chine, and if
they are only transmitted to following generations so as to become
_permanent_ modifications, they will be most important agencies in the
machine building. If, on the other hand, they are not transmitted by
heredity, they can have no permanent effect. We have here thus again the
problem of the inheritance of acquired characters. We have already
noticed the uncertainty surrounding this subject, but the almost
universal belief in the inheritance of such characters requires us to
refer to it again. It is uncertain whether such direct effects have any
influence upon the offspring, and therefore whether they have anything
to do with this machine building. Still, there are many facts which
point strongly in this direction. For example, as we study the history
of the horse family we find that an originally five-toed animal began to
walk more and more on its middle toe, in such a way that this toe
received more and more use, while the outer toes were used less and
less. Now that such a habit would produce an effect upon the toes in any
generation is evident; but apparently this influence extended from
generation to generation, for, as the history of the animals is
followed, it is found that the outer toes became smaller and smaller
with the lapse of ages, while the middle one became correspondingly
larger, until there was finally produced the horse with its one toe only
on each foot. Now here is a line of descent or machine building in the
direct line of the effects of use and disuse, and it seems very natural
to suppose that the modification has been produced by the direct effect
of the use of the organs. There are many other similar instances where
the line of machine building has been quite parallel to the effects of
use and disuse. If, therefore, acquired characters can be inherited to
_any_ extent, we have, in the direct influences of the environment an
important agency in machine building. This direct effect of the
conditions is apparently so manifest that one school of biologists finds
in it the chief cause of the variations which occur, telling us that the
conditions surrounding the organism produce changes in it, and that
these variations, being handed down to subsequent generations,
constitute the basis of the development of the machine. If this factor
is entirely excluded, we are driven back upon the natural selection of
congenital variations as the only kind of variations which can
permanently effect th