FIGURE 1. The optimal maximum likelihood tree of the combined rps4/nad5/nuc26S data set under the general timereversible model with site rate variation (GTR+I+Γ: -ln likelihood = 54575.7; optimal parameters: A/C = 1.1166, A/G = 5.0600, A/T = 0.2165, C/G = 1.0611, C/T = 5.6457, G/T =1(fixed); f(A) = 0.3360, f(C) = 0.1836, f(G) = 0.1781, f(T) = 0.3027; α = 0.5730; pinvar = 0.3410). Nodes support by>95% posterior probability are highlighted. Taxonomic labels are applied to clades at the lowest rank applicable to the clade following the classification of Buck and Goffinet (2000): A) acrocarpous mosses - the two exemplars marked with an asterisk were originally labeled Hymenostylium recurvirostre and Pottia truncata and subsequently identified as Ceratodon purpureus, B) pleurocarpous mosses - part 1, C) pleurocarpous mosses - part 2.

Published as part of Cox, Cymon J., Goffinet, Bernard, Wickett, Norman J., Boles, Sandra B. & Shaw, A. Jonathan, 2010, Moss diversity: A molecular phylogenetic analysis of genera, pp. 175-195 in Phytotaxa 9 on page 186, DOI: 10.11646/phytotaxa.9.1.10, http://zenodo.org/record/10084149