elephant capital to any other site. It is, however, quite possible that the capital may have been stopped on its way to the temple of Mahadeva, near the Nága mound and tank. The temple of Bisári Devi would then be the site of one of the ten ancient Brahmanical fanes which are described by Hwen Thsang. Altogether, this is, perhaps, a more probable solution of the difficulties of the case than that first described.*

In his description of Sankisa, Hwen Thsang mentions a curious fact, that the Brahmans who dwelt near the great monastery were "many tens-of-thousands" in number. As an illustration of this statement, I may mention that the people have a tradition that Sankisa was deserted from 1800 to 1900 years ago, and that 1300 years ago, or about A. D. 560, it was given by a Kayath to a body of Brahmans. They add also that the population of the village of Paor-Kheria is known to have been wholly Brahman until a very recent period.

X. KANOJ.

Of the great city of Kanoj, which for many hundred years was the Hindu Capital of Northern India, the existing remains are few and unimportant. In A. D. 1016, when Mahmud of Ghazni approached Kanoj, the historian relates that "he there saw a city which raised its head to the skies, and which in strength and structure might justly boast to have no equal."+ Just one century earlier, or in A. D. 915, Kanoj is mentioned by Masudi as the capital of one of the Kings of India, and about A. D. 900 Abu Zaid, on the authority of Ibn Wahab, calls "Kaduge, a great city in the kingdom of Gozar." At a still earlier date, in A. D. 634, we have the account of the Chinese pilgrim Hwen Thsang, who describes Kanoj as being 20 li, or 33 miles, in length, and 4 or 5 li, or three-quarter of a mile, in breadth. The city was surrounded by strong walls and deep ditches, and was washed by the Ganges along its eastern face. The last fact is corroborated by Fa Hian, who states that the city touched the River Heng (Ganges) when he visited it in A. D. 400.§

I have already noticed, p. 272, that the Barmese Life of Buddha fixes the point of descent at the "gate of the city," and this position seems also to be indicated by the still existing name of Paor-kheria, or "Staircase Village," which is situated just outside the south-east opening, or gate, in the earthen ramparts.-See Plate No. XLV.

+ Briggs's Ferishta, I., p. 57.

Julien's Hwen Thsang, II., p. 243.

§ Beal's Fa Hian, C. XVIII., p. 70.

Kanoj is also mentioned by Ptolemy, about A. D. 140, as Kanogiza. But the earliest notice of the place is undoubtedly the old familiar legend of the Purânas, which refers the Sanskrit name of Kanya-Kubja, or the "hump-backed maiden" to the curse of the sage Vayu on the hundred daughters of Kusanabha.

At the time of Hwen Thsang's visit, Kanoj was the capital of Raja Harsha Vardhana, the most powerful sovereign in Northern India. The Chinese pilgrim calls him a Fei-she, or Vaisya, but it seems probable that he must have mistaken the Vaisa, or Bais, Rajput, for the Vaisya, or Bais, which is the name of the mercantile class of the Hindus; otherwise Harsha Vardhana's connexion by marriage with the Rajput families of Malwa and Balabhi would have been quite impossible.* Baiswâra, the country of the Bais Rajputs, extends from the neighbourhood of Lucknow to Khara Manikpur, and thus comprizes nearly the whole of Southern Oudh. The Bais Rajputs claim descent from the famous Sáliváhan, whose capital is said to have been Daundia-Khera, on the north bank of the Ganges. Their close proximity to Kanoj is in favour of the sovereignty which they claim for their ancestors over the whole of the Gangetic Doab from Delhi to Allahabad. But their genealogical lists are too imperfect, and most probably also too incorrect, to enable us to identify any of their recorded ancestors with the Princes of Harsha Vardhana's family.

The vast empire which Harsha Vardhana raised during his long reign of 42 years, between A. D. 607 and 648, is described by Hwen Thsang as extending from the foot of the Kashmir Hills to Assam, and from Nepal to the Narbada River. He intimidated the Raja of Kashmir into surrendering the tooth of Buddha, and his triumphal procession from Pâtaliputra to Kanoj was attended by no less than 20 tributary Rajas from Assam and Magadha on the east, to Jâlandhar on the west. In the plenitude of his In the plenitude of his power, Harsha Vardhana invaded the countries to the south of the Narbada, where he was successfully opposed by Raja Pulakesi, and after many repulses was obliged to retire to his own kingdom. This account of Hwen Thsang is most singularly

* I have no doubt on this subject now (1871), as there is Indian Sanskrit authority for the intermarriage with the Malwa family.

corroborated in every particular by several ancient inscriptions of the Chalukya Rajas of Kalyan. According to these inscriptions, Raja Vikramaditya, the grandson of Pulakesi Vallabha, gained the title of Parameswara, "by the defeat of Sri Harsha Vardhana, famous in the north countries." Now Vikramaditya's reign is known to have commenced in Sake 514, or A. D. 592, as one of his inscriptions is dated in Sake 530, or A. D. 608, which is called the 16th year of his reign ;† and as his grandson did not succeed to the throne until the Sake year 618, or A. D. 696, it is certain that Vikramaditya must have been a contemporary of Harsha Vardhana throughout the greater part, if not the whole, of his reign. The unusually long reigns of the earlier Chalukya Princes have led Mr. Walter Elliot to suspect the accuracy of the dates, although, as he points out, "the succeeding dates tally with each other in a way that affords the strongest presumption of their freedom from any material error." ." The question of the accuracy of these dates is now most satisfactorily confirmed by the unimpeachable testimony of the contemporary record of Hwen Thsang, which I have quoted above.

In determining the period of Harsha's reign, between the years 607 and 648 A. D., I have been guided by the following evidence: 1st, the date of his death is fixed by the curious reported fulfilment of Hwen Thsang's dream,‡ and by the report of the Chinese embassy§. 2nd, in speaking of Harsha's career, the pilgrim records that from the time of his accession Harsha was engaged in continual war for 5 years, and that afterwards for about 30 years he reigned in peace. This statement is repeated by Hwen Thsang, when on his return to China, on the authority of the King himself, who informed him that he had then reigned for upwards of 30 years, and that the quinquennial assembly then collected was the sixth which he had convoked. From these different statements it is certain that at the date of Hwen Thsang's return to China, in A. D. 640, Harsha had

*Bombay Asiatic Society's Journal, III., 206.

p. 569.

Royal Asiatic Society's Journal, IV., 10.

See the discussion on this date in my "Ancient Geography of India,” Appendix,

§ Journal, "Asiatic Society," Bengal, 1837, p. 69,-anonymous translation. See also Journal Asiatique, 1839, p. 398, French translation by M. Pauthier.

reigned upwards of 30 years, and somewhat less than 35 years. His accession must, therefore, be placed between A. D. 605 and 610. 3rd, now, in the middle of this very period, in A. D. 607, as we learn from Abu Rihan, was established the Sri Harsha era, which was still prevalent in Mathura and Kanoj in the beginning of the 11th century. Considering the exact agreement of the names and dates, it is impossible to avoid coming to the conclusion that the Harsha, who established an era in Kanoj in A. D. 607, was the great King Harsha Vardhana who reigned at Kanoj during the first half of the seventh century.

Hwen Thsang adds some particulars regarding the family of Harsha Vardhana, which induce me to think it probable that it may be identified with one of the dynasties whose names have been preserved in the genealogies of the Râjavali. The names differ in the various copies, but they agree generally in making Raj Sing, who reigned only nine years, the predecessor of Hara or Hari Sing, who is recorded to have reigned for 44 or 45 years. Now, according to Hwen Thsang, the predecessor and elder brother of Harsha Vardhana was Rajya Vardhana, who was assassinated shortly after his accession. Here both the names of these two Kings and the lengths of their reigns agree so well together as to suggest the probability of their identity. In most copies of the Rajavali, this dynasty of six Kings, of which Raja and Hara are the 3rd and 4th names, is made the immediate predecessor of the Great Tomar dynasty, whose accession has already been assigned in my account of the Kings of Delhi to the year 736 A. D. The following lists give the names of all the Kings of this dynasty according to the various authorities in my possession:

According to Sayid Ahmad the accession of Shir Sing, who is the Hara or Hari of the other lists, took place in A. D. 611,