ARNOTT ON SOUTH AFRICAN PLANTS, 263 



E:ist India: while the only instance o^ Bryonia, in the Prodr, 

 Fl. Penins. I. O., wouUl be B. laciniosa* Linn. But if Bryonia 

 is to be broken down, the other species must be disposed of. 

 Thus B. scabrella, Linn, has the style arising out of a fleshy 

 disk as in Bryonia and PHogyne; it is undivided, and lias 

 three ovate erect stigmas more or less united together; but 

 the male flower has all the anthers united, the cells posti- 

 cous, linear, and straight : it thus approaches Cephahmdray 

 but then the anthers are gyrose. In B. tiibifora, W. and 

 A. (of which the male only exhibits flowers in my specimens) 

 there are three slender filaments, with the anthers cohering 

 into one conical mass covered on the outside with slender 

 linear anfractuose anther-cells; it thus also approaches to 

 Cephalandra, but the tube of the perianth is slender and 

 long : not having seen the male of Cephalandra, I am uncer- 

 tain whether the stamens be united or free at the apex ; they 

 are however united at the base, according to Schrader. Br. 

 rostrata^ Rpttl. belongs yet to another group : here the style 

 is entire, stigma large, deeply lobed, lacerated, and recurved; 

 anthers three, anticous, nearly sessile with the connectivum 

 produced beyond them at the back into a short beak. Br. 

 epigcea, Rottl. has a similar style, but my male specimens are 

 not suflSciently perfect for examination. Now if we adopt 

 Schrader's tabular view (Linn. xii. p. 403,) B. rostrata, 

 epigaa, and deltoidea, Arn., would form a new genus {Aech- 

 mandra) between Coniandra and Cyrtonema ; B. tubifora 

 would form another (Gymnopetalum,) near Trichosanthes ; 

 and B. Scabrella would not agree with any of his sections, 

 but might be placed under the name of Mukia, in a section 

 intermediate between those to which Pilogyne and Bryonia 

 belong, in which last the anther-cells are flexuose, gyrose, or 

 anfractuose. 



28. Among the Coniferce, we find inserted Ophiria stricta, L., 

 with which it has certainly no affinity. This genus is entirely 



* B. laciniosa. Anther-cells anfractuose or rather sinuose along the 

 margin (at the back) of the sinuated dilated apex of the filament : there is 

 no gland in the bottom of the perianth. 



