Bias in Phylogenetic Estimation and Its Relevance to the Choice between Parsimony and Likelihood Methods

Swofford et al show that results such as those of Siddall (1998) should not be take as a vindication of parsimony with respect to one particular problem--sensitivity to long-branch-attraction artifacts. When interpreted properly, simulation studies such as those of Siddall merely reinforce arguments for the utility model-based methods, including maximum likelihood, for phylogenetic analysis of molecular sequence data.