404 * . BOTAMCAL GAZETTE [may 



Yunnan boreali-occidentalis : ad latera orientalia montium niveorum prope 

 Lichiang-fu, in dumetis apertis, alt. circ. 3500 m., Octobri 1914, C. Schneider 

 (no. 2829 et 3912; fructus maturi carnei); eodem loco et tempore (no. 281 1; 

 fructus carnei; gemmae apice distinctius rufo-lanatae) ; in angustiis montium 

 inter Sung-queh et Teng-chuan, 29 Septembris 1914, C. Schneider (no. 2905; 

 arbor ad 8 m. alta; fructus carnei; gemmae ut in no. 281 1 rufo-lanatae; 

 rhachis folio rum ad 9-jugorum apicem versus distinctius alata; folia surculorum 

 a me in eadem arbore abscissorum minora ad 12-juga foliolis tantum ad 2:0. 

 7 cm. magnis iis 5. Pratt ii non absimilibus) . 



The fruiting branch of no. 2905 agrees well with that of no. 281 1, both 

 showing the buds distinctly fulvous at the apex, and the narrow wings of the 

 rhachis. I do not know whether these two numbers represent another form 

 because I have not yet seen fully developed buds of typical S. hupehensis. 



In nos. 2829 and 3912 the buds are much more glabrous, and the rhachis 

 is almost wingless. I am at a loss how to distinguish these specimens from the 

 type of S. laxiflora Koehne collected by E. H. Wilson in western Szechuan, 

 northeast of Tachien-lu,on the Ta-p'ao-shan, July 4, 1908 (no. 3008), and there- 

 fore I propose the following variety: 



*/>* 1 Sorbus hupehensis, var. laxiflora, n. var. — S. laxiflora Koeh. 

 in Sargent, PL Wils. 1:466. 1913. 



It needs further investigation to determine how this variety may really be 

 distinguished from typical S. hupehensis. Koehxe himself says that S. laxi- 

 flora forms with S. hupehensis and S. aperta "a special group distinguished by 

 its small stipules, medium-sized leaves with 4-7 pairs of medium-sized leaflets, 

 and by a remarkably loose inflorescence." 



There is another group of species described by Koehne which I cannot 

 separate because the characters on which they are founded by the author are 

 too variable according to my own observations. I therefore propose to unite 

 them in the following manner: 



<*^ Sorbus Prattii Koeh., var. tatsienensis, n. var. — S. munda 

 Koeh. in Sargent, PI. Wils. 1:469. 1913, includ. f.a. tatsienensis 

 et f.b. subarachnoidea. — S. pogonopetala Koeh., I.e. 473. — A typo 

 nonnisi foliolis paullo maioribus saeDissime basi tantum integer- 



rimis 



type 



know how to distinguish 5. munda as a variety from S. Prattii. In describing 



type 



300 



the author has seen, came from the same locality (Pan-lan-shan, west of Kuan 

 Usien) as Wilsons no. 4323 which Koehne makes the type of his S. munda f. 

 subarachnoidea. But this fruiting specimen agrees in every respect with the 



