216 STUDY OF NATURAL HISTORY. 



arrangement. It then becomes necessary to remodel 

 the whole; and if, after this, no general results can be 

 obtained — if there is no regularity in the occurrence 

 of the same analogous forms in the different groups 

 — our circles want verification, and must of course 

 be considered hypothetical. 



(152.) The presence of such remarkable forms 

 as have been just instanced in a natural group, 

 would seem to point out at once the most obvious 

 means of deciding on the number of divisions which 

 a genus contains ; and as every distinct modification 

 is the type of a sub-genus, we derive, in that first or 

 lowest stage of induction which we are now sup- 

 posing, great help in determining the number of 

 divisions in a genus. It may so happen, that in one 

 we reckon four, in another five, in a third seven, or 

 even more. It then becomes an important question, 

 whether these assumed divisions or types can be 

 augmented or reduced so as to bring them to a 

 definite number in each group. If this can be accom- 

 plished, it is clear that another principle of harmony 

 will be discovered ; and we shall have good reason to 

 conclude that the number of divisions into which the 

 majority of our groups can be divided, will be that 

 most prevalent, if not universal, in all others. The 

 verification, however, of such a theory cannot be satis- 

 factorily attained until we quit genera, and ascend to 

 higher generalisations ; and for the following reasons. 



(153.) Suppose, for instance, we looked to the 

 genus Trichius among coleopterous insects, and 

 agreed with an eminent entomologist in dividing it 

 into seven principal sections or sub-genera; and we 

 assume these to be natural. But on turning to the 



