Deep-history Biogeography: Retrieving the Historical Pattern of Evolving Continental Biotas

This paper discusses some methodological problems associated with reconstructing the evolutionary history of continental biotas through a biogeographic analysis of areas of endemism. Present evidence suggests continental biotas have had a long, complex history asjudged by patterns of relationship among areas of endemism. Thus, an analysis of speciescladograms,in which component taxa share areas of endemism, reveals, first, that more thanone general area-cladogram typically exists for the areas of a continental biota, and second, thatthere is often incongruence among these general area-cladograms. Under the vicariance hypothesisthat general area-cladograms describe real historical patterns for components of thebiota, such incongruence implies that the biota has been subjected to a number of processes thatcan produce highly complex histories among areas of endemism, including area-hybridization,vicariance events of different ages located in the same areas, extinctions, and loss of barriers,among others. Current methods of biogeographic pattern analysis—specifically component analysisand biogeographic parsimony analysis—are reductionist in the sense that they attempt toresolve multiple conflicting patterns across species-cladograms to a singular, less complex pattern,or general area-cladogram. As incongruence increases among general area-cladograms, componentanalysis will resolve fewer shared components. Biogeographic parsimony analysis, in contrast,will always be able to produce parsimonious solutions, but those solutions may not explainmuch of the original biogeographic data satisfactorily. This latter method is predicated onviewing areas as analogues of taxa and taxa as analogues of characters. The analogy may notalways be appropriate, however, because, among other things, areas can have multiple histories,but taxa cannot. It is suggested that “nonreductionist” methods need to be developed withinhistorical biogeography in order to generate hypotheses that can unify the historical patternsimplied by incongruent general area-cladograms. These problems are illustrated by biogeographic patterns among the areas of endemism of the Amazon basin.