ABSTRACT

The fact that combined analyses at the level of both nucleotides and codons yield topologies that are different from most of the other analyses is unsettling. It underscores concerns voiced in the literature that it may be poor practice to simply assume that combining data sets will result in accurate phylogenetic estimates (Barret et a/. 1991; de Queiroz 1993). The fact that combined parsimony analyses yields different solutions for a data set comprising nearly 6000 base pairs should be a matter for concern.