Aim Our objective was to evaluate the influence of distinct macroecological factors (space, floristics and environment) on the variation in seed dispersal strategies of shrubs and trees in the Araucaria forest biome in the southern and south-eastern Brazilian highlands. We hypothesize that history-related factors (space, floristics) are major determinants of the proportion of endo- and syn-zoochorous species in Araucaria forests, despite the current gradients in environmental conditions.
Location Araucaria forest is the main forest biome in southern Brazil, at altitudes above 500 m a.s.l. (latitude ≤ 30° S). Their northern limit in Brazil is at latitude 20° S, where forests occur at elevations above 1000 m a.s.l.
Methods We compiled information from 27 floristic checklists comprising shrub and tree species distributed along the geographical range of the Brazilian Araucaria forest biome. We classified species as zoochorous and non-zoochorous, based on morphological attributes of their diaspores. Sites were described by geographical coordinates (latitude, longitude) and five environmental variables. We evaluated separately the influence of floristic, spatial and environmental variables on zoochory using correspondence analysis and linear regressions. Further, we evaluated causal connections between these variable groups using Mantel tests and path analysis.
Results Zoochory increased with both latitude and longitude. Regression analysis showed that rainfall seasonality was the only environmental variable explaining the variation in the proportion of zoochorous species. All pairwise Mantel correlations between space, rainfall seasonality, floristics and zoochory were significant. Path analysis showed that rainfall seasonality was strongly determined by spatial distances between sites, and floristics was directly determined by rainfall seasonality. Further zoochory was mostly determined by floristics.
Main conclusions Taking into account historical factors in the interpretation of macroecological patterns improves our understanding of biodiversity gradients. Hypotheses based on long-term dynamics of distinct floristic groups provide some useful insights into patterns shown by studies elsewhere. Here we offer an analytical solution to incorporate history-related factors into macroecological analyses. While history-based hypotheses do not replace any other ideas concerning macroecological patterns, they are likely to improve our understanding on factors determining present-day ecological patterns.