It is clearly in his mind that, since the various corps or columns which are the objects of reconnaissance may be "advancing to battle" on a total front of 50 to 100 miles (this is his own estimate, p. 81), it will be advisable to explore their zones of approach at once. But there are other passages which support the opposite principle: for example, on page 15: "The circumstances of modern war demand that great masses of mounted men shall be used as Army Cavalry and concentrated in the decisive direction.... The front of the army, therefore, can never be covered throughout its entire length by the Army Cavalry," etc. On page 87 also he is quite decisive in the same sense: "The universal principle most always good for Cavalry, that when a decisive struggle is in prospect all possible strength must be concentrated for it"--an unexceptional truism, applicable as it stands to all struggles, great or small, by land or sea, but in its context only too suggestive of the gigantic shock-duel.[6] But on the whole he stands committed to nothing more definite than the following: "It remains for the leader to make his preparations in full freedom, and to solve the task confided to him in his own way." Profoundly true, but not very helpful in an instructional treatise on war. III.--DIVISIONAL RECONNAISSANCE. The chapter on "Divisional Reconnaissance" is still less intelligible. It would be interesting to know how Sir John French would sum up its "logical" and "convincing" doctrines. The divisional Cavalry are in all cases to "cleave to the Infantry" (p. 75) of their respective divisions, yet they are to take the place of the Army Cavalry "when a concentration of that force in a decisive direction takes place" (another hint of the gigantic preliminary shock-duel), and are even to indulge in "strategical exploration" (pp. 72-75). In fact, these amazing super-Cavalry are to perform physical feats in reconnaissance analogous to the feats designed for them in the pre-arranged battle of all arms (_vide_ p. 149). Yet they cannot "fight independently" even with the hostile divisional Cavalry, nor clear the way for their own patrols, nor find their own outposts (pp. 75-76). And then we come to a passage which, quite parenthetically and as it were by accident, throws a searching light upon the many dark places of this volume. The divisional Cavalry, _inter alia_, is to perform the "close reconnaissance along by far the greater part of the front of the army." But the close reconnaissance, owing to the range of modern firearms, is "considerably more difficult." "It thus becomes possible for the Cavalryman in general to get no closer to the enemy than his rifle will carry" (p. 80). "_His_ rifle," be it noted. And the hostile Cavalryman (surely an "enemy") is presumably in the same case. What, then, of the charging patrols and squadrons?