Phylogenetic Placement of Rhopalobrachium fragrans (Rubiaceae): Evidence from Molecular (rpsl? and trnT-F) and Morphological Data

Rhopalobrachium is an endemic New Caledonian genus of Rubiaceae with controversial tribal position, because its protologue was apparently based on two distantly related species, Rhopalobrachium fragrans and Rhopalobrachium congestum. Despite its new circumscription as a monotypic genus (Rhopalobrachium fragrans) the tribal placement of Rhopalobrachium remained unknown. We investigated, for Rhopalobrachium fragrans, all morphological characters currently used for circumscribing tribes in the subfamily Ixoroideae sensu lato. The species has axillary inflorescences, with valvate flower aestivation, drupaceous fruits with a single pendulous ovule per locule, and 3-pororate pollen grains. We also performed Bayesian and parsimony analyses based on two chloroplast markers, rpsl6 intron and trnT-F region, to assess the phylogenetic position of Rhopalobrachium within Ixoroideae sensu lato. These analyses produced well-resolved and strongly supported topologies. The results of both the morphological survey and phylogenetic analyses indicated a position of Rhopalobrachium fragrans in the tribe Vanguerieae and a close relationship with the Pacific genus Cyclophyllum. Rhopalobrachium Schltr. & K.Krause is an enig matic genus of the large cosmopolitan family Rubiaceae (Robbrecht 1988), which was originally described by Schlechter and Krause (1908) to accommodate two New Caledonian species, R. congestum Schltr. & K.Krause and R. fragrans. These authors did not select the type species of their new genus. In the original publication, Rhopalobrachium was described with the following characters: "stipules entire, flowers hermaphro dite, ovary 2-locular with a single pendulous ovule per locule, corolla infundibuliform"; the fruits were unknown. Later, Guillaumin (1930) emended the generic description with fruit char acters: "fruits bacciform, 1.5 cm wide, with black ish, persistent calyx lobes, and containing four seeds". The two seeds per locule in Rhopalobra chium reported by Guillaumin (1930) were, how ever, inconsistent with the single ovule per locule cited in the protologue.