13 subeaually or unequally S-IoTdgcI dov/n to its union vdth the ovary; 

 lobes vjith broad bases (the three inner with membranous margins) 

 and subulate or terete tips. Petals very numerous, free, in several 

 series, linear, often cilia te at the basal part. Steminodes numer- 

 ous, filiform or hair-like, at first more or less incurved, otcmens 

 numerous, at first inciirved around or over the stigmas, afterwards 

 spreading and exnosinp; the latter; filaments bearded at the base. 

 StiCTias 10-20, filiform; style none. Ovary inferior or paartly 

 sucerior, with a flattish top rising at the centre into a short 

 cone, 10-20-celled; placentas on the outer Vi^all of the cells; 

 ovules several in each cell. 'Capsule dry, half superior, with the 

 conical top separating into 10-SO narrow valves or segments, 

 which stand more or less erect, not spreading when wetted and are 

 v/ithout expanding-keels, but each valve bears a pair of thin, sub- 

 membranous, wing-like cell-appartitions on the inner face. Seeds 

 rather large for this group, subglobose or somevirhat lenticular, 

 • o-^ten slisrhtlv keeled at the margin, but not winged, smooth, — 

 IT.-E.Br., in The Gardeners' Chronicle, 1925, Vol. LaX^/HI, p. 4 33, 

 and in Phillips* Gen. S. Afr. Fi. P].., p. 247. 



Species 11, and doubtless several more, natives of South 

 Africa, the type of the genus being 0. pugioniformis, N. ^. Br. 



The name is derived from the Greek, konikos, conical, in 

 allusion to the conical, top of the fruit. 



Conicosia is easily distinguished from, all its allies by 

 its very flong and narrow, soft and dotless leaves, which tiaworth 

 likened to daggers, and by the pecuiar capsule dividing into nar- 

 row, erect valves at the top when ripe, instead of remaining closed 

 and expanding vrhen wetted by means of exT>anding-keels, which are 

 absent in this genus. 



On account of the great resemblance most the the species beer 

 to each other, their flowers especially, being all very much alike 

 except in details, I amdisposed to believe that several species 

 yet remainto bediscovered thet have been mistaken for old species, 

 and have not been collected. For a close study of the characters 

 of* the s-^ecies that have hitherto been discribed and figured has 

 revealed the fact that several wrong identifications have been 

 made, causing much confusion of nomenclature. Haworth began the 

 confusion by his mididentification of ^•^. pugionif orme, and 3alm 

 Dyck and others have followed. But it is quite impossible to 

 understand hov: Kiinsmenn, in his Clavis ^reyniana , could have iden- 

 tified Breyne t. 80 (which is C, fusiformls) as being i>i. bicolorum, 

 ".'illd., a shrubby species, and Breyne t. 81 (which 1 believe to 

 be ^. ''uirii) as being M. molle, 5preng. This confusion I have en- 

 deavored to disentagle, end have made explanations under the species. 



Both Haworth and myself have supposed the leaves of this genus 

 to be alternate, and vrhe a large main tuft of leaves is examined It 

 is difficult to form any other conclusion, but having nov; observed 

 seedlings of four species develop, I find that the leaves are really 

 oT>nosite (see figure under 0. R-^odiae) , but by a mode of displacement 

 by succeeding leaves the^'" become apparently alternate. 



The fuit of this genus is very' different in structure from 

 that of any other. In the first place the cells or carpels of the 

 ripe fruit" are so lightly adherent to the calyx-tube that they can 

 easily be separated from it. V/hen quite ripe the valves slightly 

 separte from each other, and the v/ing-like membranses (really cell 

 partitions) attached to their inner face also become more or less 

 detached from the central column, -^s the cells containing the seeds 



