PRINCIPLES OF SCIENTIFIC ARRANGEMENT. 135 



collectively the largest numbers of forms that in their 

 several combinations are the most nearly equivalents, 

 and may be almost paralleled in that quality to the alli- 

 ance of species. Ascending from species, the naturalist 

 scarcely hopes to find in the groups formed above them 

 strict parallelism, although, to be logical, it should be 

 so, and, vihere the combinations are most natural, it 

 is most nearly so. Thus we do not again distinctly 

 reach equivalents until we arrive at these families, which 

 from linking together associations usually combined 

 by an identity of instinct and functions, attach to them- 

 selves greater interest, and form alliances pointed out 

 by the finger of nature itself, which are therefore 

 exempted from the arbitrary caprice of the constructive 

 systematist. 



It does not follow that families should be even nearly 

 numerically equivalent, for a family may contain a few 

 or a multitude of genera and species, or a multitude of 

 genera and few species, or also a multitude of species 

 and few genera. Families comprise groups of forms to 

 which nature delegates the execution of certain duties 

 and offices, and whether specifically numerous or few, 

 we may assume they are sufficient for the object in- 

 tended. If we can reach the motive that controls the 

 peculiarities of the group, it is a golden key to the 

 explanation of the structure of its constituents, and,, 

 perhaps might furnish us, if not with a positive clue, 

 yet with a surmise as to the functions of the collateral 

 groups of which it forms a member, and which diligent 

 observation may accurately determine. 



Families, to be natural divisions, should stand in the 

 same relationship to genera as species do, but from the 

 opposite sidC; whatever the subdivisions are into which 



