New locality in Yunnan is occurred in Tongbiguan Provincial Nature Reserve (Dehong, 24°19'55.68"N, 97°43'45.51"E, alt., 1443m). The Yunnan accession (Fig. 1A–E) shared with the Hainan accession (Mt. Yinggeling, 19°24'15.9"N, 109°32.07'E, alt., 1749m; Fig. 1F–H) the characters of short, erect, radial rhizome (Fig. 1C, 1F), non-articulated leaves with simple lamina and short petioles, concolorous and glabrous rhizome scales, free veins, absence of hydathodes, branches hairs attached to the lamina (Fig. 1D, 1G), and coenosori arranged in parallel to the margins (Fig. 1E, 1H). However, the Yunnan accession differed from the Hainan accession in the lamina texture and positioning of the coenosori. The Yunnan accession (YunAcc) has a thick lamina margin with the coenosori arranged in submarginal grooves (Fig. 1E), whereas the Hainan accession (HaiAcc) shows a thin lamina margin with the coenosori arranged in grooves on the abaxial surface of the lamina (Fig. 1H). Compared with the established species differentiation in the genus Scleroglossum, the Hainan accession should be identified as S. sulcatum, whereas the Yunnan accession should be identified as S. pusillum. The Yunnan accession differs from S. pyxidatum because the lamina was not narrowed below the sori and wider than 2 mm.

Figure 1. 

The grammitid genus Scleroglossum in China. A–E Scleroglossum pusillum in Yunnan (YunAcc) and F–H Scleroglossum sulcatum in Hainan (HaiAcc). A habitat of S. pusillum occurrences in Yunnan B germinated green spore recovered from opened mature sporangia of S. pusillum. The remains of the spore wall are visible at the lower part of the larger of the two cells that both contain mature chloroplasts. The smaller cell is the first daughter cell formed by the first cell division C habit of S. pusillum D close up of the dorsal surface of the simple leaves showing the location of the submarginal sori and the occurrences of brown branched hairs of S. pusillum E close up of the sorus structure showing the placement at the submargin of the leaves in S. pusillum. F habit of S. sulcatum G close up of the dorsal surface of the simple leaves showing the location of the not submarginal sori and the occurrences of brown branched hairs of S. sulcatum H close up of the sorus structure showing the placement of the sori in dorsal grooves and a distinct lamina margin in S. sulcatum. Scale bars: 0.02 mm (B); 1 mm (E, H).

Pairwise comparisons of the rbcL sequences of the seven distinct Scleroglossum sequences recovered a sequence variation between 99.7% and 100% (Table 2). The two accessions of S. wooroonooran (KM218809, KM218810) showed a similarity of 99.4%, whereas the accessions of S. pusillum from Malay Peninsula (KM218812) and Thailand (KM211812) showed a similarity of 99.4%. The accession obtained in Yunnan (YunAcc in Table 2) showed a similarity of 99.9% with the S. pusillum from Thailand and a 99.3% similarity with S. pusillum of the Malay Peninsula. The Hainan accession (HaiAcc) and Yunnan accession (YunAcc) showed a similarity of 98.9%. Accessions of S. sulcatum obtained from Pacific Islands and Vietnam showed a similarity between 100% and 99.5%. The Yunnan accession and Hainan accession showed similarities to accessions of S. sulcatum below 99.0% (Table 2).

The phylogenetic reconstruction recovered the Yunnan accessions to be nested in clade comprising the two Scleroglossum pusillum accessions (Fig. 2), whereas the Hainan accessions formed an independent lineage that was part of the polytomy involving the S. pusillum and S. wooroonooran. The Vietnam accessions nested together with previously obtained accessions of S. sulcatum occurring on the Pacific Islands (Moorea, Pohnpei).

Figure 2. 

Reconstruction of the phylogenetic relationships of Scleroglossum species occurring in China including all accessions available in GenBank (October 2018). Newly generated accessions are printed in bold: HaiAcc = Hainan Accessions, two sequences obtained from individual A and B; YunAcc = Yunnan Accessions; VieAcc1 and VieAcc2 = two accessions obtained from the herbaria collections at KUN representing two independent collections (see Suppl. material 1: Table S1). The areas of origin are given for Scleroglossum GenBank accessions. The presented consensus phylogram is based on a Bayesian inference of Phylogeny including 690 accessions represented by rbcL sequences. Clades composed by non-Scleroglossum accessions were pruned. Two stars mark clades with a posterior confidence of p = 1.00 whereas one star marks clade with a posterior confidence p ≥ 0.95 and < 1.00. p.p. = pro parts to indicate the polyphyly of S. sulcatum.

The Hainan and Yunnan accession are distinct in both phenotype and genotype. Thus, two instead of one species of Scleroglossum occur in mainland China. Both genotypic and phenotypic evidence support the conclusion that the Yunnan accessions belong to S. pusillum. This is consistent with the original report of this taxon in Yunnan (Lu and Zhang 1994). Phylogenetic analyses supported S. pusillum occurring in Thailand as the closest relative of the Yunnan occurrences. In contrast, the taxonomic treatment of the Hainan accession is less clear. Phenotypic evidence supports the treatment as S. sulcatum, as established previously in the FOC treatment (Moore and Parris 2013). However, the genotypic evidence did not support this conclusion because the Hainan accessions were distinct from all other accessions obtained so far for Scleroglossum including not only S. pusillum (KM218812; KY712079) and S. wooroonooran (KM218809, KM218810), but also S. sulcatum. The phylogenetic analyses recovered a clade comprising accessions with the S. sulcatum morphology collected in the Pacific Islands of Moorea (KY099861) and Pohnpei (AY40664, AY40665) as well as the two newly obtained accessions from Vietnam. Thus, the Hainan accession may represent a new species which is morphological distinct from the Yunnan accession but not from the S. sulcatum accessions from Vietnam. Morphological evidence supports the occurrence of the S. sulcatum morphotype also in Guangxi, Guangdong, and Taiwan. However, we currently lack genotypic evidence to assign these specimens to the genotypic differentiated forms of S. sulcatum.

Species of Scleroglossum occurring in mainland China can be identified using the morphological key presented for Vietnam species (Parris et al. 2015). However, we encountered a conflict between the phenotypic and genotypic variation because S. sulcatum was recovered as polyphyletic (Fig. 2). This conflict may be explained as a consequence of cryptic speciation (Bickford et al. 2007). Conflicting results between genotype- and phenotype- based species recognition have been reported for several recent diverging fern lineages (Paris et al. 1989) and epiphytic liverworts colonizing similar habitats as Scleroglossum (Yu et al. 2013). However, the taxonomic treatment of cryptic or semi-cryptic species is challenging as a consequence of theoretical and practical issues (Jörger and Schrödl 2013). Given the small sample size, we recommend a treatment that is based on the phenotypic variation, meaning S. sulcatum, until more evidence will be obtained. This approach avoids unstable taxonomic solutions which are important especially given the lack of evidence concerning the genotype of the type collection of this widespread species. The type was collected on the island of Sri Lanka (lectotype: Thwaites 3807; see Parris et al. 2015) however none of the currently available rbcL sequences of this species has been obtained from Sri Lanka. Given the cryptic phenotypic differentiation of the two genotypes, we cannot conclude to which the type may belong. Thus, the introduction of a new name may create more confusion but we must stress that the taxonomic status of the populations currently treated as S. sulcatum require further study.

Besides this problem, DNA barcoding using the rbcL region appears to be sufficient to resolve the DNA based identification of grammitid ferns in China with all species included having a distinct rbcL sequence. However, several species such as the Hainan endemic Oreogrammitis hainanensis Parris require to be studied to confirm that this conclusion is true for all grammitid ferns occurring in China. So far, rbcL sequences have been obtained for only 45.9% of the Chinese grammitid species. Representatives of the four genera including more than one species in China that were presented with at least two species – namely Calymmodon C. Presl, Micropolypodium Hayata, Oreogrammitis Copel., and Prosaptia C. Presl – were distinct from each other in our dataset. Some of the Chinese species were not clearly distinct from closely related non-Chinese species, e.g. P. contigua and P. obliquata. A further limitation of rbcL based barcoding is the problem differentiating some of the proposed generic concepts, such as Oreogrammitis Copel., Radiogrammitis Parris, and Themelium (E. Four.) Parris (see also Sundue et al. 2014). The generic concept of these ferns requires arguably further attention besides the urgent need to focus more on the species taxonomy.

In turn, the application of DNA barcoding recovered a conflict between taxonomic treatments based exclusively on phenotypic evidence with the genotypic evidence. This is consistent with the hypothesis that the frequent employment of these approaches will not only help to resolve conflicting interpretations of phenotypic differentiation (see Yu et al. 2013, Zhou et al. 2016) but also recover more cases of cryptic/ semi-cryptic species differentiation (Pickford et al. 2006). Given the species richness of the fern floras of the southern provinces of China, future studies integrating phenotypic and genotypic evidence may recover more instances of cryptic or semi-cryptic taxa. Some of these taxa may turn out to be the consequence of hybrid-speciation (e.g. Liu et al. 2018) but other processes such as limitations to the accessible morphospace may contribute also to the accumulation of cryptic fern diversity (see Schneider 2016). The genus Scleroglossum as well as other grammitid ferns is arguably well suited to explore the role of rapid diversification and ecological conservatism in the decoupling of the accumulation of species diversity and morphological disparity (Schneider 2016). As shown by Liu et al. (2018), the interpretation of DNA barcode based species identification may result in misleading species assessment if the ploidy level has not been assessed. However, relatively little attention has been given to the study of polyploidy in grammitid ferns so far. About 23% of the published 43 chromosome counts of grammitid ferns indicate the occurrence of polyploidy (Schneider unpublished), which is substantially lower than the frequency of polyploidy in Asplenium (Schneider et al. 2017). However, no chromosome counts have been published for the genus Scleroglossum yet.

The recovered success of DNA based identification of Scleroglossum species may enable future studies to report not only the distribution of the sporophytic generation but also the distribution of the gametophytic stage of the life cycle. A recent study provides distinct distribution patterns observed for fern gametophyte and sporophyte generations of the same species on the Pacific island of Moorea (Nitta et al. 2017). This is especially important to studies focusing on grammitid ferns because this fern lineage includes gametophytes reproducing independently of their sporophytes in geographically disjunct populations (Farrar 1967; Dassler and Farrar 2001).

Given the global distribution of grammitid ferns (see Sundue et al. 2014; Bauret et al. 2017), Chinese grammitid ferns occur mainly in the two larger islands namely Hainan (10 spp., 27% of the Chinese species diversity) and Taiwan (28 spp., 76% the Chinese diversity) and preferably in the southern provinces of Guangdong (8 spp., 22% the Chinese diversity) and Guangxi (9 spp., 24% the Chinese diversity). The remaining provinces with wet subtropical to tropical climate house four (Fujian, Guizhou, Hunan, Yunnan, Zhejiang), two (Jiangxi, Sichuan, Xizang), or one (Anhui) species (Table 1). The low number of grammitid species recorded until now in Yunnan is a bit surprising given the remarkable fern richness of this province especially in the tropical southern parts (Zhou et al. 2016).

Analyses focusing on the similarities among the grammitid floras of Chinese provinces recovered similarities (Fig. 3) between the grammitid rich islands of Hainan (10 spp.) and Taiwan (28 spp.) and the two southern provinces Guangdong (8 spp.) and Guangxi (9 spp.). The East China provinces Fujian (4 spp.), Zhejiang (4 spp.), and Jiangxi (2 spp.) formed together a cluster (Fig. 3), whereas Anhui (1 spp.) was nested in cluster including Southwest China provinces, namely Sichuan (2 spp.), Xizang (2 spp.), Yunnan (4 app.), Guizhou (2 spp.), and Hunan (2 spp.). These results are consistent with the prediction that the grammitid diversity is mainly shaped by differences in the climatic conditions but this requires further analyses.

Figure 3. 

Similarity clustering of the Chinese provinces that are home to at least one grammitid ferns. The clustering analysis was carried out using Jaccard distances and the NJ-tree building algorithm. The underlying scoring (Table 1) was carried out as absence (0) and presence (1).

Several species of grammitids including Scleroglossum sulcatum were listed in the Red List of Chinese Plants (Qin et al. 2017), but none of the Chinese species have been listed in the list of rare and threatened species of Asia (Ebihara et al. 2012). Furthermore, the more recent list did not mention any species of Scleroglossum as rare or threatened. This difference is arguably the consequence of the context differences of a national (Qin et al. 2017) and a global (Ebihara et al. 2012) focus of the assessment. As a consequence of our results, the Red List of Chinese Plants needs some changes. In 2017, S. sulcatum was listed as VU D1+2 (Qin et al. 2017), but this assessment is likely based on the assumption of a single species with occurrences in five provinces of China (Yan et al. 2016). Given the recognition of the Yunnan occurrences as S. pusillum, the range of S. sulcatum has been reduced from five to four provinces of China. However, the current range is sufficient to maintain the status reported (Qin et al. 2017). In contrast, the Chinese populations of S. pusillum may be highly vulnerable but further assessments are needed. We recorded this species for the first time to occur in western Yunnan namely Mt. Tongbiguan, Dehong but were unable to confirm the previous records in southeastern Yunnan namely Mt. Laojunshan, Malipo. However, both species are considered as least concern (LC) considering the distribution range comprising most of tropical Asia and some islands in the Pacific and Indian Oceans.

In our study, we are able to confirm the occurrences in both Yunnan and Hainan. As mentioned above we obtained a new record from western Yunnan but failed to reconfirm the occurrence in southeastern Yunnan. During fieldwork we confirmed the occurrence of these ferns in one of the two locations recorded in Hainan namely the Mt. Yinggeling population but did not recollect the Mt. Wuzhishan population. The Guangxi and Guangdong occurrences have been surveyed in recent years (Zhou and Li 1993, Yan et al. 2007, Zhang and Li 2011, Jiang and Liu 2014). Thus, all recorded populations of these ferns in China are arguably stable despite the restriction to isolated localities result in local threats. All known Chinese occurrences are within national parks that will arguably enable effective protection of these populations. However, we still recommend that the number of individuals need to be regularly assessed. At the current stage, the IUCN red list assessment are based on species identification using the phenotypic differentiation only and thus they are arguably misleading for lineages including cryptic- or semi-cryptic species as recovered in S. sulcatum (see above). Thus, assessments of the distribution of S. sulcatum are arguably inaccurate. Furthermore, very little information exists about the number of individuals per site which limits any assessment of the potential dynamics of population size shrinking or expansion. Finally, the distribution of the two Scleroglossum species throughout tropical SE Asia and Pacific Islands may lack the accuracy requested to enable reliable IUCN assessments. Many older collections of these species lack detailed information about the collecting sites and several putative occurrences require to be re-confirmed by recollecting.