Generic Concepts of Aster (Asteraceae): A Comparison of Cladistic, Phenetic, and Cytological Approaches

Abstract

Many attempts have been made at splitting certain groups from the large and complex genus Aster. For North American species, however, most authors have adhered to the broad generic concept espoused by A. Gray in his Synoptical Flora of North America. In response to the recent segregation of Virgulus from Aster by Semple and Brouillet (under the name Lasallea), members of all sections of North American Aster s.l. (except Machaeranthera), plus six of the Old World groups, were subjected to phenetic and cladistic analyses. Nine aster-like species of Erigeron subg. Erigeron were selected to represent an out-group, and 47 OTUs (operational taxonomic units) were scored for 41 characters. The results of this study may not decisively favor a broad or a narrow generic concept of Aster but, in the shortest tree obtained to date, Aster subg. Oreostemma is the sister group of all other subgenera of Aster included in the investigation. At least six groups emerge that seem as well defined as Virgulus and deserving of equal rank: 1) Macrocephali, 2) Symphyotrichum (linked with Conyzopsis), 3) Tripolium (including Oxytripolium), 4) Aster s.str. (linked with Sericocarpus), 5) Galatella (including Doellingeria), and 6) Oreostemma. Symphyotrichum, comprising the entire x = 8 chromosome number assemblage, certainly could be recognized at the generic level with as much justification as Virgulus. Furthermore, Heleastrum consistently is linked with Virgulus. If the two taxa were segregated as a single genus or placed in the same subgenus, the name Heleastrum would have priority. However, because the supporting evidence for the breakup of Aster rests primarily upon karyotypic characters (i.e., chromosome number and structurecharacters not suitable for identification of plants in the field or herbarium), we propose to retain a more traditional generic concept. The above-mentioned potential segregates, including Virgulus, are considered distinct at the subgeneric level. Beginning with the first monograph of the genus by Nees von Esenbeck (1832), several attempts have been made at segregating certain definable species groups from the large and complex genus Aster L. Nees, for example, established the genus Machaeranthera. Species allied to Aster solidagineus were recognized by him (perhaps justifiably) under Sericocarpus, and the species related to A. umbellatus Miller were included in his new genus Doellingeria. De Candolle established Heleastrum to include A. paludosus Sol. in Aiton, as well as A. ptarmicoides Nees [= Solidago ptarmicoides (Nees) Boivin]. He also treated as generically distinct the Biotian complex, i.e., species related to A. macrophyllus L. Rafinesque (1837) established Virgulus to accommodate A. concolor L., and Greene (1903) described Lasallea, typified by the closely related A. sericeus Vent. Several other generic names were applied by Greene to species traditionally considered as belonging in Aster, e.g., Leucosyris (1897), to include the species now known as A. intricatus (A. Gray) S. F. Blake and A. spinosus Benth.; lonactis (1897), for the species placed by A. Gray (1882) in subg. Ianthe; and Oreostemma (1900), for taxa related to A. alpigenus (Torrey & A. Gray) A. Gray. Nuttall (1841) described Eucephalus to include the western North American species related to A. perelegans Nelson & J. F. Macbr. After extensive study and much agonizing, Gray (1880, 1882, 1884) decided that, on morphological grounds, most of the above-mentioned segregates of Aster could not be maintained at the generic level. Except for Greene, most later authors adopted Gray's broad generic concept for North American taxa. Machaeranthera, however, was reconstituted by Cronquist