only way to steer clear of both these mistakes will be, then, immediately one has chosen one's historical epoch, to study thoroughly the interests which moved the three classes of society—the people, the nobility and royalty—at that time; to choose from among the principal personages of those classes such as took an active part in the events to be comprised in the narrative; and to enquire minutely concerning their appearance, character and temperament, so that, whilst making them live, speak, and act in this triple unity, one may show the development in these historical types, of the passions which brought about those catastrophes which are recorded in the pages of the century by dates and facts and in which one can only interest one's public by showing them the actual living manner in which the same deeds were added to history."
Such was Dumas's view of the romance in the lays of "Isabel de Bavière," and "La Comtesse de Salisbury." We have already explained how the former "chronique" came to be written. Dumas selected the most effective portions of Barante, and vivified them. He was destined in the future to make a brilliant success by the way in which he painted romance on a foundation of history; but on this occasion, as Mr Saintsbury pithily puts it, "the canvas shows through." There is a want of coher-
