Our conclusion, therefore, is, that a small post in mountains possesses great strength. How this may be of decisive utility in all cases which depend entirely on a relative defence is plain of itself; but will it be of the same decisive utility for the absolute defence by a whole army? This is the question which we now propose to examine.
First of all we ask whether a front line composed of several posts has, as has hitherto been assumed, the same strength proportionally as each post singly. This is certainly not the case, and to suppose so would involve one of two errors.
In the first place, a country without roads is often confounded with one which is quite impassable. Where a column, or where artillery and cavalry cannot march, infantry may still, in general, be able to pass, and even artillery may often be brought there as well, for the movements made in a battle by excessive efforts of short duration are not to be judged of by the same scale as marches. The secure connection of the single posts with one another rests therefore on an illusion, and the flanks are in reality in danger.
Or next it is supposed, a line of small posts, which are very strong in front, are also equally strong on their flanks, because a ravine, a precipice, etc., etc., form excellent supports for a small post. But why are they so?—not because they make it impossible to turn the post, but because they cause the enemy an expenditure of time and of force, which gives scope for the effectual action of the post. The enemy who, in spite of the difficulty of the ground, wishes, and in fact is obliged, to turn such a post, because the front is unassailable requires, perhaps, half-a-day to execute his purpose, and cannot after all accomplish it without some loss of men. Now if such a post can be succoured, or if it is only designed to resist for a certain space of time, or lastly, if it is able to cope with the enemy, then the flank supports have done their part, and we may say the position had not only a strong front, but strong flanks as well. But it is