2(^2 t-'.^-ne Of the' penus '■'esembr'^rgnthemum, the follov\rinf^ appears to iv.e 

 the best method of errivinp at a satisfactory answer to this 

 somewhat complicated problem. 



In the ■*'irst rilace it is clerr from his ossigiing this 

 p-enus to the Class ^cosandria end '^rder -^entagynia that the 

 primary essential chpracters of the genus are that it should 

 have numerous stamens and five stigmas (or styles as he calls 

 the'^). Therefore v;e may eliminate from the list of species des- 

 cribed by Linne the seven species which have eight to seventeen 

 stip'mas and cells to the ovary. These are — !>-. calamiforme, 

 I-. pcinaciforme, I', rostratun, I--, loreiom, ^•-. difforme, i'^. lingui- 

 forme, and M. pugioniforme, which as ^inne himself pointed out 

 do not f^ccord with the generic character of "stif^nas and cells 

 of the ovpry 5." '^f some of these L^p^e does not mention the 

 mimber of their stigmas; indeed it is certrin that he never had 

 sriecimens of many of the species he described and that he only 

 knev; them from the descriptions and figures in ^illenius'^s tortus 

 elthamensis. 



The next characters of importance to take into consideration 

 are that the petals should be united rt the base into a short 

 tube, and the ovary be inferior, -^-"ere v;e are faced vith a some- 

 what curious problem of contradictory characters, v/hich again 

 demonstrates that Linne could not have examined the flov/ers of 

 the species he described. Out of the thirty-five species he 

 described, only one-sixth of them have a monopetalous corolla, 

 viz.: — M. nodiflorum, I'--, crystallinum, ^■-. geniculiflorum, 

 !'. noctiflorum, J-. splendens , and i'-. tortuosuia. -^^11 the others 

 have the netals free to the base, -t^-nd thene six species do not 

 accord with his character of the ovary being inferior, for they 

 all have the ovary partly superior, or more than half-superior, 

 and m.oreover, they all have axile placentas, a chars cter not men^- 

 tioned by Linne, while all the other species have free petals, 

 en inf'^rior ovr^r end the rlacentps on the outer v-'all or floor 

 o^ the cells. Therefore, either the character of the monopetel- 

 oras corolla or that of the inferior ovary must be discarded as 

 one o"!" the rrincinal characters of the genus. I propose to re- 

 ject that of the m.onopetalous corolla (petals united into a 

 short tube at the base), since it applies to such a small pro- 

 portion of the species, and to eliminate these six species from 

 • the frenus, as their nlacentation is also quite different. 



■'■he fruit is described by Linne as fleshy, but excepting 

 -'. ecinaciformp, the ^'^ottentot ^ie' and allies, the fruit of all 

 knovm species is a dry capsule. ^ very much doubt if Linne 

 ever sa^' the rine fruit of any species, for there is no fruit 

 preserved in his herbarium, and ^ sus'^ect that he ascribed this 

 character to the renus from the statements of others concerning 

 the fleshy edible fruit o-f* the ^-^ottentot fig. The stf tement 

 thpt the fruit is roundish must have been taken from the figures 

 in Hortuselthamensis, for the fleshy fru.its of I'^. acinaciforme 

 and allies are not ''roundish,'* but elon.p-^ted obconicrl or obovoid. 

 'Tiere'^ore, as the only species v/ith fleshy fruit (I-', acinaciforme, 

 Linn., v.'hich ct that date as he then understood the species 

 included also I', edule as a variety) does not accord either in its 

 shape or the number and character of its stigmas Uvhich are 

 plumose) with the Linneen definition, this species, as previosly 

 mentioned, must be eliminated from the genus end the character 

 of the fruit beinp- fleshy.'" also discarded as a generic chrrecter 

 for I.'esembrvanthemum. 



