Testing Relationships Between Biotic Components - Evolutionary Biogeography - page 210

not be associated with paralogous nodes, preventing artifactual results, if
not completely then at least to a significant degree (Nelson and Ladiges
2003; Parenti 2007).
Nelson and Ladiges (1996) developed an algorithm that constructs
paralogy-free subtrees, starting off at the most terminal groups of the clado-
gram. The procedure reduces complex cladograms to paralogy-free sub-
trees, meaning that geographic data are associated only with informative
nodes, and areas duplicated or redundant in the descendants of each node
do not exist ( fig. 5.18 ) . These are the only data relevant for cladistic biogeo-
graphy. When obtained, paralogy-free subtrees are represented in a com-
ponent or a three-item matrix and analyzed with a parsimony algorithm. Be-
fore paralogy-free subtrees are obtained, the transparent method (Ebach et
al. 2005a) may be implemented to resolve widespread taxa.
Figure 5.18 Paralogy-free subtree analysis. (a-e) Original taxon-area clado-
grams, with paralogous nodes; (f) the single paralogy-free subtree that can be de-
rived from them.