Phylogenomic and morphological evidence supports the reinstatement of the bamboo genus Clavinodum from Oligostachyum (Poaceae: Bambusoideae)

One of the most intractable problems in bamboo systematics concerns the three-branched bamboos of tribe Arundinarieae (Poaceae: Bambusoideae), which are collectively characterized by having three branches per mid-culm node. Previous phylogenetic studies based on double-digest restriction-site associated DNA sequencing (ddRAD-seq) data confirmed that Oligostachyum, a member of this group, is non-monophyletic, and in particular that Oligostachyum oedogonatum is a problematic species deserving further attention, as it appears to be morphologically and phylogenetically distinct from the other three-branched bamboos. Here we aim to define and confirm the phylogenetic position of O. oedogonatum, by including representatives from across its geographic range. We also provide new insights into the overall phylogeny of the three-branched bamboos and closely related genera, using multiple phylogenomic datasets. While a plastid genome-based tree is very poorly supported, phylogenies inferred using two sets of conserved nuclear genes and single nucleotide polymorphism (SNP) data yield generally well-supported and congruent topologies using coalescent-based approaches. The tree inferred from the largest concatenated gene set is the most dissimilar to other inferences. The nuclear-based data sets all recover a major clade that includes all of the three-branched bamboos and four other genera, which can be distinguished from related taxa due to their possession of three stigmas per floret and leptomorph rhizomes. Notably, eight O. oedogonatum samples form a clade that is distantly related to other members of Oligostachyum, including the type species of the genus (Oligostachyum sulcatum). Population genetic approaches and multi-species coalescent-based analysis of the nuclear data sets imply that seven of these populations can be treated as a single species, O. oedogonatum, but that one population from Jinggangshan is likely an intergeneric hybrid between O. oedogonatum and Pleioblastus. Morphologically, O. oedogonatum differs from all other three-branched bamboos, in having strongly asymmetrically swollen supra-nodal ridges, a laterally compressed spikelet, and rachilla segments that disarticulate readily below fertile florets. Because of its morphological distinctiveness and molecular phylogenetic position, we propose that this taxon should be reinstated as Clavinodum oedogonatum. We provide a new description for this monotypic genus here.