cases are "embarrassing to his theory." If, he adds, the auxiliary force
were placed unreservedly at the disposal of the chief belligerent, the
problem would be simple enough. It would then, in effect, be the same thing
as unlimited war with the aid of a subsidised force. But in fact, as he
observes, this seldom happened, for the contingent was always more or less
controlled in accordance with the special political aims of the Government
which furnished it. Consequently, the only conclusion he succeeded in
reaching was that it was a form of war that had to be taken into account,
and that it was a form of limited war that appeared to differ essentially
from war limited by object. We are left, in fact, with an impression that
there must be two kinds of limited war.

But if we pursue his historical method and examine the cases in which this
nature of war was successful, and those in which it was unsuccessful, we
shall find that wherever success is taken as an index of its legitimate
employment, the practical distinction between the two kinds of limited war
tends to disappear. The indications are that where the essential factors
which justify the use of war limited by object are present in war limited
by contingent, then that form of war tends to succeed, but not otherwise.
We are brought, in fact, to this proposition, that the distinction "Limited
by contingent" is not one that is inherent in war, and is quite out of line
with the theory in hand--that, in reality, it is not a _form_ of war, but a
_method_ which may be employed either for limited or unlimited war. In
other words, war limited by contingent, if it is to be regarded as a
legitimate form of war at all, must take frankly the one shape or the
other. Either the contingent must act as an organic unit of the force
making unlimited war without any reservations whatever, or else it should
be given a definite territorial object, with an independent organisation
and an independent limited function.