y no means follows that they are less important. On the
contrary, if the assailant succeed in gaining possession of one of these
centers of communication between the large valleys upon the line of
retreat of the enemy, it will be more serious for the latter than it
would be in an open country; since the occupation of one or two
difficult defiles will often be sufficient to cause the ruin of the
whole army.

If the attacking party have difficulties to overcome, it must be
admitted that the defense has quite as many, on account of the necessity
of covering all the outlets by which an attack in force may be made
upon the decisive points, and of the difficulties of the transversal
marches which it would be compelled to make to cover the menaced points.
In order to complete what I have said upon this kind of marches and the
difficulties of directing them, I will refer to what Napoleon did in
1805 to cut off Mack from Ulm. If this operation was facilitated by the
hundred roads which cross Swabia in all directions, and if it would have
been impracticable in a mountainous country, for want of transversal
routes, to make the long circuit from Donauwerth by Augsburg to
Memmingen, it is also true that Mack could by these same hundred roads
have effected his retreat with much greater facility than if he had been
entrapped in one of the valleys of Switzerland or of the Tyrol, from
which there was but a single outlet.

On the other hand, the general on the defensive may in a level country
concentrate a large part of his forces; for, if the enemy scatter to
occupy all the roads by which the defensive army may retire, it will be
easy for the latter to crush these isolated bodies; but in a very
mountainous country, where there are ordinarily but one or two principal
routes into which other valleys open, even from the direction of the
enemy, the concentration of forces becomes more difficult, since serious
inconveniences may result if even one of these important valleys be not
observed.

Nothing can better demonstrate the difficulty of strategic defense in
mountainous regions than the perplexity in which we are involved when we
attempt simply to give advice in such cases,--to say nothing of laying
down maxims for them. If it were but a question of the defense of a
single definite front of small extent, consisting of four or five
converging valleys, the common junction of which is at a distance of two
or three short marches from the su