Abstract: The classification of species in the genus Pterocyclus has been a topic of concern, as they were initially considered members of Pleurospermum (Apiaceae) in the Flora of China. However, the intergeneric and infrageneric classifications of Pterocyclus and its allies, such as Hymenidium and Pleurospermum, have been controversial. To address these issues, we performed phylogenetic analyses using one nuclear marker (nrITS) from 167 accessions, including 137 species (including five species of Pterocyclus), and 105 plastid genes from 82 accessions representing 82 species (including five species of Pterocyclus) of Apiaceae. We used both maximum likelihood and Bayesian approaches to deduce phylogenetic relationships of Pterocyclus and its allies. Furthermore, we conducted morphometric analyses that specifically targeted the mericarp morphology, and compared the structural differences in their plastid genomes. Based on these findings, we suggest the following revisions to the classification of Pterocyclus and its allies: (i) Pterocyclus should encompass five species (Pterocyclus angelicoides, Pterocyclus forrestii, Pterocyclus rotundatus, Pterocyclus tibeticus, and Pterocyclus wolffianus), all of which are all monophyletic and placed in the Komarovieae; (ii) Pterocyclus rivulorum shows both phylogenetic and mericarp morphological similarities to Hymenidium apiolens and Hymenidium dentatum, which belongs to the Hymenidium Clade, thus we provisionally transfer it to Hymenidium; (iii) Pleurospermum longicarpum is considered conspecific with Pterocyclus angelicoides based on morphological descriptions in the Flora Xizangica; (iv) we describe and illustrate Pterocyclus tibeticus as a new species; and (v) Pterocyclus wolffianus should be recognized as a valid species and not considered a synonym of Pterocyclus forrestii.

Key words: Apiaceae, morphology, phylogeny, plastid genome, Pleurospermum, Pterocyclus, taxonomy