Pseudosasa was confirmed as polyphyletic by recent phylogenetic analyses, with Chinese species of Pseudosasa distantly related to those from Japan. Among the Chinese species of Pseudosasa, Pseudosasa pubiflora is a morphologically unique as well as taxonomically problematic species endemic to South China, of which the generic designation is still uncertain. Molecular analyses based on both plastid and nuclear genomic data demonstrated that this species is closest to the recently published genus Sinosasa. Morphologically, the two are somewhat similar to each other in flowering branches developing at the nodes of every order of branches, raceme-like units of inflorescence with 3–5 short spikelets, each spikelet with few florets including a rudimentary one at the apex, and each floret with 3 stamens and 2 stigmas. However, P. pubiflora is very different from Sinosasa species in many reproductive and vegetative characters, such as the morphology of paracladia (lateral spikelet “pedicels”), the absence or existence of pulvinus at the base of paracladia, the relative length of the upper glume and the lowest lemma, the shape of lodicules and primary culm buds, the branch complement, the morphology of nodes, culm leaves and dried foliage leaf blades, and the number of foliage leaves per ultimate branchlet. The morphological and molecular evidence warrants recognition of a new genus to accommodate this unique species, which is here named Kengiochloa. After consulting related literature and examination of herbarium specimens or specimen photos, a taxonomic revision of K. pubiflora and its synonyms was made, and it was confirmed that four names, viz. P. gracilis, Yushania lanshanensis, Arundinaria tenuivagina and P. parilis, should be merged with K. pubiflora, while Indocalamus pallidiflorus and Acidosasa paucifolia are distinct species.

morphology, phylogeny, plastome, synonyms, taxonomy

Pseudosasa Makino ex Nakai, with ca. 20 species mainly distributed in East Asia and Vietnam, belongs to the subtribe Arundinariinae of the tribe Arundinarieae (Poaceae: Bambusoideae) (Zhang et al. 2020; Soreng et al. 2022; POWO 2022). Many molecular phylogenetic studies have shown that this genus is polyphyletic, as the Japanese species including the generic type, P. japonica (Siebold & Zucc. ex Steud.) Makino ex Nakai, and those from China are always distributed in different clades in the phylogenetic trees (Triplett and Clark 2010; Zhang et al. 2012; Guo et al. 2021; Triplett and Clark 2021). Morphologically, the two groups of species are also different: P. japonica has persistent culm leaf sheaths, a solitary branch at each node, or sometimes three branches at the top nodes of the culm, and three to five stamens (which also suggest that P. japonica has a hybrid origin between Japanese Pleioblastus Nakai and Sasa Makino & Shibata or Sasamorpha Nakai, as indicated by Triplett and Clark (2010, 2021)), while Chinese Pseudosasa species usually have deciduous culm leaf sheaths, one to three branches per node, and three stamens (Chen et al. 1996; Zhu et al. 2006; Zhang et al. 2012; Kobayashi 2017). Among those sampled Pseudosasa species, one species endemic to South China, viz. P. gracilis S. L. Chen & G. Y. Sheng, drew the attention of many researchers. This species was consistently associated with some Chinese Sasa species (which were recently transferred to the newly established genus, Sinosasa L. C. Chia ex N. H. Xia et al., see Qin et al. 2021), and far related with other Pseudosasa species in previous studies (Zhang et al. 2012; Guo et al. 2021) (Fig. 1). Zhang et al. (2012) argued that P. gracilis was not a typical member of Pseudosasa in terms of morphology, and suggested that more work was required to clarify the ascription of this species and its phylogenetic placement. In Flora of China, Zhu et al. (2006) considered that P. gracilis was possibly a synonym of P. pubiflora (Keng) Keng f. ex D. Z. Li & L. M. Gao, which they thought may not belong to Pseudosasa but Indocalamus Nakai.

Figure 1. 

Phylogenetic position of Pseudosasa gracilis, adapted from Guo et al. 2021.

Pseudosasa pubiflora has a relatively complicated taxonomic history. In 1936, Yi-Li Keng, the first Chinese botanist who studied the bamboo classification in China, described Arundinaria pubiflora Keng from northern Guangdong based on only one collection (To & Tsang 12284) with flowers (Keng 1936). His son, Pai-Chieh Keng then transferred it to Pseudosasa Makino ex Nakai without giving any reason (Keng 1957 & 1959). However, the new combination was invalid due to lack of a clear and direct reference to the basionym (Zhu et al. 2006). Yang and Chao (1994) recognized this species in Arundinaria Michx.., and considered several other names, including P. gracilis, Indocalamus pallidiflorus McClure, Yushania lanshanensis T. H. Wen and Acidosasa paucifolia W. T. Lin as well as several combinations based on these names, as synonyms of this species.

The P. pubiflora published in Flora of China (Zhu et al. 2006) was a combination largely based on Yang and Chao’s (1994) treatment, except that P. gracilis was suggested as “possibly a synonym of P. pubiflora”. Besides, two more names, viz. Arundinaria teuivagina W. T. Lin and P. parilis T. P. Yi & D. H. Hu, were further synonymized with P. pubiflora. Consequently, the distribution of this species was extended to include three provinces, i.e., Guangdong, Hunan and Jiangxi. Moreover, Zhu et al. (2006) argued that, for P. pubiflora, “the woolly internode apex is unusual in Pseudosasa and rather suggestive of Indocalamus Nakai”. Thus, the relationships between P. pubiflora and its synonyms and the generic designation of this species need a further study.

Field trips to the type locality of P. pubiflora as well as its several synonyms including I. pallidiflorus, Arundinaria teuivagina and P. parilis were conducted from 2018 to 2022. The complete specimens including rhizomes, culms with culm leaves, branches and foliage leaves as well as fresh young foliage leaves used for molecular analyses were collected. The types and other specimens of the pertinent species housed at CANT, IBSC, N, SYS and ZJFI, as well as photographs of specimens housed at K, US and W, were examined. Herbaria acronyms follow Thiers (2022). The morphological description is based on specimens and referred to related literature (Keng 1936; Chen et al. 1983; Wen 1985, 1986; Lin and Wu 1990; Yi 1995; Chen et al. 1996; Zhu et al. 2006). General morphological terms follow Beentje (2016). The terms applied to the flowering structure in this study mainly follow the synflorescence concept applied to grasses (Vegetti and Anton 1996; Tivano et al. 2009; Cai and Xia 2021).

To study the phylogenetic position of P. pubiflora within the tribe Arundinarieae, the whole chloroplast genomes were used for building the phylogenetic tree. Five Pseudosasa species including one from Japan, also the type species of the genus, viz. P. japonica, and four from China were sampled. In total, there were 29 species represented by 30 samples of all the five subtribes in the tribe Arundinarieae (two samples for P. pubiflora) and one sample of Bambusa bambos (L.) Voss from the tribe Bambuseae was the outgroup. Species names, voucher information and GenBank accession numbers are provided in Table 1.

Total genomic DNA was isolated from silica gel-dried leaves using TIANGEN Genomic DNA Extraction Kit (TIANGEN, Beijing, China). The extracted genomic DNA was fragmented randomly by Covaris M220 (Covaris, Woburn, MA), and the fragments with insert size of 350 bp were selected by using AxyPrep Mag PCR Clean Up Kit. The selected fragments were enriched by PCR after undergoing end repair, the addition of polyA tail and adaptor ligation. The paired-end (2 × 150 bp) libraries were constructed on NovaSeq 6000 platform. Finally, 20 Gb genome skimming data were generated for each sample.

After quality control of 20 Gb clean data by Trimmomatic v 0.39 (Bolger et al. 2014), the whole chloroplast (cp) genomes were assembled using the software GetOrganelle v 1.7.4 pipeline (Jin et al. 2018), with the reference cp genome sequence of Ampelocalamus sinovietnamensis (MW525255) and K-mer sizes of 45, 65, 85, 105 and 125. The generated whole cp genome was annotated using the program Plastid Genome Annotator (Qu et al. 2019) based on the annotation of A. sinovietnamensis. The transferred RNAs (tRNAs) were adjusted using the software tRNAscan-SE service (Schattner et al. 2005). The initiation and termination codons of all the coding genes were manually verified in Geneious Prime v 9.1.4 (Kearse et al. 2012).

All the cp genomes were aligned using MAFFT v. 7.450 (Katoh and Standley 2013). Maximum Likelihood (ML) analysis was conducted by RAxML v 8.2.12 (Stamatakis 2014) with the rapid bootstrap algorithm. GTRGAMMAI was selected as the best-fit model recommended by jModelTest v2.1.6 (Darriba et al. 2012). The number of 12345 was specified as the random seed of parsimony tree inference with 1000 replicates performed. For Bayesian Inference (BI), the data matrix was calculated using MrBayes v3.2.2 software (Ronquist et al. 2012). The best-fit model was selected as SYM+G under the Akaike information criterion (AIC) using MrModeltest v 2.3 (Nylander 2004). Rates of variations across sites were trimmed as gamma. For each analysis, two simultaneous runs of four Monte Carlo Markov Chains (three heated and one cold) were run for 20 million generations with a random tree as the starting point and saving trees every 1000 generations. After discarding the first 25% samples as burn-ins, the optimized topology was generated. The final results were visualized with Figtree 1.4.3 (http://tree.bio.ed.ac.uk/software/figtree/).

The chloroplast genomes ranged from 139,093 bp (Pleioblastus chino Makino) to 140,064 bp (Gaoligongshania megalothyrsa (Hand.-Mazz.) D. Z. Li, Hsueh & N. H. Xia) with an alignment length of 144,771 bp. The plastid data matrix was characterized by sequence divergence with 9,751 variable sites (6.74%), including 6,477 parsimony informative sites (4.47%) and 3,274 singleton variable sites (2.27%).

The reconstructed phylogenetic tree is shown in Fig. 2. Similar to some previous analyses on plastid sequence data (Triplett and Clark 2010; Zhang et al. 2012), Pseudosasa was resolved as polyphyletic in this analysis. Pseudosasa pubiflora was sister to the clade of Sinosasa with strong support (maximum likelihood bootstrap value/Bayesian inference posterior probability BS/PP = 100%/1.00). The other Pseudosasa taxa form a clade with five other genera of the subtribe Arundinariinae, namely Oligostachyum Z. P. Wang & G. H. Ye, Acidosasa B. M. Yang, Sinobambusa Makino, Indosasa Maclure and Pleioblastus (MLBS/PP = 100%/1.00).

Figure 2. 

Phylogeny of Arundinarieae based on whole chloroplast genome derived from maximum likelihood and Bayesian analyses. Only bootstrap values (BS) ≥ 70% and posterior probabilities (PP) ≥ 0.95 are indicated at each node.

We are grateful to Dr. Si-Jin Zeng, Mr. Bing-Mou Wang, Mr. Yuan-Qiu Li, Ms. Xue-He Ye and Mr Chuan Long for help during field trips. Mr. Xi-Rong Zheng is thanked for providing the photo of Indocalamus sinicus. This work was supported by the Natural Science Foundation of Guangdong Province (no. 2114050002139) and National Natural Science Foundation of China (grant no. 32270227).