wing case is
slightly different, but still shows the same principle: Naudin[199]
raised numerous hybrids between the yellow _Linaria vulgaris_ and the
purple _L. purpurea_, and during three successive generations the
colours kept distinct in different parts of the same flower.

From such cases as the foregoing, in which the offspring of the first
generation perfectly resemble either parent, we come by a small step to
those cases in which differently coloured flowers borne on the same
root resemble both parents, and by another step to those in which the
same flower or fruit is striped or blotched with the two parental
colours, or bears a single stripe of the colour or other characteristic
quality of one of the parent-forms. With hybrids and mongrels it
frequently or even generally happens that one part of the body
resembles more or less closely one parent and another part the other
parent; and here again some resistance to fusion, or, what comes to the
same thing, some mutual affinity between the organic atoms of the same
nature, apparently comes into play, for otherwise all parts of the body
would be equally intermediate in character. So again, when the
offspring of hybrids or mongrels, which are themselves nearly
intermediate in character, revert either wholly or by segments to their
ancestors, the principle of the affinity of similar, or the repulsion
of dissimilar atoms, must come into action. To this principle, which
seems to be extremely general, we shall recur in the chapter on
pangenesis.

It is remarkable, as has been strongly insisted upon by Isidore
Geoffroy St. Hilaire in regard to animals, that the transmission of
characters without fusion occurs most rarely when species are crossed;
I know of one exception alone, namely, with the hybrids naturally
produced between the common and hooded crow (_Corvus corone_ and
_cornix_), which, however, are closely allied species, differing in
nothing except colour. Nor have I met with any well-ascertained cases
of transmission of this kind, even when one form is strongly prepotent
over another, when two races are crossed which have been slowly formed
by man's selection, and therefore resemble to a certain extent natural
species. Such cases as puppies in the same litter closely resembling
two distinct breeds, are probably due to super-