234 



THE 



[vol. V 



vel elliptico-oblonga, rarius elliptica, 3-5 cm. (pleraque 8-12 cm.) longa 

 et 1.5-6 cm. (pleraque 3.5-5 cm.) lata, integra vel rarius dentibus paucis 

 remotis supra medium instructa, acuminata, rarius obtusa, mucronulata, 

 basi rotundata vel truncata, interdum angustata, rarius leviter subcordata, 

 utrinque nervis 3-6 curvatis et ascendentibus subtus elevatis supra leviter 

 impressis, venulis reticulatis, supra sparse subtus dense pilis stellatis 

 flavido-cinereis obtecta; petioli 1-3.5 cm. longi, villosi. Flores numerosi, 

 in paniculam 3-9 cm. altam dispositi; pedicelli 0.5-1.5 cm. longi et rhachis 

 stellato-pubescentes ; bracteae et bracteolae membranaceae, lineares, 

 0.2-0.5 cm. longae, caducae; calyx infundibuliformis, 0.6-0.8 cm. longus, 

 5-dentatus, dentibus ovatis obtusis vel acutis, cinereo-stellato-pubescent- 

 ibus; corolla alba, calyce duplo longior, petalis patentibus obovato-spatu- 

 latis margine undulatis et saepe sparse ciliatis, extus pubescentibus; 

 columna staminalis longe exserta, 2-3.5 cm. longa, antheris flavis. 

 Fructus lignosus pyriformis, 3.5 cm. longus et 2.5 cm. latus, sulcatus, 

 apice depressus, basi acutus. 



Western China. Szechuan: Monkong Ting, descent of the Pan-lan- 

 shan, side of stream, alt. 2300 m., E. H. Wilson, no. 4395, October, 1910 

 (fruiting; type); cultivated Hort. Berckmans, Augusta, Ga., raised from 

 Wilson's no. 4395, received Herb. Arnold Arboretum May 9, 1917 (flowers) ; 

 cultivated Hort. J. C. Williams, Caerhays Castle, Cornwall, England, 

 raised from Wilson's no. 4395, flowered August, 1924. Yunnan: Mengtsze, 

 alt. 1800 m., A. Henry, no. 11510; without precise locality, G. Forrest, 

 nos. 15897, 17640. 



In Sargent's Plantae Wilsonianae this tree is identified with R. pubescens 

 Mast, from the mountains of Sikkim and Botan, but we had only fruiting 

 specimens. Now with flowering material available obvious differences are 



apparen t . 



Mast 



that species the leaves are thinner and cordate or rarely truncate at the 

 base and the flowers, which are said to be pink, have linear-spatulate 

 petals and shorter calyx teeth. The other Himalayan species R. Wal- 

 lichii Br. with which Dunn confused this Chinese plant differs markedly 

 in its thinner very sparingly stellate-pubescent leaves, truncate or rounded 

 at the base, much smaller flowers and oblong fruit. The Chinese R. 

 thyrsoidea Lindl., native of Hongkong, with which the Formosan R. for- 

 mosana Kanehira is probably conspecific has smooth, shining, lanceolate 

 to oblong, rarely elliptic leaves, narrowed or rounded at the base and quite 

 glabrous and flowers about one-third the size of our new species, and a 

 smaller fruit. In Fedde, Rep. Spec. Nov. iv. 330 (1907) 

 and Vaniot briefly describe a plant as R. Cavaleriei but the description is 



too incomplete for definite identification, indeed, it is doubtful if it belongs 

 to the genus. 



I saw only one specimen of this interesting new Iteevesia during my 

 travels in China and that was growing in a remote district in western 

 Szechuan where it was known as the "Soh-lou" tree. Both Henry and 



Mes 



