TY - DATA T1 - Going over the edge: Why are there more thrips per flower when there are more flowers in a patch? PY - 2014/02/11 AU - Russell Dinnage UR - https://figshare.com/articles/poster/Going_over_the_edge_Why_are_there_more_thrips_per_flower_when_there_are_more_flowers_in_a_patch_/928651 DO - 10.6084/m9.figshare.928651.v1 L4 - https://ndownloader.figshare.com/files/1379316 KW - mathematical modeling KW - Thrips KW - Thysanoptera KW - animal movement KW - mcmc KW - simulated annealing KW - habitat size KW - island biogeography KW - patches KW - Ecology N2 - Poster I presented at Canadian Society for Ecology and Evolution meeting in 2011. Still haven't published the study, which I conducted as an undergraduate during a field course in Southern Ontario, Canada. I will post the raw data soon. I would be happy to answer any questions about it. Abstract Many organisms occur in habitat which is naturally fragmentary or patchy, and possibly increasingly so, due to human induced changes. It is therefore important to understand the impact of patch properties such as size on population dynamics of individual organisms. Thrips (Thysanoptera) are a common and widespread generalist herbivore which most often feed in flowers, a naturally patchy resource. I looked at whether the density of a common thrip species in Oxeye Daisy (Asteraceae: Chrysanthemum leucanthemum) flowerheads was related to the size of a daisy patch, measured as the number of flowerheads contained within it. I counted all thrips found in 10 flowerheads at both the centre and edge of 15 patches of varying sizes. Thrips density was positively related to patch size for both central and edge flowers, but the relationship was significantly weaker for the edge. To see whether this relationship could be explained by random movement alone, I developed a simple movement simulation, and fit its results to the data from daisy patches. The model fit well and reproduced many features of the data. It did however, consistently underpredict the densities at patch edges, suggesting that somewhat counterintuitively, densities at the edge may show the strongest signal of any deterministic factors which may increase density as a function of patch size. Overall, the model suggests that a positive relationships between patch size and density should perhaps be considered the null expectation, as opposed to a lack of a relationship, as is more often assumed. ER -