possible to prevent the selection of incidents to head off any other set of incidents selected to prove anything else? Thus, in the first case, by dwelling upon the resemblances in the tactics at Naupaktis and the Yalu, and fortifying it with other instances, might not a man prove (with plenty of show of reasoning) that tactics do not change, and that, by studying history carefully, a modern admiral would be fully equipped for war? The case is extreme, of course: still in both battles there was the stronger force on the defensive, and defeat was brought about in either case by the loss of cohesion in this formation. More, the statement would have truth enough in it, but he who would propose our study of modern tactics to be based on history would be swiftly accused of landing us into quagmires. And if this be true of tactics, what assurance have we that it is not true of strategy also—strategy that fades imperceptibly into tactics? Take the main objections:
(1) Our history may be incorrect.
(2) Our theorist may select his instances by a process of eliminating any facts that go to contradict his pet theories, and it must be borne in mind that the average naval officer has neither the time nor the qualifications to study history for himself enough to say whether this method has been followed or not.
(3) There is no theory, based on history, that cannot be plausibly upset by a judicious selection of contra-
