ctual approximation of these
parts before their union; or whether, as in yet other cases, the
integration is of that indirect kind which arises when, out of a number
of like organs, one, or a group, discharges an ever-increasing share of
the common function, and so grows while the rest dwindle and
disappear;--the general fact remains the same, that there is a tendency
to the unification of parts having similar duties.

The tendency, however, acts under limiting conditions; and recognition
of them will explain some apparent exceptions. In the human foetus, as
in the lower vertebrata, the eyes are placed one on each side of the
head. During evolution they become relatively nearer, and at birth are
in front; though they are still, in the European infant as in the adult
Mongol, proportionately further apart than they afterwards become. But
this approximation shows no signs of further increase. Two reasons
suggest themselves. One is that the two eyes have not quite the same
function, since they are directed to slightly-different aspects of each
object looked at; and, since the resulting binocular vision has an
advantage over monocular vision, there results a check upon further
approach towards identity of function and unity of structure. The other
reason is that the interposed structures do not admit of any nearer
approach. For the orbits of the eyes to be brought closer together,
would imply a decrease in the olfactory chambers; and as these are
probably not larger than is demanded by their present functional
activity, no decrease can take place. Again, if we trace up the external
organs of smell through fishes,[7] reptiles, ungulate mammals and
unguiculate mammals, to man, we perceive a general tendency to
coalescence in the median line; and on comparing the savage with the
civilized, or the infant with the adult, we see this approach of the
nostrils carried furthest in the most perfect of the species. But since
the septum which divides them has the function both of an evaporating
surface for the lachrymal secretion, and of a ramifying surface for a
nerve ancillary to that of smell, it does not disappear entirely: the
integration remains incomplete. These and other like instances do not
however militate against the hypothesis. They merely show that the
tendency is sometimes antagonized by other tendencies. Bearing in mind
which qualification, we may say, that as differentiation of parts is
connected with difference of func