358 MOXOECIA. 



not, and it might puzzle an adept to determine the 

 question. For instance, whether the nectary in SalLv, 

 different in the barren and fertile flowers of some 

 species, should lead us to keep that genus in Dioecia, 

 though in other species the nectary is precisely alike in 

 both the kinds, and occasionally an abortive germen 

 occurs in the barren flowers, as stamens do, more 

 rarely, in some fertile ones. Considering all this, I 

 should refer Salir to Dlandna Monogynla. 



With respect to those Monoecious or Dioecious 

 genera whose barren flowers are decidedly unlike the 

 fertile ones, the former being in a catkin, the latter 

 not, as Corylus, Quercus, &c., I conceive nothing 

 more pernicious or troublesome can be attempted than 

 to remove them to the Classes of united flowers. They 

 meet with no allies there, but, on the contrary, form 

 so natural an assemblage by themselves, as to be una- 

 nimously kept separate by the authors of every natural 

 system that has appeared. But even if this were not 

 the case, there is a most important reason for keeping 

 them as they are, which regards the artificial system 

 more particularly, and of which its author was well 

 aware; they are of all plants most uncertain in the 

 number of their stamens. Now this uncertainty is of 

 little moment, when we have them primarily distin- 

 guished and set apart from other plants by their Mo- 

 noecious or Dioecious character; because the genera 

 being few, and the Orders constructed widely as to 

 number of Stamens, we find little difficulty in deter- 



