relatively less as time went on, and the part assigned to others became correspondingly greater. There is some evidence to show that when writing began on the first volume of the Central American History in 1873, the director of the project actually had in mind the plan which he gives in the Literary Industries, that of writing with the aid of assistants who were to be responsible for "the study and reduction of certain minor sections" which he was to "employ" in his own writing. Thus we find, according to the information left by Nemos, that Bancroft actually wrote half of the volume, that Oak at first took out notes, and that Nemos prepared his work in the rough, leaving a considerable part of it to be rewritten. For the next volume undertaken, the first of the six on Mexico, we see that the chief was unable to prepare so much material in its final form, and rested with but two chapters completely to his credit, together with the rewriting of part of Nemos' work on the remainder. In four or five years, he expresses the determination of writing what he can himself and leaving the rest to his aids. This as we shall see amounted in the end to his doing about one seventh of the history, slightly revising the work of the other authors, often by the aid of critics in his employ, and preparing most of the material for the supplementary volumes.
Thus it came about that the original plan, the plan as published, was exactly reversed, and instead of Mr. Bancroft's doing all the work in final form, except some minor sections assigned to those whom he called his assistants, it was the so-called assistants who really wrote the History of the Pacific States, and Mr. Bancroft who did a few minor, or at any rate less difficult parts. Nor is it at all true, as one authority has said (Appleton's Encyclopædia of American Biography, I, 156), that Mr. Bancroft wrote the most important chapters. Of course, the sur-
