Testing the null model for polyandry: the need to breed explains multiple mating and constrains trading-up

In most species, unmated individuals run the risk of dying with zero fitness. This strong selection on virgin females to mate may also explain why females subsequently remate, despite fitness costs; all that is required is a genetic correlation between virgin and non-virgin mating propensity. Despite being the null model for the evolution and maintenance of polyandry, this hypothesis has received no empirical test. We performed separate quantitative genetic and artificial selection experiments to test the presence of this cross-context genetic correlation in the cow-pea weevil, Callosobruchus maculatus. A quantitative genetic experiment did not find evidence of the hypothesised genetic correlation. However, after 13 generations of artificial selection on virgin mating latency, we found strong evidence for evolutionary divergence in remating latency. Females from lines selected for longer virgin mating latency took approximately twice as long to remate and, were less polyandrous if their virgin mating latency was longer. There was no evidence that females could mate indiscriminately and then trade-up, rather, trading up could only occur if virgin discrimination was present. Selection against virgin death will thus constrain both the evolution of non-virgin discrimination and trading up, increasing rates of polyandry. These findings reveal a genetic correlation between virgin and non-virgin latency to mate suggesting that polyandry may be maintained because of the need to breed.