Genus Aster s.l. has been the object of much controversy, particularly in North America (NA). Recently, splitting the aggregate genus has been suggested, notably in the thorough morphological review of Nesom (1994, 2000), but students of the group have yet to agree on delimitation and relationships of the segregate taxa. Also, relationships with Eurasian taxa, and among them, have been controversial. Recent ITS phylogenetic analyses by Morgan (1997) and Noyes (2000, & Rieseberg 1998) provide a framework to test hypotheses of relationships among members of the Aster complex. In this phylogenetic analysis, we are adding more than 60 ITS sequences of many critical taxa representing all major lineages that have been recognized within Aster s.l., in order to intensify taxon sampling both within and among groups in the NA Astereae and in Eurasian Aster. All North American and derivative taxa form a monophyletic clade, while Eurasian taxa also form a single clade excluding all NA taxa. Thus, we confirm that Aster is polyphyletic; the name should be applied only within subtribe Asterinae (Eurasian clade); Doellingeria is strictly North American and not disjunct with east Asia. Each major clade of NA Astereae have an aster segregate at its base: Eucephalus and Doellingeria are basal to all NA Astereae; Sericocarpus is sister to the Solidago lineage within Solidagininae; the Symphyotrichinae (all asters) are sister to Machaerantherinae, where Eurybia and Oreostemma are basal; Ionactis and Oclemena are sister to the Conyzinae-Chrysopsidinae clade. Within the Symphyotrichinae, Canadanthus, Ampelaster and A. chapmanii form a grade basal to Symphyotrichum (including Psilactis). Biogeographically and ecologically, it appears that on both sides of North America, temperate, mesic taxa have developed first, followed by multiple radiations into more xeric habitats.

Key words: Aster s.l., Asteraceae, Astereae, Biogeography, ITS, Molecular phylogeny