Independently evolved sets of sympatric ecotypes sometimes exhibit a high degree of repeatability, suggesting some similarity across assemblages in the underlying processes driving ecotype evolution and coexistence. Fish in post-glacial lakes have repeatedly evolved sets of sympatric, phenotypically divergent ecotypes. In sets of two co-occurring ecotypes, one commonly occupies a pelagic habitat while the other occupies a littoral habitat. Two Salmonid genera, Salvelinus and Coregonus, are unique in that they commonly exhibit assemblage of three or more co-occurring ecotypes in post-glacial lakes. In these genera, we were therefore able to address both whether there is repeatability in two ecotype assemblages, and whether it extends to assemblages with three or more ecotypes. We used a meta-analysis to evaluate patterns of repeatability across assemblages with different levels of diversity in these genera. Specifically, we assessed repeatability of niche occupation, using diet and habitat categories, as well as phenotype distributions across assemblages. We then tested whether ecotypes in particular diet or habitat categories tend to be phenotypically differentiated in a repeatable way.
Results/Conclusions
We found mixed evidence for repeatability in Coregonus and Salvelinus ecotype evolution, with degree of repeatability depending on both genus and the scale at which repeatability was assessed. Coregonus ecotype assemblages showed evidence for repeatability and nestedness in which niche categories were filled while Salvelinus ecotype assemblages were highly variable. In most lakes, with Coregonus ecotypes defined by diet, if there were two ecotypes there were benthivore and planktivore ecotypes, the third ecotype was a second benthivore, and the fourth ecotype was a generalist invertivore ecotype. With ecotypes defined by habitat, two-ecotype lakes generally contained pelagic and littoral/benthic ecotypes, three-ecotype lakes additionally contained a profundal ecotype and the habitat used by a fourth ecotype was variable. In both genera, there was no evidence for repeatability in mean trait values of entire assemblages. There were similar magnitudes of divergence in trait means between ecotypes in several cases both within and across genera. This intermediate degree of repeatability is consistent with some degree of shared ecological, phenotypic, or genetic constraints in different populations. However, ecological context and lineage-specific factors are clearly also important in the evolution of ecotypes in these genera.
