Phylogenetic analysis of the latitude-niche breadth hypothesis in the butterfly subfamily Nymphalinae

1. One possible explanation for the latitudinal gradient in species richness often demonstrated is a related gradient in niche breadth, which may allow for denser species packing in the more stable environments at low latitudes.

2. The evidence for such a gradient is, however, ambiguous, and the results have varied as much as the methods. Several studies have considered the non-independence of species, but few have performed explicit phylogenetic analyses.

3. In the present study, we tested for a correlation between diet breadth and latitude of distribution in Nymphalinae butterflies using generalised estimating equations (GEE) and accounting for phylogenetic independence.

4. Using a simple model with only latitude of distribution as a predictor variable revealed a significant positive relationship with diet breadth. Previous studies, however, have shown that diet breadth is also correlated with butterfly range size, and in turn, that range size may be correlated with latitude of distribution. Including geographical range size in the model also turned out to have a profound effect on the results – to the extent that the relationship between latitude of distribution and diet breadth was effectively reversed.

5. We conclude that, at least for this group of butterflies, there is no evidence for a positive correlation between latitude of species distribution and diet breadth when controlling for range size, and that the effect may actually even be reversed.