15^ dried pnd curled un vhnlst bein^ rhotoprerhed, but cpn be seen on 

 one of the loT^er lobes, and slmiler vdngs ?re better seen on the 

 specimens A, 3, 0, K (^'i^. 62), These keels are frequently of a 

 dr.rVrer colour than the rest of the corsule, rnd consist of e horny 

 water-absorbin? tissue that forms the' mechanism by v;hich the valves 

 are o'-ened end closed. As no design? tinf term has yet been app- 

 lied to "^hen ■•"hey may be celled exnandinp keels, ^n some si^ecies, 

 as in D, the keels ere ouite separate, in others, as in Ji-, B, C^ 

 ■S, thej pre n]osely contiguous or united st their basal part and 

 more or Tess seTiprate or divergent at the ar^ic^l ^^art, or occesion- 

 all7r ere fused Into a single central keel. Sometimes the expan- 

 ding" keels are entirely Y.'ithout vings or avrns, but more often they 

 are •?"5no-ed at the um^er pprt on their outer margin or are tipped 

 with ^n pvm, or are both vjinged and fvned. In some species the 

 •'■'•inffj' a'^'T^e^r to arise from the inner surfece of the valve, rs in 

 F, v^here they are infolded and do not appear very clearly in the 

 ficure, 



152 but vnc'er t-^'o of them a seed h^s become lodged, each represented 

 in t^e fie-u.re by a black snot. The central part of the capsule is 

 occur>ier' by the cells containing the seeds. The cells are of the 

 same num.ber as the lobes and alternate with the latter. Usually 

 they are completely covered or roofed over by tvo v.dngs (marked b), 

 vfhich arise from the repartitions separatinfr the. cells, 'flaese me^'' 

 be termed cell-^'ings, although often seni-transr^arent , are of a 

 someT^'hat horney nature and slightly stiff; in others they are more 

 membranous and more flexible, as in A, B^ G^ G, II. Their inner 

 ed.c"es are usually closely contiguous or overlap, and their outer 

 mara-ins either leave a small gar) betv-ieen then and the outer bound- 

 ary o"" each cell, as in all the figures except ^, S, F; or the gap 

 is nearly filled "by a hard boss as at c i^) in Ffg. 62; or they are 

 somewhat r>inched together as at the T^laced marked d in ^, fig. 62, 

 But in a com.rara tively small number of species the cell-wings are 

 entirely absent, ?s in F, or are very rudimentary so that the seeds 

 are entirel""' ^incovered. N. S. h^ovm 



(To be continued,) 



Mesembryantheraum and some nexv genera separated from it. 

 Gard. ^hron. HI. 70 : 172. 1921. 

 (Continued from vrifre 152.) 



172 '''^hen the dispersion o" the seeds of J^^esembryanthemum is inves- 

 tip-ated, r.-e find it stated in books that they are vashed out by 

 the rp^n, and this is the only information given concern jnr their 

 disnersal. In the case of the F tjrpe of carsule (see ^igs. 61, 

 62, T>. 151 ) it is obvious th^^t the seeds after the expansion of 

 the lobes, -"ould either be vrafahed out of the carsule by the rain, 

 or, after the rain has ceased and whilst the capsule is drying 

 UT> and closing, might be shaken or blown out by v/ind. But in the 

 case o-f* such extra ordi'narily comnlicated structures of the types 

 "^ pnd '5, this exT^lanati on is anythinr but convincing or satisfac- 

 tory, for although it is evident that rain has something to do 

 wi t^ the release of the seeds from the canaules, I have as yet 

 nui^= fa^'ied to discover ho"" this is effected. For in the ^ type 

 the entrance to the cells containing seeds is so ne'^rly blocked 

 UT> tv the boss (c) thpt there is only ouite a narrow space left 

 ■'^et'-'ee'^ the boss and the cell-T-inP's and the sides of the cells, 

 whic>i :' s not wide enough to nerrait the seeds to pass through it 

 without -force being used. And the wings that cover the cell like 

 a roo:^, are, in this tyre, rather to sere rate their contiguous 



