No one would claim that the rules by which scientific historians of to-day are judged should be applied to those that wrote when history was young, when the boundaries between the possible and the impossible were less clearly defined, or when, in fact, such boundaries hardly existed in men's minds. In this connection, even while we vaunt, we smile. After all, how much of our modern and so-called scientific history must strike the reasoning reader as mere theorizing or as special pleading based upon the slenderest evidence! Among the ancients the work of the historians whom we consider trustworthy--such writers, for instance, as Cæsar, Thucydides, Xenophon, Polybius, and Tacitus--may be said to fall generally within Rawlinson's canons 1 and 2 of historical criticism--that is, (1) cases where the historian has personal knowledge concerning the facts whereof he writes, or (2) where the facts are such that he may reasonably be supposed to have obtained them from contemporary witnesses. Canon 2 might be elaborated and refined very considerably and perhaps to advantage. It naturally includes as sources of knowledge--first, personal interviews with contemporary witnesses; and, second, accesses to the writings of historians whose opportunities brought them within canon 1. In this latter case the evidence would be less convincing, owing to the lack of opportunity to cross-question, though even here apparent lack of bias or the existence of biased testimony on both sides, from which a judicious man might have a fair chance to extract the truth, would go far to cure the defect.