Sedentary songbirds maintain higher prevalence of haemosporidian parasite infections than migratory conspecifics during seasonal sympatry

Abstract

Long-distance migrations influence the physiology, behavior, and fitness of migratory animals throughout their annual cycles, and fundamentally alter their interactions with parasites. Several hypotheses relating migratory behavior to the likelihood of parasitism have entered the literature, making conflicting, testable predictions. To assess how migratory behavior of hosts is associated with parasitism, we compared haemosporidian parasite infections between two closely related populations of a common North American sparrow, the dark-eyed junco, that co-occur in shared habitats during the non-breeding season. One population is sedentary and winters and breeds in the Appalachian Mountains. The other population is migratory and is found in seasonal sympatry with the sedentary population from October through April, but then flies (≥ 900 km) northwards to breed. The populations were sampled in the wild on the shared montane habitat at the beginning of winter and again after confining them in a captive common environment until the spring. We found significantly higher prevalence of haemosporidian parasite infections in the sedentary population. Among infected juncos, we found no difference in parasite densities (parasitemias) between the sedentary and migrant populations and no evidence for winter dormancy of the parasites. Our results suggest that long-distance migration may reduce the prevalence of parasite infections at the population level. Our results are inconsistent with the migratory exposure hypothesis, which posits that long-distance migration increases exposure of hosts to diverse parasites, and with the migratory susceptibility hypothesis, which posits that trade-offs between immune function and migration increase host susceptibility to parasites. However, our results are consistent with the migratory culling hypothesis, which posits that heavily infected animals are less likely to survive long-distance migration, and with the migratory escape hypothesis, which posits that long-distance migration allows host populations to seasonally escape areas of high infection risk.

Fig 1. A maximum likelihood phylogeny of the haemosporidian parasites.

The phylogeny is based on sequences from samples collected in December. Branch tips with filled boxes represent parasites sequenced from sedentary junco blood samples. Branch tips with unfilled boxes represent parasites sequenced from migrant junco blood samples. Numbers represent individual blood sample IDs. Horizontal branch length represents phylogenetic distance (substitutions/site). For reference, the parasite sequences from the described morphospecies on the MalAvi data base that most closely matched each parasite lineage in our phylogeny are included in the phylogenetic tree.

Fig 2. Haemosporidian infection prevalence in the blood.

Prevalence estimates are based on qPCR (A) or nested PCR (B), in the sedentary population (filled bars, n = 19) and the migrant population (unfilled bars, n = 18). Error bars represent 95% binomial confidence intervals. Asterisks between filled and unfilled bars indicate significant differences between populations at each time point (* = p < 0.05, ** = p < 0.01). Brackets (top) indicate within-population cross time point comparisons where infection prevalences differed significantly (no brackets shown where infection prevalences did not differ across time points).

Fig 3. Haemosporidian parasitemias of individual juncos.

Haemosporidian parasitemias (the number of haemosporidian cyt b gene copies per 10 ng of total DNA (host + parasite), as determined by qPCR) in the bloodstream of individual sedentary (panel A, n = 19) and migrant (panel B, n = 18) juncos across time points. Each colored line represents an individual junco. Overlapping lines representing uninfected birds are stacked on top of each other at the bottom for visualization.

Fig 4. Boxplot of mean parasitemias.

The mean parasitemias (the number of haemosporidian cyt b gene copies per 10 ng of total DNA (host + parasite), as determined by qPCR, averaged across the three sampling time points: December, early March, and late March) of sedentary (filled, n = 17) juncos and migrant (unfilled, n = 9) juncos. Only infected juncos were included in this analysis. The horizontal lines inside the boxes represent the sample medians. The length of the box represents the interquartile range. Whiskers span all the data except for statistical outliers. Statistical outliers are points that are at least 1.5 box lengths away from the edge of the box.