138 ffrevish-freen leeve'^ densely covered '-ith minute, dull pepillae. 

 Stems l-lr lines thick; flowers in more or less compact cymes, 

 ouicl<-ly succeeding one 'another, ench with r flower bud formed on 

 one or 'both si'^es o-^ it hefore the flower f&des ; petels Isx, not 

 touching one another; stigmps alvrays srreeding. 



Tedlcels 4-6 lines lonf?; petals entirely bright violet-purple. 

 Rhodesip, 



M. i^honi 

 Pedicels 1-3 lines lonp; petpls ^-hite, or v-hite with pale purple 

 tirs. Transvaal. 



^ y-, herbeum. 



Stems p-t line thicV; flowers solitary st distent intervals, 

 not ■^or'^Jnc distinct cymes, nor with a flov/er bud forming beside 

 each flower while the flower is open. 



Pedicels 8-20 lines lonp", retals touching or overlapping one 

 another, hriffh"'- rosy-Tur^-ile ; stigmas collected into an erect 

 cone, never spreading. 



M. macellum, 

 i.r. macellum is evidently self-f ertilised, as each of my three 

 plants of this species has expanded its first flower (the only one 

 then on m-^'- T>lr>nts) at a time "'hen no other plant of any species of 

 the itrenus had a sinful e floxver (even in bud) uPOn it, yet that 

 flower in each case r^roduced and ripened a capsule containing good 

 seed, that has germinated, end since then I have several times had 

 each Q-f these Talents in flo^-er separatel37", so that they had no chance 

 of beinff riollinated by each others* pollen, yet all the flowers .S'^em 

 to be fertile. On the other hand, ^■^. i«iahoni and J^^. herbeum seem 

 to reouire to be cross-fertilised, other^/ise neither produce cap- 

 s\iles, and the flowers of ^^'-. ^'-ehoni often fall off the pedicels', 

 leavinr the latter to wither and harden. I have not observed this 

 fa-T', in"- o"^ the flower from the redicel to be the case either in i^^. 

 herbaujn or ^•'. macellum, 



The three above snecies -^orm a little group that appears to 

 me to be best i^ipced under the section Digitiflora, but seem also 

 to have connection v.'ith such s'^ecies as ^■-. hjrtur' pnd i-i. Goo^e"^i, 



"^ron the J^ardener's -point of view, the merits of the genus 

 ''esembrvanthemum rest u^on the attractiveness of its flcv/ers. 

 Yet, apart from their floral merits, there are many species that 

 have hitherto been included in the genus Llesembryanthemum that ep- 

 T^e°l stroTly to those who are interested in -"lent life, on ac- 

 coun"^ o-^ marked neculiarities they have in a^^^eF ranee, habit, mode 

 of e-rowth, or structure. ?or it is not generally known that this 

 fenus, as at present constituted, contains some of the most remark- 

 able Plants that exist. Certain of them well deserve a place in 

 the collections of those who have a likinr for curiosities of the 

 Plant '"orld, for, besides bein^ rem-prkable, some of them produce 

 very charmin-'? flowers. It is to some of these extraordinary plants 

 that I wish to direct attention, because I think the time has nov; 

 come vrhen they shoul'-'' be ^enericallv separated from ^'^esembryanthemum. 



(To be continued.) 



N. S, 3rov;n 



l-esembryanthemum end some nev; genera separated from it. 

 Gard. Chron. HI. 70 : 151. 1921. 

 (Continued from page 138.) 



151 In l'^^!!, just 100 yeprs ago, -^-drian Hpv/orth published his 



"^evisiones Plantarura Succulent a rura, and described therein 310 



