This is somewhat less than eight miles; but if the ascents and descents are taken into account, the actual length will correspond very closely with the statement of Hwen Thsang when corrected to 50 li. The old walls forming this exterior line of rampart are still to be seen in many places. I traced them from Vipula-giri over Ratna-giri to the Nekpai embankment, and thence onwards over Udaya-giri, and across the southern outlet of the valley to Sona-giri. At this outlet, the walls, which are still in good order, are 13 feet thick. To obtain a circuit of 25 miles, as given in Hwen Thsang's text, it would be necessary to carry these ramparts as far as Giryek on the east. As similar ramparts exist on the Giryek Hill, it is perhaps possible that Hwen Thsang intended to include it in the circuit of his outer walls. But this immense circuit would not at all agree with his statement that "high mountains surround the city on four sides," for the distant Hill of Giryek cannot in any way be said to form one of the sides of old Râjagriha.

The new town of Râjagriha is said to have been built by King Srenika, otherwise called Bimbisára, the father of Ajátasatru, the contemporary of Buddha. Its foundation

cannot, therefore, be placed later than 560 B. C. according to Buddhist chronology. In Hwen Thsang's time (A. D. 629642), the outer walls had already become ruinous, but the inner walls were still standing, and occupied a circuit of 20 li, or 3 miles. This statement corresponds tolerably well with the measurements of my survey, which make the circuit of the ramparts somewhat less than 3 miles. Buchanan calls new Rajagriha an irregular pentagon of 12,000 yards in diameter. This is clearly a misprint for 1,200 yards, which would give a circuit of 11,303 feet, or 23 miles; but this was probably the interior measurement, which, according to my survey, is 13,000 feet. The plan of new Râjagriha I make out to be an irregular pentagon of one long side and four nearly equal sides, the whole circuit being 14,260 feet outside the ditches, or rather less than three miles.*

On the south side towards the hills a portion of the interior, 2,000 feet long and 1,500 feet broad, has been cut off to form a citadel. The stone walls retaining the earthen ramparts of this work are still in good order in many places.

* See Plate XIV.

It is possible that this work may be of later date, as suggested by Buchanan, but I am of opinion that it was simply the citadel of the new town, and that its walls have suffered less from the effects of time, owing partly to their having been more carefully and more massively built than the less important ramparts of the town, and partly to their having been occasionally repaired as a military position by the authorities, while the repairs of the town walls were neglected as being either unnecessary or too costly.

The existing remains at Râjagriha are not numerous. The place has been occupied at different times by Musalmâns and Brahmans, by whom the Buddhist stupas and vihârs were pulled down to furnish materials for tombs, masjids, and temples. All the eminences that must once have been crowned by objects of Buddhist worship are now covered with Muhammedan graves; and all the Brahmanical temples about the hot springs have been constructed with the large bricks of Buddhist stupas. One of these last monuments can still be traced outside the south-west corner of the town in a large circular hollow mound, which attracted the notice of both Buchanan and Kittoe. I examined this mound carefully, and I was satisfied that the hollow represented the original site of a stupa from which the bricks had been carried off, while the surrounding circular mound represented the mass of earth and broken brick rubbish left by the workmen. The excavated stupa at Sârnâth, near Banâras, now offers almost exactly the same appearance. According to Hwen Thsang's account, this circular hollow was the site of a stupa 60 feet in height, which was built by Asoka. Beside it there was a stone pillar 50 feet high, on which was inscribed the history of the foundation of the stupa. The pillar was surmounted by an elephant.*

On Mount Baibhâr there are five modern Jain temples, besides the ruins of an old Saiva temple, of which four granite pillars, 10 feet in height, are still standing, and 50 or 60 smaller pillars are lying confusedly about. At the southern foot of the mountain, the rock has a natural scarp for about 100 yards in length, which, at the western end, has been smoothed to a height of 19 feet, in front of which the rock has been cut away to form a level terrace 90 feet in length by

Julien's Hwen Thsang, III., 38.

upwards of 30 feet in breadth. Two caves have been excavated out of the solid rock behind; that to the west, now called the Son Bhândâr, or "Treasury of gold," being 34 feet long by 17 feet broad, and that to the east perhaps somewhat less in length, but of the same breadth. This cave has either fallen in naturally through the decay of the rock, or, which is more probable, was blown up by a zemindar in search of treasure, as related by Major Kittoe of the other cave.

The Son Bhândâr Cave has one door and one window. Inside there are no traces of seats, or of pedestals of statues, and the walls and roof are quite bare, excepting where a few scarcely legible inscriptions have been cut. There are several short inscriptions on the jambs of the doorway, as well as on the outside. In the principal inscription, which is on two lines outside, the author speaks of this cave as the "auspicious cave," evidently alluding to the fact of its former occupation by Buddha for the purpose of meditating after his noonday meal. This inscription, which is not later than A. D. 200, and is perhaps earlier, records that a certain "Muni, named Vaira Deva, of powerful dignity, was able to obtain emancipation, having shut himself up for spiritual enjoyment in this auspicious cell, a retired abode of Arhantas, fitted for an ascetic for the attainment of liberation." On the east jamb of the door also the same epithet is applied to this cave, as if it was a well known name for it. This cave is excavated in the south face of the hill, where there is a natural scarp for about one hundred yards in length. The face of the cliff at the west end has been smoothed to a height of 19 feet, in front of which the ground has been levelled to form a platform of more than 30 feet. The cave itself is 34 feet long by 17 feet broad and 11 feet high. To the east there has been a second cave, about 22 feet long by 17 feet broad; but one half of the roof fell in long ago, and the cave is now filled with masses of rock and earth. The floor of this cave is on a lower level than that of the Son Bhándár, but the front is in the same line. Both caves had some building or verandah in front, as there are numerous socket holes cut in the rock above the door for the reception of the ends of beams. The whole length of level clearing in front of the caves is 90 feet.

In the centre of the valley between the five hills, and in the very midst of the old city of Râjagriha, there is a ruined

D

brick mound 19 feet 8 inches in height, which my excavations proved to be an ancient stupa. A diminutive Jain temple, called Maniâr Math, stands on the top of the mound. It was built in A. D. 1780. As I expected to find a solid brick building, I sank a shaft outside the Maniâr Math with the intention of inclining gradually towards the centre; but I soon found that the core of the mound was a mere mass of rubbish, filling a well 10 feet in diameter. This rubbish was so loose that its removal was dangerous; but by propping up the portion immediately below the little temple, and removing the bricks cautiously, I was enabled to get down to a depth of 21 feet. At 19 feet I found three small figures. One of them represents Mâyâ lying on a couch in the lower compartment, and the ascetic Buddha and two attendants above. The second is a naked standing figure, with a sevenheaded snake forming a canopy over the head. This is clearly not a Buddhist, but a Jain sculpture. The third is so excessively rude, that it is difficult to identify it. The figure is four-armed, and is seated upon a recumbent animal, which looks more like a bull than anything else. It probably, therefore, represents Mahadeva and his bull Nandi. As all three figures formed only a part of the rubbish, it seems to me certain that the well must once have been empty; and further, that the rubbish was most probably thrown in when the little Jain temple was about to be built.

The natives of the place call this well the Treasury, and they assert that it has never been opened. On my arrival I found a Punjab Sepoy, with a servant, making an excavation on his own account. He had sunk a shaft 3 feet in diameter at 7 feet from the little temple. The shaft was then 17 feet deep. I examined the bricks which had been taken out, and on finding some with bevelled and rounded edges, and others thickly coated with plaster, I guessed at once that the original structure had been covered with an outer wall, and that the shaft had been sunk just outside the original work. To ascertain whether this conclusion was correct, I laid bare the top of the mound, and soon discovered that the well was surrounded by a wall only 6 feet in thickness. This would give the original stupa a diameter of 22 feet. The Punjab Sepoy continued his shaft down to the stone foundation without finding anything, and then gave up the work.

Having observed that the slope of the mound on the north side was very gentle, I thought it probable that the building must have been approached on this side by a flight of steps. I therefore made an excavation in a line due north from the centre of the mound, and within a couple of hours I found a doorway. Continuing the excavation to the east and west, as well as to the north, I found a small room with brick walls and granite pillars containing two middle-sized sculptured slabs of middle age. Outside the doorway a flight of steps led downwards towards the north; I therefore turned to the south, and continued my excavation until I reached the main building. On examining the wall I found three recesses, the middle one being roofed by overlapping bricks. On clearing out the rubbish, this opening proved to be a carefully built passage only 2 feet 2 inches wide, and 3 feet 4 inches in height, right through the outer wall of the building. Behind it, but a few inches out of line, there was a similar passage through the original wall, only 2 feet in width. At the end of the passage I found the well filled with the same rubbish as on the south side.

The discovery of this passage shows that the Buddhist Monks had easy access to the interior of the building. I conclude, therefore, that it must originally have contained some relic that was occasionally shown to visitors, and to the public generally, on certain fixed days. I cannot, however, discover in the accounts of Fa-Hian and Hwen Thsang any mention of a stupa inside the walls of old Râjagriha.

The hot springs of Râjagriha are found on both banks of the Sarsuti rivulet; one-half of them at the eastern foot of Mount Baibhâr, and the other half at the western foot of Mount Vipula. The former ar enamed as follows: 1, GangâJumna; 2, Anant Rikhi; 3, Sapt Rikhi; 4, Brahm-kûnd; 5, Kasyapa Rikhi; 6, Biâs-kûnd; and 7, Markand-kûnd. The hottest of these are the springs of the Sapt Rikhi. The hot springs of Mount Vipula are named as follows: 1, Sitakûnd; 2, Suraj-kûnd; 3, Ganes-kûnd; 4, Chandrama kund; 5, Râm-kund; and 6, Sringgi-Rikhi-kund. The last spring has been appropriated by the Musalmâns, by whom it is called Makhdum-kûnd, after a celebrated Saint named Chilla Shâh, whose tomb is close to the spring. It is said that Chilla was originally called Chilwa, and that he was an Ahir. He must, therefore, have been a converted Hindu.