Upon a closer inspection of the language and grammar of these mountain records of the Achæmenian dynasty, a curious fact came to light which seemed to disturb the historical relation between the language of Zoroaster and the language of Darius. At first, historians were satisfied with knowing that the edicts of Darius could be explained by the language of the Avesta, and that the difference between the two, which could be proved to imply a considerable interval of time, was such as to exclude for ever the supposed historical identity of Darius Hystaspes and Gushtasp, the mythical pupil of Zoroaster. The language of the Avesta, though certainly not the language of Zarathustra,[37] displayed a grammar so much more luxuriant, and forms so much more primitive than the inscriptions, that centuries must have elapsed between the two periods represented by these two strata of language. When, however, the forms of these languages were subjected to a more searching analysis, it became evident that the phonetic system of the cuneiform inscriptions was more primitive and regular than even that of the earlier portions of the Avesta. This difficulty, however, admits of a solution; and, like many difficulties of the kind, it tends to confirm, if rightly explained, the very facts and views which at first it seemed to overthrow. The confusion in the phonetic system of the Zend grammar is no doubt owing to the influence of oral tradition. Oral tradition, particularly if confided to the safeguard of a learned priesthood, is able to preserve, during centuries of growth and change, the sacred accents of a dead language; but it is liable at least to the slow and imperceptible influences of a corrupt pronunciation. Nowhere can we see this more clearly than in the Veda, where grammatical forms that had ceased to be intelligible, were carefully preserved, while the original pronunciation of vowels was lost, and the simple structure of the ancient metres destroyed by the adoption of a more modern pronunciation. The loss of the Digamma in Homer is another case in point. There are no facts to prove that the text of the Avesta, in the shape in which the Parsis of Bombay and Yezd now possess it, was committed to writing previous to the Sassanian dynasty (226 A.D.). After that time it can indeed be traced, and to a great extent be controlled and checked by the Huzvaresh translations made under that dynasty. Additions to it were made, as it seems, even after these Huzvaresh translations; but their number is small, and we have no reason to doubt that the text of the Avesta, in the days of Arda Viraf, was on the whole exactly the same as at present. At the time when these translations were made, it is clear from their own evidence that the language of Zarathustra had already suffered, and that the ideas of the Avesta were no longer fully understood even by the learned. Before that time we may infer, indeed, that the doctrine of Zoroaster had been committed to writing, for