Populations occurring in similar habitats and displaying similar phenotypes are increasingly used to explore parallel evolution at the molecular level. This generally ignores the possibility that parallel evolution can be mimicked by the fragmentation of an ancestral population followed by genetic exchange with ecologically different populations. Here we demonstrate such an ecological vicariance scenario in multiple stream populations of threespine stickleback fish divergent from a single adjacent lake population. On the basis of demographic and population genomic analyses, we infer the initial spread of a stream-adapted ancestor followed by the emergence of a lake-adapted population, that selective sweeps have occurred mainly in the lake population, that adaptive lake–stream divergence is maintained in the face of gene flow from the lake into the streams, and that this divergence involves major inversion polymorphisms also important to marine-freshwater stickleback divergence. Overall, our study highlights the need for a robust understanding of the demographic and selective history in evolutionary investigations.

Speciation can be promoted by phenotypic plasticity if plasticity causes populations in ecologically different habitats to diverge in traits mediating reproductive isolation. Although this pathway can establish reproductive barriers immediately and without genetic divergence, it remains poorly investigated. In threespine stickleback fish, divergence in body size between populations represents a potent source of reproductive isolation because body size often influences reproductive behavior. However, the relative contribution of phenotypic plasticity and genetically based divergence to stickleback body size evolution has not been explored. We here do so by using populations residing contiguously in Lake Constance (Central Europe) and its tributaries, a system where lake fish exhibit strikingly larger size and greater age at maturity than stream fish. Laboratory experiments reveal the absence of substantial genetic divergence in intrinsic growth rates and maturation size thresholds between lake and stream fish. A field transplant experiment further demonstrates that lake fish display the life history typical of stream fish when exposed to stream habitats for one year, confirming that life history divergence in this system is mainly plastic. This plasticity appears to be driven by restricted food availability in the lake relative to the stream habitat. We thus propose that in this stickleback system, the exploitation of different trophic niches immediately promotes reproductive isolation via resource-based plasticity in life history.

How rapidly natural selection sorts genome-wide standing genetic variation during adaptation remains largely unstudied experimentally. Here, we present a genomic release-recapture experiment using paired threespine stickleback fish populations adapted to selectively different lake and stream habitats. First, we use pooled whole-genome sequence data from the original populations to identify hundreds of candidate genome regions likely under divergent selection between these habitats. Next, we generate F2 hybrids from the same lake-stream population pair in the laboratory and release thousands of juveniles into a natural stream habitat. Comparing the individuals surviving one year of stream selection to a reference sample of F2 hybrids allows us to detect frequency shifts across the candidate regions toward the genetic variants typical of the stream population-an experimental outcome consistent with polygenic directional selection. Our study reveals that adaptation in nature can be detected as a genome-wide signal over just a single generation.

The recombination rate around the chromosomal inversion in Fig. 6f of this Article was inadvertently omitted during the production process. The correct version of Fig. 6f appears below.