During an infection, malaria parasites compete for limited amounts of food and enemy–free space. Competition affects parasite growth rate, transmission and virulence, and is thus important for parasite evolution. Much evolutionary theory assumes that virulent clones outgrow avirulent ones, favouring the evolution of higher virulence. We infected laboratory mice with a mixture of two Plasmodium chabaudi clones: one virulent, the other avirulent. Using real–time quantitative PCR to track the two parasite clones over the course of the infection, we found that the virulent clone overgrew the avirulent clone. However, host genotype had a major effect on the outcome of competition. In a relatively resistant mouse genotype (C57Bl/6J), the avirulent clone was suppressed below detectable levels after 10 days, and apparently lost from the infection. By contrast, in more susceptible mice (CBA/Ca), the avirulent clone was initially suppressed, but it persisted, and during the chronic phase of infection it did better than it did in single infections. Thus, the qualitative outcome of competition depended on host genotype. We suggest that these differences may be explained by different immune responses in the two mouse strains. Host genotype and resistance could therefore play a key role in the outcome of within–host competition between parasite clones and in the evolution of parasite virulence.