The explanation is that one of the opposing Cavalry masses is assumed to have been already absolutely and completely overthrown--that is, during the pre-battle reconnaissance phase, whose central incident, as described in pp. 192-194 and 200-212 of the Manual, and criticized by me in the last chapter, is the great shock-duel of the two Independent Cavalries--a duel which is to result in the annihilation of one side or the other, and to which I shall have to return once more in the next chapter. The thread is resumed on p. 224 with the words, "Once the Independent Cavalry has defeated its opponent," etc., and from that point onwards nothing is heard of the hostile Independent Cavalry. The explanation of Sir John French's expression is the same. On p. xv he, too, assumes that _before the battle_ the hostile Cavalry has been disposed of, and says, somewhat vaguely, that the "true rôle of Cavalry on the battle-field is to reconnoitre, to deceive, and finally to support"--functions which he distinctly suggests should be carried out mainly through fire-action by troops "accustomed to act in large bodies dismounted." And we seem to recognize this view in the functions outlined in the Manual on p. 225. "Reconnoitre," it is true, disappears. We find no echo of von Bernhardi's chimerical conception of a double reconnaissance, distant and close, along the whole battle-front; nor, we may add, of his injunction to "occupy defiles and other important places to the flanks and front" of the Army. The rôles suggested for the flank Cavalry mass are: 1. To "act against the enemy's flanks." 2. To combine _fire_ concentrically with the main attack. 3. To pursue on parallel lines--a function which it is laid down on p. 229 is to be performed mainly with the rifle. 4. To force the enemy away from his direct line of retreat; which is merely a corollary of No. 3.