Abstract
Polyphagous herbivores are generalists able to exploit a large range of host-plants. In this thesis, using experimental evolution, I study the evolution of life history traits in different populations of a crop pest, the spider mite Tetranychus urticae, after colonization of a new host, the tomato plant. In the first two chapters, measures of life history traits were collected from the egg stage to the death of individuals allowing the precise description of the life cycle of mites in response to the exploitation of tomato plants. As described chapter 1, I found that juvenile survival on the new host was influenced by maternal effects. In particular, mothers having developed on tomato had offspring surviving better on this host as juvenile. However, these juveniles left more often the plant as adults and had a reduced fecundity (for females). Moreover, the evolutionary history of populations affected the proportion of females produced by mothers, potentially enhancing demography of populations having previously evolved on this host. In chapter 2, I show that several life history traits (juvenile survival, male adult lifespan and fecundity) had evolved in all populations exposed to tomato for several generations relative to control populations evolving on bean. Nevertheless, the study of female lifetime reproductive success integrating the effect of several successive traits only showed, paradoxically, an adaptive signal for populations having evolved on this host for approximately 38 generations and not for populations having evolved on tomato for approximately 78 generations. The study of genetic diversity with microsatellite markers suggests that these latter populations may have suffered from bottlenecks, which could have compromised their adaptation. Finally, as shown chapter 3, I test whether populations newly adapted to tomato are more tolerant to the host-plant defenses, induced by herbivore attacks, and/or whether these populations differentially trigger the defences of damaged plants, affecting the performance of mites that subsequently colonize this host. For that purpose, I recorded the fecundity and mortality of female adult mites put on clean tomato plants, or on tomato plant pre-infested by tomato-adapted mites or by mites not adapted to this host. Populations adapted to tomato had a higher fecundity on this host irrespective of the plant treatment, while mortality increased for all populations on plants pre-infested with tomato-adapted mites. My results thus suggest that evolution on tomato leads to an increased capacity to tolerate host-plant defences and also to a higher induction of such defences. My studies describe a continuum of evolutionary responses in a polyphagous species from failure of adaptation, leading to the eventual extinction of populations, to the presence of adaptive plastic responses and finally to host-plant adaptation via the possible acquisition of higher tolerance to host-plant defenses. My work also underlines the relevance of integrating the effects of different life history traits with contrasted variations within a same fitness measure to describe host-plant adaptation. Further investigations in this direction should enrich our understanding of mechanisms of herbivore persistence on a new host-plant but also allow developing better crop pest management strategies.