mistake in placing Mayûra and Gangâdwâra to the east of the Ganges.

There is a dispute between the followers of Siva and Vishnu as to which of these deities gave birth to the Ganges. In the Vishnu Purâna' it is stated that the Ganges has its rise "in the nail of the great toe of Vishnu's left foot; "* and the Vaishnavas point triumphantly to the Hari-ki-charan, or Hari-ki-pairi (Vishnu's foot-prints), as indisputable evidence of the truth of their belief. On the other hand, the Saivas argue that the proper name of the place is Hara-dwára, or "Siva's Gate," and not Hari-dwára. It is admitted also, in the 'Vishnu Purâna,' that the Alakananda (or east branch of the Ganges) "was borne by Mahadeva upon his head." But in spite of these authorities, I am inclined to believe that the present name of Haridwâr or Haradwâr is a modern one, and that the old town near the Gangâdwara temple was Mayapura. Hwen Thsang, indeed, calls it Mo-yu-lo, or Mayura, but the old ruined town between Haridwâr and Kankhal is still called Máyápur, and the people point to the old temple of Máyá-Devi as the true origin of its name. It is quite possible, however, that the town may also have been called Mayura-pura, as the neighbouring woods still swarm with thousands of peacocks (Mayúra), whose shrill calls I heard both morning and evening.

Hwen Thsang describes the town as about 20 li, or 33 miles, in circuit, and very populous. This account corresponds very closely with the extent of the old city of Mâyâpura, as pointed out to me by the people.

*Book ii. 8. Hall's edition of Wilson's translation, ii. 272.
t Ibid.

These traces extend from the bed of a torrent which enters the Ganges near the modern temple of Sarvvanâth to the old fort of Raja Ben, on the bank of the canal, a distance of 7500 feet. The breadth is irregular, but it could not have been more than 3000 feet at the south end, and, at the north end, where the Siwâlik hills approach the river, it must have been contracted to 1000 feet. These dimensions give a circuit of 19,000 feet, or rather more than 3 miles. Within these limits there are the ruins of an old fort, 750 feet square, attributed to Raja Ben, and several lofty mounds covered with broken bricks, of which the largest and most conspicuous is immediately above the canal bridge. There are also three old temples dedicated to Narayana-sila, to Máyá-Devi, and to BhaiThe celebrated ghat called the Pairi, or "Feet Ghat," is altogether outside these limits, being upwards of 2000 feet to the north-east of the Sarvvanâth temple. The antiquity of the place is undoubted, not only from the extensive foundations of large bricks which are everywhere visible, and the numerous fragments of ancient sculpture accumulated about the temples, but from the great variety of the old coins, similar to those of Sugh, which are found here every year.

rava.

The name of Haridwara, or "Vishnu's Gate," would appear to be comparatively modern, as both Abu Rihân and Rashid-ud-din mention only Gangá-dwára. Kâlidâs also, in his 'Meghaduta,' says nothing of Haridwâra, although he mentions Kankhal; but as his contemporary Amarasinha gives Vishnupadi as one of the synonyms of the Ganges, it is certain that the legend of its rise from Vishnu's foot is as old as the fifth

century. I infer, however, that no temple of the Vishnupada had been erected down to the time of Abu Rihân. The first allusion to it of which I am aware is by Sharif-ud-din,* the historian of Timur, who says that the Ganges issues from the hills by the pass of Cou-pele, which I take to be the same as Koh-pairi, or the "Hill of the Feet" (of Vishnu), as the great bathing ghat at the Gangâdwara temple is called Pairi Ghat, and the hill above it Pairi Pahár. In the time of Akbar, the name of Haridwâr was well known, as Abul Fazl speaks of "Mâyâ, vulgo Haridwâr, on the Ganges," as being considered holy for 18 kos in length. In the next reign the place was visited by Tom Coryat, who informed Chaplain Terry that at "Haridwára, the capital of Siba, the Ganges flowed amongst large rocks with a pretty full current." In 1796 the town was visited by Hardwicke, who calls it a small place situated at the base of the hills. In 1808, Raper describes it as very inconsiderable, having only one street, about 15 feet in breadth, and a furlong and a half (or three-eighths of a mile) in length. It is now much larger, being fully threequarters of a mile in length, but there is still only one street.

Iwen Thsang notes that the river was also called Fo-shui, which M. Stanislas Julien translates as l'eau qui porte bonheur, and identifies with Mahabhadrá, which is one of the many well-known names of the Ganges. He mentions also that bathing in its waters was sufficient to wash away sin, and that if corpses were thrown into the river the dead would escape the

** History of Timur,' translated by Petis de la Croix, iii. 131.

Ayin Akbari,' ii. 516.

+

Julien's Hiouer Thsang,' ii. 217.

punishment of being born again in an inferior state, which was due to their crimes. I should prefer reading Subhadra, which has the same meaning as Mahâbhadrâ, as Ktesias mentions that the great Indian river was named ὕπαρχος, which he translates by φέρων πάντα τὰ ἀγαθὰ.* Pliny quoting Ktesias calls the river Hypobarus, which he renders by "omnia in se ferre bona." A nearly similar word, Oibares, is rendered by Nicolas of Damascus as dyabúyyeλos. I infer, therefore, that the original name obtained by Ktesias was most probably Subhadra.

5. BRAHMAPURA.

On leaving Madâwar, Iwen Thsang travelled northward for 300 li, or 50 miles, to Po-lo-ki-mo-pu-lo, which M. Julien correctly renders as Brahmapura. Another reading gives Po-lo-hi-mo-lo, in which the syllable pu is omitted, perhaps by mistake. The northern bearing is certainly erroneous, as it would have carried the pilgrim across the Ganges and back again into Srughna. We must therefore read north-cast, in which direction lie the districts of Garhwal and Kumaon that once formed the famous kingdom of the Katyuri dynasty. That this is the country intended by the pilgrim is proved by the fact that it produced copper, which must refer to the well-known copper mines of Dhanpur and Pokhri in Garhwâl, which have been worked from a very early date. Now the ancient capital of the Katyuri Rajas was at Lakhanpur or Vairál-pattan on the Râmgangâ river, about 80 miles in a direct line from Madawar. If we might take the measurement

* Ctesiæ Indica, Excerp. ab Photio, 19, e lit. Lion.

Hist. Nat. xxxvii. 11.

Julien's Hiouen Thзang,' i. 431, and ii. 231.

from Kot-dwâra, at the foot of the hills on the northeast frontier of Madâwar, the distance would agree with the 50 miles recorded by Hwen Thsang. It occurs to me, however, as a much more probable explanation of the discrepancy in the recorded bearing and distance that they must properly refer to Govisana, the next place visited by Hwen Thsang, from which Bairat lics exactly 50 miles due north.

According to the history of the country, Vairál-pattan or Lakhanpur was the ancient capital, as the Sombansi dynasty of Kumaon and the Surajbansi dynasty of Garhwal date only from the Samvat years 742 and 745, which, even if referred to the era of Vikramaditya, are posterior to the time of Hwen Thsang. I think, therefore, that Brahmapura must be only another name for Vairát-pattan, as every other capital in these provinces is of much later date. Srinagar on the Alakananda river was founded so late as s. 1415, or A.D. 1358, by Ajaya Pala of Garhwâl, and is besides nearly as far from Madâwar as Vairat-pattan; while Chándpur, the earlier capital of Garhwal, is still more distant, and dates only from s. 1216 or A.D. 1159. The climate is said to be slightly cold, and this also agrees with the position of Bairát, which is only 3339 feet above the sea.

Hwen Thsang describes the kingdom of Brahmapura as 4000 li, or 667 miles, in circuit.* It must, therefore, have included the whole of the hill-country between the Alakananda and Karnâli rivers, which is now known as British Garhwâl and Kumaon, as the latter district, before the conquests of the Gorkhas, extended to the Karnâli river. The boundary of this tract measured on the map is between 500 and Julien's Hiouen Thsang,' ii. 231. See Map No. X.