freely and more rsridly than "H, Heethil cloes, is .alvjays rauch 

 sra8.11er, and its rrov/ths are longer in proportion to their bre&dth 

 and thickness. 3een side by side, the two plants certainly seera 

 distinct, rnd vrhen the f louvers of R. dubia are kno'.vn, definite 

 riistinctive ch'-rpcters may he found. 



It has long been in cultivation and is known in gardens 

 under the name of Mesembryanthemum fissum, and is the plant I 

 have mentioned under thst name in the •J'ournal of the Linnesn -^oc- 

 iet^T-, ?s ebove quoted. For as the plent v-es known in gardens 

 ^s r. fissum and agreed ffiirly v;ell with Haworth's description, 

 1st that time mistook it for that species. But I have since 

 discovered that a drawing exists at I-ew of the real I", fissum, 

 which demonstrates thet sr^ecies to be entirely different generical- 

 ly end sr^ecificell'"' from this which I noAv name R. dubia , 



3.--R. Roodi^e, N.E. Br. {?ig. 76). Plant about l|-£ inch- 

 es high, consisting of a clump of growths crowded upon the short 

 divisions of a deeply descending rootstock 6-8 lines thick. 

 Each p-ro-'-'t'^ 1-1 3/4 inch high, v/lth the basal part of the leaves 

 united into a bod^^- 3/4-1 inch long and ultimately withering in- 

 to « sheath, their free part 4-9 lines long, 4^--10 lines broad 

 and 4-9 lines thick, ovBte or somewhat hemispherical, flat on 

 the free, usually very broadlj?- rounded on the beck, with an 

 obscure or distinctly evident keel, and with or without smsll 

 teeth i--^- line long on the edges end keel; surface smooth, micro- 

 scoT^icrlly r-uberulous, green. Pedicels partly exserted, up to 

 9 lines long and about 2 lines thick, end together with the calyx 

 nuberulous like the leaves, '^eljx 5-lobed; iobes 3-";' lines 

 lon.o-, bropdiy ovpte, obtuse, 3 with a membranous margins. Corolla 

 onl^'- seen in a withered condition, 1^ inch or more in diameter; 

 petals about 5-7 lines long end ■|-5/4 line broad, linear, obtuse 

 or no+ched at the anex, orange-yellow. '3-(;,oj»;ens numerous, collec- 

 ted into a cone, with some of the outer ones loose and spreading 

 from the cone, apparently whitish. Stigmss 9-11, about i-|- line 

 lon.'^, end about equalling the st?mens, filiform, apparently of a 

 dsrker colour then the stamens. Capsule v/hen closed about 5 

 lines in diometer, flattened above and beneath; valves pallid 

 within, with orrnge-broim expanding-keels . Seeds about * line in 

 ^^■'"leter, v^hitis^, with a psle brovm ninple. 



Bushmanlf'nds Precise locality not stpted, i-rs. S. Hood. 



This STiecies differs from R. ^-^eathii by its greener colour, 

 microscopically puberulous surf&ce, the minute teeth on the edges 

 and keel of the leaves, end orange-yellow flowers. It so much 

 resembles e dried si^-ecimen in I'ew ^--erbsrium collected by the 

 Ipte Professor Pearson (iTo. 4071 ) at ^nriet in Little -^ushman- 

 land, that at first I thought the tv;o might be specif icclly i- 

 den-^ical, but I find the structure of the fruit of the two plsnts 

 is so different that according to the modet^i^ conception of ' 

 genera they ought not to belong to the sc. -us . j^ot J-etrson's 

 pl^nt he s a half-globose cri^sule whose Vfrlves have stout (not 

 thin) entire exrandin^-keels, '"ith narrow ^^^ ""^'^y small wings; 

 the cells hr;ve stiff Tnot membranous) eel"- ' forming a roof- 

 like (not flat) cover to them, v.-ith their outer ends turned back 

 and resembling the mouth of a trumpet, and the opening is nearly 

 closed bv a lar^-e tubercle. Ther is nothing of the kind in the 

 ^ruit of HiraariF . It a^^pe^rs to be a case of one genus mimicking 



