¹ Cf. Akurgal, , Bayrakli 54 f.Google Scholar

² Owing to the sharp fall of the rock eastward at this point it was impossible, without opening an additional trench to a great depth, to ascertain whether there is a yet earlier stratum of occupation on the site.

³ On the north part of the site occasional second-millennium sherds were found in classical levels, perhaps from the spoil of wells; but none of the third millennium came to light here. The sounding in Square Jxviii–xix was carried to a depth of nearly 3 metres below modern sea-level without reaching the bottom of the habitation on the spot.

⁴ Ancient Smyrna (1938). It is appropriate here to record the indebtedness of the excavating party to this book, which is distinguished alike by profound scholarship and conscientious accuracy.

⁵ xiv. 634. Cadoux, op. cit. 58 n. 3, does not consider the location in Strabo to be specific, and it is possible that Strabo was not using his words precisely.

⁶ Bittel has suggested that prehistoric settlement may underlie the presumed Metröon on the low hump of Tepecik at the edge of the modern city (Kleinasiatische Studien (1942) 175); but so far as I know this is not supported by actual discoveries, and I suspect that the hillock is formed of rock and not a settlement mound. A bronze thrusting sword, similar, as Miss D. H. F. Gray tells me, to those current in the earlier phases of the Mycenaean (L.H. phases I–IIIA), was found in 1942 in the fill of the Roman agora at Smyrna, (AA 1943, 203, fig. 3Google Scholar; Naumann and Seláhattin Kantar, , Belleten vii. 213–25).Google Scholar Bittel's conjecture that it came from a grave in the vicinity seems just (Kleinasiat. Studien 175); but by itself it does not constitute evidence of a settlement at this point in prehistoric times. Bittel was of course writing before the discovery of the prehistoric site at Old Smyrna.

⁷ Nicolas, of Damascus, , Jacoby, FGH 90, F10.Google Scholar

⁸ Cf. Cadoux, op. cit. 33 ff.; Bittel, Die Reliefs am Karabel (Archiv für Orientforschung xiii); Cook, J. M., Türk Arkeoloji Dergisi vi–2, 3 ff.Google Scholar

⁹ I am indebted to Mr. V. R. d' A. Desborough for an opinion based on examination of sample bags.

¹⁰ Bericht VI. Int. Kongr. 327 ff.

¹¹ For this material cf. Larisa iii. 169 f. The identification of this site at Buruncuk is by no means secure, and the absence of early occupation there, though ascertained by extensive excavation, cannot be safely used to support a general lowering of the date of the primary Greek settlements on the coast. Similarly the lack of Protogeometric at sites like Ephesus and Priene carries little weight when no systematic attempts have yet been made to locate the earliest Greek settlement there; at Ephesus and Priene these may lie concealed under river silt (for Priene see Hiller, Inschr. Priene p. iv). Old Smyrna in fact shows most clearly how deceptive appearances can be; even after the Miltners had probed the site in a number of places in 1930 the existence of rich Protogeometric deposits was still unsuspected. This elusiveness of the Protogeometric and earlier Geometric is probably to be explained in some degree by the small size of the original Ionic foundations, but much more by the unsubstantial character of the settlements in their early stages. Sealed by the accumulation of successive deposits, the lowest Greek strata were already too thoroughly covered over to be disturbed by the large-scale constructional work of archaic times, so that for the most part the earliest sherds encountered in the uppermost levels on Ionic sites are of the Late Geometric. For Larisa see below, p. 20 n. 47.

The number of Protogeometric sherds noted in Samos and Chios is very small (Desborough, Protogeometric Pottery 215 ff.); but the excavations were at sanctuaries, not settlement sites. At the foot of Mycale Protogeometric vases have been recovered from graves at Canli (Desborough, op. cit. 221), perhaps belonging to the city of Melia that perished before the end of the dark ages; and at a few hours' walk north of this, strata of occupation going back apparently to Protogeometric are exposed by the sea on the little peninsula site at Kusadasi (p. 20 n. 46), which may be identified with Pygela. Though the time has not yet come to forget Wilamowitz's warning that the early Greek finds at Miletus did not justify the assumption of an equally early date for other foundations on this coast (SBBerl. 1906, 77 = Kl. Schriften v. i. 173), there is no good reason for lowering the date of the Ionic migrations to advanced Geometric times.

¹² A few late sherds, probably of Hellenistic date, have been picked up in the low-lying pasture immediately to the south-west of this road.

¹³ In a deep sounding in the trench in Square Jxviii–xix occasional prehistoric potsherds, apparently not waterworn, were still being recovered at the bottom of the shaft 2·70 m. below modern mean sea-level when work was abandoned. Well-preserved fig-wood poles were, however, found at 1·80 m. below sea-level, indicating that at this level the ground had been sufficiently damp to preserve the wood; thus, allowing a mean ground water-level in accordance with our observations at something over one metre above sea-level, the sea-level of the beginning of the second millennium B.C. should not have lain more than 3 metres below the modern sea-level. For sea-level in classical times the best evidence is provided by the stone-lined well in Square Fxii and the Fountain House (Square Lxxiii). In the well the pottery at 0·10 m. below modern sea-level was stained with green slime; from 0·20 m. below modern sea-level downward fragments of wine-jars preserved their original washed surface, and much of the pottery was bleached as though it had been constantly in wet ground since the well was abandoned in the fourth century B.C.; resin was preserved in the bottom of wine-jars here. The wooden raft supporting the stone shaft lay about half a metre below this. Allowing a full metre for the height of ground water above sea-level at this point we can suggest a minimum figure of 1·20 m. for the rise in sea-level. In the Fountain House the outlet pipe and water mark in the draw basin are at 0·76 m. below modern sea-level: presumably late summer ground-water level in classical times was slightly higher than this (perhaps c. 0·70 m. below modern sea-level), and since ground water lies lower at this end of the site (minimum 0·30–0·40 m. above modern sea-level) the classical sea-level is estimated to have been not much over 1 metre below the modern. Since observations at other points on the edges of the site conform to this estimate, a figure of between 1 and 1½ metres for the sinking of the coast here since the fourth century B.C. should not be wide of the mark.

In a series of papers published half a century ago Ph. Negris made a very strong case for a current phase of marine transgression in the Greek seas, and on a basis of submerged building traces and constructional levels he concluded that the land has sunk as much as 3–3·50 m. since ancient times in its relation to the sea (AM xxix. 360 ff.). Published archaeological observations on the east shores of the Aegean do little more than confirm the fact of general submergence (e.g. Antissa in Lesbos, , BSA xxxii. 42Google Scholar; Chios, , BSA xli. 33Google Scholar, ILN 30. 1. 1954, p. 159; Samos, , AM lv. 34, lviii. 158 f.Google Scholar; Gulf of Syme, , Archaeologia xlix. 350Google Scholar, ‘probably several feet’, p. 355), and this can indeed be observed at a great many points on the west coast of Asia Minor (e.g. Cyme, Clazomenae, the Erythraea, Kuşadasi (Pygela), Myndus, Halicarnassus, Old Cnidus (Burgaz), Loryma). Von Gerkan has given a figure of 1·50 m. at Miletus, (Milet ii. 3. 5)Google Scholar, but it is open to question whether this (like Schweinfurth's estimate of 2 m. at Alexandria, (AM xxv. 273 n. 2)Google Scholar, Evans's at Khersonesos and Komo in Crete (upwards of 2 m., P. of M. i. 298; ii. 1. 87), and Negris's own of 2·10 m. from a mausoleum at the Telmessus, Lycian, Délos et la transgression actuelle (1907) 21)Google Scholar can be regarded as anything more than a minimum calculation of the submergence since ancient times. There is therefore an alternative possibility that ground-water level on the site at Old Smyrna has not maintained a constant relation to sea-level, and that the figures quoted above for sea-level in classical times are on the low side. It is at the same time worth remark that the calculations here offered suggest that the rate of submergence at Smyrna may have been rather more rapid in the second and early part of the first millennium B.C. than it has been since classical times, whereas at Delos Negris inferred that it was slower in pre-Hellenistic times (1 m. per millennium against a subsequent rate of 1·50 m.).

I have unfortunately failed to trace an article by von Gerkan entitled ‘Meereshöhen u. Hafenanlagen im Altertum’. Mr.Hammond, N. G. L. (JHS lxxvi. 35)Google Scholar quotes evidence for a rise of the sea by 5 to 6 feet in the Saronic Gulf and on the coast of the Epirus, and considers that the rise may be referred to a general alteration in the level of the oceans.

¹⁴ The second epithet ποντοτίνακτον printed in modern texts is Pierson's emendation of the MSS. ποτνιάνακτον. Neither ποτνιάνακτος nor ποντοτίνακτος is otherwise known. It seems clear that the composer of the epigram, as also the author of the Herodotean Life in which it was incorporated, was a native of the Southern Aeolis and likely to be familiar with the ground; so that it is relevant to object that the shallow waters of the innermost recess of the Hermeian Gulf could not fairly be described as πόντος, as the sea here is not open enough to allow Old Smyrna to be conceived as smitten by the main. From the excavations at the temenos on the site it appears that the principal deity of Old Smyrna was a goddess, probably an Artemis or Μήτηρ; the epithet ποτνιάνακτγος (‘ruled by the Lady’) would therefore be appropriate (for the passive use of the verb ἀνάσσω cf. Od. iv. 177).

¹⁵ The Phaeacian migration and settlement on Scheria is described in Od. vi. 3 ff.; the peninsular position with the fortification, the narrow causeway, the two anchorages with the agora, and the sanctuary in vi. 262; in vii. 39 ff. Odysseus descends to the town, admiring the anchorages, agoras, and wall circuit. For sites of this type cf. Thuc. i. 7.

¹⁶ xvii. 4. He made this city intermediate both in time and position between the shattered original foundation of Tantalus on Sipylus and the city of his own day. Pliny also refers to Tantalus' city of Sipylum and to Archaeopolis which succeeded it (NH v. 117); and the name Naulokhon, which Stephanus Byz. (s.v. Σμύρνα) gives as an earlier name of the city, seems most applicable to the position at Bayrakli. Cf. the discussion of the relevant testimonies and monuments in Cadoux, Ancient Smyrna 36 ff.

¹⁷ xiv. 646. It is of course possible that Strabo may have had in mind the fortified positions and monuments on the mountainside above Bayrakli (pp. 3, 5f.) rather than the peninsula site below. His measurement would then be too short unless its southern terminal was provided by a suburb well outside the limits of the city proper; but many of Strabo's measurements on this coast are demonstrably short.

¹⁸ Hdt. i. 150; Strabo xiv. 634, 646.

¹⁹ Hdt. i. 149 f.

²⁰ Strabo xiv. 634. Cf. pp. 27 f.. The phrase θεῶν βουλη̨̃ with which Mimnermus justifies the Colophonian capture, is answered by that applied in the Homeric Epigram iv to the original Cymaean foundation, ἥν ποτ᾿ ἐπύργωσαν βουλη̨̃ Διὸς αἰγιόχοιο λαοὶ Φρίκωνος: and there are other echoes of the Mimnermus passage in the epigram.

²¹ Strabo xiv. 634.

²² Cf. Cadoux, , Ancient Smyrna 48.Google Scholar

²³ i. 143, 3. Cf. Cadoux, op. cit. 67 f.

²⁴ For this see Keil, RE s.v. Melia (3); a new discussion of the problems by Roebuck, , CP 1955, 26 ff.Google Scholar

²⁵ Hdt. i. 4.

²⁶ Pausanias speaks of Gyges and the Lydians as holding the πόλις and being expelled from it (iv. 21, 5); but Hdt. i. 14 does not imply more than an invasion of the territory, and the anecdote related by Plutarch on the authority of Dositheus shows the Lydians as unable to enter the city (Parallela 30). Cf. Cadoux, , Ancient Smyrna 80 f.Google Scholar

²⁷ Comminuted potsherds, badly worn, but apparently covering approximately the whole range of the seventh century, have been picked up on the east part of the Bayrakli hill-side, and were recovered in considerable quantity at the foot of the mountain slope on the north during house-building operations in 1952; seventh-century pottery, including fine pieces of the late years of the century, was found in the sinking of two wells about 500 metres to the south-east of the peninsula site in 1951–2. See the plan, Plate 1, and drawing, Fig. 3.

^27a See my article on Greek bathtubs in Greece and Rome 1959.

²⁸ Arist. Pol. 1290b.

²⁹ The festival of Dionysus was held outside the city, in Aeolic times at least (Hdt. i. 150). There may also have been cults of the Nemeseis at the Pagos (Paus. vii. 5, 2), which became the acropolis of Alexander's new foundation, and of Artemis at Halka Pinar on the north-eastern out-skirts of the modern city (p. 31).

³⁰ Paus. v. 8, 7; Euseb. Chron. Ol. 23.

³¹ JHS lxxi. 248, fig. 8. Friis Johansen has now published a Rhodian kotyle, probably rather earlier in date, with a portrait of a lyre showing five strings (Exochi 122, 33 fig. 57); I have assumed that the Delos sherd with a lyre-playing scene (AM lvii, pl. 5.5.) is of a later date.

³² JHS i. 67 f.

³³ i. 149. The names Killa, Notion, and Aigiroessa come after Temnos and before Pitane in Herodotus' list; in so far as Herodotus' order is significant they should have been situated in the eastern or northern parts of what he treats as the Aeolis. Aigiroessa is not otherwise known (unless possibly the fourth-century coins of the Lesbian region with legend ΑΙΓΙ, BMC Troas p. lxxvii, are to be attributed to it), nor (if they are distinct from the towns which bear these names but lie outside the limits of the S. Aeolis) are Killa and Notion either.

³⁴ See Seylaz-Keil, , ÖJh xxviii Beibl. 121 f.Google Scholar (from the village below); Bean, , Jahrbuch für Kleinasiatische Forschung iii. 45.Google Scholar

³⁵ ΚλαЗομενίοις δὲ λοτρά ἐστιν· ἐν δὲ αὐτοῖς ᾿Αγαμέμνων ἔχει τιμάς(Paus. vii. 5, 12). πηγαὶ θερμαὶ ἐν ᾿ἰωνία, ἃς ἔτι καὶ νῦν ᾿Αγαμεμνονείους καλοῦσιν οἱ Σμύρναν οἰκοῦντες · ἀπέχουσι δὲ, οι̇̄μαι, τετταράκοντα στάδια τοῦ ἄστεος (Philostratus Her. 691). Cf. also Keil, , ÖJh xxiii. 86 f.Google Scholar There is in fact another spa (the Urla İçmeleri) west of Clazomenae at the head of the gulf of Hypokremnos, to which Pausanias might be thought to be alluding; and it is notable that Aelius Aristides seems to mention the cult of Asclepius, but not of Agamemnon, in speaking of the baths near Smyrna; cf. Cadoux, , Ancient Smyrna 17.Google Scholar It is therefore just possible that the title ‘Agamemnonian’ may have been assumed at the baths near Smyrna after the time of Pausanias and Aristides. But there is nothing inherently improbable in Clazomenian possession of the long coastal strip. The only point where the mountains descend to the shore so as to interpose a barrier between the plains of Smyrna and Clazomenae is east of Balçova on the western edge of the modern city of Smyrna. The old city of Smyrna lay some miles away from here; and if they were not already in possession of this riviera, the Clazomenians might well have advanced their frontier with Lydia to this point after repelling Alyattes.

³⁶ Bean, , Jahrbuch für kleinasiatische Forschung iii. 51 f.Google Scholar, where the situation of the Mormondas is fixed with some probability; cf. Kondoleon, , AM xiv. 93, xv. 337Google Scholar, Fontrier, , BCH xvi. 397.Google Scholar

³⁷ Ramsay, , JHS ii. 296 ff.Google Scholar Cf. Keil, , ÖJh xvi Beibl. 163 ff.Google Scholar, Robert, , Études anatoliennes 115.Google Scholar Keil conjecturally identifies Melampagos as the site at Gökkaya. Robert considers this Herakleia to be the one whose territory was ravaged by Prusias in his war with Attalus II, and that it was an independent city in 155 B.C.

³⁸ Cf. Keil, and Premerstein, , Denkschr. Akad. Wien lvii (1915) 6 ff.Google Scholar On the hill now occupied by the ruins of the great Byzantine fortress there are traces of an ancient citadel and outer circuit, apparently of equally great extent; some scraps of black glaze pottery picked up on the surface here go back at least to the third quarter of the fourth century B.C.

³⁹ JHS i. 63, 86; ii. 52. Herodotus' mention of the τύποι of Sesostris in connexion with the route from Sardis to Smyrna (ii. 106) proves that in the fifth century the road ran over the pass of Bel Kahve; for not only were the reliefs discovered in the Karabel a little way south of this road (cf. Cadoux, , Ancient Smyrna 33 ff.)Google Scholar, but the less direct route leading from Sardis along the north side of the Manisa Daği to the Hermus gorge would more naturally have been described as leading to Magnesia, or to Phocaea and Cyme. For more recent references, see p. 10 n. 8.

⁴⁰ C. 9, πορευόμενος δὲ διὰ τοῦ ῾᾿Ερμου πεδίου ἀπικνέεται ἐς Νέον τεῖχος. This seems to imply the possibility of his taking another road.

⁴¹ Cf. JHS ii. 286. Ramsay, discussed Aelius Aristides' journey from Smyrna to Pergamon at length in JHS ii. 44 ff.Google Scholar, and used it in building up his system of identifications of ancient sites in the Southern Aeolis. The discovery of a milestone, measured from Ephesus, at Bornova (ibid. 51) perhaps supports the case for a mountain road; the figure (7 or 8 miles) on the milestone said to have come from Menemen (ibid. 52 f.), which has caused serious difficulty, can then be more satisfactorily explained as giving the distance from the main road at the crossing of the Hermus by Emiralem. An incomplete inscription recording repairs under Trajan was found near Ulucak to the south of Menemen, (BCH xvi. 403 n. 2).Google Scholar

⁴² There may also have been a village at Bel Kahve by this time (pp. 4, 17 f.).

⁴³ The Gergithes of the Milesia may have been serfs (Heracleides ap. Athen. 524). The Pedieis mentioned in inscriptions of Priene are, rightly or wrongly, commonly understood in this sense (cf. RE, Pedieis (4)); and the πύργοι of Teos were interpreted by Ed. Meyer and Wilamowitz as the seats of a landed aristocracy (Ruge, RE s.v. Teos 556, cf. Hunt, , JHS lxvii. 70 f., 75).Google Scholar At Colophon also a suppressed class is suggested by Mimnermus' boast βίην ὑπέροπλον ἔχοντες ἑЗόμεθ ἀργαλέης ὕβριος ἡγεμόνες (fr. 9 Bergk, 12 Diehl). Wilamowitz' suggestion that the crime of the people of Melia, which occasioned their destruction, was the raising of the native population is more conjectural and hardly fits with Vitruvius' phrase ‘propter civium adrogantiam’ (SBBerl. 1906, 78 = Kl. Schriften v. i. 174).

⁴⁴ Fr. 3 Bergk, 3 Diehl.

⁴⁵ Schuchhardt, in Pergamon i. 1, 98.Google Scholar

⁴⁶ Keil, ÖJh. xi. Beibl. 146 ff., fig. 97. Cf. ATL i. 515 (Marathesion, but admitting the possibility of an identification with Pygela), JHS lxxiii. 125. The main strata my wife and I noted on the site are Byzantine (on the beach south-east of the isthmus), classical (on the north at the landward end of the isthmus), Geometric (on the northeast side of the peninsula), and prehistoric (cf. Mellaart, Anat. Studies vii. 83) on top of the peninsula. This sequence, taken in conjunction with the classical and medieval literary testimonia, leaves no doubt that this site is Pygela, which had an Amazonian tradition and its own foundation-legend, was independent between the time of the Athenian empire and the Hellenistic era, and was evidently a place of some account in Byzantine times. In Pliny, , NH v. 114Google Scholar the ‘Phygela fuit’ of our texts should be corrected to ‘Phygela, fuit et Marathesium oppidum’ (cf. my remarks on this section of Pliny in CQ 1959).

This site deserves attention. Its history is written on its sea-washed banks and could be read in detail with no more equipment than a ladder and pick.

⁴⁷ Boehlau-Schefold, Larisa i. Whether or not the site is Larisa, its access of prosperity in later archaic times must have been due to the Egyptian mercenaries (presumably of Greek or Greco-Carian origin) settled by Cyrus the Great in the Hermus Plain (Xen. Cyrop. vii. i. 45). If not Larisa, this site should be Cyllene.

The identification of Buruncuk with Larisa goes back to Ramsay (see above, p. 19 n. 41). It depends on (a) the distance (70 stades) given by Strabo (xiii. 622) as separating Larisa from Cyme, (b) the assumption that the Hermus crossing lay south of Menemen (as it did at the time of Ramsay's journey, before the diversion of the river in 1886), and that the road taken by Aelius Aristides led from there across the plain directly to Buruncuk. It is true that (a) favours the identification with Buruncuk, but it is not cogent since Strabo's distances on this coast are habitually underestimated. But (b) depends on the view, which is almost certainly false, that the Hermus in antiquity flowed in its pre-1886 bed (southward from Menemen) and that it was forded in the Hermus plain near the modern Ulucak (see above, p. 19); in fact the ‘new’ 1886 bed seems to have been an older one (cf. von Diest, , Pet. Mitt., Erg. 94 (1889) 35)Google Scholar; in all probability it was the ancient one.

The objections to placing Larisa at Buruncuk are (i) that if the Hermus ford (as I suppose it to have been) was at Emiralem, the Yanikköy site (commonly called ‘Neon Teichos’) would fit better with Aristides' data for Larisa—and conversely, if the ford was in the plain in the vicinity of the modern girder bridge, the Buruncuk site would be too near to the ford; (ii) I think that the surface archaeological evidence suggests that at the beginning of the fourth century B.C.—the time when Thibron spent much time in unsuccessfully besieging Larisa—the most powerful and extensive site in the Hermus plain was not that at Buruncuk, but that at Yanikköy; (iii) if the site at Buruncuk, which dominates the passage from Cyme sleeve into the Hermus plain, was that of Larisa (which was evidently a place of some consequence in the early fourth century B.C.), it seems extraordinary that it was Cyme (and not Larisa) that disputed with Clazomenae the possession of Leucae near the Hermus mouth after Tachos' death c. 382 B.C. (Diod. xv. 18; cf. Judeich, , Kleinas. Studien 191)Google Scholar; (iv) in the Vita Herodotea, which is evidently written by a person familiar with the S. Aeolis, the blind Homer is made to walk from Neon Teichos to Cyme διὰ Λαρίσης τὴν πορείαν ποιησάμενος· η̇̃ν γὰρ οὕτως αὐτῷ εὐπορώτατον (c. 11): this clearly implies that for a traveller who was not blind there was a recognized alternative route; but from Yanikköy the way to Cyme must inevitably lead past Buruncuk, and it would be absurd to think of an alternative route; (v) Larisa is recorded as one of the original Aeolic foundations on this coast, but despite extensive excavation the Buruncuk site has yielded no early Greek deposit (see above, p. 11 n. 11).

These arguments are of unequal value, but (iv) and (v) seem to me cogent. If the coins could be used as evidence that this Larisa remained independent in the fourth century B.C., (iii) would be equally cogent; and it could also be objected that if Buruncuk is Larisa, it is extraordinary that fourth-century coins of Larisa did not come to light in the excavations there. But since the publication of Larisa iii I have been very dubious whether this Larisa (the Phrikonid one) can ever have struck coins; and, exhaustive though it is, Robert's, L. recent discussion in his Études de numismatique grecque 36–68Google Scholar does not convince me that the bull coins of Larisa are to be assigned to the Phrikonid Larisa. In 1902 Boehlau, who was then excavating Buruncuk, seems to have noted a peculiar bronze coin of this series, but it is significant that that most conscientious of archaeologists did not enter it among the finds from the excavations (cf. Ét. num. gr. 66; it is tempting to suggest that what Boehlau jotted in his note-book was a coin shown him by a dealer). Otherwise the identification of this series rests solely on the assumption that the almost identical bull coins with legend Βοιωνιτικόν are to be attributed to an otherwise unknown city (universally called ‘Boeone’) in the S. Aeolis; and this assumption rests on a report that coins of this place were found in the Hermus valley (westward toward Cyme? or eastward towards Sardis?); cf. Ét. num. gr. 49 f., with references. This evidence is difficult to assess. Against it must be set the general improbability that there was an independent city of this name in the S. Aeolis near Larisa in the fourth century (for we know the names of all the old Aeolic cities here and have a considerable range of other testimonia for the region) and the fact that no coins of this ‘Boeone’ came to light in the excavations at Buruncuk; though not decisive (as the analogy of the Tisnaean coin shows), this objection is weighty. It is worth recalling that in later times there were two κατοικίαι close to one another in the Cayster valley up-country from Ephesus: the well-known Ephesian Larisa (at Güzelim Tepe) and the Βωνῖται (at Büyük Kadife, , Keil, and Premerstein, , Dritte Reise 99 ff.Google Scholar, cf. Robert, L., Études anatoliennes 102 n. 6Google Scholar; Robert draws a comparison with the name Βωνιτώ in the cadastral survey of Magnesia ad Maeandrum, Inschr. v. Magn. no. 122 e, l. 11; cf. also Rev. Phil. 1958, 55 n. 2). Perhaps, pending more precise evidence of finds, it might be as well to consider these two places as alternative sources of the bull issues, and not to base any arguments on the existence of a fourth-century coinage of Larisa Phrikonis.

⁴⁸ Schuchhardt, in Pergamon i. 1, 103 f.Google Scholar

⁴⁹ Unfortunately there is no indication at Cyme of the distance to which the crest was occupied in the inland direction.

⁵⁰ Miletus, , Clazomenae (mainland site, AE 1953–1954, 152 f.)Google Scholar, and Phocaea were certainly capable of a considerably greater extent, and the limits of Polycrates' city of Samos were much greater (AM ix, pl. 7). On the other hand Eresos and the proasteion of Pyrrha in Lesbos (Koldewey, , Antike Bauresten der I. Lesbos 22 ff.Google Scholar, 27 ff.) give the appearance of being no larger, even in classical times, than the peninsula site of Smyrna, while the towns of the Lelegian octapolis (BSA 1. 116 ff.) were distinctly smaller.

⁵¹ Named by Lady Beazley (who with her husband discovered the site) ‘Donkey Point’ on account of the profusion of skeletons of donkeys there; on the Turkish maps ‘Kalem Burun’.

⁵² This problem, being more complicated than that of Larisa Phrikonis, would require a more detailed argument than is in place here; but I take this opportunity of drawing attention to a known inscription of Erythrae which appears to be an ordinance regulating the laying-out in the mid-fourth century of a new system of public streets and lanes and should therefore relate to a new foundation or the complete remodelling of an old one (Smyrna Mus. no. 1032, said to have been found at the Agora of Smyrna, but in fact previously published as found at Ildir, , BCH viii. 346 ff.Google Scholar; Bechtel, , Ion. Inschr. pl. 4/5Google Scholar; SGDI 5690; Wilamowitz, , Nordion. Steine 28 f.).Google Scholar

⁵³ ATL i. 446 f., 486.

⁵⁴ i. 143, ἀσθενέος δὲ ἐόντος τοῦ παντὸς τότε ῾Ελληνικοῦ γένεος πολλῷ δὴ η̇̃ν ἀσθενέστατον τῶν ἐθνέων τὸ ᾿Ιωνικόν.

⁵⁵ Roebuck, (CP 1953, 12)Google Scholar has recently made an acute attempt to calculate the free population of some Ionic states at the beginning of the fifth century on a basis of the muster of triremes at Lade, and suggests Chios c. 80,000, Miletus c. 64,000, Samos c. 48,000, as also Lesbos c. 56,000; if extended to other mainland Ionic cities this reasoning would give Myous c. 2,400, Priene c. 9,600, Teos c. 13,600, Erythrae c. 6,400, Phocaea c. 2,400. Ephesus, Colophon, Lebedus and Clazomenae (the last-named already overrun by the Persians) are not included in the list. These figures are naturally defective in that cities with no maritime tradition are underestimated—Erythrae, for instance, must have had a much bigger population; the Athenian tribute lists, on a basis of perhaps 3,000–3,500 people per talent, offer a useful countercheck. But the total figure for the Ionic muster may well be an accurate guide within its geographical limits; Erythraeans and Phocaeans, for instance, may have helped to man the Chian fleet. Assuming that Colophon and Ephesus had by this time been outstripped by Miletus, the total free Greek population of mainland Ionia in 494 B.C. might be reckoned, in accordance with Roebuck's reasoning, at around a quarter of a million—a figure that would not be out of keeping with that of 200,000 in the seventh century.

⁵⁶ Cf. Cadoux, , Ancient Smyrna 73, 209 ff.Google Scholar

⁵⁷ This topic is discussed, with a bibliography, by Cadoux, , Ancient Smyrna 10 ff.Google Scholar

⁵⁸ Conon ap. Phot. Bibl., Jacoby, FGH 26Google Scholar, F1, p. 207.

⁵⁹ Hymn. Hom. ix. 3–5.

⁶⁰ As Prokesch and others have done, cf. Cadoux, op. cit. 13.

⁶¹ Paus. vii. 5, 3, οἳ πάγον οἰκήσουσι πέρην ἱεροǐο Μέλητος. (evidently as seen from Old Smyrna, not from Claros).

⁶² i. 16, 2.

⁶³ JHS lxxii. 107, fig. 12.

⁶⁴ ÖJh xxvii. Beibl. 167 f.

⁶⁵ Hdt i. 16, 2, προσταίσας μεγάλες.

⁶⁶ SBBerl. 1906, 52 = Kl. Schriften v. i. 145 n. 2. Wilamowitz' statement that Colophon was conquered by Gyges is questionable, since Herodotus says only that Gyges captured the ἄστυ. Polyaenus (vii. 2, 2) relates a stratagem by which Alyattes, having made an alliance with the Colophonians, overpowered a body of their cavalry; this at least seems to imply that Colophon was not permanently subjected by Gyges. Xenophanes fr. 3 (Bgk) also seems to imply that the Colophonians had until recently been free.

⁶⁷ Hdt. i. 25, 1, βασιλεύσας ἔτεα ἑπτὰ καὶ πεντήκοντα. Alternative dates for Alyattes' accession are 609 (Eusebius) and, apparently, 604 (Parian Chronicle).

⁶⁸ i. 18, 2.

⁶⁹ Now dated 612 B.C. by the Nabopolassar Chronicle (Gadd, The Fall of Nineveh). Alyattes' eagerness to break off hostilities with Miletus in that year and placate the Greeks on his western frontier would then be explained by a desire to be free to intervene in the east.

⁷⁰ As Payne himself recognized (NC 57). See Hopper, , BSA xliv. 169–84Google Scholar; also a new discussion of these problems by Dunbabin in the second memorial volume for Professor Oikonomos (AE 1953–4), which by the kindness of the author I have been able to read in advance of the publication.

⁷¹ In addition to Hopper's and Dunbabin's discussions (n. 70) the problems associated with the Corinthian pottery at Smyrna are considered by Mr. Anderson in the section of this report devoted to the Corinthian pottery (below, p. 148). Comparison of the East Greek pottery from the two sites gives a similar result.

The comparison between the pottery of the destruction level at Smyrna and the earliest pottery from Selinus is now set on a different footing by Vallet and Villard's discovery of unpublished vases from the Malophoros sanctuary (BCH lxxxii. 16 ff.). Vallet and Villard publish some fine Corinthian Transitional pieces, and some smaller vases which they consider of Late Protocorinthian date. On this evidence Payne's dating of Corinthian Transitional to 640–625 (and Early Corinthian 625–600) hardly agrees with the Thucydidean date of 629/8 for the foundation of Selinus. Vallet and Villard adhere to Payne's absolute chronology of Corinthian and prefer to follow Diodorus and Eusebius in placing the foundation of Selinus in 650; on this assumption a date towards the end of the seventh century is required or the capture of Smyrna by the Lydians. At the same time I feel some uncertainty whether any of the Selinus pottery which has been published up to date is earlier than Transitional, and the capture of Smyrna seems to have occurred before the end of the Early Corinthian phase; so that thirty years could be a fair estimate of the interval separating the capture of Smyrna from the foundation of Selinus. Thus, if the Thucydidean date for Selinus were retained and the dating of Transitional and Early Corinthian were lowered slightly to conform to the new evidence, the capture of Smyrna could well be placed about the turn of the century, or not much after it.

⁷² Von Olfers, , Abh. Berl. Akad. 1858, 539 ff.Google Scholar The tumulus can be identified with certainty as that described by Herodotus (i. 93) on account of its size.

⁷³ Bowra, , Early Greek Elegists 17.Google Scholar

⁷⁴ xiv. 634, from ‘Nanno’ (Fr. 9 Bergk, 12 Diehl). In their editions Kramer, Meineke, and Müller all affirm their belief that the whole passage is an interpolation in Strabo; and if this were so less reliance should be placed in the context. But the publication of the Vatican Palimpsest has shown that the mid-nineteenth-century editors of Strabo were altogether too audacious in hunting for interpolations of this sort in Strabo's text (see Aly, , De Strabonis Codice Rescripto 240 f.).Google Scholar

The contexts of the two fragments about Helios (from ‘Nanno’) and Jason are not clear; but in each passage the opening words suggest that the couplets preserved are not part of a longer continuous narrative but self-contained examples of courage and endurance.

⁷⁵ Wyss, Antimachi Col. Reliquiae fr. 180.

⁷⁶ ix. 29, 4. Pausanias also pictures Aristomenes as exhorting the people of Messene to emulate the exploits of the Smyrnaeans, who expelled Gyges and the Lydians from their city (iv. 21, 5); and no doubt here too he had the elegies of Mimnermus in mind. If this story is to be credited, it may be argued that the final issue of the Second Messenian War was subsequent to Alyattes' attack on Smyrna and that Tyrtaeus imitated Mimnermus rather than the reverse. But the reference to Smyrna here is not likely to depend on any genuine historical tradition.

⁷⁷ iv. 370 ff. Cf. Jacoby, , Hermes lxiii. 288.Google Scholar The analogy could be extended to show that Mimnermus' exhortation does not necessarily imply any lack of valour on the part of the Smyrnaeans.

⁷⁸ This is of course disputed: see the most recent discussion by Barigazzi, A., Hermes lxxxiv. 162 ff.Google Scholar For the Callimachus passage see Ox. Pap. xvii. 2079, 11–12; CR 1929, 214.

⁷⁹ Fr. 9 Bergk, 12 Diehl. Cf. Jacoby, , Hermes, liii. 268 ff.Google Scholar

⁸⁰ If Gyges' attack on Smyrna (Hdt. i. 14, 4, ἐπείτε ἡ̑ρξε) occurred in the early part of his reign (i.e. in the early years of the seventh century), Mimnermus' statement that he had heard of the unnamed champion's prowess from eye-witnesses would fit with a date about or shortly after the middle of the seventh century for the poet's birth. That this is not precluded by Solon's poem to Mimnermus is shown by Jacoby, op. cit. 278 n. 2.

⁸¹ Comminuted seventh-century pottery, presumably drifted from the surface of the mound, was found in the make-up of all building stages down to the end of the occupation in the adjacent trench in Squares K–Lxi–xii, as also on the surface in the vicinity of the mound.

⁸² From the context it is clear that the sigloi are of Darius I. Besides two worn Lydian coins, the hoard contained fourteen diobols of a single type similar to the Phocaean, which were in fresh condition and presumably recently struck.

It is worth noting that Prokesch von Osten remarked that Darics were found at the site (together with staters of Phocaea, Clazomenae, and Cyme) and used them as evidence that Old Smyrna was not entirely defunct after Alyattes, (Jahrbücher der Literatur lxviii (1834)Google Scholar, Anzeige-Blatt, p. 59).

⁸³ Cf. Mr. Boardman's observation in the section on Attic wares, (below, p. 152).

⁸⁴ The course inferred for this wall is marked on the plan Plate 1. The ground here is for the most part hopelessly denuded, and no wall line can now be traced with certainty; but so far as it has been possible to fix the exact provenience of cist blocks and chance finds, the most westerly graves of the classical cemetery appear to lie close outside this assumed line, whereas corroded fourth-century pottery (mainly fragments of wine amphorae) is found in the denuded ground on the inside only of this line. For the description and discussion of this wall see Mr. Nicholls' account, below pp. 93, 135 f.

⁸⁵ Strabo may have been erroneously calculating from the time of Gyges' attack; but the excavations have borne out Herodotus' statement by showing that Smyrna was a flourishing city until the time of Alyattes. More probably the village period is meant to start with Alyattes; Head has also miscalculated this period by a hundred years (HN ² 592). It is, of course, possible that Strabo misread Τ′ as υ′ in a Greek source such as Artemidorus (for similar errors cf. Aly, De Strabonis codice rescripto 246 f.).

⁸⁶ Cf. Strabo xiv. 640.

⁸⁷ This is suggested by the presence of Lydian wares (marbled ware, lydions, &c.) and one or two Lydian graffiti among the sixth-century pottery, and a mid-sixth-century tumulus burial with offerings of a Lydian character.

⁸⁸ Cf. the marble lion's-head spout from Bornova (p. 4, Plate 3 b).

⁸⁵ Also the literary evidence for a sanctuary of the Nemeseis on the Pagos with statues of the Graces by Boupalos (cf. Cadoux, , Ancient Smyrna 89 n. 2).Google Scholar

As regards the capital reported to have been brought to the Smyrna Museum from Halka Pinar, there may be some uncertainty about its provenience if it is the same capital which Schefold remarks as being from Miletus, (Larisa i. 147 f.Google Scholar; cf. Martin, R., BCH lxviii–lxix. 361 f.).Google Scholar But remains of a temple were noted by Spon and by Prokesch at Halka Pinar. Mr. Boardman drew my attention in the Ashmolean Museum to a marble block c. 6 ft. long with late archaic egg and tongue moulding, which was brought from Smyrna by Hyde Clarke; unfortunately no more precise provenience is given.

⁹⁰ Hdt. v. 123.

⁹¹ On the evidence of the surface pottery, cf. my article, ‘The Topography of Clazomenae’, in the second memorial volume for Professor Oikonomos, (AE 1953–1954, 153 f.).Google Scholar

^91a To both Xenophon and Ps.-Scylax the Hermus Plain on the north side of the Gulf of Smyrna is inland from Cyme (ὑπὲρ Κύμης) In Xenophon this is quite clear (Hell. iii. 4, 27). In Ps.-Scylax (97c–98) the reading must be πόλις Μύρινα καί λιμὴν, Κύμη καὶ λιμὴν (ὑπὲρ δὲ Κύμης ἐν μεσογείᾳ πόλις ῾Ελληνίς ἐστιν Αἰγαὶ καἱ Λεῦκαι καὶ λιμένες καὶ Σμύρνα ἐνὴ̨̃ ῾᾿Ομηρος η̇̃ν), Φώκαια καὶ λιμὴν καὶ ῾´Ερμος ποταμός, ΚλαЗομεναὶ καὶ λιμήν, ᾿Ερυθραὶ καὶ λιμήν. . . . The editors, by ending the parenthesis at Αἰγαί have made Leucae and Smyrna places of the coast, presumably on the ground that since they lie on the sea they could not be spoken of as being in the interior behind Cyme. But Leucae and Smyrna are then out of place on the coast before Phocaea; and a study of the uses of καὶ and of asyndeton in Ps.-Scylax shows the accepted punctuation to be false. The periplus of the coast is resumed by asyndeton at the word Φώκαια, and Cyme, Phocaea, Clazomenae, Erythrae represents the correct sequence of coastal cities.

If the evidence of the tribute lists is taken into account, it seems probable that in the fifth century even the road-stead of Clazomenae lay outside the normal range of the Athenian marine, and that before 427 B.C. communication between the Athenians and Clazomenae may have been maintained through Airai; Clazomenae is exposed to the full force of the Etesians, and in the confined waters of the Hermeian Gulf, where there were no snug coves and the coasts were hostile, the danger to a lone trireme would have been grave. It is notable that the raising of the tribute of Clazomenae from a paltry one and a half talents to six talents in 427 B.C. (in itself a very remarkable increase) coincides with the activity of the squadron of the ἀργυρολόγοι on these coasts at the time of the Mitylenaean Revolt; no doubt the raising of the tribute was the response of the Clazomenian demos to the visit of the Athenian fleet. The smallness of the tribute before 427 B.C. and the subsequent bitterness of party politics there would fit with the assumption that Clazomenae, like Colophon, was a border city between the Persian and Athenian empires, and that political harmony there depended as much on good relations between the oligarchs and the Persians as upon adherence to the Athenian empire. But the question cannot be fully discussed here.

⁹² Cf. Hesperia xiii. 91 ff.

⁹³ Bürchner, RE s.v. Smyrna 739, refers to a Marburg dissertation, Weismantel, Die Erdbeben Kleinasiens; cf. also RE Suppl. iv. 353.

^93a See also my article on bathtubs in Greece and Rome 1959.

⁹⁴ Kondis, , Συμβολὴ εἰς τὴν μελέτην τῆς ῥυμοτομίας τῆς Ῥόδου (Rhodes 1954).Google Scholar

⁹⁵ A count made by Mr. D. T.-D. Clarke revealed that some 400 amphora toes were found in fourth-century levels in the trench Squares K–Lxi–xii in 1948 alone, and many more came to light when the excavation of these levels here was completed in 1949. The amphorae may of course have been re-used for other purposes after the wine was consumed.

⁹⁶ SIG ³ 136.

⁹⁷ A single position for the settlement that preceded the new foundation is implicit in Paus. vii. 5, 2 ((ἀναστήσαντα ἐκ τῆς προτέρος (sc. πόλεως cf. πέρην ἱεροῖο Μέλητος in the oracle ibid. 3, p. 23; cf. also the survival of the name πόλις ἀρχαία until Pausanias' day, ibid. 1).

⁹⁸ See n. 91a. It may be questioned whether Ps.-Scylax would have mentioned Smyrna at all, had it not been Homer's birth-place.

⁹⁹ The well-known silver tetradrachm published by Weber, Corolla numismatica Head (1906) pl. 15. 6 (cf. Cadoux, , Ancient Smyrna 92)Google Scholar, whatever the occasion of its minting, can hardly rank as evidence of the political status of Smyrna in the fourth century so long as it is unsupported by bronze issues. Apart from the peculiarity that the lyre had nine or ten strings, the obverse and reverse types are pure Colophonian ones of the first half of the fourth century B.C. (cf. Milne, , Kolophon and its Coinage, pls. 3–4)Google Scholar, and the coin should perhaps rather be considered a Colophonian issue struck on behalf of the Smyrnaeans.

¹⁰⁰ vii. 5. 1–2.

¹⁰¹ Cf. Cadoux, op. cit. 95.

¹⁰² NH v. 118.

¹⁰³ The lack of street drainage, which is remarked by Strabo (xiv. 646) as a serious omission in the planning of the new city of Smyrna, might possibly be argued as showing that Smyrna was an earlier foundation than the new cities of the diadochs.

¹⁰⁴ CAH vi. 429, Alexander the Great i. 133.

¹⁰⁵ In the writing of this section on the history of Old Smyrna I have had the benefit of advice and assistance from Professor Sir F. E. Adcock on the historical issues, of Mr. J. K. Brock, of Mr. J. Boardman, and above all of Mr. Nicholls, to whom many of the observations incorporated in it are due.