Do replicates of independent guppy lineages evolve similarly in a predator-free laboratory environment?

Revisiting old truths: The evolution of male coloration in guppies as a function of predation

The Trinidadian guppy, Poecilia reticulata, is a long-studied model for the evolution of trade-offs in male nuptial coloration as a function of female mating preferences versus predation risk. Previous work suggests that female mating preferences favor the evolution of increased conspicuous male coloration in low predation populations. In contrast, high predation risk shifts the balance towards reduced conspicuous coloration. In a From the Cover article in this issue of Molecular Ecology, Yong, Croft, Troscianko, Ramnarine, and Wilson (2021) use visual detection models to estimate the 'conspicuousness' of male color patterns as seen by guppies and their predators. The study fails to find robust patterns of increased conspicuousness in low predation populations. Only one of eight measures of conspicuousness showed parallel changes between high and low predation regimes, forcing us to re-consider the validity and repeatability of this classic example of parallel evolution.

Sensory-based quantification of male colour patterns in Trinidadian guppies reveals no support for parallel phenotypic evolution in multivariate trait space

Parallel evolution, in which independent populations evolve along similar phenotypic trajectories, offers insights into the repeatability of adaptive evolution. Here, we revisit a classic example of parallelism, that of repeated evolution of brighter males in the Trinidadian guppy (Poecilia reticulata). In guppies, colonisation of low predation habitats is associated with emergence of 'more colourful' phenotypes since predator-induced viability selection for crypsis weakens while sexual selection by female preference for conspicuousness remains strong. Our study differs from previous investigations in three respects. First, we adopted a multivariate phenotyping approach to characterise parallelism in multitrait space. Second, we used ecologically-relevant colour traits defined by the visual systems of the two selective agents (i.e., guppy, predatory cichlid). Third, we estimated population genetic structure to test for adaptive (parallel) evolution against a model of neutral phenotypic divergence. We find strong phenotypic differentiation that is inconsistent with a neutral model but very limited support for the predicted pattern of greater conspicuousness at low predation. Effects of predation regime on each trait were in the expected direction, but weak, largely nonsignificant, and explained little among-population variation. In multitrait space, phenotypic trajectories of lineages colonising low from high predation regimes were not parallel. Our results are consistent with reduced predation risk facilitating adaptive differentiation, potentially by female choice, but suggest that this proceeds in independent directions of multitrait space across lineages. Pool-sequencing data also revealed SNPs showing greater differentiation than expected under neutrality, among which some are found in genes contributing to colour pattern variation, presenting opportunities for future genetic study.