Although there is considerable controversy regarding the phylogenetic structure and infrafamilial classification of Chenopodiaceae, the tribe Atripliceae generally has been recognized as a natural group. Relationships within this tribe remain controversial, especially with regard to the taxonomic delimitation and infrageneric classification of the largest genus, Atriplex. Most of the genera that have been segregated from Atriplex are monotypic or include few species that are variously distributed in Africa. Australia, or North America: however, one of these genera, Obione, includes half of all species of Atriplex sens. lat., and has a worldwide distribution. Seventy terminals were included in a cladistic analysis, based on morphological characters, to test the monophyly of Atripliceae and its subtribes, and of Atriplex, Obione, and several infrageneric groups. The taxonomic sample includes representatives of 22 putative genera of the tribe Atripliceae, and all sections of Atriplex (except Austrobione) and Obione, following Ulbrich's (1934) classification, from throughout the geographic range of the tribe. Primary hypotheses of homology were postulated for 78 characters reflecting variation in gross morphology, leaf anatomy, and chromosome number. Results, as reflected in the strict consensus tree, suggest that both Atripliceae and Atriplex are paraphyletic, with three outgroup genera from tribe Chenopodieae (Chenopodium, Monolepis, and Suckleya) nested among species of Atriplex. Also, monophyly of the subtribes of Atripliceae is not supported. The results suggest that the deepest branch within Atripliceae is between Ceratocarpus and all other members of the tribe. Kruscheninnikovia, the next genus to diverge, is the sister of a diverse clade that consists of two subclades. One of these includes Theleophyton plus Endolepis, Zuckia, and Gravia, and the other includes all species of Atriplex in the sample, plus all sampled species of Obione, the remaining genera of Atripliceae, and the three sampled genera of Chenopodieae, with Halimione pedunculata as the sister of all other members of this clade. Resolution generally is poor among species of Atriplex, and neither Obione nor most of the sections of Atriplex is resolved as monophyletic, though there is support for the monophyly of section Spongiocarpus. Also nested among species of Atriplex and Obione is a well resolved clade with four defined subgroups: Suckleya by itself, Atriplex hortensis and Chenopodium; Archiatriplex, Axyris, and Microgynoecium; and Monolepis, Proatriplex, Manochlamys, and Exomis. The results support the recognition of Endolepis, Theleophyton, and Zuckia (sensu Standley 1915) as segregates of Atriplex, but not Blackiella, Hatimione, Haloxanthium, Morrisiella, Neopreissia, Obione, Pachypharynx, or Senniella. In addition, the results suggest that the continued recognition of Archiatriplex, Axyris, Exomis, Manochlamys, Microgynoecium, Proatripex, and Spinacia (Atripliceae), as well as Chenopodium. Suckleya, and Monolepis (Chenopodieae) should be reconsidered. If these results are confirmed by further analyses using other character sets and a wider taxonomic sample, it will be necessary either to transfer several genera to Atriplex or to circumscribe Atriplex narrowly. and segregate a number of additional genera from it.
