A prickly puzzle: generic delimitations in the Carduus-Cirsium group (Compositae: Cardueae: Carduinae)

Abstract

Generic delimitations within the cosmopolitan Carduus‐Cirsium group (i.e., “thistles”) have a long history of taxonomic confusion and debate. We present the most comprehensive molecular phylogeny of the group to date to test generic limits, reconstruct the evolution of pappus type, and elucidate the role of chromosomal evolution. We offer two solutions for the recognition of monophyletic genera: (1) consolidate all taxa into one large genus (Carduus or Cirsium), or (2) recognize each major clade as a genus (Carduus, Cirsium, Eriolepis, Notobasis, Picnomon, Silybum, and Tyrimnus). Under the second proposal, the cryptic genus Eriolepis is segregated from Cirsium, and the African Carduus are included within Cirsium. The best diagnosable morphological character to delimit the genera is pollen type, which is not practical in field‐based application. We caution that prior to implementing either solution, a thorough, comprehensive morphological analysis of all current members of Cirsium sect. Epitrachys (= genus Eriolepis) be completed. Future morphological studies may find additional achene or leaf surface characters that could be used for practical field identification of the segregate genera. The data show that the plumose pappus state is symplesiomorphic for the group, with one transition to barbellate pappus, likely followed by a reversal to its ancestral state as the group colonized Eurasia. The data are consistent with a North African origin in the region of the Mediterranean and a single colonization event to North America. An ancestral chromosome state of n = 17 is hypothesized for the group, and a descending dysploidy series in Carduus is hypothesized to correspond with the aridification of the Mediterranean region. The Carduus‐Cirsium group highlights the difficulty of delimiting morphologically similar, cryptic genera.