¹ The looseness of usage is well illustrated in Strabo XIII 593 f., where Ilium is spoken of variously as a ϰώμηbefore Alexander the Great, as made a πόλις by him and Lysimachus, and then indiscriminately called a ϰωμόπολις, a πόλις, and a ϰατοιϰία around 190 B.C. In fact its city status from about 425 B.C., even when under the control of imperial Athens, Persia, or a despot, is not in doubt.

² For Marpessos see Troad 280–82. Unfortunately the site is covered with fallen pine needles and would need to be cleared before archaeologists could tell whether a population of that size in Roman times has left any recognisable impression on the scatter of potsherds on the site.

³ Its status after 188 B.C. was that described by Lactantius who gives his eighth (Hellespontine) Sibyl's birthplace as ‘in agro Trojano, vico Marpesso, circa oppidum Gergithum’ (16).

⁴ The epigram (Paus. X 12) ends πατρις δέ μοί ἐστιν ἐρυϑρή Μαρπησσὸ.ς, μητρὸς ἱερὴ,ποταμὸς δ᾿ Ἀϊδωνεύς. For the rival claims see most recently Graf, F., Nordionische Kulte (1985) 337 ff.Google Scholar The Erythraeans could produce no parallel when they rejected the connection between Ιδογενής in the epigram and the mountain of the Troad; also, Phlegon (ap. Steph. Byz. s.v. Gergis) seems to have known the Sibyl's tomb as being in the sanctuary of Apollo at Gergis – a tradition distinct from that which Pausanias received.

⁵ Most recently in BCH 106 (1982) 330 n. 70.

⁶ It may be noted that Franke, P.R., in Buchholz, H.G., Methymna (1975) 163Google Scholar, leaves the question of the supposed Methymnaean origin open but remarks that there are ‘keine Anhaltspunkte’ for their minting at Methymna.

⁷ Monnaies ant. en Troade 112. My view was that the legend θΥ on the coins probably represents some other name than Thymbra (Thyatira, for instance, was an outlying city of Mysia south-east of Pergamon, and Pergamon seems to be the focus of such of these coins as have a provenience).

⁸ Monnaies ant. en Troade 106–08.

⁹ BCH 106 (1982) 330 n. 70.

¹⁰ In the same way Robert constantly accepted as being beyond dispute the findings of V. Cuinet in modern Turkey (Troad 42 f., and Robert's defence of him in A travers l'Asie mineure (1980) 111 n. 25; contrast for instance H.B.F. Lynch, Armenia (1901) at various points and especially II 79 n. 2, ‘Vital Cuinet, whose statistics I have rarely found reliable’).

¹¹ This is probably right; but we should not forget that Stephanus Byzantius cities eight cities named Athenai of which the sixth, in Euboea, is entered in the Athenian tribute lists (Athenai Diades).

¹² The purpose of the list, inscribed and set up at Delphi, was perhaps mainly to gratify visiting thearodokoi rather than, as Robert more than once insisted, to serve as a political record.

¹³ For Incorporated and Subject Peraea see P.M. Fraser and G.E. Bean, Rhodian Peraea (1954), and more generally Bean, , Turkey beyond the Maeander (1980) 128–38.Google Scholar For the location of Kallipolis and Robert's discussion of it see Bean, and Cook, , BSA 52 (1957) 81–85Google Scholar, Bean, op. cit. 130, and in Princeton Encyclopedia of Classical Sites (1981) 287. The location at Gelibolu depended on the name, that inland on Hula and Szanto's discovery of a sanctuary of the Kallipolitai and on archaeological and topographical study. A point that is too often overlooked deserves mention here. When an inscription giving an ancient place name is discovered, the archaeological context should be considered. To take comparable cases from the same region, the siting of Old Cnidus at Datça is not proved by the classical decree of the Cnidians found there but the fact that it was lying in the middle of a classical site creates a strong presumption (BSA 47, 187 and 52, 85); on the other hand the siting of the Artemis Kindyas temple at the point where the inscription was unearthed receives no support from the other finds on the spot since they were Early Christian (AR 1964–65, 56 and 1970–71, 48). As regards the Kallipolitai, Hula and Szanto's case for recognising a sanctuary at Duran Çiftlik where they found the dedication is strengthened by the fact that remains of an ancient sanctuary came to light there.

If for any reason this Kallipolis was not in Rhodian possession at the time of the theoroi's visit, we must enquire why Idyma, for instance, was not also visited. We know of no reason why Kallipolis should have been regarded as more Hellenic than other places here.

¹⁴ Cf. Fraser and Bean, Rhodian Peraea 110 f.

¹⁵ Cook, , BSA 63 (1968) 38.Google Scholar An inscription from Denizli, (Chiron V (1975) 59 f.)Google Scholar attests the existence of another Neon Teichos much further to the south which could account for the single Neoteichite recorded (in Egypt) in full Hellenistic times. My very tentative suggestion BSA 63, 38 n. 9, that Olympos (Robert, L., Hellenica X 179 ff.Google Scholar) could be mentioned in Pliny, , NH V 121Google Scholar, is unlikely; it now seems to me more probable that the MSS read something like ‘Aegae itide(m) Posidea’; it is not sufficiently realised that, in spite of differences indicated in their critical apparatuses, the standard texts of Pliny in the geographical books comprise a vulgate rather than critical editions (cf. Cook, CQ 53 (1959) 116–25).

¹⁶ Études anatoliennes (1937) 111 ff. For the places known from boundary stones there, see Cook, , BSA 53–54 (1958–1959) 18Google Scholar. For Temnos see now Herrmann, P., 1st. Mitt. 29 (1979) 239–71.Google Scholar

¹⁷ Troad 342 f., BCH 106, 330 n. 70.

¹⁸ The site of Atarneus (Plate 1a, from the west, with a stretch of the east circuit wall Plate 1b) has in the past been known as Dikeli Kale. In 1960 we heard it called ‘Aristo Kale’ – a name that brings to mind Aristotle, who once lived there. Further enquiry locally elicited the information that the proper name of the site was ‘Aterna’ and that a great philosopher (‘Sokrat Bey’) had lived there; it emerged that the information stemmed from a party of students from Izmir who had visited Dikeli. Such fortuitous ‘survivals’ of ancient names have been reported by travelling archaeologists in the present century; some may have been of similar origin.

The abundant surface pottery on the Atarneus site in 1960 was mainly fourth-century, with tile of that date and one or two black glaze sherds that might be fifth-century; but on the German excavators’ dump at the south-west corner of the citadel (where houses had been cleared) there were also Hellenistic sherds, including West Slope, Megarian, and early eastern sigillata, giving a date around or after 100 B.C. for the end of occupation.

The site of Kanai is at Bademli Pilaji (formerly A(n)canoz) on a peninsula which in 1960 was called Yahudi (formerly çifit) Kale (Plate 1c, from the south); the sandy modern plage lies to the east of the peninsula. Ancient potsherds were dense on the isthmus and the east side of the hill from which Plate 1c was taken: two fragments of perhaps fourth-century black glaze but nothing earlier, a good deal of Hellenistic including early eastern sigillata, and abundant Roman (especially Late Roman red ware and combed ware). There was no sign of marble buildings.

¹⁹ In 1960 I was told by a former retainer on the spot that the family ended with Despoina Trikoupi, who settled in Kephissia near Athens when the land was lost to the Greeks; the çiftlik was then bought from the Turkish government by a pasha who lived in Istanbul, but after his death it had been divided into seven separate farms.

²⁰ ‘“Arab” Camel Drivers in Western Anatolia in the fifteenth Century’ Revue d'Histoire Maghrebine (Tunis) 10 nos. 31–32 (1983) 256–70. The name Araplar with its cognates is of particular interest since the Arab or associated immigrant groups were commonly concerned with working the camel transport routes. Thus in the Troad the village of Araplar (now Koca Köy) near Lekton (Troad 236f.) lies on the track from Gülpinar to Behram Köy which seems to follow the line of the Roman road from the Smintheum to Assos, while the now deserted Araplar in the Scamander gorge behind Pinarbaşi (Troad 127) could imply that the river bed provided a regular route from Ezine to the Trojan plain in the dry season. One caution is however necessary: e.g. the name Arap (Kale) which I heard and recorded (Troad 65 and 78) has been corrected by Aşkidil Akarca to Harap (=ruin).

²¹ Inalcik 261 (32 Greek peasants working çiftlik units at Mahram and serving as fortress guardians, but apparently none extant by A.D. 1500), 258 (at the Tuzla saltworks two converts in the later fourteenth century – Karaca son of Mikhal and Muhammad son of Vasil). But the obligation of Greeks who had been assigned to make oars for the navy under Mehmet II (1451–81) re-appears in 1522.

²² For Menemen, which received its name from the ‘mad’ river (Gediz Çay, Hermus) when it changed its course southward in mediaeval times (and was only prevented from closing the inner gulf of Smyrna by a timely diversion into its ancient bed in 1886) see Cook, , BSA 53–54 (1958–1959) 18 f.Google Scholar

For Peçin (‘Old Mylasa’), the seat of the former beys of Menteşe, see Cook, , BSA 56 (1961) 98–101Google Scholar (with references to which should be added Pococke, , Description of the East II 2 (1745) 63Google Scholar and Turner, Google Scholar, Journal of a Tour III (1820) 72Google Scholar; also now Bean, G.E., Turkey beyond the Maeander (1980) 30 ff.Google Scholar, A. Akarca in Belleten 1971, L. Güçer (as cited in Troad ad loc. and in Inalcik) for the saltworks). The name Tuzla has been applied in modern times to the lagoon at near-by Bargylia.

²³ It is true that a careful reading of Strabo XIII 597–606 shows that while he speaks of all five former cities as being incorporated and more than once speaks of the territory of Neandria, Hamaxitos, and Cebren equally as subdivisions of that of Alexandria Troas (X 472, XIII 604 and 606), he does not refer expressly to any of them save Cebren (together with Scepsis) as being included in the original synoecism, and geographically only Neandria would necessarily be involved in it. It would therefore be possible to contend that Demetrius of Scepsis (first half of second century B.C.), whom Strabo is quoting here, did not explicitly state that Larisa and Hamaxitos were incorporated at the outset; but it is certainly implied.

The need for care in using Strabo will become obvious. But Neandria is a good example of the need for caution in using Stephanus Byzantius. His statement (s.v.) that it is ἐν Ἑλλησπόντω̨ (ὡς Χάραξ), though misleading at first sight, is not seriously incorrect (geographical writers would have been more likely to indicate either the Troad or Mt Ida). But his remark that some spell it Leandros shows how confused he can be. And when he goes on to cite the ethnic Νεανδρεῑς from Strabo he cannot have had in front of him (as the most recent editors of Strabo would have us believe) a text that was the ancestor of our Byzantine MSS but was quite different from the largely notional Vatican Palimpsest because all the Byzantine MSS of Strabo read Νεανδριεῑς, which Meineke was unaware of because he had finished editing Stephanus before he turned to Strabo in Kramer's new recension). Some of Stephanus’ errors are obviously faults of transmission; but not a few are blunders, e.g. Meineke s.v. Arteatai (a gross misreading of Herodotus), and in the East Aegean Assos (1 = 2), Astypalaia (1 = 3), Kyme (1 = 2), Sminthe, Cherronesos init., ib. p. 721 f. (Halesion, Tragasai, Chyton), and in his Index Rerum s.v. Kyklades and Sporades. The use of Stephanus’ text to correct Strabo's is especially hazardous; Strabo and Homer are the authors whom Stephanos cites most frequently except for Hecataeus, but the assumption (which Sbordone, F., Mélanges E. Tisserant V (1964) 345 ff.Google Scholar seemed to make) that when he cites no authority he is following Strabo is demonstrably mistaken. Cf. Cook, , JHS 79 (1959) 19 ff.CrossRefGoogle Scholar

²⁴ For the description and discussion of the Liman Tepe site and the history of Larisa see Troad 216–21 and 196–98. The mention of a Larisa which sent theoroi to Samothrace (ib. 221) might have been brought into play by Robert (I first suggested the Larisa, Troadic in CR XII (1962) 100Google Scholar).

²⁵ Études de numismatique grecque (1951) 34 f.

²⁶ BCH 106 (1982) 319–33.

²⁷ Normally a preponderance of coins is needed for the identification of an ancient site, as (in the Troad) at Cebren, Scepsis, Neandria, and initially Ilium. But there are occasions on which it is not unduly bold to identify a site on the evidence of a single bronze coin if it is of a sufficiently rare issue. Lamponia (Troad 263) is a clear case, as also is Kolonai (Troad 220); and my suggestion that the site on the Balli Kaya was Gentinos (Troad 139) has now been reinforced by Akarca's report of a second coin of Gentinos there. But in those cases there were also other pointers to the identification.

We should of course remember that there are other maritime cities in the Aeolic region whose coins had a reverse type of a vase.

²⁸ In general, though on occasion (as at Gelibolu, above P. 10) regarding his own observations in the field as decisive, Robert dismissed surface finds such as tile and potsherds as invalid; but progress in archaeological exploration largely depends on such evidence.

²⁹ Thus G.K. Jenkins was unable to determine whether my bronze coin from Liman Tepe (Troad 219, pl. 24a) is of Hamaxitos or Alexandria because the types do not differ. What does seem clear is that the Alexandria ones were more or less immediate successors of the Hamaxitos ones.

³⁰ This then would be the date at which the Sminthia superseded the Πύϑια ἐν Τρωάδι as the main agonistic festival celebrated by Alexandria Troas.

³¹ Strabo's mentions (from Demetrius) of the territories of the former cities of Neandria, Hamaxitos, and Cebren as subdivisions of that of Alexandria Troas (above P. 14n. 23) might seem suggestive, though Neandria at least can never have been detached.

³² I noted a few Byzantine sherds in 1959, and coins of Roman date are found. But the absence of ancient tile and potsherds later than classical on the surface must mean that Hellenistic was never there.

³³ Cf. Troad 338 and 344. The facsimile drawing given by Sterrett suggested a third rather than second-century date; presumably Robert's reason for preferring the second century was that on his theory the place should have been named Antiocheia, not Cebren, in the third century. Merkelbach in the meantime has republished the inscription with a photograph (Die Inschriften von Assos (1976) no. 4 and pl. 1) and dates it ‘wohl 3. Jahrh. v. Chr?; that date should of course have been seen to be decisive against Robert's Antiocheia.

³⁴ In BCH 106 (1982), especially 330 n. 70, where the dating of the Assos inscription is brushed aside.

³⁵ BCH 106, 322 n. 23, where my argument is accidentally misrepresented: the point that I made was that no coin of Birytis seems to have come from the Ada Tepe site (not the çal Daĝ). For the discussion of the situation of Birytis see Troad 353–57, with my reasons for supposing that it was not incorporated in the synoecism of Alexan dria Troas but (at a later date) in that of Ilium.

³⁶ Monnaies ant. en Troade (1966) 113, cf. Troad 324 and 387. In fact, so extraordinarily few non-Assian city coins were published by the American excavating team (none whatever of the neighbouring cities of Lamponia, Hamaxitos, Cebren, or Antandrus) that there is a doubt whether the share that went to them in the division of the finds was representative of the whole. For Clarke's Pionia site it is the archaeological evidence that is decisive.

³⁷ The total number of Hellenic city mints whose coins have been recorded at different times in the Troad is nearly 70 according to my notes. Hellenistic cities not represented are all fairly distant, the nearest being Kios, Astakos, Perinthos, Abdera, Pyrrha in the south of Lesbos, Gryneion.

Typical of the various anomalies that come to light when Robert's contention is examined is this: he assumes that Birytis had the same history as Cebren (incorporation in Antigonus' synoecism and transference to Antiocheia) and therefore also lay on what became the territory of Alexandria Troas in the middle Scamander valley; yet ten bronze coins of Birytis seem to have been recorded as found at Ilium at different times (Troad 356 f.) but not a single one of Cebren (or of Neandria, which was also absorbed in Antigonus' synoecism). Points like this may be minute but they have a cumulative effect. Robert's last sentence in his footnote 23 on his p. 323 of BCH 106 had long since ceased to be even partially true.

³⁸ What seems to me to be at issue is this. It is believed that the decade 310–301 B.C. saw the incorporation of cities like Neandria in the new foundation of Antigonus, the establishment of the festival of Athena Ilias, and the first striking of coins of Ilium itself. Neandria and Cebren were evidently issuing bronze coins in quantity until their incorporation, but none of their coins have been found at Ilium. Coins of Sigeum and Birytis, on the other hand, are frequent at Ilium; it is therefore reasonable to suppose that they continued to be minted after the festival was instituted and therefore that they were still flourishing then.

³⁹ I did show illustrations of Troadic coins to the jeweller Riza Ürgüvin in Bayramiç and occasionally of types (as those of Birytis) to other handlers of coins; but clearly such enquiries need to be pursued more systematically if proveniences are being sought.

⁴⁰ I am indebted to Dr S. Mitchell for reading a draft of this article and making pertinent suggestions.