144 long, filiform, pale yallor.dsh, ascending-spreading, vvith recur- 

 ved tips. CK/pry inferior, the top et the centre rising into a cone, 

 with the^psrt p round its base concave, light green. Capsule very 

 large, l^-S inches in diameter, the rira-part shallovvly saucer- 

 shaped, vith the 16-13 valves, grooved on the hack, rising in a 

 cone at about 3 lines v/ithin the ma.rgin, glabrous, pale brovm; 

 each valve besring on its inner face tvrb large, v/lng-like, trian- 

 gular pallid membranes (cell-partitions) closed together, 9-10 

 lines long, 6 lines broad at the base, narrov.lng upwards to an acute 

 point. Seeds few in each cell, neprly 1 line in diameter, slightly 

 compressed-globose and slightly keeled all round, smooth, pale 

 brown. 



Van Hhynsdorp Div.: near ''J'an Hhynsdorp, i^rs. E. Rood. 



This species differs from C. elongata, N.S.Br., by the tuber 

 being very long and s;}.ender, not swelling out into an irregularly 

 ovoid mess, by its leaves being concave above nearly to their tips, 

 by its smaller flowers, and by the petals (v;hich apceer to be less 

 heir-like) being very incons^^icuously ciliate, with minute straight 

 (not long and flexuose) hairs. 



I'ossibly the plent figured in the Botanicpl Register, t. 493, 

 as ^-. elong8t\im var. minus, Hav;. , as to the flowering branch only 

 (not the root) may belong here. It is not t^npicel C. elongata, 

 and no descrirtion is given of its rootst'^ck,^ the figure of the 

 tuber on thpt plate being mede from tynical ^. elongate. The 

 figure is made' from a plant raised at Kev7 from seed in 1819, and 

 W8S 'published on I'^ovember 1, 1820, so could have been only one 

 year old. It agrees with C. Hoodiae in its characters, except that 

 the stigmas are described as being deep yellov/,. It differs from 

 0. elongata, petals being straight and minute so as to be sc&rcely 

 visible to the n?>ked eye. 



In the ^operberg "ountains, Little Ngmaqualend, there is a 

 srecies v.dth a tuberous rootstock stated to be 12 inches long, that 

 has a large fruit very similar to that of C, Roodiae in size, but 

 the cone of valves is shorter, the valves differently curved end 

 the vnng-like partitions adherent to the valves much shorter, so 

 that 3s probably a different species. 



Since the above was in type, I have received from I'rofessor 

 O.O.^Tel a living specimen of the plant upon v/hich the genus Herrea 

 vras founded by Schv/antes in '^-filler's I^eutsche G^'rtner-'^eiting,. 1927, 

 p. 436, UPon imperfect material and vdthout giving it a specii'lc 

 name. This plant proves to be merely a species of Gonicosia and in 

 ell a'^pearpnce is identical V7ith C. Roodiae, under which I rlece it 

 as a synon:^nn. Its flower has been described by ^-rs. ^olus u^otes 

 on ''esemb. II, 233) v.dthout ha'"'ing identified it as being a Conicosia 

 and without a specific name, 



11. 0. robusta, N.3.Er. , in ^avd. Chron,, 1928, LlOCXr/, 253.— 

 Rootstock not seen nor mentioned onnthe label but Judging from the 

 specimens seen) doubtless tuberous, ^ried flowering branches only 

 seen. 3^ lines thick. Leaves o^^posite, 6-9 inches long, 3-3^ lines 

 thick, the ultimate or so-called bracts 4-4^ inches long, their 

 sectional form indeterminable, but apparently not acutely angular, 

 obtuse at the apex. Sach branch seen has three flowers, the central 

 one v/ith a pedicel 4^ inches long, 1-1^ line thick; lateral flowers 

 in bud only. OgTyx with the longer lobes exceeding the pfetels and 

 1-1^ inch long, the 3 inner v;ith broad, blfckish, membranous margins, 

 Corolla probably about 3 inches in diameter; petals very numerous. 



