¹ I follow the history as given in Prof.Gardner's, P.The Coins of the Greek and Scythic Kings of Bactria and India in the British Museum, 1886 (cited as P.G.)Google Scholar. For other recent accounts of the history proper, so far as it can be deduced, I may refer to von Gutschmid, , Geschichte Irans, 1888Google Scholar; M. E. Drouin in the Grande Encyclopédic, s.v. ‘Bactriane’ and W. Tomaschek in Pauly- Wissowa, s.v. ‘Baktrianoi.’

² Under Eukratides, Apollodoros ap. Strabo. 15, 686—this might refer to the Punjâb. Under Diodotos, Justin 41, 1, 8; 41, 4, 5—this cannot refer to the Punjâb.

³ ᾿Ευθυδήμου ἄνασσα (see W. Tomaschek in Pauly-Wissowa art. ‘Baktriane’); Eukratideia; Demetrias in Arachosia. I omit Euthymedeia.

⁴ E.g. coins of Philoxenos, Hippostratus, Azes, Zeionises; and a coin of Peukelaos published by Mr.Smith, V. A., J.A.S.B. 1898, p. 132Google Scholar. This of course proves nothing as to whether the burghers were Greek, native, or both. There is nothing that corresponds to the Seleukid city coinages; unless it be at Taxila.

⁵ Justin, 12, 5, deduced from the revolt of the Greeks after Alexander's death. Arrian 4, 4, says mercenaries, and barbarians who volunteered, and time-expired Macedonians (of Alexandreschate). Curtius 7, 7, 27 (of the same town) ‘captivi, quos…liberavit.’ The captivi would be from Cyropolis. As Cyropolis seems to haverisen again and superseded Alexandreschate, , (see post, p. 282Google Scholar) Curtius's version, which would help to explain this, may be correct. Diodoros 17, 83, (of the cities near Alexandria of the Caucasus), bears out Arrian. The reference in Diodoros to mercenaries who volunteered is of importance. Curtius (7, 3, 23) seems to imply that volunteers settled in Alexandria of the Caucasus, ‘permissum…considere.’ These notices are not all in agreement, and, so far as they go, do not agree with the great number of Greeks settled in Bactria and Sogdiana, who rose on Alexander's death. There must have been a later importation of Greek settlers; ‘nuper deducti,’ says Curtius, 9, 7, 1.

⁶ The Branchidae town, settled with Greek men and women under peculiar circumstances, became bilingual in about six generations (Curtius, 7, 5, 29). The Barkaeans, settled at the same time in Bactria by Darius (Herod. 4, 204), are not again heard of. Some remarks on the orientalisation of the new towns in Droysen, , Hellenismus, III. 69Google Scholar. Livy, 38, 17, in Syros degenerarunt, &c., is special pleading.

⁷ A considerable legend has grown up round Alexandria of the Caucasus, seemingly based on nothing but the one well known reference to ‘Alasaddá the capital of the Yóna country’ in the Mahavanso, which may not refer to this Alexandria at all; the Egyptian capital is also a candidate, though a most unlikely one (Levi, S., ‘Le Bouddhisme et les Grecs,’ Rev. de ľ Hist des Religions, vol. 23 (1891)Google Scholar: cf. the Ptolemaic gravestone with wheel and trisula found by Prof.Petrie, , J.R.A.S. 1898, p. 875)Google Scholar; and why not Bactra-Alexandria? The legend makes this town a centre of Greek life in the East, the birthplace of Menander, and the last town ruled by a Greek king (Hermaios = Yin-muf-foo ruling in Yung-keu = Younaki, Greek town). All that is known about it is that it was twice founded by Alexander.

⁸ As to this, Specht, M., ‘Les Indo-Scythes et l'Époque du Règne de Kanichka” in J.A. ser. 9, vol. 10, pp. 159–161Google Scholar. It may be the real meaning of Pliny, 6, 25 (23). It would probably be a workable hypothesis that Alexander intended the capital of each satrapy (anyhow in the East) to bear his own name. Hence he founded no Alexandrias beyond the Indus; for he intended to establish there not satrapies, but protected native rulers. Macedonian fondness for renaming places, Strabo 11, 518.

⁹ As appears from the annals of the lesser Han. See Specht, in J.A. ser. 8, vol. 2, p. 321Google Scholar.

¹⁰ Converse instance of the double name in the case of Merv, Gr. Antiocheia; the Chinese preserved the native name in the form Mu-lu, (for Muru).

¹¹ e.g. the ‘Bactrians’ of the Periplus. Sometimes the Greek and Kushan rule is even confused together, as Amm. Marc. 23, 6, 55. Perhaps even in Justin; 2, 13, the Scyths founded the Parthian and Bactrian kingdoms—this must refer to the Yue-tche. Tomaschek (Pauly-Wissowa, ‘Baktrianoi’) says compendiously, that when classical writers from 140 B.C. to 560 A.D. say Bactrians they mean Tochari (Yue-tche).

¹² Polyb. 29, 6a, 8: Gesch. Irans, p. 37.

¹³ She wears a golden crown with eight rays and a hundred stars, and is clothed with the skins of thirty beavers of the sheen of silver and gold. Her statue set up in Bactra, Clem. Alex., Protr. p. 57Google Scholar. The description, a lengthy one, is in the Âbân Yast, §§ 126—129, see Darmesteter's trans, , of the ‘Zend-Avesta,’ (in Sacred Books of the East), vol. 2, p. 82Google Scholar; also p. 63 for M. Halévy's suggestion that this description was taken from a consecrated type of statuary.

¹⁴ Polyb. 10, 49; no troops of Euthydemos are mentioned except the Bactrian horse.

¹⁵ Taken together, these particulars may carry some weight; in particular, it is difficult to see where else a royal Laodike can have come from; unless indeed it were from some dynasty in Arachosia or Aria, concerning which great provinces our knowledge is a blank, but which are treated as separate kingdoms in the Annals of the elder Han (if indeed Arachosia be Kepin, as M. Drouin supposes). The coins, indeed, know nothing of such dynasties; but they would leave us equally ignorant of Ta-yuan. As to the connection of Eukratides's coin-type with the Syrian Cabiri, Babelon, , Bois de Syrie, xxxi.—Seleukos II (246–226 B.C.)Google Scholar had taken the Dioscuri as a type.

¹⁶ Onesikritos ap. Strabo. 11, 517, ἐνταφιασταί The town was fair without, but within ‘full of men's bones.’

¹⁷ It is conceivable that, if Trogus were recovered, it would be found that Eukratides's offence against Bactria was religious. Alexander's edict against the dogs nearly brought on a revolt; Porph., de abst. 4, 21Google Scholar.—Onesikritos (St. 11, 517) says Alexander stopped the custom, with which a rhetorical passage in Plutarch agrees. But the version that he tried to must be correct, as von Gutschmid takes it; Zoroastrianism was excessively tenacious of customs.

¹⁸ Σάγαλα ἥ καὶ Ευθυμεδεία see p. 273. Its site does not appear to have been identified; see McCrindle, J. W., ‘Ancient India; its invasion by Alexander the Great,’ p. 347Google Scholar, note M. Lahore is one conjecture.

¹⁹ P.G. xxii, as to Euthydemos. P.G. xxv, “The coins of Demetrius come in almost all cases from Bactria.”

²⁰ Num. Chron. 1890 p. 110. Adopted by Mr.Rapson, E. J., Ind. Coins, § 30Google Scholar (in Bühler's, Grundriss der Indo-ar. Philol. 1898)Google Scholar. If “Moga” of the Manikyala copper-plate be Maues (P.G. xlix), this becomes almost a certainty as regards Taxila.

²¹ Prof. Rhys Davids's translation of the ‘Milinda’ in Sacred Books of the East, vol. i, p. xxvi.

²² “A historical basis for the questions of king ‘Menander’ from the Thibetan,” by Dr.Waddell, , J.R.A.S. 1897, p. 227Google Scholar. “Chinese translations of the Milinda Panho,” by Takakusu, J., J.B.A.S. 1896, p. 16Google Scholar. The form “Ananta” known to the Lamas; the Chinese translation—date given as between A.D. 317–420—gives Nanda; query, Nanda of Magadba? Thibetan sources make Nagasena and Nanda contemporary.—Criticism by Count d'Alviella, Goblet, Bull, de ľAcad. Royale de Belgique, 1897, vol. 33, p. 688Google Scholar n, to the effect that Prof. Rhys Davids takes the Pâli back to 1st cent. A.D., i.e. prior to the Chinese version. I do not find that he takes it further back than its citation by Buddhagosa as of conclusive authority, about 430 A.D. D'Alviella however does not deal with that part of Dr. Waddell's article which attempts to show, by tables, that the rainfalls mentioned in the “Milinda” do not suit the Punjâb at all; and no criticism can carry much conviction which does not first dispose of this definite matter of the rains.

²³ We are interested in the Greek for his art and literature. But to his contemporaries he must have meant, chiefly, the best of all known fighters; until the Roman came. The Koman, having beaten him in the field, could afford to exalt his art and literature.—Apollodoros ap. Strab. 11, 516, attributes to the Bactrians (principally to Menander) the conquest of more nations than Alexander; and it is of interest to notice that Alexander's name is said not to occur in Indian literature, which possibly records Demetrios as well as Menander.

²⁴ That is, if Prof. Rhys Davids is correct in calling it an ‘island.’ SirCunningham, A. however would translate Alasandadipa as “the country of which Alasanda was the capital” (J.A.S.B. 1893, vol. 62, part 1, p. 86Google Scholar, communicated to Mr. V. A. Smith).

²⁵ In the introduction to his translation of the ‘Milinda.’

²⁶ See post, under ‘Taxila’; and see note 19. Assuming him later than Demetrios, Taxila appears to be the furthest east yet attained: the coins in fact do not bear out the tradition of Demetrios's conquests in India, which may only mean that he was the first to cross the Hindu-Kush; unless Demetrios' elephant-scalp refers to this. Whatever the legend Hidujasaine means (‘Just to those born on the Indus,’ Bendall ap. P.G. lxxiii; ‘King of Indians,’ von Sallet; Of me, Agathokles, ‘Indian by birth,’ Levi, S. doubtfully in “Le Bouddhisme et les Grecs,” Rev. de ľHist. des Reliyions, vol. 23 (1891), p. 41Google Scholar, criticising the older interpretations), it appears to refer to some close connection of Agathokles with India.

²⁷ This would of course have nothing to do with their own belief.

²⁸ It is no objection that the word would be poetical. There are at least two undoubted poetical words on the coins, ἀνικητου and τυραννουντος the former as early as Demetrios, and common; add perhaps νικηφορου and the obscure θεοτροπου Also the poetical name of Hermaios's queen, Kalliope. (On the adoption of Muse-names in late Hellenism, see von Gutschmid, , Gesch. Irans. 116Google Scholar.)

²⁹ A very similar ease is that of the Bactrian town Εὐθυδήμου ἄνασσα which must in fact have borne the queen's actual name.

³⁰ See McCrindle, , Ancient India, before cited, p. 342Google Scholar, note I, “Taxila.” Its site has been identified with Shah-deri.

³¹ Agathokles's coins, post.

³² Rapson, , Indian Coins, § 56Google Scholar.

³³ The star is obliterated on the specimen here illustrated, but clear on others, e.g. P.G. No. 15.

³⁴ Cf. the use of δίκαιος see p. 273.

³⁵ Levi, M. S., in Le Bouddhisme et les Grecs, already cited, at p. 43Google Scholar; Agathokles, “soit par conviction, soit par politique, aurait élevé un stûpa.” This stupa coin is perhaps imitated by a copper coin from Khotan, which appears to bear traces of a stûpa; Dr.Hoernle, A. R. in Ind. Ant. 1898, p. 227Google Scholar.

³⁶ It used for a long time to be believed that a species of maneless lion existed in Gujerat. This is now said to have been conclusively disproved, the individuals in question being only immature specimens. (See e.g. Ene. Brit. s. v. ‘Hon.’) I do not see therefore why they should figure on the coinage. Mr.Rapson, E. J. (J.R.A.S. 1900, p. 103Google Scholar) gives a seal which he compares at length with the square coins of Agathokles and Pantaleon; the lion has a mane. The figure on the coins is certainly a poor tiger; but it would also be a poor lion; and as no Greek could ever manage a good lion, it is unreasonable to suppose that a designer on the fringe of Hellenism would succeed better with a tiger.

³⁷ ‘Stûpa of Barhut,’ p. 27. It has been noticed that the dancing girl is the only purely Hindu type that occurs, prior to Siva on the coins of Ooemo Kadphises (P.G. 124); and this is an additional reason for finding aii explanation for her.

³⁸ Rapson, , Ind. Coins, Pl. 1, 13Google Scholar; caitya on both obv. and rev.

³⁹ Strabo, 15, 714.

⁴⁰ The name is Tusâspa; Levi, M. S., ‘Quid de Graecis veterum lndorum monumenta tradiderint,’ 1890, p. 4Google Scholar, and generally. These Iranian settlers would be of more assistance to the Bactrian invaders than would Alexander's Indian foundations, if subsisting.—The Graeco-Bactrians seem to have found a Persian weight system established in the Punjab; Rapson, , Ind. Coins, § 8Google Scholar.

⁴¹ Dr.Stein, A. in Ind. Ant. for 1888 (vol. 17), p. 89Google Scholar, 98, ‘Zoroastrian deities on Indo-Scythian coins.’

⁴² It might be objected that the gold coins of the Kushans are not strack on the Persian standard, but approximate to the weight of the Roman aurei. Very likely, however, they are aurei restruck. (Cunningham, in Num. Chron. 1889, p. 277Google Scholar.)

⁴³ In the paper above referred to.

⁴⁴ Eustath. ad Dionys. Perieg. 976, Bernhardy. Tigris swiftest of all rivers. διό φασὶ καὶ Τίγρις καλει̑ται, ἤγουν ταχὺς ὡς βέλος. Μη̑δοι γὰρ Τίγριν καλου̑σι τὸ τόξενμα But some say it is called from the tiger; (follows a story); καὶ ἄλλως δὲ δίχα τη̑ς μυθικη̑ς ταύτης ἐπιβολη̑ς τη̑ν πρὸς τὸ ζω̑ον ὁμνυμίαν ὁμονυμίαν ὁ ποταμὸς ἔχει διὰ τὸ του̑ ῥεύματος ὀξύ For the tiger is very swift. Cf. the supernaturally swift tigers in Herodotos.

⁴⁵ Cf. an interesting suggestion of Dr. Hirth, that the metal mirrors with Bacchic symbols imported into China under the Han emperors might refer to a joint cult of Dionysos and Haoma; Ueber fremde Einflüsse in der Chines. Kunst, p. 25 seq.

⁴⁶ I purposely refrain from attempting to use Philostratos. Yet he must at least be evidence of a belief that Taxila would be a reasonable location for such a story as his.

⁴⁷ For the Annals of the Elder Han I use Wylie's, A. trans, of ‘Notes on the Western Regions,’ J. Anthrop. Inst. 1881Google Scholar, cited as ‘Wylie’; for. Sze-ma-ts'een, Kingsmill's, T. W. translation of ch. 123 in J.B.A.S. 1883, vol. 14Google Scholar, ‘Intercourse of China with Eastern Turkestan in the 2nd century B.C.’ cited as ‘Kingsmill.’ Prof. Chavannes' translation has not yet, unfortunately, reached ch. 123; but his introduction, pp. lxx to lxxviii, deals with its subject-matter. Every Chinese scholar seems to transliterate the proper names differently; where I can, I have used Prof. Chavannes' spelling.

⁴⁸ Wylie, 67.

⁴⁹ See note 53.

⁵⁰ Wylie, 45; Kingsmill, 94. Cf. Strabo, 15, 724. Ariana includes parts of Media and Persia as well as Bactria and Sogdiana; εἰσὶ γάρ πως καὶ ὁμόγωττοι παρὰ μικρόν

⁵¹ Kingsmill, 83. This seems to be applied to Parthia as well as to Ta-yuan and the Ta-hia; but elsewhere he knows of the power of Parthia, Kingsmill, 81.

⁵² Wylie, 41, ‘weak and afraid to engage in war.’ Kingsmill, 82, ‘weak and cowards in battle.’

⁵³ 41, 6. siquidem Sogdianorum et Arachotorum et Drangianorum Indorumque bellis fatigati, ad postremum ab invalidioribus Parthis, velut exsangues, oppressi sunt.—Justin clearly contemplates separate states here. But one cannot depend on him as accurate merely because he is scanty. Strabo as clearly mentions a preponderance of Bactria, 11, 516, 517; the Hellenes who held Bactria held Sogdiana also, and conquered Ariana and the Indians—this last from Apollodoros.

⁵⁴ Wylie, 45; Kingsmill, 94.

⁵⁵ Kingsmill, 91.

⁵⁶ Wylie, 45; Kingsmiil, 94.

⁵⁷ Kingsmill, 83.

⁵⁸ Kingsmill, 82. The parallel passage (Wylie, 41) reads that the Ta-hia ‘were originally without a chief paramount, and were accustomed to set up petty chiefs over their cities.’

⁵⁹ Wylie, 46. ‘Women are honourably treated among them, and their husbands are guided by them in their decisions.’ Kingsmill, 94, ‘They held their women in high estimation, and the husband commonly took his wife's advice before coming to a decision.’ This statement appears to be made of all the countries westward from Ta-yuan as far as Parthia.

⁶⁰ ‘Ueber das Arimaspische Gedicht von Aristeas,’ in Abh. d. k. Ak. der Wiss. in Wien. h. phil. Clas., 116, (1888), p. 751. Some writers treat the Yue-tche as Turks.

⁶¹ But not impossible. There must be a similar error, whatever its nature, in the contradictory statements as to the use of silk; see post p. 290.

⁶² In ‘Notes on the Western Regions’ they do not even have a separate section from the Yuetche.

⁶³ The name Ta-hia is so far unexplained. It does not even seem to be certain whether it means Great Hia or not. But the common explanation that Ta-hia = Dahae seems impossible. Ta hia may be good Chinese for Dahae; but unless it can only mean Dahae, which is clearly hot the case (see e.g. Dr.Hirth, , Ueber. fremde Einflüsse, p. 23Google Scholar), it is worthless without some fact to support it. No connection of the Dahae with Bactria is known. The theory is, that they may have joined in the Saka invasion; but (1) the Chinese only mention the Sakas, (2) if so they were driven out with the Sakas before Tchang K'ien came. As a fact, the Dahae remained in their original seats, beyond Margiana, and contributed a refuge for Parthian pretenders, and mercenaries for Parthian and Seleukid kings; Artabanus III. (c. 8 A D.) lived among them, Tac., Ann. 2, 3Google Scholar; they fought at Magnesia, Livy, 37, 40, and at Raphia, Polyb. 5, 79; see also Strabo, 2, 718 and Ptol. 6, 10, and Prof.Gardner, P. ‘The Parthian coinage,’ 1877 (in Marsden's Numismala Orientalia), pp. 12Google Scholar, 13—Identifications by similar sound are worthless in themselves, unless used to support deductions from facts.

⁶⁴ Specht, in J.A. 1883, p. 321Google Scholarseq.

⁶⁵ See note 68.

⁶⁶ Coin representing Macedonian horseman charging two riders on an elephant; Prof. P. Gardner takes the riders to be Yue-tche, and the coin to commemorate a victory of either Eukratides or Heliokles, ; Num. Chron. 1887, p. 177Google Scholar; but its genuineness is said to be doubtful.—Isit possible that the ‘Bactriaiis’ of Justin, 36, 1, 5, allies of Demetrios Nikanor, were really an advanced horde of the Yue-tche? See note 11.

⁶⁷ Wylie, 39; Kingsmill, 81.

⁶⁸ Wylie, 41; Kingsmill, 81. Spechtin, J.A. 1883, p. 322Google Scholar, ‘passèrent an delà de Ta-Ouaii.’

⁶⁹ This does not exclude the possibility of Ta-yuan being one of. the kingdoms formed by the Sse (Sakas); ef. the horse-archers, and the coins referred to, note 85.

⁷⁰ Strabo, 11, 517.

⁷¹ Tomaschek, in Pauly- Wissowa, s.v. ‘Baktriane.’ Strabo however is clear that the Oxus was the boundary.

⁷² For instance, Dr.Budge's, Life and Exploits of Alexander the Great, from the Ethiopian, pp. 183,·186Google Scholar. For volunteers at Alexandreschate, note 5.

⁷³ Kingsmill, 83 to eud.

⁷⁴ Kingsmill transliterates Urh-shi, Dr. Hirth Ir-schi: de Lacouperie wished to read Nise.

⁷⁵ It is interesting to compare this account with the siege of the same town, (Cyropolis), by Alexander (Arr. 4, 3). Alexander took the outer city by thirst. As to the identification, see note 83.

⁷⁶ The same threat that Enthydemos used to Antiochos III.

⁷⁷ de Lacouperie, T., Western origin of early Chinese Civilisation, pp. 220Google Scholar, 221.

⁷⁸ As the revival of the letter San on the Kushan coins appears to he generally accepted, and as this letter, (sound sh), is known as used in Greece for branding horses, it ought to be suggested, to complete the list, that the Shen horses were σαμφόραι For a suggestion that Ta-yuan = Strabo's Τουριούαν (the province beyond Mervlost by Eukratides to the Parthians, and translated by Brunnhofer, , vom Aral bis zum Gangâ, 61Google Scholarseq., as ἱππόβοτος i.e. Nisaean fields which he places between Balkh and Merv) see Hirth, Ueber fremde Einflüsse, &c.,-p. 24Google Scholar; it is geographically quite impossible, as Dr. Hirth sees. A considerable number of places called Nisaea are known: but the ‘fields’ were certainly in Media.

⁷⁹ P'u-tao, vine = βότρυς; Muh-tuk, lucerne = μηδικὴ (πόα)

⁸⁰ By Prof. Chavannes in his Introduction before cited.

⁸¹ The only case that occurs to me is the coins of the so-called Nameless king, P.G. xlvii., Kabul valley, circ. A.D. 30–50; the inscription is generally βασιλεὺς βασιλύων σωτἢρ μέγας possibly Kushan.

⁸² See p. 287.

⁸³ Eul-che = Uratube; Prof. Chavannes in the Introduc tion before cited, p. lxxv; Dr.Hirth, , Ueber f remde Einflüsse, &c, p. 21Google Scholar; both on a consideration of Chinese evidence. Cyropolis = Uratube; von Schwarz, , Alexander des G. Feldzüge in Turkestan, (1893), pp. 51, 52Google Scholar. The stream and citadel are there; the town gave more trouble to the Russians than any other in Khokand and Bokhara.—Mr. D. G. Hogarth (Philip and Alexander of Macedon) does not accept von Schwarz's identifications as sufficient, in the absence of excavation.

⁸⁴ Hardly the last, having regard tothe date.

⁸⁵ In the British Collection of Central Asian Antiquities are seven silver tetradrachms from ‘Samarkand, Tashkend, and other places in Western Turkestan,’ which imitate coins of Heliokles and Euthydemos, and some of which are referred by Dr.Hoernle, (Ind. Ant. 1898, p. 225Google Scholarseq.), to circ. 150 and 130 B.C. Are they Saka?

⁸⁶ For discussions of this question, see (among other things) Weber, , ‘Die Griechen in Indien,’ Sitz. d. Ak. d. Wiss. Berlin, 1890Google Scholar; Levi, Quid de Graecis, &c.,; Count d'Alviella, Goblet, Ce que ľInde doit à la Grèce, 1897Google Scholar, and his series of articles in the Bull. de. ľAcad. Royale dos Sciences de Belgique, vols. 33 and 34 (1897), (strongly prp-Greek); Mb Smith's, V. A. three articles, ‘Graeco-Roman Influence on the Civilisation of Ancient India’ in J.A.S.B. 1889 (vol. 58), 1892 (vol. 61), and 1893 (vol. 62)Google Scholar; and a clear summary in Prof.Macdonell's, A. A. recent History of Sanskrit Literature, p. 411Google Scholar onwards.—Greek or Graeco-Roman influence is of course generally treated as a whole.—Bibliography of the large literature relating to the architecture and sculpture is given by Mr. V. A. Smith, and by Dr. Burgess in his recent edition, with translation, of Prof. Griinwedel's Buddhistische Kunst in Indien. My references to Grünwedel are to the second German edition (1900), as this paper was practically completed bef ore I saw the translation.

⁸⁷ Notes on Hindu Astronomy, by Dr.Burgess, , J.R.A.S., 1893, p. 717Google Scholar.

⁸⁸ See also Oldenberg's, essay on ‘Buddhistische Kunst in Indien’ in Aus Indien und Iran (collected essays, 1899), esp. pp. 116, 117Google Scholar; and cf. Mr. Vincent Smith's Indo-Hellenic school.

⁸⁹ Arch. Survey of India, vol. 5, 189.

⁹⁰ Grünwedel, pp. 17, 51, 57. They may be meant, he thinks, to symbolise the doctrine of transmigration and rebirth; in each stage of animal existence the human may be concealed, to be released through good works. It is interesting to meet the celebrated βουγενὲς ἀνδρόπρωον in India serving the use of an alien philosophy.

⁹¹ Certain traces of Greek or Graeco-Roman influences appear in the art of Khotan, which was so largely influenced by that of India; see for instance the seals from Taklamakah given by Dr.Hoernle, (‘A collection of Antiquities from Central Asia,’ J.A.S.B. 1899, vol. 68Google Scholar, part 1, nos. 24, 26, 32 and 33 on plate 3, and no. 11 on plate 19), which include two figures of Athene; and the clay seals representing Athene and Eros referred to by DrStein, in his recent Preliminary Report of his excavations in Chinese Turkestan, at p. 53Google Scholar. Dr. Stein says. ‘There is good reason to believe that this influence was exercised, partly through Bactria, partly through Gandhara and the adjoining regions on the N. W. frontier of India.’ I do not know if any date has yet been suggested for these figures of Athene, or if they may be earlier than the Gandhara school.

⁹² Grünwedel, 81; 184 ‘direkt als griechische Göttin ist dargestellt Athene Promachos.’

⁹³ Grünwedel 80. Mr. V. A. Smith has suggested tentatively 200 A.D. for this school.

⁹⁴ Cunningham, , Archaeol. Survey of India, vol. 3 (1870–1872), p. 14Google Scholar; coins of Apollodotos, Menander, Strabo, and Antimachos.

⁹⁵ Perhaps the fact that a station on the silk toute became wellknown as ‘The Stone Tower,’ (Ptolemy), may even suggest that in the neighbourhood, Sogdiana for example, stone architecture was rare and remarkable.—I have not met anywhere with a description of the ‘Greek’ pillars at Oosh in Ferghana, mentioned by Vambèry (Central Asia).

⁹⁶ From Julien's translation. D'Alviella, , (Ce que ľInde, &c. p. 82)Google Scholar cites this passage, together with Philostratos, for a continuous Greek art.

⁹⁷ Wylie, 34, 35 the people of Kepin are ingenious in building palaces and mansions.

⁹⁸ Curtius, 7, 5, 29.

⁹⁹ Academy, 10th September, 1887.

¹⁰⁰ Doric; Atti. 466 f. Pindar ap. Ath. 467 b, complains that singers would use san; and as this is generally true, (a Germansings Ish for Ich), the fact that the sound could not die might help to keep the letter alive.

¹⁰¹ P. G. xlvi.

¹⁰² Prof.Gardner, P.Num. Chron., 1887, p. 177Google Scholarseq.) suggests, on grounds of style, that the Kushan kings got their artists from Bactria.

¹⁰³ In ‘Quid de Graecis,’ &c. before referred to.

¹⁰⁴ One would seem to date from the middle of the first century B.C., but refers only to astrology.

¹⁰⁵ Spiegel referred these ‘Kronen tragenden Griechen’ to some section of the Greek race; (Erân. Alt., 1,223); but if the Ionian satrapy had been divided, some notice of it should have been given upon the first occurrence of the name in its usual place, beside Sparda. Clearly the epithet is to distinguish these Yunas from the ‘Ionians.’

¹⁰⁶ Hirth, , ‘China and the Roman Orient,’ (1885)Google Scholar.

¹⁰⁷ Prof.Bury, has suggested that the Ionians got their common name from an original people of Iavones in Asia Minor; (‘Prehistoric Ionians,’ Eng. Hist. Rev., 1900, p. 288)Google Scholar; but the supposed occurrence of the name in the fifteenth and thirteenth centuries in Egyptian records appears to be a mistake, see Mr.Hall, H. R., The Oldest Civilisation of Greece (1901), p. 129Google Scholar. The connection between ᾿ΙάFων and ᾿´Ιων is not known (Busolt, , Gr. Gesch. ²1Google Scholar, 283), and it would be tempting to compare the two forms with Yavana and Yona, but these latter seem to be identical; Levi, , ‘Quid de Graecis, p. 3Google Scholar, (n), ‘Yona nomen pracritice idem quod Yavana sanscritice scribitur.’

¹⁰⁸ Dr.Justi, F., Gesch. Irans, p. 444Google Scholar, in Geiger, and Kuhn's, Grundriss der Iranischen Philologie (1900)Google Scholar.

¹⁰⁹ Since the above was written, I see that Mr. V. A. Smith frankly calls the Yonas of Asoka's Rock Edicts 5 and 13 one of the ‘semi-independent foreign tribes on the north-western frontier’ of Asoka's dominions; ‘Asoka’ (1901), pp. 120, 132. But as he naturally translates the word in Rock Edict 2 as ‘the Greek king Antiochos,’ the confusion I have noticed would be as old as the time of Asoka at least.

¹¹⁰ Levi, p. 4, translates ‘Tusâspa, Açoki Mauryeiisis Yonorum rex.’

¹¹¹ Levi, p. 16. Cunningham interprets Mâdhyamiki as the people of the middle country, that is, the Gangetic provinces above the Delta, (Num. Chron, 10Google Scholar, 225).

¹¹² Strabo, 11, 576.

¹¹³ Kingsmill, 83.

¹¹⁴ Wylie, 20.

¹¹⁵ Richthofen, , China, p. 464Google Scholar.

¹¹⁶ Wylie, , ‘History of the South-western Barbarians,’ J. Anthrop. I. 1880, p. 59Google Scholar; de Lacouperie, , Western Origin of Chinese Civilisation, p. 50Google Scholar. This book, though requiring critical use, contains a mass of information about the overland route to China.

¹¹⁷ It may be noted that in the message of Euthydemos to Antiochos (Polyb. 11, 34) he speaks of ‘admitting’ the nomads (προσδέχεσθαι) as if through some barrier, which can hardly be the Jaxartes, as they are said to be close at hand.

¹¹⁸ Pauly-Wissowa, ‘Baktrianoi.’

¹¹⁹ Ueber das Arimaspische Gedicht des Aristeas before cited. Issedon = Sera Metropolis = Χουβδάν = S'ian-fu. Issedones, a northern branch of the Thibetan race: Arimaspi, the Huns; Hyperboreans, the Chinese: Iranian influences in Tarim valley, and Iranian trade with Issedon, which may have possessed a merchants’ quarter.

¹²⁰ Coins of Menander and Antimaehos II. and the ‘iron’ coin of Hermaios; also Roman coins of Constans II., Justinus, Theodosius. SirForsyth, T. D., J.E.G.S. 47Google Scholar, p. 12; Prof.Gardner, P., Num. Chron. 1879, 274Google Scholar; Dr.Hoernle, A. F. R., ‘Indo-Chinese coins in the British collection of Central Asian Antiquities,’ Ind. Ant. 1899, p. 46Google Scholar; also J.A.S.B. 1899, vol. 68, part 1; the ‘iron’ coins are really of copper. As to the seals from Khotan, see note 91.

¹²¹ Gold currency in Kepin, in the time of Tchang K'ien; on one side a man on horseback, on the other a man's face; Wylie, 34. Not a very gocd description of Enkratides's money, but it does not appear to what else it can refer.

¹²² Ptolemy, 1, 11, § 7 (from Marinus).

¹²³ Kingsmill, 94.

¹²⁴ Wylie, 46.

¹²⁵ Pliny, , N.H. 34Google Scholar, 14, § 145; Periplus, § 39.

¹²⁶ Cunningham, in Num. Chron. IX, 298Google Scholar. But the description given by Prof. P. Gardner (P.G. 57) is ‘clad in chiton, himation and boots.’ The boots are plain; but it does not clearly appear on the plate what the garment is.

¹²⁷ See note 13.

¹²⁸ Wylie, 46.

¹²⁹ Figured and described by Dr. Hirth, Ueber fremde Einflüsse, &c.

¹³⁰ De Lacouperie, , Western Origin, &c., p. 217Google Scholar, note 933.

¹³¹ More will be known about this if chemical analysis should ever prove that the jade objects found, e.g. in Assyria, must be Khotan jade.

¹³² Alexander-legends attach themselves along this trade route to places where Alexander certainly never was, and are not all due to Islam; e.g. the story of the foundation of Taugast and Chubdan, given by Theophylaet, is prae-Mahommedan. But even if they in fact referred to the Graeco-Bactrian kings, they cannot be dated, and so would be of little value for the present subject.

¹³³ Strabo, 11, 516. He calls it the greatest ornament of Iran (τη̑ς συμπάσης ᾿Αριανη̑ς πρόσχημα) and says the Greeks made their conquests by growing strong διὰ τὴν ἀρετὴν τη̑ς χώρας It was πάμφορος πλὴν ἐλαίου which is borne out by Sze-ma-ts'een, with a natural alteration of the important thing missing; Kingsmill, 94, ‘Their country produced everything but silk and varnish’ (lac).

¹³⁴ Wylie, , 46, ‘They applied the Chinese gold and silver to make vessels, instead of using them for state presents’; Kingsmill, 94Google Scholar, ‘They obtained from China gold and silver surreptitiously to make various utensils.’

¹³⁵ Head, , Hist. Num. pp. 238, 239Google Scholar.

¹³⁶ Cunningham's, deduction. Num. Chron., 1888, p. 217Google Scholar.

¹³⁷ Apoll. ap. Strab. 11, 516 μέχρι Σηρῶν καὶ Φρυνῶν. Tomaschek, , Ueber das Arim. Gedicht before cited, p. 769Google Scholar reads Φο῀υνοι = Χο῀υνοι of Marinus = Ο῏υννοι of Cosmas = Arimaspi = Huns.

¹³⁸ Cunningham, in Num. Chron. 10Google Scholar, 224 seq.: Levi, , Quid de Graecis, 15–17Google Scholar.

¹³⁹ Analogies drawn from Anglo-Indian life seem to me most misleading. It might be more in point to compare the history of the Portuguese in the Indian Ocean, more especially their later prazo system in South-East Africa.

¹⁴⁰ Of the founders of new dynasties, Euthydemos certainly (Polyb. 11, 34), and Eukratides probably (p. 271), came from the west, and perhaps represented two movements of new settlers or mercenaries.

¹⁴¹ As to his dependence on Parthian (and Roman) sources of information, see 1, 14; 2, 118; 11, 508.

¹⁴² Vindusara's request for a sophist, which Antiochos put off with a jest; probably none would go. (Hegesandros ap. Ath. 652 f.)

¹⁴³ Strabo, 1, 66. The same feeling seems to underlie all Alexander's dealings with India, and was well expressed by the mediaeval romance writer who made of Alexander and Poros, two knights tilting in the ring.

¹⁴⁴ Cf. Cunningham, in Num. Chron. IX. 293Google Scholar.

¹⁴⁵ Cf. Oldenberg's most interesting parallel between Orphism and Buddhism, , Aus Indien. und Iran, pp. 85Google Scholar to 100.

¹⁴⁶ A curious similarity may be noted here. King Paerisades at Panticapaeum had to bring up a Scythian prince as his heir (Holm. 4, 532, Eng. Tr.); while their joint coius show that Kadphises succeeded Hermaios peaceably.