Comparative seed shadows of animal- and wind-dispersed species in a karst forest in Taiwan

The spatial distributions of seeds dispersed from a parent plant, known as seed shadows, translate spatial patterns of parent plants to those of their offspring, thus critically influencing ecological processes of plant populations and communities.  Seed shadows vary greatly among species.  Such variation may, to a certain extent, reflect the variation in dispersal agents. Our objective of this study is to compare seed shadows of animal- and wind- dispersed species in a karst forest in Taiwan. We set up 72 seed traps (0.5 m²) along 4 transects within a 10-ha permanent plot (400 x 250 m) in Kenting in summer, 2006.  These seed traps have been checked weekly since August 2006. We choose one animal-dispersed species, Aglaia formosana, and its wind-dispersed counterpart, Radermachera sinica, based upon density of seeds and fruit characteristics.  Applying the inverse-modeling approach, we develop seed dispersal functions via maximum likelihood methods.  Exponential models with Poisson or negative binomial errors are used.  The Akaike’s Information Criterion (AIC) is used to determine the best-fit model.  From August to November, 2006, mean number of seeds per trap is 5.3 for Aglaia formosana and 20.0 for Radermachera sinica.  The results of maximum likelihood methods indicate that Radermachera sinica, the wind-dispersed species, has higher fecundity and longer dispersal distances.  Furthermore, models with negative binomial errors fit better than Poisson errors for both species. In the future, we will conduct a community-wide comparison in dispersal functions to determine to which extent differences in dispersal functions can be explained by the variation in dispersal agents.