^{page 438 note 1} See Cunningham, Ancient Geography of India, p. 404, for references.

^{page 438 note 2} Chap, v, śl. 31.

^{page 438 note 3} Buddhist India, p. 39.

^{page 438 note 4} Griffith's transl., book i, canto v.

^{page 439 note 1} Anc. Geog., p. 405.

^{page 439 note 2} Bigandet: Life of Gaudama, p. 142, quoted Anc. Geog., p. 404.

^{page 440 note 1} Buddhist India, p. 39.

^{page 440 note 2} Asiatick Researches, 1836, vol. xx, p. 72.

^{page 440 note 3} Certainly Sāketa.

^{page 441 note 1} I suppose to the north of the range for these reasons. Setavya lay on the road from Śrāvastī to Kapilavastu (S.B.E., vol. x, p. 188 ; Buddhist India, p. 103). Therefore it might be supposed that Setavya lay to the south-east of Śrāvastī city, as Kapilavastu city lay in this direction from Śrāvastī (J.R.A.S., 1903, p. 98). But Setavya really stood to the west or north-west of Śrāvastī city (J.R.A.S., 1903, p. 513). To reconcile the seeming discrepancy in the Pālī and Chinese accounts of the bearing from Śrāvastī city to Setavya, I believe that the former city will be discovered within the hills, and that to avoid crossing the southernmost range when travelling to Kapilavastu it was found easier going, and perhaps in other ways more convenient, to follow the Rāpatī River westward from Śrāvastī as far as Setavya, and so reach level country. I should expect to find the ruins of Śrāvastī city at approximately lat. 28° 3′ and long. 82°. The supposed site of Śrāvastī visited by Mr. Vincent Smith and myself (J.R.A.S., 1898, p. 527) I am now inclined to think may have been that of Setavya, the Too-wei of Fa-hian, which lay 60 li, or just short of 8 miles (J.R.A.S., 1903, pp. 98, 102), to the westward of Śrāvastī city. The distance between the cities of Kapilavastu and Srāvastī was 740 li (800 – 60 li) or 97·8 miles. The site we visited is apparently too far to the north-west of the rains of Kapilavastu to correspond to Śrāvastī city.

^{page 441 note 2} J.R.A.S., 1900, p. 15.

^{page 442 note 1} J.R.A.S., 1903, p. 98.

^{page 442 note 2} Arohæology in India, p. 110.

^{page 442 note 3} J.R.A.S., 1898, p. 524.

^{page 442 note 4} J.A.S.B., 1900, p. 75. Pasha is a misprint. There is another place named Pasakā, 14 miles from Jaunpur, on the road to Azamgaḍh.

^{page 442 note 5} p. 274.

^{page 444 note 1} Route Book, 1900.

^{page 444 note 2} J.A.S.B., 1895, p. 71.

^{page 444 note 3} Al Bīrūnī (Sachau : India, p. 201) gives the distance from Bārī in the Sītāpur District through Dugām, which he calls Dūgum, to what I think was Śrāvastī, ‘the city of Bhut,’ that is the city to which Bhuṭṭiya traders came from Tibet (Bod) and Nepal. Śrāvastī appears slightly disguised as , Sarrast (Elliot: History of India, vol. i, p. 57, and note 2), an error, I suppose, for , Sarvast, or Śrāvastī. He gives the distance from Bārī to Dūgum as 45 farsakhs, an error, I believe, for 45 Arabian miles, that is, 45 x 2148 English yards, or 54–92 English miles, which distance must be about right, but is perhaps a little, not more than a mile or two, in excess of the actual distance to Dugām. From Dūgum to Sarrast was 22 farsakhs,—again I read Arabian miles,—or 26·85 English miles. From the east side of the Dugām ruins to Nānpārā is about 4 miles; Nānpārā to Bkī by road is 17 miles (Route Book), leaving 5·85 English miles onwards to the city of Sarrast, which must have been situated nearly in the position chosen by Mr. Vincent Smith for Śrāvastī, if Al Bīrūnī's informants' figures are reliable, and my interpretation of the real distances is correct.

^{page 443 note 4} Dutt: A History of Civilization in Ancient India, vol. iii, p. 206. The Sanskrit passage, with an English translation, from Dr. Kern's Bṛhat Saṁhitā, will be found in Growse's Mathura· (1874 ed.), p. 73.

^{page 445 note 1} Wilson (Hall): Vishńu Puráiáa, vol. iv, p. 218 ; quoted in Indian Antiquary, 1902, p. 258, note 7.

^{page 445 note 2} Legge has 3 yojanas instead of 10 yojanas; but I learned from Mr. Thomas Watters that 10 yojanas is the correct distance, and that 3 yojanas is a blunder. Even if we allow that 6 yojanas (3 + 3) is the distance to the city of Sāketa from Kanauj, and that Sāketa is represented by Sañcānkoṭ, the distance of 42·3 miles ;(6 yojanas × 7·105 miles) cannot he made to agree, as Sañcānkoṭ is 9·75 miles by road from Bgarmaū, and exactly 30 miles from Kanauj viâ Sarāi Mīran and Nānāmaū-ghāṭ (Route Book, Nos. 266, 418a, and 154a).

^{page 446 note 1} Neval, not Naval, is the correct spelling.

^{page 446 note 2} Hardy: Manual, p. 247.

^{page 446 note 3} See Mon. Antiq. for these places.

^{page 447 note 1} A.S.R., vol. xi, p. 64, and pl. xix.

^{page 447 note 3} A.S.R., vol. xi, p. 63.

^{page 447 note 3} A.S.R., vol. xi, p. 48.

^{page 448 note 1} Beal, Si-yu-Ki, vol. i, p. 230, and Life, p. 90. I state, J.R.A.S., 1904, p. 249, that the Jamunā and Ganges were intended by “two branches of the river,” but I now see that this was a mistake.

^{page 448 note 2} Beal, vol. i, p. 224.

^{page 448 note 3} Wilson: Mackenzie Collection, p. 176.

^{page 449 note 1} Mémoires, vol. i, p. 290.

^{page 449 note 2} Op. cit., vol. ii, p. 522.

^{page 449 note 3} Méthode, No. 1410.

^{page 449 note 4} Vāi in Vāiśākha is noted (Méthode, No. 1410) as occurring also in Vaiśya and Vaiśravaṇa, from which it is seen that a may be a or ā. So in Pi-so-kia may represent sa, sā, so, sya, ṣa, or svā (Méthode, Nos. 1597–1603), but Julien gives no example of so equal to śa or śā, although he has restored the name to Vāiśakha. Kia may be long as in Kāśyapa, or short as ia Śeaśāṅka (Mémoires, vol. ii, p. 509, and Méthode, No. 582).