t though it is quite true that all those groups we call species _may_,
according to the known laws of reproduction, have descended from a
single stock, and though it is very likely they really have done so, yet
this conclusion rests on deduction and can hardly hope to establish
itself upon a basis of observation. And the primitiveness of the
supposed single stock, which, after all, is the essential part of the
matter, is not only a hypothesis, but one which has not a shadow of
foundation, if by "primitive" be meant "independent of any other living
being." A scientific definition, of which an unwarrantable hypothesis
forms an essential part, carries its condemnation within itself; but
even supposing such a definition were, in form, tenable, the
physiologist who should attempt to apply it in Nature would soon find
himself involved in great, if not inextricable difficulties. As we have
said, it is indubitable that offspring _tend_ to resemble the parental
organism, but it is equally true that the similarity attained never
amounts to identity, either in form or in structure. There is always a
certain amount of deviation, not only from the precise characters of a
single parent, but when, as in most animals and many plants, the sexes
are lodged in distinct individuals, from an exact mean between the two
parents. And indeed, on general principles, this slight deviation seems
as intelligible as the general similarity, if we reflect how complex the
co-operating "bundles of forces" are, and how improbable it is that, in
any case, their true resultant shall coincide with any mean between the
more obvious characters of the two parents. Whatever be its cause,
however, the co-existence of this tendency to minor variation with the
tendency to general similarity, is of vast importance in its bearing on
the question of the origin of species.

As a general rule, the extent to which an offspring differs from its
parent is slight enough; but, occasionally, the amount of difference is
much more strongly marked, and then the divergent offspring receives the
name of a Variety. Multitudes, of what there is every reason to believe
are such varieties, are known, but the origin of very few has been
accurately recorded, and of these we will select two as more especially
illustrative of the main features of variation. The first of them is
that of the "Ancon," or "Otter" sheep, of which a careful account is
given by Colonel David Humphreys, F.R.S., i