The evolution of host specialization: Are trade-offs overrated?

The idea that traits that adapt a population to one habitat or host are deleterious in other habitats or on other hosts is central to most theories for the evolution of ecological specialization. The trade-off concept has played an especially prominent role in discussions of the evolution of host specialization in phytophagous insects. The evidence for genetically based trade-offs in phytophagous insect populations is ambiguous, however. Cross-host genetic correlations for fitness traits are seldom negative, but this does not preclude the existence of trade-offs at a subset of loci controlling the fitness variation. One clear result is that cross-host genetic correlations for fitness traits are often less than one, which implies that genotypes have different fitness rankings on different hosts (i.e., that reaction norms for fitness cross). Using verbal arguments and a mathematical model, I show that crossing of reaction norms for fitness is a sufficient condition for selection to favor specialized host choice, in the absence of search costs and other disadvantages of specialization. In other words, selection will favor specialization if alleles that are positively selected on one host are less strongly positively selected, or neutral, on other hosts; it is not necessary for the alleles to be deleterious on the other hosts. Search costs and other factors may oppose the evolution of specialization, but introducing trade-offs into a model does not result in a quantum jump in the strength of selection favoring specialization. and thus in the likelihood of specialization evolving, compared to the situation in which alleles that affect fitness on one host are neutral on others. There is therefore no justification for focusing on the qualitative presence or absence of trade-offs as the critical issue in predicting or explaining the evolution of specialization. Furthermore, the quantitative genetic data, rather than give little information on why specialization evolves, indicate that the potential for selection to favor specialization exists in many phytophagous populations.