The attack on fortresses cannot of course come before us here in its aspect as a branch of the science of fortification or military works; we have only to consider the subject, first, in its relation to the strategic object with which it is connected; secondly, as regards the choice among several fortresses; and thirdly, as regards the manner in which a siege should be covered.
That the loss of a fortress weakens the defence, especially in case it forms an essential part of that defence; that many conveniences accrue to the assailant by gaining possession of one, inasmuch as he can use it for magazines and depôts, and by means of it can cover districts of country cantonments, etc.; that if his offensive at last should have to be changed into the defensive, it forms the very best support for that defensive—all these relations which fortresses bear to theatres of war, in the course of a war, make themselves sufficiently evident by what has been said about fortresses in the book on the Defence, the reflection from which throws all the light required on these relations with the attack.
In relation to the taking of strong places, there is also a great difference between campaigns which tend to a great decision and others. In the first, a conquest of this description is always to be regarded as an evil which is unavoidable. As long as there is yet a decision to be made, we undertake no sieges but such as are positively unavoidable. When the decision has been already given—the crisis, the utmost tension of forces, some time passed—and when, therefore, a state of rest has commenced, then the capture of strong places serves as a consolidation of the conquests made, and then they can generally be carried out, if not