Because phylogenies play a central role in many evolutionary and popul
ation genetic questions, it is vital to have confidence in phylogeny r
econstructions for particular data sets. An algorithm recently introdu
ced by Templeton et al. (1992, Genetics 132:619-633) simultaneously es
timates cladograms from a given set of restriction site or nucleotide
sequence data and calculates a confidence set of alternatives while al
lowing for the possibility of recombination within the data set. This
method was developed for use at the intraspecific level, where mutatio
nal differences between operational taxonomic units are few. Because o
f the strengths of this method, it would be desirable to use it at hig
her levels of phylogenetic analysis. I tested the accuracy of the meth
od of Templeton et al. using a known phylogeny from bacteriophage T7,
which simulates higher levels of diversification. This method has grea
ter statistical power and more accurately reconstructs phylogenetic re
lationships than does maximum parsimony when few molecular characters
are available for analysis, even at higher levels of diversification.