Although parasitism has been acknowledged as an important selective force in the evolution of host life histories, studies of fitness effects of parasites in wild populations have yielded mixed results. One reason for this may be that most studies only test for a linear relationship between infection intensity and host fitness. If resistance to parasites is costly, however, fitness may be reduced both for hosts with low infection intensities (cost of resistance) and high infection intensities (cost of parasitism), such that individuals with intermediate infection intensities have highest fitness. Under this scenario one would expect a non-linear relationship between infection intensity and fitness.
Using data from blue tits (Cyanistes caeruleus) in southern Sweden, we investigated the relationship between the intensity of infection of its blood parasite (Haemoproteus majoris) and host survival to the following winter. Presence and intensity of parasite infections were determined by microscopy and confirmed using PCR of a 480bp section of the cytochrome-b-gene. While a linear model suggested no relationship between parasite intensity and survival (F = 0.01, p = 0.94), a non-linear model showed a significant negative quadratic effect (quadratic parasite intensity: F = 4.65, p = 0.032; linear parasite intensity F = 4.47, p = 0.035). Visualization using the cubic spline technique showed maximum survival at intermediate parasite intensities.
Our results indicate that failing to recognize the potential for a non-linear relationship between parasite infection intensity and host fitness may lead to the potentially erroneous conclusion that the parasite is harmless to its host. Here we show that high parasite intensities indeed reduced survival, but this effect was masked by reduced survival for birds heavily suppressing their parasite intensities. Reduced survival among hosts with low parasite intensities suggests costs of controlling parasite infections; however, the nature of such costs remains to be elucidated.