^{page 183 note 1} In regard to the points discussed in this paper, neither the new edition of the Old Persian inscriptions nor the edition of the so-called Scythian nor that of the Babylonian version (all included in the Assyriologische Bibliothek) supplies any divergent information.

^{page 186 note 1} See the map in Tomaschek's Centralasiatische Studien, i, and Ptolemy's Geographia, vi, xiv, 13.

^{page 186 note 2} Cf. Strabo, xi, c. viii, 2: οἱ μν δ πλεους τν Σκυθν π τς Κασπας θαλττης ρξμενοι Δαι προσαγρεονται. τοὺς δ προσεῴους τοτων μλλον Μασσαγτας, κα Σκας νομζουσι. τοὺς δ’ἄλλους κοινς μν Σκθας νομζουσιν, ἰδᾳ δ ὡς κστους.

^{page 187 note 1} There seems to be no real proof that the Sse of the Chinese, though the original pronunciation was Sek or Sok (see M. Lévi's very interesting note, Journal Asiatique, sèr. ix, vol. ix, 1897, pp. 10, 11), were our Sakas. The Tibetan Sog means Mongol.

^{page 188 note 1} Bartholomae, Altiranisches Wörterbuch, s.v. zrayah-.

^{page 188 note 2} According to Strabo, xi, c. vii, 2, the Caspian was ἄπλους τε κα ργς, ‘ unsailed and idle.’

^{page 188 note 3} Behistun Inscr., ii, § 35 (xvi)–iii, 38 (iii).

^{page 188 note 4} Drangiana = ‘ Seelandschaft’ (Geiger, , Grundriss d. Iran. Philologie, ii, p. 393Google Scholar, doubted by Foy, , Kuhns Zeitschrift, xxxv, p. 22).Google Scholar

^{page 188 note 5} This remark is also made by Foy, , Kuhns Zeitschrift, xxxvii, p. 536.Google Scholar

^{page 189 note 1} Geiger, , Grundriss, ii, p. 393Google Scholar; Justi, ibid., p. 489.

^{page 189 note 2} Stathmoi Parthikoi, § 18.

^{page 189 note 3} See Mr. Vincent Smith's article, J.R.A.S., 1903, pp. 1–64, esp. pp. 18–24 and reff.

[I find that the above statement requires modification. Ma-twan-lin's work (thirteenth century) does, in the account of Ki-pin, affirm that when the Yué-tchi moved west “ the king of the Saï went to the south to dwell in Ki-pin. The “ tribes of the Saï divided and dispersed so as to form here and there different “ kingdoms. From Sou-le on the north-west, all the dependencies of Hieu-Siun “ and Siun-tu (Sind) are inhabited by former Saï tribes” (Rémusat, , Nouveaux Mélanges, i, pp. 205–206)Google Scholar. Whether this account goes back to an earlier source I am not in a position to say. Sakastāna, though not mentioned, might be held to be included. But the whole story seems to me incorrect.]

^{page 189 note 4} The ‘German Empire’ and ‘ Hindustan ’ are rather later than the Germans and Hindus!

^{page 190 note 1} Other Ikhthuophagoi and a people named Makai are placed by Ptolemy (vi, c. vii, 14) on the Arabian side of the Gulf of Oman.

^{page 190 note 2} Cf. Zend kadrva = Skt. kadru, ‘brown,’ kadrvaspa, ‘ a certain mountain,’ acc. to Brunnhofer, Iran u. Turan, pp. 109, 168.

^{page 191 note 1} The second part of Kāpisakānish, ‘ a fort in Arachosia,’ is supposed by Justi (Grundriss, ii, p. 430) to correspond to modern Persian khānī, ‘ spring ’ (= Sanskrit khāni, ‘ mine ’), or khundah, ‘ ditch of a fort.’ But, whether it is to be explained so or as a fusion of the two common suffixes ka and āna, at any rate it occurs in several names of towns noted by Ptolemy in this region and in Persia, e.g., Artakāna (Persis), Sourogāna, Astakāna (Bactria), Sarmagāna, Zamoukhāna, Ortikāna (Herat), Daroakāna, Tarbakāna (Paropamisadæ).

Kāpisakānish is therefore the Kāpisa in Ghorband, which was destroyed by Cyrus (Cunningham, , Numismatic Chronicle, xiii (1893), pp. 97 and 99Google Scholar; Justi, , Grundriss, ii, p. 420)Google Scholar, although Cunningham seems to distinguish the two. The identity of Kavisiye nagara with Kāpisa, suggested by Marquardt (Ērānšahr, p. 280), is now vindicated by Professor Rapson (J.R.A.S., 1905, pp. 783–4). The Arachosian Κτισα of Ptolemy should surely (though I do not find it suggested) be the same, and perhaps the (Σκαι κα) Κσπιοι of Herodotus are really Κπισοι.

^{page 191 note 2} Mentioned with ref. by Tomaschek, Zur Historischen Topographic von Persien, p. 188.

^{page 192 note 1} Trans. West, Sacred Books of the East, v.

^{page 192 note 2} Le Strange, Lands of the Eastern Caliphate, p. 334.

^{page 194 note 1} Professor Rapson (Indian Coins, pp. 7, 16) and Mr. Vincent Smith, whom 1 name honoris causa, are therefore in contradiction with this, the latter very sharply: “ The flood of barbarian invasion. … finally extinguishing the “ Hellenistic monarchy, which must have been weakened already by the growth of “the Parthian or Persian power” (Early History, p. 201). What Mr. Vincent Smith ascribes to the Sakas, Professor Rapson attributes to the Kushans. This latter view seems to me incorrect, though only slightly. I conceive that the Kushans conquered the Kabul valley not from the Greeks, but from the Parthians, who had themselves taken it from the Greeks. Nor is this a mere inference or conjecture. The Chinese History of the Second Hans (25–220 a.d.) states in a passage cited by M. Specht (Études sur l’ Asie Centrale, i, p. 10) as follows:—

“ They” (the people of Kabul) “ have been successively under the dominion “ of the Thièn-tehou (Hindus), of Ki-pin, and of the A-si (Parthians). These “ three realms at the time of their greatness had conquered this country, and “ they lost it at the moment of their decay. The book of the Han (Han-chou) “ is therefore mistaken in counting Kao-fou among the five principalities of the “ Yué-tchi. It had never belonged to these last, since it was at that time under “ the dominion of the A-si. But when the Yué-tchi attacked the A-si, they “ became in that way possessors of Kao-fou.” From the circumstances it is clear that the people of Ki-pin to whom reference is made in this extract must be the Greeks.

^{page 195 note 1} From another passage (xi, 2) we learn that it was two satrapies (τν τε Ασπιώνου κα τν Τουριοαν) that they took from Eucratidas.

^{page 195 note 2} Grundriss, ii, pp. 488–9. It is at this period that von Gutschmid considers that the Scythians “ must have ” occupied Sakastan, although the “ too favourable ” accounts of the dealings of the Parthians with their disloyal Scythian allies do not mention the fact. (Encycl. Brit., 9th ed., vol. xviii, p. 594b.)

^{page 196 note 1} Megasthenes (ap. Strabo, xv, 44) places the scene among the Δρδαι (Dards).

^{page 196 note 2} See Bartholomae, Altiran. Wörterbuch, s.v. aγrya.

^{page 196 note 3} See Smith's Dictionary of Ancient Geography, s.v.

^{page 197 note 1} Grundriss d. Iran. Philologie, ii, p. 383.

^{page 197 note 2} Unless Ptolemy's Τατακην in Drangiana is really Σακαστην.

^{page 198 note 1} Regarding the European Scythians.

^{page 198 note 2} Grundriss, ii, p. 400.

^{page 198 note 3} We may perhaps hope to learn something bearing on the subject of this paragraph from Dr. Grierson's forthcoming work on the Paisācī dialect.

^{page 199 note 1} Altiran. Wörterbuch, s.v. Haumavarka. The old Persian form of the name need not, however, be more correct than the Ἀηργιοι and the Umurj Umamarga (i.e. Umavarga) of the Greeks and of the Babylonian and Scythian versions of Darius' inscriptions. It may be due to popular etymology. What if the original form of the word was Hāmavarka, i.e. the Varka of the Hāmūn ?

^{page 199 note 2} In Badakshān acc. to Justi, but in Margiana ace. to Ptolemy (vi, c. x, 2).

^{page 200 note 1} Aral bis zur Gaṅgā, p. 120. “ So müssen die Çaka schon einmal in der Urzeit, nicht erst im zweiten Jahrhundert vor Christus, die mitteliranische Tiefebene besetzt haben.”

^{page 200 note 2} Vedische Mythologie, i, pp. 101 sqq., questioned by Oldenberg, Religion des Veda, p. 145, n. 1, and Foy, , Kuhns Zeitschrift, xxxv, p. 51.Google Scholar

^{page 201 note 1} Chronology of Ancient Nations, trans. Sachau, p. 109.

^{page 201 note 2} Cf. σατρπης (ξατρπης, ξαιθρπης), the initial vowel in Ἰσσδονες being, no doubt, prothetic, whence its variation.

^{page 201 note 3} See also below.

^{page 201 note 4} Rémusat, , Nouveaux Mélanges Asiatiques, i, p. 203.Google Scholar

^{page 202 note 1} Op. cit., p. 94.

^{page 202 note 2} Op. cit., i, p. 203.

^{page 202 note 3} Cf. Lucian, Macrobioi, § 5. His Omanoi, § 17, will be the Yamāma of Albīrūnī, loc. cit.

^{page 202 note 4} Ancient enough to be disputed by Democritus (Strabo, xv, 38). For the Chinese account of the ‘ weak water ’ see Rémusat, op. cit., i, pp. 216–17.

^{page 202 note 5} Iran u. Turān, p. 139. For the Chinese version see Rémusat, op. cit.

^{page 202 note 6} Megasthenes, xxi–xxiii (trs. McCrindle), has Silas.

^{page 202 note 7} For the genealogẏ see Justi, Iran. Namenbuch, p. 394.

^{page 202 note 8} Zendavesta, trans. Darmestater, ii, p. 62 and note; Bundahiš, trans. West, xxix, p. 5, note.

^{page 203 note 1} See Stein, , White Huns and Kindred Tribes, etc., Indian Antiquary, vol. edxxviii, 1905, pp. 73 sqq.Google Scholar

^{page 203 note 2} Journal Asiatique, sér. ix, vol. vii, p. 12. ‘ Ephthalite ’ also is stated by M. Specht (Études sur l’ Asie Centrale, i, p. 33) to be properly a family name.

^{page 204 note 1} Cited by Boyer, M., Journal Asiatique, sér. ix, vol. x, p. 150Google Scholar, and used by Cunningham. Cf. M. Lévi's note mentioned above, p. 187.

^{page 204 note 2} We may refer to M. Chavannes’ very valuable work, Documents sur les Tou-kiue, St. Petersburg, 1903.

^{page 204 note 3} Rājataraṅginī, i, pp. 168–70, see Dr. Stein's observations in his translation, i, p. 31, and Introd., p. 76. Dr. Stein, however, seems to hold that the Kushans were by race Turuṣka: see his paper on the ‘ White Huns ’ (Ind. Antiquary, 1905).

^{page 204 note 4} Rājataraṅginī, iv, p. 179.

^{page 205 note 1} Journal of the Royal Asiatic Society, vol. xxxvi, 1904, pp. 703 sqq., v. infra, p. 216.

^{page 205 note 2} Journal Asiatique, sér. ix, vol. vii, pp. 12 sqq. Among other points he suggests that Spala in Spalahora is a Scythian word denoting ‘victory.’ Some etymologies are proposed by Cuno, Die Skythen (1871), p. 211.

^{page 205 note 3} The confusion of y and j between vowels is in the inscriptions of Aśoka rare and almost confined to the words rājā, pūjā, and mayūra. We have to distinguish between y for j as in rāyā and pūyā, and j for y as in majūla. It is not likely that both changes took place in the same dialect at the same time, but the occurrence of either might lead to confusion in writing. For the early period the matter still needs investigation. But as regards the time and place of the Saka, Pahlava, Kushan dynasties, I am inclined to believe that the choice between j and y is not quite haphazard, and that the y properly represents the intermediate sound ž = French j.

^{page 205 note 4} Journal of the Royal Asiatic Society, n.s., xxvi, 1894, p. 549.

^{page 206 note 1} Which was, of course, not the native, even if an official, language of the Parthians themselves.

^{page 206 note 2} Geiger, , Grundriss d. Iran. Philologie, i, pp. 205, 236, 300–1.Google Scholar

^{page 206 note 3} Ptolemy (McCrindle), pp. 263 and 269. Cf. Avestan Varena ?

^{page 206 note 4} Ibid., p. 275.

^{page 206 note 5} Ibid., pp. 35 and 268.

^{page 206 note 6} Geiger, , Grundriss d. Iran. Philologie, i, p. 299, § 19.Google Scholar

^{page 216 note 1} For a discussion of the matter see Dr. Fleet's articles in this Journal, 1904, pp. 703 sqq.; 1905, pp. 643 sqq.

^{page 216 note 2} Professor Rapson, Indian Coins, p. 8, § 29.

^{page 216 note 3} For ref. see Professor Rapson, loc. cit.

^{page 216 note 4} Archœological Survey of India, New Imperial Series, vol. xx, pls. xliii–1.

^{page 216 note 5} See figures, pp. 48, 49, 68, 124, 134, of Inscriptiones Palœo - Persicœ Achœmenidarum by Dr. Cajetanus Kossowicz (St. Petersburg, 1872).