¹ The only Mediterranean source of tin where the existence of ancient workings can be proved is Monte Valerio, near Campiglia in Tuscany. These workings were very difficult and exiguous (Daubrée, , Revue Archéologique, 1881, p. 235Google Scholar).

Modern prospectors have found tin, but no evidence of ancient exploitation, near Granada (Bérard, , Les Phéniciens, i. 445Google Scholar), Almeria and Cartagena (Fawns, , Tin Deposits, p. 181Google Scholar). As has often been pointed out, the tin deposits of the river Tartessus (Scymnus, ll. 164–6; Stephanus, s.v. Τάρτησσος), and on two islands at the head of the Adriatic (Scymnus, ll. 392–3) are a myth. The Adriatic tin may have come from Bohemia or Saxony by one of the old amber routes (Déchelette, , Manuel d'Archéologie, i. 626Google Scholar), or, more probably, from the Atlantic via Gaul and the Po valley. The Tartessus tin was certainly of Atlantic origin.

Several tin workings, presumably of ancient origin, have been discovered in Central France. Those at Vaulry (Haute Vienne) and Montebras (Creuse) were fairy extensive (Daubrée, p. 274 sqq.) But these mines could not have gone far to meet the Mediterranean demand for tin.

² Some old tin workings have been found on Tresco Island in the Scillies (Borlase, , Observations, p. 73Google Scholar). But these are almost microscopic.

Smith, G. (The Cassiterides, pp. 52–3)Google Scholar and Torr, C. (The Academy, 1895, pp. 342–3)Google Scholar suggest that the Greeks were misled by a Phoenician omnibus word which did duty for our ‘island,’ ‘headland’ and ‘overseas land.’ This is unnecessary. Explorers ancient and modern have often mistaken mainlands for islands. Cf. Ptolemy, ‘Island of Scandia’ (Geogr. II. ii. 35)Google Scholar, Marco Polo's ‘Island of Zenzibar,’ i.e. Central Africa (Travels, ch. xxxvii.), and the sturdy belief of the discoverers of America that there was a waterway between the new lands to the East.

For the opposite error, cf. the obsession in a continuous ‘Terra Australis,’ which was not dispelled till the time of Bass and Flinders (1800).

³ Reinach, S., L'Anthropologie, 1899, p. 401Google Scholar; Holmes, Rice, Ancient Britain, pp. 483–98Google Scholar; Haverfield, in Pauly-Wissowa, s.v. Κασσιτερίδες.

⁴ See Déchelette, i. pp. 428, 626. The use of bronze in Western Gaul spread, not from Brittany, but from Gascony. This implies that it was imported from Spain.

⁵ Strabo, p. 147:

⁶ Nat. Hist. xxxiv. 16 (47), § 156: nunc certum est plumbum album in Lusitania gigni et in Gallaecia.

⁷ This may be inferred from the great scale of the Roman workings, and the depth to which they were eventually excavated (Borlase, , Tin Mining in Spain, p. 15Google Scholarsqq.).

⁸ The ‘Oestrymnis’ of Avienus, (Ora Maritima, 1. 90Google Scholarsqq.), where ‘tin and lead are found in plenty,’ clearly is Brittany. (See Siret, , L'Anthropologie, 1908, p. 130Google Scholar, and Schulten's recent edition of Avienus.) It matters little whether Avienus drew upon the Carthaginian Himilco, or upon an earlier Massiliote writer, as Schulten contends on good grounds. In any case his ultimate source goes back to about 500 B.C.

⁹ See the map at the end of Vol. ii. Pt. 1 of Déchelette's Manuel.

¹⁰ Siret (loc. cit.) points out that when the Breton miners had scoured away the alluvium they would have had to follow the tin into some very hard quartz lodes. Small wonder if their industry migrated to other sites where the tin could still be streamed or the stanniferous rock was more friable.

¹¹ If the prominence of Ireland in Avienus' poem (ll. 108–110) is not due to a mere fancy of Avienus but to its commercial importance, we may assume that its exports were not so much of tin, as Schulten, (Tartessos, p. 67)Google Scholar supposes, but of gold (Coffey, The Bronze Age in Ireland, ch. iv.).

¹² Diodorus, v. 22; Pliny, iv. (16), 30, § 104. On these passages see p. 174.

¹³ Bellum Gallicum, v. 12, 5: nascitur ibi (sc. in Britannia) plumbum album in mediterraneis regionibus. Caesar's error as to the situation of the mines is immaterial. Presumably the small quantities of tin used by his British informants (in the Thames valley region?) reached them by a land route.

¹⁴ See Haverfield, , Mélanges Boissier, pp. 249–53.Google Scholar His archaeological evidence, though not conclusive in itself, is supported by Pliny, xxxiv. 16 (47), § 156: fabulose narratum (plumbum album) in insulas Atlantici maris peti … nunc certum est in Lusitania gigni et in Gallaecia. Pliny could hardly have ignored the British supplies as he here does, had they been forthcoming in his own day.

Presumably the lode tin of Cornwall was beaten by the alluvial tin of Galicia. This, being some way inland, may not have been discovered till the campaigns of P. Crassus (95 B.C.) or Augustus' Cantabrian Wars (24–19 B.C.). In any case, it required a large hydraulic apparatus such as probably none but a Roman engineer would have troubled to instai (Borlase, loc. cit.). But once put into operation, these Spanish stream-mines must have been more lucrative than lode mines.

¹⁵ Ridgeway, (Folklore, 1890, p. 91)Google Scholar assumes a traffic from Spain to Bordeaux and Narbo, but adduces no clear evidence.

¹⁶ Avienus, ll. 113–14: Tartessiisque in terminos Oestrymnidum negotiandi mos erat; Strabo, pp. 175–6. Strabo escapes his own notice in proving that the ‘Cassiterides’ in this context are British land.

¹⁷ Strabo, p. 147. This passage will be discussed later.

¹⁸ Déchelette, ii. 393; Schuchardt, , Alteuropa, pp. 57–8Google Scholar; Schulten, , Tartessos, pp. 9–15Google Scholar; Peake, , The Bronze Age and the Celtic World, pp. 41–5Google Scholar; Haverfield in Pauly-Wissowa, s.v. Κασσιτερίδες.

¹⁹ On the importance of excavation at this Cnossus of the west, and on its probable situation, see Schulten, , Tartessos, pp. 81–90.Google Scholar

²⁰ Herodotus, iv. 152. For the date, see Busolt, , Griechische Geschichte, i. p. 482Google Scholar, n. 3. No doubt the Greeks had heard of the Straits long before 630 B.C., and their knowledge of them may even have been quite accurate, as M. Bérard has contended. But it does not follow that they could have found their way to them. Similarly the authors of the Sindbad story possessed some accurate information concerning Ceylon and Madagascar, but could not have drawn a chart of the Indian Ocean.

²¹ Herodotus, i. 163.

²² vii. 56 (57), § 197.

²³ L'Anthropologie, 1899, pp. 401–6.

²⁴ Hyginus, Fab. 274: Midas rex, Cybeles fflius, Phryx, plumbum album et nigrum primus invenit.

Cassiodorus, , Variarum, iii. 51Google Scholar: aes enim Ionos, Thessaliae rex, plumbum Midas regnator Phrygiae reppererunt.

²⁵ It is significant that both the Phrygian and the Galatian Ancyra struck coins with an anchor-type (Catalogue, B.M., Galatia, p. 9Google Scholar; Phrygia, p. xxxi.).

²⁶ Else how could, e.g., the Lacedaemonians appear on the list?

²⁷ Schulten, (Tartessos, pp. 25–6)Google Scholar points out that the cognate form Μειδόκριτος is found elsewhere. A name formed from Midas would be nothing unlikely in a Phocaean.

²⁸ So Schulten.

²⁹ Did the Carthaginians show the way to Midacritus or vice versa? There is no evidence of the Phoenicians having sailed out into the Atlantic before the time of Himilco (Avienus, 1. 117), whose probable date is c. 500 B.C. (See esp. Schulten, ch. iii., on the relations between the Phoenicians and Tartessus.) Midacritus' cruise probably belongs to the later rather than the earlier half of the sixth century, but it cannot have been later than 500 B.C. Thus it is quite possible that Midacritus explored the Atlantic for the benefit of the Carthaginians.

³⁰ Hence Herodotus' agnosticism about the Cassiterides (iii. 115), which finds its parallel in Polybius' diatribes against Pytheas. These cases, like that of the Norsemen in America, or of Torres in Australian waters, show that an isolated cruise, however successful, may not make any impression on the map.

³¹ The Massiliote voyager Euthymenes, who was probably Herodotus' informant, or rather misinformant, about West Africa (Jacoby, in Pauly-Wissowa, s.v. Euthymenes) should be placed in the sixth century rather than in the fifth.

³² As has often been pointed out, Pytheas' cruise was unknown to Aristotle (d. 322 B.C.) and known to Dicaearchus (d. 285 B.C.). Probably it was made possible by Alexander's preparations for a campaign against Carthage, which gave the Carthaginians some more urgent work than to picket the Spanish Main. The date of his cruise would thus be about 325–3 B.C.

³³ Diodorus, v. 22. On this passage see p. 174.

³⁴ So Rice Holmes, pp. 220, 507; Berger, , Geschichte der Erdkunde,² pp. 353–8Google Scholar; Blazquez, Pyteas de Marsella.

³⁵ Strabo, p. 104 (on the authority of Polybius). Polybius' opinions on Pytheas are generally worthless, but he would hardly have invented a detail like this.

³⁶ Cf. Eudoxus' successful appeal for contributions to meet the expenses of his cruise round Africa (Strabo, p. 100). See also Damsté, , Mnemosyne, 1917, pp. 181–5.Google Scholar

³⁷ Strabo, pp. 175–6. If, as seems most probable, this passage has been taken bodily from Poseidonius, the explorer P. Crassus must be the consul of 97 B.C., not the legatus of Julius Caesar (Schulten, p. 49 and n. 3).

³⁸ Hatzfeld, , Les trafiquants italiens dans l'Orient hellénique, pp. 238–56.Google Scholar

³⁹ Ibid., loc. cit.; Frank, , Economic History of Rome, p. 261.Google Scholar

⁴⁰ Frank, loc. cit.

⁴¹ Ibid., pp. 261–5.

⁴² C.I.L. ii. 1724–1922.

⁴³ Haverfield, , Mélanges Boissier, pp. 252–3.Google Scholar

⁴⁴ L'Anthropologie, 1899, p. 401.

⁴⁵ Schuchhardt, Alteuropa, passim; Déehelette, ii. 393.

⁴⁶ Déehelette, i. 626.

⁴⁷ M. Reinach himself admits the greater expensiveness of land transport (p. 400).

⁴⁸ See the map at the end of Déehelette, Vol. ii. Pt. 1.

⁴⁹ L'Anthropologie, 1908, p. 153 sqq.

⁵⁰ Strabo, p. 654.

⁵¹ Rhoda, like Emporiae, probably dealt in Spanish silver and iron. (Jullian, , Histoire de la Gaule, i. 412.Google Scholar)

⁵² Déchelette, ii. 1006.

⁵³ Deutsche Altertumskunde, i. 223.

⁵⁴ LI. 148–151.

⁵⁵ So Schulten in his edition of Avienus, , and in Tartessos, p. 39.Google Scholar

⁵⁶ Avienus' estimate of time squares well with Poseidonius' estimate of distance (less than 3000 stades from Narbo to Ocean: Strabo, p. 188). Possibly Poseidonius was Avienus' ultimate source in this instance.

⁵⁷ Until 500 B.C. the main source of amber seems to have been Jutland (Dèchelette, ii. 873). From here to Cornwall or Brittany is a far cry.

⁵⁸ Histoire de la Gaule, i. 410–11.

⁵⁹ Tartessos, pp. 49–50.

⁶⁰ Arqueologia de España, p. 7.

⁶¹ In Scylax, chs. iii.–iv. (c. 350 B.C.) we find no mention of Celts in S. France. Provence is Ligurian country, Languedoc is shared by Ligurians and Iberians. In Hecataeus, Narbo, curiously enough, is called a Celtic city (fr. 19). But the remaining places on the French coast are called Ligurian (frs. 20–24). Possibly the of fr. 19 is a blunder by the excerptor Stephanus.

⁶² See Clerc, , Aquae Sextiae', Pt. 1Google Scholar, and Jullian, in Journal des Savants, 1922, pp. 99–113.Google Scholar Jullian points out that Arles, the key of the Rhône valley, did not come securely into Massiliote hands until the second century.

⁶³ Déchelette, ii. pp. 582–3.

⁶⁴ Head, (Historia Numorum, p. 6)Google Scholar attributes the earliest issues to 450 B.C. Déchelette (ii. 1562) dates back the earliest Massiliote hoard (No. 14 in Blanchet's list) beyond 450 B.C.

⁶⁵ Blanchet, , Revue Numismatique Belge, 1913, p. 292Google Scholar, and the list of finds on p. 308 sqq.

⁶⁶ See the numerous instances in Blanchet, Traité des Monnaies Gauloises, or Hucher, L'art Gaulois.

⁶⁷ Folklore, 1890, p. 107. According to Sir William Ridgeway the traffic previous to Pytheas was in Gaulish, not in Greek hands.

⁶⁸ Ancient Britain, p. 507.

⁶⁹ Op. cit. p. 353.

⁷⁰ This passage has caused much trouble, owing to the words ‘introrsum sex dierum navigatione,’ which defy all rational explanation. Presumably ‘introrsum’ is a translation of i.e. nearer the Mediterranean or Could this be due to the persistent notion that the Tin Islands, being reached from Spain, were near Spain? In any case, Pliny's Mictis or Ictis may be safely identified with Diodorus' tin depot.

⁷¹ Rice Holmes, loc. cit.; Müllenhoff, i. p. 472.

⁷² Blanchet, , Rev. Num. Belge, 1913, pp. 320–22.Google Scholar

⁷³ It is generally admitted that Timaeus' authority can have been none other than Pytheas.

⁷⁴ On this point see esp. Unger, , Rheinisches Museum, 1883, p. 164.Google Scholar

⁷⁵ Head (loc. cit.) gives some striking differences of weight between the earlier and the later issues.

⁷⁶ Déchelette, loc. cit.

⁷⁷ Celtic Britain (3rd ed.), p. 7.

⁷⁸ Timaeus (fr. 36, ed. C. Müller) attributed the tides of the Atlantic to the inflow of the French rivers. His knowledge of these was almost certainly derived from Pytheas.

⁷⁹ The incredulity of the Greeks in regard to Pytheas may to us seem incredible. But the same misplaced scepticism was shown by their contemporaries to Marco Polo and to Sebastian Cabot. So long as there existed no authoritative tribunal like the Royal Geographical Society to test the claims of explorers, it was equally possible for an impostor to gain credence and for an honest man to be turned down.

⁸⁰ Déchelette, ii. pp. 584, 875, 999, 1450, 1577.

⁸¹ The copiousness of these issues is probably due to the fact that Emporiae lay closest of Greek towns to the Spanish silver mines.

⁸² The Tarentine coins, and possibly some of the ‘Philips,’ entered Gaul by way of the Adriatic and Po valley (Déchelette, ii. p. 1570). But the chief distributing agent was Massilia (Blanchet, pp. 305–6).

⁸³ For the dates of the Rhoda and Emporiae pieces, see Head, p. 10. Head assigns the Massilian bronze to the second and first centuries. Blanchet holds that it was issued copiously towards the end of the third century and was first struck in imitation of the bronze coins of Syracuse (317 B.C. sqq.).

⁸⁴ Strabo, p. 147. As we have seen (p. 174), Poseidonius is probably Diodorus' authority for saying that the transit took thirty days. Hence we cannot follow Blazquez, (Las Casiterides, p. 55)Google Scholar in ascribing this time schedule to Julius Caesar.

⁸⁵ Strabo, p. 190.

⁸⁶ Blanchet, p. 301.

⁸⁷ The Seine route is advocated by Déehelette (ii. p. 584), Jullian (i. p. 410) and Frank (p. 254); the Loire route by Rice Holmes (p. 507); the Garonne route by Haverfield, (Proceedings of the Society of Antiquaries, Vol. xviii, p. 119).Google Scholar Siret (p. 159) votes for the Loire and Garonne, Ridgeway (p. 107) and Schulten (p. 50) for the Loire and Seine, Besnier (Daremberg-Saglio, s.v. stannum, col. 1461) for all three.

⁸⁸ If the Ictis of Diodorus and Pliny was St. Michael's Mount near Penzance, as Dr. Rice Holmes has made practically certain (pp. 499–507), a port on the Loire would be the natural receiving-point for the tin on the French coast.

Jullian (p. 410, n. 5) and Besnier (loc. cit.) suggest that Narbo drew its tin from the more northerly routes via Aries. But any tin that went down the Rhône valley to Aries would hardly take the long road from that point to Narbo instead of the short cut to Massilia.

⁸⁹ Blanchet, pp. 318–19.

⁹⁰ Déchelette, ii. 584.

⁹¹ Corot, , Bulletin Monumental, 1901, p. 572.Google Scholar

⁹² Besnier, , Revue Archéologique, 1921, p. 65.Google Scholar

⁹³ Bunbury, , History of Ancient Geography, ii. pp. 235–6.Google Scholar

⁹⁴ Blanchet, pp. 319–22.

⁹⁵ The Narbo route may have passed into Italian hands after the establishment of the Roman colony at that point (120 B.C.). But it was probably originated by Greek traders, for the merchants whom Scipio interviewed at Narbo were clearly not Romans, and the Massiliotes had for centuries had a small colony named near the mouth of the Aude.