Phylogenetic relationships among fantail darters (Percidae : Etheostoma : Catonotus): Total evidence analysis of morphological and molecular data

Results of a phylogenetic analysis of the complete mitochondrial cytochrome b gene (1140 base pairs) for all species of Catonotus are presented along with a synthesis and phylogenetic analysis of published morphological data. The two datasets are combined in a total evidence analysis, and results fr om the molecular, morphological, and total evidence datasets are compared w ith each other and with previously published hypotheses. Phylogenetic relat ionships suggested by morphological data are similar to those from previous studies. The cytochrome b and total evidence analyses also produced trees that are generally congruent with precious hypotheses. The monophyly of the Etheostoma squamiceps group and the monophyly of a dade including members of the E. virgatum and E.flabellare groups are well supported. However, in contrast to traditional classification, E. barbouri usually clustered with the E. flabellare group, and E. percnurum usually clustered with the E virg atum group. Etheostoma percnurum and E. barbouri possess many autapomorphie s, and it is possible that they share fewer cytochrome ii characters with c lose relatives than they share with other species due to homoplasy resultin g from accelerated rates of evolution. Also, in contrast to earlier hypothe ses, the molecular and total evidence analyses suggested that E. squamiceps , E. crossopterum, and E. olivaceum are closely related. An earlier hypothe sis based on morphology suggested that E. olivaceum was basal to other memb ers of the E. squamiceps group and that E. squamiceps was related to E. chi enense, E. pseudovulatum, E. oophylax and E neopterum. Etheostoma olivaceum has been considered basal because it lacks putative synapomorphies of all other members of the E. squamiceps group. Although reversals (in E: olivace um) and (on vergence tin E. squamiceps) in character states are possible, a test of the two hypotheses of relationship requires additional data.