duction of this estimate. No probable alteration of the estimate of the extent of the Hadha-mãthric Nasks, which is the most uncertain, would materially affect the total.
Another matter of interest to the readers of translations from the Pahlavi, especially to those who are aware of the ambiguities of the original text, is the degree of confidence they can place in the correctness of the translation. In the case of the Dînkard it is fortunately possible to consult manuscripts written in Persia, and descended through only four or five intermediate copies from the work of the original writer, so that the text is remarkably free from copyists' errors. The eighth and ninth books also contain very few of those involved sentences, with long parenthetical clauses, which, owing to the habitual absence or misplacement of stops, are very perplexing to a translator. The chief difficulties of the text arise from its synoptical character, and the consequent want of connection between its sentences; there being often too little context to define the meaning of a doubtful word. The number of words of doubtful meaning in Pahlavi is, however, fast diminishing, in proportion to the advancing study of the texts; and the certainty of a translator, as to the correctness of his work, is increasing in a like proportion. At any rate, the reader may safely rely upon the general accuracy of these translations, even if a few errors should hereafter be discovered.
As an instance of such possible errors I will here correct one that exists in my translation of the Epistles of Mânûskîhar, which was pointed out to me by Môbad Tehmuras Dinshawji Ankalesaria, in a letter dated 28th October, 1887. In Ep. II, ii, 9–11, there occurs an illustration of what should be done when commentators differ, derived from the use that can be made of different observations of the stars, and containing three names that were difficult to identify. These names were doubtfully read as corruptions of the names of three of the lunar mansions, but it now appears that they were the names of three sets of astronomical tables (zîk); so that Shatro-ayârân, Hindûk, and Ptolemêôs should be read, instead of Satvâharân, Avênak, and Padramgôs; both sets of readings expressing the same
