ARNOTT ON SOUTH AFJIJCAN PLANTS. 265 



Having endeavoured to elucidate the synonyms of these 

 plants, I shall advert to the structure of the ovary and their 

 place in the system. Endlicher states the ovary to be 1 -cel- 

 led, with 2-3 ovules suspended from the apex of a free central 

 placenta. Klotzsch gives the same structure to the ovary, but 

 attributes only one ovule to Grubbia, and two to Ophiria. 

 Endlicher with doubt, and Klotzsch with certai/ity, refer them 

 to Santalaceae, and were there indeed a free central placenta, 

 such an affinity would be at once acknowledged; but my ex- 

 amination leads to a different conclusion. 



In neither genus can I discover the least trace of a genu- 

 ine free central placenta. But M. Decaisne in an excellent 

 memoir on these and other plants in the 12ih volume of the 

 new series of the Attn, des Sc. NatureUes, observes : " Hitherto 

 the ovary of Grubbia has been described as unilocular ; never- 

 theless, on examining the flower before or even at the period 

 of its expansion, we see the ovary divided into two portions by 

 a thin and membranous dissepiment at the summit, and on 

 each side of which is suspended an anatropal ovule; afterwards 

 one only of these ovules becomes developed, pressing the dis- 

 sepiment against one oi the sides of the ovary cell. In Ophi- 

 ria, this structure is observable in tfie ovarj', and resembles 

 exactly that described and figured by M. Brongniart in the 

 genus Berzelia^ belonging to the Brimiacea;" 



My observations on these genera do not precisely coincide 

 with those of Decaisne; but in both there is a decided tendency 

 towards a bilocular ovarium. In Ophiria, the dissepiment I have 

 always found to be imperfect, and attached only to the one side 

 of the cell, constituting an elevated internal ridge : there is one 

 pendulous ovule from each side of this dissepiment or ridge, at 

 the apex. In Grubbia I also iind constantly two ovules; and al- 

 though 1 have never been so fortunate as detect the complete 

 membraneous dissepiment mentioned by Decaisne, I find a 

 free very small and thin membrane separating the ovules, 

 which are pendulous from its apex ; -^ind along each side of 

 the inner surface of the ovary are two slightly elevated lines, 

 to which it is highly probable the membrane was attached in 

 Vol. III.— No. 21. 2 m 



