West Wind Drift revisited: testing for directional dispersal in the Southern Hemisphere using event-based tree fitting

Recent studies suggest that if constrained by prevailing wind or ocean currents dispersal may produce predictable, repeated distribution patterns. Dispersal mediated by the West Wind Drift (WWD) and Antarctic Circumpolar Current (AAC) has often been invoked to explain the floristic similarities of Australia, South America and New Zealand. If these systems have been important dispersal vectors then eastward dispersal - from Australia to New Zealand and the western Pacific to South America - is expected to predominate. We investigate whether phylogenies for Southern Hemisphere plant groups provide evidence of historical dispersal asymmetry and more specifically whether inferred asymmetries are consistent with the direction of the WWD/AAC. Southern Hemisphere. We assembled a data set of 23 published phylogenies for plant groups that occur in New Zealand, Australia and/or South America. We used parsimony-based tree fitting to infer the number and direction of dispersals within each group. Observed dispersal asymmetries were tested for significance against a distribution of expected values. Our analyses suggest that dispersal has played a major role in establishing present distributions and that there are significant patterns of asymmetry in Southern Hemisphere dispersal. Consistent with the eastward direction of the WWD/ACC, dispersal from Australia to New Zealand was inferred significantly more often than in the reverse direction. No significant patterns of dispersal asymmetry were found between the western Pacific landmasses and South America. However, eastward dispersal was more frequently inferred between Australia and South America, while for New Zealand-South American events westward dispersal was more common. Our results suggest that eastward circumpolar currents have constrained the dispersal of plants between Australia and New Zealand. However, the WWD/ACC appear to have had less of an influence on dispersal between the western Pacific landmasses and South America. This observation may suggest that differences in dispersal mechanism are important - direct wind or water dispersal vs. stepping-stone dispersal along the Antarctic coast. While our analyses provide useful preliminary insights into dispersal asymmetry in the Southern Hemisphere we will need larger data sets and additional methodological advances in order to test fully these dispersal patterns and infer processes from phylogenetic data.