Before we proceed, however, a few words may be said regarding the general nature of the historical account of Japan. In ancient times Japan possessed neither literary script nor a regular system of calendar and chronology. All events had to be transmitted from generation to generation by oral tradition. The use of writing was imported from China, probably one or two hundred years before Christ, but the general use of letters for the purpose of recording events dates from the fifth century A. D., while the compilation of the national annals began two centuries later still. But such historical records as were then compiled suffered almost total destruction a short time afterward by fire, and the oral records of remote antiquity must have already been greatly disfigured by omissions, errors, and confusion of facts. Regretting this, and perceiving that unless steps were then taken to correct the annals, subsequent generations would be without any trustworthy record of remote events, the Emperor Temmu ordered an eminent scholar, Hieda-no-Are, to prepare a brief chronicle of sovereigns and important events. Unfortunately the death of the emperor, which occurred in 689, interrupted this work. Some twenty years later the Empress Gemmyō instructed Ō-no-Yasumaro to continue the compilation of Hieda-no-Are's annals. The work thus completed in 712 is the "_Kojiki_" as we now possess it. It must be regarded as the most trustworthy record extant of the events of ancient times. Eight years after the appearance of the "_Kojiki_," or in 720, the "_Nihongi_" was compiled. We find, on comparing these works, that although, on the whole, they agree, certain discrepancies exist between them. In these two works, however, is found the chief material for the reconstruction of the history of ancient Japan. If curious supernatural incidents figure in their pages, it should be remembered that literature being then in its infancy and a long interval having elapsed since the time of many of the events recorded, the annalists were untrained in the selection of matters worthy of a place in authentic history, while, at the same time, in the oral traditions on which they relied, errors had doubtless been included, and, ordinary events drifting out of sight, extraordinary incidents and supernatural stories had alone survived. As regards the ancient chronology of Japan it is recorded that almanacs first came into use in Japan in 604 A. D., although it appears that the Chinese calendar had been imported about the middle of the preceding century. The compilation of annals, as already stated, had preceded the latter event by a considerable interval. We may therefore conjecture that some method of reckoning months and years had been practiced from an early era, but no certain knowledge of this matter is available.