Graduation Year


Document Type




Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Biology (Integrative Biology)

Major Professor

Marc J. Lajeunesse, Ph.D.

Committee Member

Earl McCoy, Ph.D.

Committee Member

Christina Richards, Ph.D.

Committee Member

Peter Stiling, Ph.D.


experimental evolution, meta-analysis, spider mite, Tetranychus, trade-offs


The distribution of insect herbivores among plant hosts is largely nonrandom: most herbivores have limited sets of hosts within one or a few plant families. This host use specialization is reinforced by traits that confer differential fitness across host plant species. Classic explanations for herbivore specialization predict that evolutionary trade-offs reinforce these relationships by imposing costs in the form of reduced potential fitness on alternative hosts, due to negative genetic correlations in fitness across hosts. This prediction that trade-offs constrain host use in herbivores can be tested with experimental evolution, by showing the direct evolutionary effects of host manipulation on population mean fitness across hosts, or alternatively, with quantitative genetic tests based on split-family experiments, which provide evidence for how genetic architecture contributes to cross-host fitness relationships.

To date, selection experiments have demonstrated that herbivores, particularly Tetranychus spider mites, can rapidly evolve increased fitness on initially challenging novel hosts, but that these increases do not result in decreased fitness on the ancestral host. However, effects on alternative hosts (neither the ancestral host or novel host on which selection has occurred) remain unclear, and no studies have investigated the role that host phylogenetic relationships have on mediating these multivariate cross-host effects. Further, split-family experiments reveal that high fitness on one host does not entail low fitness on another. However, the heterogeneity in effects has never been synthesized with meta-analysis.

Here, I used experimental evolution to test how populations of Tetranuchus urticae and T. evansi responded to evolution on novel hosts, in terms of fecundity and development, across multiple host species, including the novel and ancestral hosts. Further, I tested the impact that phylogenetic distance from the novel hosts had on these cross-host effects. I additionally synthesized the results of split-family herbivory experiments to measure the impact that various aspects of herbivore biology had on moderating the sign and magnitude of cross-host genetic correlations in fitness.

In Chapters 2 and 3, I present evidence that increases in fitness observed on novel hosts post-selection do not come with costs to fitness on alternative hosts in Tetranychus spider mites. Contrary to the predictions based on classical explanations for herbivore specialization, these findings indicate that increases in fitness can extend to alternative host species, though the impact of phylogeny on these effects were unclear. These trends are consistent with the results, presented in Chapter 4, of split-family experiments that I synthesized via meta-analysis, which revealed largely positive correlations in cross-host fitness among herbivore genotypes, except for among aphids. Further, I found that important aspects of herbivore biology, such as adult fitness effects, the role of host preference, and complex trade-offs involving multiple traits, were underemphasized in quantitative genetics studies of trade-offs in herbivores, and thus represent significant gaps in research. Overall, my thesis emphasizes the importance of positive cross-host effects of host selection, and points to aspects of herbivore biology other than genetic constraints, such as mating success and predator escape, and the potential interaction between these factors, that are key to explaining the maintenance of genetic variation among herbivore populations.