Because phytogenies play a central role in many evolutionary and population genetic questions, it is vital to have confidence in phylogeny reconstructions for particular data sets. An algorithm recently introduced by Templeton et al. (1992, Genetics 132:619-633) simultaneously estimates cladograms from a given set of restriction site or nucleotide sequence data and calculates a confidence set of alternatives while allowing for the possibility of recombination within the data set. This method was developed for use at the intraspecific level, where mutational differences between operational taxonomic units are few. Because of the strengths of this method, it would be desirable to use it at higher levels of phylogenetic analysis. I tested the accuracy of the method of Templeton et al. using a known phylogeny from bacteriophage T7, which simulates higher levels of diversification. This method has greater statistical power and more accurately reconstructs phylogenetic relationships than does maximum parsimony when few molecular characters are available for analysis, even at higher levels of diversification.