. Our results suggested that a) major seed predators or dispersers of leatherwood (Peromyscus spp.) appear to use both the density of parental plants and seeds as foraging cues, b) survival of leatherwood seeds is very low, and c) seed survival can be enhanced significantly by dispersal to areas with neither extant shrubs nor seeds.