nally taking leave of the difficulty
from the uselessness of incipient organs, as this difficulty has been
presented, in varying degrees of emphasis, by the Duke of Argyll, Mr.
Mivart, Professors Naegeli, Bronn, Broca, Eimer, and, indeed, by all
other writers who have hitherto advanced it. For, as thus presented, I
think I have shown that it admits of being adequately met. But now, I
must confess, to me individually it does appear that behind this
erroneous presentation of the difficulty there lies another question,
which is deserving of much more serious attention. For although it
admits of being easily shown--as I have just shown--that the difficulty
as ordinarily presented fails on account of its extravagance, the
question remains whether, if stated with more moderation, a real
difficulty might not be found to remain.

My quarrel with the conclusion, like my quarrel with the premiss, is due
to its universality. By saying in the premiss that _all_ incipient
organs are _necessarily_ useless at the time of their inception, these
writers admit of being controverted by fact; and by saying in the
conclusion that, _if_ all incipient organs are useless, it necessarily
follows that in _no_ case can natural selection have been the cause of
building up an organ until it becomes useful, they admit of being
controverted by logic. For, even if the premiss were true in
fact--namely, that all incipient organs are useless at the time of their
inception,--it would not necessarily follow that in no case could
natural selection build up a useless structure into a useful one;
because, although it is true that in no case can natural selection do
this by acting on a useless structure _directly_, it may do so by acting
on the useless structure _indirectly_, through its direct action on some
other part of the organism with which the useless structure happens to
be correlated. Moreover, as I believe, and will subsequently endeavour
to prove, there is abundant evidence to show that incipient characters
are often developed to a large extent by causes other than natural
selection (or apart from any reference to utility), with the result that
some of them thus happen to become of use, when, of course, the supposed
difficulty is at an end.

But although it is thus easy to dispose of both the propositions in
question, on account of their universality, stated more carefully they
would require, as I have said, more careful consideration. Thus, if it