Temporal variability and local adaptation

Abstract

Gene flow is often regarded as a constraining force because it can counteract selective forces that lead to local adaptationl. Local adaptation (adaptive deme formation) can be studied relatively easily in the context of host-parasite and plant-herbivore interactions through transplant experiment+4. The presence or absence of local adaptation can then be related to the dispersive ability of the herbivore or parasite. In a recent TREE review, Mopper4 discussed transplant experiments that have been done with phytophagous insects. Contrary to her expectation, insect mobility seems not to be related to the finding of deme formation (see Table 1 in Ref. 4). In particular, some sessile insects do not present adaptive deme formation, while some highly dispersive species do. We present some arguments that could contribute to explaining this discrepancy, by taking the host’ s and herbivore’ s population structure into account. The main point in Mopper’ s article is that deme formation of phytophagous insects is adaptive because host populations are highly heterogeneous with respect to their quality as ‘ insect habitat’ . Mopper points out that high levels of migration should prevent the parasite from being locally adapted. Indeed, migration is maladaptive when the environment is variable in space5,6 but, when the environment is variable in time this is not necessarily the case. Environmental variability in time may occur because of various factors. First, host plants are ephemeral and therefore herbivore population extinction would be unavoidable unless migration occurs at sufficiently large rates. As a consequence, herbivore demes must be seen as part of a metapopulation in which each host plant is an ephemeral habitat. Large habitat extinction rates would select for relatively large migration rates’ . Second, within-host growth of herbivores may lead to a decrease in habitat quality (host-plant saturation) and to competition between close relative@. Therefore, dispersing insects have a chance to find a new under-exploited host plant and to compete against non-relatives. Thus, the rate of habitat loss (resulting from host death and from host-plant saturation) relative to the parasites’ generation time and kin competition are important variables to consider when assessing migration. The important point here is that within-host competition selects for local adaptation, since the best competitors are those that are locally adapted. Thus, the evolution of migration rates and of local adaptation depend interactively on each other, but are both dependent on other factors as well, such as host demography and host population structure. In this respect, the absence of deme formation in sessile insects as discussed by Mopper, could be the result of low within-host competition and low host-plant death rates. Deme formation of highly mobile insects could then be the result of strong competition among host plants. A further factor contributing to environmental variability in time are the natural enemies of the herbivores (entomophagous predators or parasites), as discussed by Mopper4. If we assume coevolution between the insect and its enemies, then a given genotype that is selected for in a given site at a given time will be selected