^{page 84 note 1} Exact number uncertain. Diod. XIX, 14, 8 gives 120; id. 28, 4 (Paraitakene) 114, but the detailed numbers add up to 125.

^{page 84 note 2} Jerome's 400 and Appian's 300 (Prooim. 10) for Ptolemy II are worthless; the Appian passage is full of gross exaggerations. But his statement (Lib. 95) that Carthage had stabling for 300 might perhaps be true.

^{page 84 note 3} Polyb. XI, 34, 12.

^{page 84 note 4} Przyluski, J., however, in La légende de l'Empereur Açoka (Açoka–Avadāna), 1923Google Scholar, has in his index a considerable special entry under cinq cents.

^{page 84 note 5} There must of course be a very great deal more. But what I give should suffice.

^{page 84 note 6} I am indebted to my daughter for going through Jātakas, Cowell's (The Jātaka, trans, by various hands, ed. Cowell, E. B., 6 vols., 1895–1907) and taking out the figuresGoogle Scholar.

^{page 84 note 7} Trans, by Davids, T. W. Rhys, 2 vols., in Sacred Books of the BuddhistsGoogle Scholar.

^{page 84 note 8} Trans, by Davids, T. W. Rhys, 2 vols., in Sacred Books of the East, XXXV, XXXVIGoogle Scholar.

^{page 84 note 9} Trans, by J. Przyluski, op. cit., from the Chinese translation; extracts only from the Sanscrit original survive.

^{page 85 note 10} Jātaka, I, 1, 16, 27, 33, 35, 100, 109, 115, 159, 173, 207, 237, 241, 260, 273, 274, 305. 321Google Scholar; II, 28, 65, 119 (twice), 239; III, 12 (twice), 14, 35, 76, 83, 93, 94 (twice), 113, 154 (twice), 240, 319; IV, 9, 47, 61, 74, 91, 95, 168, 189, 205, 227, 268, 275; V, 66, 136, 219, 254; VI, 20 (often repeated). Dialogues of the Buddha, II, 237, 286, 287, 349Google Scholar. Milindapañha, II, 227 (160)Google Scholar. Przyluski, op. cit. 309, 403.

^{page 85 note 11} Jātaka, III, 281Google Scholar; IV, 227; V, 22. Przyluski, op. cit. p. 309.

^{page 85 note 12} Przyluski, op. cit. p. 406.

^{page 85 note 13} Jātaka, I, 6, 7, 10, 73, 135, 212, 220, 239Google Scholar; II, 90; III, 132, 243; IV, 221, 222; V, 12, 42, 86, 256; VI, 20. Dialogues of the Buddha, II, 139, 141Google Scholar. Milindapañha, I, 27 (17), 289 (202)Google Scholar.

^{page 85 note 14} Jātaka, I, 13, 253Google Scholar; II, 290, 294; IV, 9, 157; VI, 54, 123. Przyluski, op. cit. p. 387.

^{page 85 note 15} Jātaka, I, 52, 64, 106, 109 (2 cases), 166 (3 cases), 314Google Scholar; III, 228; IV, 98, 101; VI, 40. Milindapañha, I, 289 (202)Google Scholar. Przyluski, op. cit. p. 401.

^{page 85 note 16} Jātaka, I, 19, 64, 255Google Scholar; II, 14, 65; III, 261; IV, 102. Przyluski, op. cit. pp. 392, 402.

^{page 85 note 17} Jātaka, I, 121, 185Google Scholar; II, 264; III, 243; IV, 268. Przyluski, op. cit. p. 390.

^{page 85 note 18} Jātaka, III, 40, 261Google Scholar; V, 5; VI, 140.

^{page 85 note 19} Id. II, 65, 278, 287; IV, 32, 71, 183.

^{page 85 note 20} Id. I, 273; V, 7. Przyluski, op. cit. pp. 415, 417.

^{page 85 note 21} Jātaka, II, 65, 218Google Scholar; IV, 94; VI, 253. Dialogues of the Buddha, I, 65Google Scholar.

^{page 85 note 22} Jātaka, II, 267 (warriors)Google Scholar; IV, 94 (chariots); VI, 68 (archers).

^{page 85 note 23} Id. II, 64 (nymphs), 94 (goblins).

^{page 85 note 24} Foucher, A., Notes sur la géographie ancienne du Gandhāra, p. 17Google Scholar.

^{page 85 note 25} Jātaka, V, 221Google Scholar.

^{page 85 note 26} Milindapañha, I, 182 (121)Google Scholar.

^{page 85 note 27} Id. I, 130 (85); Jātaka, V, 250Google Scholar.

^{page 85 note 28} Jātaka, VI, 86Google Scholar.

^{page 85 note 29} Milindapañha, I, 171 (114)Google Scholar; II, 129 (283).

^{page 85 note 30} Demiéville, P., Les versions chinoises du Milindañha, BEFEO, XXIV, 152 (LXXXVI)Google Scholar.

^{page 85 note 31} Horses, Jātaka, I, 19, 22Google Scholar; II, 22, 66. Asses, id. II, 67. Oxen, id. I, 73. Dogs, id. III, 319. Monkeys, id. III, 218; Przyluski, op. cit. p. 309. Stags and deer, Jātaka, I, 35, 39Google Scholar; Przyluski, op. cit. 293 (in the Avadāna of Kunala).

^{page 85 note 32} Dialogues of the Buddha, I, 65Google Scholar.

^{page 85 note 33} Jātaka, II, 15Google Scholar.

^{page 85 note 34} Dialogues of the Buddha, I, 108, 144, 173, 288, 300Google Scholar; II, 284. Milindapañha, I, 298 (207)Google Scholar. Przyski, op. cit. p. 266.

^{page 86 note 35} Dialogues of the Buddha, II, 180, 182, 183, 185, 186Google Scholar.

^{page 86 note 36} Milindapañha, I, 186 (131)Google Scholar.

^{page 86 note 37} On the treaty itself see Tarn, , The Greeks in Bactria and India, 1938, p. 174, n. 3Google Scholar, and on what Seleucus actually ceded, id. p. 100. ******Appian, Syr. 55 ultimately goes back to the same source, but does not mention the elephants. Fresh light has since been thrown on the circumstances by Newell, E. T., The coinage of the Eastern Seleucid mints from Seleucus I to Antiochus III, 1938Google Scholar, on which see my review in JHS 1939; but it is not material here.

^{page 86 note 38} It is strange that O. Stein, in his enormously full article Megasthenes in PW, never considers the source of this treaty, though he ascribes to Megasthenes almost every mention of anything Indian in classical literature between the Alexander-historians and Claudius Ptolemy.

^{page 87 note 39} FHG II, frs. 20, 21, p. 416, and the end of fr. 23, p. 420.

^{page 87 note 40} The argument that follows would apply equally to any other early Greek writer on India.

^{page 87 note 41} Though Seleucus could never have manœuvred 500, or 400, elephants on the spur of the moment, I am not using this passage as an argument, as it is possible that he laid a trap: his cavalry may have had orders to fly, which, as they were unbroken, would inevitably draw Demetrius on and enable the elephants, who were waiting for this, to intervene.

^{page 87 note 42} I take it to be a sound rule that if two texts, derived from a common source, give, on one matter, conflicting accounts, then what the common source gave on that matter is, in the absence of evidence aliunde, unknown.

^{page 87 note 43} The fragment of a Babylonian Chronicle concerning Antiochus I, in Smith, Sidney, Babylonian Historical Texts, 1924Google Scholar. I gave a full account in JHS XLVI, 1926, p. 157Google Scholar.

^{page 87 note 44} Lucian, , Zeuxis IIGoogle Scholar; cf. Suidas, , and the Myrina terracotta which I cited, op. cit. p. 157.

^{page 88 note 45} Pint., Dem. XXVGoogle Scholar.

^{page 88 note 46} A Maurya, not a local ruler: Tarn, , Bactria and India, p. 130 and notesGoogle Scholar.

^{page 88 note 47} Polyb. XI, 34, 12: . It is not clear whether ἄπαντας means those he had in India at the moment, or all he possessed; but it makes little difference, for nothing shows that when he met Sophagasenos he had any in the West. The 102 he had at Raphia must have been lost there; the demotic version of the trilingual stele concerning Raphia, l. 14, says they were all captured (Spiegelberg, , SB Bayer Ak. 1925, Abh. 4Google Scholar; the Greek version has a lacuna). Spiegelberg, p. 19, indeed says that this cannot be taken literally in view of Polybius' statement (V, 85, 5) that Antiochus had only 5 elephants killed; but this, even if true, only refers to his breaking of Ptolemy's left (which he continued to claim as a personal victory), as is shown by the added statement that he captured most of Ptolemy's elephants, whereas Ptolemy's right had been completely victorious. If one reads Polybius'. account of the battle as a whole, one sees that Antiochus could not have saved many elephants (his own or captured), if any, in the universal flight at the end; the inscription, in saying he lost them all, is substantially correct.

^{page 88 note 48} Polyb. XI, 34, 10.

^{page 88 note 49} Tarn, , Bactria and India, p. 309Google Scholar.

^{page 89 note 50} Megasthenes (named ) in Arr., Ind. VIII, 6Google Scholar; Pliny, , H.N. VI, 76Google Scholar, gives the same story. But there is nothing to warrant the belief, which Schwanbeck started (though he himself treated the matter as uncertain) that all the vires gentium in Pliny VI, 66 sqq. are from Megasthenes; the Heracles story stands apart, since it is nothing real, and Pliny, as Müller noticed (FHG II, p. 410, note at the end of fr. 8), was using a number of sources, as he so often did.

^{page 89 note 51} Polyaenus, IV, 9, 3.

^{page 89 note 52} Tarn, , JHS XLVI, p. 157Google Scholar.

^{page 89 note 53} FHG III, p. 474, fr. 30 = Aelian, , N.An. IX, 58, Google Scholar.

^{page 89 note 54} Id. p. 476, fr. 34 = Philostratus, , Vita Apoll. 13Google Scholar. The brand, τοὐπίσημον, on the tusk was .

^{page 89 note 1} Tarn, , The Greeks in Bactria and India, 1938, pp. 118sqGoogle Scholar.

^{page 89 note 2} It comes from a Thibetan translation of a Sanscrit work: Lévi, S., Journ. Asiatique, 1933, p. 271, n. 1Google Scholar. The original mentions a town Dharmamitra (Demetrias), and the translator says that the name was the origin of Tarmita on the Paksu (Oxus). See further below.

^{page 90 note 3} Ptol. VI, 12, 6.

^{page 90 note 4} Barger, E., Geographical Journal, XCIII, 1939, p. 387Google Scholar.

^{page 90 note 5} Tarn, op. cit. chs. IV–VI, esp. pp. 131–2, 181.

^{page 90 note 6} References for these cities, Tarn, op. cit., Index under Alexandria.

^{page 90 note 7} The Peutinger Table also gives a place Thimara in Orissa, near the mouths of the Ganges (see Miller, K., Itineraria Romana, 1916Google Scholar, maps 255, 256 and col. 789), which has sometimes been confused with Tharmata; consequently Tomaschek, , Antiochia, 13 in PWGoogle Scholar, and Tscherikower, V., Die Hellenistische Städtegründungen, p. 111Google Scholar, said that Antioch Tharmata was near the Ganges mouth (a strange place for an Antioch), though Droysen, , Hellenismus,² III, 2, p. 323Google Scholar, had already shown that this idea was wrong. Thimara has nothing whatever to do with Antioch Tharmata.

^{page 90 note 8} Ravennatis Anonymi Cosmographia, ed. Pinder, M. and Parthey, G., p. 141Google Scholar.

^{page 90 note 9} Miller, op. cit. col. 798.

^{page 90 note 10} Id. col. 795. Miller himself noted this.

^{page 90 note 11} Tarn, op. cit. p. 118, n. 8: ‘Antioch Tharmata … possibly a conflation of Antioch–Merv and Demetrias–Termedh which has been misplaced.’ This is wrong. I also think now that it was probably wrong to connect Tharmata with Stephanus' ἄρματα πόλις Ἰνδικῆς; for some MSS of Diodorus, XVII, 103, 1, give Ἄρματα instead of Ἁρματήλια, a town in Sambos' country on the lower Indus.

^{page 90 note 12} Alexander had traversed the southern part, from Herat to Zarin; and the Suren, in the liquidation of the Saca invasion of 129 B.C., had driven the enemy back along the northern part, Herat–Traxiane–Merv: Tarn, , Seleucid–Parthian Studies (Proc. Brit. Acad. 1930), pp. 16–18Google Scholar, Bactria and India, p. 499.

^{page 91 note 13} Pliny, , H.N. VI, 47Google Scholar; Strabo, XI, 516.

^{page 91 note 14} Tarn, , Bactria and India, p. 118Google Scholar.

^{page 91 note 15} Id. p. 291. Junge, J., Saka-Studien, Klio, Beiheft 41, 1939, p. 67Google Scholar, would extend this expression greatly north-eastward. It is not necessary to consider that here.

^{page 91 note 16} Arr. IV, 1, 3.

^{page 91 note 17} Syr. 57.

^{page 91 note 18} Arr. IV, 4, 1.

^{page 91 note 19} See the reproduction of this section in Miller, op. cit. at cols. 639–42.

^{page 91 note 20} The name Araxes originally came from the Oxus; it is Ar-axes, ‘river Oxus,’ just as Plutarch's Orexartes (Alex. XLV) is Ar-yaxartes; see Roesler, R., Wien SB LXXIV, 1873, p. 256, n. 3Google Scholar, who gives many parallels of river names beginning with Ar-. But ‘Araxes’ was often used of the Jaxartes after that river became known; and in the Peutinger Table there is no doubt, as the Oxus is given separately by its name.

^{page 91 note 21} Tarn, op. cit. pp. 231 sq. Of course a town ‘refounded’ did not always stand on the old site.

^{page 91 note 22} Strabo, XI, 524.

^{page 91 note 23} Presumably meaning ‘corn-land,’ ‘The Sown.’

^{page 92 note 24} This story, that the gods forbade Alexander to go any farther here, as they also did at the Beas, is not preserved in any extant Alexander-historian.

^{page 92 note 25} For the decree see Holleaux, , Rev. E.G. 1923, p. 1Google Scholar.

^{page 92 note 26} OGIS 213.

^{page 92 note 27} Holleaux, id.

^{page 92 note 28} Pliny, H.N. VI, 49 (from Demodamas)Google Scholar. Dux is στρατηγός.

^{page 92 note 29} So again Pliny, VI, 58, Seleuco et Antiocho praefectoque classis eorum Patrocle.

^{page 92 note 30} Pliny, VI, 49: transcendit eum amnem Demodamas, Seleuci et Antiochi regum dux, quern maxime sequimur in his, arasque Apollini Didymaeo statuit. Fr. Stähelin, Seleukos I n PW, 1921, col. 1231, said that according to Rehm's dating in Milet, I, 3, p. 262Google Scholar, Demodamas' crossing of the Jaxartes must be before 299/8; and Hennig, R., Terrae Incognitae, 1936, p. 173Google Scholar, citing the same passage in Rehm, said that Demodamas' expedition was connected with the campaign which Seleucus undertook in 299/8 to strengthen and widen his rule over Asia. But Rehm has not a word about Demodamas' campaign; what he said, after dating the decree for Apama in 299/8, was ‘wir kommen auf das Jahr nach der Ehrung des Antiochos, für die Anregung des Demodamas (his proposal of the decree) vielleicht sogar in eben dieses Jahr.’ The theory of Stähelin and Hennig is completely refuted by the words Seleuci et Antiochi regum dux; long ago, U. Wilcken dated Demodamas' campaign correctly, Antiochos I in PW, col. 2451. No such campaign of Seleucus' as Hennig supposes is known; 299 was the year of his alliance with Demetrius the Besieger, their meeting at Rhossos, and his marriage to Demetrius' daughter Stratonice.

^{page 92 note 31} Arr. IV, 4.

^{page 92 note 32} VI, 49: oppidum … in ultimis eorum (the Sogdians) finibus Alexandria ab Alexandra Magno conditum. Arae ibi sunt & c.

^{page 92 note 33} I, 2, 5: (Tanais) praeteriens aras ac terminos Alexandri Magni in Rhobascorum finibus sitos.

^{page 92 note 34} Ibi in Pliny, VI, 49 is ambiguous; it could mean either Alexandria or the fines generally.

^{page 93 note 35} The wall took only 20 days to build, Arr. IV, 4, 1.

^{page 93 note 36} Habakkuk, i, 10.

^{page 93 note 37} Justin, XV, 4, 11.

^{page 93 note 38} Diod. XIX, 48, 1.

^{page 93 note 39} The Coinage of the Eastern Seleucid mints from Seleucus I to Antiochus III, 1938, p. 230Google Scholar.

^{page 93 note 40} Patrocles' date: Pliny, VI, 58, Seleuco et Antiocho (see note 29 above). Strategos of Bactria–Sogdiana: Strabo, II, 74, .

^{page 93 note 41} Plut. Dem. 47.

^{page 93 note 42} Memnon 15 (FHG III, 534).

^{page 94 note 43} Hennig, op. cit. p. 172, thought the list of peoples in Pliny, VI, 50, was from Demodamas. But some names come from Herodotus and the Alexander-story; and it is only one of Pliny's usual lists, combining his notes from many sources; see auctorum, VI, 51.

^{page 94 note 44} Salles, G., Rev. de Paris, XLIII, 1936, p. 144Google Scholar.

^{page 94 note 45} In 1932 a fragment of a Graeco-Buddhist sculptured cornice was found in the Oxus, 13 km. below Termez, but seemingly it came from a stupa and is later than the Christian Era: Field, H. and Prostov, E., Amer. Anthropologist, XXXIX, 1937, p. 475Google Scholar. In 1938 M. Masson published from Tashkent four square Greek bases, from places near Termez, which once carried circular disengaged columns, though but little of these remains; thanks to Professor E. H. Minns I have seen his drawing, but it is too small to show details. Of course Greek coins have often been found in the ruins of Old Termez.