Observe that, for the sake of argument, I am accepting as historically accurate Sir John French's statement about the advantage possessed by the Boers owing to the release of their prisoners. It is almost superfluous to add that the statement, in the sense he uses it, has no historical foundation. The truth is exactly the opposite. The advantage was immensely on our side. The Boers took many thousands of British prisoners, but permanently retained none, because they had no means of retaining them. During the last year of the war prisoners were released on the spot. A large proportion of these men fought again, some several times. No Boer prisoner of war--that is, captured in action--was released. In December, 1900, we had about 15,000 in our possession; in May, 1902, about 50,000. It was mainly by this attrition of the Boer forces that we reduced them to submission. The element of historical truth in Sir John French's proposition is this: that in 1900, after the fall of Bloemfontein, a considerable number of Boers surrendered voluntarily, not in action, and were dismissed to their farms under a pledge not to fight again--a pledge which they broke, under circumstances into which we need not enter. There are no exact statistics as to the numbers of these men, but at an outside estimate they cannot have amounted to more than 5 per cent. of the total number of Boers engaged in the war. In any case, the point is totally irrelevant to the question of shock-tactics. That is a question of combat, and in combat, as Sir John French is aware, the Boers were, nine times out of ten, greatly outnumbered. Such are Sir John French's reasons for the failure of the lance and sword in South Africa. They constitute an instructive revelation of the mental attitude of the advocates of those weapons. Is it not plain that we are dealing here with a matter of faith, not of reason; of dogma, not of argument; of sentiment, not of technical practice? The simple technical issue--what happens in combat?--is persistently evaded, and refuge sought in vague and inaccurate generalizations, which, when tested, turn out to throw no light upon the controversy.