All who have written on the subject are agreed that the Hâthi Gumpha or Elephant Cave, is the oldest that exists in these hills. It is, however, only a natural cavern of considerable extent, which may have been slightly enlarged by art, though there is no distinct evidence that this was so. At all events there is certainly no architectural moulding or form, to show that it was ever occupied by man and not by wild animals only, except a long inscription in 17 lines engraved on the smoothed brow of the rock above it. It is consequently of no value whatever in an architectural object, and from an archaeological point of view its whole interest resides in the inscription, which, so far as is at present known is the earliest that has yet been found in India.

A very imperfect attempt to copy this inscription accompanies Mr. Stirling's paper on Cuttack in the 15th volume of the Asiatic Researches, but so badly done as to be quite illegible. The first real copy was made by Lieutenant Kittoe in 1837, and though only an eye sketch was done with such marvellous exactness, that Mr. Prinsep was enabled to make a very correct translation of the whole, which he published in the sixth volume of the Bengal Asiatic Journal (pp. 1080 et seq.). From the more matured and priestly style of composition with which it commences, he was inclined to consider it more modern than the edicts of Aśoka, and assumed the date to be about 200 B.C., a date which I, and everyone else, was at the time, led to adopt in deference to the opinion of so distinguished a scholar. It has since, however, been more carefully re-examined by Babu Rajendralàla Mitra, by personal inspection on the spot, and with the aid of photographs. For reasons which seem to me sufficient to establish his conclusion, he places it about a century earlier, B.C. 300 or 325. One of the more important data for the earlier date is the occurrence in the 12th line of the name of Nanda, king of Magadha, of which Mr. Prinsep does not seem to have been aware; and as it is used apparently in the past tense, it looks as if the king Aira who caused this inscription to be written, came after these predecessors of the Mauryan dynasty. It may, however, be that he was only contemporary with the Nandas and with the first Mauryan kings. At the same time all the historical allusions which this inscription contains seems to show that he must have lived before the time when Aśoka carved his edicts at Dauli.

The Hâthi Gumpha inscription represents the king as oscillating between the Brahmanical and the Buddhist forms of faith, and though he finally settled down to the latter belief, the whole tenor of the narrative is such, that we are led to believe that the Brahmanical was the prevalent faith of the country, and that he was, if not the first, at least one of the earliest converts to Buddhism. This could hardly have been the case had Aśoka's inscriptions at Dauli—almost in sight of this cave—been in existence when it was engraved, and he could hardly have failed to allude to so powerful an emperor, had he ruled in Orissa before his time. Altogether, it seems from the contents of the inscription so much more probable that Aira should have ruled before the rise of the great Maurya dynasty, than after their establishment, that I feel very little hesitation in coming to the conclusion that 300 B.C., or thereabouts, is the most probable date for this inscription.1

In so far as the history of cave architecture is concerned, the determination of the age of this inscription is only a political question, not affecting the real facts of the case. As it is avowedly the earliest thing here, if its date is 200 B.C., all the caves that show marks of the chisel are more modern, and must be crowded into the period between that date, and the epoch at which it can be ascertained that the most modern were excavated. If, on the other hand, its date is about 300 B.C., it allows time for our placing the oldest and simplest caves as contemporary with those just described in Behar, and allows ample time for the gradual development of the style in a manner more in conformity with our experience of cave architecture in the west of India.

Though I am myself strongly of opinion that the true date of this inscription is about 300 B.C., the question may very well be left for future consideration. The important lessons we are taught by the peculiarities of the Hâthi Gumpha are the same that we gathered from the examination of those in Behar. It is that all the caves used by the Buddhists, or held sacred by them anterior to the age of Aśoka, are mere natural caverns unimproved by art. With his reign the fashion of chiselling cells out of the living rock commenced, and was continued with continually increasing magnificence and elaboration for nearly 1,000 years after his time.

Before proceeding to describe the remaining excavations in these hills, it may be as well to advert to a peculiarity we learn as much from the sculptures of the Bharhut Tope as from the caves of Behar. It is that during the reign of Aśoka, and for 100 years afterwards, it was the fashion to add short inscriptions to everything. Not only as already pointed out are all the Behar caves inscribed, but almost all the Bharhut sculptures are labelled in the most instructive manner, which renders these monuments the most valuable contribution to Buddhist legendary history that has been brought to light in modern times. By the time when the gateways of the Sanchi Tope were erected, the fashion had unfortunately died out. It still continued customary for donors of pillars, or of parts, to record their Danams or gifts, but no description of the scenes depicted, nor is any other information afforded, beyond the name and condition of the donor, who generally, however, was a private person, and his name consequently of no historical value.2

• 1. It seems that the vowel marks in the word which Prinsep read as "Suke" in the first line are so indistinct, that it is more probable the word ought to be read Saka: and if this is so it may lead to an interesting national indication. I submitted the passage to Professor Eggling, of Edinburgh, and in reply be informs me that the passage may very well be read "By him who is possessed of the attributes of the famous Saka (race)." If this is so, he may have been either one of those Yavanas who came from the north-west, or at least a descendant of some of those conquerors.
• 2. In the old temple of Pâpanâth, at Pattadkal, this fashion seems to have been revived, for once at least, for all the sculptures on its walls are labelled in characters probably of the fifth century. Arch. Survey of West. lndia, lst Report, p. 36.