Abstract
BACKGROUND AND AIMS: Like numerous patterns in ecology and evolution, the latitudinal diversity gradient varies across phylogenetic levels. Yet, studies that investigate systematically how patterns and processes change at different phylogenetic levels, from the tips to the root, are still relatively scarce. Here, we test the hypothesis that, despite the high long-distance dispersal capacities of liverworts, which would be expected to result in the homogenization of their distributions, an increase of diversity with latitude persists at increasing phylogenetic level owing to macroclimatic niche conservatism since the earliest evolutionary history of the group. METHODS: Liverwort distributions were scored for 450 operational geographical units worldwide. From the tips to the root, the phylogeny was sliced continuously to examine how taxonomic and phylogenetic diversity are correlated with latitude in a standardized way. Taxonomic diversity and mean phylogenetic distance among taxa were computed for each operational geographical unit at different phylogenetic levels and correlated with macroecological factors using spatial linear models. KEY RESULTS: The correlation between taxonomic diversity and latitude shifted progressively from significantly negative at species level to non-significant, then significantly positive at the highest phylogenetic levels. Taxonomic diversity and mean phylogenetic distance were both significantly correlated with macroclimatic factors across all phylogenetic levels. CONCLUSIONS: In contrast to the marked increase of angiosperm family diversity towards the tropics, the latitudinal diversity gradient evidenced at species level in liverworts decayed progressively at increasing phylogenetic level, suggesting that phylogenetic niche conservatism has played a much weaker role in liverworts than in angiosperms. The inverted latitudinal diversity gradient towards the deepest phylogenetic levels lends support to the hypothesis that the earliest lineages diversified in extra-tropical conditions, explaining why, unlike in angiosperms, high species richness in the tropics is not associated with high phylogenetic diversity in liverworts. Our results highlight the extent to which a phylogenetically deconstructive approach allows for a better understanding of the accumulation of biodiversity through time.