218 STUDY OF NATURAL HISTORY. 



cidence, we must yet bear in mind that we have 

 advanced but one step in the scale of induction; 

 that it is not very difficult, even with a strict atten- 

 tion to the foregoing rules, to divide two genera, 

 each into the same number of sections. To assume, 

 on such slender premises, the existence of a general 

 law, that all genera will be found similarly constituted, 

 would be a total departure from that mode of enquiry 

 which is absolutely essential to the prosecution of 

 all physical science. 



(154-.) When, therefore, we have verified the 

 prevalence of a definite number in the divisions of 

 a genus, by comparing the contents of several, we 

 may then advance a step further, and are at liberty, 

 from the facts already elicited, to form a theory. 

 It is essential, however, that, in so doing, we over- 

 step not those inductions which lie before us, and 

 which can be appealed to as instances of particular 

 verification, and as presumptive evidences of the 

 universality of the law assumed. By the process of 

 investigation we are now pursuing, all deviations 

 from the law we assume, must be accounted for on 

 sound principles, or by the probable operations of 

 known effects. Thus, for instance, many groups 

 which, from having been already analysed and de- 

 monstrated, we know to be genera, contain but two 

 or three divisions, others four, and many but one. 

 Now, as this occasional paucity of forms in a genus 

 may be accounted for by various natural causes 

 (148.), we are not hastily to conclude that there is no 

 definite number in nature, or that a genus may 

 contain from one to twenty sub-genera, for aught 

 we know to the contrary. Such imperfect groups, — 



