Maximum Parsimony Vs Maximum Likelihood Estimation