Whenever different types of data are used in a scientific analysis, a concern is whether they all point to the same answer except for sampling error. This is no less the case for phylogenetic analysis. If rates of evolution exceed a certain threshold, and edge (branch or internode) lengths and branching pattern meet certain configurations, then maximum parsimony will be inconsistent (Felsenstein 1978Citation ). That is, the data will indicate the wrong tree with increasing support as sequence length tends to infinity. Accordingly, it is possible that faster-evolving characters strongly disagree with more slowly evolving characters in recovering the underlying tree, and the combined data set may also indicate an incorrect tree (Bull et al. 1993Citation ). Furthermore, inconsistency has been shown to be a problem for all classes of phylogenetic analysis, including distance methods and maximum likelihood (Waddell 1995Citation ; Lockhart et al. 1996Citation ).