308 WILSON EXPEDITION TO CHINA 



Attara at 1800 m. altitude, and named Rosa abyssinica R. BrowTi, f. glandvlosa,^ 

 differs from R. ruscinonensis Grenier only in being more prickly and more glandular, 

 and in the leaflets being less membranous and pale on the underside. These two 

 specimens we regard as representing the two extreme geographical forms of the 

 Musk Rose of our ancestors, the R. moschaia Miller. 



The Musk Roses of China and India and that of the high mountains of Persia, 

 R. moschata, var. nasturana Christ {R. Pissarti Carridre), distinguished at a glance 

 by their larger leaves, much more numerous flowers in large rounded or paniculate 

 corymbs, were unknown to botanists and to western gardens in Miller's time. In 

 the latter half of the 19th century Cr^pin and other botanists believed that the 

 range of Miller's R. moschata extended through Persia and northern India to the 

 extreme east coast of Cliina. Most of them admitted that the new forms discovered 

 differed, however, from their conception of the original Musk Rose, and gave them 

 varietal names. Our study, however, of this perplexing subject has convinced us 

 of two things: — (1) That the Musk Rose of Miller is little if at all different from 

 Rosa ruscinonensis Grenier and Deseglise apud Deseglise (in Billotia, I. 33 [1864]), 

 and from Rosa abyssinica R. Brown, and their intermediate forms: (2) that the 

 Musk Roses of India and China are distinct from those knowTi to Miller and his 

 predecessors and that they are best considered geographical segregates under 

 separate names since they display constant characters sufficient to distinguish 

 them as species. 



The Chinese Musk Roses referred by various botanists to R. moschata belong 

 to several of the undermentioned species, but without seeing the specimens it is 

 impossible to place them where they actually belong. 



The Rose figured by Miss Willmott (Gen. Rosa, 1. 33, t. [1910]) as R. moschata 

 Miller is a form of R. Brunonii Lindley. Since this last named Rose is superior 

 as a garden plant to the old Musk Rose, it has verj' generally supplanted it and it 

 is doubtful if R. moschcUa Miller is now in general cultivation. 



Rosa Rubus Leveille & Vaniot in Bull. Soc. Bot. France, LV. 55 

 (1908). — Willmott, Gen. Rosa, II. 507, t. (1914). — Leveille, Fl. 

 Kouy-Tcheou, 354 (1915). 



Rosa moschata, var. hupehensis Pampanini in Nuov. Giorn. Bot. Ital. n. ser. 

 XVII. 295 (1910). 



Western Hupeh: Patung Hsien, thickets, alt. 600-1300 m., No- 

 vember 1907 (No. 431"; scandent bush 2.5-4 m. tall, fruit dull red); 

 Hsing-shan Hsien, woodlands, alt. 1300 m., December 1907 (No. 666^; 

 bush 3-5 m. tall, fruit dark scarlet); without locality, A. Henry (Nos. 

 5550, 7007). Western Szech'uan: Chien-chi Hsien, roadside, alt. 

 1800 m., October 1910 (No. 4174; scandent bush 3-6 m. tall, fruit red) ; 

 Mt. Omei, June 1904 (Veitch Exped. No. 4880; bush 6 m. tall); with- 

 out precise locality, banks of Yangtsze River, May 1903 (Veitch Exped. 

 No. 3536; large bush, flowers fragrant). Kweichau : " route de Pin- 

 yang. May 12, 1899, L. Martin (No. 2603, type). 



1 This is apparently a manuscript name, as this specimen is enumerated by 

 Schweinfurth in Bull. Herb. Boissier, IV. app. 2, 205 (1896) under R. abyssinica 

 without any varietal designation. 



