need
not be absolutely simultaneous. Many variations, however, are from the
first connected by the law of correlation. Hence it follows that even
closely-allied species rarely or never differ from each other by some one
character alone; and this same remark applies to a certain extent to
domestic races; for these, if they differ much, generally differ in many
respects.

Some naturalists boldly insist[928] that species are absolutely distinct
productions, never passing by intermediate links into each other; whilst
they maintain that domestic varieties can always be connected either with
each other or with their parent-forms. But if we could always find the
links between the several breeds of the dog, horse, cattle, sheep, pigs,
&c., the incessant doubts whether they are descended from one or several
species would not have arisen. The greyhound genus, if such a term may be
used, cannot be closely connected with any other breed, unless, perhaps, we
go back to the ancient Egyptian monuments. Our English bulldog also forms a
very distinct breed. In all these cases crossed breeds must of course be
excluded, for the most distinct natural species can thus be connected. By
what links can the Cochin fowl be closely united with others? By searching
for breeds still preserved in distant lands, and by going back to
historical records, tumbler-pigeons, carriers, and barbs can be closely
connected with the parent rock-pigeon; but we cannot thus connect the
turbit or the pouter. The degree of distinctness between the various
domestic breeds depends on the amount of modification which they have
undergone, and especially on the neglect and final extinction of the
linking, intermediate, and less valued forms.

It has often been argued that no light is thrown, from the admitted changes
of domestic races, on the changes which natural species are believed to
undergo, as the former are said to be mere temporary productions, always
reverting, as soon as they become feral, to their pristine form. This
argument has been well combated by Mr. Wallace;[929] and full details were
given in the thirteenth chapter, showing that the tendency to reversion in
feral {416} animals and plants has been greatly exaggerated, though no
doubt to a certain extent it exists. It would be opposed to all the
principles inculcated in this work, if domestic animals, when exposed to
new conditions and compelled to struggle for their own wants against a host
of foreign c