the end of March; and this
shows that the period of hatching was inherited. But the grandchildren of
these Aylesbury ducks completely lost their early habit of incubation, and
hatched their eggs at the same time with the common ducks of the same
place.

Many cases of non-inheritance apparently result from the conditions of life
continually inducing fresh variability. We have seen that when the seeds of
pears, plums, apples, &c., are sown, the seedlings generally inherit some
degree of family likeness from the parent-variety. Mingled with these
seedlings, a few, and sometimes many, worthless, wild-looking plants
commonly appear; and their appearance may be attributed to the principle of
reversion. But scarcely a single seedling will be found {26} perfectly to
resemble the parent-form; and this, I believe, may be accounted for by
constantly recurring variability induced by the conditions of life. I
believe in this, because it has been observed that certain fruit-trees
truly propagate their kind whilst growing on their own roots, but when
grafted on other stocks, and by this process their natural state is
manifestly affected, they produce seedlings which vary greatly, departing
from the parental type in many characters.[65] Metzger, as stated in the
ninth chapter, found that certain kinds of wheat brought from Spain and
cultivated in Germany, failed during many years to reproduce themselves
truly; but that at last, when accustomed to their new conditions, they
ceased to be variable,--that is, they became amenable to the power of
inheritance. Nearly all the plants which cannot be propagated with any
approach to certainty by seed, are kinds which have long been propagated by
buds, cuttings, offsets, tubers, &c., and have in consequence been
frequently exposed during their individual lives to widely diversified
conditions of life. Plants thus propagated become so variable, that they
are subject, as we have seen in the last chapter, even to bud-variation.
Our domesticated animals, on the other hand, are not exposed during their
individual lives to such extremely diversified conditions, and are not
liable to such extreme variability; therefore they do not lose the power of
transmitting most of their characteristic features. In the foregoing
remarks on non-inheritance, crossed breeds are of course excluded, as their
diversity mainly depends on the unequal development of characters derived
from either parent, modified by the princ