ARNOTT ON SOUTH AFRICAN PLANT, 259 



Fructus inflatus? 2 — 3-spermus. Frulex ramosus. Folia 

 ad ramulorum apices fasciculafa, impart pinnata ; rachis in- 

 terrupte alatus, alis anguste ohovato-oblongis bast attenuatis ; 

 foliola obovata, complicata, undulata, integerrima, muerotiata, basi 

 aciita^ subsessilia. Flores corymhoso-racemosi, rubri. 



My observations on the female flower vvere made on a very 

 advanced ovary ; the petals had fallen away, but the filaments 

 of the stamens remained, and presented the same appearance 

 as in the male flower ; hence they are probably fertile. The 

 only fruit I have seen is far from mature, and is so much 

 pressed by the process of drying that I cannot ascertain its 

 form; the pericarp is thinly crustaceous, much larger than 

 the seeds, and as there seems no trace of pulp I presume it 

 is allied to that of Cardiospermum, a.nd Aitonia^ although this 

 last genus cannot be united to the order oi Sapindacece. 



21. As these notes principally relate to the genera, I shall 

 pass on to Lythraricce. The only new genus here is Toly- 

 peuma {T. Jloridum, E. M.), but how this differs from Neseea, 

 I cannot discover. 



22. Myrtacece : of these Jambosa cyminifera, E. M., is a 

 Syzygiutn ; No. 5366, is Eugenia Zeyheri, Harv., and No- 

 5367, is Eugenia? Capensis Harv. No 3576, is Phoberos 

 Eckloni (Eriudaphiis Echloni N. ab E.), as I have already 

 noticed. 



23. LoasecB: the genus Cm'rfbne (C. menizelioides, E. M.) is, I 

 am informed by my friend Mr Bentham, the same as Fis- 

 senia (F. arabica) Brown mst. 



24. OnagraricE require no remark, farther than Vahlia is 

 now removed to a very different order. 



25. Bruniacenc; on the ffenera of this order few alterations 

 have been made since these genera were determined by Brong- 

 niart, but it nevertheless appears tome that some modifications 

 are required. Ttaspallia is described with a perfectly free ova- 

 rium, upon the lower half of which the petals and stamens 

 are inserted ; now in the original species, R. microphylla, the 

 lower half of the ovary certainly does cohere with the calyx- 



