Uneconomical Diagnosis of Cladograms: Comments on Wheeler and Nixon's Method for Sankoff Optimization.

We provide three simple examples demonstrating that Wheeler and Nixon's method of recoding "stepmatrix' characters can fail to yield most parsimonious reconstructions of character evolution under specified cost (transformation-weight) schemes. These examples variously indicate undercounting or overcounting of tree lengths due to an inappropriate assumption of independence among the recoded characters. Their method is therefore not equivalent to Sankoff's dynamic programming algorithm, contrary to their claim.