then as now the Kiṣangaǹgā Valley about Gurez and the neighbouring territories to the north. From Kalhana's description it is evident that this frontier fort stood on, or close to, the summit of a pass. Thanks to the indications of the Chronicle I was able to identify its site on the top of the Dudokhut Pass.1 The Pass (shown on the map by its ancient name Dugdhaghāta) is approached on the Kasmir side from the valley of the Band pōr stream, still known to the Brahmans by its old name Madhumati. At the small village of Atavuth (map Atawat') a side valley is entered which is narrow and somewhat difficult below, but higher up widens. Its highest portion which forms the immediate approach to the pass, is an open alpine valley known to the mountain shepherds as Vijje Marg.

The term Marg which denotes any high alpine grazing ground frequented in the summer by herdsmen, is the modern Kaśmiri equivalent, and direct derivative, of Skr. maṭhikā. It designated originally the small shelter-huts of stone or wood usually erected on such high plateaus or valleys by their summer occupants. It is probably that Vijje Marg represents the Prajimaṭhika which Kalhana mentions as the position occupied by the Kaśmir forces during their unsuccessful siege of the fort.

As a characteristic point it may be mentioned that the garrison depended for its water-supply on the storage of snow. This had become exhausted at the late summer season when the siege took place, but, luckily for the Darad defenders, was replaced by a fresh fall of snow. The latter is explained by the elevation of the pass which I estimated at about 11,500 feet. Snow-storms occur sometimes on the neighbouring Trag bal Pass so early as September.

From the Dudekhut Pass an easy track over the ridge marked 'Kiser' on the map leads down to Gurēz, the chief place of the Valley. The latter corresponds probably to the DARATPURI of the Rājatarangiņī. The route over the Dudkhut, being very direct and comparatively easy during the summer, was much frequented by Dard traders until the recent construction of the 'Gilgit Transport Road.' It was used by the Sikhs for military convoys until a disaster caused by an avalanche above Ātavuṭh induced them to change it for the Tragabal route. It also seems to have been mentioned to Baron HÜGEL.8 In Muhammadan

1 For detailed evidence regarding this location and a description of the site, see Rājat. vii. 1171 note.

2 Skr. maṭhikā is the diminutive of matha 'hut', 'Sarai.' The Ks. derivative of the latter term, mar, is still used regularly for the rude shelter-huts which are found on the higher passes particularly towards the north.

3 See Kaschmir, ii. p. 169.

times both routes were in charge of a 'Malik' who resided in the castle of Bandskoth, not far from the ancient Mātṛgrāma shown on the map.

In ancient times there probably existed in the same neighbourhood a watch-station or Dranga. Ou-k'ong when speaking of the 'gate to the north' through which the road led to Poliu or Baltistān, may have meant either this Dranga or the fort of Dugdhaghāta.

57. To the east of the Dud khut Pass the summits of the range gradually get higher and higher until we reach Mount Haramukuţa. the great mountain-mass of the Haramukh Peaks. Rising to close on 17,000 feet and surrounded by glaciers of considerable size, these Peaks dominate the view towards the north from a great part of the Kasmir Valley. Sacred legends have clustered around them from early times. The lakes below their glaciers belong still to the holiest of Kaśmirian Tirthas. The ancient name of the Peaks is HARAMUKUȚA, 'Siva's diadem.' This is explained by a legend which is related at length in the Haracaritacintāmaņi.1 Their height is supposed to be Siva's favourite residence. Hence Kaśmirian tradition stoutly maintains that human feet cannot reach the Peaks' summit.3

The lake which lies at the foot of the north-eastern glacier, at a level of over 13,000 feet, is looked upon as the true source of the Kasmir Gangā or Sind River. It is hence known as UTTARAGANGA or popularly Gangabal. It is the final goal of the great Haramukuṭaganga' pilgrimage which takes place annually in the month of Bhadrapada and is attended by thousands of pilgrims. The bones of those who have died during the year, are on that occasion deposited in the sacred waters. A short distance below this lake is another also fed by a glacier and now known as Nundkōl. Its old name Kalodaka or Nandisaras is derived from a legend which makes the lake the joint habitation of Kāla, i.e., Siva, and of his faithful attendant Nandin. From the

1 See Haracar. iv. 62 sqq.

2 The legends relating to Siva's residence on Mount Haramukuța and his connection with the several sacred sites of Nandikşetra, are given at great length in the Nilamata 1049 sqq.

8 Owing to this superstition I had great difficulty in inducing any of my Kasmiri Coolies (Muhammadans!) to accompany me on the ascent I made to the Peaks in September, 1894. My Brahman friends could not give credence to my having reached the summit. According to their opinion the very fact of my having reached the Peak was a sufficient proof of this not having been Haramukuța. An argument as simple as incontrovertible to the orthodox mind.

See my note Rājat. i. 57. Another name often used in the Nilamata and other texts is Uttaramānasa; see Rājat, iii, 448 note.

latter the whole collection of sacred sites takes the name of Nandikṣetra by which Kalhana usually designates it.'

In the valley of the Kankanai stream (Skr. KANAKAVAHINI) which issues from these lakes, lies the sacred site of Siva BHŪTESVARA (now Buthisēr). It is closely connected with the legends of Mount Haramukuta and often mentioned in the Rajatarangini. A series of interesting temple ruins marks the importance of this Tirtha and that of the ancient Jyeṣṭheśvara shrine which immediately adjoins it.3 Bliūteśvara is passed by the pilgrims on their way back from the sacred lakes, while on their way up they reach the latter by another route, passing the high ridge known as BHARATAGIRI and the smaller lake of BRAHMASARAS.

From the Ganga lake a track passable for ponies leads over the Satsaran Pass to Tilel, a Dard district on the Kişanganga. It is probably the route by which King Harsa's rebel brother Vijayamalla escaped from Lahara (Lar) to the Darad territory.

58. Eastwards from the Haramukuta Peaks the range does not overlook on the south the main Valley of North-eastern range. Kaśmir, but that of the Sind River. The gene

ral level of the summits rises, and glaciers of fair size become frequent on their northern slopes. Close to the head of the Sind Valley, the range we have been so far following joins on to the great chain of snowy mountains which stretches from Mount Nanga Parvat in a southeasterly direction to the Nankun Peaks in Sūru. A few miles south of this junction we arrive at a gap in the mountains which forms the lowest watershed between the Indus and the Vitastā basins. It is the Pass known generally by its Ladākhi name of Zōji-Lā. It leads at an elevation of 11,300 feet from Baltal, on the headwaters of the Sind, to a high-level valley draining into the Dras River and hence into the Indus. The route leading over the Zōji-Lā undoubtedly has been already in ancient times a most important thoroughRoute over Zōji-Lā. fare. It connects Kaśmir with Ladākh and thence with Tibet and China. Here too the natural watershed has in old as in modern times been also the ethnic boundary. Beyond the Pass begins the land of the Bhauṭṭas or Bhuṭṭas, as the Tibetan inhabi

1 See Rajat. i. 36 note.

2 See regarding the history and remains of Bhūtesvara, Rājat. i. 107; v. 55 notes. The Tirtha was rich enough to attract a special expedition of marauding hillmen in Kalhana's time; see viii. 2756.

8 See Rajat. i. 113 note.

4 See Rajat. vii. 911.

5 Compare regarding this great range which may fitly be called the main range of the mountain system around Kaśmir, DREW, Jummoo, pp. 194 sqq.

tants of the Indus region are uniformly designated in our Kasmirian texts (modern Kś. But).1

Ou-k'ong is the first who refers distinctly to this route when speaking of the road which leads through the gate in the east to Tou-fan or Tibet. Kalhana has scarcely occasion to refer to it, as the regions beyond the Pass lay quite beyond the reach of the political power of the later Kasmirian kings. He probably means, however, the Zōji-La when mentioning the route of the Bhutta-land (Bhuṭṭarāṣṭrādhvan) by which the Darads offered to pass the pretender Bhoja into Kaśmir, while the more direct routes from their own territory were closed by the winter. An easy pass connects Tilēl at the head of the Kiṣangangā Valley with the Drās territory to the east. From there Bhoja could then have entered Kaśmir via the Zōji-Lā.

This Pass, the ancient name of which is not known to us, has more than once witnessed successful invasions of Kasmir. Through it came early in the 14th century the Turk (?) Dulca and the Bhautta Riñcana whose usurpation led to the downfall of Hindu rule in the Valley.3 About two centuries later Mirzā Muḥammad Haidar with his small Mughal force successfully fought here his entrance into Kasmir (A.D. 1532). The account he gives of this exploit in his Tārīkh-iRashidi, is not without topographical interest.4

59. A high peak situated about 10 miles east-southeast of the Zōji-Lā, marks the point where the range forming the eastern boundary of Kaśmir branches off from the main chain. This range runs in an almost southerly direction until it reaches the southernmost headwaters of the Vitasta. It then turns to the north-west and at the Bānahāl Pass joins on to the Pir Pantsal Range. Through this range there lead routes connecting Kaśmir with the Maḍivāḍvan Valley (see § 84) which drains into the Cinab, and with Kaṣṭavār, the ancient Kāṣṭavāṭa, on the Cināb itself. Both these Valleys are confined, difficult of access, and scantily populated. They have hence never played an important part either in the foreign relations or trade of Kasmir. On this account our notices regarding the old topography of the dividing range are extremely meagre.

1 Compare Rajat. i. 312-316 note.

& Compare Rājat. viii. 2887.

8 See Jonar. 142 sqq., and for the stratagem by which Riñcana forced his way into Lahara (Lar), 165 sqq. The Laharakoṭṭa mentioned in the last passage probably represents the old watch-station of this route, but its position is uncertain.

4 See Tarikh-i-Rashidi, transl. by Messrs. Elias and Ross, pp. 423 sqq., and below, § 131.

Tīrtha of
Amaranātha.

At its northern end and close to the great snowy peak already mentioned, is the Tirtha of AMARESVARA or AMARANATHA, known by its Kaśmiri name as Amburnath. Together with the sacred Gangālake on Mount Haramukuța, it is now the most popular of Kaśmirian pilgrimage places. Its Yātrā in the month of Sravana attracts many thousands of pilgrims not only from Kasmir but from all parts of India. Their goal is a cave situated at a considerable altitude and formed by a huge fissure on the south side of a snowy peak, 17,300 feet high (marked 'Ambarnath' on map). In this cave there is a large block of transparent ice formed by the freezing of the water which oozes from the rock. It is worshipped as a self-created (svayaṁbhū) Linga, and is considered the embodiment of Siva-Amareśvara.

Judging from the scanty references made to this Tirtha in the Rajatarangiņi and the Nilamata, it appears doubtful whether it could have enjoyed in old times quite such great celebrity as now. But Jonarāja already relates a visit to this sacred site paid by Sultan Zainu-l-abidin, and in the Mahatmya literature Amareśvara receives its due share of attention. The pilgrims' route described in great detail by the Amaranathamāhātmya ascends the valley of the eastern branch of the Lidar or Ledari.

There the lake of the Naga SUSRAVAS, now known as Suśramnāg or (with a popular etymology) Sesanag, is visited at the north foot of a great glacier descending from the Kohenhär Peak. In this lake and a small rock-bound inlet of it called JAMATṚNĀGA (Zāmaturi Nag), the local legend, related by Kalhana, i. 267 sqq., and connected with the ancient site of Narapura, has placed the habitation of the Nāga Suśravas and of his son-in-law. The route then crosses a high pass, known as Vāvajan (Skr. Vayuvarjana in the Mahatmya), into a high-level valley drained by five streams which bear the joint designation of PAÑCATARANGIŅI. From there the pilgrims toil up a lofty spur to the northeast and descend into the narrow gloomy valley which lies at the foot of the Amburnāth Peak. It is watered by a stream (Amarāvati) which comes from the glacier of a still higher peak to the east. Joining the Pañcatarangiņi it flows through an inaccessible gorge down to the head of the Sind Valley near Baltal.

1 See for the old notices of the Tīrtha, Rajat. i. 267 note; for a description of the modern pilgrimage, VIGNE, Travels, ii. pp. 10 sqq., and BATES, Gazetteer, pp. 121 sq. 2 Compare Jonar. (Bombay ed.) 1233 sqq.

3 Compare Rajat. i, 267 note.