302 
Proceedings of the Royal Society 
fication being granted, and the nnsatisfactoriness of existing and 
proposed schemes of classification being shown, we must now pro- 
ceed to consider the desiderata of a system of classification. Our 
classification must be natural, not artificial ; must be capable of 
complete specialisation, so as to include the minutest details, and 
capable, too, of the widest generalisation ; it must be universal in 
application, and it must be, as far as possible, simple of under- 
standing, and convenient in use. But how shall we obtain such a 
classification ? What phenomena in the whole field of human 
knowledge have as yet been classified in this way A moment's 
consideration will show that it is biological science which alone 
answers the question satisfactorily. ATone of the preliminary 
sciences is wanting in order or definiteness, but to biology above all is 
presented the problem of an innumerable multitude of actual pheno- 
mena demanding arrangement. Take an instance from zoology. Birds 
and reptiles, fishes and worms, are groups which the common sense 
of probably every rational human being enables him to form with 
considerable approach to correctness ; yet at this point the task of 
the zoologist is only beginning. He has to work in two directions, 
to specialise until every member of these groups is known in the 
greatest detail, and also to generalise these groups into larger and 
larger ones. The two lines of research may be represented thus : — 
letting the dots represent the details of the various groups, and 
the large rectangles the successive generalisations which combine 
model might often he constructed. A geologist, for instance, by piling map 
upon map of a given island at successive times (the margin being of course 
removed) would thus construct a solid model which would clearly exhibit the 
changes throughout the whole period. Where the area was increasing per unit 
time, the solid would widen upwards and overhang its base ; where decreasing 
it would narrow, and thus even the minutest local increase or decrease would 
be represented with extreme vividness. 
