THEORY OF REPRESENTATION. 237 
discovered, it was often most erroneously applied. The 
result of our researches in following up this law will 
now be given. 
(292.) No law of the natural system is more calcu- 
lated to keep in check the ardour of imagination than 
this. So numerous are the resemblances between ob- 
jects, that, without a better guide than the return of a 
series into itself, we may form circles ad infinitum 
under the idea that they are natural, when, in truth, they 
are artificial. We could even cite many instances where, 
by the help of much ingenuity, parallel relations of 
analogy between artificial groups have been made out, 
and where, in truth, the whole theory has been mis- 
applied. But when, superadded to these, we apply the / 
theory of representation in all its bearings, as a third 
test to the accuracy of our groups, it is next to impos- 
sible that we should err or violate the series of nature., 
It is, in fact, as we have elsewhere demonstrated*,’ 
* the only certain test of a natural group.” This will’ 
be evident when we exemplify the theory by a reference 
to acknowledged facts. 
(293.) The class of Birds, as being that which of all 
others in the animal kingdom has been most analysed, is, 
in consequence, best calculated for our present purpose. 
Every natural group, as we have seen (285.), contains re- 
presentations of the divisions composing a neighbouring 
group. Thus the tribes of the order Rasores { repre- 
sent, by analogy, the tribes of the order Znsessores ; and 
these tribes, in a similar way, represent the primary 
orders of birds. Now this principle pervades every 
natural group, whatever may be its value, or size, or 
denomination. It extends not only to orders, tribes, 
and families, but even to genera and sub-genera. So 
that, if a sub-genus is sufficiently numerous in species, 
it will contain types of representation of the remaining 
sub-genera composing the entire genus, and, conse- 
quently, of every natural division in the whole class of 
* North. Zool. + See Linn, Trans, vol. xvi. p. 45. 
