ON GENERIC CHARACTERS. 243 



the characters by which such divisions are to be 

 denned, are most judicious when they are drawn 

 from two structures: first, from that structure of 

 the animal itself, which it possesses in common with 

 all others of the genus ; and secondly, from that 

 peculiarity which renders it analogous to many 

 others out of its own genus. 



(167.) An instance, taken from an ornithological 

 group, will illustrate the foregoing remarks. The 

 genus Sylvicola {North. Zool. ii. 205.), or titmice 

 warblers, comprehends so large an assemblage of 

 species, that we are able to trace, and demonstrate, its 

 circular succession. All these birds agree, more or 

 less, in the form of their wings; the first quill of 

 which is nearly as long the second. This being the 

 most prevalent character, we select it as the chief 

 mark of discrimination, without stopping to enquire 

 whether other groups are to be also characterised 

 by these organs ; for this, we see, most assuredly is. 

 We find, however, that among all the birds thus 

 brought together, we have different modifications of 

 the other organs : some have the bill more conic 

 and entire, others have it depressed and notched ; 

 in two or three, it is very sharp pointed, and even 

 curved ; while a few depart from all the rest, in 

 having feet adapted for climbing. Now, it is ma- 

 nifest, that if we had set out with presupposing that 

 all birds were to be classed by the bill and the 

 feet, and that all other characters were of little or 

 no moment, this most natural group would never 

 have been detected. It is also clear, that, being 

 discovered, we cannot draw our essential and primary 

 characters from any of those organs which, as above 

 r 2 



