ABSTRACT
Cladistic biogeography searches for general patterns of distribution among multiple organisms as evidence of a common biogeographic history, but does not explicitly consider the evolutionary processes that created such patterns or the timing of divergence of the organisms studied. This makes biogeographic results often difficult to interpret. Here, I review recent methodological developments on biogeographic analysis based on the recognition of processes (“the event based approach”), which have led to an extraordinary revolution in biogeographic studies. Event-based methods explicitly include all biogeographic processes into the analysis as possible explanations for the observed biogeographic pattern. Each process (vicariance, duplication, extinction, and dispersal) is associated with a cost inversely related to its likelihood. The optimal biogeographic reconstruction is the one that minimizes the total cost of the implied events (most parsimonious). Significance of results can be assessed by comparing them with those derived from random data sets under the null hypothesis that distributions are not phylogenetically constrained. I compare two event-based methods, dispersal-vicariance analysis and parsimony-based tree fitting in relation to alternative biogeographic scenarios: hierarchical vs. reticulate, widespread vs. endemic, etc. I describe recent studies in which event-based methods, combined with large data sets of phylogenies and estimates of divergence times, have been used to test large-scale biogeographic patterns on the Holarctic and Southern Hemisphere biotas. Despite claims by cladistic biogeography that dispersal is a random, stochastic event, these studies show that dispersal can also generate congruent distribution patterns among multiple taxa if coordinated in direction, e.g., by prevailing winds and ocean currents (“concerted dispersal”). They also suggest major biogeographic differences between animals and plants, with plant patterns more influenced by recent dispersal.
