Mixed–genotype infections of microparasites are common, but almost nothing is known about how competitive interactions within hosts affect the subsequent transmission success of individual genotypes. We investigated changes in the composition of mixed–genotype infections of the rodent malaria Plasmodium chabaudi clones CR and ER by monoclonal antibody analysis of the asexual infection in mice, and by PCR amplification of clone–specific alleles in oocysts sampled from mosquitoes which had fed on these mice. Mixed–clone infections were initiated with a 9:1 ratio of the two clones, with ER as the minority in the first experiment and CR as the minority in the second experiment. When beginning as the majority, clones achieved parasite densities in mice comparable to those achieved in control (single–clone) infections. When they began as the minority, clones were suppressed to less than 10 per cent of control parasitaemias during the early part of the infections. However, in mosquitoes, the frequency of the initially rare clone was substantially greater than it was in mice at the start of the infection or four days prior to the feed. In both experiments, the minority clone in the inocula produced as many, or more, oocysts than it did as a single–clone infection. These experiments show that asexual dominance during most of the infection is poorly correlated to transmission probability, and therefore that the assumption that within–host population size correlates to transmission probability may not be warranted. They also raise the fundamental question of why transmission rates of individual genotypes are often higher from mixed than single–clone infections.