our maps. It is, however, a very petty place; and, although it boasts of a small khera, or mound of ruins, it cannot, I think, have ever been more than one-fourth of the circuit of two miles which Hwen Thsang attributes to Pi-lo-shan-na. But there are two strong points in its favour, namely, 1st, its position which agrees both in bearing and distance with the Chinese pilgrim's account; and 2nd, its name, which is almost identical with the old name, sh being very commonly pronunced as kh, so that Hwen Thsang's Piloshanna would usually be pronounced Pilokhana.

In proposing Atranji-Khera as the site of the ancient Piloshanna, I am influenced solely by the fact that this is the only large place besides Soron of any antiquity in this part of the country. It is true that the distance from Sankisa is somewhat greater than that recorded by the Chinese pilgrim, namely, 45 miles, instead of 33 miles, but the bearing is exact; and as it is quite possible that there may be some mistake in Hwen Thsang's recorded distance, I think that Atranji-Khera has a better claim than any other place to be identified with the ancient Piloshanna. I have not visited the place myself, as I was not aware of its importance when I was in its neighbourhood. I have had it inspected by a trustworthy servant, whose report shows that Atranji must once have been a place of considerable extent and importance. According to him, the great mound of Atranji is 3,250 in length, and 2,550 in breadth at the base. Now, these dimensions would give a circuit of about two miles, which is the very size of Piloshanna as recorded by Hwen Thsang. Its highest point is 44 feet 9 inches, which, if my identification is correct, should be the ruins of the great Stupa of Asoka, upwards of 100 feet in height, as this loftly tower is said to have been situated inside a monastery in the middle of the town. Outside the town there were two other monasteries, inhabited by 300 monks. These may, perhaps, be represented by two small mounds which still exist on the east side of the Great Khera. To the south there is a third mound, 165 feet in length by 105 feet in breadth, which may possibly be the remains of one or more of the five Bramanical temples described by Hwen Thsang.

Atranji-Khera had two gates,-one to the east, towards the Kali Nadi, and the other to the south. The foundation of the place is attributed to Raja Vena Chakravartti.

The mound is covered with broken bricks of large size and fragments of statues, and old coins are said to be frequently found. All the existing fragments of statues are said to be Brahmanical. There is a temple of Mahadeo on the mound, and there are five lingams in different places, of which one is 6 feet in height. The principal statue is that of a fourarmed female called Devi, but which, as she is represented treading upon a prostrate figure, is most probably Durga.*

The only objection to the identification of Atranji with Piloshanna is the difference between the distance of 200 li, or 33 miles, as stated by Hwen Thsang, and the actual distance of 43 miles direct, or about 48 or 50 miles by road. I have already suggested the possibility of there being some mistake in the recorded distance of Hwen Thsang, but perhaps an equally probable explanation may be found in the difference of the length of the yojana. Hwen Thsang states that he allowed 40 Chinese li to the yojana; but if the old yojana of Rohilkhand differed from that of the Central Doab as much as the kos of these districts now differ, his distances would have varied by half a mile in every kos, or by two miles in every yojana, as the Rohilkhand kos is only 1 mile, while that of the Doab is two miles-the latter being onethird greater. Now, if we apply this difference to Iwen Thsang's measurement of 200 li, or 33 miles, we increase the distance at once to 44 miles, which agrees with the direct measured distance on the map. I confess, however, that I am rather inclined to believe in the possibility of there being a mistake in Hwen Thsang's recorded distance, as I find exactly the same measurement of 200 li given as the distance between Sankisa and Kanoj. Now, the two distances are precisely the same, that is, Sankisa is exactly midway between Atranji and Kanoj; and as the latter distance is just 50 miles by my measurement along the high road, the former must also be the same. I would, therefore, suggest the probability that both of these distances should be 300 li, or 50 miles, instead of 200 li as recorded in the text. In favor of this proposed correction I may cite the testimony of the earlier Chinese pilgrim Fa Hian, who makes the distance from San

*At my request Atranji was visited in 1865 by my friend Mr. C. Horne, then Judge of Manipuri, whose account of the ruined mound will be found in the Bengal Asiatic Society's Journal, 1866, p. 165. The mound has been dug up in all directions for many centuries in search of bricks, and it was with difficulty that an entre brick was found for measurement.

kisa to Kanoj 7 yojanas, or 49 miles. At IIwen Thsang's own valuation of 40 li to the yojana, this measurement would give 280 li; and as Fa Ilian does not record half yojanas, we may increase the distance by half a yojana, or 20 li, which brings the total up to 300 li, or exactly 50 miles.

But whatever may be the true explanation of the difference between the actual distances and those recorded by Hwen Thsang, there still remains the important fact that Sankisa was exactly midway between Kanoj and Piloshanna just as. it now is midway between Kanoj and Atranji. If we couple this absolute identity of position with the fact that Atranji is the only old place in the part of the country indicated by IIwen Thsang, we can scarcely arrive at any other conclusion than that the great ruined mound of Atranji is the site of the ancient Piloshanna,

IX. SANKISA.

The site of Sankisa was discovered by me in 1842, but it was not until the end of 1862 that I got an opportunity of exploring the ruins at leisure. The name of the place is written Seng-kia-she by the Chinese pilgrims, a spelling which is well preserved in the Sankisa of the present day, and which represents, with considerable faithfulness, the Sankasya of Sanskrit. Hwen Thsang calls it also by the name of Kie-pitha, or Kapitha, of which I was unable to discover any trace.* Sankisa was one of the most famous places of Buddhist pilgrimage, as it was there that Buddha was believed to have descended from the Trayastrinsa heaven by the ladder of gold or gems, accompanied by the gods Indra and Brahma. According to this curious legend, Máyá, the mother of Buddha, died seven days after his birth, and ascended at once to the Trayastrinsa heaven, the abode of the 33 gods, of whom Indra was the chicf. But as she had no opportunity in this abode of the gods of hearing the law of Buddha, her pious son ascended to the Trayastrinsa heaven and preached for three months in her behalf. He then descended to the earth with the gods Brahma and Indra by three staircases, one of which was formed either of crystal or precious stones, another of gold, and the third of silver. According

* Julien's Hwen Thsang, II., 237.-In the Brihaj-Játaka it is said that the famous astronomer, Varaha Mihira, "obtained the gracious favour of the sun at Kapitthaka," I presume that this is the Kic-pi-tha of the Chinese pilgrim. Dr. Kern thinks that Varaha Mihira was very probably educated there. Sankisa must at any rate have been a place of considerable importance in the 6th century.