Introduction

Working within the framework of the Flora of Pan-Himalaya project, we have examined much relevant literature on Salix L. (Salicaceae) since 2010. In Plantae Wilsonianae, Schneider (1916) described 32 new Chinese taxa in Salix, including 30 species and two varieties based on E. H. Wilson's collections during the years 1907, 1908 and 1910. We examined the protologues of these taxa and all the original material that we could trace. We found that the original material of the names S. mictotricha C. K. Schneid., S. polyclona C. K. Schneid. and S. rhoophila C. K. Schneid., which Wilson collected from two neighbouring counties of Hubei Province (Xingshan County and Fang County), represent two currently recognized species: S. alfredii Goerz ex Rehder & Kobuski and S. rhoophila. This issue has been almost completely overlooked by previous taxonomists (Hao 1936; Chou & al. 1984; Fang & al. 1999). A. K. Skvortsov examined the original material of S. polyclona (Wilson 2116 and 2116^bis), and labelled Wilson 2116 (A barcode 00031199) as S. hylonoma C. K. Schneid. and Wilson 2116^bis as S. rhoophila. Unfortunately, he did not discuss this issue in his publications. The purpose of the present work is to clarify the application of the names S. alfredii, S. mictotricha, S. polyclona and S. rhoophila as well as their taxonomy.

Material and methods

We examined the protologues, relevant herbarium specimens in BJFC, E, FJFC, HIB, K, P, PE and WUK, and high-resolution digital images of specimens in A, BM, S and US (herbarium codes according to Thiers 2020+).

Taxonomic and nomenclatural treatment

1. Salix rhoophila C. K. Schneid. in Sargent, Pl. Wilson. 3(1): 54. 1916. – Lectotype (designated here): China, Hubei, Fang Hsien [Fang County], 7000–8000 ft [2300–2600 m], 16 May 1909, E. H. Wilson 2117 [♀] & 2117^a [♂] (A  barcode 01536135 [♀]!; isolectotypes: A  barcode 00404477 [♂]!, BM barcode BM000958022 p.p. [♂]!, US barcode 00105210 [♂]!).

Description — Shrubs to 3.5 m tall. Branches dull brown, glabrous; branchlets pubescent, glabrescent, or glabrous. Petiole 0.5–1.4 cm, pubescent, glabrescent or glabrous; leaf blade elliptic or elliptic-oblong, rarely obovate-elliptic, 2.9–6.8 × 1.2–3.3 cm, abaxially glaucous, pubescent when young, glabrescent, or glabrous, adaxially dull green, glabrous, base broadly cuneate to rounded, margin entire, apex acute; veins raised abaxially. Catkins flowering as leaves emerge, moderately densely flowered. Male catkin 1.1–2.9 × 0.2–0.5 cm; peduncle 0.2–0.7 cm, with (0 or)1 or 2 leaflets; rachis pilose. Female catkin 1.7–3.4 (to 6.8 cm at maturity) × 0.3–0.5 cm; peduncle 0.3–0.7 cm, with (0 or)1 or 2 leaflets; floral bract brown, obovate, glabrous or subglabrous. Male flower: adaxial gland oblong, c. ½ as long as bracts, apex retuse, abaxial gland absent; stamens 2; filaments distinct, villous proximally; anthers yellow, globose. Female flower: gland adaxial; sessile or shortly stipitate; ovary long ovoid; styles 2-cleft; stigmas short. Capsules ovoid.

Phenology — Flowering from mid-April to mid-May; fruiting in May and June.

Distribution — Hubei and Shaanxi provinces of China.

Habitat — Mountain slopes, roadsides and valleys at altitudes of 1345–2300 m.

Remarks — Schneider (1916) described Salix rhoophila and cited Wilson 2117 as the “type”, Wilson 2117^a as the “cotype”, and another collection with sterile plant fragments, Wilson 2117^bis, which had mature leaves. Three duplicates of Wilson 2117 were found: A barcode 01536135 (♀), BM barcode BM000958022 (♀) and E barcode E00301602 (♀). In addition, three duplicates of Wilson 2117^a were traced: A barcode 00404477 (♂), BM barcode BM000958022 (♂) and US barcode 00105210 (♂); and one duplicate of Wilson 2117^bis was located: A barcode 01536133 (sterile). A barcode 01536135 and A barcode 0404477 are mounted together on a single sheet, and BM barcode BM000958022 includes material of both sexes. In Plantae Wilsonianae, Schneider (1916) always cited one gathering as the “type” and another gathering with branch(es) of the opposite sex as the “cotype”; if he had material of only one sex, he always only indicated one gathering as the “type”. Because Wilson 2117 and Wilson 2117^a were collected by the same collector at one place and time, they belong to a single gathering as defined by Art. 8.2 footnote of the International Code of Nomenclature for algae, fungi, and plants (Turland & al. 2018, hereafter “Code”), even though they were given different numbers. Hence, all the specimens (duplicates) of Wilson 2117 and Wilson 2117^a are syntypes of S. rhoophila (Art. 9.6 of the Code).

Our examinations indicated that duplicates of the type gathering (Wilson 2117 and Wilson 2117^a) belong to two different taxa as currently recognized: Salix rhoophila and S. polyclona. Schneider (1916) was uncertain whether the sterile 2117^bis with mature leaves belonged to S. rhoophila as he defined it; we tentatively identify it as a hybrid of S. rhoophila. However, that specimen collected on a different date is not part of the gathering to which the syntypes belong. The duplicates of Wilson 2117 (A barcode 01536135 [♀]) and Wilson 2117^a (A barcode 00404477 [♂], BM barcode BM000958022 [pro parte], US barcode 00105210 [♂]) represent the species to which the name S. rhoophila is usually applied and best correspond to the description in the protologue. The other two duplicates of Wilson 2117 (BM barcode BM000958022 [pro parte], E barcode E00301602 [♀]) have hairy bracts that are not in agreement with the protologue of S. rhoophila (“Amenta ♀, … bracteae late ovatae, … brunneae, glabrae”) and are identified as S. polyclona. Schneider noted that S. rhoophila appeared closely related to S. hypoleuca Seemen, which better applies to the former group of sheets (A barcode 01536135 [♀], A barcode 00404477 [♂], BM barcode BM000958022 [pro parte], US barcode 00105210 [♂]). Schneider (1916) did not specify that typifications were restricted to the existing material at A, so the A, BM and US duplicates of Wilson 2117 including Wilson 2117^a (A barcode 01536135 [♀], A barcode 00404477 [♂], BM barcode BM000958022 [pro parte], US barcode 00105210 [♂]) are all candidates for the lectotype of S. rhoophila. We designate Wilson 2117 (A barcode 01536135 [♀]) as the lectotype, considering it is in good condition with Wilson's handwritten label. The isolectotypes are limited to those duplicates belonging to the same species; the duplicates identified as S. polyclona are excluded.

Representative specimens examined — China: Hubei: Fang County, 1700–2300 m, 16 May 1909, E. H. Wilson 2116[2118]^bis (A barcode 00055980 [♀]); Shennongjia, 2280 m, 20 Jun 1976, E Shennongjia Botany Expedition 10456 (HIB, PE); Shennongjia, Dayanwu, 3 May 2015, L. J. Liu 20150503-1 & 2 & 3 & 4 & 11 & 12 & 13 & 18 & 19 & 21 & 26 & 27 & 28 & 30 & 32 & 33 & 35 & 38 & 39 (BJFC). — Shaanxi: Zhenping County, Shangzhu, 1345 m, 11 Apr 2004, Y. S. Chen, Z. H. Wu, B. Li & X. S. Guo 376 (WUK).

2. Salix polyclona C. K. Schneid. in Sargent, Pl. Wilson. 3(1): 55. 1916. – Lectotype (designated here): China, Hubei, Fang Hsien [Fang County], 7000–8000 ft [2300–2600 m], 16 May 1909, E. H. Wilson 2116 (A  barcode 00031199 [♀]!; isolectotype: US barcode 00105232 [♀]!).

= Salix mictotricha C. K. Schneid. in Sargent, Pl. Wilson. 3(1): 56. 1916, syn. nov. – Lectotype (designated here): China, Hubei, Hsing-Shan Hsien [Xingshan County], 4000–5000 ft [1300–1700 m], 14 May 1907, E. H. Wilson 2118 (A  barcode 00404476 [♂]!).

= Salix alfredii Goerz ex Rehder & Kobuski in J. Arnold Arbor. 13: 403. 1932, as ‘Alfredi’, syn. nov. –Lectotype (designated here): China, Gansu, Diebu County, 11000 ft [3300 m], Jun 1925, J. F. Rock 12522 (A  barcode 00031139 [♀]!; isolectotypes: P barcode P00760894!, S No. S13-9538 [♀]!, PE barcode 00704548 [♀]!).

= Salix wuana K. S. Hao ex C. F. Fang & A. K. Skvortsov in Novon 8: 469. 1998, as “Wuiana”, syn. nov. – Holotype: China, Shaanxi, Tai Pei Shan [Taibai Mountain], 1910, W. Purdom 979 (A  barcode 00056025 [♀]!; isotype: US barcode 00503929 [♀]!).

Description — Shrubs or trees. Branches purplish brown or yellowish green, thin, glabrous, shiny; branchlets pilose when young, glabrous. Petiole 0.2–0.9 cm; leaf blade elliptic, ovate-elliptic or lanceolate, 3.4–6.4 × 1.3–2 cm, abaxially greenish or greyish blue, white villous when young, glabrescent, adaxially green, pubescent when young, glabrescent, base convex to rounded, margin entire, rarely gland-dotted, apex acute or acuminate. Catkins flowering as leaves emerge, moderately densely flowered. Male catkin 1.2–4.8 × 0.2–0.8(–1) cm; peduncle 0.2–0.3 cm, with 1 or 2 leaflets. Female catkin at maturity to 4.4 cm; peduncle 0.4–0.9 cm, with 0–3 leaflets; bracts obovate, c. 1 mm, abaxially pilose, long ciliate, adaxially subglabrous, hairs white and ferruginous. Male flower: abaxial gland small or absent, adaxial gland narrowly ovate, c. 0.4 mm, rarely divided; stamens 2; filaments distinct, c. 2 mm, puberulent on proximal half; anthers yellow or reddish, globose. Female flower: gland adaxial, 0.4–0.5 mm, shortly ligulate; ovary ovoid, c. 2.2 mm, pubescent or subglabrous, shortly stipitate; styles c. 0.3 mm; stigmas 2-lobed. Capsules subglobose, c. 3 mm, puberulent or subglabrous; stipe short (c. 0.2 mm).

Phenology — Flowering from late April to early June; fruiting in June and July.

Distribution — Gansu, Henan, Hubei, Qinghai, Shaanxi and Sichuan provinces of China.

Habitat — Mountain slopes, at altitudes of (?1300–) 1880–3300 m.

Remarks — Schneider (1916: 56) described Salix polyclona based on Wilson 2116 and Wilson 2116^bis and stated the former gathering to be the type, expressing doubt about the identity of the latter (“I do not know if No. 2116^bis really belongs to S. polyclona”). We traced two specimens of Wilson 2116 (A barcode 00031199 [♀], US barcode 00105232 [♀]) and one sheet of Wilson 2116[2118]^bis (A  barcode 00055980 [♀]). Schneider's (1916) work was based on material collected by Wilson “for the Arnold Arboretum”. However, the publication did not specify that typifications were restricted to the existing material at A, so the A and US duplicates of Wilson 2116 were syntypes under Art. 9.6 of the Code. The duplicate A barcode 00031199 is designated as the lectotype of S. polyclona because it is in far better condition, with several inflorescences, whereas the duplicate US barcode 00105232 is fragmentary and has no intact inflorescences.

As for the other cited gathering, Wilson 2116^bis (A barcode 00055980), we note some confusion as to its numbering. Wilson originally wrote “2118 Bis” in grey on the label of the sheet, which someone later changed to “2116 Bis”, scrawling “6” over “8” in black pen on the label without any note of explanation. Skvortsov noted on the sheet in 1997 that the collecting number on that sheet should be Wilson 2118^bis. However, we do not know who changed the number, when or why; Wilson himself might have made an error and later corrected it. The two numbers, Wilson 2116 and Wilson 2116[2118]^bis, were collected on the same date, 16 May 1909, but on Wilson's labels of the two specimens at A the altitude of Wilson 2116 was given as “7-8000 ft” and that of Wilson 2116[2118]^bis was given as “5-7000 ft”; therefore, they were not from the same gathering.

Skvortsov observed part of Schneider's original material in 1997 and labelled Wilson 2116 (A 00031199) as “Salix hylonoma Schneid.” and Wilson 2116[2118]^bis (A barcode 00055980) as “Salix rhoophila Schneid.” However, this was not mentioned in the subsequent Flora of China treatment (Fang & al. 1999) or other publications. By comparing the original material (Wilson 2116 and Wilson 2116[2118]^bis) of S. polyclona with all recognized taxa of the genus Salix previously described from relevant regions of China, we concluded that Schneider's type Wilson 2116 does not correspond to S. hylonoma C. K. Schneid. Salix polyclona differs from S. hylonoma by its shortly ligulate flower glands (shorter than the floral bracts), abaxially white villous young leaves, and entire leaf blade margins. By contrast, S. hylonoma has a narrowly terete flower glands (as long as the floral bracts), golden downy indumentum on the young leaves, and indistinctly serrulate leaf margins. Instead, Wilson 2116 corresponds to S. alfredii, while Wilson 2116[2118]^bis appears to be identifiable as S. rhoophila. We consider S. polyclona to be the correct name for the species formerly called S. alfredii, because it has priority over that name. In addition, S. alfredii has not previously been typified as far as we know. In the protologue, Rehder & Kobuski (1932) cited three gatherings (J. F. Rock 12147, J. F. Rock 12149 and J. F. Rock 12522) as original material for S. alfredii, but they did not designate a type from among them. We traced several duplicates of these gatherings: J. F. Rock 12147 (A barcode 00031138 [♂], P barcode P00760893 [sterile], S No. S13-9540 [sterile]), J. F. Rock 12149 (A barcode 00055783, P barcode P00760895, S No. S13-9539, US barcode 00105058 [all ♀]) and J. F. Rock 12522 (A barcode 00031139, P barcode P00760894, S No. S13-9538, PE barcode 00704548 [all ♀]). All these duplicates are syntypes, and all appear to belong to the species now called S. alfredii. The specimen J. F. Rock 12522 (A barcode 00031139) is designated as the lectotype of S. alfredii because it is fertile and has mature leaves, and there is a duplicate in a Chinese herbarium (PE).

Schneider (1916) described Salix mictotricha and indicated Wilson 2118 as “type” and Wilson 2118^a as “cotype”. These specimens belonged to a single gathering, because they were collected by the same person at one time and place. One specimen of Wilson 2118 was traced: A barcode 00404476 (♂). Four duplicates of Wilson 2118^a were found: A barcode 00055963, BM barcode BM000958023, E barcode E00301605 and US barcode 00105201 (all ♀). The single sheet at A includes fragments with inflorescences of both sexes: the main label gives the number 2118 and is marked “♂” and the barcode above it is 00404476, while elsewhere on the sheet, near a female fragment, is the annotation “♀ 2118a” and the barcode 00055963. Our examination shows that the four duplicates of Wilson 2118^a are identifiable as S. rhoophila and the single fragment of Wilson 2118 is identifiable as S. polyclona. However, Hao (1936) overlooked this issue when he published an image of the sheet at A (i.e. A barcode 00055963 [♀] + barcode 00404476 [♂]) in his monograph. The protologue of S. mictotricha (Schneider 1916: 57) says of the male catkins: “bracteis obovato-rotundis utrinque fulvo-sericeis (pilis albis intermixtis)” and adds “This species may be easily distinguished by the brownish silky bracts of the ♂ plant” and “The specific name is derived from µικτός, mixed, and θρίξ, hair.” Therefore, Wilson 2118(A barcode 00404476 [♂]) best fits Schneider's concept of S. mictotricha and is designated as the lectotype of that name. Salix mictotricha and S. polyclona, now synonyms, were described simultaneously by Schneider in the same monograph and therefore have equal priority. No choice between the two names has previously been made under Art. 11.5 of the Code. We choose to adopt the name S. polyclona because the original material for this name is in better condition.

Chou & al. (1984) treated “Salix Wuiana K. S. Hao” (Hao 1936, not validly published) as a synonym of S. alfredii before it was validated by Fang & Skvortsov (1998). However, the validly published name S. wuana K. S. Hao ex C. F. Fang & A. K. Skvortsov was not recorded in Flora of China (Fang & al. 1999). After examining the type specimens of S. wuana and S. polyclona, no relevant differences between these taxa were found. Therefore, the name S. wuana is synonymized with S. polyclona. Neither Hao (1936) nor Fang & Skvortsov (1998) specified an etymology for the epithet. However, Hao (1936: 13, 95, fig. 72) gave the epithet as “Wuiana”, proving that it was named for a person, because throughout this work he capitalized the first letter of epithets only when they were based on people's names. Wu is a very common surname in China, shared by dozens of botanists, while “Wui” is not a Han name nor indeed a syllable normally used in Mandarin Chinese. Therefore, we assume that the species honoured a person named Wu and that the original spelling of the epithet is to be treated as an error to be corrected to wuana in accordance with Art. 60.8(c) of the Code.

Representative specimens examined — China: Gansu: Diebu County, Sangbagou, 2961 m, 20 Jun 2012, L. He & X. X. Mao PH20120620-02 (BJFC); Lianhua Mountain, 3103–3234 m, 11 Jun 2014, L. He, Q. Guo & Z. X. Zhang PH20140611-12 & 18 (BJFC). — Henan: Luanchuan County, Laojun Mountain, 1880 m, 28 May 1957, Anonymous s.n. (NAS barcode NAS00281514). — Hubei: Shennongjia, Laojun Mountain, 2540 m, 6 Jun 2019, F. Y. Guo & C. L. Yang 060611 (BJFC). — Qinghai: Menyuan County, Xianmiqi, Hankegou, 2850 m, 14 Jun 1960, Chinese Academy of Sciences Qinghai-Gansu Expedition 2485 (PE). — Shaanxi: Xi'an, Guangtou Mountain, 2458 m, 17 Jul 2016, L. He & J. Zhao HL 38 (FJFC). — Sichuan: Jiuzhaigou, Changhai to Wucaichi, 2995–3085 m, 26 Jul 2017, L. He & J. Zhao HL 114-2 & 3 & 4 & 5 (FJFC); Jiuzhaigou, Yuanshisenlin, 2974 m, 26 Jul 2017, L. He & J. Zhao HL 115 (FJFC).