394 
DARWINISM 
CHAP. 
feet and teeth, as in the American fossil horses; or in the in¬ 
creased development of the branching horns, as in the true 
deer. In each of these cases specialisation and adaptation to 
the conditions of the environment appear to have reached their 
limits, and any change of these conditions, especially if it be 
at all rapid or accompanied by the competition of less developed 
but more adaptable forms, is liable to cause the extinction of 
the most highly developed groups. Such we know was the 
case with the horse tribe in America, which totally disappeared 
in that continent at an epoch so recent that we cannot lie 
sure that the disappearance was not witnessed, perhaps caused, 
by man; while even in the Eastern hemisphere it is the 
smaller species — the asses and the zebras — that have persisted, 
while the larger and more highly developed true horses have 
almost, if not quite, disappeared in a state of nature. So we 
find, both in Australia and South America, that in a quite 
recent period many of the largest and most specialised forms 
have become extinct, while only the smaller types have sur¬ 
vived to our day ; and a similar fact is to be observed in many 
of the earlier geological epochs, a group progressing and reach¬ 
ing a maximum of size or complexity and then dying out, 
or leaving at most but few and pigmy representatives. 
Cause of Extinction of Large A nimals. 
Now there are several reasons for the repeated extinction 
of large rather than of small animals. In the first place, 
animals of great bulk require a proportionate supply of food, 
and any adverse change of conditions would affect them more 
seriously than it would smaller animals. In the next place, 
the extreme specialisation of many of these large animals 
would render it less easy for them to be modified in any new 
direction suited to changed conditions. Still more important, 
perhaps, is the fact that very large animals always increase 
slowly as compared with small ones—the elephant producing 
a single young one every three years, while a rabbit may have 
a litter of seven or eight young two or three times a year. 
Now the probability of favourable variations will be in direct 
proportion to the population of the species, and as the smaller 
animals are not only many hundred times more numerous than 
the largest, but also increase perhaps a hundred times as 
