95th ESA Annual Meeting (August 1 -- 6, 2010)

COS 85-2 - Temporal turnover and local affinities of the bees of Gardiners Island, New York

Thursday, August 5, 2010: 8:20 AM
410, David L Lawrence Convention Center
Sarah Kornbluth, Biological Sciences, Rutgers University, Newark, NJ
Background/Question/Methods

Information about the distribution of bees is critical to understanding the health of bee communities, as well as the pollination services they provide, given current trends of climate change and habitat loss. Our primary data come from Gardiners Island, Suffolk County, New York. Gardiners Island is one of the only relatively undisturbed coastal island environments in New York State. Bee specimens were net collected during 1976, 1977, and 2005-2009, with 147 species recorded and just one additional species, Bombus pensylvanicus, occurring in historical collections. First New York State records were made for Colletes speculiferus, Hylaeus schwarzii, and Perdita bradleyi, known ranges of other species were expanded and co-occurrence provided evidence of specific Andrena-Nomada host-parasite relationships. Temporal species turnover was assessed within and across years, and spatial turnover by comparing the collection to those made at other local sites with different habitat types (Black Rock Forest, NY, for example). Rarefaction was used to correct for differing sample size.
Results/Conclusions

Collectors curves and estimation of true richness suggests that our sampling is relatively complete. Comparing community turnover between the same two-week intervals in adjacent years and different intervals in the same year suggests that the sequence of bee emergence in each year can vary by a few weeks. Time-shifting each year to align the communities for maximum similarity reduced the year-to-year turnover and confirmed that the communities in each year follow a very similar sequence. Comparing the Gardiners Island communities to those in different local habitats showed that between-habitat differences are much larger than the between-year differences at any one site. Comparison of the most common species at each site shows distinct communities. The known resource preferences of the species present correlate well with the communities observed adding to specific habitat types known to support significant pollinator species.