STIGMA AND PARIETAL PLACENTzE IN PLANTS. 559 



being obviously so in many cases ; and in one genus ut least, 

 Tasmannia, it extends nearly the whole length of the ova- 

 rium, so as to bo commensurate with and placed exactly 

 opposite to the internal polyspermous placenta. 



That the stigma is always double appears probable from 

 those cases in which it is either completely developed, as in 

 the greater part of Graminece. where the ovarium is simple ; 

 in the compound ovarium in TJrena ; and from those in which 

 the development, though less complete, is still sufficiently 

 obvious, as in many Eupliorbiacecs and in several Tridece. 

 This degree of development, however, is comparatively rare, 

 confluence between the two stigmata of each carpel being 

 the more usual structure ; and in the compound pistillum a 

 gTeater degree of confluence often takes place in the stigmata 

 than in the placentae ; — a fact, which in all such cases is 

 obviously connected with adaptation of surface to the more 

 complete performance of function. 



Another difference frequently occurs between the mode of 

 confluence of placentae and stigmata, namely, that in the 

 compound but unilocular ovarium, while the placentae of the 

 adjoining carpels are united, the stigmata of each carpel are 

 generally confluent. But this rule admits of exceptions, as 

 in Parnassia, in many Cruciferae, and in Papaveracea ; in all 

 these cases the stigmata as well as placentae of the adjoin- 

 ing carpels are confluent, a structure satisfactorily proved in 

 CrucifercB by several cases of monstrosity, in which the 

 stamina are transformed into pistilla; and in PapaveracetB 

 by a series of modifications of structure as well as by a like 

 transformation of stamina. 



A similar confluence of stigmata in the compound mul- 

 tilocular pericarpium is of much rarer occurrence ; it is 

 found, however, in the majority of Iridece, in which the three 

 stigmata alternate with the cells, and consequently with the 

 placentae of the trilocular ovarium. That this is the correct 

 view of the composition of the stigmata in Iridecs is at least 

 probable from their occasional deep division, and more par- 

 ticularly still from the bifid petal-like styles or stigmata 

 which are opposite to the cells of the ovarium in other genera 

 of the same family, as in Iris and Morma. In both these 



