and Amaranatha is much frequented even at the present day. The Māhātmyas of Vijayeśvara do not fail to name a considerable number of minor Tirthas to be visited along with the main site now marked by the new temple above referred to. But apart from Cakradhara and Gambhirasamgama I am unable to trace any of these in the older texts.

Turning to that portion of the Dachünpōr district which lies in the Lidar Valley we have but few old localities to notice. The village of Liver, some ten miles to the north-west of Vijayeśvara, is the Levāra of the Rajatarangiņi, mentioned as an Agrahāra established by King Lava. Kular, about four miles higher up the Valley, is identified by an old gloss with KURUHĀRA, said to have been an Agrahāra of Lava's son Kuśa.*

Close to Pahalgam where the Lidar Valley divides into two branches, lies the hamlet of Mamal. A small temple of the usual Kasmir style built by the side of a fine spring is visited by the pilgrims to Amaranatha. It is designated in the Mahatmya called Amareśvarakalpa as MAMMESVARA. It is in all probability identical with the shrine of this name mentioned in the Rājatarangiņi.3

District of
Vāmapārsva.

110. As we have already before noticed the several sacred sites of the Amaranatha pilgrimage, we may now turn back and descend to the left or eastern portion of the Lidar Valley. It forms the modern Pargana of Khōvurpōr. The latter name meaning 'left side' reproduces the earlier designation VAMAPĀRSVA, of the same significance, found in Jonaraja's Chronicle, the Lokaprakāśa and elsewhere. In the upper portion of the Pargana I am not able to identify any particular old locality, though ancient remains in the form of sculptures of some interest are found near several Nagas of this tract, e.g., at Lokut1pār and Sali (Papaharaṇanāga).

The large village of Hut mar is undoubtedly an old site. Its modern name seems to identify it with the SAKTAMATHA which Kṣemendra names as one of the stations in the peregrinations of his heroine Kankali. The chief mosque of the place is built with the remains of a Hindu temple and preserves in its walls some sculptured fragments of remarkable beauty."

1 See Rajat. i. 87.

& Rajat. i. 88.

8 See viii. 3360.

Jonar. (Bo. ed.) 79,1232.

5 See Samay. ii. 43. The change of Sakta > Hute is in accordance with the phonetic laws of Kaśmiri; mar is the regular derivative of matha, see above, § 56. [When preparing my map, I had not noticed the local name of Kşemendra's text; it is hence not shown on the map].

Shrine of Bhimakesava.

Sahib is nothing but a

About one mile below Hutamar and on the bank of a branch of the Lidar, lies the hamlet of Bumizu, which contains an ancient structure of considerable historical interest. The Ziarat of Bāba Bām din well-preserved temple, converted, with a liberal use of plaster, into the supposed resting place of a Muhammadan saint. I have shown elsewhere that there is good reason to identify this shrine with the BHIMAKESAVA temple which Bhima Sahi, king of Kabul, the maternal grandfather of Queen Didda, is said to have erected during the rule of her husband Kşemagupta (A.D. 950-958).1

The legendary of the Ziarat relates that the saint was originally a Hindu and bore before his conversion to Islam the name of Bhima Sadhi. It is easy to recognize in this name an adaptation of Bhima Sahi. Also the name of the locality Bum zu which the Mārtāṇḍamāhātmya renders by Bhimadvipa, is clearly derived from the old name of the shrine. Bhima is an abbreviation of Bhimakesava to which Ks. zu, 'island,' has been added with reference to the several islands formed here by the Lidar immediately in front of the hamlet.

Kalhana tells us a curious anecdote regarding the fate of Bhima Sahi's temple in King Harga's time who confiscated the great treasures, with which it was endowed. Close to the present Ziarat of Bāmadin Sahib is a small cave in the cliff containing a well-preserved little temple which is still used for Hindu worship. Another smaller shrine outside has been turned into the tomb of Rishi Ruknu-d-din Ṣāḥib. 111. About one mile south of Bum zu we reach the Tirtha sacred to Mārtāṇḍa which has from early times to the Tīrtha of Mārtāṇḍa. present day enjoyed a prominent position among the sacred sites of Kaśmir. It is marked by a magnificent spring traditionally represented as two, Vimala and Kamala. An ancient legend connects them with the birth of the sun-god MARTĀNDA. The Tirtha is visited at frequent intervals by crowds of pilgrims and is well-known also in India proper.

The popular name of the Tirtha, Bavan, is derived from Skr. bhavana, '[sacred] habitation.' This somewhat general appellation seems to have come into use already at an early date, as Srīvara employs it, and is in itself an indication of the great popularity of the Tirtha. A

1 See Rājat. vi. 178 note. For an accurate description of the temple, see Bishop COWIE's paper, J. A. S. B., 1866, pp. 100 sq.

2 See Rajat. vii. 1081 sqq.

3 Compare for a detailed account of the Tirtha, Rājat. iv. 192 note. The Vimala Naga is named by the Nilamata, 963; Sriv. i. 377, etc.

♦ Sriv. i. 376, 387.

more specific designation is Mats bavan, Skr. Matsyabhavana'; this. owes its origin to the abundance of sacred fish which swarm in the large basins filled by the spring.1

The ancient remains at the sacred spring itself are very scanty. All the more imposing are the ruins of the great temple which King Lalitaditya erected at a short distance in honour of the presiding deity of the Tirtha.2

They are situated a little over a mile to the south-east of Bavan,' near the northern edge of the Uḍar which stretches towards Anatnāg. It can scarcely be doubted that the site was chosen with a view to the prominent position it assured to the great temple. Kalhana duly praises "the wonderful shrine of Martāṇḍda with its massive walls of stone, within a lofty enclosure.' Its ruins though much injured by the ravages of time and earthquakes, form still the most impressive specimen of ancient Kaśmir architecture. They have been much admired by European travellers and often described. They are the earliest ruins in Kasmir the date of which is fixed with approximate accuracy,8

The name Mārtāṇḍa, in the form of Martand or Matan, still attaches to the ruins though they have long ago ceased to be an object of religious interest. King Kalasa had sought this great fane at the approach of death and expired at the feet of the sacred image (A.D. 1089). Harga, his son, respected this temple in the course of the ruthless confiscations to which he subjected the other rich shrines of the country. Subsequently in Kalhana's time the great quadrangular courtyard of the temple with its lofty walls and colonnades was used as a fortification. The destruction of the sacred image is ascribed to Sikandar Butshikast.

Kalhana dicinctly mentions the town "swelling with grapes" which Lalitāditya founded near his temple; but of this no trace remains now. It is probable that at that time a canal supplied water from the Lider to the naturally arid plateau on which the temple stands. This canal seems to have been repaired by Zainu-l-'ābidin whose irrigation works on the Martand Uḍar are described at length by Jonarāja.1 The

1 Comp. Ain-i-Akb., ii. p. 358.

2 See Rajat. iv. 192 and for details my note on the passage. For a description of the temple compare, e.g., CUNNINGHAM, J. A. §. B., 1848, pp. 258 sqq.; COLE, Ancient Buildings, pp. 19 sqq. FERGUSSON, Ind. Architecture, pp. 285 sqq.

3 Lalitȧditya's rule falls in the first half of the eighth century. Gen. Cunningham's assumption that the temple was built by the earlier King Raṇāditya, and only the enclosure by Lalitäditya, rests on a misinterpretation of the Rājatar. passages

iv. 192 and iii. 462.

♦ See Jonar. 1245 sqq.

J. 1. 23

plateau has since become once more an arid waste though the course of the old canal can still be traced above Hutamar.

The town of Mārtāṇḍa had left its name to the small Pargana of Matan which comprised this plateau as well as the villages situated along the foot of the hills further east. It is referred to as Mārtāṇḍadeśa by Jonarāja. Abu-l-Fazl notices the large temple of Mațan and the well or pit close by, which a Muhammadan legend represents as the place of captivity of the 'angels Hārūt and Mārūt.' 2

SECTION VII.-SOUTHERN DISTRICTS OF MADAVARAJYA.

112. At the foot of the western extremity of the Martāṇḍa plateau lies the town of Islāmābād or by its Hindu Anantanāga. name Anatnāg. The latter is derived from the great spring of the ANANTANAGA which issues at the southern end of the town. The Naga, though no Tirtha of particular repute, is mentioned in the Nilamata, Haracaritacintamaņi and some Māhātmyas. Of the town, however, I cannot find any old notice, and it is in all probability, as its Muhammadan name implies, a later foundation. To the north of the town and on the way to Bavan is the Gautamanāga, named by the Nilamata and the Mārtāṇḍamāhātmya.

The modern name of the small district which comprised besides Anatnag the tract immediately south and west of it, is Anyech. This is represented in some Mahatmyas of recent composition by Anekākṣa. This name occurs also once in Srivara's Chronicle, but the locality there meant is not certain.

Tīrtha of
Kapatesvara.

The valley of the Arapath or Harṣapatha which opens to the east of Islāmābād, forms the Pargana of Kuṭhār. This name is in all probability connected with that of the ancient Tirtha of KAPATEŚVARA, situated on the southern side of the valley close to the village of Kōṭhēr. The name of the latter is undoubtedly a derivative of Kapaṭeśvara, as the analogy of Jyeṭhēr < Jyeşṭheśvara, Triphar < Tripureśvara, etc., clearly shows.

1 Jonar. 1310.

2 See Ain-i-Akb., ii. p. 358. For the Muhammadan story, see also VIGNE i. p.

361.

8 See Nilamata, 902; Vitastā,-Trisuṁdhyāmāhātmya, etc., also Haracar. x. 251 sqq.

(Anantabhavana).

Sriv. iii. 184.

See for a detailed account, Rājat. i. 32 note.

The place of pilgrimage is the sacred spring of Pāpasūdana ('sinremoving'), situated a short distance above Kōṭhēr. In it Siva is believed to have shown himself in the disguise (kapaṭa) of pieces of wood floating on the water. The legend is related at length in the Nilamata, and the author of the Haracaritacintāmaṇī devotes to it a separate canto which has now become the official Mahatmya of the Tirtha. The importance of the latter is shown by the fact that Kalhana mentions it in his Introduction first among the sacred sites of Kaśmir.

Before him already Albērūni had heard of the story that pieces of wood sent by Mahadeva appear annually "in a pond called Kūdaishahr to the left of the source of the Vitasta, in the middle of the month of Vaiśākha." Kūdaishahr (5), is an easily explained corruption for i.e., *Kavadēśvar, a prakritized form of the name. The map shows that the description of the position of the Tirtha is accurate enough with reference to the Nilanaga as the Vitastā's traditional The date named by Albērūni is identical with that prescribed for the Kapateśvara Yātrā.

source.

The sacred spring rises in a large circular tank, enclosed by an ancient stone-wall with steps leading into the water. According to Kalhana's account this enclosure was constructed about a century before his own time at the expense of the well-known King Bhoja of Mālava. The latter is said to have taken a vow to always wash his face in the water of the Papasūdana spring which he caused to be regularly supplied to him in jars of glass. In my note on the passage I have shown that local tradition at Kōṭhēr still retains a recollection of this story though in a rather legendary form. A small temple which stands. to the east of the tank, and some other remains probably belong to the period of Bhoja. Abu-l-Fazl too knows, "in the village of Kōtihār, a deep spring surrounded by stone temples. When its water decreases an image of Mahadeva in sandal wood appears."

Samāngāsā.

About four miles to the north-east of Kōṭhēr and on a branch of the Arapath river lies the populous village of Sangas, the ancient SAMĀNGĀSĀ. The modern name can be traced back to Samāngāsā through a course of regular phonetic conversion, one stage of which is preserved in the form Svāngas supplied by the old glossator of the Chronicle.

1 Haracar. xiv.

& See India, ii. p. 181.

8 See Rajat. vii. 190 sqq.

See Rajat. i. 100; viii. 651.

6 Compare Rūjat. i. 100 note and the analogy of Sanăra > Sār.

Some old