Why humans cooperate in large groups and with non-kin remains a puzzle for researchers across the natural and social sciences. Investigating whether cooperation is sexually selected could contribute to an understanding of the evolution of human cooperation. Competition for access to mates could indeed select for cooperation. Using controlled laboratory experiments, we analyse whether and how the sex composition of a social environment, testosterone level, and relationship status affect contributions to a public good. The results show that variation in sex composition alters the amount of money that single men (but not men in a couple or women) contribute to a public good. Notably, in line with the competitive helping hypothesis, awareness of the presence of a woman leads to larger contributions by single men, most likely by triggering their competitiveness to be the most cooperative man in the group. However, we find no link between basal testosterone level and cooperativeness. We argue that men, notably single men, adopt cooperative behaviours as a signalling strategy in the context of mate choice and hence that cooperation is partly sexually selected. Our findings highlight the need to consider sexual selection as an additional mechanism for cooperation. The evolution of cooperation between unrelated individuals remains a challenging issue for evolutionary biologists 1. The puzzle is that cooperative behaviour is beneficial to recipients but is costly to the actor. Theory suggests that cooperative acts can only be selected if the costs of cooperation are compensated by fitness benefits 2. Some theoretical 3–5 and experimental studies 6,7 have shown that sexual benefits could maintain cooperation in a group. Investigations into whether cooperation is sexually selected could thus contribute to our understanding of the evolution of cooperation. Based on controlled and incentivized laboratory experiments, we show that variation in the sex composition of a group of four individuals alters the amount of money that single men (but not men in a couple or women) contribute to a public good, when the sex composition is common knowledge. Notably, in line with the competitive helping hypothesis, awareness of the presence of a woman in the group likely increases single men's competitiveness in their willingness to appear as the most cooperative individual, leading to larger contributions by those men. We also investigate the potential influence of basal testosterone as a proximate mechanism of this behaviour. However, no association between male testosterone level and contribution to a public good is detected. Despite the absence of testosterone influence on cooperation, we argue that men, especially single men, adopt cooperative behaviours as a signalling strategy in the context of mate choice, and hence that cooperation is – at least partly-sexually selected in humans. Observations in non-human animal species suggest that sexual selection could be implicated in the evolution of cooperation. Male Lance-tailed Manakins (Chiroxiphia lanceolata) perform cooperative efforts, such as singing duets and dancing with other males, which increases their chances of mating with females 8. In chimpanzees (Pan troglodytes), males exchange political support from the dominant male for mating opportunities 9 : dominant males, in turn, tolerate males who support them most frequently in conflicts, with the result that supporters copulate more often than other males. In humans, there is also evidence that cooperative behaviour toward non-kin could be sexually selected. For example, several studies suggest that cooperative traits are implicated in mate choice 10–12 and that women are more sensitive to those traits in sexual partners than men are 10 .