喜山葶苈 xi shan ting li

Draba algida Adams var. brachycarpa Bunge; D. alpicola Klotzsch; D. alpina Linnaeus var. rigida Franchet; D. kizylarti (Korshinsky) N. Busch; D. oreades prol. alpicola (Klotzsch) O. E. Schulz; D. oreades prol. chinensis O. E. Schulz; D. oreades var. ciliolata O. E. Schulz; D. oreades var. commutata (Regel) O. E. Schulz; D. oreades var. dasycarpa O. E. Schulz; D. oreades var. depauperata O. E. Schulz; D. oreades var. estylosa O. E. Schulz; D. oreades prol. exigua O. E. Schulz; D. oreades var. glabrescens O. E. Schulz; D. oreades var. occulata O. E. Schulz; D. oreades prol. pikei O. E. Schulz; D. oreades var. pulvinata O. E. Schulz; D. oreades var. racemosa O. E. Schulz; D. oreades var. tafelii O. E. Schulz; D. pilosa Adams var. commutata Regel; D. pilosa var. oreades (Schrenk) Regel; D. qinghaiensis L. L. Lou; D. rockii O. E. Schulz; D. tianschanica Pohle; Pseudobraya kizylarti Korshinsky.

Herbs perennial, (0.5-)1.5-14(-20) cm tall, cespitose, scapose. Caudex simple to many branched, ultimate branches terminated in rosettes and covered with petiolar remains of previous years. Stems erect, simple, often densely pubescent with a mixture of simple and subsessile forked trichomes, sometimes subhirsute with almost exclusively simple trichomes, rarely glabrous. Basal leaves rosulate, persistent; petiole absent or short, rarely to 2 cm, persistent, often ciliate with simple and/or long-stalked forked trichomes; leaf blade suborbicular, obovate, spatulate, oblanceolate, or lanceolate, (0.3-)0.5-2(-3) cm × (1-)2-6(-8) mm, sparsely or densely pubescent with simple trichomes, these often mixed abaxially with stalked forked and subsessile, 3- or 4-rayed stellate ones with unbranched rays, adaxially with predominantly simple trichomes, rarely glabrous except for margins, base cuneate to attenuate, margin entire or rarely 1- or 2-toothed on each side, apex obtuse to acute. Cauline leaves absent, sometimes 1, very rarely 2, sessile, similar to basal. Racemes (2-)4-15(-25)-flowered, ebracteate, subumbellate and not elongated or rarely subracemose and slightly elongated in fruit. Fruiting pedicels 1-7(-10) mm, divaricate or divaricate-ascending, straight or rarely slightly curved, tomentose or pilose abaxially, glabrous adaxially or rarely throughout. Sepals oblong or ovate, 1.5-2.5(-3) × 0.8-1.5(-1.8) mm, erect, caducous or rarely persistent, abaxially sparsely pilose or glabrous, base of lateral pair not saccate, margin narrowly membranous. Petals yellow, obovate to narrowly spatulate, 2.5-5(-6) × (0.9-)1.5-2.5(-3) mm, apex emarginate or rounded; claw absent, rarely 1(-2) mm. Filaments 1.5-2.5(-3) mm; anthers ovate, 0.2-0.4(-0.6) mm. Ovules (4-)6-12 per ovary. Fruit ovate to suborbicular, rarely ovate-lanceolate, (3-)4-9(-12) × 1.5-4.5(-6) mm, latiseptate and basally inflated, not twisted; valves glabrous or rarely puberulent with simple or forked trichomes, not veined, base obtuse, apex acute to subacuminate; style (0.1-)0.3-0.8(-1) mm. Seeds black to dark brown, ovate, (0.7-)1-1.5 × 0.5-0.9(-1) mm. Fl. and fr. Jun-Aug. 2n = 40*.

Rock crevices, moraine, scree, alpine meadows and tundra, glacier margins, hillsides, grassy slopes, swampy meadows, muddy gravel, rocky outcrops, cliffs; 2300-5500 m. Gansu, Nei Mongol, Qinghai, Shaanxi, Sichuan, Xinjiang, Xizang, Yunnan [Bhutan, India, Kashmir, Kazakhstan, Kyrgyzstan, Mongolia, Pakistan, Russia, Sikkim, Tajikistan].

Draba oreades is probably the most variable of all Asian species in the genus. A critical examination of the type collections of all except one of the taxa synonymized above (D. tianschanica) clearly reveals the following facts: firstly, all of the 14 infraspecific taxa recognized by Schulz are based on quantitative differences of highly variable characters (e.g., density of indumentum, scape length, petal and fruit length, and presence vs. absence of fruit trichomes); secondly, the infraspecific taxa recognized by Schulz occur sporadically throughout most of the range of the species, and none has a distinct geographic distribution; thirdly, Schulz repeatedly recognized two or more varieties from a single collection and often from a single sheet; and fourthly, some of Schulz's varieties (see below) are based on erroneously interpreted characters. Therefore, none of Schulz's varieties merits recognition, a conclusion already reached by several workers who wrote accounts of Draba for countries bordering China.

In the protologue of Draba rockii, Schulz (Notizbl. Bot. Gart. Berlin-Dahlem 10: 555. 1929) compared his new species with D. involucrata, and he was correct in pointing out the differences between the two taxa. However, had the type collection of D. rockii been compared with material of D. oreades, no differences would have been found. Indeed, D. rockii is indistinguishable from any collection that Schulz in his monograph cited under D. oreades var. oreades or D. oreades prol. chinensis. The records in FRPS, Fl. Qinghai. (1: 442. 1997), and Fl. Xinjiang. (2(2): 117. 1995) of D. rockii from Qinghai and Xinjiang are likely based on misidentified plants of D. korshinskyi.

The allegedly slightly longer scapes and slightly larger flowers and fruits that Schulz used to distinguish Draba oreades var. oreades from prol. chinensis are unrealistic. In fact, there is a considerable overlap in every character, and the type collection of prol. chinensis can easily be accommodated under var. oreades.

Schulz indicated that both Draba oreades var. ciliolata and var. dasycarpa are distinguished from the rest of the species by having pubescent instead of glabrous fruits. However, the holotype specimen of var. ciliolata has four plants, of which three have pubescent fruits and one has glabrous ones, whereas the holotype of var. dasycarpa has four plants in flower/fruit, of which one has pilose fruits and three have glabrous ones. These facts clearly show the artificiality of Schulz's varieties.

Plants that are completely glabrous, which correspond very well with the holotype of D. qinghaiensis, occur sporadically in China and in bordering countries. They are indistinguishable from typical plants of D. oreades in other characters and, therefore, the two taxa are here considered to be conspecific.

Draba alpina Linnaeus was recorded from China in several treatments (e.g., FRPS, p. 135; Fl. Qinghai. 1: 439. 1995; Fl. Xinjiang. 2(2): 114. 1995; Fl. Xizang. 2: 348. 1985; Fl. Yunnan. 6: 39. 1995), but these records are based on misidentified plants of D. oreades. True D. alpina is distributed in arctic and subarctic Alaska, N Canada, Europe, Mongolia, and Russia (Far East and Siberia), and the present authors have examined no Chinese material of it. Draba alpina is easily distinguished from D. oreades by having fruits typically oblong and flattened, instead of ovate and basally inflated.