﻿Molecular phylogeny and taxonomy of Hosta (Asparagaceae) on Shikoku Island, Japan, including five new species, one new subspecies, and two new status assignments

Tetsukazu Yahara; Shun K. Hirota; Seiko Fujii; Yasushi Kokami; Kengo Fuse; Hiroyuki Sato; Shuichiro Tagane; Yoshihisa Suyama

doi:10.3897/phytokeys.235.99140

Abstract

Japan has 16 native species of the genus Hosta Tratt. (Asparagaceae). A recent study on Hosta based on field surveys and molecular phylogenetic analyses resulted in the discovery of six unknown taxa in Kochi Prefecture, Shikoku Island, southwestern Japan. We aimed to identify these unknown taxa. Therefore, we constructed a finely resolved phylogeny for 320 Hosta samples collected from the Honshu, Shikoku, and Kyushu Islands using multiplex inter-simple sequence repeat genotyping by sequencing (MIG-seq). Based on this phylogenetic analysis and related morphological observations, we describe five new species, H. longipedicellata sp. nov., H. minazukiflora sp. nov., H. polyneuronoides sp. nov., H. samukazemontana sp. nov., and H. takiminazukiflora sp. nov. and one new subspecies, H. takiminazukiflora subsp. grandis subsp. nov. In addition, we propose two new status assignments, H. tardiva subsp. densinervia comb. and stat. nov. and H. scabrinervia stat. nov. We also propose classifying H. kikutii var. tosana as a species, H. tosana. Further studies that combine MIG-seq with careful morphological observations are needed for Hosta plants on all Japanese islands, which may result in the discovery of even more undescribed species.

Key words

anacladogenetic speciation, flowering season, MIG-seq, next generation sequencing, reproductive isolation, threatened species

Introduction

The genus Hosta Tratt. (Asparagaceae) comprises 22–25 species endemic to East Asia and Russia (Jones 1989). Japan has 16 native species of Hosta (Fujita 1976; Tamura 2015; Tamura and Fujita 2016; Yahara et al. 2021); this is a higher species richness than is present in China (four species), Korea (six species), and Russia (one species) (Chen and Boufford 2000; Lee et al. 2019). Recently, Yahara et al. (2021) conducted a molecular phylogenetic study of Hosta species on Kyushu Island, Japan, and described a new species, Hosta alata Hatus. ex Yahara, based on a finely resolved phylogenetic tree reconstructed using multiplexed inter-simple sequence repeat genotyping by sequencing (MIG-seq) as well as morphological and ecological observations.

Here, we examined the molecular phylogeny and taxonomy of Hosta on Shikoku Island, located east of Kyushu Island, which has the highest diversity of Hosta in Japan. According to Fujita (1976) and Kochi Prefecture & Makino Memorial Foundation of Kochi Prefecture (2009), the following nine species (11 taxa) were recorded in Shikoku: H. alismifolia F. Maek., H. capitata (Koidz.) Nakai, H. kikutii F. Maek. var. caput-avis (F. Maek.) F. Maek. and var. polyneuron (F. Maek.) N. Fujita, H. longipes (Franch. & Sav.) Matsum. var. caduca N. Fujita and var. gracillima (F. Maek.) N. Fujita, H. longissima F. Maek., H. shikokiana N. Fujita, H. sieboldiana (Hook.) Engl., H. sieboldii (Paxton) J. W. Ingram, and H. tardiva Nakai. Among these species, H. kikutii in Shikoku is highly polymorphic. Fujita (1976) noted that eight H. kikutii populations distributed in different river basins of Shikoku showed remarkable variation in leaf and flower morphologies and flowering phenology. Fujita and Tamura (2008) described two new varieties of H. kikutii from Shikoku: H. kikutii var. densinervia N. Fujita & M. N. Tamura and var. scabrinervia N. Fujita & M. N. Tamura. These two varieties were described as having different leaf vein characteristics: abaxially slightly papillose (densinervia) and prominently papillose (var. var. scabrinervia). These two varieties were distinguished from H. kikutii var. tosana (F. Maek.) F. Maek. distributed in eastern Shikoku by their straight flowering stems (Fujita and Tamura 2008 treating this variety as H. kikutii var. caput-avis (F. Maek.) F. Maek., Tamura and Fujita 2016 as var. tosana).

This study was initiated based on the findings of our previous research, which identified Hosta alata Hatus. ex Yahara of Kyushu Island as a new species (Yahara et al. 2021) as well as preliminary studies on plants identified as H. kikutii on Shikoku Island. These findings led to the development of the following hypotheses. First, we hypothesized that three varieties of H. kikutii in Shikoku are not closely related to H. kikutii var. kikutii collected from Kyushu Island, which is known to be related to H. longipes var. caduca (Yahara et al. 2021). Second, we hypothesized that H. kikutii var. tosana is not closely related to H. kikutii var. densinervia and var. scabrinervia. Third, because we found that two neighboring populations of “H. kikutii var. densinervia” distributed along the Asemi River on Shikoku Island are morphologically distinct and have different flowering seasons (one population flowering in June and fruiting in August, whereas the other flowering in August and fruiting in September), we hypothesized that the two populations have diverged into two distinct biological species.

To elucidate the taxonomy of Hosta plants previously identified as H. kikutii in Shikoku, we collected DNA samples of Hosta plants from Shikoku, Kyushu and Honshu. Following a previous study on H. alata (Yahara et al. 2021), we utilized MIG-seq to obtain a finely resolved phylogeny. The objective of this study is to present the results of molecular phylogenetic analyses and subsequent morphological observations, and revise the taxonomy of Hosta spp. on Shikoku.

Materials and methods

Field surveys and samples examined

We collected 320 DNA samples and voucher specimens from 70 localities for 30 taxa of Hosta in Japan (Suppl. material 1). These collections include specimens of three varieties distinguished by Tamura and Fujita (2016), H. kikutii var. densinervia, var. scabrinervia , and var. tosana, from their respective type localities. We attempted to identify our specimens based on the diagnostic traits described by Fujita and Tamura (2008) and Tamura and Fujita (2016), and the observation of the type localities for these varieties. However, we encountered many specimens that could not be definitively classified into any of the three varieties. As a result, we assigned tentative names to these specimens during our field surveys and subsequent molecular analyses. The MIG-seq data for the H. alata group determined in a previous study (Yahara et al. 2021; accession number DRA011465) were also included in the phylogenetic analysis.

DNA isolation, sequencing, and construction of SNP-based phylogenetic trees

Total DNA was extracted from the dried leaves using the CTAB method (Doyle and Doyle 1990). De novo single nucleotide polymorphism (SNP) discovery was performed using MIG-seq (Suyama and Matsuki 2015). A MIG-seq library was prepared using a two-step PCR amplification process based on the method described by Suyama et al. (2022). Sequencing was performed using the Illumina MiSeq platform (Illumina, San Diego, CA, USA) and MiSeq Reagent Kit v3 (150 cycles, Illumina). Low-quality and extremely short reads containing adapter sequences were removed using Trimmomatic 0.39 (Bolger et al. 2014). The Stacks 2.62 pipeline software (Catchen et al. 2013; Rochette et al. 2019) was used to obtain individual genotypes with the following parameters: minimum depth of coverage required to create a stack (m) = 3, the maximum distance between stacks (M) = 2, and maximum mismatches between loci when building the catalog (n) = 2. Three different filtering criteria were considered for quality control of the SNP data. First, any SNP site where one of two alleles had less than three counts was filtered out because it was difficult to distinguish polymorphisms from sequencing errors when the minor allele count of SNPs was too low (Roesti et al. 2012). Second, SNPs with high heterozygosity (Ho ≥ 0.6) were removed because excess heterozygosity may have resulted from artifactual loci built from several paralogous genomic regions. Third, a SNP was excluded if the number of samples shared by the SNP was below the reference value R, the minimum percentage of samples that retained an SNP. Although phylogenetic reconstructions using datasets with high R values tend to neglect the presence of SNPs unique to each lineage, phylogenetic reconstructions using datasets with low R values can contain noise with artifacts. Yahara et al. (2021) compared the performance of phylogenetic reconstructions for Hosta alata and its allies using datasets with R=0.1, 0.3, 0.5, and 0.8 and showed that the most reliable result was obtained when R=0.5. In this study, we used R=0.3 for phylogenetic reconstructions because the performance was higher at R=0.3 than at R=0.5. The major differences between the results at R=0.3 and R=0.5 are described in the Results section. The SNP detection using R=0.3 was repeated hierarchically: the first detection for all samples, the second detection for 159 samples of Clade 1 (see Fig. 1), and the third detection for 45 samples of Clade 4. Pairwise F_ST between taxa of Clades 1 and 4 was calculated using the SNPs with R=0.3 for all samples by the populations program of the Stacks pipeline. In population genetic analysis using STRUCTURE 2.3.4 (Pritchard et al. 2000), SNP detections were executed using R=0.5 to reduce SNPs with defects in some samples as much as possible. If a locus had two or more SNP sites, only one SNP site was used to avoid the inclusion of linked SNPs.

Figure 1.

An unrooted Maximum Likelihood tree of 28 Japanese species of Hosta, reconstructed using MIG-seq data. Bootstrap values (50% or larger) are indicated on internodes in or above the species level. Numbers in parentheses indicate the IDs of 6 clades.

The Maximum Likelihood phylogeny of the Japanese species of Hosta based on SNPs was inferred for all samples using RAxML 8.2.10 (Stamatakis 2014), as well as for two subsets of samples: Clade 1 and Clade 4. We used a GTRCAT model with an ascertainment bias correction using the Lewis method and performed 1,000 replicates of parallelized tree search bootstrapping. Phylogenetic networks using the Neighbor-Net method were performed by SplitsTree4 4.14.6 (Huson and Bryant 2006) using the uncorrelated P distance matrix calculated from the SNP matrix with R=0.3 for Clades 1 and 4. Hosta gracillima, which was used as an outgroup in the phylogenetic reconstruction, was not included in the phylogenetic networks of Clade 1. Population genetic structures of Clades 1 and 4 were examined using STRUCTURE 2.3.4 (Pritchard et al. 2000). Furthermore, hierarchical STRUCTURE analyses were performed for Clade 1 to detect more detailed population structure. We performed 30 independent runs with a burn-in of 100,000 steps and an additional 100,000 steps with the admixture model and estimated log-likelihoods for each number of clusters (K = 1–10). Optimal K values were determined using the ΔK method of Evanno et al. (2005) in the STRUCTURE Harvester (Earl and von Holdt 2012). Graphical results were obtained using CLUMPAK (Cluster Markov Packager Across K; Kopelman et al. 2015).

Morphological observations

Based on the results of phylogenetic analyses using MIG-seq, we reassessed the morphological traits utilized for classifying Hosta taxa in previous studies (Maekawa 1940, 1948, 1952; Fujita 1976; Fujita and Tamura 2008; Tamura and Fujita 2016) and identified morphological traits that can serve as diagnostic markers for newly discovered or revised taxa in this study. Using specimens subjected to phylogenetic analyses, we measured the following traits: the number of flowers per scape, length and width of the flower bract (a bract located in the axil from which a pedicel arises), pedicel length, perianth length, length of the narrowed part of the perianth tube, length and width of perianth lobes, stamen length, pistil length, anther-sac length, scape length, raceme length, length and width of leaf blades, length and width of petioles, and the number of lateral veins on leaves. We also measured the length of the peduncle bract (a bract located on a peduncle, not subtending a flower), if it is present. To complement these measurements, the same sets of traits were measured on confidently identified specimens deposited in KYO and MKB. All measurements were made to the nearest 1 mm, except for anther-sac length measured to the nearest 0.5 mm.

Data resources

All raw MIG-seq data were deposited in the DDBJ Sequence Read Archive (DRA) under accession numbers DRA011621, DRA013286, and DRA015301.

Results

Phylogenetic tree reconstructed using MIG-seq

A total of 96,556,838 raw reads (301,740 ± 10,362 reads per sample) were obtained from MIG-seq, and 79,861,049 reads (249,566 ± 8,620 reads per sample) were used for further analysis. After de novo SNP detection and filtering, 13,328 SNPs on 2,060 loci selected by R=0.3 were used for the phylogenetic reconstruction of all samples. Phylogenetic analyses of infrageneric groups and STRUCTURE analyses were performed using different SNP datasets selected by R=0.3 and R=0.5, respectively (see Suppl. material 2).

In the following descriptions of the results obtained from phylogenetic analyses, we employ new names as determined by the conclusions of this study for five new species (H. longipedicellata sp. nov., H. minazukiflora sp. nov., H. polyneuronoides sp. nov., H. samukazemontana sp. nov., and H. takiminazukiflora sp. nov.), one new subspecies (H. takiminazukiflora subsp. grandis subsp. nov.), and two new taxonomic status assignments (H. tardiva subsp. densinervia comb. & stat. nov. and H. scabrinervia stat. nov.) (See the Taxonomy section for authorities of these new names). Hosta tardiva subsp. densinervia and H. scabrinervia are based on H. kikutii var. densinervia and var. scabrinervia, respectively (Fujita and Tamura 2008; also see Tamura 2015, Tamura and Fujita 2016). On the other hand, concerning H. tosana previously treated as H. kikutii var. caput-avis by Fujita and Tamura (2008), and H. kikutii var. tosana by Tamura and Fujita (2016), we adhere to the earlier classification proposed by Maekawa (1948), who differentiated between H. tosana F. Maek. var. tosana and H. tosana var. caput-avis F. Maek. This is done to facilitate a comparison of the results of the present phylogenetic analysis with the earlier classification, concluding that these two varieties were indistinguishable. Furthermore, the results include two unknown species, designated as H. sp. 3 and H. sp. 4, in addition to two other unknown species, H. sp. 1 and H. sp. 2, reported by Yahara et al. (2021).

In the Maximum Likelihood tree reconstructed using the MIG-seq dataset at R=0.3 (Fig. 1), 28 Japanese Hosta species were clustered into the following six clades. The monophyly of Clade 1 was supported by a 100% bootstrap value and Clade 1 included H. tardiva subsp. tardiva, H. tardiva subsp. densinervia (previously classified as H. kikutii var. densinervia), H. scabrinervia (previously classified as H. kikutii var. scabrinervia), H. shikokiana, and six undescribed species (H. longipedicellata, H. takiminazukiflora, H. samukazemontana, H. minazukiflora, H. polyneuronoides, and H. sp. 3). The monophyly of Clade 2 was supported by a 100% bootstrap value with H. gracillima F. Maek. (=H. longipes var. gracillima). The monophyly of Clade 3 was supported by a 100% bootstrap value; Clade 3 comprised H. alata, H. pulchella N. Fujita, H. kikutii, H. longipes var. caduca, and two unknown species designated as H. sp. 1 and H. sp. 2 by Yahara et al. (2021). The monophyly of Clade 4 was supported by a 100% bootstrap value and Clade 4 comprised H. tosana. Monophyly of Clade 5 was supported by a 100% bootstrap value and Clade 5 comprised H. aequinoctiiantha Koidz. ex Araki (=H. longipes var. aequinoctiiantha (Koidz. ex Araki) Kitam.). Finally, monophyly of Clade 6 was supported by a 98% bootstrap value and Clade 6 comprised H. longipes (var. latifolia and var. longipes), H. sieboldiana, H. capitata, H. tibae F. Maek., H. tsushimensis N. Fujita, H. sieboldii, H. alismifolia, H. longissima, and an unknown species from Kii Peninsula, Honshu Island, designated as H. sp. 4, which had been identified as H. kikutii var. caput-avis (Maekawa 1948, 1952; Fujita 1976).

In the full data set tree (Fig. 1), a clade with a bootstrap support of 99%, comprising H. takiminazukiflora and H. longipedicellata was sister to the other species of Clade 1. To obtain a better resolution of phylogenetic relationships within Clade 1, we reconstructed the Maximum Likelihood tree for Clade 1 using H. gracillima (the closest related species) as an outgroup (Fig. 2). In this tree, the monophyly of conspecific samples was supported by a 100% bootstrap value for each of six species (H. tardiva, H. samukazemontana, H. minazukiflora, H. shikokiana, H. longipedicellata, and H. takiminazukiflora, arranged in order from the top of Fig. 2), and by a bootstrap value of 82% for H. scabrinervia. The clade including H. minazukiflora and H. shikokiana was supported by a 99% bootstrap value. Hosta shikokiana comprised two distinct clades, designated as H. shikokiana 1 and H. shikokiana 2, which were supported by bootstrap values of 100% and 99%, respectively. The samples of H. shikokiana 2 were placed between H. shikokiana 1 and H. minazukiflora in the SplitsTree (Fig. 3). Hosta sp. 3 from Honshu (FJI12268, Wakayama Prefecture in the Kii Peninsula) was sister to a clade with a bootstrap support of 99%, comprising H. shikokiana and H. minazukiflora. In the full data set tree (Fig. 1), H. minazukiflora was sister to a subgroup including H. sp. 3 and H. shikokiana, but the resolution was low; the bootstrap support for the latter subgroup was as low as 46%.

Figure 2.

A rooted Maximum Likelihood tree for Hosta Clade 1 reconstructed using MIG-seq data. Bootstrap values are indicated on internodes. Based on the Maximum Likelihood tree of 28 Japanese species (Fig. 1), H. gracillima is used as an outgroup.

Figure 3.

A SplitsTree for Hosta Clade 1 reconstructed using MIG-seq data.

The overall topology of the Maximum Likelihood tree reconstructed using the MIG-seq dataset at R=0.5 (not shown) was identical to Fig. 1, where the 28 Japanese Hosta species were clustered into six clades. The topology of Clade 1 in the Maximum Likelihood tree at R=0.5 was also identical to Fig. 1, except for H. scabrinervia. In this species, the samples did not form a cluster but instead scattered between a clade comprising H. longipedicellata and H. takiminazukiflora and another clade comprising H. shikokiana, H. minazukiflora, and H. sp. 3.

The monophyletic relationship between H. tardiva subsp. tardiva and subsp. densinervia was supported by a 92% bootstrap value (Fig. 2). The monophyly of H. tardiva subsp. tardiva was supported by a 100% bootstrap value, but H. tardiva subsp. densinervia was not monophyletic; a cluster including three samples from the type locality of H. tardiva subsp. densinervia (published as H. kikutii var. densinervia) in Tokushima Prefecture (JPN4071, 4072, 4075) and an additional sample from its vicinity (FJI17566) was sister to a clade including both the other samples of H. tardiva subsp. densinervia from Kochi Prefecture and H. tardiva subsp. tardiva, and the monophyly of the latter clade was supported by a 100% bootstrap value (Fig. 2). The closer relationship of H. tardiva subsp. tardiva with the ten samples of subsp. densinervia was also supported by SplitsTree (Fig. 3).

STRUCTURE analysis for Clade 1 indicated that ΔΚ was highest at K=3 (Fig. 4A). At K=3, H. tardiva subsp. tardiva exhibited a distinct genetic identity represented by light blue, while H. polyneuronoides was primarily associated with the second genetic identity depicted by dark purple. Hosta tardiva subsp. densinervia exhibited a mixture of the first and second identities. The remaining seven species showed higher probabilities of a genetic component derived from the third identity illustrated by orange (Fig. 4B). Considering that H. tardiva subsp. tardiva is a highly sterile taxon (Fujita 1976), hierarchical STRUCTURE analysis was performed for H. tardiva subsp. densinervia and H. polyneuronoides, excluding H. tardiva subsp. tardiva. This analysis revealed that ΔΚ was highest at K=2 (Suppl. material 3). At K=2, H. tardiva subsp. densinervia was predominantly characterized by the first identity depicted by orange, while H. polyneuronoides was dominated by the second identity represented by light blue (Fig. 4C). Four samples of H. tardiva subsp. densinervia collected from Miyoshi city (JPN4071, 4072, 4075, FJI17566) exhibited admixture with H. polyneuronoides. Hierarchical STRUCTURE analysis for the other seven species showed that ΔΚ was highest at K=3 and second highest at K=5 (Suppl. material 3). At K=3, H. shikokiana was primarily associated with the first identity depicted by dark purple, H. takiminazukiflora was dominated by the second identity represented by orange, and H. samukazemontana, H. scabrinervia, and H. longipedicellata were primarily associated with the third identity represented by light blue (Fig. 4D). At K=5, H. samukazemontana, H. scabrinervia, and H. longipedicellata were predominantly characterized by different identities represented by red-purple, light blue, and green, respectively. Four samples of H. takiminazukiflora subsp. takiminazukiflora collected at Mt. Inamura (JPN3930, 6513–6515) exhibited a mixture of two identities, one identity predominantly found in H. takiminazukiflora (orange) and another identity dominant in H. samukazemontana (dark purple). Hosta minazukiflora exhibited a mixture of two identities, predominantly the ones found in H. shikokiana (dark purple) and H. scabrinervia (light blue). Hosta sp. 3 showed a mixture of three identities at both K=3 and K=5.

Figure 4.

Population genetic structure of Hosta Clade 1 and Clade 4 A, E changes of ΔΚ with K in Clades 1 and 4, respectively B a diagram showing the result of STRUCTURE analysis for Clade 1 at K=3 C a diagram showing the result of hierarchical STRUCTURE analysis for H. polyneuronoides and H. tardiva subsp. densinervia D diagrams showing the results of hierarchical STRUCTURE analysis for H. shikokiana and six related species at K=3 and 5 F diagrams showing the results of STRUCTURE analysis for Clade 4 at K=2, 3, and 8.

Clade 4 consisting of H. tosana was more fully examined (Fig. 5). Seven samples collected from the Kagami River basin formed a distinct clade with a bootstrap support of 100% in the Maximum Likelihood tree. This clade, designated as the Kagami River lineage, formed the basal-most position of the 38 samples in Clade 4 (Fig. 1) and also formed a distinct cluster in the SplitsTree analysis (Fig. 6). The other samples of H. tosana var. tosana and var. caput-avis formed a cluster, but neither var. tosana nor var. caput-avis was monophyletic (Fig. 5). A clade including four samples collected from the type locality of H. tosana var. caput-avis (JPN12536–12539) and another sample from its vicinity (JPN12653) was sister to a clade including 12 samples of H. tosana var. caput-avis from Kami City, ca. 24 km NW of the type locality, and 21 samples of H. tosana var. tosana.

Figure 5.

A rooted Maximum Likelihood tree for Hosta Clade 4 reconstructed using MIG-seq data. Bootstrap values are indicated on internodes. Based on the Maximum Likelihood tree of 28 Japanese species (Fig. 1), the tree is rooted by the Kagami River lineage of H. tosana.

Figure 6.

A SplitsTree for Hosta Clade 4 reconstructed using MIG-seq data.

STRUCTURE analysis for Clade 4 showed that ΔΚ was highest at K=3 and second highest at K=8 (Fig. 4E). At K=3, the Kagami River lineage of H. tosana was dominated by the first identity depicted by dark purple. For H. tosana var. caput-avis, out of the 12 samples collected from Kami City (depicted in the upper section), the second identity, represented by the orange color, was predominant. However, among the five samples taken from the vicinity of the type locality (depicted in the lower section), a mixture of three identities was observed, indicated by dark purple, orange, and light blue colors. For H. tosana var. tosana, the third identity (light blue) was predominant. However, three samples (depicted in the uppermost section; from Befu Valley, Kami City, JPN6553–55) were mixed with the second (orange) and third (light blue) identities, and two samples from Kami City (other than Befu Valley, JPN3934 and 6603) showed a mixture of three identities. At K=8, the Kagami River lineage of H. tosana was dominated by the first identity (dark purple), but H. tosana var. caput-avis and var. tosana showed a complicated mixture of five or six identities.

Morphological and phenological divergence between taxa with adjacent distribution ranges

(1) Hosta tardiva subsp. densinervia and H. polyneuronoides

We observed five populations of H. tardiva subsp. densinervia at elevations from 90 m to 900 m (Fig. 7B). Although the populations are geographically isolated in three river basins, the plants in these populations are morphologically indistinguishable. We collected H. polyneuronoides at elevations from 237 m to 1980 m (Fig. 7C). These populations are located in the upper reaches and headwaters of the Yoshino River and on the ridgeline from Mt. Ishizuchi to Mt. Komochi-gongen and its vicinity where H. tardiva subsp. densinervia is not distributed (see a dotted quadrilateral in Fig. 7B). The specimens of H. polyneuronoides collected from Mt. Ishizuchi were morphologically identical to another specimen collected from the same location (Takahashi & Fujita 226; see the Taxonomy section for details of this specimen record). This specimen was identified as H. shikokiana in its original description by Fujita (1976).

Figure 7.

Maps of collection locations for Hosta species A Shikoku Island (black) in relation to Japan B detail of Shikoku Island with population locations for some species and varieties C details of collection locations for H. polyneuronoides (solid circles) and H. scabrinervia (an open square), and approximate ranges (circles) of H. minazukiflora (M), H. longipedicellata (L), H. samukazemontana (Sa), H. shikokiana (Sh), and H. takiminazukiflora (T). Considering conservation concerns, we refrained from disclosing the precise collection locations of these five species. The contour map was reproduced from a webpage. (https://maps.gsi.go.jp/#11/33.726624/133.351364/&base=english&ls=english&disp=1&vs=c1g1j0h0k0l0u0t0z0r0s0m0f1&d=m) in accordance with the usage regulations of the Geospatial Information Authority of Japan.

Hosta polyneuronoides is morphologically similar to H. tardiva subsp. densinervia, but is distinguished by smaller inflorescence (3–10 flowers per inflorescence vs. more than 20 per inflorescence in H. tardiva subsp. densinervia) and shorter anther sacs (3.5–4 mm long vs. 5 mm long). Hosta polyneuronoides tends to flower earlier than H. tardiva subsp. densinervia: flowering specimens of H. tardiva subsp. densinervia were collected between August 13 and September 27, whereas flowering specimens of H. polyneuronoides were collected between July 25 and August 20 (see specimen records cited in the Taxonomy section). While H. tardiva subsp. densinervia was observed to grow on soil or rocks in open habitats along riverbanks, H. polyneuronoides was found to grow on wet rocks in shaded habitats along streams.

(2) Hosta tardiva subsp. densinervia, H. polyneuronoides, and H. scabrinervia

The type locality of H. scabrinervia (Fig. 7, open rectangular) is located 15.5 km south of the type locality of H. tardiva subsp. densinervia, both of which are located along the Yoshino River, at elevations of 90 m (H. tardiva subsp. densinervia) and 180 m (H. scabrinervia). Fujita and Tamura (2008) distinguished H. kikutii var. densinervia (i.e., H. densinervia subsp. densinervia) from H. kikutii var. scabrinervia (i.e., H. scabrinervia) by the slightly papillose abaxial surface of the leaf nerves (vs. prominently papillose in var. scabrinervia). However, in the type locality population of H. scabrinervia, Hosta plants are highly variable in the state of abaxial nerve surface; some plants have leaf nerves prominently papillose adaxially (e.g., JPN4080, 4084–85), but other plants have leaf nerves almost smooth adaxially (e.g., JPN4082–83, 4086) or intermediately papillose adaxially (e.g., JPN 4087). All seven samples clustered into a single clade in both the Maximum Likelihood tree and SplitsTree (Figs 2, 3). While H. tardiva subsp. densinervia is closely related to H. tardiva subsp. tardiva, H. scabrinervia is positioned between H. samukazemontana and a clade comprising H. minazukiflora, H. shikokiana, and H. sp. 3 (Fig. 2). The flowering specimens of H. scabrinervia and H. tardiva subsp. densinervia were collected from July 13 to July 25 and from August 13 to September 27, respectively. Morphologically, H. tardiva subsp. densinervia and H. scabrinervia are very similar, but distinguished by their anther sac length (5 mm in the former vs. 3 mm in the latter). Additionally, H. tardiva subsp. densinervia has more than 20 flowers per scape (except for small plants growing on rocks), while H. scabrinervia has 15–20 flowers per scape.

Two specimens (MBK0319724 and MBK0179549) collected at an elevation of 200 m along the middle reach of the Yoshino River, at 7.5–9 km south of the type locality of H. scabrinervia, respectively, were morphologically identical to H. scabrinervia in the type locality (Fig. 7B). Five other specimens of H. scabrinervia (JPN12523–12525, 15231, 15232) were collected at an elevation of 550 m, 33 km WSW of the type locality of H. scabrinervia, within the range of H. polyneuronoides (Fig. 7B, C). At 5–6 km W of the locality of the five specimens, the specimens of H. polyneuronoides (JPN12529, 12530, 12532, and 12535) were collected. In this area, H. scabrinervia was found to grow on wet cliffs, whereas H. polyneuronoides was found on rocks along the stream.

(3) Hosta polyneuronoides and H. shikokiana

In addition to H. scabrinervia, five other species (H. minazukiflora, H. longipedicellata, H. shikokiana, H. samukazemontana, and H. takiminazukiflora) are densely distributed within or near the range of Hosta polyneuronoides (Fig. 7C). Among them, H. shikokiana, endemic to the serpentine area of Mt. Higashi-akaishi (Sh in Fig. 7) and its surroundings, is most distinct in having leaf blades strongly undulating along the margin, lustrous below when dried, each leaf blade decurrent to a winged petiole, purple perianths with three dark purple veins inside each perianth lobe, and usually purplish scapes and flower buds (Fig. 11). At Mt. Higashi-akaishi, we collected H. shikokiana in open habitats at an elevation of 1660 m along the rocky ridgeline (JPN6197, 6228, 6239), and H. polyneuronoides in shaded habitats on wet rocks along streams at elevations of 970 m (JPN 6269–71) and 1040 m (JPN6153–56).

(4) Hosta minazukiflora and H. shikokiana

The type locality of H. minazukiflora (M in Fig. 7) is located 20 km SE of the habitat of H. shikokiana at the ridgeline of Mt. Higashi-akaishi. Hosta minazukiflora was observed to have flower buds on May 23 and flowers on June 13 and 25. In contrast, H. shikokiana was observed to have flower buds on June 9, and flowers from July 14 to 21 (see specimens cited in the Taxonomy section). Morphologically, H. shikokiana has leaf blades strongly undulate along margin, usually purplish scapes, and purple perianths with dark purple inside veins (Fig. 11), whereas H. minazukiflora has leaf blades plain or only weakly undulate along margin, usually green scapes, and lavender perianths with darker colored inside veins (Figs 12, 13). Phylogenetic analyses revealed that H. shikokiana consisted of two distinct clades: H. shikokiana 1, collected from the mountain range of Mt. Higashi-akaishi, and H. shikokiana 2, collected from the other mountains located northeast of the type locality of H. minazukiflora (JPN3937, 9952). Both H. shikokiana 1 and H. shikokiana 2 exhibited the diagnostic traits mentioned above. While the SplitsTree analysis (Fig. 3) indicated that H. shikokiana 2 was positioned between H. minazukiflora and H. shikokiana 1, the Maximum Likelihood tree (Fig. 2) supported a sister relationship between H. shikokiana 1 and H. shikokiana 2 with a bootstrap value of 100%.

(5) Hosta polyneuronoides and H. minazukiflora

A population of H. polyneuronoides is located at an elevation of 480 m at 7.5 km north of the type locality of H. minazukiflora (at an elevation of 280 m), both growing on rocks along the stream of a tributary of the Yoshino River (see M in Fig. 7C). This is one of the two neighboring populations of “H. kikutii var. densinervia’ along the Asemi River mentioned in the Introduction, and the other is H. minazukiflora. While H. polyneuronoides grows on wet rock, H. minazukiflora grows on a vertical cliff. While flowering specimens of H. polyneuronoides were collected from July 28 to August 20 (see specimens cited in the Taxonomy section), flowering specimens of H. minazukiflora along the Asemi River were collected on June 23 and 25. On August 19, 2020, we collected H. polyneuronoides specimens with flower buds on scapes (JPN1327–29, Fig. 9A) and a specimen of H. minazukiflora with fruits (JPN1310, Fig. 13). In late August 2020, we also observed and collected mature plants of H. polyneuronoides that had just begun to bloom in three other localities: Niyodogawa-cho, Ino-cho, and Ochi-cho in Kochi City (Fig. 7). We compared morphological traits between these specimens and the type specimens of H. minazukiflora deposited in MBK (Table 1). Hosta minazukiflora had shorter leaf blade length, narrower leaf blade width, and shorter petiole lengths than H. polyneuronoides. For the number of flowers and length of flower bracts, H. minazukiflora had distinctly smaller values than H. polyneuronoides. The two species were also distinguished by their anther sac length (3.5–4 mm in H. polyneuronoides and 3 mm H. minazukiflora).

Table 1.

Download as

CSV

XLSX

Measurements of 14 morphological traits in nine taxa of Hosta in Shikoku. Aspect ratio is defined as leaf length divided by leaf width.

	H. t. densinervia	H. polyneuronoides	H. scabrinervia	H. minazukiflora	H. shikokiana	H. t. takiminazukiflora	H. t. grandis	H. longipedicellata	H. samukazemontana
Leaf blade length	12–23 cm	7.5–28.5	16.6–32	11.8–17	8.1–9.3	11–26.5	22–25	13.4–30.2	7–22
Leaf blade width	3.2–11.5 cm	3.3–15.3	7.5–14.8	2.9–6.2	2.9–5.0	3.6–10.5	13–14	5.4–11.8	3.6–10.7
Aspect ratio	2–3.6	0.9–2.9	1.6–3	2.4–4.3	1.9–2.9	2.1–3.3	1.7–1.8	2.2–3	1.8–2.9
Lateral veins	6–11 pairs	8–9	7–15	6–9	5–7	6–10	12–13	7–10	6–12
Petiole length	9–36 cm	5–40	5.5–47	8–19	6.5–8.5	3.4–22	35–38	15–34.3	3.2–14.5
Scape length	27–44 cm	15–51	28.5–45	18–37	20.5–29	8.6–28.5	31–36	26.5–32.5	20–43
Raceme length	6–9 cm	4.3–9	5–12	10–15	5.3–13	5.9–18.5	12–15.5	7.6–9	7–9
Flower number	15–25	3–10	15–17	3–6	2–10	4–18	10–12	7–15	5–18
Floral bract length	2–3 cm	1.2–3.1	2–4.6	1.3–1.6	1.3–2.2	2–4.7	2.8–4.3	2–2.5	1.7–2.1
Floral bract width	0.4 cm	0.1–0.8	0.4–1.3	0.2–0.4	0.1–0.3	0.2–0.7	0.4–0.5	0.3–0.4	0.3–0.5
Pedicel length	0.8–1.8 cm	0.6–1.3	1.5–3.4	1.1–1.2	0.6–1.5	1.4–2.5	1.9–2.0	2.5–3.3	1–1.8
Perianth length	3.9–5.3 cm	4.5	5–6.8	4.1–4.7	4.4–5.0	4.0–5.7	3.8–4.9	4.4–6.4	4.5–5
Perianth lobe length	1.1–2.0 cm	0.7–1.2	0.8–1.6	1.2–1.4	1.4–1.7	1.2–1.8	1.2–1.3	1–1.5	1–1.3
Perianth lobe width	0.4–0.8 cm	0.4–1	0.5–1	0.6–0.8	0.8–1.0	0.5–1.0	0.7–0.8	0.6–0.9	0.7–0.8
Anther-sac length	5 mm	3.5–4	3	3	2.5–3	2–3	3	3	3

(6) Hosta takiminazukiflora and H. minazukiflora

The type locality of Hosta takiminazukiflora was only 14 km west of that of H. minazukiflora (T and M in Fig. 7C, respectively). Hosta takiminazukiflora was observed to grow on wet cliff faces. Flowering specimens of H. takiminazukiflora were collected from June 24 to August 8. Morphologically, H. minazukiflora (Figs 12, 13) has petioles and scapes not or only sparsely dotted and lavender perianths with three darker colored inside veins per lobe, whereas H. takiminazukiflora (Figs 14, 15) has petioles and scapes densely dotted with purple and whitish perianths with a darker colored midvein. The two species are similar in appearance, but are not closely related: H. takiminazukiflora is sister to H. longipedicellata and H. minazukiflora is sister to H. shikokiana (Figs 1, 2).

(7) Hosta takiminazukiflora subsp. takiminazukiflora and subsp. grandis

A small population of H. takiminazukiflora subsp. grandis (JPN6487–89, JPN12518–12521) was found alongside a larger population of H. takiminazukiflora subsp. takiminazukiflora (JPN6483–86 and JPN6490), with the former growing upright from the soil between the rocks under the waterfall and the latter deflected from the wet cliff of the waterfall. Both subspecies were found in flower on June 24, 2021.

Hosta takiminazukiflora subsp. grandis (Fig. 16) is easily distinguished from subsp. takiminazukiflora in having erect scapes 31–36 cm in height (vs. deflected scapes less than 30 cm long in subsp. takiminazukiflora), wider leaves (13–14 cm wide vs. 3.6–10.5 cm wide), cordate at the base (vs. cuneate), glaucous abaxially (vs. not glaucous but light green), lateral veins 12–13 pairs (vs. usually 6–10 pairs), and petioles green (vs. dotted with purple). These differences are maintained under fertilization and cultivation conditions in the Kochi Prefectural Makino Botanical Garden. Despite its morphological distinctiveness, subsp. grandis is phylogenetically included within a clade of H. takiminazukiflora (Figs 2, 3) because it is sister to the sympatric population of subsp. takiminazukiflora and is more remotely related to the other populations of subsp. takiminazukiflora. In the SplitsTree analysis of Clade 1 (Fig. 3), four samples of H. takiminazukiflora subsp. grandis formed a cluster, while the other three samples clustered with five samples of H. takiminazukiflora subsp. takiminazukiflora collected from the same locality. However, in the further SplitsTree analysis of a clade comprising H. takiminazukiflora and H. longipedicellata, seven samples of H. takiminazukiflora subsp. grandis formed a separate cluster from the cluster containing the five samples of H. takiminazukiflora subsp. H. takiminazukiflora from the same locality (not shown). Furthermore, in the Maximum Likelihood tree, all seven samples of H. takiminazukiflora subsp. grandis formed a clade with a bootstrap support of 99%, which is significantly differentiated from the sympatric population of subsp. takiminazukiflora (JPN6483–86, 6490) (Fig. 2). Hosta takiminazukiflora subsp. takiminazukiflora, collected from three other localities (Mt. Inamura: JPN3930, 6513–15; along Road 6: JPN6478–81; Mt. Higashikado: JPN3928, 3938), clustered into three separate subclades. However, these subclades were morphologically indistinguishable from the lineage of Hosta takiminazukiflora subsp. takiminazukiflora that is sympatric with H. takiminazukiflora subsp. grandis.

(8) Hosta longipedicellata and H. takiminazukiflora

The type locality of H. longipedicellata is located at 22.5 km west of that of H. takiminazukiflora and the two species are distributed in two mountain regions separated by an upper reach of the Yoshino River (L and T in Fig. 7C, respectively). The flowering type specimens of H. longipedicellata and H. takiminazukiflora were collected on August 1, 2006 and July 25, 1983, respectively. Both H. takiminazukiflora subsp. takiminazukiflora and H. longipedicellata grows on wet cliffs. Morphologically, H. longipedicellata (Figs 17, 18) is similar to H. takiminazukiflora subsp. takiminazukiflora, but it can be distinguished by longer pedicels (2.5–3.3 cm long in H. longipedicellata compared to 1.4–2.5 cm long in H. takiminazukiflora), shorter flower bracts (1.5–2.5 cm long compared to (2.2–)2.5–3.7 cm long) that wither during flowering (as opposed to being fresh), and leaf veins papillose on the abaxial surface (as opposed to being smooth). Phylogenetically, the sister relationship between H. longipedicellata and H. takiminazukiflora was supported by a bootstrap value of 99% (Fig. 1) or 100% (Fig. 2). However, in SplitsTree (Fig. 3), H. longipedicellata was distinct from H. scabrinervia and H. takiminazukiflora.

(9) Hosta samukazemontana, H. shikokiana, and H. polyneuronoides

The type locality of H. samukazemontana (Sa in Fig. 7C) is located 13 km SW of the type locality of H. shikokiana (Sh in Fig. 7C) to which H. samukazemontana was included by Fujita (1976). Hosta samukazemontana was observed growing on wet cliffs of Mt. Kanpu (called Mt. Samukaze in the old days), while H. shikokiana was found on rocky slopes along the ridgeline of Mt. Higashi-akaishi and its vicinity. Morphologically, H. samukazemontana is similar to H. shikokiana with leaves shorter than 20 cm and scapes shorter than 40 cm, but distinguished by deflected scapes upwardly curved at the tip when flowering and fruiting (in contrast to being straight when flowering and fruiting in H. shikokiana), green scape color (as opposed to the usual purplish color, dotted with purple in the lower part), whitish perianths (vs. purple), narrowly winged petioles (instead of widely winged, particularly in the upper part), and a weakly undulate leaf margin (in contrast to a strongly undulate margin).

Along the Kamegamori forest road, we collected both H. samukazemontana (JPN6325, with flower buds on June 22, 2021) and H. polyneuronoides (JPN6342 and 6361, without flower buds on June 22, 2021). The two populations were approximately 500 m apart along the road. Flowering specimens of H. samukazemontana were collected from Mt. Kanpu on June 19, July 24 and July 25 (see specimens cited in the Taxonomy section) and flowering specimens of H. polyneuronoides from Mt. Ishizuchi, located in the vicinity of Mt. Kanpu and the Kamegamori forest road, were collected on July 28 and August 7. Morphologically, H. samukazemontana is distinguished from H. polyneuronoides by deflected scapes upwardly curved at the tip (vs. straight) and anther sacs 3 mm long (vs. 3.5–4 mm long).

Morphological and phenological observations on H. tosana

Two populations of the Kagami River lineage of H. tosana were observed on cliffs along the Kagami River in Kochi City. This location is 50 km SSW of the population of H. tosana var. tosana along the Monobe River. The specimen of the Kagami River lineage with flowers (Fig. 21A) was collected on July 14, 2013 at 130 m elevation. Near this locality, three additional flowering specimens were collected in mid-July (MBK0104375, MBK0247212, MBK0247214), two specimens with young fruits were collected in late July (MBK0247386, MBK0247387), and a fruiting specimen was collected in early October (MBK0208327). These populations were found below 300 m elevation, whereas the localities of typical H. tosana var. tosana were above 1000 m elevation. Morphologically, the Kagami River lineage of H. tosana is distinguished from typical H. tosana var. tosana (Fig. 22) by upwardly curving scapes at the apex when flowering and fruiting (vs. curved downward at apex) and leaves with fewer lateral veins (5–10(–11) compared to 10–13) running at wider intervals (usually 1 cm in contrast to usually 0.7 cm).

In the Befu Valley of Kami City, we collected H. tosana var. caput-avis from two populations: the upstream population (JPN6524–6526; Fig. 23A) and the downstream population (JPN6571, 6572), which were 1 km apart along the upper reach of the Monobe River. We also collected H. tosana var. tosana (JPN6553–55; Fig. 23B) within a distance of 100 m from the downstream population. These samples of two varieties in sympatry are depicted by red circles in the Splits Tree (Fig. 6). On June 25, 2021, plants of var. caput-avis were flowering in the downstream population and just before flowering (in flower buds) in the upstream population, whereas var. tosana had young scapes before flowerings. The two varieties were distinguished by scapes (strongly bent at the base and strongly curved at the apex in var. caput-avis vs. usually gently bent like a bow in var. tosana; Figs 23, 24) and peduncle bracts enclosing young flower buds (upward curved in var. caput-avis vs. straight in var. tosana; see enlarged photographs in Fig. 23). Additional specimens of var. caput-avis (JPN6586, 6587, 6793–6796) were collected from the higher elevations of Mt. Ishidate, located east of the above locality, whereas additional specimens of var. tosana (JPN6603, 6611–6616, 6619, and 6620) were collected at lower elevations located west of the above locality. Some plants of var. tosana (JPN6614, 6616) were flowering at a lower elevation. In flowering specimens, var. caput-avis (JPN6571–6572) and var. tosana (JPN6614, 6616) were distinguished by the length of the flower bracts (3–4.7 cm in var. caput-avis vs. 1.5–2 cm in var. tosana), the length of flowers (5–6.5 cm vs. 4.3–5 cm), and the length of pedicels (1.5–2.3 cm vs. 0.9–1.3 cm).

In Yanase, Aki-gun, the type locality of H. tosana var. caput-avis, we only found H. tosana var. caput-avis. Four samples from Yanase and an additional sample from its vicinity formed a clade with a bootstrap support of 100%. This clade was sister to another clade with a bootstrap support of 100%, comprising H. tosana var. caput-avis from Kami City and var. tosana from Kami City and three other localities including its type locality at Kajigamori (Fig. 5). The samples from Yanase and its vicinity also formed a distinct cluster in the SplitsTree analysis (Fig. 6).

Genetic divergence between taxa

Genetic divergence between taxa, measured by F_ST (Table 2), varied from 0.07 to 0.52. Specifically, F_ST values between infraspecific taxa were observed as 0.10 and 0.14 between H. tosana var. tosana and var. caput-avis, and H. takiminazukiflora subsp. takiminazukiflora and subsp. grandis, respectively. Additionally, F_ST values were found to be 0.15 and 0.16 between the Kagami River lineage and either H. tosana var. tosana or var. caput-avis, respectively. Between species within Clade 1, F_ST values exhibited variability, ranging from 0.07 (between H. polyneuronoides and H. tardiva subsp. densiflora) to 0.33 (between H. minazukiflora and either H. tardiva subsp. tardiva or H. samukazemontana), and 0.38 (between H. takiminazukiflora subsp. grandis and either H. tardiva subsp. tardiva, H. shikokiana 2, or H. minazukiflora). Notably, F_ST values between H. takiminazukiflora subsp. grandis and other species within Clade 1 (ranging from 0.18 to 0.38) tended to be larger than values observed between other species (ranging from 0.10 to 0.33).

Table 2.

Download as

CSV

XLSX

Genetic divergence between species or subspecies measured by F_ST. The abbreviations in the first line represent taxon names from H. t. densiflora (H. tardiva subsp. densiflora) to H. tosana var. tosana, and the last capu represents H. tosana var. caput-avis.

	dens	poly	shi1	shi2	mina	grand	taki	samu	scab	long	Kaga	tosa	capu
H. t. tardiva	0.16	0.15	0.29	0.32	0.33	0.38	0.27	0.30	0.23	0.31	0.44	0.40	0.40
H. t. densinervia		0.07	0.18	0.18	0.18	0.25	0.16	0.16	0.12	0.19	0.37	0.34	0.33
H. polyneuronoides			0.14	0.12	0.12	0.18	0.14	0.10	0.09	0.15	0.30	0.29	0.28
H. shikokiana 1				0.17	0.22	0.30	0.19	0.24	0.18	0.24	0.41	0.36	0.36
H. shikokiana 2					0.27	0.38	0.17	0.31	0.17	0.24	0.50	0.40	0.41
H. minazukiflora						0.38	0.18	0.33	0.16	0.25	0.52	0.43	0.42
H. t. grandis							0.14	0.37	0.22	0.28	0.51	0.42	0.43
H. t. takiminazukiflora								0.18	0.14	0.17	0.37	0.33	0.32
H. samukazemontana									0.17	0.25	0.49	0.41	0.41
H. scabrinervia										0.17	0.38	0.34	0.33
H. longipedicellata											0.42	0.37	0.36
H. tosana Kagami River lineage												0.15	0.16
H. tosana var. tosana													0.10

Discussion

Our phylogenetic analysis of Japanese Hosta species showed that H. tardiva subsp. densinervia and H. scabrinervia, previously classified as H. kikutii var. densinervia and var. scabrinervia by Tamura and Fujita (2016), belong to Clade 1 (Fig. 1). In contrast, H. kikutii, previously classified as H. kikutii var. kikutii belongs to Clade 3, while H. tosana, previously classified as H. kikutii var. tosana, belongs to Clade 4. Therefore, “H. kikutii” in the sense of Tamura and Fujita (2016) are polymorphic.

Among taxa previously treated as varieties of H. kikutii, H. tardiva subsp. densinervia formed a clade with H. tardiva subsp. tardiva, supported by a bootstrap value as high as 99% (Fig. 1) or 92% (Fig. 2). This result was unexpected because they are placed in different sections due to their significant morphological differences. According to Fujita’s key (1976), Japanese Hosta species can be classified into two main groups based on the presence or absence of darker colored veins inside the perianths. Hosta tardiva subsp. tardiva belongs to Sect. Tardanthae of the former group, while H. tardiva subsp. densinervia belongs to Sect. Helipteroides of the latter group. Hosta tardiva subsp. tardiva has attractive purple perianths with darker colored inside veins and is widely cultivated in gardens (Fujita 1976). This subspecies shows high sterility in flowers (Fujita 1976), and is suggested to be of hybrid origin (Tamura and Fujita 2016). It has a diploid chromosome number of 2n=60, but has two pairs of homologous chromosomes that are unequal in size and morphology which is believed to be the cause of the species high sterility (Kaneko 1970).

Considering significant morphological differences, it is puzzling that the cluster of H. tardiva subsp. tardiva is a single offshoot of H. tardiva. Consequently, if H. tardiva subsp. tardiva is separated, H. tardiva subsp. densinervia cannot be considered monophyletic. The result of the STRUCTURE analysis provides a clue to explain this puzzling result. Ten samples of H. tardiva subsp. densinervia collected from the lower reach of the Niyodo River exhibit a mixture of two genetic identities: one identity is dominant in H. tardiva subsp. tardiva, and another identity is dominant in H. polyneuronoides, as well as in other samples of H. tardiva subsp. densinervia, including three samples from the type locality population in Tokushima Prefecture. In the Maximum Likelihood tree (Fig. 2) and the SplitsTree (Fig. 3), these ten samples are positioned between H. tardiva subsp. tardiva and five other samples of H. tardiva subsp. densinervia. Although the flowers of H. tardiva subsp. tardiva are highly sterile (Fujita 1976), Fujii (2018) reported that H. tardiva subsp. tardiva can cross with other species, producing fertile pollen grains. This finding suggests that the ten samples of H. tardiva subsp. densinervia from the Niyodo River originated through hybridization events between H. tardiva subsp. tardiva and ancestral lineages of H. tardiva subsp. densinervia. Despite this hybrid origin, these ten samples are morphologically identical to other samples of H. tardiva subsp. densinervia and lack morphological traits indicative of their hybridization with H. tardiva subsp. tardiva. Therefore, we propose that the Niyodo River population of H. tardiva subsp. densinervia has become stabilized through repeated backcrossing with ancestral lineages of H. tardiva subsp. densinervia.

Hosta tardiva is likely an instance of anacladogenetic speciation, wherein a new species originates through budding, initially rendering the ancestral taxon paraphyletic. In such instances, it becomes necessary to acknowledge a paraphyletic subspecies when both derived and ancestral lineages display distinct diagnostic traits that set them apart (Carnicero et al. 2019). Theoretically, these subspecies are referred to as diachronic subspecies, constituting segments of species-level clades that are differentiated from other parts of the clade by evolutionarily significant characteristics (Reydon and Kunz 2021). Our classification of H. tardiva subsp. tardiva and H. tardiva subsp. densinervia is grounded in this subspecies concept.

Hosta tardiva subsp. densinervia and H. scabrinervia were initially described as H. kikutii var. densinervia and var. scabrinervia by Fujita and Tamura (2008). Morphologically, H. scabrinervia is similar to H. tardiva subsp. densinervia but distinguished by anther length (3 mm long compared to 5 mm long). These taxa are also distinct in their phylogenetic positions within Clade 1 and their flowering seasons. Flowering specimens of H. tardiva subsp. densinervia were collected between August 13 and September 27. On the other hand, the flowering specimens of H. scabrinervia were collected from July 13 to July 25. Therefore, H. tardiva subsp. densinervia and H. scabrinervia are considered to be reproductively isolated, and we propose treating H. scabrinervia as a separate species from H. tardiva.

In the STRUCTURE analysis, H. scabrinervia exhibited a mixture of two genetic identities, one identity dominated in H. tardiva subsp. densinervia and another identity shared by seven species. This result seems to suggest a hybrid origin of H. scabrinervia. Among the seven species, H. takiminazukiflora subsp. takiminazukiflora is most similar to H. scabrinervia; however, it can be distinguished by several characteristics. It has purplish green or purple flower bracts instead of white or purplish white in H. scabrinervia, and its petioles are shorter (3–22 cm long as opposed to (20–)22–40 cm long). Hosta polyneuronoides is also similar to H. scabrinervia but can be distinguished by longer anther-sacs (3.5–4 mm long vs. 3 mm in H. scabrinervia). If H. scabrinervia originated from a hybridization, H. polyneuronoides and H. takiminazukiflora subsp. takiminazukiflora are most likely candidates of parental species. While available specimen records showed that H. polyneuronoides flowers earlier (July 25 to August 20) than H. scabrinervia (July 13 to July 25), these flowering records are close and two taxa may be able to hybridize in late July. However, at K=5 in the hierarchical STRUCTURE analysis, H. scabrinervia exhibited a unique genetic identity depicted by light blue, and did not exhibit another identity depicted by orange, which is dominant in H. takiminazukiflora subsp. takiminazukiflora. This result suggests that genetic variation accumulated after the origin of H. scabrinervia, even if it originated through a hybridization event between H. polyneuronoides and H. takiminazukiflora subsp. takiminazukiflora. In the SplitsTree, H. scabrinervia was separated from H. takiminazukiflora, which is closely related to H. longipedicellata, but connected with H. polyneuronoides. This relationship as well as the result of STRUCTURE analysis at K=3 suggest that H. scabrinervia is a species separated from H. takiminazukiflora and H. polyneuronoides, but could have a history of introgression with H. polyneuronoides.

Hosta polyneuronoides is closely related to H. tardiva, but its monophyly was supported by a bootstrap value of 82% (Fig. 2). It is morphologically similar to H. tardiva subsp. densiflora, having whitish perianths without distinct inside veins. However, H. polyneuronoides is distinguished by having 10 or fewer flowers per scape (compared to 20 or more in H. tardiva subsp. densiflora), and anther lengths of 3.5–4 mm (as opposed to 5 mm). Two taxa grow in different habitats: H. tardiva subsp. densinervia thrives in more open habitats on soil or crevices of dry rocks along riverbanks at elevations ranging from 90 m to 900 m, while H. polyneuronoides prefers shaded habitats on wet rocks along streams at elevations ranging from 237 m to 1980 m. While H. tardiva subsp. densinervia is widely distributed in the lower reaches of the Yoshino River, Niyodo River, and Shimanto River, H. polyneuronoides is more restricted to a narrow area at the headwaters of the Yoshino River (Fig. 7). These two taxa may have differentiated due to geographical isolation and disruptive selection in distinct habitats. Despite an F_ST value of 0.07 between H. tardiva subsp. densinervia and H. polyneuronoides, which is lower when compared to the F_ST values between other species of Clade 1, the two taxa exhibited distinct identities in the hierarchical STRUCTURE analysis (Fig. 4C). Considering these findings, we propose distinguishing H. polyneuronoides as a separate species from H. tardiva.

Our phylogenetic analysis also indicated that another morphologically distinct species, H. shikokiana, belonged to Clade 1. Due to its morphological distinctiveness, Fujita (1976) placed H. shikokiana in Section Eubryocles F. Maek. together with a Chinese species H. ventricosa Stearn. However, H. shikokiana is closely related to a newly discovered species, H. minazukiflora. In addition to molecular phylogenetic analyses, the STRUCTURE analysis revealed a unique genetic identity shared between H. shikokiana and H. minazukiflora (Fig. 4). Thus, there is some justification for considering these two as intraspecific taxa within a single species. However, it is challenging to identify a diagnostic trait that categorizes H. shikokiana and H. minazukiflora as a single species. Considering the morphological differences between H. minazukiflora and H. shikokiana, we propose treating them as separate species. This proposition is also supported by a relatively high F_ST value of 0.22, indicating limited gene flow between H. shikokiana and H. minazukiflora. The clade comprising H. shikokiana and H. minazukiflora was found to be the sister group to Hosta sp. 3 collected from Wakayama Prefecture, Honshu. This result indicates that the divergence of H. shikokiana and its related species occurred in a larger geographic area, including regions beyond Shikoku. Further studies with more extensive sampling across a broader area are needed to fully understand the diversity of H. shikokiana and its related species.

Our phylogenetic analyses, morphological observations, and field investigations have led to the discovery of three additional new species: H. samukazemontana, H. longipedicellata, and H. takiminazukiflora. Among them, H. samukazemontana occupies an intermediate position between a clade consisting of H. shikokiana, H. minazukiflora, and H. sp. 3, and another clade comprising H. tardiva subsp. tardiva, H. tardiva subsp. densinervia, and H. scabrinervia (Fig. 2). This separate position supports the notion that H. samukazemontana is a distinct species. Hosta samukazemontana is endemic to Mt. Kanpu and its surrounding area where it thrives on cliffs. Morphologically, H. samukazemontana is characterized by its scapes deflected from cliffs and curving upwards at the tip. In the STRUCTURE analysis, H. samukazemontana as well as H. scabrinervia exhibited a mixture of two genetic identities, one identity dominated in H. tardiva subsp. densinervia and another identity shared among seven species. However, at K=5 in the hierarchical STRUCTURE analysis, H. samukazemontana exhibited a unique genetic identity. These findings, along with the results of phylogenetic analyses and morphological observations, provide support for considering H. samukazemontana as a distinct species.

Hosta longipedicellata and H. takiminazukiflora formed a clade with a bootstrap support of 100%, and this clade, along with another clade comprising seven other species, originated at the base of Clade 1. This phylogenetic relationship supports the differentiation of H. longipedicellata and H. takiminazukiflora from the seven other species. We propose treating them as two distinct species, taking into account their genetic differences (Fig. 4, Table 2) and morphological characteristics (Figs 14, 15, 18, 19). The distribution ranges of H. longipedicellata and H. takiminazukiflora are isolated by an upstream reach of the Yoshino River (L and T in Fig. 7C) and their geological substrates are mafic schist and pelitic schist, respectively (Geological Survey of Japan 2023). These two species may have differentiated due to geographical isolation and disruptive selection on distinct geological substrates.

The divergence between the two sympatric subspecies of H. takiminazukiflora is intriguing. Hosta takiminazukiflora subsp. grandis is characterized by several distinct features including a cordate leaf base, a glaucous lower surface, green petioles and erect scapes without purple dots. These characteristics remain consistent under fertilized cultivation. In the Maximum Likelihood tree (Fig. 2), seven samples of H. takiminazukiflora subsp. grandis and three samples of H. takiminazukiflora subsp. takiminazukiflora formed two monophyletic groups with bootstrap support of 99% and 97%, respectively, even though they were collected at the same location. Consequently, considering these samples as variations within a single population is challenging. The observed genetic differentiation, coupled with distinct morphological variations, is likely sustained by strong disruptive selection and some degree of reproductive isolation. Based on this evidence, we treat them as two separate subspecies. This is likely another instance of anacladogenetic speciation, which initially resulted in the ancestral taxon, H. takiminazukiflora subsp. takiminazukiflora, being paraphyletic.

Notably, the F_ST values between H. takiminazukiflora subsp. grandis and the other species were relatively high, ranging from 0.22 to 0.51, while F_ST values between the two subspecies were as low as 0.09. This suggests that H. takiminazukiflora subsp. grandis may have originated through hybridization between an unknown species and a lineage of H. takiminazukiflora subsp. takiminazukiflora. Further investigations are necessary to test this hypothesis and elucidate the origin and taxonomic status of H. takiminazukiflora subsp. grandis.

While Clade 1 was classified into nine separate species, Clade 4 was considered to be a single species, H. tosana. Following the earlier classification proposed by Maekawa (1948), we morphologically identified two varieties, H. tosana var. tosana and var. caput-avis, but both varieties were not monophyletic (Fig. 5). The F_ST value between H. tosana var. caput-avis and var. tosana was of 0.10, the second smallest following the 0.07 F_ST value between H. polyneuronoides and H. tardiva subsp. densinervia. This suggests a relatively weak differentiation between these two “varieties”. Furthermore, the results of the STRUCTURE analysis (Fig. 4F) showed that the JPN6553–55 samples identified as H. tosana var. tosana exhibited a mixture of genetic identities from H. tosana var. caput-avis and var. tosana, suggesting a possible origin through hybridization. The SplitsTree analysis (Fig. 6) placed the JPN6553–55 samples between other samples of H. tosana var. tosana and H. tosana var. caput-avis, supporting a hybrid origin. These findings suggest that H. tosana var. caput-avis and H. tosana var. tosana hybridize in the sympatric population along the Befu Valley. In contrast, H. tosana var. caput-avis in Yanase, the type locality (JPN12536-12539) and its vicinity (JPN12653) showed clear differentiation from H. tosana var. caput-avis in Kami City (Figs 5, 6), suggesting that “H. tosana var. caput-avis ” is polyphyletic. The available evidence indicates that var. caput-avis is not a distinct variety.

Additionally, we discovered another lineage designated as the Kagami River lineage, which was monophyletic and distinct from other lineages. The F_ST values between the Kagami River lineage and the two “varieties” were 0.15 and 0.16, respectively, indicating significant differentiation of the Kagami River lineage from the two “varieties”. It is likely that the Kagami River lineage can be distinguished as an infraspecific taxon. However, since H. tosana is widely recorded in the eastern part of Shikoku, further studies encompassing the entire range of H. tosana are needed to elucidate the taxonomic status of the Kagami River lineage.

Remarkably, seven species within Clade 1 (H. longipedicellata, H. minazukiflora, H. samukazemontana, H. scabrinervia, H. shikokiana, H. takiminazukiflora, and H. polyneuronoides) are densely distributed in the upper reaches and headwaters of the Yoshino River (Fig. 7B, C). This pattern suggests that speciation within Hosta Clade 1 was uniquely accelerated by factors specific to this region. Within Clade 1, H. polyneuronoides occupies a broader elevation range from 237 m to 1980 m. While H. polyneuronoides grows on rocks along streams and H. shikokiana thrives on the soil on the ridgeline of Mt. Ishizuchi, five other species grow on cliffs in geographically isolated localities (L, Sa, T, M, and an open square in Fig. 7C). Thus, spatial isolation between cliff habitats is likely a driving force behind the speciation of these five species. In addition to spatial isolation, the geological diversity of this region could also contribute to the speciation of Clade 1. For instance, H. shikokiana is restricted to Mt. Higashi-akaishi and its surroundings, where serpentine rock is exposed. Some endemic taxa associated with serpentine soil, such as Adenophora triphylla (Thunb.) A. DC. var. puellaris (Honda) H. Hara and Euphrasia microphylla Koidz., are known in this area (Yamanaka 1952, 1958). Furthermore, H. longipedicellata and H. takiminazukiflora are found on different geological substrates: mafic schist and pelitic schist, respectively (Geological Survey of Japan 2023). Studying how these species have adapted to cliff habitats and different substrates will yield a deeper understanding of the factors propelling the diversification of these species.

In the Taxonomy section, we provide a key to the taxa and update the taxonomy of Hosta species found on Shikoku Island, belonging to Clades 1 and 4. This update includes descriptions of six new taxa, encompassing five new species and one new subspecies, along with revised status for two existing taxa. However, we have refrained from revising the taxonomy of Clade 2, Clade 3, and Clade 6, pending further studies. Further comprehensive studies combining MIG-seq with meticulous morphological observations are necessary for these clades.

Below, we first treat H. tardiva subsp. tardiva and H. tardiva subsp. densinervia comb. & stat. nov., located on the top of the Maximum Likelihood tree for Clade 1 (Fig. 2). Secondly, we treat H. scabrinervia stat. nov. because it was compared with H. tardiva subsp. densinervia in the original description (Fujita and Tamura 2008). Thirdly, we treat two closely related species of Clade 1, H. shikokiana and H. minazukiflora sp. nov. Fourthly, we describe three species of Clade 1, H. takiminazukiflora sp. nov., H. longipedicellata sp. nov., and H. samukazemontana sp. nov. Fifthly, we treat a species of Clade 4, H. tosana.

Taxonomy

Key to the species of Hosta from Shikoku Island

1	Bracts ascending or spreading before flowering	Clade 2, 3, and 6; not treated below
–	Bracts tightly closed before flowering	2
2	Scapes deflected and gently bent like a bow, curved upwards at the tip	3
–	Scapes straight	4
3	Perianths 4.5–5 cm long, whitish. Flower bracts whitish purple, 1.7–2.1 cm long.	8. Hosta samukazemontana
–	Perianths 6.1–6.9 cm long, purple. Flower bracts whitish green, 2.2–5.2 cm long	9. Hosta tosana
4	Leaf margin strongly undulated. Leaves with decurrent to winged petioles, lustrous below when dried	4. Hosta shikokiana
–	Leaf margin either not undulated or weakly undulated. Leaves not or only slightly decurrent to petioles, not lustrous below when dried	5
5	Perianths purple or lavender; inside veins distinct	6
–	Perianths whitish; inside veins indistinct	7
6	Perianths purple; five distinct veins inside each lobe	1A. Hosta tardiva subsp. tardiva
–	Perianths lavender; three distinct veins inside each lobe	5. Hosta minazukiflora
7	Anthers 3.5–5 mm long	8
–	Anthers 2.5–3 mm long	9
8	Flowers more than 20 per scape. Anthers 5 mm long	1B. Hosta tardiva subsp. densinervia
–	Flowers 3–10 per scape. Anthers 3.5–4 mm long	2. Hosta polyneuronoides
9	Leaf blades glaucous abaxially, broadly ovate	6A. Hosta takiminazukiflora subsp. grandis
–	Leaf blades light green abaxially, not glaucous, ovate or oblong-ovate	10
10	Flower bracts 2–2.5 cm long, withering during flowering. Leaf veins papillose on abaxial surface	7. Hosta longipedicellata
–	Flower bracts (2–)2.5–4.6 cm long, fresh during flowering	11
11	Flowers 15–20 per inflorescence. Flower bracts white or purplish white. Petioles (20–)22–40 cm long. Leaf veins papillose or smooth on abaxial surface	3. Hosta scabrinervia
–	Flowers 4–13 per inflorescence. Flower bracts purplish green or purple at anthesis. Petioles 3–22 cm long. Leaf veins smooth on abaxial surface	6. Hosta takiminazukiflora subsp. takiminazukiflora

1. Hosta tardiva Nakai, Bot. Mag. (Tokyo) 44: 513. 1930.

Type

Japan. Prov. Awa (Ehime Pref.): tractu Myōtō, in oppido Kamomyō, n.d., J. Nikai 1377 (holotype TI, n.v.).

1A. Hosta tardiva subsp. tardiva

Phenology

August to September.

Distribution and habitat

Japan (Shikoku Island). This species is widely distributed in Shikoku and typically thrives on herbaceous slopes near villages. It is cultivated in Honshu but often naturalized when it escapes cultivation.

Conservation status

Using criterion D1 for IUCN Red List categories (IUCN 2012, IUCN 2022), we recommend that this species be qualified as LC (Least Concern) because its population is estimated to exceed 1000.

Japanese name

Nankai-giboshi (Nakai 1930).

Specimens examined

Japan. Kochi Pref.: Takaoka-gun, Kubokawa-cho, 8 Aug. 2001, Se. Fujii 8725 (JPN10003, FU!), ditto, 23 Aug. 2002, Se. Fujii 9271 (JPN10002, FU!); Cultivated at Makino Botanical Garden (32 specimens listed in the Suppl. material 1; FU!).

Note

In Shikoku, it usually grows near mountain villages, suggesting that its wild populations escaped from cultivation.

1B. Hosta tardiva subsp. densinervia (N. Fujita & M.N. Tamura) Yahara & Se.Fujii, comb. &, stat. nov.

urn:lsid:ipni.org:names:77331104-1

Fig. 8

Hosta kikutii var. densinervia N. Fujita & M. N. Tamura, Acta Phytotax. Geobot. 59: 34 (2008); Tamura & Fujita in Iwatsuki et al., Fl. Jap. IVb: 143 (2016). Type. JAPAN, Shikoku: Tokushima Pref, Miyoshi-gun, Ikeda-cho, Shikino-kami Ferry, 27 Sep. 1965, with flowers, S. Takafuji 253 (holotype KYO!, isotype KYO!).

Hosta kikutii var. polyneuron sensu Fujita, Acta Phytotax. Geobot. 27: 80. 1976, p.p.

Phenology

Mid-August to September.

Distribution and habitat

Japan (Ehime, Kagawa, Kochi, and Tokushima Prefectures). This subspecies is distributed in the lower reaches of the Yoshino River, the middle and lower reaches of the Niyodo River, the headwater of the Shinjyo River (near the headwaters of the Shimanto River), and the lower reaches of the Shimanto River (Fig. 7). It grows in open habitats on rocks and soil along the riverbanks.

Figure 8.

Flowering plants of H. tardiva subsp. densinervia A JPN1263,18 Aug. 2020 B JPN1332, 20 Aug. 2020. Scale bars: 20 cm (A); 2 cm (B).

Conservation status

Using criterion D1 for IUCN Red List categories (IUCN 2012, IUCN 2022), we recommend that this species be classified as LC (Least Concern) because its population is estimated to exceed 1000.

Japanese name

Sudare-giboshi (Maekawa 1952, excluding Yakushima-giboshi), Awa-giboshi (Fujita and Tamura 2008).

Specimens examined

Japan. Ehime Pref.: Kamiukena-gun, Omogokei, 13 Aug. 1970, N. Fujita 269, with flowers (KYO!); Saijyo City, 450 m elev., 6 Aug. 1971, H. Takahashi & N. Fujita 273 with young flower buds (KYO!); Nii-gun, 500 m elev., 30 Aug. 1955, G. Murata 9233 with flowers (KYO!). Kagawa Pref.: Mt. Otaki, 24 Sep. 1961, S. Sakaguchi s.n. with flowers (KYO!). Kochi Pref.: Agawa-gun, Niyodogawa-cho, Kuki, 80 m elev., 18 Aug. 2020, T. Yahara et al. JPN1255, 1258, 1263 with flowers (FU!); ditto, in the headwater of a tributary of Shinjyo River, 900 m elev., cultivated in Makino Botanical Garden, 8 Apr. 2021, T. Yahara et al. JPN3929 sterile (FU!); Agawa-gun, Ino-cho, Kashiki, 40 m elev., 20 Aug. 2020, T. Yahara et al. JPN1331–1333 with flowers (FU!); Takaoka-gun, Ochi-cho, Kataoka, 50 m elev., 20 Aug. 2020, T. Yahara et al. JPN1375–1377 with flowers (FU!); ditto, Tokoroyama, 13 Sep. 1962, G. Murata 17106 and 17107 with flowers (KYO!); Hata-gun, Nakamura-cho, along the Shimanto River, 20 Aug. 1913, Z. Tashiro s.n. with flowers (KYO!). Tokushima Pref.: Miyoshi City, Shikino-kami, 90 m elev., 11 Apr. 2021, sterile, T. Yahara et al. JPN4071, 4072, and 4075 (FU!); Ikeda-cho, along the Yoshino River, 16 Sep. 1971, C. Abe 43049 with flowers (KYO!); Miyoshi-gun, Minawa-mura, Kawasaki, along the Iya River, 8 Sep. 1958, T. Yamanaka 26270 with flowers (KYO!).

Note

This subspecies is morphologically more similar to H. polyneuronoides, rather than to H. tardiva subsp. tardiva.

2. Hosta polyneuronoides Yahara & Se.Fujii, sp. nov.

urn:lsid:ipni.org:names:77331106-1

Fig. 9

Diagnosis

Hosta polyneuronoides is distinguished from H. tardiva subsp. densinervia in having 3–10 flowers per inflorescence (in contrast to 15–25 in subsp. densinervia) and anthers 3.5–4 mm long (compared to 5 mm long).

Type

Japan. Kochi Pref.: Agawa-gun, Ino-cho, near Nitaki Bridge, 237 m elev., 20 Aug. 2020, T. Yahara, K. Fuse & H. Sato JPN1378 with flowers (holotype FU!).

Figure 9.

Flowering plants of H. polyneuronoides A JPN1378, 20 Aug. 2020 B JPN1327, 20 Aug. 2020. Scale bars: 2 cm (A, B).

Description

Herbs perennial. Leaves basal, spiral, long petiolate, 3–15 per ramet; blades ovate or oblong-ovate, 7.5–28.5 cm long, 3.3–15.3 cm wide, 0.9–2.9 times longer than width, thinly papery, glabrous on both surfaces, base cuneate to subcordate, apex acute to short acuminate, margin entire, veins 8–9 pairs, smooth on the lower surface; petioles 5–40 cm long, narrowly winged, 0.3–0.5 cm wide, glabrous. Scape 15–51 cm long, rachis terete, bract lanceolate, 4 cm long, 0.3–0.5 cm wide, light green, glabrous. Raceme 4.3–9 cm long, 3–10-flowered; flower bracts vivid (not withering) in anthesis, erect or diagonally spreading, whitish green, oblong-lanceolate, boat-shaped, 1.2–3.1 cm long, 0.1–0.8 cm wide, membranous, glabrous, apex acuminate. Flowers not fragrant, 5.7–6.0 cm long; pedicels 0.6–1.3 cm long, glabrous. Perianths white or very pale purple-white, funnel-form, 2.1–4.9 cm long, glabrous, 6-lobed; tube abruptly dilated from apical 2/3, lobes narrowly triangular, 0.7–1.2 cm long, 0.4–1 cm wide, apex acute. Stamens 6, 0.1–0.4 cm exerted from perianth; filaments white, free, 4.5–4.8 cm long, glabrous, anthers purple to yellow when fresh, dark blue-grey to light yellow when dried, 3.5–4 mm long. Ovary ellipsoid, 0.5 cm long, glabrous style 5.5–5.7 cm long, upwardly curved at the distal part, subequal to 1.5 cm exerted from perianth, glabrous, stigma capitate. Capsules or seeds not observed.

Phenology

Flowering from July to August.

Distribution and habitat

Japan (Kochi and Ehime Prefectures). Hosta polyneuronoides grows on open or shaded rocks and rock cliffs along rivers in the upper reaches and headwaters of the Yoshino River, and on the rocky ridgeline from Mt. Ishizuchi to Mt. Komochi-gongen and its vicinity.

Etymology

This subspecies was named for its resemblance to H. tardiva subsp. densinervia identified as Hosta kikutii var. polyneuron by Fujita (1976).

Conservation status

Using criterion D1 for IUCN Red List categories (IUCN 2012, IUCN 2022), we recommend that this species be classified as NT (Near threatened) because its population is estimated to exceed 1000 but may be declining under the influence of river bank construction and collection for horticulture.

Japanese name

Oku-sudare giboshi (new).

Additional specimens examined

Japan. Ehime Pref.: Niihama City, Mt. Higashi-akaishi, 970 m elev., 21 Jun. 2021, T. Yahara et al. JPN6269–6271 sterile (FU!); ditto, 1040 m elev., 21 Jun. 2021, T. Yahara et al. JPN6153–6156 sterile (FU!); Mt. Ishizuchi, 28 Jul. 1972, S. Takafuji 799 with flowers (KYO!); Kamiukena-gun, Mt. Ishizuchi, 1600 m elev., 23 Jun. 2021, T. Yahara et al. JPN6390 sterile (FU!); ditto, 1980 m elev., 23 Jun. 2021, T. Yahara et al. JPN6407 sterile (FU!); Saijyo City, Mts. Ishizuchi, from Yoake-toge to Mt. Nishinokanmuri-dake, 1600 m elev., 7 Aug. 1971, H. Takahashi & N. Fujita 226 with flowers (KYO!). Kochi Pref.: Nagaoka-gun, Motoyama-cho, Along Asemi River, 480 m elev., 19 Aug. 2020, T. Yahara et al. JPN1327–1329 with flower buds (FU!); Tosa-gun, Okawa-mura, 570 m elev., 24 Jun. 2021, T. Yahara et al. JPN6493–6495 sterile (FU!); Tosa-gun, Ohkawa-mura, Kawasaki, 400 m elev., 24 Jun. 2021, T. Yahara et al. JPN6501–6503 with flower buds (FU!); Tosa-gun, Ohkawa-mura, Takano, 24 Jun. 2022, K. Fuse et al. JPN12529, 12530, 12532, 12535 sterile (FU!); Tosa-gun, Ohkawa-mura, near Okina Waterfall, 443 m elev., Jul. 25, 2023, Se. Fujii, FJW49-1, 2 with flowers (MBK!); Tosa-gun, Tosa-cho, 24 Jun. 2022, K. Fuse et al. JPN12506-12508 sterile (FU!); Agawa-gun, Ino-cho, 750 m elevation, 23 Jun. 2021, T. Yahara et al. JPN6449–6451 sterile (FU!); Agawa-gun, Ino-cho, 900 m elevation, 23 Jun. 2021, T. Yahara et al. JPN6446–6448 sterile (FU!); Agawa-gun, Ino-cho, Kamegamori Forest Road, 1650 m elev., 22 Jun. 2021, T. Yahara et al. JPN6342, 6361 sterile (FU!); ditto, Mt. Komochi-gongen, 29 Jun. 2022, K. Fuse et al. JPN12663, 12664 sterile (FU!); Agawa-gun, Ino-cho, Ohmori-gawa dam, 755 m elev., 3 May 2022, Se. Fujii JPN12427 sterile (FU!); Agawa-gun, Ino-cho, Terakawa-Ohdaru, 29 Jun. 2022, K. Fuse et al. JPN12666–12668 sterile (FU!). The sterile specimens were identified using MIG-seq.

Note

Two specimens collected from Mt. Ishizuchi (S. Takafuji 799 and H. Takahashi & N. Fujita 226) were cited in the original description of H. shikokiana (Fujita 1976), but were identified as H. polyneuronoides by molecular evidence and also morphologically in having leaves not undulate along the margin and not lustrous below, perianths not purple and without dark purple veins, and anthers 0.4 mm long.

3. Hosta scabrinervia (N. Fujita & M.N. Tamura) Yahara & Se.Fujii, stat. nov.

urn:lsid:ipni.org:names:77331107-1

Fig. 10

Hosta kikutii var. scabrinervia N. Fujita & M. N. Tamura, Acta Phytotax. Geobot. 59: 34. 2008. Type: JAPAN. Tokushima Pref., Miyoshi-gun, Ohboke, along the river, 13 Jul. 1968, C. Abe 33197 (holotype KYO!).

Hosta kikutii var. polyneuron sensu Fujita, Acta Phytotax. Geobot. 27: 80. 1976, p.p.

Phenology

Flowering in July.

Distribution and habitat

Japan (Tokushima and Kochi Prefectures). The typical lineage of this species grows in rock crevices in the open habitats of riverbanks along the middle reach of Yoshino River. The upstream lineage (JPN12523-12525, FJS00006, FJS00007) grows on the wet cliffs in the upper reaches of the Yoshino River.

Figure 10.

Flowering plants of H. scabrinervia. A Abe 33197 (holotype KYO) B, C JPN15231, 23 Jul. 2022. Scale bars: 2 cm (A, B).

Conservation status

Using criterion D1 for IUCN Red List categories (IUCN 2012, IUCN 2022), we recommend that this species be classified as VU (Vulnerable) because its population is estimated to be between 250 and 1000.

Japanese name

Zaratsuki-giboshi (Fujita and Tamura 2008).

Additional specimens examined

Japan. Kochi Pref.: Nagaoka-gun, Ohtoyo-mura, Isodani, along Yoshino River, 8 Sep. 1958, T. Yamanaka 26222 with fruit (KYO!); Nagaoka-gun, Ohtoyo-cho, Higashidoi, Yoshino River, 4 Jul. 2002, H. Sasaoka FOK-603540 with flowers (MBK0170549!); left bank of Yoshino River, Okubo, Ohtoyo-cho, Nagaoka-gun, 26 Nov. 2017, A. Sakamoto et al. FOS-017749 with fruit (MBK0319724!); ditto, 18 Jul. 2004, M. Matsumoto et al. FOK-067597 with flowers (MBK0087490!); Tosa-gun, Ohkawa-mura, Takano, 24 Jun. 2022, K. Fuse et al. JPN12523–12525 (FU!); Tosa-gun, Ohkawa-mura, Kawasaki, Jul. 25, 2023, Se. Fujii FJW-50-1, 2 with flowers (MBK0342373!, MBK0342374!); Tosa-gun, Ohkawa-mura, Kogane Waterfall, 23 Jul. 2022, Se. Fujii JPN15231, 15232, with flowers (FU!); ditto, Jul. 25, 2023, Se. Fujii FJW-51-1–4 with flowers (MBK!); Nagaoka-gun, Motoyama-cho, in front of Kizenzan Park, 8 Aug. 1971, H. Takahashi & N. Fujita 214 with flowers and fruit (KYO!); . Tokushima Pref.: Miyoshi-gun, Oboke, 11 Apr. 2011, T. Yahara et al. JPN4080, 4082–4087 sterile (FU!); ditto, 22 Jul. 1967, C. Abe 33157 with fruits (KYO!).

Note

Fujita and Tamura (2008) distinguished var. scabrinervia from var. densinervia by the papillose lower surface of the lateral veins, but some plants of the type locality population have smooth lower lateral vein surfaces. Morphologically, H. scabrinervia is similar to H. tardiva subsp. densinervia but distinguished by the flowering season (July in contrast to August to September in the latter) and anther length (3 mm long vs. 5 mm long). The specimens collected from Kogane Waterfall (JPN15231, etc.) have exceptionally long pedicels (2.9–3.4 cm long), compared to 2–2.7 cm in other specimens. In this aspect, they are similar to H. longipedicellata, but can be distinguished by their flower bracts which remain fresh during flowering (unlike the withering flower bracts of H. longipedicellata). In the original description of H. kikutii var. scabrinervia, Fujita and Tamura (2008) cited specimens from Ehime and Kochi Prefectures, but further studies combining molecular analyses with subsequent morphological observations are required to confirm their identities. As far as examined, the range of H. scabrinervia is restricted to a narrow region in the middle reach of Yoshino River (Fig. 7).

4. Hosta shikokiana N. Fujita, Acta Phytotax. Geobot. 27: 93. 1976. Plants on Mt. Higashi-akaishi, Mt. Akaishi, Mt. Nishi-akaishi, and Mt. Shiragayama only.

Fig. 11

Type

Japan. Ehime Pref., Mt. Higashi-akaishi, 7 Aug. 1957, T. Yamanaka 22475 (holotype KYO!).

Phenology

Flowering in July and fruiting in August.

Distribution and habitat

Japan (Ehime and Kochi Prefectures) (endemic). This species grows in open habitats on the rocky ridgeline of Mt. Higashi-akaishi and its surrounding areas on Shikoku Island.

Figure 11.

Plants of H. shikokiana A JPN5249 B a flowering plant (photo taken on 12 Jul. 2018 by Se. Fujii; not collected). Scale bars: 10 cm (A); 5 cm (B).

Conservation status

Using criterion D1 for IUCN Red List categories (IUCN 2012, IUCN 2022), we recommend that this species be classified as VU (Vulnerable) because its population is estimated to be between 250 and 1000.

Japanese name

Shikoku-giboshi (Fujita 1976).

Additional specimens examined

Japan. Ehime Pref.: Mt. Higashi-akaishi, 15 Jul. 1952, T. Yamanaka 8901 with flowers (KYO!); ditto (as Mt. Akaishi), 9 Jul. 1928, G. Koidzumi s.n. with flower buds (KYO!); ditto, 1650 m elev., 21 Jun. 2021, T. Yahara et al. JPN6197, 6228, 6239, 6249, and 6252 with flower buds (FU!); ditto, 1640 m elev., 14 Jul. 1960, K. Tsuchiya 500 with flowers (KYO!); ditto, Mt. Hachimaki (a peak of Mts. Higashi-akaishi), 22 Jul. 2022, Se. Fujii JPN9953, 31 Jul. 2022, Se. Fujii JPN9954 with flowers (FU!); derived from Mt. Higashi-akaishi, cultivated in Makino Botanical Garden, 8 Apr. 2021, T. Yahara & Se. Fujii JPN3939–3942 sterile (FU!); between Mt. Higashiakaishi and Mt. Hutatsudake, 1600 m elev., 8 Sep. 1961, G. Murata 14981 with fruit (KYO!); ditto, cultivated stock of G. Murata 14981, G. Murata s.n. with flowers (KYO!). Kochi Pref.: Mt. Shiraga (cultivated stock), T. Yahara & Se. Fujii JPN3937, 8 Apr. 2021, sterile (FU!); ditto, Se. Fujii JPN9955, 23 Jul. 2021 with flowers (FU!).

Note

The following specimens cited by Fujita (1976) are not this species; the specimen from Mt. Kanpu (Yamawaki s.n., TI) is H. samukazemontana, and the specimens from Mt. Ishizuchi (Takahashi & Fujita 226, KYO; Takafuji s.n., KYO) are H. polyneuronoides.

5. Hosta minazukiflora Se. Fujii & Yahara, sp. nov.

urn:lsid:ipni.org:names:77331108-1

Figs 12, 13

Diagnosis

Hosta minazukiflora is similar to H. longipedicellata in having leaves shorter than 20 cm, straight scapes shorter than 43 cm, and perianth lobes 0.6–0.8 cm wide. However, H. minazukiflora is distinguished from H. longipedicellata by its lavender flowers (vs. whitish), shorter pedicels (1.1–1.2 cm vs. 2.5–3.3 cm), narrower leaves (2.9–6.2 cm wide vs. (5.4–)9–12 cm wide) smooth on the lower surface (vs. papillose), and the occurrence at elevations of 270–280 m (vs. 1700–1750 m).

Figure 12.

The type specimen of H. minazukiflora: Se. Fujii & K. Yabe FOS-012016 (holotype MBK).

Figure 13.

Flowering plants of H. minazukiflora. Photos taken in the type locality (on 25 Jun. 2017 by Se. Fujii; not collected). Scale bars: 20 cm (A); 2 cm (B).

Type

Japan. Kochi Pref.: Nagaoka-gun, Motoyama-cho, Asemi River, 25 Jun. 2017, Se. Fujii & K. Yabe FOS-012016 with flowers (holotype MBK0318414!).

Description

Herbs perennial. Leaves basal, spiral, long petiolate, 4–6 per ramet; blades oblong-ovate to oblong-lanceolate, 11.8–17.0 cm long, 2.9–6.2 cm wide, 2.4–4.3 times longer than width, thinly papery, glabrous on both surfaces, base cuneate, decurrent, slightly folded, apex acuminate, margin entire, veins in 6–9 pairs, smooth on the lower surface; petioles 8–19 cm long, narrowly winged, wing 0.1–0.3 cm wide, glabrous. Scape 18–43 cm long, terete. Raceme 10–15 cm long, 3–6-flowered; flower bracts vivid (not withering) in anthesis, erect, whitish purple, oblong-lanceolate, boat-shaped, 1.3–1.6 cm long, 0.2–0.4 cm wide, membranous, glabrous, apex acute. Flowers not fragrant, 5.2–5.9 cm long; pedicels 1.1–1.2 cm long, glabrous. Perianth whitish purple, funnel-form, 4.1–4.7 cm long, glabrous, 6-lobed; abruptly dilated from apical 2/3, lobes narrowly triangular, 1.2–1.4 cm long, 0.6–0.8 cm wide, apex acute. Stamens 6; filaments 4.6–5.4 cm long, white, free, glabrous, 0.5–0.7 cm exerted from perianth, upwardly curved at the distal part, anthers purple when fresh, dark blue-grey when dried, 3 mm long. Style 5.2–6.1 cm long, upwardly curved at the distal part, 1.1–1.4 cm exerted from the perianth, glabrous, stigma capitate. Ovary ellipsoid, 0.6 cm long, glabrous. Capsule green, cylindrical, 1.8–2.3 cm long, 0.3–0.4 cm wide, shallowly 3-angled. Seeds ellipsoid-ovoid, 2 mm long, with wings 4 mm long, black when dry.

Phenology

Flowering from mid to late June and fruiting in August.

Distribution and habitat

Japan (Kochi Prefecture). This species grows on rock cliffs along the Asemi River, a branch of the Yoshino River, in the central part of the Kochi Prefecture on Shikoku Island.

Etymology

The specific epithet was derived from the flowering season in June. Minazuki refers to June in Japanese.

Conservation status

Using criterion D1 for IUCN Red List categories (IUCN 2012, IUCN 2022), we recommend that this species be classified as EN (Endangered) because its population is estimated to be between 50 and 250.

Japanese name

Minazuki-giboshi (new).

Additional specimens examined

Japan. Kochi Pref.: Nagaoka-gun, Motoyama-cho, Asemi River, 280 m elev., 19 Aug. 2020, T. Yahara et al. JPN1308 sterile, JPN1309 sterile, and JPN1310 with fruit (FU!); ditto, 13 Jun. 2004, Y. Yamashita et al. FOK-066868 with flowers (MBK0083845!); ditto, 23 May 2002, Y. Kokami et al. FOK-055767 with flower buds (MBK0146148!); ditto, 270 m elev., 25 Jun. 2017, Se. Fujii & K. Yabe FOS-012017 with flowers (MBK0318415!).

Note

Hosta minazukiflora is sister to H. shikokiana (Figs 2, 3), but morphologically more similar to H. longipedicellata (see diagnosis) and H. scabrinervia. The type specimen Se. Fujii & K. Yabe FOS-012016 was first identified as H. kikutii var. scabrinervia. While the two species are distributed in the Yoshino River system, H. minazukiflora is easily distinguished from H. scabrinervia by its smaller leaves (12–17 cm long and 3–6 cm wide compared to 17–32 cm long and 7.5–15 cm wide in H. scarbrinervia), lavender perianths (in contrast to whitish perianths), and earlier flowering season in June (rather than July).

6. Hosta takiminazukiflora Se.Fujii & Yahara, sp. nov.

urn:lsid:ipni.org:names:77331109-1

Figs 14, 15

Diagnosis

Hosta takiminazukiflora is similar to H. longipedicellata and H. minazukiflora. It is distinguished from H. longipedicellata by leaves smooth on the lower surface (vs. papillose), pedicels 1.4–2.5 cm long (compared to 2.5–3.3 cm long), and flower bracts being fresh during flowering (in contrast to being withering), and from H. minazukiflora by perianths with a distinct midvein (as opposed to three distinct veins) and pedicels 1.4–2.5 cm long (compared to 1.1–1.2 cm long).

Figure 14.

The type specimen of H. takiminazukiflora: N. Inagaki et al. FOK-080097 (holotype MBK).

Figure 15.

A flowering plant of H. takiminazukiflora subsp. takiminazukiflora (JPN6481, 24 Jun. 2021). Scale bars: 5 cm (A); 2 cm (B).

Type

Japan. Kochi Pref.: Tosa County, Ookawa village, Mt. Higashikado-yama, 22 Jul. 2007, with flowers, N. Inagaki et al. FOK-080097 with flowers (holotype MBK0189902!).

Description

Herbs perennial. Leaves basal, spiral, long petiolate, 3–12 per ramet; blades oblong-ovate, 11–26.5 cm long, 3.6–10.5 cm, 2.1–3.3 times longer than wide, thinly papery, glabrous on both surfaces, base cuneate to obtuse, often decurrent, apex long acuminate, margin entire, veins in 6–10 pairs, smooth on the lower surface; petioles 3.4–22 cm long, narrowly winged, wing 0.1–0.6 cm wide, glabrous. Scape 8.6–28.5 cm long, terete. Raceme 5.9–18.5 cm long, 4–13-flowered; flower bracts vivid (not withering) in anthesis, erect, purplish light green, light bluish purple, ovate oblong-lanceolate, boat-shaped, 2–4.7 cm long, 0.2–0.7 cm wide, papery, glabrous, apex acute to acuminate. Flowers not fragrant; pedicels 1.4–2.5 cm long, glabrous; perianths 4–5.7 cm long, funnel-form, pale white-purple to light bluish-purple outside, almost white to light pale purple inside, midveins more or less purplish, glabrous, 6-lobed; tube dilated from apical 1/2, lobes triangular, 1.2–1.8 cm long, 0.5–1 cm wide, apex acute. Stamens 6, same or ca. 0.5 cm exerted from perianths; filaments 3.9–5.8 cm long, upwardly curved at the distal part, white, free, glabrous, anthers purple when fresh, dark blue-grey when dried, 3 mm long. Ovary ellipsoid, 0.6–0.7 cm long, style 4.5–6.5 cm long, upwardly curved at the distal part, up to 1 cm exerted from perianth, glabrous, stigma capitate. Young capsules 2.8 cm long (for MBK0087737).

Phenology

Flowering in late June to early August.

Distribution and habitat

Japan (Kochi Prefecture: Tosa County, endemic to Mt. Inamura, Mt. Higashikado, and the surrounding area). It grows on cliffs.

Etymology

A specific epithet is derived from its habit of growing on rock cliffs (called ‘taki’ in Kochi dialect) and flowering in June (Minazuki).

Conservation status

Using criterion D1 for IUCN Red List categories (IUCN 2012, IUCN 2022), we recommend that this species be classified as VU (Vulnerable) because its population is estimated to be between 250 and 1000.

Japanese name

Taki-minazuki-giboshi (new).

Additional specimens examined

Japan. Kochi Pref.: Tosa County, Ookawa village, 900 m, 27 Jul. 2004, N. Inagaki et al. FOK-067742 with flowers and young fruits (MBK0087737!); Tosa County, Tosa Town, 620 m elev., 24 Jun. 2021, T. Yahara et al. JPN6478–6486, 6490 with flowers (FU!); Agawa-gun, Ino-cho, Mt. Inamura, 1390 m elev., 2 Aug. 2019, Y. Oohira 14695 with flowers (MBK0314655!); Agawa-gun, Ino-cho, north cliff of Mt. Inamura, 23 Jul. 2022, Se. Fujii JPN15229, 15230 with flowers (FU!).

Note

The clade comprising this species and H. longipedicellata is sister to all the other species in Clade 1 (Fig. 2), and its morphological similarities to H. longipedicellata and H. minazukiflora are considered to be derived from a common ancestor.

6А. Hosta takiminazukiflora subsp. grandis Se.Fujii & Yahara, subsp. nov.

urn:lsid:ipni.org:names:77331110-1

Fig. 16

Diagnosis

Hosta takiminazukiflora subsp. grandis is distinguished from subsp. takiminazukiflora in having scapes 31–36 cm long (compared to 8.6–28.5 cm long in subsp. takiminazukiflora), broader leaf blades (13–14 cm wide compared to 3.6–10.5 cm wide), cordate leaf base (vs. cuneate to obtuse), glaucous lower leaf surface, lateral veins 12–13 pairs (as opposed to 6–10 pairs), and petioles green, not dotted with purple.

Figure 16.

A, B flowering plant of H. takiminazukiflora subsp. grandis (JPN6487, 24 Jun. 2021) C its lower leaf surface. Scale bars: 10 cm (A); 1 cm (B).

Type

Japan. Kochi Pref.: Tosa-gun, Tosa-cho, along Seto River, 24 Jun. 2021, T. Yahara et al. JPN6487 with flowers (holotype FU!).

Description

Herbs perennial. Leaves basal, spiral, long petiolate, 3–5 per ramet; blades oblong-ovate, 22–25 cm long, 13–14 cm wide, 1.7–1.8 times longer than width, thinly papery, glabrous on both surfaces, base usually cordate, often decurrent, apex acuminate, margin entire, veins in 12–13 pairs, smooth or slightly papillose on the lower surface; petioles 35–38 cm long, narrowly winged, wing 0.3 cm wide, glabrous. Scape 31–36 cm long, terete. Raceme 12–15.5 cm long, 4–18-flowered; flower bracts vivid (not withering) in anthesis, erect, purplish light green, oblong-lanceolate, boat-shaped, 2.8–4.3 cm long, 0.4–0.5 cm wide, papery, glabrous, apex acuminate. Flowers not fragrant; pedicels 1.9–2 cm long, pale purple, glabrous; perianths 3.8–4.9 cm long, funnel-form, glabrous, 6-lobed, pale blue-purple outside, almost white to pale purple inside, midveins faintly purplish adaxially; tube dilated from apical 2/3, lobes oblong-triangular, 1.2–1.3 cm long, 0.7–0.8 cm wide, apex acute. Stamens 6, ca. 0.5 cm exerted from perianth; filaments 4.3–4.5 cm long, upwardly curved at the distal part, white, free, glabrous, anthers purple when fresh, dark blue-grey when dried, 3 mm long. Ovary ellipsoid, 0.6–0.7 cm long, style 4.5–5.1 cm long, upwardly curved at the distal part, ca. 1 cm exerted from perianth, glabrous, stigma capitate. Capsules or seeds not observed.

Phenology

Flowering in late June.

Distribution and habitat

Japan (Kochi Prefecture) (endemic). This species grows on soil near waterfalls, whereas subsp. takiminazukiflora grows on rock cliffs.

Etymology

The specific epithet is derived from its larger plant size rather than that of the typical subspecies.

Conservation status

Using criterion D1 for IUCN Red List categories (IUCN 2012, IUCN 2022), we recommend that this species be classified as CR (Critically endangered) because its population is estimated to be fewer than 50.

Japanese name

Setogawa-giboshi (new).

Additional specimens examined

Japan. Kochi Pref.: Tosa-gun, Tosa-cho, along Seto River, 24 Jun. 2021, T. Yahara et al. JPN6488–6489 with flowers (FU!); ditto, 24 Jun. 2022, K. Fuse et al. JPN12518–12521 with flowers (FU!).

Note

In the type locality, subsp. grandis grows side by side with subsp. takiminazukiflora, and both subspecies flower simultaneously in late June, but the two subspecies are morphologically distinct and subsp. grandis formed a monophyletic group significantly separated from the sympatric population of subsp. takiminazukiflora. Therefore, subsp. grandis appears to be an evolutionary distinct lineage. This lineage is of particular interest as a research material for studying rapid speciation under disruptive selection and also serves as a valuable resource for breeding Hosta cultivars. Taking into account these characteristics, as well as its morphological distinctiveness, we distinguish it as a subspecies. Due to its limited population size, conservation measures are urgently necessary.

7. Hosta longipedicellata Se.Fujii & Yahara, sp. nov.

urn:lsid:ipni.org:names:77331162-1

Figs 17, 18

Diagnosis

Hosta longipedicellata is similar to H. scabrinervia, but distinguished by its flower bracts 2–2.2 cm long, withering during flowering (compared to (2–)2.2–4.6 cm long, fresh during flowering), smaller and narrower leaf blades (compared to up to 25 ×15 cm), and the occurrence at elevations at 1700–1750 m (in contrast to at 90–550 m).

Figure 17.

The type specimen of H. longipedicellata: S. Kobayashi & Y. Kurahashi FOK-077809 (holotype MBK).

Figure 18.

Plants of H. longipedicellata cultivated at the Makino Botanical Garden (photos taken on 3 Jun. 2022 by Se. Fujii). Scale bars: 10 cm (A); 1 cm (B).

Type

Japan. Kochi Pref.: Tosa-gun, Hongawa-mura, Mt. Tsutsujo, 1750 m elev., 1 Aug. 2006, S. Kobayashi & Y. Kurahashi FOK-077809 with flowers (holotype MBK0164413!).

Description

Herbs perennial. Leaves basal, spiral, long petiolate, 3 per ramet; blades ovate or oblong-ovate, 13.4–30.2 cm long, 5.4–11.8 cm, 2.2–3 time longer than wide, thinly papery, glabrous on both surfaces, base rounded to subcordate, apex acute to short acuminate, margin entire, veins in 7–10 pairs, papillose on the lower surface; petioles 15–34.3 cm long, narrowly winged, wing 0.1–0.2 cm wide, glabrous. Scape 26.5–32.5 cm long, terete. Raceme 7.6–9 cm long, 7–15-flowered; flower bracts withering in anthesis, erect, purplish light green in the upper part, light green in the lower part, oblong-lanceolate, boat-shaped, 2–2.5 cm long, 0.3–0.4 cm wide, membranous, glabrous, apex acute. Flowers not fragrant; pedicels 2.5–3.3 cm long, glabrous; perianth 4.4–6.4 cm long, funnel-form, glabrous, 6-lobed, whitish purple outside in flower buds, almost white outside and inside when flowering, midveins purplish adaxially; tube dilated from apical 2/3, lobes triangular, 1.0–1.5 cm long, 0.6–0.9 cm wide, apex obtuse. Stamens 6, slightly shorter than perianth, not exerted; filaments 4.6–4.8 cm long, upwardly curved at the distal part, white, free, glabrous, anthers purple when fresh, dark blue-grey when dried, 3 mm long. Ovary ellipsoid, 0.6–0.7 cm long, style 5.1–5.6 cm long, upwardly curved at the distal part, 1.1 cm exerted from perianths, glabrous, stigma capitate. Capsules or seeds not observed.

Phenology

Flowering from late July to early August.

Distribution and habitat

Japan (Kochi Prefecture). This species grows on rock cliffs at Mt. Tsutsujo, Mt. Tebako, Yasui Valley, and its vicinity in Kochi Prefecture on Shikoku Island.

Etymology

The specific epithet is derived from the long pedicel.

Conservation status

Using criterion D1 for IUCN Red List categories (IUCN 2012, IUCN 2022), we recommend that this species be classified as VU (Vulnerable) because its population is estimated to be between 250 and 1000.

Japanese name

Kamuro-giboshi (new).

Additional specimens examined

Japan. Kochi Pref.: Tosa-gun. Hongawa-mura, Mt. Tsutsujo, 28 Jun. 2022, K. Fuse et al. JPN12637–12639, 12647, 12654 sterile (FU!); Agawa-gun, Niyodogawa-cho, 612 m elev., 3 May 2022, Se. Fujii JPN12425, 12426 sterile (FU!); Agawa-gun, Niyodogawa-cho, Yasui Valley, 470 m elev., 30 Jun. 2022, K. Fuse et al. JPN12674, 12676, 12680 sterile, JPN12675 with flower buds (FU!); Agawa-gun, Niyodogawa-cho, Miyagahira, 612 m elev., Jul. 22, 2023, Se. Fujii FJW48-1, 2, 3 with flowers (MBK!); cultivated at Makino Botanical Garden, derived from Mt. Tsutsujo, 1750 m elev., T. Yahara et al. JPN3927 sterile (FU!); cultivated at Makino Botanical Garden, derived from Mt. Tebako, 1700 m elev., T. Yahara et al. JPN3936 sterile (FU!).

Note

The plant shown in Fig. 18 is a clone of JPN3927 that grew larger than the wild state after fertilization and was not used for the above description.

8. Hosta samukazemontana Se.Fujii & Yahara, sp. nov.

urn:lsid:ipni.org:names:77331111-1

Figs 19, 20

Diagnosis

Hosta samukazemontana is distinguished from H. polyneuronoides by anther size (3 mm vs. 3.5–4 mm), and inflorescences (deflected and curved upward at the top in contrast to erect and straight). Hosta samukazemontana is also similar to H. tosana in having inflorescences deflected and curved upward at the top, but distinguished from H. tosana by whitish perianths 4.5–5 cm long (in contrast to purple perianths 6.1–6.9 cm long) and whitish purple flower bracts 1.7–2.1 cm long (as opposed to whitish green peraiths 2.2–5.2 cm long).

Figure 19.

The type specimen of H. samukazemontana: Y. Kokami M83-255 (holotype MBK).

Figure 20.

Plants of H. samukazemontana A JPN1381 B JPN6325 C photo taken on 17 Jul. 2013 by Y. Kokami; not collected. Scale bars: 10 cm (A, B); 2 cm (C).

Type

Japan. Kochi Pref.: Tosa County, Hongawa Village, Mt. Kanpu, 25 Jul. 1983, Y. Kokami M83-255 with flowers (holotype MBK0175579!).

Description

Herbs perennial. Leaves basal, spiral, long petiolate, 2–6 per ramet; blades oblong-ovate, elliptic-ovate, ovate, 7–22 cm long, 3.6–10.7 cm wide, 1.8–2.9 time longer than wide, thinly papery, glabrous on both surfaces, base cuneate, rounded, cordate, decurrent, slightly folded, apex acuminate, margin entire, veins in 6–12 pairs, smooth on the lower surface; petioles 3.2–14.5 cm long, narrowly winged, wings 0.1–0.2 cm wide, glabrous. Scape 20–43 cm long, terete, deflected and curved upward at the top. Raceme 7–9 cm long, 5–18-flowered; flower bracts vivid (not withering) in anthesis, erect, whitish purple, oblong-lanceolate, boat-shaped, 1.7–2.1 long, 0.3–0.5 cm wide, membranous, glabrous, apex acuminate. Flowers 5–6 cm long; pedicels 1–1.8 cm long, glabrous. Perianths 4.5–5 cm long, purple-white outside, almost white inside, funnel-form, glabrous, 6-lobed, tube abruptly dilated from apical 2/3, lobes narrowly triangular, 1–1.3 cm long, 0.7–0.8 cm wide, apex acute, three veins distinct inside. Stamens 6; filaments 4.8 cm long, white, free, glabrous, almost as long as perianth, upwardly curved at the distal part, anthers purple when fresh, dark blue-grey when dried, 3 mm long. Ovary ellipsoid, 1 cm long, style 5 cm long, upwardly curved at the distal part, 0.5 cm exerted from perianth, glabrous, stigma capitate. Capsules green, cylindrical, 2.3–2.4 cm long, 0.4 cm wide, 3-angled. Seeds 2.2 mm long, with wings ca. 4 mm long, blackish brown.

Phenology

Flowering in July, and fruiting in August.

Distribution and habitat

Japan (Ehime and Kochi Prefectures) (endemic). This species grows on rock cliffs in the vicinity of Mt. Kanpu in the central part of the Kochi Prefecture on Shikoku Island.

Etymology

The specific epithet was derived from the old name of the type locality (Mt. Samukaze).

Conservation status

Using criterion D1 for IUCN Red List categories (IUCN 2012, IUCN 2022), we recommend that this species be classified as VU (Vulnerable) because its population is estimated to be between 250 and 1000.

Japanese name

Samukaze-giboshi (new).

Additional specimens examined

Japan. Kochi Pref.: Agawa-gun, Ino-cho, along Kamegamori Forest Road, 1660 m elev., 22 Jun. 2021, T. Yahara et al. JPN6325 with flower buds (FU!). Ehime Pref.: Saijyo City, W slope of Mt. Kanpu, 1050 m elev., 20 Aug. 2020, T. Yahara et al. JPN1380–1382 with fruit (FU!); Saijyo City, Mt. Kanpu 1600 m, 24 Jul. 2022, Se. Fujii JPN15235 with flowers (FU!); Mt. Kanpu, June 19, 1943, S. Yamawaki s.n. with flowers (TI!).

Note

This species was identified as H. shikokiana by Fujita (1976), but is distinguished from H. shikokiana in its whitish perianth color and scapes deflected and curved at the top when flowering and fruiting (Fig. 20). The distribution record of “H. kikutii var. caput-avis” from Mt. Kanpu in the Flora of Kochi Prefecture (Kochi Prefecture & Makino Memorial Foundation of Kochi Prefecture 2009) is a misidentification of this species.

9. Hosta tosana F. Maek., J. Fac. Sci. Univ. Tokyo, Sect. 3, Bot. 5: 376. 1940.

Fig. 21, 22, 23, 24

Hosta tosana F. Maek., J. Fac. Sci. Univ. Tokyo, Sect. 3, Bot. 5: 376. 1940. Type. Japan, Shikoku: Kochi Prefecture, Mt. Kajigamine (now called Kajigamori), T. Yoshinaga s.n. (unknown; not deposited in TI).

Hosta kikutii var. tosana (F. Maek.) F. Maek., Engei-daijiten 2:633. 1950; Tamura & Fujita in Iwatsuki et al. Fl. Jap. IVb: 143. 2016.

Hosta tosana var. caput-avis F. Maek., J. Jap. Bot. 22: 64. 1948. Type. JAPAN. Kochi Pref., Yanase, K. Yasui 332 (unknown; not deposited in TI).

Hosta caput-avis (F. Maek.) F. Maek. in Nakai, Iconogr. Pl. Asiae Orient. 5: 495. 1952.

Hosta kikutii var. caput-avis (F. Maek.) F. Maek., Engei-daijiten (Encycl. Hort.) 2: 633. 1950; Fujita, Acta Phytotax. Geobot. 27: 79. 1976.

Phenology

Flowering in June to July.

Distribution and habitat

Japan (Kochi and Tokushima Prefectures). This species grows on wet slopes, rocky cliffs, rocky riverbanks, and tree trunks in the elevations from 77 m to 1420 m.

Figure 21.

Plants of H. tosana Kagami River lineage A photo taken on 7 Jul. 2021 by Se. Fujii, not collected B JPN1307. Scale bars: 5 cm (A, B).

Conservation status

Using criterion D1 for IUCN Red List categories (IUCN 2012, IUCN 2022), we recommend that this species be classified as NT (Near threatened) because its population is estimated to exceed 1000 but may be declining under the influence of deer browsing, river bank construction, and collection for horticulture.

Figure 22.

A flowering plant of H. tosana (a form corresponding to var. tosana, JPN6616). Scale bars: 5 cm (A); 1 cm (B).

Japanese name

Tosa-no-giboshi (Maekawa 1940), Unazuki-giboshi (Maekawa 1948).

Specimens examined

Morphs corresponding to “var. tosana ”:

Kochi Pref.: Locality not specified, cultivated in Tokyo, 1 Aug. 1943, Maekawa 7043 with flowers (TI; the cultivated plant may have been collected by T. Yoshinaga from the type locality, as this is the only specimen of this taxon deposited in TI, where Maekawa studied Hosta); Nagaoka-gun, Otoyo-cho, Mt. Kajigamori, 30 Apr. 2022, Se. Fujii JPN12424 sterile (FU!); Kami City, Monobe Town, O-dochi, 500 m elev., 25 Jun. 2021, T. Yahara et al. JPN6619–6620 with flowers (FU!); ditto, Go-o-do, 553 m elev., cultivated at Makino Botanical Garden, 8 Apr. 2021, T. Yahara & Se. Fujii JPN3931–3933 sterile (FU!); ditto, 600 m elev., 25 Jun. 2021, T. Yahara et al. JPN6614 & 6616 with flowers, JPN6611–6613, 6615 sterile (FU!); Miyanose, 400 m elev., T. Yahara et al. JPN6603 sterile (FU!); Befu, 550 m elev., cultivated at Makino Botanical Garden, 8 Apr. 2021, T. Yahara & Se. Fujii JPN3934 sterile (FU!); Befu Valley, 700 m elev., 25 Jun. 2021, T. Yahara et al. JPN6553–6555 with flowers (FU!); Aki City, 77 m elev., cultivated at Makino Botanical Garden, 8 Apr. 2021, T. Yahara & Se. Fujii JPN3935 sterile (FU!); Aki-gun, Kitagawa-mura, Shima, along the Nahari River, 28 May 2022, K. Fuse et al. JPN12653 sterile (FU!). Tokushima Pref.: Mt. Tsurugi, 14 Aug. 1931, Z. Tashiro s.n. with flowers (KYO!); ditto, below Minokoshi, 1400 m elev., 14 Aug. 1954, G. Murata 7977 with flowers (KYO!); ditto, 1420 m elev., 23 Jun. 2022, T. Yahara et al. JPN12821–12823 sterile (FU!); Higashiiyayama-mura, Inter Sugeoi et Minokoshi, 800 m elev., 12 Aug. 1954, G. Murata 7814 & 7816 with flowers and young fruits (KYO!); Naka County, Wajiki Town, 14 Jul. 1965, S. Takafuji 216 with flowers (KYO!); ditto, Kizawa Village, 24 Jul. 1980, S. Takafuji 1442 with flowers (KYO!); ditto, Riu-toge in Miyahama-mura, 2 Jul. 1952, G. Murata 5752 with flowers (KYO!); ditto, Kizu Village, 24 Jul. 1974, S. Takafuji 950 with flowers (KYO!).

Figure 23.

Plants of H. tosana A a form corresponding to var. caput-avis (JPN6524) B a form corresponding to var. tosana (JPN6555). Scale bars: 5 cm (A, B).

Morphs corresponding to “var. caput-avis ”:

Kochi Pref.: Aki-gun, Yanase, 25 Jun. 2020, K. Fuse et al. 12536–12539 sterile (FU!); Kami City, Monobe Town, Befu Valley, 25 Jun. 2021, T. Yahara et al. JPN6524–6526 with flower buds, 6571 with flowers, 6572 with flower buds (FU!); Mt. Ishidate, 25 Jun. 2021, T. Yahara et al. JPN6585–6587, 6593–6596 with flower buds (FU!); on the river bank of Yanase, 1000 m elev., 20 Jul. 1958, S. Hatusima 22009A with wilted flowers (KAG056576!); transplanted from Yanase, 29 Jun. 1960, S. Hatusima s.n. with flowers (KAG056577!); ditto, 10 Jun. 1967, Hatusima s.n. with flowers (KAG056578!).

Figure 24.

Plants of H. tosana (a form corresponding to var. caput-avis) A JPN6571 B JPN6572. Scale bars: 10 cm (A); 1 cm (B).

The Kagami River lineage:

Kochi Pref.: Kochi City, Tosayama, Hirose, 85 m elev., 3 Oct. 2008, Hamaguchi & N. Shintani PRC-00118 with fruits (MBK0208327!); ditto, 14 Jul. 2001, K. Hosokawa et al. FOK-001722 with flowers (MBK0104375!); Kuwao, 130 m elev., 14 Jul. 2013, A. Sakamoto FOS-004959 with flowers (MBK0247214!, MBK0247215!); from Kuwao to Tsuami, 27 Jul. 1968, N. Naruhashi & M.Wakabayashi 223 with flowers (KYO!); Namekawa, 30 m elev., 19 Aug. 2020, T. Yahara et al. JPN1265 with fruit, JPN1266 sterile, JPN1267 sterile, JPN1269 with fruit (FU!); Kajitani, 160 m elev., 29 July 2013, A. Sakamoto FOS-005031 with young fruits (MBK0247386!) and FOS-005032 with young fruits (MBK0247387!); Miyanokubo, 160 m elev., 14 Jul. 2013, A. Sakamoto FOS-004958 with flowers (MBK0247213!); ditto, 170 m elev., 14 Jul. 2013, A. Sakamoto FOS-004957 with fruits (MBK0247212!); Oh-ana Valley, 110 m elev., 19 Aug. 2020, T. Yahara et al. JPN1298 sterile, JPN1299 sterile, JPN1307 with flowers (FU!); Kagami-mura, Kawaguchi, along the Kagami River, 9 Jul. 1958, T. Yamanaka 25548 with flowers (KYO!).

Note

The Kagami River lineage is sister to a clade comprising all other samples (Fig. 5). It is morphologically distinguished from other samples by having leaves with fewer lateral veins (5–11 compared to 10–13) that run at wider intervals (1 cm vs 0.7 cm), and it is distinct in its occurrence at elevations below 300 m (compared to elevations ranging from 77 m to 1420 m, usually above 300 m). It is likely that this lineage could be classified as a subspecies of H. tosana or a separate species. However, further studies are necessary to arrive at definitive conclusions regarding this lineage because H. tosana is a polymorphic species widely distributed in the eastern part of Shikoku. Nonetheless, our collections did not cover the entire range of this species.

According to the MIG-seq tree, two genetically distinct lineages are distributed along the Monobe River in Kami City (Figs 5, 6). The lineages were identified as H. tosana var. tosana and H. tosana var. caput-avis. However, population genetic evidence shows that the two varieties hybridize in the sympatric population along the Befu Valley. The available evidence indicates that var. caput-avis is difficult to distinguish from var. tosana in Kami City.

While the two varieties appeared to lack distinction in Kami City, the population of Yanase, which serves as the type locality for H. tosana var. caput-avis, exhibited clear differentiation from the populations in Kami City. However, just like in the case of the Kagami River lineage, further studies using a larger sample size are necessary to definitively conclude about the Yanase lineage.

A lineage distributed in SE Kinki, Honshu (H. sp. 4 in Fig. 1) was identified as H. kikutii var. caput-avis by Fujita (1976) and Fujita and Tamura (2008), and later treated as H. kikutii var. tosana by Tamura and Fujita (2016). However, it is not related to H. tosana of Clade 4, but to H. longipes of Clade 6 (Fig. 1).

Acknowledgements

DNA samples and voucher specimens of Hosta spp. were collected in the protected areas of Ashizuri-Uwaumi National Park, Ishizuchi National (Kokutei) Park, Tsurugi National (Kokutei) Park, and Akaishi Mountains Nature Reserve, with permission from the Ministry of Environment, Ehime Prefecture, and Tokushima Prefecture. We thank Satoru Kinoshita and Yasushi Ibaragi for identifying the location of H. tosana in Tokushima Prefecture, and Shinji Fujii for providing DNA samples of H. densiflora subsp. densiflora (FJI17566) and H. sp. 3 (FIJ12268). We thank the Ministry of the Environment’s Rare Species Conservation Promotion Office and Saki Funamoto of Kyushu Open University for their assistance in obtaining collection permits. We are also grateful to the curators of FU, KAG, KYO, MBK, TI and VNMN for their help to access specimens for our studies.

Additional information

Conflict of interest

The authors have declared that no competing interests exist.

Ethical statement

No ethical statement was reported.

Funding

This study was supported by the Environment Research and Technology Development Fund (JPMEERF20204001 and JPMEERF20234001) of the Ministry of Environment, Japan.

Author contributions

Data curation: TY, SF. Formal analysis: YS, TY, SKH. Investigation: HS, KF. Resources: YK, KF, HS, SF. Writing - original draft: TY, ST. Writing - review and editing: TY.

Author ORCIDs

Tetsukazu Yahara https://orcid.org/0000-0001-5105-7152

Shun K. Hirota https://orcid.org/0000-0002-6104-1119

Hiroyuki Sato https://orcid.org/0000-0001-6003-6492

Shuichiro Tagane https://orcid.org/0000-0002-1974-7329

Yoshihisa Suyama https://orcid.org/0000-0002-3136-5489

Data availability

All of the data that support the findings of this study are available in the main text or Supplementary Information.

References

Bolger AM, Lohse M, Usadel B (2014) Trimmomatic: A flexible trimmer for Illumina Sequence Data. Bioinformatics 30(15): 2114–2120. https://doi.org/10.1093/bioinformatics/btu170

Carnicero P, Schönswetter P, Garcia-Jacas N, Galbany-Casals M (2019) Is there a need for accepting paraphyletic taxa? A case study in the Sardinian endemic Cymbalaria muelleri (Plantaginaceae). Botanical Journal of the Linnean Society 191(3): 325–338. https://doi.org/10.1093/botlinnean/boz052

Catchen J, Hohenlohe PA, Bassham S, Amores A, Cresko WA (2013) Stacks: An analysis tool set for population genomics. Molecular Ecology 22(11): 3124–3140. https://doi.org/10.1111/mec.12354

Chen X, Boufford DE (2000) Hosta. Flora of China 24: 204–205. http://flora.huh.harvard.edu/china/PDF/PDF24/hosta.pdf

Doyle JJ, Doyle JL (1990) Isolation of plant DNA from fresh tissue. Focus (San Francisco, Calif. ) 12: 13–15.

Earl DA, von Holdt BM (2012) STRUCTURE HARVESTER: A website and program for visualizing STRUCTURE output and implementing the Evanno method. Conservation Genetics Resources 4(2): 359–361. https://doi.org/10.1007/s12686-011-9548-7

Evanno G, Regnaut S, Goudet J (2005) Detecting the number of clusters of individuals using the software STRUCTURE: A simulation study. Molecular Ecology 14(8): 2611–2620. https://doi.org/10.1111/j.1365-294X.2005.02553.x

Fujii S (2018) A travelogue on Hosta in Shikoku. Shumino-sanyaso 2018(12): 3–13. [Wild plants for hobbies]

Fujita N (1976) The genus Hosta in Japan. Acta Phytotaxonomica et Geobotanica 27: 66–96. https://doi.org/10.18942/bunruichiri.KJ00002992172

Fujita N, Tamura MN (2008) Two new varieties and one change of status in Hosta (Asparagaceae). Acta Phytotaxonomica et Geobotanica 59(1): 31–36. https://doi.org/10.18942/apg.KJ00004899884

Geological Survey of Japan (2023) GeomapNavi. https://gbank.gsj.jp/geonavi/geonavi.php#12,33.72341,133.26548 [Accessed on August 13, 2023]

Huson DH, Bryant D (2006) Application of phylogenetic networks in evolutionary studies. Molecular Biology and Evolution 23(2): 254–267. https://doi.org/10.1093/molbev/msj030

IUCN (2012) IUCN Red List Categories and Criteria, Version 3.1, 2^nd edn. IUCN Species Survival Commission. IUCN, Gland and Cambridge. https://portals.iucn.org/library/node/10315

IUCN (2022) Guidelines for using the IUCN Red List Categories and Criteria, Version 15.1. IUCN, Gland and Cambridge. https://nc.iucnredlist.org/redlist/content/attachment_files/RedListGuidelines.pdf

Jones Jr SB (1989) Hosta yingeri (Liliaceae/Funkiaceae): A new species from Korea. Annals of the Missouri Botanical Garden 76(2): 602–604. https://doi.org/10.2307/2399504

Kaneko K (1970) Cytological studies on some Species of Hosta VI. Karyotypes of H. undulata, H. undulata var. erromena, H. undulata var. albomarginata, H. tardiva and H. sacra. Botanical Magazine Tokyo 83(979): 27–35. https://doi.org/10.15281/jplantres1887.83.27

Kochi Prefecture & Makino Memorial Foundation of Kochi Prefecture (2009) Flora of Kochi. Kochi Prefecture & Makino Memorial Foundation of Kochi Prefecture, Kochi, 844 pp.

Kopelman NM, Mayzel J, Jakobsson M, Rosenberg NA, Mayrose I (2015) CLUMPAK: A program for identifying clustering modes and packaging population STRUCTURE inferences across K. Molecular Ecology Resources 15(5): 1179–1191. https://doi.org/10.1111/1755-0998.12387

Lee S, Kim K, Lee B, Lim C (2019) Complete chloroplast genomes of all six Hosta species occurring in Korea: Molecular STRUCTUREs, comparative, and phylogenetic analyses. BMC Genomics 20(1): 833. https://doi.org/10.1186/s12864-019-6215-y

Maekawa F (1940) The genus Hosta. Journal of Faculty of Science, Imperial University of Tokyo, section 3 Botany 5: 317–425.

Maekawa F (1948) A new variety, Hosta kikutii var. caput-avis. Shokubutsu Kenkyu Zasshi 22: 64.

Maekawa F (1952) Hosta caput-avis. In: Nakai T (Ed. ) Iconographia Plaritarum Asiae Orientalis 5(2): 495–497.

Nakai T (1930) Notulæ ad Plantas Japoniæ & Koreæ 39. Botanical Magazine Tokyo 44: 507–537. https://doi.org/10.15281/jplantres1887.44.507

Pritchard JK, Stephens M, Donnelly P (2000) Inference of population STRUCTURE using multilocus genotype data. Genetics 155(2): 945–959. https://doi.org/10.1093/genetics/155.2.945

Reydon TAC, Kunz W (2021) Classification below the species level: When areinfraspecific groups biologically meaningful? Biological Journal of the Linnean Society, Linnean Society of London 134(1): 246–260. https://doi.org/10.1093/biolinnean/blab067

Rochette NC, Rivera-Colón AG, Catchen JM (2019) Stacks 2: Analytical methods for paired-end sequencing improve RADseq-based population genomics. Molecular Ecology 28(21): 4737–4754. https://doi.org/10.1111/mec.15253

Roesti M, Salzburger W, Berner D (2012) Uninformative polymorphisms bias genome scans for signatures of selection. BMC Evolutionary Biology 12(1): 94. https://doi.org/10.1186/1471-2148-12-94

Stamatakis A (2014) RAxML version 8: A tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics 30(9): 1312–1313. https://doi.org/10.1093/bioinformatics/btu033

Suyama Y, Matsuki Y (2015) MIG–seq: An effective PCR–based method for genome-wide single-nucleotide polymorphism genotyping using the next–generation sequencing platform. Scientific Reports 5(1): 16963. https://doi.org/10.1038/srep16963

Suyama Y, Hirota SK, Matsuo A, Tsunamoto Y, Mitsuyuki C, Shimura A, Okano K (2022) Complementary combination of multiplex high-throughput DNA sequencing for molecular phylogeny. Ecological Research 37(1): 171–181. https://doi.org/10.1111/1440-1703.12270

Tamura MN (2015) Hosta. In: Ohashi H, Kadota Y, Murata J, Yonekura K (Eds) Wildflowers of Japan vol. 1p. Heibonsha, Tokyo, 250–253.

Tamura MN, Fujita N (2016) Hosta. In: Iwatsuki K, Yamazaki T, Boufford DE, Ohba H (Eds) Flora of Japan IVb. Kodansha Scientific, Tokyo, 140–147.

Yahara T, Hirota SK, Fuse K, Sato H, Tagane S, Suyama Y (2021) Validation of Hosta alata (Asparagaceae) as a new species and its phylogenetic affinity. PhytoKeys 181: 79–93. https://doi.org/10.3897/phytokeys.181.64245

Yamanaka T (1952) Some problems on the flora and the vegetation of serpentine region. Acta Phytotaxonomica et Geobotanica 14: 195–197.

Yamanaka T (1958) Serpentine flora of Mt. Nishi-akaishi, Shikoku. Acta Phytotaxonomica et Geobotanica 17: 94–96.

Supplementary materials

Supplementary material 1

20 DNA samples and voucher specimens from 70 localities for 30 taxa of Hosta in Japan

Tetsukazu Yahara, Shun K. Hirota, Seiko Fujii, Yasushi Kokami, Kengo Fuse, Hiroyuki Sato, Shuichiro Tagane, Yoshihisa Suyama

Data type: xlsx

This dataset is made available under the Open Database License (http://opendatacommons.org/licenses/odbl/1.0/). The Open Database License (ODbL) is a license agreement intended to allow users to freely share, modify, and use this Dataset while maintaining this same freedom for others, provided that the original source and author(s) are credited.

Download file (27.16 kb)

Supplementary material 2

Sample sets

Tetsukazu Yahara, Shun K. Hirota, Seiko Fujii, Yasushi Kokami, Kengo Fuse, Hiroyuki Sato, Shuichiro Tagane, Yoshihisa Suyama

Data type: xlsx

This dataset is made available under the Open Database License (http://opendatacommons.org/licenses/odbl/1.0/). The Open Database License (ODbL) is a license agreement intended to allow users to freely share, modify, and use this Dataset while maintaining this same freedom for others, provided that the original source and author(s) are credited.

Download file (10.50 kb)

Supplementary material 3

Changes of ∆K with K in the samples of H. polyneuronoides and H. tardiva subsp. densinervia (A) and the other species of group 1 (B)

Tetsukazu Yahara, Shun K. Hirota, Seiko Fujii, Yasushi Kokami, Kengo Fuse, Hiroyuki Sato, Shuichiro Tagane, Yoshihisa Suyama

Data type: docx

This dataset is made available under the Open Database License (http://opendatacommons.org/licenses/odbl/1.0/). The Open Database License (ODbL) is a license agreement intended to allow users to freely share, modify, and use this Dataset while maintaining this same freedom for others, provided that the original source and author(s) are credited.

Download file (126.89 kb)

﻿Abstract

Key words

﻿Introduction

﻿Materials and methods

﻿Field surveys and samples examined

﻿DNA isolation, sequencing, and construction of SNP-based phylogenetic trees

﻿Morphological observations

﻿Data resources

﻿Results

﻿Phylogenetic tree reconstructed using MIG-seq

﻿Morphological and phenological divergence between taxa with adjacent distribution ranges

(1) Hosta tardiva subsp. densinervia and H. polyneuronoides

(2) Hosta tardiva subsp. densinervia, H. polyneuronoides, and H. scabrinervia

(3) Hosta polyneuronoides and H. shikokiana

(4) Hosta minazukiflora and H. shikokiana

(5) Hosta polyneuronoides and H. minazukiflora

(6) Hosta takiminazukiflora and H. minazukiflora

(7) Hosta takiminazukiflora subsp. takiminazukiflora and subsp. grandis

(8) Hosta longipedicellata and H. takiminazukiflora

(9) Hosta samukazemontana, H. shikokiana, and H. polyneuronoides

﻿Morphological and phenological observations on H. tosana

﻿Genetic divergence between taxa

﻿Discussion

﻿Taxonomy

﻿Key to the species of Hosta from Shikoku Island

﻿1. Hosta tardiva Nakai, Bot. Mag. (Tokyo) 44: 513. 1930.

Type

﻿1A. Hosta tardiva subsp. tardiva

Phenology

Distribution and habitat

Conservation status

Japanese name

Specimens examined

Note

﻿1B. Hosta tardiva subsp. densinervia (N. Fujita & M.N. Tamura) Yahara & Se.Fujii, comb. &, stat. nov.

Phenology

Distribution and habitat

Conservation status

Japanese name

Specimens examined

Note

﻿2. Hosta polyneuronoides Yahara & Se.Fujii, sp. nov.

Diagnosis

Type

Description

Phenology

Distribution and habitat

Etymology

Conservation status

Japanese name

Additional specimens examined

Note

﻿3. Hosta scabrinervia (N. Fujita & M.N. Tamura) Yahara & Se.Fujii, stat. nov.

Phenology

Distribution and habitat

Conservation status

Japanese name

Additional specimens examined

Note

﻿4. Hosta shikokiana N. Fujita, Acta Phytotax. Geobot. 27: 93. 1976. Plants on Mt. Higashi-akaishi, Mt. Akaishi, Mt. Nishi-akaishi, and Mt. Shiragayama only.

Type

Phenology

Distribution and habitat

Conservation status

Japanese name

Additional specimens examined

Note

﻿5. Hosta minazukiflora Se. Fujii & Yahara, sp. nov.

Diagnosis

Type

Description

Phenology

Distribution and habitat

Etymology

Conservation status

Japanese name

Additional specimens examined

Note

﻿6. Hosta takiminazukiflora Se.Fujii & Yahara, sp. nov.

Diagnosis

Abstract

Introduction

Materials and methods

Field surveys and samples examined

DNA isolation, sequencing, and construction of SNP-based phylogenetic trees

Morphological observations

Data resources

Results

Phylogenetic tree reconstructed using MIG-seq

Morphological and phenological divergence between taxa with adjacent distribution ranges

Morphological and phenological observations on H. tosana

Genetic divergence between taxa

Discussion

Taxonomy

Key to the species of Hosta from Shikoku Island

1. Hosta tardiva Nakai, Bot. Mag. (Tokyo) 44: 513. 1930.

1A. Hosta tardiva subsp. tardiva

1B. Hosta tardiva subsp. densinervia (N. Fujita & M.N. Tamura) Yahara & Se.Fujii, comb. &, stat. nov.

2. Hosta polyneuronoides Yahara & Se.Fujii, sp. nov.

3. Hosta scabrinervia (N. Fujita & M.N. Tamura) Yahara & Se.Fujii, stat. nov.

4. Hosta shikokiana N. Fujita, Acta Phytotax. Geobot. 27: 93. 1976. Plants on Mt. Higashi-akaishi, Mt. Akaishi, Mt. Nishi-akaishi, and Mt. Shiragayama only.

5. Hosta minazukiflora Se. Fujii & Yahara, sp. nov.

6. Hosta takiminazukiflora Se.Fujii & Yahara, sp. nov.

6А. Hosta takiminazukiflora subsp. grandis Se.Fujii & Yahara, subsp. nov.

7. Hosta longipedicellata Se.Fujii & Yahara, sp. nov.

8. Hosta samukazemontana Se.Fujii & Yahara, sp. nov.

9. Hosta tosana F. Maek., J. Fac. Sci. Univ. Tokyo, Sect. 3, Bot. 5: 376. 1940.

Acknowledgements

Additional information