I 



MUTUAL COiS^NECTIONS OF TLANTS. 101 



fruit, and the Syngenesistae by their pappus, we cannot, in 

 the former case, associate a winged with a solid fruit, nor, in 

 the latter, a pinnated pappus with one that is bristly. 



It is much to be wished, that in all classes, or in all tribes of 

 plants, we had such characters as have a constant value, and 

 which could hence be employed in the formation of new gene- 

 ra. But that this is by no means the case, we see, in particular, 

 from the numerical proportion in very many genera and fa^ 

 milies ; that is to say, this proportion is often of so little value, 

 or it fluctuates so much, that we cannot employ it for the dis- 

 cTimination of genera. It is also obvious, that the character, 

 taken from one organ, cannot be applied to several families. 

 Important as the inflorescence is in the Grasses, we cannot use 

 it for the discrimination of genera in otlier families. 



151. 



Tlie second rule, by following which we may avoid the 

 above-mentioned errors, is, that as much as possible natural 

 genera should be constructed, and, when that cannot be ac- 

 complished, artificial genera ; but that the variation of form 

 and of proportion in one part, ought not to be considered as 

 sufficient for the construction of a genus, unless this difference 

 be also expressed by other marks. " The character does not 

 constitute the genus is a very wise saying of the founder of 

 Scientific Botany. A genus is not on this account firmly esta- 

 blished, because one or another difference of structure occurs 

 in individual parts, but because the plants actually exhibit 

 striking differences in their whole vegetation. We might 

 easily, by giving to a genus a very circumscribed character, 

 be able to separate from it all the species, and to unite them 

 to new genera, which have not that circumscribed character ; 

 but it is the business of Art, to evolve a more comprehensive 

 and generally availing character, so as to shew the value of 

 the principle, That the genus forms the character. 



152. 



Lastly, we must remark, that the further the knovv^ledg-e of 

 plants is extended, the more do we find that many great ge- 



