The course of infection of a malaria parasite (Plasmodium mexicanum) is highly variable in its host, the fence lizard (Sceloporus occidentalis). However, a seasonal trend is superimposed on this variation such that gametocyte production is intensified during mid- to late summer. Host testosterone levels follow a similar seasonal fluctuation and are variable among individual lizards. We sought to determine if testosterone levels affect seasonal and among-host variation in 11 P. mexicanum life history traits: rate of increase in level of infection (3 measures), peak parasitemia (3 measures), duration of increase (3 measures), time to detectable infection, and timing of production of gametocytes. We followed the course of infection in 125 male S. occidentalis, each randomly assigned to 1 of 4 treatment groups: castrated, castrated and implanted with exogenous testosterone, sham implanted, and unmanipulated controls. Median values for the 11 life history traits did not differ among treatment groups, and variances were homogeneous among the treatment groups for 10/11 traits. However, elevated testosterone significantly reduced the variation in timing of the onset of gametocyte production. Therefore, testosterone does not appear to be a primary regulator of P. mexicanum life history, yet testosterone may have some effect on when gametocytes first become detectable.