from the class of possible facts (to which we must always confine ourselves). We say a relatively short length of the river , by which we mean a length which does not very much exceed that which the same number of troops would usually occupy on an ordinary position in line without a river. Such cases, we say, do not occur, and every direct defence of a river always becomes a kind of cordon system, at least as far as regards the extension of the troops, and therefore is not at all adapted to oppose a turning movement on the part of the enemy in the same manner which is natural to an army in a concentrated position. Where, therefore, such turning movement is possible, the direct defence of the river, however promising its results in other respects, is a measure in the highest degree dangerous.
Now, as regards the portion of the river between its extreme points, of course we may suppose that all points within that portion are not equally well suited for crossing. This subject admits of being somewhat more precisely determined in the abstract, but not positively fixed, for the very smallest local peculiarity often decides more than all which looks great and important in books. Besides, it is wholly unnecessary to lay down any rules on this subject, for the appearance of the river, and the information to be obtained from those residing near it, will always amply suffice, without referring back to books.
As matters of detail, we may observe that roads leading down upon a river, its affluents, the great towns through which it passes, and lastly above all, its islands, generally favour a passage the most; that on the other hand, the elevation of one bank over another, and the bend in the course of the river at the point of passage, which usually act such a prominent rôle in books, are seldom of any consequence. The reason of this is, that the presumed influence of these two things rests on the limited idea of an absolute defence of the river bank—a case which seldom or never happens in connection with great rivers.