The Utility of the Incongruence Length Difference Test

Abstract

Conditional combination of phylogeneticdatarequiresde” nitionofexplicitcriteriaforcombinability (Bull et al., 1993). In this con-text,combinabilityreferstothemethodolog-ical validity of combining multiple sourcesof phylogenetic data, given the underly-ing assumptions (explicit or otherwise) ofthe analysis. Combinability has been eval-uated by the effect of data set combina-tion on phylogenetic accuracy: Combinabledata sets increase accuracy (Bull et al.,1993; Cunningham, 1997b). When inferen-tial methods are statistically consistent, thisconvergent property is guaranteed by sta-tistical homogeneity of the data sets to becombined: Increasing sample size increasesprecision. In a phylogenetic context, datahomogeneity can be de” ned as the shar-ing of a single history (topological pat-