226 | B. NATURAL HISTORY. 
7. On the Systematic Arrangement of the Reptilia. 
‘ 
a. On Systematic Classification in general. 
The rationale of systematic classification, in zodlogy at least, 
is a problem unsolved in the minds of many. As Agassiz has 
observed, it reposes in most cases on a purely empirical basis; 
and such are the difficulties that a resolution of its true nature 
presents, that some of the best naturalists have been fain to 
admit that it does not rest on any basis of principle of natural 
order, but on the convenience of the student alone. Yet I pre- 
sume that even these will hardly admit their position to be true, if 
brought face to face with such a legitimate deduction from it as 
that a classification based purely on coloration or size would be as 
satisfactory as that they adopt. Believing that a true classification 
of species of organic beings based on their structure will be the 
expression of some of the laws according to which their creation 
has been conducted, as well as of some of those which govern 
their mutual relations in the scenes of active life, I would propose 
to state the principle which I imagine to lie a the basis of a system 
which fulfils such requirements. 
In practice, so general is the coincidence of external and readily 
visible characters with the deeper and more significant ones, that 
the usual practice of arranging groups of animals in accordance 
with some readily observed tangible character of the former kind 
is generally justified by the more conclusive test of an examination 
of the whole structure. Yet this method fails to stand such tests 
sufficiently often to render it obvious that external characters are 
not enough for the resolution of the problem of aflinity, and that 
they may be deceptive in cases where we little suspect it. Asan ex- 
ample of the first, the genus Sphenvdon is sufficient. In characters 
usually employed by naturalists for distinguishing the families of 
Reptilia, it is an Agamoid Lacerfilian: a complete examination of 
its anatomy has shown that it is not even a member of the order 
Lacertilia. In the second case of deceptive characters, those of 
the corresponding genera of different homologous series may be 
mentioned, where the characters determining the series are rarely 
visible externally. 
Some valuable propositions respecting alaesiicalinn are made by 
Professor Gill, in an essay on the Mammalia, read before the Ameri- 
can Association (to be published in abstract in its volume, and in 
the “ American Naturalist ” for October, 1870). His first two prop- 
