150 ARNOTT ON SOUTH AFRICAN PLANTS. 



very different ; and also of Casearia capensis, Ant. Spr. : I do 

 not find it in Ecklon and Zeyher's works. The other is Eri- 

 udaphus, N. ab E. ; this last is still retained by Endlicher, in 

 his Gen. PL p. 923, in Homalinece. This order has a fascicle of 

 one or more stamens opposite the inner divisions of the peri- 

 anth, while opposite to the outer segments there ought to be 

 no stamens, but a mere gland ; and the ovary coheres at the 

 base with the bottom of the perianth. In Eriudaphus, 

 however, the ovary is perfectly free, and there are stamens 

 also opposite to the outer pieces of the perianth. In the true 

 HomnlinecR, the fruit is capsular ; in Eriudaphus, it is baccate. 

 These circumstances intluce me to remove this genus to Bixi- 

 7ie(s, nor do I perceive the slightest difference between it and 

 Phoberos of Loureiro, of which several species ai-e described 

 in Wight and Arn. Prod. Fl. Penin. I. O. p. 'IQ. In that 

 work, from the resemblance of the genus to Flacourtia, we 

 improperly referred it to Flacourtianecs, but the placenta) are 

 simple, and Endlicher (^Gen. p. 919) has correctly brought 

 it near Prockia. Nees ab Esenbeck takes no notice of the 

 beak to the anthers in his Eriudaphus, but in three or four 

 Cape species before me, I observe this character more or less 

 distinctly ; it is particularly so in Drege's No. 3576, inserted 

 among his Mi/rtacecs, but which I presume is Eriudaphus 

 Eckloni, N. ab. E., and still more in another species which I 

 consider a variety of Phoberos (or Eriudaphus) Zeyheri ; 

 indeed it agrees better with Esenbeck's description of P'^- 

 Zeyheri, than specimens I have from Zeyher (No. 858), m 

 which last the leaves are often ovate and acute ; it is known 

 by the name of Wolf's Thorn, and the spines are in one speci- 

 men before me eight inches long, and three-eijxhths of an inch 

 thick. The structure of the anthers to which I have alluded 

 induces me to refer Rhinanthera of Blume likewise to Pho- 

 beros, and to consider his description of the fruit to be erro- 

 neous. The placentae are, as I have said above, simple, ana 

 seem to be constantly two in number. 



4. Rhus oblongifolia, E. M. Of this the following is the 

 analysis:— Sef»o/a 5, rotundata, concava, fimbriato-ciliata, 



