^{page 213 note 1} Life of Gaudama, 1868, p. 291.

^{page 213 note 2} Vol. IV. (N.S.), pp. 81–137.

^{page 214 note 1} Foě-kouě-ki. Translated by Rémusat, Abel. 4to. Paris, 1836Google Scholar. Travels of Fah-IIian and Sung-Yun. Translated by the RevBeal, S.. London, 1869Google Scholar.

^{page 214 note 2} Mémoires sur les contrées occidentales traduites du Sanscrit en Chinois l'an 648, par Hioueu-Thsang, etc. This work in the text I propose to call by the Chinese name Si-yu-ki, in order to distinguish it from “La vie de Hiouen-Thsang par Hoéi-li,” translated also by M. Stanislas Julien, and published in 1853.

^{page 215 note 1} The results of these researches were first published in four annual reports which General Cunningham sent to the Government of India, and were printed as they came to hand. The results were afterwards embodied by him in a work on the Ancient Geography of India, published by Trübner & Co., in 1871, which, as his maturest and last work, is the one that will be principally quoted from in the following pages.

^{page 215 note 2} In order that there may be no mistake whether I am quoting literally or merely giving the sense of a passage, I have in all instances where the former was intended quoted the original French of the text. By doing this, I have also, I hope, avoided any difficulty that might arise as to the correct interpretation of any passage in the original.

^{page 217 note 1} 329 mètres, Si-yu-ki ii. 258.

^{page 218 note 1} In a paper on the Central Asian portion of Hiouen-Thsang's travels by Yule, Colonel, which appeared in the last number of this Journal (pp. 92—120)Google Scholar, I observe that the author considers the expression 100 li as meaning only “a day's journey.” Had Colonel Yule followed the traveller's route through Bengal with the same care he has shown in examining his routes in Central Asia, I feel convinced he would have modified this statement. He would there have found distances, which are quoted in the following pages, of 3, 4, 5, or 10, 20, 30 li; longer distances are quoted, as “1400 to 1500” ( Vie, p. 185), or as “2400 to 2500” ( Vie, p. 202); and frequently the word “environ” is used by the Pilgrim in speaking of distances he himself travelled over,—all clearly showing, I think, that he spoke of lis as we speak of miles. Besides this, 100 li, or 20 miles, in a country where there were no roads, must be more nearly two days' journey than one. Of course, when our traveller speaks of the boundaries of countries he never perambulated, and routes he never traversed, his measurements must be received with the utmost caution as mere hearsay statements; and Colonel Yule may probably be correct in this, if he intends to limit his valuations only to such vague estimates of untraversed distances. But this has, I conceive, no bearing whatever on his statements regarding routes he travelled over himself, on which alone reliance is placed in the following pages.

^{page 219 note 1} Village Communities in the East and West, p. 119.

^{page 221 note 1} Si-yu-ki, i., pp. 439–456. Hoeï-li, p. 139.

^{page 221 note 2} Ancient Geography of India, pp. 455, 456.

^{page 225 note 1} Report to Government, 1861–2, pages 16 to 21. At page 20, General Cunningham twice expresses his astonishment that Hiouen-Thsang should not have visited these caves. His attention being therefore drawn to the subject, it is curious he should have overlooked these facts.

^{page 226 note 1} Hoeï-li, p. 143.

^{page 226 note 2} Vide Buchanan Hamilton's account of this place in Eastern India, vol. i., p. 95.

^{page 227 note 1} J.A.S.B., 1847, p. 956.

^{page 227 note 2} Si-yu-ki, ii., p. 382.

^{page 227 note 3} Ancient Geography, p. 468.

^{page 227 note 4} Ruins of the Nalanda Monastery, by Broadley, A. M. Esq, Calcutta, 1872Google Scholar.

^{page 227 note 5} Fa-Hian, chap. xxviii. Hoeï-li, p. 153. Si-yu-ki, ii., p. 28.et seqq.

^{page 227 note 6} Si-yu-ki, i., p. 413. Hoeï-li, p. 153.

^{page 228 note 1} Foě-kouě-ki, p. 259. Ancient Geography, p. 453.

² Si-yu-ki, ii, pp. 51–64. Hoeï-li, pp. 161–175.

^{page 229 note 1} Travels of Fah-Hian and Sung-Yun. Translated by the RevBeal, S.. London, 1869. pp. 110, 111Google Scholar.

² Montgomery Martin's Reprint, vol. i., p. 85.

^{page 230 note 1} Report to Government, 1861–2, p. 14.

^{page 230 note 2} Kittoe, loc. cit. Vivien de St.-Martin. Si-yu-ki, ii., p. 382. Cunningham, p. 469.

^{page 230 note 3} J.R.A.S. Vol. IV. (N.S.), p. 116.

^{page 231 note 1} Beal, , Fah-Hian, pp. 114–115Google Scholar. Ancient Geography, p. 473.

^{page 231 note 2} In a paper communicated to the “Indian Antiquary” (No. i., p. 18, et seqq.), Mr. Broadley clearly enough perceives General Cunningham's mistake in this matter, and points out that Fa-Hian's “isolated rock” was that in the town of Behar. But, curiously enough, he quotes the distance from Patna (9 yôdjanas) as 54 miles, and therefore quite correct. The real distance is 33 or 34, and this was General Cunningham's principal reason for taking it to Giryëk. Mr. Broadley also states (p. 19) the exact distance from Baragaon to Giryëk at 6 miles. It is more than 9. Notwithstanding these mistakes, he is right on the main point, though he fails to perceive all the discrepancies between his and Cunningham's determinations which his correction involves.

^{page 232 note 1} General Cunningham says “on an isolated hill,” p. 472, but this is a mistake.

^{page 235 note 1} Si-yu-ki, i. pp. 71–81.

^{page 236 note 1} Francklin, . Inquiry concerning the site of Palibothra, part ii., pp. 13, 26, 72, 78Google Scholar. Bose, Babu Rasbihari. Indian Antiquary, part ii., p. 51, et seqq.Google Scholar

^{page 237 note 1} Both the road and the railroad follow the course of the Ganges, and so far as I know, the former always has done so. The distance between Rajmahal and Bhagulpur being by the one 63, by the other 66 miles. I am not aware of any practical road ever having existed through the hills, and if our traveller had got entangled in these impenetrable jungles, I feel sure he would have mentioned it.

^{page 238 note 1} Six centuries afterwards we find from the Mahomedan historians that the city of Gaur was divided into two parts. One half, called Laknauti, was situated on the right or western bank of the river; the other, Deo Kote, on the left hank, near where the present city stands. The first may therefore represent the city we are looking for, and if so, it could not have been far from Rajmahal.

From the same authorities we learn that the sovereign of that day erected or repaired bunds extending for ten days' journey from Nagore, in Birbhûm, to the city of Deo Kote, to protect the country from inundation, and also to form a road for travellers through a country which was otherwise an impassable swamp during the rainy season.

page 238 note * Stewart's, History of Bengal, p. 57Google Scholar. SirElliot, H., Historians of India, ii. p. 318Google Scholar.

^{page 240 note 1} Martin, Montgomery. Eastern India, iii., p. 358Google Scholar.

^{page 240 note 2} l. c. ii., pl. 1, p. 426.

^{page 240 note 3} In the fourth volume of the Oriental Quarterly Magazine, published in 1824, there is an account of Pundra Desa, abstracted from the Brahmananda section of the Bhavishyat Purana. Among the divisions of that country therein specified, is one called Nivritti, and said to comprise Dinagepur, Rungpur, and Cooch Behar (p. 190),—consequently the whole of the Eastern half of Hiouen-Thsang's kingdom of Poundravarddhana. Its chief towns are said to be Verddhana Kuta, which may be the very town we are looking for, but then governed by a Yavana or Musselman; Kachhapa, on the banks of the Guru or Teesta river; and Sriranga or Vaharica, where the women are remarkable for flat noses. This last was no doubt the capital of Kooch Vihar, or Behar; but whether Komatipur or not, there is nothing to show:—most probably it was.

^{page 241 note 1} Asiatic Researches, xv., p. 54.

^{page 241 note 2} Greschichte des Buddhismus, pp. 99, 101, 111, 143.

^{page 241 note 3} J.A.S.B. xxiv., p. 1 et seqq.

^{page 241 note 4} J.A.S.B. xx., p. 291 tt seqq.

^{page 242 note 1} Si-yu-ki, ii., p. 80.

^{page 242 note 2} Si-yu-ki, pp. 76 to 91.

^{page 242 note 3} J.A.S.B. vi., p. 973.

^{page 243 note 1} Hunter's Orissa, i., p. 311.

^{page 243 note 2} Travels, translated by Beal, p. 147. His remark indeed is nearly final on the controversy, for he says, “Here it is that the river (the Ganges) empties itself into the sea,”—the same remark that was made by the Ayeen Akbari, and no way applicable to Tumluk.

^{page 245 note 1} Gladwin's translation, ii., p. 5.

^{page 245 note 2} I am afraid it is impossible to found any argument on the position assigned by Ptolemy to “Tamlites.” The name may be the same as Tamralipti, but calculating the distance from Palibothra, his latitude and longitude would place it near Monghir. Perhaps we may infer from this that his informants described it as distant from the sea, but that is as much as we dare.

^{page 245 note 3} J.A.S.B., 1847, p. 394 et seqq.

^{page 246 note 1} The description given by Mr. Hunter in his “Orissa” leaves no doubt but that Tumluk was an important city in ancient times, and certainly represented one of these two places.

^{page 247 note 1} From the authorities quoted ahove, p. 26, it is evident that Nagore was a place of importance in the 13th century, and that it was then necessary to protect it by embankments against inundation.

^{page 248 note 1} In a paper by Layard, Captain, in vol. xxii. of the Journal of the Asiatic Society of Bengal, it is stated, at p. 281Google Scholar, that the town of Rungamutti, situated 12 miles south of Murshidabad, stands on the site of an old city whose name was Kuru-sona-ka-ghur, which is evidently the modern Bengali translation of Karna Souvarna Nagara. The two first words at least are quite certain, and they are those which are important in the present case.

^{page 248 note 2} In the extracts from the Bhavisayat Purâna, in the 4th vol. of the Quarterly Oriental Magazine, above referred to, it is said (p. 191), “In that part of the district of Virabhumi called Viradesa is the city Nagara (Nagore), also Sipulya, and other towns. On the western borders, among others, the village of Suverna di, and on the southern confines towards the Odra country is Kinda Vishna. Suvernamukhya is in the forest,” etc. Prom this description it is hardly probable that either of these Suvernas is the one we are looking for; but it is curious to find the name Suverna twice repeated in this small district, where, from Hiouen-Thsang's indication, we would place it. The passage is also curious in bringing up the frontiers of Odra or Orissa so far north. In the following page (192) Verddhamana, or Burdwan, is distinctly described, and according to our usual notions it intervened between the two; but the compiler of the Purâna places Birbhûm and Odra in juxtaposition, as Hiouen-Thsang does.

As, however, Murshidabad was built (p. 189) when the Purâna was written, the old capital was deserted, and probably removed bodily to the new; so we should not be surprised at its not being mentioned.

^{page 249 note 1} Mr. Beames, it is true, has recently discovered and described the remains of several ancient cities in this district which were hitherto unknown. Neither from their size nor position do any of them seem to answer to our Pilgrim's description, but their existence goes far to show that others may be found when looked for.

^{page 249 note 2} Asiatic Researches, xv., p. 265. Hunter. Orissa, i.

page 249 note * Indian Antiquary, ii., p. 33; iii., p. 75; viii., p. 254.

^{page 250 note 1} J.R.A.S. xx. p. 105.

^{page 250 note 2} Ancient Geography, p. 516.

^{page 250 note 3} I am rather inclined to adopt General Cunningham's suggestion that Coringa is the Calingon of Pliny, but not for the same reasons. He states that“in book vi. ch. 23, Pliny states that the distance of this place from the mouths of the Ganges, is 625 Roman, or 574 British miles, and then that Dandagula must be Dantapura, and may with much probability be identified with Raja Mahendri, which is only 30 miles N.E. of Coringa,” etc. I find in the 20th chapter of book vi., “Abostio Gangis ad Promontorium Calingon et oppidum Dandagula DCCXXV., m. pas.,” 725, instead of 625, and certainly to the town as well as to the promontory, so it is impossible to add 30 miles without violence to the text. Nor can I admit the nominal similarity between Dandagula and Dantapura is worth anything, as taking the scene so far south would violate every probability of the narrative of the Danta dhatu wanso. It could Hot, as I read the story, be further south than Puri, and I miss the passage which says that Dantapura was situated on a great river. Notwithstanding all this, we come to the same conclusion with regard to the promontory, because I measure from Satgaon, he from Saugur.

^{page 253 note 1} If my boundary lines between the different kingdoms are correct, Hiouen-Thsang must have passed through a corner of the country of Kama Souvarna on his way between Poundravarddhana and Samatata, which may account for the mistake, but he certainly did not approach the capital on that occasion.

^{page 254 note 1} J.A.S.B. iii., 213.

^{page 254 note 2} History of Architecture ii., p. 713 et seqq. Garnier's Exploration du Mekon, i. Thomson's Photographs. (Edmonston & Douglas, 1869.)

^{page 255 note 1} Geschichte des Buddnismus in Indien, p. 262. Translated by Schiefner, . (St. Petersburg, 1869Google Scholar.)

^{page 256 note 1} Si-yu-ki, pp. 92–116. Hoeï-li, pp. 185–189.

^{page 256 note 2} The Gazetteer of the Central Provinces of India, edited by Grant, Charles Esq, 2nd edit. (Nagpore, 1870.) sub voceGoogle Scholar.

^{page 257 note 1} Beal's Fah-Hian, pp. 139, 140. Foě-kouě-ki, p. 314, et seqq.

^{page 259 note 1} Transactions R.A.S. iii., p. 470, pl. 16.

^{page 259 note 2} Si-yu-ki, ii., p. 102.

^{page 260 note 1} The suggestion that Fa-Hian meant Ellora is easily disposed of, by one among many facts that might be adduced. The oldest excavations there—the Dherwarra and Vishwaoarma—were not commenced till the seventh century, if indeed so early. The Poloyu monastery belonged to the first century of our era.

^{page 260 note 2} Gazetteer Central Provinces, Introduction, p. lvi.

^{page 261 note 1} Po-lo-mo-lo-ki-li (Paramalagiri).—Si-yu-ki, ii., p. 101.

² J.R.A.S., New Series, I., p. 264.

^{page 263 note 1} As neither General Cunningham nor I were aware, or had any means of knowing, of these local peculiarities when we wrote, neither of us are, I conceive, to blame that we accepted Daranicotta as the capital, instead of Bezwarra. But General Cunningham was certainly to blame when he called the district Majerika, and mixed up the history of the place with the legends of the Rama Grama relics and other untenable positions. Majerika is, I believe, only mentioned in the Mahawanso, and its position may be gathered from the following extracts:—When

Dutthagamini had erected the Ruanwelli Dagoha for the reception of the Kama Grama relics, which had heen carried off to Majerika by the Nagas (Mahawawo, p. 185), a priest, Sonuttara, was told off on the day previous to the ceremony to go to M ajerika and fetch them, “when he heard, for the first time, the burst of the musical sounds which announced the procession to be in motion, instantly diving into the earth, and proceeding subterraneously to the land of the Nagas, then presented himself to the Naga-raja.” The latter, fearing the relics were to be taken from him, sent his nephew off to the foot of Mount Meru with them; but the thero, making a supernaturally long arm, possessed himself of the casket. Then saying (to Kálo), “‘Naga, rest thou here,’ rending the earth, he reascended to his pariweno at Anuradhapura” (p. 188), which he reached before the procession whose start he had listened to, before descending. One point that seems to have influenced General Cunningham was the wonderful magnificence of the last step of the Naga Stupa at Majerika, as it is described in the Mahawanso (p. 188). This, however, is a peculiarity of Ceylonese architecture. All the great buildings in Anuradhapura and Pollonarua have their last steps carved in the most elaborate manner. One is represented in the third volume of the Transactions of the Royal Asiatic Society, pl. 17, and I have photographs of half a dozen others even more elaborate than that one. But no such carved step occurs at Amravati, nor, so far as I know, in any temple on the Continent of India. It is, consequently, just such a hyperbole as a Ceylonese writer would use in describing the magnificence of the underground palace of the Naga Raja, bat means nothing.

When General Cunningham assisted Mr. Prinsep in copying in vol. vi., pl. x., the Amravati inscriptions in the Calcutta Museum, both of them thought Amravati was Omrawutty in Berar. General Cunningham retained the same belief in 1848, when he described the place as west of Nagpore (J.A.S.B., xvii., p. 43), and I believe was not aware of its true situation, or of the legends bearing upon it, till I told him in 1867. Yet, though my book was published two years before his, he abstains from all allusion to it, though he himself most kindly assisted in its production, and added an appendix. This is a matter of literary etiquette, and I should not allude to it if it were a merely personal question; but it is necessary, in the interests of science, to protest against the mixing up Majerika and Rama Grama legends with the facts as stated in my work, and confounding events which took place 157 B.C. with those which occurred nearly 500 years afterwards in the beginning of the fourth century of our era.

^{page 264 note 1} Si-yu-ki, ii., pp. 116–153. Hoeï-li, pp. 189–204.

^{page 265 note 1} Hamilton, Gazetteer of India, sub voce.

^{page 265 note 2} Tod, , in his “Travels in Western India,” p. 277Google Scholar, mentions that the Mount Satrunja, or Palitana, had 21 appellations, some of which he gives. I am afraid this is far from being a solitary instance of such exuberance.

^{page 266 note 1} The Book of Marco Polo, by ColHenry Yule, C.B. (Murray, 1871.) vol. ii., p. 273Google Scholar.

^{page 267 note 1} According to an inscription partly deciphered by Dr. Bhau Daji, J.B.B.R.A.S. vol. ix., p. cxcix., the capital of Pulakesi, at the time of Hiouen-Thsang's visit, is said to have been Vatapipuri. The name is new to me, and I do not find it on any map I have access to; so, until the Doctor chooses to enlighten us, we are not much wiser than we were. The one condition, if it is to be the place at which Pulakesi was residing when Hiouen-Thsang visited him, is, that it should be within 150 or 170 miles of Baroache, which is one of the fixed points in our itineraries, and situated on the banks of a river. So new a name, however, springing up in so authentic a form, is another of the 1001 instances which occur in this investigation, to show how worthless nominal similarities are. It may either be a new site altogether, or only a new name for an old city.

It was mentioned by SirElliot, Walter, J.R.A.S., Vol. IX., p. 9Google Scholar, but even he was ignorant of its position.

^{page 269 note 1} Hoeï-li, pp. 204–206. Si-yu-ki, ii., pp. 154–162.

^{page 272 note 1} Travels in Western India, p. 268.

^{page 272 note 2} He mentions, however, one tradition, which, as far as it goes, seems to point to the truth. The sons of Govindaso, who received the image of Parsnath from the thief, resided at Puran Puttun, which was 40 coss distant from Wullee Puttun, the name by which the city is known which Tod identified with Ballabhi.

^{page 273 note 1} This seems certainly to have been the case in Al Biruni's time, for he says, “Ballaba, qui a donné aussi son nom à une ère, était prince de la ville de Ballaba, au midi de Anhalouara, à environ trente yodjanas de distance.” (Thomas's Prinsep, i., 269.) If my view is correct, the first was the old, the second the new, city of that name.

^{page 273 note 2} When an error once gets introduced in a standard work, it is copied and repeated without examination, till it becomes so completely established that it is almost impossible to get it afterwards corrected. I hope, therefore, I may be allowed again to protest against Colonel Tod's date for the destruction of Ballabhi. All he did assert was, “that the records preserved by the Jains,” apparently the Satrunjiya Mahatmaya—which for such a purpose is absolutely worthless—assert that Ballabhi was destroyed by the Parthians (properly the barbarians) in 205 of the Ballabhi Samvat, for which he adds 319, making 524 A.D. It is quite needless, therefore, General Cunningham speculating whether the event was dated from the Samvat or Saka era (p. 318). It is from the Ballabhi era, and from no other, and it rests on no inscription, no extraneous history or tradition; simply on the authority of some worthless and nameless Jain scribe, and it is contradicted by the most positive evidence, the one we are at present concerned with being quite sufficient. In 640, 116 years after its reputed destruction, Hiouen-Thsang visited it, and found, as one would expect from the chronology (J.R.A.S., new series, iv., p. 90), that a King Dhrouvapatu was on the throne, and no symptom of decay was visible in its long period of prosperity.

The truth seems to be, and Tod himself admits it (pp. 149–152), that Ballabhi was destroyed, and Anhulwara-puttun rose on its ruins in the Samvat year 802, or A.D. 746. Every scrap of information we have, every inscription, every tradition, point to the middle of the eighth century as the period of a revolution which changed the face of India. The great Chalukya and Canouge dynasties were then eclipsed—the Moslems, who were fast advancing to the conquest of India, were driven back to the Indus—Buddhism practically disappeared, and Siva and Vishnu take the place of the mild ascetic. The only written record of this great event is a distorted paraphrase of it in the Raja Tarangini, where it figures as the Indian wars and conquests of Lali aditya (A.R. xv., p. 44 et seqq.); but Elliot's inscriptions (J.R.A.S. iv., 7, et seqq.), and above all Tod's Annals, everywhere point to this as the end of the old and the beginning of the new state of things in India. It may have been, as above stated, in 746, or it may be 10 years later; but till we have more exact information, it will be better to speak in round numbers, and take 750 A.D. as the date of the great revolution.

page 273 note * See also Travels in Western India, p. 149.

^{page 724 note 1} If Dr. Stevenson's conjecture could be maintained, it would be an easy solution of the difficulty. In J.B.B.R.A.S., vol. viii., p. 51, he says, “I suppose it refers to Anandapoora, so called by the Jains, and afterwards by the Hindus named Valabhi, the capital of the well-known sovereignty of that name, on ihe north side of the Gulf of Cambay.” The objection to this theory is that both Hoeï-li (p. 207) and the Sirju ki (ii., p. 164) mention Anandanoura as a separate capital 700 li N.W. from Ballabhi. The whole, however, it appears to me, is only another instance of perpetual shuffling of names of cities which are so frequently met with in these inquiries, and which, in fact, is what makes them so difficult.

² Thomas, , Prinsep's Indian Antiquities, i., p. 252, et seqq.Google Scholar