The effects of topological inaccuracy in evolutionary trees on the phylogenetic comparative method of independent contrasts

Abstract.—Computer simulations were used to test the effect of increasing phylogenetic topological inaccuracy on the results obtained from correlation tests of independent contrasts. Predictably, increas-ing the number of disruptions in the tree increases the likelihood of signicant error in the r values produced and in the statistical conclusions drawn from the analysis. However, the position of the disruption in the tree is important: Disruptions closer to the tips of the tree have a greater effect than do disruptions that are close to the root of the tree. Independent contrasts derived from inaccurate topologies aremore likely to lead to erroneous conclusionswhen there is a true signicant relationship between the variables being tested (i.e., they tend to be conservative). The results also suggest that random phylogenies perform no better than nonphylogenetic analyses and, under certain conditions, may perform even worse than analyses using raw species data. Therefore, the use of random phy-logenies is not benecial in the absence of knowledge of the true phylogeny. [Comparative method; computer simulation; independent contrasts; phylogenetic error; random phylogenies; topological inaccuracy.] Perhaps the most important assumption associated with all phylogenetic compara