DIVERGENT EVOLUTION THROUGH SEGREGATION. 283 
NATURAL SELECTION WITH GREAT DIFFERENCE IN EXTERNAL CON- 
DITIONS NOT SUFFICIENT TO EXPLAIN DIVERGENT EVOLUTION. 
The insufficiency of natural selection without se-generation to account 
for divergent evolution in an area where the external conditions are 
nearly uniform may be admitted by some who will claim that the case 
is quite otherwise when a species ranges freely over an area in which 
it is subjected to strongly contrasted conditions. It may be claimed 
that diversity of natural selection resulting from a great difference in 
external nature is sufficient to account tor divergent evolution without 
any se-generation. 
In the discussion of this subject important light can be gained by 
referring to the experience of the breeder. This experience, in as far 
as it relates to the subject of separation in the production of divergent 
breeds, may be arranged under three heads: First, diversity of selee- 
tion without separation; second, separation without diversity of selee- 
tion; third, separation more or less complete with diversity of selection. 
As the full discussion of these points is impossible here, and as there 
is probably but little difference of opinion in regard to what the results 
would be, I shall content myself with a simple statement of what I 
believe the experience of breeders shows. Difference in the standards 
of selection without separation can avail nothing in creating diver- 
genceof types; while separation without difference in the standards of 
selection will avail something, though food and external conditions are 
kept the same; but to secure the greatest divergence in a given time, 
there must be both diversity of selection and complete separation. In 
the case of separation without diversity of selection, there is room for 
difference of opinion; for the examples that some would claim as prov- 
ing that there is often divergence without diversity of selection and 
without difference in external conditions may be attributed by others 
to unconscious selection. It is granted by everyone that no skill in 
selecting the animals that possess the desired qualities will have any 
effect in establishing a new breed unless the selected animals are pre- 
vented from breeding with others that are deficient in the desired qual- 
ities. We further find that while separation, isan absolutely essential 
condition for this divergence, diversity of selection is not so essential. 
This is illustrated in the case of the slightly different types that are 
presented by the wild cattle found in ‘the different parks of England,* 
a phenomenon which can hardly be attributed to any diversity in the 
environment. 
In artificial breeding universal experience teaches that variation and 
selection, without separation, do not produce divergence of races. The 
separate breeding of different classes of variation is a necessary condi- 
tion for the accumulation of divergent variation; and wherever the 
separate breeding of different classes of variation is secured there diver- 
* See Darwin’s “ Variation under Domestication,” chapter xv, second page. 
