34 ANIMAL PHYSIOLOGY. 
edge there can be no perfect classification of objects; hence, 
any classification must be regarded but as the temporary creed 
of science, to be modified with the extension of knowledge. As 
a matter of fact, this has been the history of all zodlogical and 
other systems of arrangement. The only purpose of grouping 
is to simplify and extend knowledge; this being the case, it fol- 
lows that a method of grouping that accomplishes this has 
value, though the system may be artificial that is based on 
resemblances which, though real and constant, are associated 
with differences so numerous and radical that the total amount 
of likeness between objects thus grouped is often less than the 
difference. Such a system was that of Linnzeus, who classified: 
plants according to the number of stamens, etc., they bore. 
Seeing that animals which resemble each other are of com- 
mon descent from some earlier form, to establish the line of de- 
scent is to determine in great part the classification. Much as- 
sistance in this direction is derived from embryology, or the 
history of the development of the individual (ontogeny); so 
that it may be said that the ontogeny indicates, though it does 
not actually determine, the line of descent (phylogeny) ; and it 
is owing to the importance of this truth that naturalists have 
in recent years given so much attention to comparative embry- 
ology. 
It will be inferred that a natural system of classification must 
be based both on function and structure, though chiefly on the 
latter, since organs of very different origin may have a similar 
function; or, to express this otherwise, homologous structures 
may not be analogous ; and homology gives the better basis for 
classification. To illustrate, the wing of a bat and a bird are 
both homologous and analogous; the wing of a butterfly is 
analogous but not homologous with these; manifestly, to clas- 
sify bats and birds together would be better than to put birds 
and insects in the same group, thus leaving other points of re- 
lationship out of consideration. 
The broadest possible division of the animal kingdom is into 
groups, including respectively one-celled and many-celled 
forms—i. e., into Protozoa and Metazoa. As the wider the 
grouping the less are differences considered, it follows that the 
more subdivided the groups the more complete is the informa- 
tion conveyed: thus, to say that a dog is a metazoan is to con- 
vey a certain amount of information; that it isa vertebrate, 
more; that it isa mammal, a good deal more, because each of 
the latter terms includes the former. 
