Systematics of North American Petasites (Asteraceae: Senecioneae). I. Morphometric analyses
Petasites Mill. (Asteraceae: Senecioneae) is a taxonomically difficult genus of polygamodioecious, perennial clonal herbs distributed across Canada, Alaska, and the northern contiguous United States. The taxonomic confusion surrounding the genus is reflected in the ambiguity of the various classification systems currently used. As many as ten species and as few as two species have been recognized (Petasites sagittatus (Banks ex Pursh) Gray and Petasites frigidus (L.) Fries with three varieties), in addition to one hybrid taxon. To determine if discrete groups could be recognized according to a combination of vegetative and (or) reproductive characters and thereby clarify the phenetic relationships among the taxa of Petasites, morphological data were subjected to principal components and discriminant analyses. Particular emphasis was placed on P. frigidus var. frigidus, P. frigidus var. nivalis sensu Bogle, and the purported hybrid between these taxa, which exhibits an intermediate morphology (the frigidus complex). Additionally, the specific status of P. sagittatus was investigated. Evidence obtained from multivariate morphometric analyses indicates that North American Petasites is comprised of four closely related polymorphic entities. While morphological variation among different taxa of Petasites is continuous for most reproductive characters, leaf characters were determined to be the most taxonomically discriminating. The entities of the frigidus complex form a cohesive group somewhat distinct from the other taxa. The morphological differentiation between P. sagittatus and other taxa of Petasites is not considered sufficient to warrant recognition of specific status. Therefore, we recommend that only one polymorphic species of Petasites in North America be recognized (P. frigidus with three varieties) in addition to one hybrid taxon originating from a cross between two of the varieties. Key words: Petasites, Asteraceae, North America, systematics, principal components analysis, discriminant analysis.