RUTACEAE. — ORIXA 135 



perhaps most closely related to E. Henryi Dode, and more especially to the 

 variety villicarpa Rehder & Wilson. This variety has turbinate fruit but the 

 beak is long and rostrate, the leaflets are glabrous above and nearly glabrous below; 

 the branches are also nearly glabrous; the inflorescence is pyramidate. 



Evodia velutina is apparently very local and is only known to us from one dis- 

 trict. It is in cultivation. 



Here may be added notes on another species of this section and on a species of 

 the section Lepta, not collected during the Arnold Arboretum Expeditions. 



Evodia Daniellii Hemsley in Jour. Linn. Soc. XXII. 104 (1886). — Dode in 

 Bull. Soc. Bot. France, LV. 706 (1908). 



Zanthoxylum Daniellii Bennett in Ann. Nat. Hist. ser. 3, X. 201, fig. 5 (1862). 

 Zanthoxylum Bretschneideri Maximowicz in Bull. Acad. Sci. St. Petersbourg, 

 s6r. 3, XXIX. 73 (1884); in Mel. Biol. XI. 655 (1884). — Hemsley in Jour. 

 Linn. Soc. XXIII. 105 (1886). 



Shantung: Lau-shan, August 1907, F. N. Meyer (No. 278); without locality, 

 September 1907, F. N. Meyer (No. 262). Chili: mountains, Peking, E. Brel- 

 schneider. Korea: Ping-yang, September 18, 1905, J. G. Jack; Seoul, September 

 24, 1905, J. G. Jack; Seoul, cultivated in Mission grounds, September 1906, 

 U. Faurie (No. 468). 



The specimens before us agree very well with Bennett's description and figure, 

 particularly the specimens from Seoul; the specimen from Ping-yang has some- 

 what broader and shorter leaflets sparingly hairy on both surfaces; that from 

 Shantung has only 5 rather large leaflets, while all the others have from 5 to 9 

 leaflets. 



There can be no doubt that Z. Bretschneideri Maximowicz belongs here; we 

 have before us a specimen collected by Bretschneider on the mountains near 

 Peking, consisting of two branches. One of the branches has all the leaves 5- 

 foliolate and agrees in every respect even in the dull seeds which are immature 

 with Maximowicz's description; the other branch has 3-7-foUolate leaves and thus 

 approaches the typical E. Daniellii. The carpels contain one fully grown and 

 one abortive seed. Sometimes, however, in E. Daniellii both ovules develop into 

 seeds and this explains the differences emphasized by Maximowicz in his original 

 description. In 1890 (in Act. Hort. Petrop.XI.pt.I. 93) Maximowicz points out 

 again the differences between the two species, but makes no mention of the differ- 

 ence in the number of seeds; the specimen from Kansuhe refers there to Z. Daniellii, 

 which belongs probably to either Evodia hupehensis Dode or E. Henryi Dode, 

 both of which occur in Shensi. 



Sect. 3. Lepta Engl. 



Evodia simplicifolia Ridley in Jour. Linn. Soc. XXXVIII. 306 (1908). — 

 Guillaumin in Lecomte, Fl. Gen. Indo-Chine, I. 633, t. 24, fig. a. (1911). 

 Yunnan: Szemao, woods, alt. 1500 m., A. Henry (No. 12092^). 



ORIXA Thunb. 



Orixa japonica Thunberg, Fl. Jap. 61 (1784). — Pritzel in Bot. 

 Jahrb. XXIX. 423 (1900). — Shirasawa, Icon. Ess. For. Jap. II. t. 

 34, fig. 18-25 (1908). — Pampanini in Nuov. Giorn. Bot. Ital. n. ser. 

 XVIII. 170 (1911). 



