EXPERIMENTAL STUDIES OF COMMUNITY EVOLUTION II: THE ECOLOGICAL BASIS OF THE RESPONSE TO COMMUNITY SELECTION.

Community selection, defined as the differential proliferation and/or extinction of communities, can bring about a response that may be qualitatively different from the response to selection acting at lower levels. This is because community selection can result in genetic changes in all of the species within the community by acting on the interaction among species. In the experiment presented here, a series of one generation assays were performed on the coevolved communities of two species of flour beetles, Tribolium castaneum and T. confusum, discussed by Goodnight (1990). Two community assays and one single-species assay were performed. Taken together, these provide insights into the genetic basis of the response to community selection. The first community assay involved measuring the selected traits on the original coevolved communities that had been subjected to community selection. This assay indicated that all of the selection treatments resulted in a significant response to selection in the original coevolved communities. The single-species assay involved separating the coevolved communities into their constituent single-species populations and again measuring the selected traits on these populations. None of the single-species populations exhibited a significant response to selection; thus the responses to community selection observed in the first community assay are expressed only in a community context. The second community assay again involved separating the coevolved communities into their constituent single-species populations; however, in this assay a competitor of the opposite species that had never been exposed to community selection was added to each population to form a "reconstructed" community. The results of this assay were that for two traits, emigration rate in T. castaneum and emigration rate in T. confusum, the genetic identity of the competing species did not affect the response to selection. This indicates that the competing species was acting like a nonevolving part of the environment. For the other two traits measured, population size in T. castaneum and population size in T. confusum, the results were very different. For these traits there was no detectable response to selection in the reconstructed communities. This indicates that for these traits the response to selection cannot be attributed to a genetic change in either species independently of the other species in the community. Rather it resides in the interaction between the two species.