10 INTRODUCTION 
proportioned to the nature of the barrier which sepa- 
rates the two regions. Transverse mountain chains, 
such as the Himalayas or Atlas, separate two different 
faunas. Similarly, lands separated by a broad and 
deep belt of ocean may have quite different faunas ; 
those separated by a shallow and narrow strait may 
have almost identical animals. Such facts led him to 
believe that each species originated in a particular 
region, and spread out from that region till it was 
stopped by some barrier to further distribution. The 
remarkable features of the animals of isolated regions 
can be explained on this hypothesis as the result of 
descent with modification from the original forms. 
One result of the direction thus given to the study of 
distribution has been to make the division of the globe 
into zoogeographical regions, and the study of the faunas 
of the different regions, of supreme importance, and it 
is this aspect which is studied in most books on the 
subject. But, obviously, another and more truly 
geographical point of view is possible. We may study 
the animals of the natural regions of the globe as form- 
ing a part of the features of those regions, and in this 
case attention is directed rather to the adaptations 
displayed by the animals, than to their zoological 
relationships. 
In deliminating the natural regions of the globe the 
botanists have played a large part, for they have shown 
that the existence of the great plant associations 
depends upon the climate, the relief, and so forth of 
the region. To the practised eye a collection of plants 
from e.g. the Mediterranean region gives a great deal 
of information in regard to the region. For example, 
the small silvery leaves of such plants, and the other 
devices for husbanding water, together with the occur- 
