Allium yingshanense D.Q.Huang, A.G.Zhen & X.X.Zhu, 2021

Huang, De-Qing, Zhen, Ai-Guo & Zhu, Xin-Xin, 2021, Allium yingshanense, a new species from the Dabie Mountains (east-central China), and taxonomic remarks on the related species, Phytotaxa 498 (4), pp. 227-241 : 230-237

Allium yingshanense D.Q.Huang, A.G.Zhen & X.X.Zhu , sp. nov. ( Figs. 1–3 )

Type: — CHINA. Hubei Province, Yingshan County, Longtan Valley, Dabie Mountains Area , moist rocky mountain slopes, 31°04′59.31”N, 115°49′03.16”E, elev. ca. 320 m, 7 Oct., 2017, ZhuXX059 (holotype KUN; isotypes KUN, CSH) ( Fig. 4 ) .

Allium yingshanense is morphologically mostly similar to A. longistylum and, at lesser extent, to A. flavovirens . It is mainly distinguished from them by its inner and outer filaments 1-toothed on each side, respectively; white tepals with yellowish-green to green midvein; boat-shaped outer tepals; oblong inner tepals usually with irregularly remotely denticulate along the margins; yellow or yellowbrown anther. Also, it flowers later (mid September) and grows at relatively low elevation (below 1,000 m).

Description: —Bulbs densely clustered, cylindric, slightly thickened and oblique at base, 3.5–6.5 mm in diam.; outer tunics scarious to subleathery, entire, blackish brown to yellowish brown, and inner ones membranous, whitish. Leaves three to five, conspicuously longer than scape; leaf blades semiterete-flat, 16–23 cm × 1–2.6 mm, solid, clearly 3- angled near the base, 1-keeled abaxially and channeled adaxially, smooth or scabrous-denticulate along angles. Scape lateral, 5–25 cm × 0.5–0.9 mm, subterete, narrowly 2-winged, solid, covered with leaf sheaths for 1/4–1/2 of its length. Spathe bivalved, membranous, persistent; valves unequal, the larger 3–5 nerved, with broadly-triangular base and slightly flat beak 1.2–1.5 mm long, the smaller 1 nerved, with long-triangular base and slender beak 0.8–1 mm long. Umbel laxly hemispheric, 4–29 flowered. Pedicels subequal, terete, 1–1.4 cm long, 2.5–3 × as long as perianth, bracteolate. Perianth stellately spreading, white; segments with yellowish-green midvein; inner tepals longer and wider than outer ones, oblong to oblong-elliptic, margin usually irregularly sparsely denticulate, obtuse or rounded at apex, sometimes irregularly denticulate, 4.5–5.5 mm × 2.3–2.8 mm; outer tepals oblong-ovate to ovate, clearly boat-shaped, abaxially keeled, acuminate at apex, slightly tinged with pale pink abaxially, 3.5–3.9 mm × 1.6–1.9 mm; base connate into an annulus ca. 0.5 mm high. Filaments equal, white, 5–6 mm long, 1.5 × as long as perianth segments, connate at base for ca. 0.5 mm and adnate to perianth segments for ca. 1 mm; outer filaments, sometimes deflexed after anthesis, basally broadened, ca. 2.5 mm wide, 1-toothed on each side, teeth 1–1.5 mm long, entire at apex; inner ones broadened at base ca. 3.5 mm wide, slightly wider than outer ones, with 1 obtuse tooth on each side, teeth 1.5–2 mm long, apex of tooth usually forming 1–2 irregularly obtuse or minute faint blunt denticules, rarely entire. Anthers elliptical, yellow or yellow-brown, 1.6–1.7 mm long, rounded at the apex. Ovary trivalved, obovoid, smooth, greenish, base constricted into a short stipe ca. 1 mm long, with hood-like appendages at base, 2.3–2.5 mm, ovules two per locule. Style slightly shorter than to equal to ovary; stigma punctiform. Capsules obovoid ellipsoid, 5 mm × 6 mm. Seeds ovoid-flattened, 1.6–1.9 mm × 1.2–1.5 mm, black, testa with densely granulous periclinal walls.

Chromosome number: —2n = 2x = 16.

Phenology: —Flowering from September to October, fruiting from October to November.

Etymology: —The specific epithet derives from Yingshan County (Hubei province), where the type material was firstly found and collected. The Chinese vernacular name for Allium yingshanense is “崖‡ (ái jiǔ, an ancient pronunciation in Chinese phonetic transcription)” by the local people of the Dabie Mountains of Hubei Province.

Distribution and habitat:— Allium yingshanense is currently known from the type locality in Yingshan County and its neighboring areas at the junction of Hubei, Henan and Anhui provinces. Geographically, A. yingshanense is widely isolated from its closest relatives, namely A. longistylum and A. flavovirens , which are also isolated from each other and distributed mainly in Taihang, Yanshan, and Daqing mountainous area, and Helan Mountains, respectively ( Fig. 6 ). This new species grows in aggregates on moist rock slopes along river banks in forests or rocky cliffs at the relatively lower elevation (below 1,000 m).

During our survey, we also learned that A. yingshanense has been consumed intentionally as a delicious wild leek by the local mountain residents, and only sporadically cultivated in the Dabie mountainous region. They believe that wild edible herbs have high nutritional value and medical or health function. Considering its narrow native range size and habitat range and small wild population size, effective conservation effort is required, such as in situ and ex situ cultivation, to relieve overexploitation pressure on its wild populations.

Systematic relationships: —Our ITS phylogenetic tree presents a topology similar to those of Friesen et al. (2006) and Li et al. (2010), viz. subgenera Allium , Cepa, Reticulatobulbosa and Polyprason are not monophyletic and their relationships are beyond the discriminatory power of the ITS region ( Fig. 5 ). Nevertheless, the phylogenetic analysis shows that the new species A. yingshanense is sister to A. flavovirens / A. longistylum in a clade well supported by posterior probabilities but weakly supported by bootstrap values (PP/BS/ML bootstrap 97/<50/88 %, Fig. 5 ) (hereafter referred to as Flavovirens clade), supporting the placement of the new species into the subgenus Cepa as delimited by Friesen et al. (2006) and Li et al. (2010).

Albeit A. yingshanense is distinct from both A. flavovirens and A. longistylum , they share several morphological characters, such as cylindric bulbs, channeled or V(U)-shaped semiterete-flat leaves, hemispheric to subglobose umbel inflorescence, confirming their close relationship. Moreover, the seed testa sculpture of A. yingshanense has densely granulous periclinal walls, which are similar to those of A. flavovirens and A. longistylum ( He 1999) , also in agreement with those belonging to A. sect. Sacculiferum Gritzenko (1979: 22) ( Choi et al. 2012) and as shown by most of the species of the subgenus Cepa ( Lin & Tan 2017, Baasanmunkh et al. 2020). Seed testa epidermal cells characters support thus its inclusion within the subgenus Cepa . Furthermore, our data show that A. yingshanense is diploid, and its karyotype is mostly characterized by nearly metacentric chromosomes ( Fig. 2A ), being similar to that of A. flavovirens , which also has 2n = 16 chromosomes, 12 of which are metacentric, 2 submetacentric and 2 subtelocentric (2 satellited chromosomes) (fig. 1A in Zhou et al. 2007).

Taxonomic remarks on A. longistylum and A. flavovirens : — Allium longistylum is endemic to north China (Hebei, Beijing, Nei Mongol, and Shanxi) according to herbarium records and Flora of China ( Xu & Kamelin 2000), but Choi et al. (2003) and Choi & Oh (2011) also reported it from the central part of Korea, and they argued that it is morphologically similar to A. thunbergii var. teretifolium Choi & Oh (2004: 79) . Based on molecular data, Choi et al. (2012) further suggested that A. thunbergii Don (1827: 84) (= A. sacculiferum Maximowicz 1859: 281 ) ( Baasanmunkh et al. 2018) is the closest relative of the Korean specimen attributed to A. longistylum . Our ITS phylogenetic tree, however, clearly indicates that A. longistylum is most closely related to A. flavovirens and A. yingshanense rather than to the species of A. sect. Sacculiferum ( Fig. 5 ).

Indeed, A. longistylum is often confused with species of A. sect. Sacculiferum , as mentioned by Wang & Tang (1937) “ A. longistylum has long been considered as a synonym of A. sacculiferum , but so far as the shape of its bulb and some other characters (e.g. leaf, filament, and tepal) are concerned, it seems hard to be conspecific with the latter”. In fact, A. longistylum is morphologically quite different from that Korean specimen in bulb (cylindric vs cylindrically ovoid), bulb tunics (scarious to subleathery/red-brown/lustrous vs papery/brown/dull), leaf blade (semiterete/adaxially channeled vs terete/non-channeled), outer tepals (oblong with slightly boat-shaped abaxially vs elliptical), inner tepals (ovate vs ovately elliptical), basal part of inner filaments (entire vs entire or with two small teeth), and anther color (deep purple vs yellowish) ( Choi & Oh 2011). Thus, we must conclude that the Korean specimen previously identified as A. longistylum belongs to a different species, probably included in the section Sacculiferum .

Another taxonomic consideration on A. longistylum is its placement into A. sect. Condensatum Friesen (2006: 390) by Li et al. (2010) on the basis of the classification of Xu & Kamelin (2000). However, according to our phylogenetic analysis ( Fig. 5 ), ITS sequence submitted by Li et al. (2010) as A. condensatum from Weixian county, Hebei province, China, (GenBank code HQ690291 ) clusters with A. longistylum sequences and is distant from all the other A. condensatum accessions. Actually, it is not difficult to discriminate A. condensatum from A. longistylum based on its bulb solitary (or paired) habit, bulb-coat usually breaking up into strips, fistulose and thread-like leaves, and pale yellow to white tepals with greenish midvein ( Xu & Kamelin 2000, Friesen et al. 2006). Moreover, the seed testa of A. condensatum has verrucose sculptures, similar to that of species within A. sect. Codonoprasum Reichenbach (1827: 538) ( Choi et al. 2012). From the foregoing evidences, A. longistylum cannot be considered as a member of A. sect. Condensatum .

Li et al. (2010) tentatively suggested a new monotypic section Flavovirens Li & He (2010: 732) typified by A. flavovirens under the subgenus Cepa , but also noted that the circumscription of the new section seemed premature, being based solely on molecular data. Our present phylogenetic analysis further reveals that the Flavovirens clade occupies a separate position next to the section Sacculiferum . Indeed, the Flavovirens clade is remarkably similar to the section Sacculiferum in gross morphology and phenology (with a late flowering period from August to November), but the latter can be distinguished from the former by ovoid-cylindric, globose to ovate bulbs, (sub-)fistulose leaves, and rose or violet-colored flowers. Also, the distribution range of the Flavovirens clade is mainly restricted to a few mountain locations from northwest to east-central China, whereas section Sacculiferum is widely distributed in Russia, China, Korea and Japan. Thus, it seems reasonable to formally recognize the Flavovirens clade as a section, viz. accepting the section Flavovirens proposed by Li et al. (2010) and including A. longistylum / A. yingshanense within it. However, considering that there is a substantial incongruence between taxonomy and molecular phylogeny of the subgenus Cepa , we advocate retaining the Flavovirens clade rather than adopting it as a formal section until a taxonomic reassessment and a comprehensive revision of this subgenus are performed using an integrative taxonomic approach.

Additional specimens examined (paratypes):— CHINA. Hubei Province, Yingshan County, Longtan Valley, Dabie Mountains Area , moist rocky slopes, 31°04.986’N, 115°49.051’E, elev. 326 m, 1 Oct., 2018, H2018100106 , D.Q Huang ( IBK, GLMC); ibidem, streamsides of the valley, 31°04′58.90″N, 115°49′03.16″E, elev. 323 m, 14 Nov., 2020, ZXX202619, X.X Zhu, J Wang & M. H Wang ( KUN, HIB) ; Henan Province, Shangcheng County, Jingangtai, Dabie Mountains Area , stone slope under the forest, 31°44′02.86″N, 115°30′12.85″E, elev. 656 m, 20 Sept., 2020, ZXX202034, X.X Zhu, J Wang & M. H Wang ( KUN, HIB) ; Anhui Province, Jinzhai County, Tiantangzhai town, Dabie Mountains Area , cliff, 31°09′01.56″N, 115°46′14.90″E, elev. 947 m, 25 Sept., 2020, ZXX202098, X.X Zhu, J Wang & M. H Wang ( KUN) .

Figures

FIGURE 1. Habitat and habitus of Allium yingshanense. A–B, natural habitat on moist rock slopes (from the type locality in Yingshan county of Hubei Province). C, general view of the plant (from Jingangtai of Henan Province).

FIGURE 2. Morphology of Allium yingshanense. A, growth habit at flowering stage. B, inflorescence. C, fruiting stage. D, single flower (teeth at the base of the inner and outer filaments indicated by the arrows). E, leaves (shape of the leaf in cross-section indicated by the arrow). F, bulbs (showing clustered growth habit).

FIGURE 3. Details of Allium yingshanense from type locality. A, single flower. B, opened perianth (teeth at the base of outer filaments indicated by the arrow). C, teeth of the inner filament. D, teeth of the outer filament. E, pistil (showing basal hood-like projections). F, spathe. G, seed showing densely granulated testa (above: ventral face, below: dorsal face). H, capsule. I, cross-section of the scape at the base. J, cross-section of the scape near the top). K, cross-section of the leaf at the base. L, cross-section of the leaf at the mid-length. M, chromosome plate from type locality (2n = 2x = 16; satellited chromosome indicated by the arrow).

FIGURE 4. Holotype of Allium yingshanense.

FIGURE 5. Phylogenetic relationship derived from Bayesian analysis of ITS sequences for the studied taxa of Allium subg. Cepa given in bold italics. Values along branches or nodes represent Bayesian posterior probabilities and bootstrap values from BI/MP/ML analyses. Asterisks and hyphens denote 100% support and no support value, respectively. Numbers following specific names represent population or individual codes. Classification arrangement according to Friesen et al. (2006) and Li et al. (2010).

FIGURE 6. Geographic distribution of Allium yingshanense, A. longistylum, and A. flavovirens.