PARENT SESSION
Symposium #17: Specialization and Generalization in Plant-Pollinator Interactions.
Organized by: NM Waser, J Ollerton, and R Zamora
Wednesday, August 7. 8:00 AM to 11:30 AM. Crystal Ballroom, TCC.
Extreme generalization in pollination systems confirmed by null models.
Vázquez, Diego*,1, Aizen, Marcelo2, 1 University of Tennessee, Knoxville, Tennessee, vazquez@utk.edu2 Universidad Nacional del Comahue, Bariloche, Río Negro, Argentina, marcito@crub.uncoma.edu.a
ABSTRACT- Several recent studies have analyzed patterns of specialization in pollination systems, suggesting that plant-pollinator interactions may be more generalized than previously thought. A problem with this proposition is the lack of null expectations against which to compare the observed patterns. Without such expectations, it is not possible to answer whether the number of specialists or generalists in a given community is unusually high or low. We carried out quantitative tests of community-wide patterns of specialization in pollination systems using null models. We developed computer algorithms to randomize plant-pollinator interactions. For each of four published datasets of plant-pollinator interaction webs, we compared the patterns arising from the randomized communities with those observed in the original datasets. In the four communities, the observed numbers of extreme specialists and extreme generalists were significantly higher than those expected in randomly-assembled plant-pollinator webs. However, this pattern could be partly attributed to a sampling artifact. In the four datasets the observed frequency of interaction of a species was related to its estimated degree of specialization, so that most frequently interacting species appear more generalized than less frequently interacting species. After statistically correcting for this bias, we found that the number of extreme specialists did not differ from null expectations, whereas the number of extreme generalists was still significantly higher than null expectations in all but one case. Thus, our quantitative results support recent arguments about the generalized nature of plant-pollinator interactions.
KEY WORDS: specialization, generalization, null models, plant-pollinator interactions