Unexpected changes in reproductive barriers between incipient species after experimental evolution in sympatry

Hybridization is generally considered a temporary phenomenon, but it is actually widespread and may last for large time periods between species that stably coexist. Here, to test whether evolving with a closely-related species modifies or maintains partial reproductive isolation, we performed experimental evolution in artificial sympatry vs. allopatry with two closely-related colour forms of spider mites (Tetranychus urticae) that exhibit an asymmetrical pattern of pre-mating isolation despite almost complete post-zygotic isolation. We assessed whether evolutionary changes occurred in traits associated to (i) pre-mating isolation, (ii) post-mating pre-zygotic and early post-zygotic isolation, and (iii) late post-zygotic isolation. Our results revealed that reinforcement did not occur even under forced long-term sympatric evolution. Instead, the strength of some reproductive barriers decreased (e.g., premating isolation and fertilization failure), and some trait changes indicated convergence rather than divergence between species (e.g., mating propensity, latency to copulation). In fact, both types of males showed the same decreased preference for red-form females across generations in sympatry. In line with this, traits underlying fertilization success evolved in the same direction and with similar amplitude in heterotypic crosses and in their homotypic control, as the offspring sex ratio of green-form females decreased in sympatry irrespective of the male they mated with. Finally, other changes in reproductive barriers resulted from trait correlations (e.g., decreased zygote mortality but increased juvenile mortality). Hence, despite very high costs of hybridization, responses occurring following evolution in sympatry were unrelated to selection directly associated to hybridization, but rather the by-product of other evolutionary forces, with cascading consequences for reproductive barriers. In particular, these results support the underappreciated hypothesis that within-species sexual interactions can constrain population diver-gence, or even drive trait convergence between species, thereby playing a role in the maintenance of partial reproductive isolation.