Mating failure shapes the patterns of sperm precedence in an insect

Estimates of last male sperm precedence (P 2) are often used to infer mechanisms of sperm competition, a form of post-copulatory sexual selection. However, high levels of mating failure (i. e. copulations resulting in no offspring) in a population can lead to misinterpretations of sperm competition mechanisms. Through simulations, García-González (2004) illustrated how mating failure could cause bimodal distributions of paternity with peaks at P 2 = 0 and 1, under a random sperm mixing mechanism. Here, we demonstrate this effect empirically with the seed bug Lygaeus simulans, a species known to exhibit high levels of mating failure (40–60 %), using a morphological marker to estimate paternity. Contrary to previous findings in a sister species, we did not find strong evidence for last male sperm precedence. There was a tendency towards last male precedence (P 2 = 0.58) but within the expected range for random sperm mixing. Instead, P 2 was highly variable, with a bimodal distribution, as predicted by García-González (2004). After taking mating failure into account, the strongest driver of paternity outcome was copulation duration. Furthermore, we found evidence that mating failure could partly be a female-associated trait. Some doubly-mated females were more likely to produce no offspring or produce offspring from two different sires than expected by chance. Therefore, some females are more prone to experience mating failure than others, a result that mirrors an earlier result in male L. simulans. Our results confirm that mating failure needs to be considered when interrogating mechanisms of post-copulatory sexual selection.