The latitudinal specialization gradient of bird–malarial parasite networks in South America: lower connectance, but more evenly distributed interactions towards the equator

Abstract

Whereas the latitudinal diversity gradient has been shown for a diverse set of taxa, a related macroecological pattern, the latitudinal specialization gradient (LSG), remains controversial. A classical expectation is that species should present more specialized interactions towards the equator, however, recent studies have provided conflicting evidence for this hypothesis. Here, we tested the LSG in a set of bird–malarial parasite networks across South America. Our analyses comprise 9763 individual birds surveyed in 52 communities, within eight biomes and spanning a gradient of 4700 km. We measured network‐level specialization through indices that account for increasingly comprehensive information. Binary specialization considers the occurrence/absence of interactions between each host–parasite pair; quantitative specialization is also affected by the frequency of interactions between each host–parasite pair; and phylogenetic specialization takes into account the phylogeny and abundances of host species. We found that, while binary specialization increases towards the equator, quantitative specialization decreases. Thus, despite each parasite lineage infecting a more restricted set from the available host species in low latitudes, infections were more evenly distributed among host species, than in higher latitudes. Additionally, using a structural equation model, we show that both direct and indirect effects contribute to this unexpected relationship. Direct effects are weak, and network‐specialization was mostly explained by host species and parasite lineage richness, evidencing that changes in network specialization along latitude are in a large part explained by the latitudinal diversity gradient for birds and malarial‐parasites. In the light of the accumulated evidence over the past years, reinforced by our findings, we suggest that the classical latitudinal specialization hypothesis should be reevaluated, making room for a theoretical framework able to encompass the conflicting results.