^{page 10 note 1} A summary of the views of modern Russian geographers will be found in an article (with map) by Kropotkin, Prince, ‘The old beds of the Amu-Daria,’ Geogr. Journ. vol. 12 (1898), p. 306.CrossRefGoogle Scholar It has always been, and still is, a Russian dream to turn the Oxus back into the Caspian.

^{page 11 note 1} Konshin's, M. results in Moser, À travers l'Asie Centrale (1886), p. 228Google Scholarseq.; ‘The old channel of the Oxus,’ by Mr.Morgan, E. Delmar, Proccedings R. G. S. vol. 14 (1892), p. 236Google Scholar; and Prince Kropotkin's article above cited. I only know them at second hand.

^{page 11 note 2} Beside the sea and river theories of the Uzboi, the view has been put forward by Bogdanovich that this channel, other than the extreme western portion, which may be due to the action of the sea, has been formed by rain. This view is examined by Komischke, W. in Das Ausland for 1893, p. 657, ‘Die Hydrographie des Oxus-Beckens’Google Scholar; he sums up that, though one cannot trace all the steps of the gradual separation between Aral and Caspian, ‘wahrscheinlich bestand der aibugirische sowie der balchanische Abfluss in seiner ursprünglichen Bedeutung als Meeresstrasse bis in die historsche Zeit hinein.’

^{page 12 note 1} ‘Das Oxus-problem in historischer und geologischer Beleuchtung,’ in Petermann's, Mittheilungen (1898)Google Scholar, No. 9.

^{page 12 note 2} Herod. 1, 202: οὐ συμμίσγουσα τῇ ἐτέρῃ θαλάσσῃ

^{page 13 note 1} Mr.Myres, J. L., in a paper read before the Geographical Society on ‘An attempt to reconstruct the maps used by Herodotus’ (Geogr. Journ. vol. 8 (1896), p. 605)CrossRefGoogle Scholar, has put forward a theory that Herodotus had two different ideas about the Caspian, based on different maps, and that in 4, 40, he (Herod.) ‘assumes that the Caspian, as a part of the undiscovered “North Sea,” corresponds with the known Red or “Southern” Sea, a conclusion which reappears in Eratosthenes,’ and which is inconsistent with Herod. 1, 202.

^{page 13 note 2} 1, 202: he calls it the Araxes. It is generally supposed to represent the Jaxartes, because of the marshes; but, a priori, it is much more likely to be the larger and better known Oxus, which must, in a natural state, have had an equally marshy mouth or mouths.

^{page 13 note 3} Alexander, in his speech at the Hyphasis (Arr. Anab. 5, 26) says: καὶ ἐγὠ ἐπιδείξω . . .τὸν μὲν ᾿Ινδικὸν κὸμπον ξὑῥῥουν ὄντα τῷ Περσικῷ τὴν δ᾿ Γρκανίαν τῷ ᾿Ινδικῷ but this, if he said anything of the sort, is clearly special pleading.

^{page 13 note 4} In Plutarch, (Alex. 44)Google Scholar Alexander is made to take the Caspian for part of the Maeotis.

^{page 13 note 5} Strabo 11, 507. This is Susemihl's, opinion (Gesch. d. Griech. Lit. in der Alexandriner-zeit 1, 657–9)Google Scholar; and though Strabo does not actually say so, we know of no other Greek who ever sailed on the Caspian, and Strabo says that it was little exploited, owing to the brief and disturbed nature of the Macedonian rule in those parts (11, 509); besides, Strabo expressly cites the measurement of one part of this periplus, the distance between the mouths of the Oxus and Jaxartes, as Patrocles’ (11, 518), and Eratosthenes (l.c.) speaks as if no other periplus were known τὸν ὑπὸ τῶν Ελλήνων γνωριζόμενον περίπλουν

^{page 14 note 1} Pliny, N.H. 6, 13Google Scholar, quoting from the same passage in Eratosthenes, has ‘ab exortu et meridie per Cadusiae et Albaniae oram.’

^{page 14 note 2} Aristohulus (Strabo 11, 509) notes a deficiency of light wood in Hyrcania, though plenty of oak.

^{page 14 note 3} μυχός

^{page 14 note 4} Strabo 11, 508: τούτων (the mountains) ἐστὶ μηνοειδὲς τὸ σχῆμα κατὰ τὰς ὑπωρείας αἴ τελευ τῶσαι πρὸς θάλατταν ποιοῦσι τὸν μυχὸν τοῦ κόλπου So Pliny 6, 13 lunatis cornibus; Curtius 6, 12. The map does not permit of identification; but Curtius shews that the crescent meant was only a blunt one, flexu modico.

^{page 14 note 5} Strabo 11, 507: ἔστι δ᾿ ὁ κόλπος ἀνέχων ἐκ τοῦ ὠκεανοῦ τρὸς μεσημβρίαν κατ᾿ ἀρχὰς μὲν ἐκανῶς στενός ἐνδοτέρω δὲ πλατύνεται προϊών καὶ μάλιστα κατὰ τὸν μυχὸν ἐπὶ σταδίους που καὶ πεντακισχίουις ὁ δ᾿ εἴσπλους μέχρι τοῦ μυχοῦ μικρῷ πλειόνων ἀν εἴη συνἀπτων πως ἤδη τῇ ἀοικήτῳ ‘Sailing-in point’ is of course not meant as a translation of εἴσπλους it is the point whence the length of the εἴσπλους is reckoned, sometimes (2, 74, 119; 11, 491) called στόμα

^{page 14 note 6} The evidence for this will appear, p. 17 seq. Here I merely wish to note that in one passage (11, 519) Strabo scems to think the στόμα is in the North. But I think, as will appear, that there is often a distinction between Strabo's view, and the true view that he has preserved without always understanding it. Even in 11, 519 the στόμα is straight opposite to the μυχός and that the μυχός is S.W. is indisputable, and (so far as I know) generally admitted.

^{page 15 note 7} Susemihl, l.c. The Greeks of about Strabo's time seem to have known nothing definitely of the northern part. Their names for the sea, Caspian, Hyrcanian, Albanian, are southern local names, originally no doubt signifying different stretches of water (Pliny 6, 21 circumvectis in Hyrcanium mare et Caspium 6, 13 ante quos mare quod est Albanum nominatur cf. Arist., Meteor. II. 1 § 8)Google Scholar; but they have no northern local names, unless Scythicus sinus (Pliny 6, 13; Pomponius Mela 3, 5) be one. Arrian, (Anab. 7, 16)Google Scholar says the ἀρχαί of the sea had not been discovered; but Strabo, by giving the length and breadth, seems to have thought it was bounded all round, subject to the question of the εἴσπλους And so, clearly, did the authorities from whom Pliny (6, 13) took the phrase ‘circuitum a freto.’

^{page 14 note 8} Negative criticism, refuting earlier attempts (based on the measurements) to locate the point reached by Patrocles, , in Wagner's, ‘Patrocles am Kara-Bugas?’ Nachr. v.d. König. Gesellschaft (Göttingen) 1885, p. 209Google Scholar. It appears to me that the writer proves his points; but that the problem has rather shifted its ground. Any system of measuring out this voyage must be vitiated (other things apart) by the fact that we do not know where to measure from; for that Patrocles started from the mouth of the Mardus (Eizil Uzen), though likely enough, is merest guesswork.

^{page 15 note 1} Pliny 6, 11. He gives no express authority for this statement; but he has used some good sources in book 6, as well as bad.

^{page 15 note 2} Strabo 11, 507 l.c.

^{page 15 note 3} He thought it as large as the Euxine: (Strabo 11, 508). I shall say something about the measurements presently.

^{page 15 note 4} Strabo 11, 501.

^{page 15 note 5} Strabo 11, 507.

^{page 15 note 6} Strabo 11, 518; Aristobulus is the other.

^{page 15 note 7} I know of nothing to warrant SirRawlinson's, H. statement (Proceedings R.G.S. 1 (1879) p. 161)Google Scholar that Patrocles ‘actually measured’ the distance.

^{page 15 note 8} A proceeding that Strabo must be criticising when he insists (11, 518), with illustrations, on the extraordinary variation of length of the parasang in different places. Elsewhere (11, 507) he says that these measurements of Eratosthenes' are to be received with caution.

^{page 16 note 1} Cf. Moser, , À travers l' Asie Centrale p. 228Google Scholar seq.: and Komischke's article before cited p. 11. n. 2.

^{page 16 note 2} E.g. the gulf of Kara Bugas; or the promontory of Mangischlak (von Gutschmid).

^{page 16 note 3} Susemihl, l.c.; Bunbury, , Hist. of Anc. Geog. 1, 644.Google Scholar

^{page 16 note 4} Strabo 11, 507, εἰσπλέοντι That this is not a figure of speech (=εἰσιόντι) is proved by the use of εἴσπλους just before: cf. 2, 121, where he balances the 4 great seagulfs, Caspian, Persian, Arabian, and Mediterranean, each with a narrow εἴσπλους from the outer sea. Cf. Pliny 6, 13 ab introitu; Pomponius Mela 3, 5, 4 introeuntium.

^{page 16 note 5} Op. c. 2, 283. The italics are in the original. Strabo has been even worse treated by the writer of the article ‘Caspian’ in the Encycl. Britannica, who refers to him à propos of a great Aralo-Caspian sea discharging into the Obi—presumably at some geological epoch.

^{page 17 note 1} Strabo 11, 508: . . . Πάρνους εἶτ᾿ ἔρημος πρόκειτι μεταξύ καὶ ἐφεξῆς ἡ Γρκανία καθ ἤν ἤδη πελαγίζει

^{page 17 note 2} Strabo 11, 515 Parnoi said to be Δάας μετανάστας from the Daai beyond Maeotis; some of them dwell on the Ochus. But the best commentary on the above is 511, where it appears clearly that Strabo imagines 3 parallel belts, (1) cultivated land, Hyrcania, Nesaia, Parthia; (2) Desert; (3) nomads, Daai, Aparnoi and others, the Aparnoi nearest Hyrcania; they raid regularly across the desert. Cf. Ptolemy 6, 10; and Agathodaemon's map, which places the Daai and Parnoi south of the Oxus.

^{page 17 note 3} Strabo 11, 507 already cited. Poinponius Mela 3, 5, 3 Mare Caspium ut angusto ita longo etiam freto primum terras quasi fluvius irrumpir, Pliny 6, 13 irrumpit autem arctis faucibus et in longitudinem spatiosis. … utrinque accolunt Seythae et per angustias inter se commeant. Ar., Pseud.de mundo 3, 11Google Scholar; Solinus 14, 18. It is not clear that Patrocles is the common source, but I am willing to assume it; anyhow he spoke of the ‘mouth’ of the Caspian (Strabo 2, 74).

^{page 17 note 4} That this sea-strait cannot be far from where I have put it, and that it has nothing to do with the north, is also shewn by Strabo's calling the ‘mouth’ of the Caspian ‘the Hyrcaniaa mouth’ (11, 519) [cf. n. 1, p. 24]. Pomponius Mela also connects the ‘os’with the ‘sinus Hyrcanus,’ 3, 5, 3. His account has become very confused; but it may be worth noticing that on the narrow strait he places the Derbikes (3, 5, 4), a tribe whom Strabo (11, 514) places near the Hyrcanians and Pliny 6, 16, on either side of the Oxus. Ptolemy 6, 10 puts them in Margiana. on the Oxus.

^{page 18 note 1} I assume this, not as necessarily being the fact, but as being most against my own view.

^{page 18 note 2} That the strait was no bar to the intercourse of the tribes on either side of it. But Turcomans swim the Oxus at its broadest. And the ‘mouth’ was looked on as narrow; Agathemerus (3, 13) says 4 stades across.

^{page 18 note 3} Curtius indeed (6, 12) hints that great intermittent floods of water came into the Caspian. After speaking of the way this sea sometimes flooded the land and then retired, he says ‘et quidam credidere, non Caspium mare esse, sed ex India in Hyrcaniam cadere.’ Solinus appears to have heard a similar story and to refer it to snow-water; 14, 18 Caspii maris ‥ fauces maciantur imbribus, crescunt aestibus (if Th. Mominsen's reading be correct). It would be interesting to know the source of this.

^{page 18 note 4} I.e. that the Caspian joined Ocean, and that one could sail from India into it.

^{page 18 note 5} Strabo 2, 74: τοῦ στόματος τῆς Κασπίας θαλάττης . . . ὄπερ . . . δοκεῖ . . . περίπλουν ἔχειν ἀπὸ τῆς Ινδικῆς δυνατόν ὤς φησιν Πατροκλῆς

^{page 18 note 6} 2, 74 same passage: ὄπερ . . . δοκεῖ αὐτῆς τῆς παραλίας μέχρι τῆς Ινδυκῆς ἀρκτικώτερον εῖναι σημεῖον

^{page 18 note 7} Strabo 11, 518: see note 4, p. 19.

^{page 19 note 1} Pliny 6, 17. Cf. Pomp. Mela 1, 2, 3; Solinus 50, 1.

^{page 19 note 2} I.e. that the cold land of the north was bounded by water. Pytheas had spoken of τῆς πεπηγυίας θαλἀττης Strabo brings the mouth of the Caspian and Ierne into connection as being both far north 2, 119. In 11, 507 the εἴσπλους of the Caspian is συνάπτων πως ἤδη τῇ ἀοικήτῳ

^{page 19 note 3} Ath. 5, 201 c ἄρκτος λευκὴ μεγάλη μία

^{page 19 note 4} The passages in Strabo are (1) 2, 73: τὸν Ωξον οὔτω φασὶν εὔπλουν εἶvαι ὤστε τὸν Ινδικὸν φόρτον ὐπερκομισθέντα εἰς αὐτὸν ῥᾳδίως εἰς τὴν Γρκανίαν κατάγεσθαι καὶ τοὺς ἐφεξῆς τόπους μέχρι τοῦ Πόντου διὰ τῶν ποταμῶν (2) 11, 509: φησὶ δὲ καὶ εὔπλουν εἶvαι (τὸν Ωξον) καὶ οὖτος (Aristobulus) καὶ ᾿Ερατοσθένης παρὰ Πατροκλέους λαβὼν καὶ πολλὰ τῶν Ινδικῶν φορτίων κατάγειν εἰς τὴν Γρκανίαν θάλατταν ἐντεῦθεν δ᾿ εἰς τὴν ᾿Αλβανίαν περαιοῦσθαι καὶ διὰ τοῦ Κύρον καὶ τῶν ἐξῆς τόπων εἰς τὸν Εὔξεινον καταφέρεσθαι (3) 11, 518: ρὐχ ὁμολογοῦσι δ᾿ ὄτι περιέλευσάν τινες ἀπὸ τῆς Ινδικῆς ἐπὶ τὴν Γρκανίαν ὄτι δὲ δυνατὸν Πατροκλῆς εἴρηκε Of these (1) and (2) represent a common original. There is nothing here about any Indian ocean, and I doubt if there ought to be anything about the Caspian either. (1) and (3) only say ‘to Hyrcania:’ (2) says ‘to the Hyreanian sea.’ Now it is not natural to say that from the Hyreanian sea ἐντεῦθεν goods were carried across (i.e. across the Hyreanian seaπεραιοῦσθαι from one side to the other) to Albania. I would suggest that θάλατταν may be a gloss, inserted by some one who had Pliny 6, 17 in his mind, where the Caspian is certainly mentioned. If it be the right reading, there is nothing to account for its omission in (1), and this appears to me conclusive. Of course, it may be contended that τὴν Γρκανίαν alone means the Hyreanian sea; but is this possible unless the context render it unmistakeable? I have been through practically every instance of ἡ Γρκανία the sea, given in Pape's Wörterbuch d. Griech. Eigennamen; there are 4 cases (Arr. Anab. 5, 26, 1 Strabo 2, 129 and 11, 519 and Ptol. 5, 13, 6) where θάλασσα is left to be supplied, and in all these passages ἡ Γρκανία θάλασσα has been mentioned just before and the wording of the context makes the meaning unmistakeable: neither of these is the case in (2).

^{page 20 note 1} E.g. von Gutschmid in art. ‘Persia’ in Enc. Brit. Tozer, , Hist. of Anc. Geog. 136Google Scholar. Pliny gives a wild story of circumnavigation 6, 21; 2, 67.

^{page 20 note 2} It may be of interest, in this connection, to note Peter the Great's orders to the ill fated Bekovitch expedition. They were to go up the old bed of the Amu to Khiva, win over the Khan, turn the Amu back into the Caspian, and sail in the Khivan boats towards India. Humboldt, , Asie Centrale 1, 425.Google Scholar

^{page 20 note 3} The story of Eudoxus of Cyzicus, Strabo 2, 98 seq.

^{page 20 note 4} Strabo 11, 507 compared with 518. Pliny 6, 13 gives 4,800 and 2,400 stades, presumably following Eratosthenes. The symmetry of these measurements is suspicious.

^{page 21 note 1} Strabo 2, 74.

^{page 21 note 2} Strabo 11, 507.

^{page 21 note 3} Strabo 11, 507, 510, 518; Arrian 7, 16; Pomponius Mela 3, 5, 6; Ptolemy.

^{page 21 note 4} Is this what Strabo means in 2, 173: τὸν ὠκεανὸν τόν τε ἔξω καὶ τὸν τῆς Γρκανίας θαλάτ της There is no difficulty in the supposition. The list of names of oeeanus in Solimis 23, 17, includes Hyrcanus and Caspius.

^{page 21 note 5} Strabo 11, 509: this story must rest upon the real confusion of the European and Asiatic Tanais (Jaxartes). It does read very much like the supposition of a waterway between the Aral and Caspian. Cf. Curtius 6, 12 ‘alii sunt, qui Maeotiam paludem in id (Caspium mare) cadere putent.’ A very clear case of Maeotis meaning the Aral is Polyb. 10, 48; the Apasiacae dwell on the middle Oxus and Tanais, the Oxus entering the Hyrcanian sea and the Tanais the Maeotis.

^{page 21 note 6} Strabo 11, 509, 510: ὐπόγλυκυ

^{page 21 note 7} Prof. Walther (l.c.) uses this avgument about Anthony Jenkinson's journey. Pliny repeats the statement (6, 17 haustum ipsius maris dulcem) on the authority of ‘Alexander Magnus’ and M. Varro, attributing the fact to the inflow of the rivers. So Solinus 19, 3. Curtius (6, 12) also gives it, as due to the inflow of the Maeotis. The Caspian is salt, the northern section (which is very shallow compared to the rest) being less so than the rest of the sea, owing to the inflow of the Volga and Ural. The Aral is generally said to be only slightly brackish. Hedin, M. Sven, however, (Through Asia, 1, 49)Google Scholar, says, that it is too salt to drink, except at the river mouths; ‘but far out in the lake there are said to exist certain fresh-water belts.’ I must thank Mr. G. F. Hill for calling my attention to this.

^{page 22 note 1} This explains why the Greeks (apparently) never mention the Aral, a fact which has led some to suppose, either that they did not know of it (Bunbury), or that it did not exist. They always mention it as something else, Caspian, Maeotis, or (perhaps) Ocean. This view also perhaps throws some light on the confused arrangement of the three gulfs of the Caspian in Pomponius Mela 3, 5; his Scythieus sinus, on the left hand as one enters by the ‘mouth,’ and receiving the Oxus and Jaxartes, must be the Aral. That the Aral existed is clear from the Chinese accounts. In A. Wylie's translation of Notes on the Western Regions, from the Annals of the elder Han, (Journ. Anthrop. Inst. 1881 at p. 44)Google Scholar the Yentsai are said to live about 2,000 le N.W. of the Khang-kiu (the nomads settled on the Polytimetus) ‘on the border of a great marsh without banks, which is the Northern sea.’ Tchang-kien, on whose report this account is based, had personally visitedthe Khang-kiu (p. 67); the date is now generally given as about 128 B.C. In the corresponding passage of the Shi-ki of Sze-ma-t'sien ch. 123, Kingsmill, T. W.'s translation (‘Intercourse of China with Eastern Turkestan,’ J.R.A.S. 1883, vol. 14, p. 80)Google Scholar gives ‘a great marsh, without defined banks, covered with reeds, and (communicating with) the Northern sea.’ Gigantic swamps (one of 2,000 sq. miles) still exist near the mouth of the Syr.

^{page 22 note 2} Strabo 11, 512, 513.

^{page 22 note 3} τὴν ἄλλην τὴν πρὸς ἄρκτοις θἀλατταν

^{page 22 note 4} πρὸς τὸν κόλπον τὸν Γρκάνιον It is clear that this will not suit the Jaxartes.

^{page 22 note 5} Incidentally, this shews the confusion in Strabo's mind as to whether this strait ended in ocean or in some sea: cf. p. 21.

^{page 22 note 6} It is perhaps interesting to note that in the same chapter in which he defines the Caspian asa lake, ‘rather like the opposite of an island,’ he calls the Peloponnese an island. On the whole, his notions of this part of the world are confused, and inferior to those of Strabo.

^{page 23 note 1} Proceedings R.G.S. vol. 20 (1876) p. 178; vol. 1. (1879) p. 161 seq. ‘The road to Merv;’ vol. 4 (1882) p. 355; vol. 5 (1883) p. 14.

^{page 23 note 2} The Aria palus of Ptolemy, the Akes pool of Herod. (3,117.)

^{page 23 note 3} Strabo, 11, 509, 518.

^{page 23 note 4} Ptolemy, 6, 10.

^{page 23 note 5} Pliny, 6, 16 difficilis aditu propter arenosas solitudines. Solinus 48, 2 has turned this into ‘paene inaccessa.’

^{page 23 note 6} Strabo, 11, 516.

^{page 23 note 7} Strabo 11, 511. The desert is ἄνυδρος So in the Zend Avesta, the ‘plague’ of Merv is an evil concourse of horsemen and robbers.

^{page 23 note 8} Strabo 11, 518.

^{page 23 note 9} This appears from the names of the five nations connected with it.

^{page 23 note 10} Strabo 11, 514.

^{page 23 note 11} Strabo's notices of the Ochus (11, 509–511, 518) come to this; it flows through Hyrcania and Nesaia, and near Parthia; and comes from the Indian mountains. Some say it falls into the Oxus, some into the Caspian, after an independent course; some say it flows through Bactria, some that it bounds it. So far as I know, no one else tells any new fact about it. But a statement is sometimes quoted from Curtius, to the effect that Alexander crossed both Ochns and Oxus marching from Samarcand to Merv: e.g. by SirRawlinson, H.Proceedings R.G.S. 20 (1876) p. 178.Google Scholar No such march of Alexander's is known, and a reference to Curtius 7, 40 shews that Curtius says he started from Bactra to punish the rebels, on the 4th day reached the Oxus, and then crossing Ochus and Oxus (superatis deinde amnibus Ocho et Oxo, one MS. Oxo et Meo) reached Margania (ad urbem Marganiam pervenir—Margianam is only a conjecture). That is to say, he crossed back into Sogdiana after the rebels. Round Margania, says Curtius, he built 6 towns. Margania seems unknown. There seems no need to make even Curtius romance unnecessarily; there is nothing here about Merv or a southern Oxus. All that can safely be said about Strabo's Ochus is that it seems to be a confusion of two rivers, one a Bactrian tributary of the Oxus; the other would be well suited by the Atrek (Bunbury).

^{page 24 note 1} There seem to be traces of such a wider meaning in Strabo. 11, 519 he speaks of the ‘mouth’ of the Caspian as the ‘Hyrcanian’ mouth; on any theory, it was not in Hyrcania proper. 11, 513 the one branch of the ‘Araxes’ enters the Hyrcanian κόλπος i.e. the narrow strait (507), between Balkan bay and the Aral. Did ‘Hyrcania’ follow the ‘Hyrcanian gulf’? It would be a tempting conjecture that somewhere in Strabo's sources a confusion had occurred of Gurgân (Hyrcania) and Gurgânj (Orgunje); but there appears to be no real authority for the name Gurgânj till much later. (Dr.Sachan, E., ‘Zur Gesch. und Chronologie von Khwârizm’ in Silz. der K. Akad. der Wiss., Wien 1873 vol. 73 at p. 472)Google Scholar. Sir H. Rawlinson had conjectured this name for Urva in the 1st Fargard of the Vendidad (verse 38); but see now Daimosteter's trans, in ‘Sacred Books of the East.’

^{page 24 note 2} Hellenismus ², III. 2, 253. Enough is known, perhaps, about Alexander and the towns he founded to make the argument from silence a fair one. It is sometimes stated (e.g. Roesler, , ‘die Aral seefrage,’ Sitzungsb. der philosophisch-hist. Classe d. K. Akad. der Wiss., Wien 1873 vol. 74 p. 186)Google Scholar that Ptolemy places a town Aspabota at the Oxus mouth. It is really put two degrees from it, and much nearer the Polytimetus (Ptol. 6, 14, 2).

^{page 25 note 1} See p. 19, note 4. In the first passage (2, 73) there cited, Strabo appears to have himself added the word ‘easily’ to his original, for purposes of controversy, his argument in that part of book 2 compelling him to insist on the fertility and resources of the provinces north of ‘Taurus.’

^{page 25 note 2} 6, 17; repeated by Solinus 19, 4.

^{page 25 note 3} The Baetrus is said to have then reached the Oxus, Strabo 11, 516. Later, a Turkish geographer says that the Balkh river entered the Oxus at Termedh (Ritter, , Erdkunde pt. 8 bk. 3, 219Google Scholar).

^{page 25 note 4} Among recent writers, who repeat without comment the statement that goods could be shipped on the Oxus and taken by its ancient course to Balkan Bay, may be mentioned Brunnhofer, , vom Aral bis zum Gangâ, (1892)Google Scholar who has a good deal about it, p. 129, p. 134–141, and who (Iran tend Turan, p. 113 seq.) speaks of the ‘Ungeheuern Transithandels;’ Skrine and Ross, , The Heart of Asia, (1899), p. 315Google Scholar; Tozer, , Hist. of Anc. Geog. (1897), p. 134.Google Scholar On the contrary, among older writers, who are generally positive about it, Roesler, , die Aralseefrage (cited above, 1873)Google Scholar, while believing the Oxus reached Balkan Bay, already suggested it was of little importance for trade (p. 215). SirHunter, W. W., History of British India, (1899) vol. 1. p. 31–33Google Scholar has a very guarded statement about this route; his map shews no trade-route to the Caspian by the Oxus, but a land route from Kashgar viâ Merv to Asterabad, thence (1) ship to mouth of Cyrus, (2) caravan through Armenia to Trebizond (3) caravan viâ Euphrates to Syria. Mr.Kennedy, J., ‘The early commerce of Babylon and India,’ J.R.A.S. 1898Google Scholar, expresses, I think, the facts of the case in saying, (p. 242), ‘Articles of commerce doubtless passed along this way from early times: but the trade was of little importance, fitful, intermittent, and passing through many intermediate hands,’ but he adds ‘until the Parthian domination forced trade into this channel.’ What is the evidence for the statement about the Parthians? And, a priori, why should they try to force trade into a channel entirely outside their own dominion or control? See note 5, p. 26.

^{page 26 note 1} 11, 509: ἄπλους τε οὖσα καὶ ἀργός Cf. Pomponius Mela 3, 5, 3: omne atrox, saevum, sine portubus … belluis magis quam cetera refertum et ideo minus navigabile.

^{page 26 note 2} Ptolemy 1, 11 and 12. See Bunbury op. cit. 2, 529 seq., who follows Colonel Yule in thinking that the silk came by this road. Ptolemy does not say so; but he does rather imply that the whole of it was one route; and of course it was the silk route in the portion east of Bactra. Frazer (note to Pausanias 6, 26, 6) says the silk went overland from N. China by Samarcand to the Caspian, citing Ptolemy, 1, 11; this may be correct, but is hardly what Ptolemy says.

^{page 26 note 3} The mere argument from silence is of little value in this history of scraps and fragments, after Alexander. As to the Macedonians, we have some little evidence in Strabo 11, 509. They had no time.

^{page 26 note 4} No coin-finds, so far as I know. And see Appendix, p. 28.

^{page 26 note 5} The Parthians, a small aristocracy of great slave-owners, did not usually bear a mercantile character: see von Gutschmid, , Gesch. Irans, pp. 56, 65Google Scholar; though no doubt glad to enrich themselves by tolls. But the fact that Vardanes, when he pursued a beaten enemy to the Tejend (lower Arius) boasted of having reduced nations who never before paid tribute to an Arsacid, is very much in point here, as shewing what strangers the Parthians had then become in this part of the world (von Gutschmid, , Gesch. Irans p. 126)Google Scholar.

^{page 26 note 6} Ptol. 1, 11 § 7.

^{page 26 note 7} Strabo 11, 506: ἐχρυσοφόρουν δὲ διὰ τὴν εὐπορίαν

^{page 26 note 8} This attempt is known only from Chinese sources, as to which see Dr.Hirth, F., China and the Roman Orient (1885), p. 42.Google Scholar

^{page 27 note 1} Wylie, A. ‘Notes on the Western Regions,’ J. Anth. Inst. 1881, p. 40Google Scholar, cited above.

^{page 27 note 2} ‘China and the Roman Orient,’ where everything bearing on the question is collected. The passage quoted is p. 39, from the Annals of the later Han.

^{page 27 note 3} Von Gutschmid, , Gesch. Irans, pp. 56, 134.Google Scholar

^{page 27 note 4} Hirth p. 146.—See also on this story de Lacouperie, T., The western origin of Chinese civilisation, pp. 222, 226.Google Scholar Among recent writers Drouin, M. E. (Art. ‘Bactriane’ in Grande Encyclopédie, p. 1120)Google Scholar still refers it to the Caspian; von Gutschmid op. c. p. 138 seq. to the Mediterranean, which is out of the question for quite a number of reasons. There is an odd parallel to what the sailors told Kanying to be found in Dionysius Periegetes, who says of the Caspian that you would not cross it in three months (719, 720); I do not think this has been noticed, but it must be mere coincidence. A missing link in his proof, on which Dr. Hirth and others lay some stress, is that according to the Hou-han-shu the rhinoceros was found in T'iao-chih (in his view Babylonia), and this cannot be proved for Babylonia. If the rhinoceros had once lived here, it would not be difficult to credit its return after the canals began to go to ruin under the Seleucids; and we know that some pachydermata had a very different range in antiquity to the present day; both Thothmes III. and Tiglath-pileser I. found wild elephants numerous about the upper Euphrates. But so far there seems to be no proof of the rhinoceros at all, in spite of the fondness of the Assyrians for representing animals; for the ‘rhinoceros’ of the black obelisk of Shalmaneser II. is an ox (Hommel, , Gesch. Babyloniens und Assyriens 602, 603)Google Scholar, like the ‘unicorns’ or ‘rhinocerots’ of Isaiah 34, 7. Is it quite certain that the animal mentioned in the Hou-han-shu is a rhinoceros? Anyhow, the same difficulty applies to any other location of T'iao-chih.