Taxonomic revision of Zingiber (Zingiberaceae) of Taiwan

. The genus Zingiber contains about 180 species distributed mainly in tropical regions. Several species of Zingiber are cultivated globally for their medicinal and culinary value, such as true gingers ( Z. officinale ), bitter gingers ( Z. zerumbet ), and Z. purpureum . In Taiwan, two endemic species and one incompletely known taxon were recorded in the last edition of Flora of Taiwan , and several taxonomic issues still remain unresolved. Therefore, we revised the Taiwanese Zingiber based on morphological, palynological, anatomical, and molecular evidence, as well as their distribution. The results showed that floral characters such as labellum, fertile bracts, and corolla tubes are of great taxonomic value in distinguishing taxa of Zingiber of Taiwan. Accordingly, five species are treated in the present study, namely Z. chengii Y.H.Tseng, C.M.Wang & Y.C.Lin, Z. mioga Thunb., Z. oligophyllum K.Schum., Z. pleiostachyum K.Schum., and Z. shuanglongense C.L.Yeh & S.W.Chung. Zingiber mioga might be a newly naturalized species to Taiwan. Zingiber kawagoii Hayata and Z. koshunense C.T.Moo are treated as synonyms of Z. pleiostachyum .

A haplotype network was reconstructed by using TCS ver. 1.21 (Clement et al. 2000) to visualize the genetic relationship of the three species between each haplotype and population. The long-fragment indels in three cpDNA fragments were all treated as single-step mutations by manual recoding. Uncorrelated genetic distances between Z. koshunense, Z. pleiostachyum, and Z. shuanglongense were calculated by MEGA 11 (Tamura et al. 2021).

Characters of Zingiber from Taiwan
We examined and compared several characters of species of Zingiber from Taiwan, and those that were shown to have taxonomic value are listed in Table 1.

Rhizome
The cross sectional color of the rhizome varies from purple in Z. shuanglongense to yellow in other taxa (Fig. 1).

Ligule
The ligule is bilobed except in Z. oligophyllum, in which it is reduced (Fig. 2).

Inflorescence
The number of flowers per inflorescence and the shape of the bract, including the bracteole, are diagnostic characters. Zingiber chengii has only one to three loosely arranged flowers in an inflorescence. The other taxa all have more than three flowers that are densely arranged. The fertile bracts are usually narrowly ovate except in Z. oligophyllum where they are broadly ovate (Fig. 3).

Flowers
Floral characters are often regarded as diagnostic by several scholars Yeh et al. 2012Yeh et al. , 2013Bai et al. 2016;Ohba 2016), and we found that to be true for the classification and identification of Taiwanese Zingiber. Each character is listed and described below.

Corolla tube
Two colors of corolla tube are classified here, creamy-white and yellow. Zingiber chengii and Z. shuanglongense have creamy-white corolla tubes, while Z. mioga, Z. oligophyllum, and Z. pleiostachyum have yellow corolla tubes.

Labellum
The labella can be separated according to color into purple and yellow. Zingiber chengii and Z. shuanglongense have obovate to widely ovate, bluish-purple labella with creamy-white patches on the basal parts. In contrast, Z. pleiostachyum has an obovate to oblong, reddish-purple labellum with yellow patches on the basal parts (Fig. 4). For the yellow type, Z. mioga and Z. oligophyllum have ovate to widely ovate, yellowish labella without patches.

Lateral staminodes
The morphology of lateral staminodes is similar to the labella; however, the dimensions of the adnate part could be of some taxonomic significance. For Z. pleiostachyum, the adnate part is ½ to ⅔ of the size of the staminodes, while all others often have adnate parts that are only ⅓ to ¼ the size of the staminodes.

Anther crest
Stamen morphology also provides taxonomic value for the Zingiber of Taiwan

Distribution pattern of taxa of Zingiber of Taiwan
Zingiber of Taiwan are often found in shady areas with high humidity, such as beneath the forest canopy, at forest margins, and in deep valleys. In terms of vertical distribution, the highest altitude habitat was found for Z. shuanglongense, which grows at heights up to 1600 m in the south (Fig. 5). Zingiber pleiostachyum is the most widely distributed species in Taiwan, found across the island and Lanyu Island (Fig. 6); in contrast, Z. chengii and Z. mioga are restricted to a very narrow area in Hsinchu and Ilan counties, respectively (Figs 5,7). Zingiber shuanglongense and Z. oligophyllum are found in the central to southern parts of Taiwan (Fig. 7).

Haplotype network and genetic distance of Zingiber koshunense, Z. pleiostachyum, and Z. shuanglongense
In order to clarify the relationship of taxa with similar morphology, i.e., Z. koshunense, Z. pleiostachyum, and Z. shuanglongense, we reconstructed a haplotype network of these taxa. Two groups emerged from the haplotype network analysis: Z. pleiostachyum and Z. shuanglongense, with nine and four haplotypes, respectively. No shared haplotypes were found between the two groups. The haplotype of a specimen of Z. koshunense from Lanyu Island was identical to that of Z. pleiostachyum (Fig. 8).
The genetic distance of Z. pleiostachyum vs Z. shuanglongense and Z. koshunense vs Z. shuanglongense was 0.0113, and that of Z. koshunense vs Z. pleiostachyum was 0.0005. These results revealed that Z. koshunense was more close to Z. pleiostachyum than to Z. shuanglongense, which was similar to the pattern of haplotype network analysis (Table 2).

Phenology
Flowering between May and July, and fruiting between July and September.

Distribution
Endemic to Taiwan.

Conservation status
This species was evaluated as Endangered (EN) by Wang et al. (2020).

Phenology
Flowering from July to September. Fruiting from September to October.

Distribution
Distributed in China, Japan, and Korea. In Taiwan, this species was naturalized in Ilan County (see Discussion).

Conservation Status
This species is evaluated as data-deficient (DD) here, due to its uncertain taxonomic status.

Phenology
Flowering from July to September and fruiting from August to December.

Distribution
In Taiwan, Z. oligophyllum grows in shaded, moist undergrowth of disturbed secondary forest (orchards and bamboo mixed forest), at elevation 300-800 m. Recently, this species was also recorded from China (Bai et al. 2016).

Conservation status
This species was evaluated as Endangered (EN) by the Editorial Committee of the Red List of Taiwan Plants (2017).

Phenology
Flowering from August to November and fruiting from October to March.

Distribution
Endemic to Taiwan.

Conservation status
This species was evaluated as Least Concerned (LC) by the Editorial Committee of the Red List of Taiwan Plants (2017)

Phenology
Flowering from August to November and fruiting from October to February.

Distribution
Endemic to Taiwan.

Conservation status
This species was evaluated as Least Concerned (LC) by the Editorial Committee of the Red List of Taiwan Plants (2017).

Notes
According to ICN art. 23.5 (Turland et al. 2018), the adjective form of the species epithet must agree with the gender of the generic name, and must be corrected when not conforming to this article. Therefore, the gender of the original spelling of ʻshuanglongensisʼ by Yeh et al. (2013) did not conform to the neuter gender of ʻZingiberʼ, thus needed to be corrected to ʻshuanglongenseʼ. Bai et al. (2015b) also made the same correction for this specific epithet.

Morphological comparison of taxa of Zingiber from Taiwan
The floral parts are often regarded as the most important diagnostic characteristic for the delimitation of species of Zingiber Yeh et al. 2012Yeh et al. , 2013Bai et al. 2016;Ohba 2016); however, flowers of Zingiber wither on herbarium specimens and fine distinctions become difficult to interpret. To add to the problem of species delimitation and identification, the flowers of Taiwanese taxa of Zingiber often hide under leaves near the ground. This fact together with the plant's preference for shady habitats and its diurnal habit makes field work difficult.
In the second edition of Flora of Taiwan (Wang 2000), all species with bilobed ligules and purplish labella were lumped into Z. kawagoii. Our study revealed that at least three species from this group could be distinguished by the corolla tube color and labellum morphology. Zingiber chengii differs from Z. shuanglongense and Z. pleiostachyum in its deciduous habit, lanceolate leaves and inflorescence with loosely arranged flowers, together with a creamy-white corolla tube and violet labellum with creamywhite patches. In addition, Z. shuanglongense resembles Z. pleiostachyum in the pseudostem and leaf morphology, and their distribution range overlaps in central and southern areas. However, the creamywhite corolla tubes and violet labellum with creamy-white patches of Z. shuanglongense could easily be distinguished from the yellow corolla tubes and purplish labellum with yellow patches of Z. pleiostachyum, even in sympatric populations.
Zingiber pleiostachyum was described by Schumann (1904). This species was characterized by its narrow bracts and by the flowers that have pubescent pedicels. He cited two gatherings (Henry 147 and 1659; both in K) collected from the Bankensing Mountains. Both Schumann's description of these specimens and the label (Henry 1659) describe this species as having red flowers similar to those of Z. kawagoii. We compared the protologue and references of Taiwanese taxa, and confirm that this species is identical to Z. kawagoii. According to ICN art. 11.3 (Turland et al. 2018), Z. pleiostachyum has priority over Z. kawagoii; therefore, we treat Z. kawagoii as a synonym of Z. pleiostachyum.
Zingiber oligophyllum was historically misidentified as Z. mioga (Wu & Chen 1981;Ohba 2016). Both species show some similarities in terms of deciduous habit and yellow corolla tubes and labellum. However, Z. oligophyllum can be distinguished from Z. mioga by having an entire rather than bilobed ligule, plicate lateral veins instead of smooth lateral veins, and widely ovate bracts and bracteoles (vs lanceolate). These differences between the two species were supported by Yeh et al. (2013) and Bai et al. (2016). Considering the differences between them, Z. oligophyllum is treated as an independent species rather than a synonym of Z. mioga in this study.

Taxonomic status of Zingiber mioga in Taiwan
Zingiber mioga was recorded as a cultivated species by Matsumura & Hayata (1906). They stated that the specimen was collected at Pachina in 1896 by Makino. The supposed provenance of this specimen suggests that this species might have been introduced before the twentieth century. Later studies by Kawakami (1910), Sasaki (1928) and Masamune (1936)  ). We also found fertile seeds and seedlings in this population, demonstrating good regeneration. The habitat of this population was similar to that of other species of Zingiber of Taiwan.
Zingiber mioga was cultivated in China for its medicinal rhizomes and in Japan for its edible young inflorescences Ohba 2016). Recently, this species was also cultivated for edible inflorescences near Fengchihu, southern Taiwan. Therefore, it was not surprising that the population of Z. mioga introduced in the northeastern section had naturalized into the surrounding areas. In fact, Z. mioga had been reported as a naturalized plant in Korea (Kim & Ou 2010;Ikeda et al. 2021). However, considering the habitat and its global distribution, Z. mioga is also possibly a native species of Taiwan. From these observations and revisions to the literature, Z. mioga might be a naturalized species of Zingiber in Taiwan.

Typification of Zingiber kawagoii and Z. pleiostachyum
Zingiber kawagoii was described by Hayata (1921) based on the specimen collected by Kawagoe from Funkiko (Kawagoe s.n.; TI, TAIF). The specimen cited by Hayata would likely have been deposited in TI or TAIF, or other herbaria in Japan or Taiwan, but it could not be found in these herbaria. Therefore, the type specimen of Z. kawagoii was considered to be lost and a lectotype or neotype had to be designated (ICN Art. 9.3) (Turland et al. 2018). However, Hayata (1921) did not cite the other specimen in the protologue of Z. kawagoii; therefore, the other original material studied by the author were candidates for the lectotype specimen (ICN Art. 9.4) (Turland et al. 2018). We found other specimens in TI and TAIF identified as Z. kawagoii and also collected by Kawagoe, but the locality of these specimens was the Urai instead of the type locality, Funkiko. One specimen was identified as being collected by Hayata based on his handwriting, and we believe this specimen could be regarded as part of the original material collected and examined by Hayata (ICN Art. 9.4) (Turland et al. 2018). Therefore, this specimen is here designated as the lectotype of Z. kawagoii (Fig. 14). Schumann (1904) cited Henry 147 (K) and 1659 (K) in the protologue of Z. pleiostachyum, however, none of them was designated as the type specimen. Zingiber kawagoii is treated as a synonym of Z. pleiostachyum here, but according to ICN Art. 9.3 (Turland et al. 2018), a lectotype specimen should still be designated. Among those cited by Schumann, Henry 1659 was an intact, related specimen and thus, is here designated as the lectotype (Fig. 15).

Taxonomic status of Zingiber koshunense
This name was first published by Sasaki (1930) in the herbarium list as Zingiber koshunense [sic] Hayata, based on a collection of T. Kawakami and S. Sasaki at an altitude of 5000 feet of Abei Line, 2 Jan. 1911. This collection was the first one of this species. Moo (1978) described Z. koshunense as a new species. He provided detailed descriptions and two specimens were cited, one which was collected by T. Kawakami and S. Sasaki, and the other one from Lanyu, which was collected by him (Moo 2340;TAI). Neither of them was designated as the type specimen. Therefore, Moo's name was still invalid because no single specimen or illustration was designated as the type specimen as required by ICN articles 8.1 and 9.1 (Turland et al. 2018).  designated a holotype (which actually was a lectotype) for this species, and ascribed its name to Moo due to his providing the morphological description of this species, and Z. koshunense Moo has been valid since then.
However, Z. koshunense was still a poorly known species in Taiwan. Because of the lack of the floral parts, the morphology could only be judged by Moo's description (1978). Fortunately, Moo (1978) provided a detailed description of the labellum of Z. koshunense, reddish-purple and lanceolate, which was identical to that of Z. pleiostachyum. As we have stated before, the labellum characters are crucial for settling Zingiber taxonomy, and based on this description, we treated Z. koshunense as a synonym of Z. pleiostachyum. The result of the haplotype network of cpDNA also supported such treatment of Z. koshunense (Fig. 8).