The experimental evolution of parasite resistance in wild guppies: artificial selection, resource availability and predation pressure.

In a recent paper, ‘Experimental elimination of parasites in nature leads to the evolution of increased resistance in hosts’, Dargent et al. [1] reported on a transplant experiment using wild guppies, Poecilia reticulata. Fish were collected from a source population in a downstream, high productivity site where they experienced parasitism and predation. They were held in the laboratory and chemically treated four times to remove parasites before being introduced into four guppy-free, upstream, low productivity sites without predators. Fish were collected after one year from two of the introduced populations, and after both one and two years from the other two introduced populations. Along with synchronized collections from the source population, these fish were again chemically treated for parasites and bred under standard laboratory conditions for two generations. The authors then used standard protocols to test whether fish descended from the introduced and source populations differed in their resistance to Gyrodactylus turnbulli, a dominant guppy parasite. If parasite resistance is costly to hosts, theory predicts that relaxed selection pressure will result in the evolution of decreased resistance [2,3], and evidence from laboratory studies supports this prediction [4]. Surprisingly, Dargent et al. [1] found that the introduced populations had repeatedly evolved greater resistance compared to the source population. After considering various alternative methodological and biological explanations, Dargent et al. [1] suggest that, ‘relaxed selection for defence against parasites (owing to parasite removal) can be overpowered by the evolution of a slower life history (owing to predator removal) that, through pleiotrophic or functional associations, leads to increased resistance’. While I agree that this is a reasonable explanation, I believe other factors may have contributed to the intriguing result presented by Dargent et al. [1]. Here, I consider additional methodological then biological factors not discussed by the authors. I conclude by suggesting an experimental design that would help elucidate the effect experimental elimination of parasites may have on the evolution of parasite resistance in guppies. As part of their methodology, Dargent et al. [1] may have imposed artificial selection for parasite resistance. Capture, laboratory conditions and chemical treatment may have caused mortality among both stress-prone individuals, which are likely to be less resistant to Gyrodactylus infection [5], and those fish already weakened by the heaviest Gyrodactylus infections. Whilst the authors state that ‘all fish were monitored to ensure that they were in good health’, they do not provide mortality data. In a mesocosm experiment, van Oosterhout et al. [6] report 14% mortality among their captured wild fish, 24% of which was after re-introduction. If, as one could predict, there was higher mortality among stress-prone, less resistant fish either before or soon after the introductions by Dargent et al. [1], or, as is implied, the authors only released fish in ‘good health’, these introduced populations would have been a non-random sample of the more resistant fish from the source population. Even in the absence of artificial viability selection, this process may have reduced the contribution of stress-prone, less resistant individuals to the next generation. Although fish from the source population were exposed to similar conditions before being transferred to the laboratory during the final collection, the effects of this artificial selection are