Sexual selection has been shown to be the driving force behind the evolution of the sometimes extreme and elaborate genitalia of many species. Sexual selection may arise before and/or after mating, or vary according to other factors such as the social environment. However, bouts of selection are typically considered in isolation. We measured the strength and pattern of selection acting on the length of the male intromittent organ (or processus) in two closely related species of lygaeid seed bug: Lygaeus equestris and Lygaeus simulans. In both species, we measured both pre- and post-copulatory selection. For L. equestris, we also varied the experimental choice design used in mating trials. We found contrasting pre- and post-copulatory selection on processus length in L. equestris. Furthermore, significant pre-copulatory selection was only seen in mating trials in which two males were present. This selection likely arises indirectly due to selection on a correlated trait, as the processus does not interact with the female prior to copulation. In contrast, we were unable to detect significant pre- or post-copulatory selection on processus length in L. simulans. However, a formal meta-analysis of previous estimates of post-copulatory selection on processus length in L. simulans suggests that there is significant stabilising selection across studies, but the strength of selection varies between experiments. Our results emphasise that the strength and direction of sexual selection on genital traits may be multifaceted and can vary across studies, social contexts and different stages of reproduction.Significance statementAnimal genitalia vary greatly in size and complexity across species, and selection acting on genital size and shape can be complex. In this study, we show that the length of the penis in two species of seed bug is subject to complex patterns of selection, varying depending on the social context and whether selection is measured before or after mating. In one of the species, we show unexpectedly that penis length is correlated with male mating success, despite the fact that the penis does not interact with the female prior to mating. Our results highlight the fact that genitalia may be subject to both direct and indirect selection at different stages of mating and that to fully understand the evolution of such traits we should combine estimates of selection arising from these multiple episodes.Electronic supplementary materialThe online version of this article (doi:10.1007/s00265-016-2082-6) contains supplementary material, which is available to authorized users.