^{page 175 note 1} i. 203–4 (Dimensions and eastern and western shores), iv. 40 (southern shore).

^{page 175 note 2} iii. 117 (the river Akes and the landlocked sea). The fact that the preceding chapter deals with the Arimaspi, who are located East of the Caspian (v. inf. p. 191) at any rate helps to orientate this story.

^{page 175 note 3} See Tarn, , J.H.S. xxi. pp. 18–19.Google Scholar

^{page 176 note 1} Particularly in i. 204.

^{page 176 note 2} iv. 13.

^{page 176 note 3} Herodotus himself in iv. 24 tells us that Scythians and Greek traders from the Black Sea are his sources.

^{page 176 note 4} For convenience the various names of Aral and the Oxus are here given, viz. Oxus = Amu Daria, Jihun, and Viadme (Gonzales de Clavijo); Aral = Lake Kitchai or Kithay and Khwarizm. For the identity of the Oxus and Araxes, see below p. 190.

^{page 177 note 1} This is done by Minns: Greeks and Scythians, p. 10.

^{page 177 note 2} ‘Cosmographorum descriptio cum Marino accordata,’ reproduced in Hedin's, SvenSouthern Tibet, vol. i. PI. XII.Google Scholar

^{page 177 note 3} Sven Hedin, op. cit. Pl. XIII. This map is dated at 1459.

^{page 177 note 4} Sven Hedin, op. cit. Pl. XIX.

^{page 177 note 5} In Kircher's China illustrata: reproduced by Sven Hedin, op. cit. Pl. XI.

^{page 177 note 6} Delisle's map, Sven Hedin, op. cit. Pl. XLII.

^{page 177 note 7} Summarised by Wood, The Shores of Lake Aral (1876), chap. x.

^{page 178 note 1} This was, of course, effected by a gradual process of erosion along a line of least resistance and not by any sudden catastrophe. See Frazer, J.R.A.S i. xlvi. p. 278.Google Scholar The most recent geological pronouncement on the question states: ‘the depression of the region, changing the Bosporus from a river to a strait, is placed in the recent past during the existence of men.’ Pumpelly, , Explorations in Turkestan, 1905, p. 26.Google Scholar

^{page 178 note 2} It is interesting in this respect to note that the project of uniting the Caspian with the Euxine by a canal existed in the time of Seleucus. See Tarn, loc. cit., p. 19.

^{page 178 note 3} ‘As regards the basin of Aral, it is evident that, after the opening of the Bosporus and its consequent separation from the Caspian, it would have been placed and would have remained in its present state of isolation had not the quantity of water it received from the tributary rivers filled it up to overflowing and so caused its junction in another way with the Caspian.’ Wood, op. cit.p. 124.

^{page 178 note 4} Encyclopaedia Britannica (ed. 1902) (‘Lake Aral’).

^{page 179 note 1} The vexed question as to the course taken by the Oxus is dealt with below.

^{page 179 note 2} It is only fair to state, however, that in the middle of last century the opinion of geologists was not unanimous. Thus Rawlinson (Proc. R.G.S. 1867, March) held that the sea of Aral separated off from the Caspian in the Middle Ages, while Murchison (Journal R.G.S. 1867) held that both seas had a separate existence in prehistoric times.

^{page 179 note 3} Quintus Curtius tells us that the Caspian has but little depth on its northern side (vi. 4. 19).

^{page 179 note 4} iv. 109 and i. 202.

^{page 179 note 5} See Sven Hedin, op. cit. Pl. XXVI., Pl. XX., p. 183. Gerhard Mercator gives a lake called Kichai, near Tashkent. See also Wood, op. cit. p. 143.

^{page 179 note 6} The evidence of cartographers would obviously be insufficient for this dating. It is, however, strengthened by Rawlinson's view (see note 2). Minns accepts this general view as to the existence in historical times of an Aralo-Caspian Sea: Greeks and Scythians, p. 10. The name Aral occurs on the Carte Nouvelle de l'Asie Septentrionale, 1726 (Sven Heclin, op. at. PI. XLIII.), and is mentioned in the poetical account of his travels by a Greek, Vassili Vatatsis, who visited Turkestan about 1725–1730. (See S. P. Lambros, Μίκται Σελίδες p. 596). He claims to have been the first ‘to bring information about Aral to the learned men of Oxford and London.’ I am indebted to Professor Andreades of the National University, Athens, for this reference.

^{page 181 note 1} J.H.S. xxi. pp. 10 and 12.

^{page 181 note 2} I.e. (i) the Uzboi channel from Lake Sarykamish; (ii) the Unguz channel, which is 44 m. below the level of the Caspian, across the Karakum sands; (iii) from Charjui along the line now followed by the Central Asiatic Railway viâ Merv and Askhabad, and with the rivers Tejend and Murghab as tributaries. There are, however, two other possible routes, viz. (iv) from Charjui to Kizil Arvat (as in No. iii) and thence between Kopet Dagh and Kuren Dagh to the bed of the present river Sumbar and to the sea at Chikishlyar; (v) from Charjui due West across the Kara kum sands to the Igdy wells, thus joining up with the Uzboi channel (No. (i) above).

^{page 181 note 3} iv. 40.

^{page 181 note 4} Arrian distinctly says that both the Oxus and the Jaxartes flow into the Caspian, (Anabasis, iii. 29, 3 and iii. 30, 7).Google Scholar This affords no evidence as to the direction of flow of the Oxus, but is strong proof of the existence of an Aralo-Caspian Sea.

^{page 182 note 1} See Le Strange, Land of the Eastern Caliphate, p. 455 et seq. Mr. Tarn's arguments against the Oxus having flowed to this outlet are based on the geological evidence of the Russian engineer M. Konshin, who examined the Dardji peninsula (the reputed outlet of the Oxus in Balkhan bay) and found no trace of fresh-water deposits or river shells. It should be remembered, however, firstly that this is negative evidence only and in consequence not infallible, and secondly that the bay was almost certainly of greater extent inland, being part of the Aralo-Caspian Sea, so that a search for fresh-water shells would in any case be fruitless. In fact, Mr. Tarn himself says that a rise in the Caspian of only 20·17 metres would take the sea up the Balkhan bay as far as the small lake called Topiatan.

^{page 182 note 2} Narrative of the Embassy of Ruy Gonzales de Clavijo to the Court of Timur, published by the Hakluyt Society, p. 118.

^{page 183 note 1} E.g. Jonas Hanway in 1743 and Reynolds and Hogg in 1640, quoted by Wood, op. cit. p. 224.

^{page 183 note 2} The whole of this evidence can be summarised conveniently as follows:—

^{page 183 note 3} Wood, op. cit. p. 228.

^{page 183 note 4} Ch. 21 et seq.

^{page 183 note 5} Τάναϊν ποταμὸν διαβάντι οὐκέτι Σκυθική

^{page 184 note 1} Ch. 22.

^{page 184 note 2} It is unlikely that much of the province was ever inhabited to any extent, as between the Volga and the northern boundary of the Stavropol province it is for the most part, except in the coastal region, a salt and unfertile plain.

^{page 184 note 3} Op. cit. p. 104.

^{page 184 note 4} Westberg, quoted by Minns, op. cit. p. 113 (note).

^{page 184 note 5} Vol. ii. p. 32, Map II.

^{page 185 note 1} I.e. the Volga between Tsaritsyn and Astrakhan.

^{page 185 note 2} iv. 22.

^{page 185 note 3} iv. 109.

^{page 185 note 4} A collection of implements used in this industry and of photographs is in the museum at Askhabad in Transcaspia. The North Caspian is usually frozen up until April, and the Volga is only open in May. For the suggestion that Herodotus here refers to the seal fisheries, see Wood, op. cit. p. 131. A remarkable parallel to this passage occurs in the works of the seventeenth century Turkish traveller Evliya Effendi. Referring to the shores of the Caspian, he says, ‘the shore is covered with bones and carcases of strange kinds with square and pentagonal heads.’ Travels (Ed. von Hammer, 1850, ii. p. 164).

^{page 186 note 1} Travels in Caucasus and Georgia (English translation by Shober, 1814), p. 224.

^{page 186 note 2} Minns (p. 104) puts Gelonus at Kazan, which seems too far to be under Greek influence in so marked a way as Herodotus describes. Wood (op. cit. p. 131) suggests Urgenj, but this is equivalent to placing the Budini in the territory of the Massagetae, and is, in any case, on the wrong side of the Caspian.

^{page 187 note 1} That he is dependent on traders for most of his information about the tribes between the Don and the Argippaei is clear from i. 24, where he tabulates his authorities as (a) Scythians, (b) Greeks from the Borysthenes, (c) Greeks from other Euxine towns.

^{page 187 note 2} i. 204.

^{page 188 note 1} iv. 24. μέχρι μέν νυν τῶν φαλακρῶν τούτων πολλὴ περιφάνεια τῆς χώρης ἐστὶ καὶ τῶν ἔμπροσθε ἐθνέων.

^{page 188 note 2} Their nomadic habits have been largely revived during the last two years owing to the insecurity caused by recent political upheavals in the province.

^{page 188 note 3} iv. 23.

^{page 188 note 4} Of. cit. p. 109.

^{page 189 note 1} If the derivation of the name Oxus from the Turkic word Aksu is correct, we may have an example of a pre-Greek Turkic word.

^{page 190 note 1} i. 201. οἰκημένον δὲ πρὸς ἠῶ τε καὶ ἡλίου ἀνατολάς πέρην τοῦ ᾿Αράξεω ποταμοῦ ἀντίον δὲ ᾿Ισσηδόνων ἀνδρῶν

^{page 190 note 2} Minns, for no very clear reason, interprets ‘opposite’ as ‘to the West.’

^{page 191 note 1} It is curious that in Bk. iii. 17 the description of the river Akes, which is somewhere near the Caspian, comes immediately after an account of the Arimaspi. Information about both has every appearance of being derived from folk stories, and it is tempting to think that Herodotus had the Arimaspi entered up in his notebooks under the heading ‘Notes from Transcaspia.’

^{page 192 note 1} Arrian, , Anabasis, iii. 27, 4.Google Scholar

^{page 192 note 2} Presumably the Euxine, though it is curious that he calls it simply ἡ θάλασσα (iv. 28).

^{page 192 note 3} I saw snow about this time on the lower hills of Kopet Dagh, and the plains were in places still covered with pools of water.

^{page 192 note 4} E.g. Minns, op. cit. (see maps). He identifies the Issedones with the Yüeh Chih on the Tarim river.

^{page 193 note 1} Arrian, , op. cit. vii. 16, 1.Google Scholar

^{page 193 note 2} Ibid. vi. I. 2.