Ecological patterns in plant defence chemistry and herbivore responses in natural populations of Brassica oleracea

Relationships between two taxonomic kingdoms; plants and herbivorous insects, are hypothesized to be a major zone of interaction for generating current day biodiversity; and coevolutionary processes between these intricately linked organisms are hypothesized to maintain diversity in plant secondary chemistry. These metabolites play a key role in plant defence against herbivory and a high degree of intraspecific variation is observed at multiple ecological scales. However, the nature of selection maintaining variation in plant defence profiles is still a major question in evolutionary biology and ecology, and progress towards a deeper understanding is hampered by a lack of studies that take into account ecological context and the multivariate nature of plant defence phenotypes. In this thesis, I employ sophisticated chemical analysis techniques to identify a suite of glucosinolate secondary chemicals, representing different biosynthetic pathways, in the wild cabbage, Brassica oleracea, in natural populations in the UK. I used model-based cluster analysis to explore patterns of association between individual glucosinolates, predicting that as simultaneous resource allocation to multiple defences is likely to be constrained; negative associations between defensive traits should be observed. However, results revealed positive associations between glucosinolates. Therefore co-expression of multiple defences may not be costly for this species. Using this information in conjunction with herbivore surveys and experiments, I show that this mixture has the potential to shape patterns of herbivore abundance and host plant utilization: species-specific responses to variation in glucosinolate phenotypes are discovered at various ecological scales. Thus there is the potential for differential selection on plant chemotypes though species-specific attractions and aversions. By conducting fine scale experiments with herbivore species, I also found that glucosinolate variation has an impact on the counter-adaptations that some brassica specialists have evolved: in order to optimally defend against their own natural enemies, Brevicoryne brassicae aphids sequestering glucosinolates from their host plants must do so selectively, and must choose plants whose chemical profile best matches this behaviour. These findings show that glucosinolate profiles may be under natural selection by herbivores in wild populations, and that reciprocal evolution between these plants and their specialists may continue to promote diversity in secondary metabolites. Together these results highlight the complexity inherent in plant-insect interactions, the importance of field studies and generate a wealth of testable hypotheses for future work.