Evaluation of secondary dispersal in a large-seeded tree Aesculus turbinata: A test of directed dispersal

Among the several hypotheses on selective advantage of seed dispersal, the directed dispersal hypothesis explains the advantage of non-random seed transportation by animals to particular patch type suitable for offspring establishment. We tested this hypothesis in dispersal of a large-seeded, rodent-dispersed tree (Aesculus turbinata) in a temperate forest. We investigated the change in location of seeds through secondary dispersal, and the survival and growth of seedlings at their destinations. Hemispherical photographs taken at the seed locations both before and after secondary dispersal were used to evaluate the consequence of dispersal. Survival and growth rates of seedlings were measured to evaluate the responses of seedlings to light, the most important factor for seedling establishment in A. turbinata. Survival and growth rates of the seedlings were both positively correlated with light conditions, indicating the advantage of dispersal to the microsites with more light available. However, light levels at seed destinations were not significantly different from those at the locations of seeds before secondary dispersal nor those of the surrounding background forest floor. Survival of newly-emerged seedlings varied as a function of light level but not seedling density. This suggests that the effect of density-dependent mortality was small relative to light-dependent mortality during the seedling stage. Therefore we conclude that the directed dispersal hypothesis for this species is rejected, and that the role of rodents in dispersing large seeds secondarily is more important for finding suitable sites merely by enlarging seed shadow (mean dispersal distance = 12.2-44.7 m during the 3 years studied, max. = 41.5-114.5 m) and relatively less important for escaping natural enemies.