Widely acclaimed but poorly named: Phylogeny and systematics of the African daisy genus Dimorphotheca (Asteraceae, Calenduleae)

Abstract

Phylogenetic relationships in the South African daisy genus Dimorphotheca have long been uncertain, with the taxonomy of the genus relying on a few morphological traits, most prominently capitulum sexual system (e.g., cypsela type) and ray colour, which may not be evolutionarily conserved. Here we present the first well‐sampled molecular phylogeny of Dimorphotheca, based on nuclear ribosomal (ITS, ETS) and plastid trnL‐trnF region DNA sequences from multiple accessions per species. Although the relationships suggested by these markers are broadly congruent, we do find some instances of incongruence which we resolve using a combined decomposition and deletion approach. Using our best estimate of phylogenetic relationships, we reconstruct the evolution of capitulum fertility and ray colour to assess the evolutionary conservatism of these traits and their taxonomic utility. We find support for the monophyly of Dimorphotheca, excluding the recently segregated Osteospermum polypterum, and our data thus support the modern, enlarged circumscription of the genus incorporating the former genus Castalis and Osteospermum sect. Acanthotheca and O. sect. Blaxium. Major subclades within Dimorphotheca are largely cohesive in terms of geographic distribution and morphological traits such as growth form and cypsela structure. While many species are resolved as monophyletic, the polyphyly of a few species suggests a need for taxonomic re‐evaluation. On the basis of both morphological and molecular data, we describe one new species, and elevate one variety to species level. A full taxonomic key to the enlarged genus is presented for the first time. Ancestral reconstructions show that capitulum sexual system and ray lamina colour are not evolutionarily conserved and that neither can therefore be used to delimit major lineages within Dimorphotheca. While our findings resolve some taxonomic problems, they also highlight the need for further species‐level taxonomic work on Dimorphotheca.