OF GENERA. 277 



species from Convolvulus, which are attached to that 

 genus by stronger ties of another kind. 



Some genera are obvious and indubitable both in 

 habit and character, as Quercus, Rosa, Euphorbia, 

 Begoyudy Exot. Hot. t, 101, and Sarraceyiia, t. 53 ; 

 others are obvious, but their character extremely dif- 

 ficult to define, as Valeriana, The greatest difficulty 

 lies in distinguishing genera that belong to such very 

 natural orders as the Grasses and Umbelliferous 

 plants ; and the ablest botanists differ about the best 

 guides in these two particular cases. Yet other orders, 

 equally natural, sometimes afford very excellent ge- 

 neric differences, as that to which Rosa, Rubus, Era- 

 garia^ &c., belong; and even in the Papilionaceous 

 plants with ten distinct stamens, a tribe hitherto judged 

 inextricable, a regular examination on scientific prin- 

 ciples has led to the discovery of very natural well de- 

 fined genera. See Annals of Botany, ^'. 1. 501. I 

 have in a preceding chapter hinted that the umbelli- 

 ferous plants seem to me very capable of being well 

 discriminated by their seeds, and other botanists have 

 held the same opinion. 



But though I feel convinced, as far as my experience 

 goes, that genera are really founded in nature, I amr 

 far from asserting that Linnaeus, or any other writer, 

 has succeeded in fixing all their just limits. This deep 

 and important branch of natural science requires the 

 union of various talents. Many persons who can per- 

 ceive a genus cannot define it ; nor do acuteness of 



