The drivers and consequences of unstable dynamics: a long-term study of three malaria parasite species infecting a tropical lizard.

Understanding the consequences of environmental fluctuations for parasite dynamics requires a long-term view stretching over many transmission cycles. Here we studied the dynamics of three malaria parasites (Plasmodium azurophilum, P. leucocytica and P. floridense) infecting the lizard Anolis gundlachi, in the rainforest of Puerto Rico. In this malaria-anole system we evaluated temporal fluctuations in individual probability of infection, the environmental drivers of observed variation and consequences for host body condition and Plasmodium parasites assemblage. We conducted a total of 15 surveys including 10 from 1990 to 2002 and five from 2015 to 2017. During the early years, a lizard's probability of infection by all Plasmodium species appeared stable despite disturbances ranging from two hurricanes to short droughts. Over a longer timescale, probability of infection and overall prevalence varied significantly, following non-linear relationships with temperature and rainfall such that highest prevalence is expected at intermediate climate measures. A perplexing result was that host body condition was maximized at intermediate levels of rainfall and/or temperature (when risk of infection was highest), yet we found no significant decreases in body condition due to infection. Plasmodium parasite species composition varied through time with a reduction and near local extinction of P. floridense. Our results emphasize the need for long-term studies to reveal host-parasite dynamics, their drivers and consequences.