270 POLYGAMIA. 



requisite character. Some of the grasses indeed have 

 awns to one kind of flower only, but that part is too 

 uncertain to establish a character upon ; and this 

 family is so natural in itself, ai:d so liable to variations 

 in the perfecting of its flowers or florets, that there 

 can be no doubt of the propriety of classing its gene- 

 ra simply by the number of their stamens and styles, 

 which are very constant. 



2. Dioecia. The different flowers on two different 

 plants. I can scarcely find a certain instance of this, 

 except Hippophae^ already mentioned under Monoecia 

 Tetrandria. 



3. Trioecia. Of the only two genera which have ever 

 been placed here, Ceratonia, Cavan. Ic. t. 1 13, be- 

 longs to Pentandria Monogynia. Ficus is so cele- 

 brated for the diversity of its flowers, as connected 

 with the history of vegetable impregnation, see /?.262, 

 that we are glad to take advantage of a trifling differ- 

 ence in the calyx of the two florets, (die barren one 

 being most frequently three-cleft, the fertile five-cleft,) 

 to keep it here. 



All things being considered, this Class may be 

 ihought scarcely worth retaining. Yet as we know 

 two or three genera entitled to a place in it, upon 

 principles which the analogy of the two preceding 

 Classes shews to be sound, we cannot tell but others 

 may exist in the unexplored parts of the globe. For 

 this reason, and for the uniformity of the system, I 

 would venture to preserve it. If the 2 1st and 22d Clas- 

 ses should hereafter be reformed by some judicious 



