^{page 641 note 1} See my article “Sāgala, Śākala, the City of Milinda and Mihirakula” in the Transactions of the Fourteenth Oriental Congress, Algiers, 1905; Indian Section, page 164 ff.

^{page 642 note 1} Memoirs of Baber, trans, by Leyden and Erskine, 413; trans, by De Courteille, 2. 413 f.

^{page 643 note 1} The names of the two kingdoms are not given there. But we know well that one of them was Magadha. And we infer that the other was Anga, from a passage in the Dīgha-Nikāya, 4. 1 (ed. Davids and Carpenter, 1. III ): we are there told that Buddha, in the course of his wanderings in the Anga country, came on a certain occasion to Champā, and that the Brāhman Sōnadanda was then dwelling at Champā in a brahmadeyya, a grant to Brāhmans, which was a rājadāya, a royal allotment, that had been given by the Māgadha king Bimbisara (father of Ajātasattu); and Bimbisāra could hardly have made such an allotment there, unless the Anga country belonged to him.

^{page 643 note 2} The translator has taken the text in that way, and has said (The Jātaka, 1. 22):—“And that though the walls of Benares were 12 leagues round by themselves, while the city and suburbs together were 300 leagues round! ” But there is no reference in the text to “suburbs” at all; the word raṭṭha, ‘country,’ precludes the idea of any such intention; and the text further on (p. 126, line 3) distinctly specifies Bārānasi-rajjam, “ the kingdom of Bārānasī.” Another writer has taken over the idea of “city and suburbs,” but has treated both the city and its suburbs as included in the 12 yōjanas: see more fully pages 647–8 below.

^{page 644 note 1} See the commentary on the Nidānakathā, in the Jātaka, ed. Fausböll, 1. 49. Compare the Sumangalavilāsinī, ed. Davids and Carpenter, 1. 173:—Ayāmatō ti-yōjana-satē vitthāratō aḍḍhatiya(v.l. °teyya)-yōjana-satē parikkhēpato navayōjana satē Majjhimapadēsē.

^{page 644 note 2} See, fully, note 3 on page 653 below.

^{page 644 note 3} Pargiter, in JASB, 66, 1897. 86 ff.; with a full statement of details, authorities, and previous views.

^{page 645 note 1} It would obviously be unpractical to coEsider the ancient divisions of India as circles. They might be treated as hexagons or as triangles; but squares and rectangles will suffice for the purposes of a tentative note.

^{page 646 note 1} See Rawlinson's translation of Herodotus (1875), 1. 298, notes.

^{page 647 note 1} We have, for instance, the case of Pāṭaliputra, cited further on. In its case, indeed, there was the important river frontage along the Ganges, with its facilities not only for traffic but also for water-supply, to account for a great length of city having been laid out or having grown up. But Taxila also, where there seems to have been nothing in the shape of a real river to account for such a feature, seems to have been of much greater length than width. And Vaisall, otherwise called Viśālā, “the spacious, broad, or wide,” may be another case to the point.

^{page 648 note 1} See, e.g. IA, 6. 131; McCrindle's Ancient India, 42.

^{page 648 note 1} See Cunningham's Ancient Geography of India, 452.

^{page 650 note 1} Julien proposed, doubtfully, Mulasambhuru; Mémoires, 2. 173. Watters has said that the word must be more like Morasampuru or Molasampul, and need not be pure Sanskrit; On Yuan Chwang, 2. 255.

^{page 651 note 1} The other parts of the circuits of M. and P. of course lay in the area of Cheh-ka; and the eastern boundary of M. coincided more or less with the western boundary of P.

^{page 652 note 1} Compare the Mahāvastu, ed. Senart, 3. 204 ff.; where, however, the end of the story is not so full.

^{page 652 note 2} Neither version of the story, however, indicates which was his territory. The Dipavamsa, 3. 40, mentions both Disampati and Rēnu, and places them among the kings either of Bārānasī or of Kapilavatthu according as we take the context. The Lalitavistara presents a confused reminiscence of Disampati, in describing Rēnu as Rēnu-bhūdiśāmpati, “Rēnu, lord of the earth and of the cardinal points thereof” (ed. Lefmann, 171, line 1): it represents him as a previous incarnation of Buddha; and perhaps the verse, taken as a whole, would mark him as a king of Kāī (Bārānasī).

^{page 652 note 3} The article entitled “Note on the Middle Country of Ancient India,” published in this Journal, 1904. 83 ff., can hardly be regarded as particularly useful. The writer of it has started by placing the eastern limit of the country, the town Kajaiigala, which he would locate about 65 to 70 miles on the east of Champā, Bhāgalpūr, in lat. 25°, long. 98° (pages 88, line 2; 90, line 5f.). Here, the 98° is of course, by origin, only a careless mistake for 88°; Bhāgalpūr being in 87° 2′. But it is no misprint: it is adopted as a fundamental error throughout the article; and, after locating the western limit of the country in lat. 31°, long. 73° (page 89, line 21), the writer has expressly told us (page 91, line 8 ff.) that “the extreme points in the east in long. 98°, and in the west in long. 73°, are 25 degrees apart” (by mistake for 15). The immediately following statement, that these 25 degrees of longitude = about 1750 miles, discloses another fundamental error; namely, an assumption that each of them represents about 70 statute miles (the real value in those latitudes, even on the assumed slant, being only about 64 miles, and the difference, about 150 miles on the 25 degrees, being a very appreciable amount). Alongside of that, the writer has told us (page 91, line 21 f.) that he had determined the value of the yōjana at “about 7 to 7¼ miles” in a previous writing (where, however, the result actually propounded by him is “between 7 and 8 miles” ). Having fixed the position of Kajangala, the writer has sketched out a boundary of the country, starting from that place and returning to it, which, he says (page 91), measures in straight lines about 3500 miles: we need not try to check this; the result would be unimportant. Then (apparently realizing, in connexion with what was to follow, that 3500 miles, if applied as a perimeter to 1750 miles, could only enclose 1750 miles of ‘length without breadth’), he has proceeded to say that, “as the boundary, both on the north and the south, follows the very irregular contour of the mountains, a traveller along the route would probably have to go twice that distance.” The perimeter is thus raised to about 7000 miles. There is then announced the denouement, that 25 degrees (taken at about 70 miles each) = “about 1750 miles in a straight line” = “exactly” 250 yōjanas (taken at about 7 miles each). And we are finally told that the given circumference is “in practical accord” with the perimeter of about 7000 miles, obtained as shewn above; because the 900 yōjanas, taken at about 7 to 7¼ miles, would = from 6300 to 6500 miles. It need only be added that, taken at 7 miles each, the 250 yōjanas, = 1750 miles, would make the breadth of the Middle Country reach— (following a slanting line, as desiderated by the writer of the article)— from the selected point in lat. 31°, long. 73°, past Bhāgalpūr, on to some 200 miles beyond Mandalay. While the 300 yōjanas, = 2100 miles, would make the length of the country reach from Śrinagar in Kashmīr to a point out at sea some 280 miles south of Cape Comorin.

The points given in the Vinaya-Pitaka, Mahāvagea, 5. 13, 12 (ed. Oldenberg, 1. 197), as marking the limits of the Buddhist Middle Country, are, no doubt, authentic and reliable: namely, Kajangala and Mahāsālā on the east; the river Sallavatī on the south-east; Setakannika on the south; Thūna on the west; and the mountain Usīraddhaja on the north. And Usīraddhaja is, no doubt, Uśīnaragiri, a mountain on the north of Kankhal, which is just below Hardwar: see Hultzsch in IA, 1905. 179. But the details about the length, breadth, and perimeter, added in the Sumangalavilāsinī and the Commentary on the Nidānakathā, are plainly imaginative and preposterous. Even with the true value, 1 yōjana = 4.54 miles, the 250 yōjanas, = 1136.36 miles, would make the Middle Country reach from Multan through Hardwār almost to Lhasa; and the 300 yōjanas, = 1363.63 miles, would make it reach from Hardwār to Pudukattai below Tanjore. we may, however, perhaps infer that the inventor of those details had a fairly good idea of the general shape of India.

Still more extravagant than the above are the measures reported by I-tsing (see this Journal, 1904. 538, and Takakusu's translation of I-tsing, introd., 33); namely, about 300 yōjanas from east to west, and 400 from north to south.

^{page 654 note 1} This is the mean diameter, from 7899·37 miles for the polar diameter and 7925·82 for the equatorial; see, e.g., Bonney's Story of Our Planet, 9,