ON THE NATURAL SYSTEM. Xlix 



first promulgated in the Horce Entomologies, where it has been exten- 

 sively applied, and the author has shown its existence both in some 

 of the highest and in one of the lowest groups of vertebrated animals : 

 yet nowhere can I perceive that it has been declared — what I think 

 it to be — the only certain test of a natural group. Circles may be, 

 and have been, formed with such a deceitful appearance of following 

 nature, that the most eminent and the most cautious have been led 

 into a belief that they were strictly natural. If such a group is thought 

 to be complete or perfect, it is very well to say, put each of its divisions 

 to the test of returning into itself*, and the fallacy will be discovered ; 

 but among groups of a certain value, genera and sub-families more 

 particularly, there is not one in three that can be so tested. This 

 inability partly arises from our superficial acquaintance with forms, 

 and partly, as we believe, from there being many real gaps in the 

 chain of continuity. Without, therefore, some other test for a natural 

 group, than the mere circumstance of its returning into itself, or even 

 its simple parallelism with a contiguous group, I consider demonstra- 

 tion not to have been attained. The theory of representation thus 

 steps in, and at once dispels the illusion, or demonstrates the correct- 

 ness of the series. In the sub-families of Myotherina and Pariana, I 

 have endeavoured to exemplify this principle of the natural system 

 in all its bearings. 



3. The results that have attended my analysis of every natural 

 group hitherto investigated, lead me to differ in the onset from all 

 who have adopted the quinarian system. And so far as regards the 

 order of Insessores among birds, I have endeavoured to shew that 

 the primary circles of each group are invariably three. These I 

 have denominated — i. The typical ; ii. The sub-typical ; and iii. The 

 aberrant. Mr. Macleay, on the contrary, considers that every group 

 is first resolvable into five minor groups, two of which he terms 

 normal, and the other three aberrant. I know not why this talented 

 writer should have chosen to have used this latter mode of division, 

 which is binary, and but ill calculated, as it appears to me, to express 

 his own definition of a natural group. Neither he, nor M. Fries, 



* Mr. Macleay's Letter to Dr. Fleming, p. 33. 



