OOS 18-5
An experimental test of gibbon and bird seed dispersal patterns

Wednesday, August 7, 2013: 9:20 AM
101F, Minneapolis Convention Center
Jackson L. Frechette, School of Natural Resources and Environment; Department of Wildlife Ecology and Conservation, University of Florida, Gainesville, FL
Douglas J. Levey, Population and Community Ecology Cluster, National Science Foundation, Washington, DC

Seed dispersal is essential for plant recruitment. Despite much attention to animals as seed dispersers, very few studies have been able to link spatial patterns of animal-generated seed shadows to patterns of tree regeneration. We asked: How do species of seed dispersers differ in the seed shadows they create, and are those differences reflected in spatial patterns of tree regeneration? Focusing on a common forest tree (Microcos paniculata) in northeastern Cambodia, we quantified dispersal kernels created by the tree’s major dispersers -- gibbons (Nomascus annamensis) and four species of bulbul (Pycnonotidae). We experimentally sowed 9360 Microcosseeds to match kernels created by those two groups of dispersers and a control (no dispersal). To determine if there were any differences in seed germination and seedling survival between the three types of kernels, we checked seed addition plots one, three, six and eleven months after sowing. 


After eleven months, 0.3% of sown seeds in the gibbon dispersal kernel, 0.9% of those in the bulbul kernel, and 1.4% of control seeds had germinated and were alive as seedlings. Preliminary analyses indicate higher mortality of seeds and seedlings at further distances from source trees. This pattern is the opposite of that predicted under the Janzen-Connell hypothesis and observed in other systems. Our data suggest that habitat near adult Microcos trees is most suitable for seed germination and seedling survival, despite presumed density-dependent mortality. We suggest that Microcos relies on seed dispersers to move seeds to novel areas and provide gene flow between sub-populations rather than to avoid distance and density dependent mortality near parent trees.