102 WILSON EXPEDITION TO CHINA 



Salix Pyi Leveille in Bull. Soc. Bot. France, LVI. 300 (1909). 

 Salix yunnanensis Leveille, I.e. 301 (1909). 



Salix tetrasperma Diels in Not. Bot. Gart. Edinburgh, VII. 91 (PI. Chin. 

 Forrest.) (non Roxburgh) (1912). 



CHINA. Yunnan : " environs de Yunnan-sen, bord des canaux dans la plaine," 

 February 15, 1897, E. Bodinier (No. 65; type of S. polyandra; tree; cf and 9); 

 same locality, " pagode de He-long-tan," March 19, 1905, April 11, 1906, F. Du- 

 cloux (Nos. 658, 669; 9 type of *S. Cavaleriei); same locality, " vallons de Tchong- 

 Chan," February 21, 1906, F. Ducloux (cf type of 5. Pyi); same locality, " plaines," 

 April 4, 1906, F. Ducloux (No. 653; c? type of S. yunnanensis) ; Li chiang fu, plain, 

 alt. 2500 m., lat. 26° 50', May 1906, G. Forrest (No. 2033; 9). 



By the kindness of Mgr. Leveille 1 have flowers of his different types except those 

 of S. yunnanensis, of which I have seen only one leaf. There seems to be no real 

 difference in the 9 and cf flowers of the different tj-pes. The cf flowers have from 

 8-12 stamens and a somewhat five-lobed disc. The bracts are ovate-deltoid, rather 

 acute, distinctly pubescent on the inner surface and on the margin and often gla- 

 brous on the outer surface. The 9 flowers have sessile stigmas, glabrous ^ ovaries, 

 and are distinctly stalked with pedicels from one-half to three-fifths as long as the 

 ovaries. The bracts are like those of the cf flowers, only in S. Cavaleriei Leveill6 

 they are obtuse or even truncate at the apex. There are two glands: the ventral is 

 broad and low, embracing the base of the pedicel; the dorsal is 2- or 3-lobed in 

 S. Cavaleriei Leveille and S. polyandra Leveille, but it is simple and rather large 

 in Forrest's plant which might represent a variety. 



An apparently nearlv related species is S. anisandra Leveille et Vaniot (in Fedde, 

 Rep. Spec. Nov. III. 22 [1906]. — L6veill4, Fl. Kouy-Tcheou, 381 [1915]), the type 

 of which was collected by /. Esquirol in Kweichou at Pia Fong, March 1905 

 [No. 362; cf ]. It has from 10 to 12 stamens and orbicular-ovate, obtuse and nearly 

 glabrous bracts. The two glands seem to be more distinctly separated. This spe- 

 cies needs further observation. 



19. Salix macrolepis Turczaninow in 5uZL 5oc. Nat. Mosc. XXVII. 371 (1854); 

 Fl. Baical.-Dahur. II. 98 (1856). — Andersson in Svensk. Velensk. Akad. Handl. 

 VI. 52, t. 3, fig. 33 [Monog. Salic.) (1867); in De Candolle, Prodr. XVI. pt. 2, 213 

 (1868), ut videtur pro parte. — Glehn in Beitr. Kennt. Russ. Reich. XXV. 211 

 (1868); in Act. Hort. P'etrop. IV. 81 (?) (1875). — Fr. Schmidt in Mem. Acad. Sci. 

 St. Petersbourg, s6t. 7, XII. 172 (Reis. Amur. Sachal.) (1868). — Herder in Act. 

 Hort. Petrop. XI. 400 (pro parte) (1891). — Wolf in Act. Hort. Petrop. XXI. 48 

 (1903). 



Salix bracteosa Turczaninow, PI. Exsicc, sed ut videtur non apud Trautvetter 

 in Middendorff, Reise Sibir. 1. pt. 2, Bot. abt. 2, 77 {Fl. Ochot.) (1856), 

 quoad specim. prope Udskoi lectum. 



NORTHEASTERN ASIA. Transbaikalia: "in virgultis subalpinis Da- 

 huriae," N. Turczaninow (type, ex Turczaninow); "ad flumen Selenga prope 

 Werchne-Udinsk," N. Turczaninow (ex Andersson et Wolf). Amur: Stanovoi 



as a nomen nudum, while Weigel (Fl. Pomerano-Rugica, 80 [1769]) first describes it. 

 This S. polyandra is the same as S. polyandra Bray (in Denkschr. Bot. Ges. Regensb. 

 I.pt. 2, 41, t. 1 [1818]), namely, the hybiid between S. pentandra Linnaeus and S. 

 fragilis Linnaeus. S. polyandra Gleditsch {Syst. Einl. Forslwissensch. II. 4 [1775]) 

 is only a sjmonym of S. pentandra Linnaeus. 



^ In Fedde, 1. c. (1909), the description is: " ovariis fusiformibus, pubescentibua, 

 vix pedicellatis." This must be a mistake. 



