BEKBERIDACEAE. — BERBERIS 361 



ovato-lanceolata v. lanceolata, 3:1 cm. ad 8.5:2 cm. v. 7:2.5 cm. 

 magna apice acuta, spinosa, basi acuta sensim in petiolum 3-6 mm. 

 longum attenuata, margine spinoso-dentata, spinis pungentibus 

 1-2 mm. longis utrinque circiter 8-20 instructa, supra viridia, nitida, 

 hypodermate instructa; praetcr costam immersam rete nervorum non 

 V. vix visibili, subtus pallidiora; praeter costam tantaum nervis- late- 

 ralibus paullo elevatis et visibilibus; flores ad 15-fasciculati, satis 

 parvi (circiter 6-7 mm. diam.), flavi, extus rubescentes, pedicelli 

 0.8-1.5 cm. longi; prophylla lanceolata; sepala interna late ovata, 

 petalis obovatis apice emarginatis, basi breviter unguiculatis nec- 

 tariis elongatis instructis majora; ovaria ovulis 1-2 subsessilibus in- 

 structa. Fructus vix maturi elliptici, nigri, pruinosi, ad 8:4-5 mm. 

 magni, stylis brevibus exclusis; semina plerumque 1. 



Western Hupeh: north and south of Ichang, alt. 1000-1300 m., 

 May and October 1907 (No. 417, type). Hupeh: Changyang, May 

 1900 (Veitch Exped. No. 535 in Arnold Arboretum Herb.; shrub, 

 1-2 m. tall). Yunnan: I-men district, A. Henry (No. 10618; shrub 

 2.5 m., flowers; seems to be a form with more lanceolate leaves, 

 4:0.8 cm.-9: 1.2 cm.). Shensi: Tai-pei-shan, 1910, W. Purdom 

 (No. 7; fruit unripe, but apparently pruinose, possibly a distinct 

 species). 



Berheris JuUanae much resembles in the leaves B. Wallichiana De Candolle, 

 of which I have seen the type specimens of Wallich. These have somewhat 

 thinner leaves, the hypoderm of which under the upper epidermis is not so well 

 developed and resembles more the epidermis, so that it may be often overlooked. 

 The leaves of B. JuUanae are furnished with a similar but rather distinct hypo- 

 derm. Besides this the ovules of the true B. Wallichiana possess a long funiculus, 

 but it is possible that this character may not be constant. 



The type of B. Wallichiana was found in eastern Nepal, and I have seen it 

 also from Shillong in the Khasi Hills in Assam, but I have not seen any Chinese 

 specimens which represent the true B. Wallichiana. 



B. Wallichiana, var. microcarpa Hooker f. & Thomson, which comes from the 

 same parts of Assam, also has stalked ovules, but I have not yet found a hypoderm- 

 like layer beneath the epidermis of the leaves, which are lanceolate, thinner, 

 somewhat shorter and smaller than those of the type. 



The plant from Shensi has leaves closely resembling those of cultivated forms 

 in the Fruticetum Vilmorinianum at Les Barres, which represent the typical B. 

 So\dieana. The seeds from which these plants were raised were collected by Abb6 

 Farges in the mountains near Cheng-kou, northeastern Szech'uan. The numbers 

 of Giraldi, cited by me in Bull. Herb. Boissier, B6r. 2, VIII. 199 (1908) under B. 

 levis, may belong to this species, but this needs further investigation. In the 

 Herbarium of the Hof-Museum at Vienna the flowering branch of No. 535 is rather 

 different and may belong to B. GriffUhiana Schneider, see p. 364. The fruiting 

 branch agrees very well with that of No. 417. 



This species is named in compliment to my wife. 



