Virulence of lizard malaria: the evolutionary ecology of an ancient parasite-host association.

The negative consequences of parasitic infection (virulence) were examined for two lizard malaria parasite-host associations: Plasmodium agamae and P. giganteum, parasites of the rainbow lizard, Agama agama, in Sierra Leone, West Africa; and P. mexicanum in the western fence lizard, Sceloporus occidentalis, in northern California. These malaria species vary greatly in their reproductive characteristics: P. agamae produces only 8 merozoites per schizont, P. giganteum yields over 100, and P. mexicanum an intermediate number. All three parasites appear to have had an ancient association with their host. In fence lizards, infection with malaria is associated with increased numbers of immature erythrocytes, decreased haemoglobin levels, decreased maximal oxygen consumption, and decreased running stamina. Not affected were numbers of erythrocytes, resting metabolic rate, and sprint running speed which is supported by anaerobic means in lizards. Infected male fence lizards had smaller testes, stored less fat in preparation for winter dormancy, were more often socially submissive and, unexpectedly, were more extravagantly coloured on the ventral surface (a sexually dimorphic trait) than non-infected males. Females also stored less fat and produced smaller clutches of eggs, a directly observed reduction in fitness. Infected fence lizards do not develop behavioural fevers. P. mexicanum appears to have broad thermal buffering abilities and thermal tolerance; the parasite's population growth was unaffected by experimental alterations in the lizard's body temperature. The data are less complete for A. agama, but infected lizards suffered similar haematological and physiological effects. Infected animals may be socially submissive because they appear to gather less insect prey, possibly a result of being forced into inferior territories. Infection does not reduce clutch size in rainbow lizards, but may lengthen the time between clutches. These results are compared with predictions emerging from several models of the evolution of parasite virulence. The lack of behavioural fevers in fence lizards may represent a physiological constraint by the lizards in evolving a thermal tolerance large enough to allow elimination of the parasite via fever. Such constraints may be important in determining the outcome of parasite-host coevolution. Some theory predicts low virulence in old parasite-host systems and higher virulence in parasites with greater reproductive output. However, in conflict with this argument, all three malarial species exhibited similar high costs to their hosts.