102°. Durbasa, Vasishta and Parasari, the fifth, sixth and seventh of these springs, have become dry.' In the wall of this place has been built an image of a Buddh or Jain, with both han is in the lap. It has lotus flowers on the soles of the feet, and lions on the throne. It has a short inscription. In the cistern

is lying a male figure somewhat like that usually called Vasudeva, but somewhat different. Notwithstanding its sex the Brahmans call it Devi or the Goddess.

Immediately west or above this reservoir and two small temples (Maths) of Siva, and south from them, [is] a reservoir containing a spout of stone which emits the the finest stream in the place. It is 110° hot, and perfectly limpid and tasteless, but not near so copious as the fine springs of the Mongger hills. In the reservoir are lying two carved stones, one a Ganes. The other, such an octagonal ornament as is so common at Buddh Gya, and containing four images of Buddhs.

Immediately south from thence is another spring, nearly as fine, and named Markunda3. In the reservoir are lying some images. A Gauri Sankar. A Gauri Sankar. A male and female standing, both called Devi. Two such as are usually called Vasudeva, with large ears as usual. In one side of the reservoir is a dark hovel called a temple of Kamaksha, but it contains no image. All these images except such as are objects of worship are said to have been brought from the hill above, and the same has probably been the case with those which are worshipped, (1) In the Saptrishi tank, the third to seventh outflows, as well as that in the Anantrishi Kund to the north-west, which Buchanan did not notice, are all connected with a common underground source, above the Veda Vyasa tank, which in turn is connected with the same source. The third outflow is closest to the source, and is therefore the hottest, though usually more than a degree cooler than the source itself. Its mean temperature since 1909, 104.6°. Maximum 107-4° in October 1914; minimum 96.3° in April, 1909, when, with the exception of Brahmakund, the whole of the springs were practically dry.

11

(2) The Veda Vyasa Kund, called Vyas by Buchanan in his drawing and in the Report. Mean temperature since 1909, 105 3°; maximum 107 00 in April, 1912.

(3) Fed from the same source as the Veda Vyasa, but the subterranean channel from the real source is longer, and the outflow is slightly cooler on this account.

2 Res. J.

as they are in the same style and all resemble those at Buddh Gya and Baragang. The whole water from all these springs unites at the bottom of the hill, and forms a stream rather larger than that of Sitakund1.

Advancing a little farther up the river on the same side with Brahmakund is a fine little spring of clear water issuing from a small square cistern cut in the rock. It is called Vanur Vanuri, from a monkey and his wife having been immediately translated to heaven from bathing in it. A little farther up, the rivulet divides into two branches, and in the fork is a small conical mound of earth and stones. On it is a small modern temple, but the traces of one more ancient and somewhat larger are observable. The size of the mound could never have admitted a large one. The image is broken and is carved on a small stone. It represents one of the most hideous forms of the destructive power that I have seen, with three heads and eight hands, dressed in armour and holding in its hands two serpents, various implements of destruction, and a human head. It seems to me clearly a male, and is probably the same deity with the chief figure in the caves of Elephanta, although that represents only the head and shoulders and this represents the whole body. By the attendant Brahmans it is considered as a female, and called Jaradevi, and to its worship it is supposed that Jarasandha owed all his power.

Beyond this is a considerable plain surrounded by five hills held sacred by the Jain, but neglected by the Brahmans. This plain with the adjacent hills is called (1) Buchanan did not notice the Ganga-Jamuna tank west of the Anant rishi, which is the third independent outflow. Mean temperature since 1909, 106 6°; maximum 107 8° in March, 1911. Being at the highest level, it is often dry. In the Report, he says "I suspect that those near Brahmakunda have, in a state of nature, been one spring; which has been subdivided and conveyed by various channels, so as to supply the various pools and spouts from which it now issues; and in this manner I account for the different degrees of heat observable, and for several of the spouts that formerly flowed being now dry." This observation is correct, The whole of the area occupied by the springs and temples has been built up artificially against the side of the hill. No substantial changes have been made during the last century, and the general agreement with Hiuen Tsang's account suggests that the subdivision of the springs dates from a very early period.

Hangsapurnagar, and is supposed to have been the situation of a city, but of this I see no traces; some zigzag structures of stone both here and at Giriyak were pointed out to me as walls of the old city, but I have not the smallest doubt that they have been roads, and it is probable that there has been a route communicating by the hills with Giriyak, as the zigzags of Rajagriho ascend the west end of the same hill on the east end of which those of Giriak are. It consists of five bends, in all 1200 cubits long, with a roundish resting place at each turn, and is four cubits wide. The people I sent to measure it could trace it no farther, but they might lose it by a very short interruption, as the hill is covered in many places with thick reeds.1 They saw no images nor traces of buildings.

A road leads through the hills, towards the south as well as to the north, and there is a narrow passage towards the valley between the two ridges. On all other sides are rugged hills. The situation is exceedingly strong and in that respect well fitted for a city, and the extent is considerable, three coses by one, but would no doubt be exceedingly unhealthy. The situation however is very grand, and well adapted for occasional visits or for inspiring religious awe, and accordingly the three great Hindu sects have all chosen it as a favourite residence and claim it as their own. The Buddhists of Ava came to it, directed by their books, and considered Rajagriho as the residence of Jarasandha, one of the most religious princes of their sect.

In the south side of the hill by which the central plain is bounded on the north and west, has been dug a cave called Sonbundar. The door is small, but there is also a window, which occasions a circulation of air and gives a light unknown in the dismal caves of Burabur. The materials here however are vastly inferior, as the rock is everywhere intersected by fissures, so that some parts have fallen down, and it admits (1) This zigzag road leads to a flat-topped stone garh on the top of Ratnagiri, not far west of the Jain temple.

water which has stained the walls with a red ferruginous crust. The stone is an imperfect Khori', variegated red and grey in veins, layers, and blotches, and is evidently the rude jasper of the hills, similarly marked, passing into an indurated clay. This cave is called Sonbundar, and is an object of worship with the Jain. In its middle is left a small kind of quadrangular pyramidical figure, on each side of which is carved an erect man with two arms. The chief figure is the same on all the four sides, but on each he is accompanied by different emblems. On the wall is a short inscription in a strange character. It probably merely contains the name of some pilgrim.

On the east side of the rivulet also there are sundry places of worship common to all sects. At the west end of the northern hill is a cluster of springs and small temples, similar to those opposite, and surrounding Surjokundo. This is a small reservoir in which the water is collected in a pool, and does not fall from a spout, so that it is beastly dirty and swarms with frogs. Its heat is 103°. An image of Surja somewhat different from that at Kasi Tirtha is built into the wall, and near it a fat male figure with two arms, and one leg hanging over the throne. It is surrounded by an inscription. This figure is in several other places intermixed with Buddhs, and seems to me to represent the cook of Gautama that I have seen in Ava. Before the feet of Surja has been placed [a] small figure of Buddh. In a small math of Siva southwest from this kund is an image of Buddh, and on the outside [are] two throne-like stones such as I saw at the west end of the great Asuren road. Each has a sitting figure of a Buddh over a standing figure, but in the two the position of the hands is reversed. Here a wretched Sannyasi has taken up his abode. He sits all day besmeared with ashes in the position of a Buddh

(1) Appendix, No. 56.

(2) Surajkund. Usually in the cold weather the level of the water in these tanks is kept above that of the spouts which lead the water into them. Mean temperature of the inflow since 1909, 104 8°, maximum 107 0° in April, 1010.

He neither moves nor speaks, and those who choose bestow alms on him. If he gets none he fasts. It was alleged that some thieves had stolen his blanket, but I suspect that this was a mere allegation to endeavour to extract a rupee from me; no thief in all probability would steal from so wretched an animal, especially as viewed as being of the utmost purity and enjoying divine favour.

West from that is a small ruined math, with a stone containing two feet, and a short inscription in relief, which is not common. The Brahmans call it the feet of Dototreya, one of the 24 Avatars of Vishnu, but from the inscription it evidently belongs to the Jain, as it commences with the character called Balaminda, which the Jain prefix as the Astik do the name of Ganesa.

South from Surjakund are two temples of Siva with one of Tulasi between them. One of them is called Halokeswar. Here are several old images. A pedestal like those already mentioned, containing a Buddh sitting above a man standing. A Gauri Sangkor. A Vasudeva. Two Sahusera Lingas, which implies 1000 Lingas. A Lion rampant, which is an ornament of Gautama. South from thence is Santonkund,' a pool similar to Surjakundo. Its heat [is] 106°. North from Surjakundo is that of Som or the moon. Its heat [is] 102°. Near it is lying an ornament similar to those of Buddh Gya, with four Buddhs on the four sides. Ganesakundo north from thence3. The heat also is 102°.

(1) The position of Santonkund in this account corresponds to that of the present Sitakund, but in Buchanan's rough plan it is shown as east and slightly north of Surajkund, thus occupying a site where in 1917 I found that the earth had fallen in, exposing channels leading from the hillside further east, and diverging to the present Sitakund, Surajkund and Chandramakund. Mean temperature of Sitakund inflow, 104 7°. Maximum 106 0° in May, 1912.

(2) Chandrama or Somkund. Mean temperature of inflow 104.7°. Maximum 106 0° in May, 1912.

(8) Ganeshkund. Mean temperature of inflow, 104 0°. Maximum, 106·0° in May, 1911. At the steps leading into this tank, as well as those of Sitakund and Chandramakund, the temperature since 1909 is between 97 and 99°, and has never reached 100.