d, and there is a tendency in both parents to
revert to some long-lost character, this tendency, for all that we can see
to the contrary, may be transmitted undiminished for an indefinite number
of generations. These two distinct cases of reversion are often confounded
together by those who have written on inheritance.

Considering, on the one hand, the improbability of the three assumptions
which have just been discussed, and, on the other hand, how simply the
facts are explained on the principle of reversion, we may conclude that the
occasional appearance in all the races, both when purely bred and more
especially when crossed, of blue birds, sometimes chequered, with double
wing-bars, with white or blue croups, with a bar at the end of the tail,
and with the outer tail-feathers edged with white, affords an argument of
the greatest weight in favour of the view that all are descended from
_Columba livia_, including under this name the three or four wild varieties
or sub-species before enumerated. {203}

To sum up the six foregoing arguments, which are opposed to the belief that
the chief domestic races are the descendants of at least eight or nine or
perhaps a dozen species; for the crossing of any less number would not
yield the characteristic differences between the several races. _Firstly_,
the improbability that so many species should still exist somewhere, but be
unknown to ornithologists, or that they should have become within the
historical period extinct, although man has had so little influence in
exterminating the wild _C. livia_. _Secondly_, the improbability of man in
former times having thoroughly domesticated and rendered fertile under
confinement so many species. _Thirdly_, these supposed species having
nowhere become feral. _Fourthly_, the extraordinary fact that man should,
intentionally or by chance, have chosen for domestication several species,
extremely abnormal in character; and furthermore, the points of structure
which render these supposed species so abnormal being now highly variable.
_Fifthly_, the fact of all the races, though differing in many important
points of structure, producing perfectly fertile mongrels; whilst all the
hybrids which have been produced between even closely allied species in the
pigeon-family are sterile. _Sixthly_, the remarkable statements just given
on the tendency in all the races, both when purely bred and when crossed,
to revert in numerous minute details of colou