ARNOTT ON SOUTH AFRICAN PLANTS. 257 



there can I think be no doubt, although supposing the char- 

 acter proposed by Eckloti and Zeyher really to apply to it, 

 the genus may be new, differing from Celastrus, by having 

 opposite leaves, and from Hartogia by the capsular fruit. 



18, Ilex crocea; this is Crocoxylon excelsum, E. & Z, I 

 find the ovary to be immersed in a 4-ang]ed thick fJeshy disk, 

 2-celled, with two erect ovules in each cell; style thick and 

 conical, and the stigma entire; whereas in the generic char- 

 acter proposed by Ecklon and Zeyher, the ovary is said to !)e 

 4-celled and 8-ovuled, and the stigma subquadrifid. My 

 plant does not seem to differ from Hartogia. — 2. Ilex flexuosa 

 has all the structure oi Ilex crocea, except having five petals 

 and stamina; the leaves also are alternate ; it seems to belong 

 to Mystroxyhn, E. & Z. — 3. No. 6745 ; this appears to me 

 to differ only from Celastrus rupestris, E. & Z., by the some- 

 what smaller leaves. Calyx-segments 5, rounded, membra- 

 naceous on the margin; petals orbicular, patent; disk 5- 

 angled ; filaments 5, short and broad, persistent, inserted 

 under the angles of the disk; anthers orbicular, with a broad 

 connectivum at their back, by the middle of which they are 

 attached to the filament, 2-ceIled, cells nearly parallel and 

 dehiscing vertically; ovary immersed in the disk, 3-celled, 

 with two ovules in eacli cell ; style short and thick ; stigma 

 very slightly 3-lobed. This may belong to Catha of Endli- 

 cher, but differs widely in habit. I possess another species, 

 closely allied to these, collected between Cape and Grahams- 

 towns; this is destitute of flower, but viith the valves of the 

 last year's capsule still adhering; it is probably a mere variety 

 with younger foliage. 4. Ilex livicla, E. M., differs in struc- 

 ture from the last only by having four petals, stamens, and 

 calycine segments ; the leaves are however much larger, more 

 lanceolate, and tapering much at the apex. 



19. Curtisia faginea requires no observations; it is now 

 generally removed far from Celastrinecs. 



From the above notes it will be seen that in ail the species 

 of Cape genera usually referred to Ilicinece, which I have 

 examined, I have never found fewer than two ovules in each 

 Vol. III.— No. 21. 2 L 



