556 ON THE RELATIVE POSfTION OF 



Secondlv. Where the production of ovula is limited to 

 the external angle of the cell or axis of the leaf supposed 

 to form the carpel. 



A case of this kind is found in a portion of one of those 

 families in which the whole surface is generally ovuliferous, 

 namely, in IlydropeUidece., wdiich 1 have always regarded as 

 merely a section of Nyinphceacece •} and from the nature of 

 these differences in placentation, wdiich are more apparent 

 than real;, an argument might even be adduced in favour of 

 that opinion. 



A placenta apparently limited to the outer angle of the cell 

 also occurs in the greater number of species of Mesemhry- 

 anfJiemiim. As this structure, however, is certainly not 

 wdthout exception in that very natural genus, several species, 

 among which are Mcsemhryanfhernum crystaUinum, cordi- 

 folium, pajmlosum and nodiflorum, having the placenta con- 

 fined to the internal angle of the cell or margins of the 

 carpel ; and as in some of those species in which the outer 

 angle is placentiferous, the production of ovula is not con- 

 fined to it, but extends to the lower half of the inner angle ; — 

 this apparent deviation from ordinary structure may perhaps 

 be explained by assuming cohesion of the inflected portion 

 of the carpel with the wall of the cell ; — an hypothesis, in 

 some degree supported by the fact, that in several species 

 the termination of the assumed inflected portion is free and 

 not ovuliferous. 



But whatever opinion may be adopted as to the relation 

 of this seemingly anomalous to the ordinary structure, it 

 cannot, as M. Fenzl proposes," be employed as the essential 

 character of a distinct natural family limited to the Linnaean 

 genus Mesembryantherimm. 



The placenta then of a simple ovarium in its usual state, 

 according to this view, is necessarily double ; though by the 

 complete suppression of ovula in one of its two component 

 parts, and their diminished production in the other, the 

 ovarium is not unfrequently reduced to a single ovulum. 

 That such is the origin of the sinde ovulum is at least 



^ Gen. Pvem. in Fiinders's Voy. vol. ii. Append, p. 59$. {Ante j}, 74.) 

 2 Aimal. des Wieii, Mus. vul.'i, p. 349, 



