Molecular phylogenetic studies of the six currently accepted species in the genusEchinocactushave partially clarified certain aspects of its phylogeny. Most of the studies lack a complete sampling ofEchinocactusand are based only in one source of data. Phylogenetic uncertainties inEchinocactus, such as the recognition ofHomalocephalaas a different genus fromEchinocactus, the exclusion ofE.grusoniior the affinities ofE.polycephalus, are here resolved. Phylogenetic relationships ofEchinocactuswere reconstructed with a maximum parsimony, a maximum likelihood and a Bayesian approach including 42 morphological characters, four chloroplast markers (atpB-rbcL,trnH-psbA,trnL-trnF andtrnK/matK) and two nuclear genes. The utility of these two nuclear regions related to the betalain cycles (DODA and5GT) are explored and discussed in relation to their potential as phylogenetic markers. Concatenated analyses with morphological and molecular data sets, plus 13 indels (2847 characters and 26 taxa), show general agreement with previous independent phylogenetic proposals but with strong support in order to propose the recognition of a reducedEchinocactusand the recognition ofHomalocephalaat the generic level. These results recovered a polyphyleticEchinocactusas currently defined. The here-named HEA clade, recovers the species ofHomalocephala,EchinocacutsandAstrophytumas a monophyletic group with strong internal support. TheHomalocephala(H.texensis, H.parryiandH.polycephala), was recovered as sister to theEchinocactusclade (E.platyacanthusandE.horizonthalonius), plus theAstrophytumclade. Consequently, we propose here to recognise a monophyleticEchinocactusand a monophyleticHomalocephalaas two distinct genera with their own molecular and morphological synapomorphies. The evolution of some morphological characters supporting these clades are discussed, the necessary new taxonomic combinations forHomalocephalaare proposed and an identification key for the genera, the species and the subspecies of the HEA clade are presented.
