A positive relationship across species between the extent to which females mate with more than one male and relative testes mass has been demonstrated in a wide range of vertebrate taxa and certain insects. At least two hypotheses, which are not necessarily mutually exclusive, could account for this pattern: (1) the numerical sperm competition hypothesis, which assumes that larger testes enable the male to transfer more sperm to each female, giving the male an advantage in sperm competition and (2) the male mating rate hypothesis, which proposes that larger testes allow the male to produce a greater number of (potentially smaller) ejaculates to engage in frequent copulations with different females. Of these hypotheses, the former has won broad acceptance, while the latter has tended to be dismissed. Here, we argue that the lines of evidence commonly used to support the numerical sperm competition hypothesis in favour of the male mating rate hypothesis are not as clear cut or generally applicable as they are purported to be and that, consequently, the male mating rate hypothesis cannot be excluded with confidence on the basis of the current evidence. Furthermore, some evidence, such as the finding that ejaculate mass and/or sperm number is negatively correlated with testes mass across species in some insects and that larger testes in Drosophila can evolve in response to an increase in the number of females available for mating in the laboratory, provides support for the male mating rate hypothesis. Further work is needed to disentangle the relative effects of these selective pressures on the evolution of testes size.