INTRODUCTION. th 
will, however, be practically good or bad, just as this single 
character expresses a genuine and fundamental distinction, or 
is of a merely trivial and superficial nature. The instance 
here chosen will serve to illustrate either case. If we insist 
upon the fact that all the four legs must be externally visible, 
unmistakable legs, never fewer in number than four, then our 
classification is a very bad one, in fact entirely “artificial.” In 
this case our group of ‘quadrupeds” will comprise only the 
ordinary four-legged mammals, such as oxen, sheep, horses, and 
suchlike—together with the very dissimilar groups of the four- 
legged reptiles and amphibians, such as tortoises, lizards, 
crocodiles, frogs, and newts. Now these different animals 
have certainly much in common, but we are not justified in 
placing them together simply upon the ground that they have 
four conspicuous legs, unless we are willing to put in a vast 
number of other animals as well. We must, in fact, put ina 
great number of animals which are not quadrupeds in the sense 
that they have four legs, but which agree with those that have 
four legs in the other fundamental and essential points of their 
structure. In this way we may arrive at a very genuine and 
natural classification by making some concessions. We must 
allow, for instance, that two of the legs or limbs, ceasing to be 
fit for walking, may be converted into organs of flight, or wzngs. 
This will let in the birds. We must allow, again, that all the 
limbs may be converted into fizs. This admits most of the 
fishes. We must further grant that two of the legs may be 
altogether absent, whilst the remaining two are converted into_ 
swimming-paddles. This will bring inthe whales and dolphins. 
Lastly—and this is the greatest admission of all—we must allow 
the total absence of a// the limbs, provided the animal only 
show those other essential characters which are invariably found 
to go along with the possession of four legs in the regular quad- 
rupeds. This will bring in the snakes and some of the fishes. 
So that, paradoxical as it may seem, it is in one sense scientifi- 
cally correct to speak of a snake as a quadruped, though in 
reality it has no legs at all. In other words, there is no reason 
why a snake should not some day be found with four legs, and 
in point of fact some snakes show rudiments of these append- 
ages. Making these allowances, and some more of a similar 
nature, we may ultimately succeed in converting our division of 
Quadrupeds into a strictly scientific group, comprising the 
