have been observed and noted very carefully at the time. It frequently
happens that a variation occurs, but the persons who notice it do not
take any care in noting down the particulars, until at length, when
inquiries come to be made, the exact circumstances are forgotten; and
hence, multitudinous as may be such "spontaneous" variations, it is
exceedingly difficult to get at the origin of them.

The second case is one of which you may find the whole details in the
"Philosophical Transactions" for the year 1813, in a paper communicated
by Colonel Humphrey to the President of the Royal Society,--"On a new
Variety in the Breed of Sheep," giving an account of a very remarkable
breed of sheep, which at one time was well known in the northern states
of America, and which went by the name of the Ancon or the Otter breed
of sheep. In the year 1791, there was a farmer of the name of Seth
Wright in Massachusetts, who had a flock of sheep, consisting of a ram
and, I think, of some twelve or thirteen ewes. Of this flock of ewes,
one at the breeding-time bore a lamb which was very singularly formed;
it had a very long body, very short legs, and those legs were bowed! I
will tell you by-and-by how this singular variation in the breed of
sheep came to be noted, and to have the prominence that it now has. For
the present, I mention only these two cases; but the extent of variation
in the breed of animals is perfectly obvious to any one who has studied
natural history with ordinary attention, or to any person who compares
animals with others of the same kind. It is strictly true that there
are never any two specimens which are exactly alike; however similar,
they will always differ in some certain particular.

Now let us go back to Atavism,--to the hereditary tendency I spoke of.
What will come of a variation when you breed from it, when Atavism
comes, if I may say so, to intersect variation? The two cases of which