criticism, in comparison with the servile and often mechanical work of pursuing statements to their root. For a historian has to be treated as a witness, and not believed unless his sincerity is established #61. The maxim that a man must be presumed to be innocent until his guilt is proved, was not made for him. For us, then, the estimate of authorities, the weighing of testimony, is more meritorious than the potential discovery of new matter #62. And modern history, which is the widest field of application, is not the best to learn our business in; for it is too wide, and the harvest has not been winnowed as in antiquity, and further on to the Crusades. It is better to examine what has been done for questions that are compact and circumscribed, such as the sources of Plutarch's Pericles, the two tracts on Athenian government, the origin of the epistle to Diognetus, the date of the life of St. Antony; and to learn from Schwegler how this analytical work began. More satisfying because more decisive has been the critical treatment of the medieval writers, parallel with the new editions, on which incredible labour has been lavished, and of which we have no better examples than the prefaces of Bishop Stubbs. An important event in this series was the attack on Dino Compagni, which, for the sake of Dante, roused the best Italian scholars to a not unequal contest. When we are told that England is behind the Continent in critical faculty, we must admit that this is true as to quantity, not as to quality of work. As they are no longer living, I will say of two Cambridge professors, Lightfoot and Hort, that they were critical scholars whom neither Frenchman nor German has surpassed.