60 CONTROLLED NATURAL SELECTION 
usually form specific characters and acquired 
characters do not, it follows that the char- 
acters to do with protection from enemies are 
likely to be very important ones. 
Much criticism has been levelled against natural selec- 
tion from a mathematical point of view, and especially as 
regards the swamping effects of chance. If chance kills 
off 950 of a 1000 individuals, then natural selection has 
50 instead of 1000 to work with. However, the effect 
must be the same, except that longer time will be re- 
quired for the origin of favourable variations. When, 
however, after perhaps many generations the favourable 
variation does arise and escape the hands of chance, then 
it will be one amongst 50 instead of one in 1000, and 
therefore less likely to be swamped. 
False examples of chance are often cited. The stepping 
of a horse on an ant’s nest is a typical one. The action 
of the horse may be the result of chance. But what can 
be said about the position of the nest? The mother ant 
in the past chose this particular spot for her nest. Other 
mothers chose in some cases no doubt places where horses 
are not likely to walk. Will not such mothers therefore 
be thus selected to carry on the race? 
In conclusion it may be said that, although 
the subject has been approached from several 
standpoints, reasoning has led in each case to 
the same two conclusions : 
1. Specific environments are more impor- 
tant than common environments as selecting 
agents. 
2. Species’ enemies are very important 
