way, can hardly be doubted; but they
probably die out even more rapidly than races propagated by grafts or
layers.

_Opinion of De Candolle._--De Candolle, whose opinion on a philosophical
question of this kind deserves the greatest attention, has observed, in
his Essay on Botanical Geography, that the _varieties_ of plants range
themselves under two general heads: those produced by external
circumstances, and those formed by hybridity. After adducing various
arguments to show that neither of these causes can explain the permanent
diversity of plants indigenous in different regions, he says, in regard
to the crossing of races, "I can perfectly comprehend without altogether
sharing the opinion, that, where many species of the same genera occur
near together, hybrid species may be formed, and I am aware that the
great number of species of certain genera which are found in particular
regions may be explained in this manner; but I am unable to conceive how
any one can regard the same explanation as applicable to species
which live naturally at great distances. If the three larches, for
example, now known in the world, lived in the same localities, I might
then believe that one of them was the produce of the crossing of the two
others; but I never could admit that the Siberian species has been
produced by the crossing of those of Europe and America. I see, then,
that there exist in organized beings, permanent differences which cannot
be referred to any one of the actual causes of variation, and these
differences are what constitute species."[834]

_Reality of species confirmed by the phenomena of hybrids._--The most
decisive arguments perhaps, amongst many others, against the probability
of the derivation of permanent species from cross-breeds, are to be
drawn from the fact alluded to by De Candolle, of species having a close
affinity to each other occurring in distinct botanical provinces, or
countries inhabited by groups of distinct species of indigenous plants;
for in this case naturalists, who are not prepared to go the whole
length of the transmutationists, are under the necessity of admitting
that, in some cases, species which approach very near to each other in
their characters, were so created from their origin; an admission fatal
to the idea of its being a general law of nature that a few original
types only should be formed, and that all intermediate races should
spring from the intermixture of those stocks.

T