viii PREFACE
of the age would have to be admitted, such as the functions of the main priests and some festivals or ritual practices. Again, certain names of perhaps purely mythological figures might have to be mentioned. The evidence is occasionally insufficient to show whether a name represents an actual historical personage: a demon or a mythical hero or priest may be meant. An undoubted demon may even have to be included, such as the one that is supposed to cause eclipses, because he belongs to the domain of primitive astronomy.
Chronological Limits.—The period which the book was intended to embrace had been decided at the outset as that of the Vedas and Brāhmaṇas. The upper limit here is the date of the oldest hymns of the Rigveda. That date is uncertain, but my conviction (set forth in my History of Sanskrit Literature, pp. 11-12) that it is not much earlier than 1200 B.C. still remains unshaken. It does not appear to me to be in the slightest degree invalidated by Professor Hugo Winckler's discoveries at Boghaz-köi, in Asia Minor, in the year 1907. That scholar has deciphered, in an inscription of about 1400 B.C. found there, the names of certain deities as mi-it-ra, uru-w-na, in-da-ra, and na-ša-at-ti-ia, which correspond to those of the important Vedic gods Mitra, Varuna, Indra, and Nasatya. Three inferences may be drawn from the occurrence of these names. They may have been derived from Vedic India ; in that case the Vedic religion must have flourished in India considerably before 1400 B.C., even though the hymns that have come down to us may not have been composed before that date. But that these names should have travelled all the way from India to Asia Minor is a hypothesis so highly improbable that it may be dismissed. Secondly, the names may belong to the early Iranian period after the Iranians had separated from the Indians, but before their language had reached the phonetic stage of the Avesta. This seems the most probable theory, both chronologically and geographically. It implies only that the Indian branch had separated from the Iranian, not that it had already entered
