¹ Cf. the sequence of coiffures now set up by van Zwet, L. Furnée, Bulletin van de Vereeniging xxxi (1956) 1 ff.Google Scholar; the author has kindly examined photographs of the Marmaris head and considers it middle Augustan, roughly contemporary with the head of Livia, Ny Carlsberg 615 (ibid. fig. 12). Prof. B. Ashmole kindly drew our attention to this study.

² See the detailed report in Türk Arkeologya Dergisi 1957.

³ Cf. Adm. Chart 1604; Med. Pilot iv (5th ed., 1918) 329; Philippson, , Reisen v. 77Google Scholar; Maiuri, , Ann. iv–v. 405, fig. 11.Google Scholar

⁴ Bericht über eine Reise in Karien, SB Wien no. 132 (1894) 31 f.

⁵ Cf. Paton, , CR ii. 329Google Scholar, JHS xi. 110.

⁶ We are indebted to Mr.Corbett, P. E. for giving the above date to the sherd and referring to Hesperia xviii. 329, 68 for comparison.Google Scholar

⁷ Op. cit. 339.

⁸ This evidence is similar to that, for example, which led to the location of Hygassus at Bayir, an identification which has since proved to be false: see Peraea 67, n. 1.

⁹ The site of Amnistus has hitherto remained quite unknown, though the demotic is fairly common in Rhodian inscriptions. Hiller, (IG xii. 1, p. 10)Google Scholar placed it in the Peraea, but without bringing arguments in support. Our no. 2 is, if we are not mistaken, the only epitaph of an Amnistian yet discovered in the Peraea.

¹⁰ Ht. 0·64, w. 0·62, thickness 0·36 m. (the top surface almost entirely broken away); relief depth 0·025 m.; the woman, apparently full-face, sits on a high-backed throne, her feet on a footstool; a small boy stands at her side; the man extends his right hand to the woman, who seems to be grasping it.

¹¹ The cone appears in the centre of the background in the view from Idyma, , Robert, , Ét. anat. pl. 32. 2.Google Scholar

¹² Cf. Med. Pilot iv (1918) 343; Newton, , Halicarnassus ii. 625.Google Scholar

¹³ JHS xi. 110.

¹⁴ See Peraea 69.

¹⁵ We are very much indebted to Mr. W. K. Smith for providing us with the foregoing information about the text of Mela.

¹⁶ Pliny, , NH v. 107Google Scholar, seems to duplicate the place, Etene (Eutene E ², R ²), Eutane (Euctane R); his location of Eutane is evidently up the Ceramic Gulf and thus lends some confirmation to Mela.

¹⁷ Possibly it is also to be identified with the of Const. Porph., De them. xiv, but this is far from certain.

¹⁸ We note that Pliny, NH v. 100Google Scholar uses the same term of Phellus in Lycia: Antiphellos, quae quondam Habesos, atque in recessu Phellos. The site of Phellus has been disputed; we believe that Pliny refers to its position on Felendaǧi some 2 miles inland. The alternative site close to the shore at Bayindir Limani is, it is true, in a small bay, but no more so than Antiphellus, which is not so described by Pliny.

¹⁹ It is, of course, possible that there was some ancient occupation between Söǧüt and Bencik. But the coast there is now quite devoid of habitation (above, p. 60), and no remains were noted in the survey for the Admiralty charts. So it seems to us unlikely that Euthena can have lain there.

²⁰ iv (1918) 339 f.

²¹ Ann. iv–v. 379 f.

²² The antiquities have been briefly described by Diehl and Cousin (BCH x. 423 f.) and by Guidi (op. cit. 380–4). Our investigations have added nothing of note. The island is also known today as Sedir Ada, i.e. ‘Cedar Island’. This may be a modern rendering of the ancient name rather than a genuine survival; it does not appear to have been known to the early travellers.

²³ The theatre is well described by Guidi, loc. cit.

²⁴ Tombs here are described by Guidi, op. cit. 384 f.

²⁵ See Robert, 's remarks, Ét. anat. 497 and his pl. xxxix.Google Scholar

²⁶ Robert in 1934 evidently climbed the hill, since his photograph pl. xxxix is taken from it; but what, if anything, he saw on it does not appear.

²⁷ Op. cit. 497, n. 1.

²⁸ Op. cit. 376 f., with photographs of the upper circuit, figs. 37–38. Cf. Robert, op. cit. 498.

²⁹ If, as we suppose, this is Guidi's fig. 38, it is certainly a part of the original design, since the two ends of the curtain are running in overlapping arcs and could not otherwise meet.

³⁰ This form of gate, presumably the predecessor of the Hellenistic round-arched one, seems to have been common in the east Aegean between the second half of the fifth century and the beginning of the third; besides the well-known examples at Assos, the ‘Acropolis of Old Smyrna’, Heracleia ad Latmum, and Lysimachian Ephesus, we have noticed such gates at Termera (Asarlik, , BSA l. 116Google Scholar), Samos, (AA 1931, 289, fig. 37)Google Scholar, Kozpinar (Calynda, , JHS lxxiii. 26, fig. 23Google Scholar), and the dummy ‘false arch’ at Söǧüt (above, p. 61).

³¹ They are shown on Admiralty Chart no. 1604 and on Kiepert's map: cf. Robert, op. cit. pl. 30. Von Diest, (Petermanns Mitteilungen 1909, 223)Google Scholar and Robert (op. cit. 483) passed close by without paying them a visit.

³² Peraea 76, with n. 1.

³³ The muhtarlik, which includes Taşbükü and Çetibeli, has 115 houses, of which (according to the muhtar) 70 are at Gelibolu.

³⁴ Med. Pilot iv (5th edition, 1918) 345. This excellent description helps to explain how a Hellenistic city was apparently able to exist at Gökbel (BSA l. 135).

³⁵ Paton and Myres, op. cit. 197; Guidi, op. cit. 385 f., fig. 44. Cf. Med. Pilot iv. 344.

³⁶ Pet. Mitt. 1909, 223; Robert, op. cit. 478.

³⁷ By softening of intervocalic k. The Greek Portolan ii speaks of the gulf here under the name H. Kosmas and also gives the name (Delatte, , Les Portulans grecs 248Google Scholar).

³⁸ For descriptions of the plain see Med. Pilot 342 f. and Robert, op. cit. 479.

³⁹ Op. cit. 190.

⁴⁰ Op. cit. 342 f.

⁴¹ Op. cit. 484, n. 3.

⁴² Op. cit. pl. 31, 1 (looking south over Kozlu Kuyu and the plain), 2 (looking south-east over the Kadin towards Duran Çiftlik); pl. 32, 1 (looking west-south-west over the gulf and the flank of the Kiran Daǧ), 2 (looking south-west over the plain to Ferek and Altinsivrisi).

⁴³ BCH x. 428 f.

⁴⁴ A view from the south, looking up towards the site, Robert, op. cit. pl. 33.

⁴⁵ One is illustrated, Robert, op. cit. pl. 34, 1; we suppose it to be the stretch of circuit that we have marked on the south side on Fig. 3.

⁴⁶ Paktyes in the Athenian tribute lists of 453/452 and 451/450 (ATL i. 288). Nesselhauf, , Klio Beih. xxx (1933) 126, n. 2Google Scholar, remarked that might be a pair of ethnics; and Robert refers to him and speaks of Paktyes' domination as being (at the most) momentary (op. cit. 473). But the only other instance of such coupling in the lists of the first assessment period is the entry and here an ethnic can be definitely ruled out. At both Idyma and Syangela we find that the dynast's name is omitted in lists 7–8 and continuously on through the third and fourth assessment periods. Pigres of Syangela reappears in list 23; Paktyes (or his successor) does not reappear, for the Idymes had dropped out of the lists in the fourth period. Thus, for the period during which Idyma was in the league, the evidence for the existence of a dynasty there is precisely the same as that for the dynasty at Syangela, and there is no reason to doubt that Paktyes is a dynast's name.

⁴⁷ Op. cit. 370 ff. For other references see Robert, op. cit. 482.

⁴⁸ Loc. cit. fig. 32.

⁴⁹ Loc. cit. figs. 33–34; Robert, op. cit., pl. 36. 2; the right-hand anta, with the corner of the entablature, has now collapsed.

⁵⁰ The fasciae here and on the antae had no doubt painted Ionic decoration, as was noted on the tomb of Amyntas at Telmessus (see below). Paton and Myres remarked painting on the façade of the upper tomb at Sandama, (JHS xvi. 261 f.)Google Scholar; cf. Guidi's ‘intonaco’ on this tomb no. 4.

⁵¹ A row of tombs here, Robert, op. cit. pl. 34. 3; the architectural one, ibid. pls. 35, 36. 1.

⁵² The distyle tomb has six graves inside; for the position cf. Paton (below, n. 54), Hula-Szanto, , Bericht über eine Reise in Karien 34Google Scholar, Laumonier, , BCH lviii. 344.Google Scholar

⁵³ Reisen in Lykien 41, 112, pl. 17 (the inscription 40 no. 9); cf. Hirschfeld, , Paphlagonische Felsengräber 31, n. 2.Google Scholar

⁵⁴ CR ii. 328 f., JHS xvi. 260 (a); from Paton's correspondence with Myres, which by the generosity of the latter we were able to read, it is certain that the phrase ‘the inscription ΝΟ’ was a misprint for ‘the inscription No. 9’.

⁵⁵ We recall that Benndorf and Niemann likewise were convinced that the very similar tomb of Amyntas at Telmessus must have been a princely tomb and not that of a private citizen (Reisen 41).

⁵⁶ Halicarnassus ii. 626.

⁵⁷ We have nothing to add to the existing descriptions of this site; cf. Robert, op. cit. 480.

⁵⁸ Cf. above, n. 37, and Tomaschek, , Zur hist. Topographie v. Kleinasien im Mittelalter 39.Google Scholar

⁵⁹ See Robert, op. cit. 481, n. 1.

⁶⁰ See especially Guidi, op. cit. 369 f., fig. 30; cf. Robert, op. cit. 481 f., pl. 34, 2.

⁶¹ iv (1918) 342. But see Wittek, , Das Fūrstentum Mentesche 172.Google Scholar

⁶² Op. cit. 488, 497.

⁶³ Pet. Mitt. 1909, 223.

⁶⁴ For Cousin's ‘Kara-Burtlen Assar’ see below, p. 74, n. 69; L. Robert in 1934 does not seem to have pursued his investigations east of Duran Çiftlik. It is not at all easy to relate this terrain to Kiepert's and Philippson's maps.

⁶⁵ Ht. 0·69, width 0·34, thickness 0·095 m. The stone is not marble, but has a rough surface.

⁶⁶ Ht. of back 0·88 m.; ht. of seat 0·36 m. The back and ends have plain mouldings at top and bottom.

⁶⁷ By the Kallipolitai, , Bericht über eine Reise in Karien (SB Wien no. 132) 34Google Scholar; see below, p. 81.

⁶⁸ Cf. the similar socket over the door of Robert's tomb above Kozlu Kuyu (p. 71).

⁶⁹ We are inclined to think that Hisarlik is probably the site described by Cousin, G. in BCH xxiv (1900) 42Google Scholar, under the name Kara-Burtlen Assar, ‘en face de Groba. Sur une colline faisant partie de la chaîne qui sépare Groba de Djova et de Kara-Agatch [this is not really intelligible] se trouvent les ruines d'une ville qui semble avoir été assez importante. En plusieurs endroits le mur est assez bien conservé et, surtout du côté de la plaine, on le suit facilement. Il est construit en appareil irrégulier, qui du côté de la montagne est en très bon état. Ce sont de gros blocs de cette pierre grise que l'on trouve partout dans le pays. Ils s'ajustent ensemble assez bien, et les vides sont remplis par des pierres de moindre dimension. … À un coude il y eut autrefois, soit un escalier, soit plutôt une tour carrée. La saillie rectangulaire est en appareil régulier. … Sur la côte ouest plusieurs redans. Il y a en tout plusieurs rangées de murs, qui semblent parallèles … Au dessous du mur vers la montagne du côté sud, il y a encore des restes de monuments et édifices importants.… Saraikli, petit village au SE de la colline qui porte ces ruines.’ This could hardly be called a good description of Hisarlik, but the points of resemblance seem enough to make the identification probable. We could learn of no other site ‘en face de Groba’, and in fact Saraikli (unknown) seems to correspond to Sarayyani.

⁷⁰ The total disappearance of the wall itself is remarkable, but we observed much the same thing in connexion with the outer perimeter wall of early Cnidus (see BSA xlvii. 176, n. 22) and elsewhere. It is hardly conceivable that the Asar site can have been unwalled; it is certainly not merely a sanctuary site.

⁷¹ No. 33 was republished in Peraea 45, no. 42.

⁷² The provenience was confirmed to us personally by a local resident who had his information from the soldiers who dug the stone up near the karakol (now the gümrük) at Gelibolu in 1949.

⁷³ Both features occur occasionally on subject territory, notably at Idyma; see SEG iv. 175 (= Robert, Ét. anat. 477, no. 4) and no. 11 below. Since the stone was actually unearthed at Gelibolu, it is hardly permissible to suppose that it has been brought there from elsewhere in modern times.

⁷⁴ A place by the name of κότα is recorded SIG ³ 46, ), but it presumably lay, like the other recognizable places in that inscription, in the region of Halicarnassus.

⁷⁵ So far as can be judged from the printed epigraphic text, which has the forms Α, Μ, Π, Σ, Ω, there is at least no apparent discrepancy.

⁷⁶ The stone is described as ‘base en pierre, brisée en haut’; if this means that the text is complete at the bottom, we must restore somewhat differently: It is then possible, though not necessary, that we have fragments of two separate lists of priests of Leto and Aphrodite.

⁷⁷ On line 2 Robert notes: ‘on restitue il me semble qu'il faut ’ But is right. The man in question has changed his place of residence upon adoption; in such a case both toponymies are given to the man himself, not to either of his two fathers: e.g. Blinkenberg, , Lindos ii. 330Google Scholar cf. IG xii. 1. 181, Lindos ii. 259, and Blinkenberg's remarks on p. 96. In line 2 Robert prints but is the Rhodian form.

⁷⁸ is most naturally taken to relate to Laodicea on the Lycus, but this is not of course certain.

⁷⁹ Natives of all parts of the Peraea are regularly described in this way; no indication is therefore afforded as to whether a particular place lay in incorporated or in subject territory, except that on subject territory Rhodian demesmen have the ethnic in place of their demotic. Idyma itself was certainly subject: see Peraea 71.

⁸⁰ A small locality by the name of Carpotheca would be perfectly possible if it lay in the vicinity of Idyma, but in this case Callinicus should have a patronymic; having none, he would necessarily be a foreigner, and a city by the name of Carpotheca is hardly conceivable.

⁸¹ See n. 79 above and Peraea 53.

⁸² There can hardly be any question at this date of a deliberate omission; forms such as &c., are common in the Imperial period, but not in the second century B.C.

⁸³ See RE s.v. Parochos, and for a more general use about the date of our inscription, IG xiv. 951 (1st C. B.C.).

⁸⁴ masculine, shows that the dedication is a private one, not by the koinon of Idyma. would be abnormal at this date. The dedicator had no doubt benefited personally by the governor's generosity.

⁸⁵ Op. cit. 494, in his chapter (ibid. 491–502) on the subject of Callipolis.

⁸⁶ JHS xi. 109 f.

⁸⁷ See Peraea 71. In the unusually abundant inscriptions of Rhodes and its incorporated Peraea and islands not a single Callipolitan is ever mentioned (the ethnic was read in IG xii. 3. 48 (Telos), but is corrected by Hiller, ibid. p. 276). This is normal if Callipolis was on subject territory (in the three main collections of Rhodian inscriptions there is a single mention of a Pladasan, (IG xii. 1.Google Scholar 962, from Chalke), but Idymians, Therans, Pisyetes, and others are totally absent); but it would be extraordinary if Callipolis was a Rhodian deme. The visit of the Delphic theoroi also suggests that Callipolis was not incorporated in the Rhodian state.

⁸⁸ Idyma, Cedreae, and the Tripolitai (who presumably also lay outside the confines of the incorporated Peraea).

⁸⁹ The situation of Thera on the plateau (Peraea 47 f.) was of course not known when Robert wrote. Theangela is not named among Orontobates’ positions; presumably it had been won over by Alexander.

⁹⁰ Cf. Cramer, , Asia Minor (1832) ii. 216.Google Scholar Newton, however, gives the name (of the stream) as Bourla, Gheli (Halicarnassus ii. 625).Google Scholar

⁹¹ The one established parallel that presents itself is Erine (Rena, Arine, now Hisarönü) at the head of the Gulf of Syme. But this identification, though attractive, is like wise uncertain. The name Rina is not rare (e.g. on Kalymnos and Leros), and on Kalymnos at least it is derived from the name H. Eirene; at Hisarönü it may have the same origin.

⁹² Thus the Greeks of Menemen, Nymphi, and Magnesia (under Sipylus) built themselves a fortress at Old Phocaea about the beginning of the fourteenth century (Tomaschek, , Zur hist. Topographie v. Kleinasien im Mittelaitsr 26Google Scholar).

⁹³ Cf. ATL i. 508, where (in connexion with the presumed situation at Gelibolu) it is suggested that Callipolis is a ‘Hellenistic euphemism for Kyllandos’.

⁵⁴ IG xii. 1. 1036; Peraea 99 q.v.

⁹⁵ See Peraea 72 f., 76. ATL's location near Gelibolu (above, n. 93) is bound up with the false location of Callipolis there.

⁹⁶ The order in which the three territories recovered by Nicagoras are cited can hardly be that in which they were reconquered. But if (as we suppose) the Cyllandia bordered the Idymia on the east, the order can fairly be called a geographical one; whereas there would be no order at all if the Cyllandia were also on the plateau north of Idyma.

⁹⁷ Festschrift für Carl Weickert (1955) 94–96. Robert, J. and Robert, L., REG lxix (1956) 163Google Scholar, refer to the ‘cadre du stoichedon’, but this inscription is not written stoichedon.

⁹⁸ Robert, J. and Robert, L., REG lxvii (1954) 169Google Scholar; cf. lxviii. 247.

⁹⁹ The phrase is on either view unsatisfactory to the modern historian as a dating formula; not so, however (if we are right), from the point of view of its own time; in the first year of the new city is in fact a precise date, and might well be acceptable in a verse inscription. The new foundation was of course the great event that filled every man's mind.

¹⁰⁰ Op. cit. 208–10 and map, pl. 37.

¹⁰¹ This ripple is clearly visible in a photograph he sent us. Since we wrote this paragraph Dr. Penfield's account of his discoveries has appeared in Proc. Amer. Philosoph. Soc. 101, no. 5, Oct. 1957, 393 ff. He discusses the situation of Triopion at Kumyer and the harbour mole connecting the island to the mainland (with photographs and sketch-plans of Datça-Burgaz and Kumyer-Palamut Bay; for the snake of his fig. 4 cf. BSA xlvii. 183 (large funerary altar = Penfield's n. 12), Keil-Premerstein, Zweite Reise in Lydien, no. 17, Laumonier, , BCH lviii. 344).Google Scholar Penfield notes that at the foot of the Kumyerkalesi hill towards Kumyer village there is a ‘remarkably level field’, and that the stadium of the Triopian sanctuary may have been there.

¹⁰² Hdt. iii. 60:

¹⁰³ A mole with an open passage in the middle, consisting in fact of two walls projecting from the mainland and the island, would fit Mr. Penfield's observations better.

¹⁰⁴ Even without a wall, an anchorage is shown on the chart near the north end of the island, in depths varying from 9 to 20 fathoms.

¹⁰⁵ BSA xlvii. 209.

¹⁰⁶ BSA xlvii. 194, 209.

¹⁰⁷ REG lxvii. 165 f.

¹⁰⁸ Perhaps in order to let it out to others, as was commonly done with temple land, though in this case it would be more natural to buy the land outright.

¹⁰⁹ The Roberts find a parallel for the present text in the well-known inscriptions from Smyrna, SIG ³ 990. These are not easily intelligible in all parts, but it is clear that the money passes from the temenos to the temple; how this can be parallel to the present case we do not see, and wonder if we can have taken our critics' meaning correctly.

¹¹⁰ A number of the Roberts' other criticisms of the epigraphical part of our article hardly seem to justify the denigratory remarks with which they are prefaced. On no. 2, for example, they have misunderstood our note concerning the date of Iphiadas' proxeny, which marks a point not in the history of the proxeny, but in the career of Iphiadas. For no. 58 their suggestion neglects our observation that about half the name is preserved. Our translation (p. 208, n. 20) of the epigram BMI 797 is said to contain nothing new; yet we note that in Robert's own Ét. épigr. et phil. 44, to which he refers us, the text is still printed in the old erroneous form, with in line 4, where is obviously required. On the single-line epitaph no. 12, we noted that the inscription, brief though it is, appears complete. The Roberts observe, ‘certainement, car elle forme un hexamètre’. Did they suppose we had not noticed this? Why must the text be complete because it forms a verse? What was to prevent the addition of a second or more verses, supplying a little more information? Are we alone in feeling that criticism of this kind is rather tiresome?

¹¹¹ Tomaschek, , Zur hist. Topographie v. Kleinasien im Mittelalter 38Google Scholar (also as a v.l. in manuscripts of the Byzantine historians).

¹¹² REG lxi. 194.

¹¹³ Plieninger, , Pet. Mitt. 1920, 162Google Scholar, mentions ‘den von Kos aus wie ein Vulkanberg erscheinenden, mit einem schönen alten Kastell gekrönten Andesitgipfel Ewräokastro oder Strobilos (oder, wie ihn die englische Seekarte nennt, Chifut Kalessi)’. The Turkish portolan of Piri-reis of the year 1521, also seems to place Strobilos at Aspat (according to Wittek, , Das Fürstentum Mentesche 172Google Scholar).

¹¹⁴ See most recently Hellenica x. 232, n. 1 (‘cette ville (sc. Termera), dont le site n'est pas connu, a disparu dès le début de l'époque hellénistique’).

¹¹⁵ BSA l. 151 ff.

¹¹⁶ Rev. phil. xv (1941) 11 ff.

¹¹⁷ Cf. BSA l. 131, n. 201, map p. 86, fig. 1.

¹¹⁸ See BSA l. 131, 159.

¹¹⁹ ‘Deux Inscriptions de Carie’, Mél. Isid. Lévy (1955) 569 ff.

¹²⁰ BSA l. 100 f. We note that our Melanthios of no. 8 may perhaps be restored in Ann. iv–v. 468, no. 8; for the name in no. 10 Robert refers to his Hellenica vii. 221, n. 4. For the inscriptions mentioned under no. 11, see Wilhelm, , ÖJh viii. 238 ff.Google Scholar

For our inscriptions nos. 33 (a), 34 cf. Grégoire, Recueil nos. 235–6.

¹²¹ Evidently one of the pair unearthed by Newton and now built into a house (BSA l. 92, n. 41, pl. 13(d); for the snake and kantharos in hero reliefs see now Andronikos, M., ‘Λακωνικὰ Ἀνάγλυφα’, Πελοποννησιακά i (1956) 303—5Google Scholar).

¹²² AM xii. 224, figs. 1–2; cf. BSA l. 134.

¹²³ BSA l. 132 and fig. 12.

¹²⁴ Judeich and Winter were disposed to believe that the building, originally a tomb, was at a later time used as a mandra, the size of the entrance being reduced for the purpose. We prefer to think that this was the original purpose.

¹²⁵ This section should be read in connexion with BSA l. 112–15, 145–7.

¹²⁶ The outline of the city wall was entirely traced with compass and tape, and may be regarded as accurate within a very small margin of error; the positions of the individual buildings in the interior were for the most part estimated, without actual measurement, and checked as far as possible by cross-bearings. The contours were judged by eye alone and make no pretence to give more than a general idea of the terrain.

¹²⁷ See BSA l. 146, n. 243. The mentioned in the Theangelan treaty with Eupolemus are accordingly the peaks A, B, and D. The tower at C measures 9·30 by 6·40 m., which is little larger than the ordinary towers in the wall; it is doubtful therefore whether this should be included in the Similar considerations apply also to E.

¹²⁸ Illustrated by Robert, , Coll. Froehner pl. xxviiih.Google Scholar

¹²⁹ The steep south slope of B is covered with a mass of collapsed wall-blocks, and it is not easy to discern any plan. RR is only 0·70 m. thick, but it appears to be continuous.

¹³⁰ Except for the south-east corner tower, which is badly ruined.

¹³¹ BSA l. 114, Robert, op. cit., pl. xxvii d.

¹³² BSA l. 113, pl. 16 c.

¹³³ Specimens were taken to the school at Milâs; those that remained at the tomb had by 1956 all been removed by the villagers.

¹³⁴ The alternative is that these stone balls were used for weight-putting or some such athletic exercise; for similar, but much larger, stones found at Colophon and Pompeii, and supposed to be used for athletic purposes, see Hesperia xiii. 147, 150 and n. 46. But forty such balls is more than could reasonably be required in the palaestra, nor is it easy to understand why in this case they came to be collected on the roof of the tomb.

¹³⁵ TAM ii. 288–9. This is at least its present condition; the ends are not preserved, but it can never, it seems, have been much longer. To judge by the old plans, the stadia at Iasus (p. 100) and Myndus (BSA l. 109) also seem likely to have been well below the standard size.

¹³⁶ AM 1887, 334; cf. Robert, op. cit. 65.

¹³⁷ Apart of course from the vases in the tomb T. It is to be remembered that no proper excavation was possible: soundings could not be taken to a depth of more than 2 metres in the time and with the resources available.

¹³⁸ We considered other possible explanations of ZZ, but none is satisfactory. An earlier Mausolan fortification seems ruled out by the style of masonry; and after the powerful tetrapyrgon was built a wall of this kind would appear unnecessary—or if it were constructed as an additional precaution, it would surely be made to conform better to the style and plan of the major fortification. We even considered the possibility that ZZ might be a modern field wall, but its thickness as well as its position make this suggestion quite unacceptable.

¹³⁹ The three peaks A, B, and D are of approximately equal altitude. The approach is steep in all parts (though rather less so between D and C) except on the west, where the hill-side slopes gently down from D towards the pass on the road from Etrim to Çiftlik. This is the obvious line of approach for a hostile force.

¹⁴⁰ Koca Kale at Gelibolu (above, pp. 65 ƒ.) affords one parallel among many. We note that, if Z were accepted as belonging to a city on B, its style of masonry, which is not seen on the Lelegian sites, would suggest that the city was not Lelegian, and therefore not Syangela. This would open up interesting speculations, into which, however, we prefer not to enter here. For early rubble walls see Scranton, Greek Walls 145, 184–5.

¹⁴¹ About 420 B.C. (BSA l. 113).

¹⁴² These coins were not found in a hoard, but close together, scattered ‘as if from an aeroplane’, as our informant expressed it.

¹⁴³ The apparent exception, Astyra, (BMC Caria xxxviii, 59–61)Google Scholar, is only apparent. There was never a city of Astyra in Caria; L. Robert's suggested identification of it with the Carian Ptolemais is chimerical: see JHS lxxiii (1953) 23, n. 61. The coins have only ΑΣΤΓ, or merely Α, which by no means postulates Astyra.

¹⁴⁴ The possibility (it is no more) that Alâkilise represents Amynanda (BSA l. 165) remains open. Alternatively, it may have been the residence of a member of the royal family.

¹⁴⁵ The situation is very clearly shown on the inset ‘Bargylia Creek’ on Adm. Chart 1531.

¹⁴⁶ Polyb. xvi. 24.

¹⁴⁷ Adm. Chart 1531, inset ‘The Ruins of Antient Bargylia’. Cf. Lebas, , Voyage arch. Itin. pl. 67.Google Scholar

¹⁴⁸ Halicarnassus ii. 604 ff.; cf. Travels and Discoveries ii. 57 f.

¹⁴⁹ Ann. iv–v. 359 ff.

¹⁵⁰ For Myndus see BSA l. 108 ff.

¹⁵¹ For Iasus see below, p. 100; for Siralik see BSA l. 131 f. For Salihadasi (Tarandos I.) see ibid. 132; the island had a town (Caryanda?) in classical times, and seems to have been deserted in Hellenistic times, but reoccupied in the Middle Ages; the fluctuation in its population in the last hundred years perhaps deserves remark. In 1865 Biliotti found a single inhabitant (a Samiot) on the island; presumably he was a lone settler. A decade or two later Haussoullier found a ‘village’ of fishermen's huts on the island (and an inscription, which he assumed to have been brought from Iasus or a neighbouring city; BCH viii. 218 ff., Hicks, , JHS viii. 103 ff., 116ff.Google Scholar). In 1953—4 we found the island uninhabited and overgrown with deep impenetrable scrub. At the present day the island is also called ‘Gölcü Ada’; the fishermen may have come from Göl.

¹⁵² Akarca, A. and Akarca, T., Milâs 166 f., pl. 53.Google Scholar

¹⁵³ BSA l. 168.

¹⁵⁴ The plan is a rough one, the distances being paced; and we did not have time for a proper investigation of the south-west slope below the citadel. Paton and Myres gave a brief description of the site ( JHS xvi. 196).

¹⁵⁵ Cf. BSA l. 132, 159.

¹⁵⁶ Loc. cit. 196, map pl. 10. Philippson's map errs in the other direction, making the sanctuary seem nearer to Çömlekçi Kalesi (Asardağ) than to Siğirtmaç Kalesi. For the site of the sanctuary cf. also Laumonier, , RA 1933, ii. 49.Google Scholar

¹⁵⁷ Hdt. v. 118.

¹⁵⁸ This is the distance as quoted to us by villagers. We have not ourselves been to the Asardağ (Çömlekçi) site.

¹⁵⁹ JHS xvi. 196 f.; A. and T. Akarca, op. cit. 168.

¹⁶⁰ Op. cit. 254.

¹⁶¹ RA 1935, ii. 163; Hellenica viii. 14, no. 11.

¹⁶² See below, pp. 145 f.

¹⁶³ See especially Texier, , Description de l'Asie min. iii. 135 ff., pls. 142–9Google Scholar; Judeich, , AM xv (1890) 137 ffGoogle Scholar. (with references to earlier accounts of the site; for the convenience of those who look at the illustrations we remind the reader that Judeich's figs. 2, 4, 7 are not at Iasus but at Halicarnassus, Thoricus, Bargylia); Guidi, , Ann. iv–v. 345 ff.Google Scholar

¹⁶⁴ The ancient authorities (Strabo xiv. 658, Steph. Byz. s.v.) speak of it as an island, but it was probably joined to the mainland in antiquity.

¹⁶⁵ There is, however, a big fish-garth (‘dalyan’) across the bay from Iasus at the outlet of the ‘Little Sea’.

¹⁶⁶ Arrian, , Anab. i. 19, 11.Google Scholar

¹⁶⁷ Diod. xiii. 104, 7. ATL i. 492 assumes that the place destroyed was really Cedreae, but we doubt whether Cedreae can have been so populous.

¹⁶⁸ Cf. Paton, , CR i (1887) 176 f.Google Scholar

¹⁶⁹ Robert, , Ét. anat. 452.Google Scholar

¹⁷⁰ Judeich op. cit. 144 fr. Judeich's figure of 3·5 km. length seems calculated to include the circuits of the towers. He quotes 1·75–2·00 m. for the wall thickness; we measured it as 1·60–1·70 at several points.

¹⁷¹ Cf. Texier, op. cit. pls. 147–8; Judeich, op. cit. figs. 3 and 5–6.

¹⁷² On the right side of the gate court, high up, is a window c. 60 cm. high; in the curtain to the left are three windows c. 1·05 m. high, set low down in the wall.

¹⁷³ Griech. Städteanlagen (1924) 115, n. 1.

¹⁷⁴ Op. cit. 116 n., fig. 14; the figure of 2·30 m. for the breadth is evidently a misprint.

¹⁷⁵ Cf. BSA l. 166 f. Guidi's fig. 9 (Ann. iv–v. 354) poses a problem which we have discussed (BSA l. 166, n. 346). Von Gerkan's huts are not mentioned in Guidi's text; but since he speaks of tombs of Lelego-Carian type at Iasus and illustrates them with his fig. 9, we are inclined to suppose that the huts did not in fact escape the notice of the Italian party and that Guidi's notice contains an oblique reference to them.

¹⁷⁶ AM xv (1890) 149 ff.

¹⁷⁷ Op. cit. 115, n. 1. Jost, G., in his dissertation Iasos in Karien (Hamburg, 1935) 77 ff.Google Scholar, attaches himself to this theory, but he does not appear to have investigated the matter.

¹⁷⁸ In a lecture reported AA 1913, 476. With his ‘camp retranché des Leleges’ Texier may be considered the originator of this explanation.

¹⁷⁹ Op. cit. 349 f.

¹⁸⁰ The compounds mentioned above (p. 102) lie both inside and outside the circuit and seem to be unrelated to it; in any case they cannot be considered to be city habitations. The piles of stones, with tile fragments and limekilns, which von Gerkan saw in the interior of the site, do not seem to us the right sort of evidence; early Greek cities do not disappear in that way.

¹⁸¹ Thuc. viii. 19, 2 shows that Amorges was capable of taking independent military action on the Ionic coast.

¹⁸² Thuc. viii. 28, 2–29, 1.

¹⁸³ Thuc. viii. 19, 2; Andoc. iii. 29.

¹⁸⁴ Op. cit., 153; cf. ATL i. 492.

¹⁸⁵ Cf. Thuc. viii. 5, 4–5; 28, 2.

¹⁸⁶ e.g. BMI iii, no. 440 (regulation of the priesthood of Zeus Megistos); AM liv ( 1929) 90 f. (woman of Iasus on a Samian funerary altar). Iasus paid a normal tribute of a talent in the Athenian league, and was one of the cities where ΣΓΝ coins were struck in the early fourth century.

¹⁸⁷ With so many barbarians included, this seems the inevitable explanation, though elsewhere, as in the Athenian lists IG ii.² 1956–7, the names of mercenaries are given without patronymic.

¹⁸⁸ Strabo i. 54, vii. 319; Mela ii. 22; Pliny, HN iv. 44.Google Scholar Stephanus gives three forms of the ethnic, and Cf. also the references in REG lxix. 139, 142 to the cult of Dionysus in bull form at Bizone.

¹⁸⁹ See the copious article in RE s.v.

¹⁹⁰ Prof. H. Hommel, to whom we are gratefully indebted for his advice in this matter, points out to us that, when in IG i.² 329 Cephisodorus is designated as this is certainly not intended as a compliment.

¹⁹¹ This seems to be the spot marked as Jaly (i.e. Yali) on Kiepert's map. There are one or two houses but no village. The 1:100,000 map of southern Ionia in Milet iii. 5 marks numerous remains on the Milesian peninsula.

¹⁹² Mentioned by von Diest: see no. 2 below.

¹⁹³ Marked ‘Ru.’ on Kiepert's map.

¹⁹⁴ It would certainly by a nahiye if its communications were better. The roads are so bad that even jeeps only rarely and unwillingly penetrate to it.

¹⁹⁵ There was perhaps an oil factory here. Olives are the main product of the district and are exported as far as Ayvalik. Ayvalik is itself noted for its olive-oil, which is used for cooking, that of Kazikh for making soap. For oil troughs in this region see Paton, and Myres, , JHS xviii. 211.Google Scholar

¹⁹⁶ For Kapakli Kuyu see Robert, , Ét. anat. 437 ff.Google Scholar, but the site is inconsiderable in comparison with the others in the neighbourhood.

¹⁹⁷ The left-hand portion of this inscription is still lying there; in line 1 the stone has in fact ΘΙΛΛΓ, as copied by the editors.

¹⁹⁸ The localities are in fact largely named from the wells: Örtülü Kuyu, Kapakli Kuyu, Tathkuyu, Yavankuyu, Yenikuyu, &c.

¹⁹⁹ A ‘naval base’ at Karakuyu is indeed little short of an absurdity. As a harbour Akbük Limam is better, but far from good. The low ground to the west affords no protection against the prevailing north-wester, and in a south wind (according to local information) it is not uncommon for boats to capsize.

²⁰⁰ Under these circumstances it is surprising that of all the scholars who have previously visited Kazikli not one seems even to have considered the possibility that this might be Teichiussa. Cousin and Deschamps, commenting (BCH xviii (1894) 21) on the inscription mentioned above (n. 197), observe that the occurrence of the stephanephoros Athenais ‘prouve que nous avons quitté le territoire de Milet pour celui de lasos, où l'on connaît déjà la stéphanéphorie éponymique des femmes’. This, of course, does not hold good: the stephanephorate is equally characteristic of Miletus. Von Diest, (Pet. Mitt. 55 (1909) 268Google Scholar), though noting the good connexion with the sea, finds no clue to an identification. Robert (op. cit. 439), by a notable understatement, says that Kazikli, ‘à l'époque impérial au plus tard’, certainly formed part of Milesian territory.

²⁰¹ On a low hill called Çamtepe, a little north of Doğanbeleni, are some rather scanty remains of a fort which may be of classical date: a wall of roughish polygonal about 40 metres long is preserved for one course above ground. The tombs of the type shown in Plate 22 ƒ are difficult to date. Everything else seems to be of the Roman period or later.

²⁰² The similar wall at Emecik on the Cnidian peninsula (Maiuri, , Ann. iv–v. 403, fig. 8)Google Scholar is dated to the sixth century by the carvings on it (BSA xlvii. 172 f., pl. 39 a); cf. also the bit of the old acropolis wall bedded on the rock at Neandria (Koldewey, , Neandria 11, fig. 9, at point ‘a’, not ‘d', on the plan).Google Scholar

²⁰³ For the red mullet of Teichiussa (T. 6, τρίλαν) we cannot speak, as the modern inhabitants are little interested in fishing. The centre of piscatorial activity at present is the dalyan near Küllük, where an abundance of all kinds of fish, including mullet, is caught. We are not very clear what is implied by the epithet ψαφαρῇ. As denoting a poor sandy soil it is hardly appropriate: the plain of Kazikli, despite the absence of running water, is fertile and supports many kitchen-gardens; the water-table is little below the surface and wells are easily dug. Probably ψαφαρῇ. is no more than a poetic epithet alluding to the position of Teichiussa by the shore: see LSJ s.v.

²⁰⁴ The difference in the form of omega observed by von Diest does not in fact exist.

²⁰⁵ Cf. Hamilton, , Researches ii. 70Google Scholar; Med. Pilot iv; Ross, , Reisen iii. 123 ff.Google Scholar; Gregoropoulos, M., Ἡ Νῆσος Σύμη (Athens, 1877) 8Google Scholar; Khaviaras, ÖJh vii. 87.Google Scholar The ‘trophy’, Maiuri, , Ann. iv–v. 456 f.Google Scholar

²⁰⁶ BSA l. 128, 131 f.; for Salihadasi see also above, p. 97, n. 151. For Karaada and Orakada see the references BSA l. 134. For Syrne, Telos, Nisyros, and other islands in their relationship to Rhodes, see Pernea 138 ff.Google Scholar

²⁰⁷ BSA xii. 159 ff., with bibliography 151 f. Some views, Ann. ii. 13 ff. A history of the island is being prepared by Mr. Anagnostides.

²⁰⁸ This is the same stretch as Dawkins–Wace's fig. 9.

²⁰⁹ Herzog, , Riv. Fil. 1942, 15 (there dated c 300 B.C.).Google Scholar

²¹⁰ See Robert, , Rev. phil. lx. 46.Google Scholar

²¹¹ The small vase seen by Ross, (Reisen iv. 44)Google Scholar, with birds and palmettes on a pale ground, might well be of an earlier date.

²¹² When L. Robert speaks of independent Telos being attested from the middle of the fifth century (op. cit. 48) he evidently has this date in mind. The only earlier mention known to us is that in Hdt. vii. 153 to a Telian, the ancestor of Gelon, who joined in the Lindian foundation of Gela.

²¹³ BSA xii. 165 ff. To their bibliography, p. 152, may now be added Papadopoulos, , Γενικὴ γεωγρφικὴ καί ίστορικὴ Περιγραφὴ Νισύρου (Nisyros, 1909)Google Scholar; and Gerola, on the medieval antiquities, Ann. ii. 20 ff.Google Scholar; for the geology and geography of Nisyros and Cos see Plieninger, and Sapper's, article (in four instalments) in Pet. Mitt. 1920 (with 1: 75,000 map and views, pls. 28–30).Google Scholar

²¹⁴ Farther away from the gate the masonry of the curtain is less regular, especially in the upper courses. The great thickness may have been intended to give added resistance against the frequent earthquake shocks.

²¹⁵ There is now no natural harbour at Mandraki; on the question of the ancient one see Ross, , Reisen ii. 72 f.Google Scholar; BSA xii. 169 f.; Pet. Mitt. 1920, 216.

²¹⁶ IG xii. 3. 86 (with drawing; the last two letters in πόδες are in fact stepped up towards the beginning of the next word, and are clear); BSA xii. 167.

²¹⁷ We saw no prehistoric ware, but a headless marble idol of Cycladic type in Berlin is reported to have come from Nisyros (Ant. Skulpt. Mus. Berl. no. 575).

²¹⁸ AE 1913,9 f.

²¹⁹ Clara Rhodos vi–vii. 471 ff. Cf. now Schiering, , Werkstätte orientalisierender Keramik auf Rhodos 22 f.Google Scholar

²²⁰ Ibid. 550 ff.

²²¹ e.g. Mollard-Besques, , Cat. raisonné C126–8, pl. 76.Google Scholar

²²² Reisen ii. 79 f. Brock's report on the towers on the adjacent islet of Pyrgoussa, ibid. 68; Plieninger and Sapper report towers on Pyrgoussa, , Pakhia, , and Strongyli, (Pet. Mitt. 1920, 196 f.).Google Scholar

²²³ Except in list (25–) 26, where it is Nesiotic.

²²⁴ In an inscription of Halasarna, , Paton-Hicks, , Inscriptions of Cos 249 (no. 368, vi. 39), 257.Google Scholar

²²⁵ RE s.v.v Nisyreites, Nisyros (4).

²²⁶ Paton, and Hicks, , Inscriptions of Cos (1891) 285.Google Scholar

²²⁷ These figures are quoted from Hatzeamallos, D., Κῶς, τὸ νησὶ τοῦ Ἱπποκράτους (Athens, 1952).Google Scholar In 1843 Ross estimated the population of the island at 7,000–8,000 (Reisen iii. 126).

²²⁸ The island is shown apportioned among the six demes and the city area in maps (Paton–Hicks, op. cit. at p. 1; Memorie i, in folder); there is of course no absolute certainty that all these demes, and only these, were flourishing at any one time. Cf. Modona, A. N., L'isola di Coo (Memorie i) 22 ff.Google Scholar

Sapper and Plieninger published an excellent map of Cos (see above, p. 118, n. 213); the Italian map used in Memorie i is printed much more legibly in Historia v (1931) at p. 604.

²²⁹ The older name might, in theory, have been Skandarion; but C. Skandarion in fact retained its name, and there is no reason to suppose that it was ever the name of an inhabited site.

²³⁰ Morricone, , Boll. d'Arte 1950, 320 ff.Google Scholar; for the Protogeometric pottery see Desborough, , Protogeometric Pottery 222 ff.Google Scholar Mycenaean tombs have also been reported at Langada somewhere to the south, and Mycenaean finds at the Asklepieion (cf. Morricone, op. cit. 323, 327).

²³¹ The symposium fragment, Clara Rhodos ix. 73 f., pl. 6; the stele and epigram are mentioned ibid. 80, 177.

²³² Ibid. 151 ff.; see below, pp. 124 f..

²³³ AA 1901, 134 ff.

²³⁴ Koische Forschungen (1899) 69, no. 36, pl. 2. 1.

²³⁵ Paton-Hicks, op. cit. no. 148.

²³⁶ Tile fragments, Hellenistic amphora sherds, and sigil lata may be picked up at many points; we found classical coarse wares and black glaze of the first half of the fourth century on the flat crest east of the Asklepieion, but the site was in any case not that of a substantial settlement.

²³⁷ Paus. vi. 14, 12; cf. Lippold, RE s.v. Philotimos (4). The sculptor Pantias seems to be late fifth-century, but the dedication read by Pausanias may have been inscribed later than 366 B.C. See now Amandry, , Charites (Studien zur Altertumswissenschaft) 66, 83.Google Scholar

²³⁸ The mention of 5,000 cities of India, of which none was smaller than Cos Meropis, goes back to one of the companions of Alexander the Great (see Hicks in Paton–Hicks, op. cit. li f.).

²³⁹ Thuc. viii. 41, 2.

²⁴⁰ Id. viii. 44, 3.

²⁴¹ Id. viii. 108, 2. Diodorus (xiii. 42, 3) says that Alcibiades sacked Meropis on this occasion; but Thucydides appears to have had no knowledge of this, and it may be an assumption on the part of Diodorus.

²⁴² Diod. xiii. 69, 5; Xen. Hell. i. 5, 1.

²⁴³ Paton–Hicks, op. cit. xxv f.

²⁴⁴ Adm. chart; Hamilton, Researches ii, map; Ross had the name from Capt. Graves and conjectured that it was the ancient Astypalaia, (Reisen ii. 89, iii. 136)Google Scholar, but subsequently placed Astypalaia elsewhere (iv. 28). Mackenzie seems to have heard the name as ‘Stympalia’; he found it applied at the ancient site; V. Bérard also speaks of the hamlet of Stampalia, (Les Phéniciens i. 38).Google Scholar

²⁴⁵ BSA iv (1897–8) 95 ff. Cf. also Ross's, description (Reisen iv. 23 ff.).Google Scholar

²⁴⁶ Cf. Segre, , Ann. xxii–xxiii. 29.Google Scholar

²⁴⁷ AA 1903, 2 f.

²⁴⁸ Historia v (1931) 625 f.

²⁴⁹ For the references see Memorie i (1933) 19 f.; Modona himself inclined towards Laurenzi's location. The specific identification that he attributes to Paton and Hicks (their p. xix) is a misunderstanding of the English. Cf. Morricone, , Boll. d'Arte 1950, 322Google Scholar, where the position of Astypalaia is said to be unknown.

²⁵⁰ These two photographs, kindly provided (together with that of Plate 25 c) by Dr. Wilder Penfield, were taken from the north; they are taken from slightly different positions and are partly overlapping.

²⁵¹ e.g. bolsals, and plates and stamped bowls of the first half of the fourth century.

²⁵² Cf. especially Mackenzie, , BSA iv. 96 f.Google Scholar; Levi, , Ann. viii–ix. 235 ff.Google Scholar, 310 ff.; cf. also Laurenzi's site (above, p. 122).

²⁵³ Paton–Hicks nos. 420 and 431.

²⁵⁴ Mycenaean chamber tombs have also been reported at Eleona (apparently at the north-east end of the coastal plain of Kardamena=Halasarna), cf. Stubbings, , Mycenaean Pottery from the Levant 21 f.Google Scholar; Morricone, op. cit. 323 f.

²⁵⁵ For this view cf. Morricone, op. cit. 322 f.

²⁵⁶ Paton-Hicks, op. cit. xxvii. Herzog also took this view.

²⁵⁷ On this see Segre, , Clara Rhodos ix. 178.Google Scholar

²⁵⁸ Memorie i. 20.

²⁵⁹ Clara Rhodos ix. 176 ff.

²⁶⁰ Op. cit. 41 f.

²⁶¹ Thus, for instance, burials in the heart of the modern town area seem to have been continuing in the fifth century; and the famous cult of Asklepios has been shown to be no older than the mid-fourth century (Herzog, , Kos, Ergebnisse i. 72Google Scholar; cf. Segre, , Riv. Inst. Arch, vi (1938) 191 ff.).Google Scholar

²⁶² Clara Rhodos ix. 176 f.

²⁶³ Presumably not (as Segre took it) the mints of the cities, but (as Robinson, E. S. G. has pointed out, Hesperia Suppl. viii. 327)Google Scholar the mint at Athens. The various texts making up the decree are now conveniently assembled in ATL ii. 61 ff.

²⁶⁴ Op. cit. 337.

²⁶⁵ For the facts see ATL i. 326; the scattered comments of the editors will be found by reference to their index (vol. iv. 68 s.v. Kos).

²⁶⁶ Cf. also below, p. 142. We have not taken into account the so-called Presbeutikos of Thessalus (Hippocrates, ed. Littré, vol. ix; Hercher's Epistolographi Graeci, Hippocr. xxvii), which relates how Cos refused to submit to the Persians and Artemisia's expedition against the island was repelled by divine intervention; we do not believe that any genuine historical tradition underlies this account.

²⁶⁷ This is in accordance with his normal practice; the city of Rhodes is likewise at the east tip of the island (xiv. 652), and Methymna, being at the point of short crossing to the Asiatic coast, is on the east side of Lesbos (xiii. 616).

²⁶⁸ Pet. Mitt. 1920, pl. 28.

²⁶⁹ ATL i. 372.

²⁷⁰ Ibid. 498; see, however, our observations in BSA l. 155 ff.

²⁷¹ Ibid. 533.

²⁷² BSA l. 155 ff., 164 fig. 14.

²⁷³ JHS xvi. 209.

²⁷⁴ See now Segre, , Ann. xxii–xxiii. 12 ff.Google Scholar; Klaffenbach, , Gnomon 1953, 457.Google Scholar

²⁷⁵ See Travels and Discoveries in the Levant i. 283 ff.

²⁷⁶ (Athens). Unfortunately we have not had access to Flegel's, C.Ἡ νῆσος Κάλυμνος (Constantinople, 1896).Google Scholar

²⁷⁷ Memorie iii. 33–55; Ann. xxii–xxiii (Tituli Calymnii).

²⁷⁸ Near Perakastro and Pothia, Maiuri, , Clara Rhodos i. 106ff.Google Scholar

²⁷⁹ In a cave in the cliff face (Daskalio) above Rina harbour, ibid, 110 ff.; neolithic pottery in the spoil from a well farther up the valley, Brown, Burton, JHS lxvii. 128.Google Scholar

²⁸⁰ At Palionesos and Sykati, Maiuri, op. cit. 105. Maiuri also reports prehistoric stone axes kept by peasants as amulets in Kalymnos and various neighbouring islands, but the sponge-fishers travel far afield and are notorious collectors.

²⁸¹ On the east side of the watershed, BMCat. Vases i. 1, 189 (for the position, Paton, , JHS viii. 446Google Scholar, CR i. 80); apparently also from Damos west of the watershed.

²⁸² In the Daskalio cave, Maiuri, op. cit. 115, with ware of Kamares type; Reïses, op. cit. 38, at Phylakai. The ancient building which goes by the name of Phylakai (cf. Segre, , Memorie iii. 39, n. 7)Google Scholar is in squared masonry with vertical drafting at the angles, and forms a square of c 12·70 by 11·60 m.; it apparently consisted of two main rooms side by side, with a shallow vestibule (roofed with flagstones) in front of each. We noted Hellenistic sherds there, and imagine the building is of that date. The building is backed against the hill slope, and so perhaps was a strong dwelling rather than a fort; in the surviving west part the exterior door is only 0·65 m. wide, whereas that leading into the main room is 1·00 m. wide (cf. the Lelegian farmhouse near Halicarnassus, , BSA l. 132 f.).Google Scholar

This building is on the north edge of the valley midway between Metokhi and Rina; Reïses seems to place Phylakai close to Metokhi and does not mention a building, so is perhaps not referring to the same place.

²⁸³ Travels i. 319 f., where the site has the descriptive name ‘Encremea’; the name is given by Reïses (op. cit. 38) and is in use at the present time. Ross speaks of the circuit as an extensive five- or six-sided peribolos (Reisen ii. 112). A photograph of the wall, Arias, , Historia ix (1935) 30, fig. 8.Google Scholar

²⁸⁴ Cf. Reïses, op. cit. 39, where other remains in this vicinity are remarked (including ruins at a place with the interesting name of Empetassos, which may be added to Segre's list of pre-Hellenic-sounding names surviving on Kalymnos); see also Maiuri, , Clara Rhodos i. 110Google Scholar, and Segre, , Ann. xxii–xxiii. 1 f.Google Scholar, where the various evidences for the old Carian population are discussed.

²⁸⁵ e.g. Newton, op. cit. 307 (archaic marble head = BMCat. Sculpt, i. 1, B323, fig. 193), 308 f. (archaic bronze griffin protome; Jantzen, , Griech. Greifenkessel pl. 41. 1)Google Scholar, 312 (fine red-figured fragments); Segre, op. cit. 7 n. 2, pl. 126 (inscribed archaic sherds); from here also, presumably, the Orientalizing plate fragment, Schiering, , Werkstätte orientalisierender Keramik 64 fig. 4.Google Scholar

²⁸⁶ Newton, op. cit. 310 f.; Segre, op. cit. 148.

²⁸⁷ Cf. Newton, op. cit. 287 ff., 299 ff., 329 ff. Newton dated the finds 330–150 B.C., or later. That, a hundred years ago, was a very remarkable achievement. The recent excavations at Olynthus and in the Athenian agora now make it possible to date with slightly greater precision. We are very greatly indebted to Mr. R. A. Higgins and Mr. P. E. Corbett for arranging for our benefit Newton's finds from Kalymnos in the British Museum; especially to Mr. Corbett for his observations on the Damos pottery, on which the following account is based.

The terra-cotta lamps from Damos start with late examples of Broneer's type VI; there are a number of type VII, one of type VIII, several of type IX, two examples of type XI, and one roughly of type XII (as in Thompson's Agora group E, Hesperia iii. 393 ff.). There are also some lamps of Roman imperial date, but it is not certain that they came from Damos; they may rather be from Pothia (see below, p. 130, n.291). The early lamps from Damos can be matched at Olynthus. Among black-glazed wares two bolsals, a skyphos, a cup-kantharos, a kantharos, and two fish-plates can also be matched at Olynthus; there is also a heavy cup-kotyle and a pair of red-figured askoi (one mentioned Travels i. 300, the other exactly matched at Olynthus). This mass of material points to a commencing date in the second quarter of the fourth century for the series of tombs excavated by Newton at Damos.

Among the Hellenistic pottery there are a number of fusiform unguentarles; some are of the earlier Hellenistic form; the latest examples are more advanced than Thompson's group C 76, though not quite so straight-sided as his group D 78. There are several lagynoi (a form most popular around the second century B.C., see Thompson's discussion, his pp. 450f.); a Megarian bowl of Delian type; a satyr'shead brazier rest found outside the city wall at Damos (examples of the type in Thompson's groups D 76 and E 150). The lamps mentioned above seem to descend towards Thompson's group E. Thompson dated his group D in the mid-second century B.C., and his group E at the end of the second century B.C. and in the early years of the first. On this evidence the burials in the cemetery at Damos seem to have been continuing well into the second century B.C. at least. Newton also excavated a grave at Damos that yielded a Coan coin of the reign of Caracalla.

On the present evidence there may have been a decline in burials at Damos in late Hellenistic times. The cemetery at Pothia in the eastern half of the valley seems to be of late antiquity (below, p. 130, n. 291). So far as it is legitimate to apply the evidence of incompletely excavated cemeteries to the problems of habitations of the living, we might infer a shift of settlement eastward in late times; but this can hardly have occurred as early as the late third century B.C., when the new deme-system was introduced.

Much pottery now in the archaeological collection in the harbour town is said to come from Damos, including Mycenaean and fourth-century vases. Cf. also Ross, , Reisen ii. 99, iv. 8 ff.Google Scholar; Arias, , Historia ix. 28, figs. 5–6.Google Scholar

²⁸⁸ Newton, op. cit. 286 f., 301 f.; cf. Ross, op. cit. ii. 99 (many foundations of ancient buildings and many coins), Reïses, op. cit. 59 (traces of a very strong fortification, many and great ruins). Ross also surmised that here (as in Vathy) there was an ancient town and believed this to be the principal one on the island; but, not knowing of the fortifications at Damos, he assumed that the habitation extended along the valley (op. cit. 110).

²⁸⁹ Gerola, , Ann. ii. 55 ff.Google Scholar

²⁹⁰ Cf. Ross, op. cit. 100, Newton, Travels i. 227, 314Google Scholar; Dubois, , BCH viii (1884) 41Google Scholar, speaks of ancient substructures, especially at the point where the old church of the Panagia stood. Reises speaks of these remains and cites the judgement of his compatriot Kallisperes that the Ionic order of the temple was not inferior to that of the Erechtheum in Athens (op. cit. 28); they concluded that the ancient city of Pothaea lay here, but we have not found evidence to support their view, which seems to be founded in large part on local pride.

²⁹¹ Late Roman and Byzantine lamps and vases in graves at Pothia near the harbour (Newton, , Excavations at Calymnos 23 f.Google Scholar (= Arch. Journal xiii (1856) 36 f.)); this cemetery seems to have been in the valley opposite Perakastro.

²⁹² Ross, op. cit. 94; Newton, , Travels i. 316Google Scholar; Reïses, op. cit. 17; Ann. xxii–xxiii, no. 117.

²⁹³ CR viii. 376; the plan, JHS xviii. 213, fig. 5; Reïses, op. cit. 47, with reference also to Flegel; photographs, Arias, , Historia ix. 30 ff.Google Scholar, figs, n–13 (gems and bone relief, ibid. figs. 14–15).

²⁹⁴ Ross, Reisen ii. 102; Dubois, loc. cit.; Reïses, op. cit. 46 ff. (with mention of Cyclopean fortifications at Vryokastro above Arginonda); Segre, , Memorie iii. 39, n. 5.Google Scholar

²⁹⁵ A stretch of walling at Kastelli, , Gerola, , Ann. ii. 59, fig. 53.Google Scholar Arias, op. cit. 29 f., figs. 9–10, describes an acropolis with dry rubble masonry hereabouts, and calls it ‘Anghinarie’ (the name does not seem to be otherwise known, and does not occur in Reïses); Segre, (Ann. xxii–xxiii. 1)Google Scholar equates it with Kastri above Emborios, but Arias says that the Kastri fort is visible from the other and to the north-west of it.

²⁹⁶ Travels i. 317; cf. also Maiuri, , Ann. viii–ix. 323Google Scholar, Segre, , Memorie iii. 34 f.Google Scholar

²⁹⁷ Ross, op. cit. 101; Newton, op. cit. 318; Reises, op. cit. 54; Segre, , Ann. xxii–xxiii. 219.Google Scholar

²⁹⁸ Cf. Paton, and Hicks, , Inscriptions of Cos 354Google Scholar; most recently Segre, , Memorie iii. 39Google Scholar, Ann. xxii–xxiii. 41 f. To relate the three demes to Pliny's notices of Calymna in the way that Segre does (cf. Ann. xxii–xxiii. 31) seems to us misleading. In NH v. 133 the names of the three towns are in fact given by Pliny (Calydne (vel sim.) cum III oppidis, Notio, Nisyro, Mendetero(s)); they bear no resemblance to the known deme names of Calymna; possibly—assuming that the reference is in fact to Calymna—Pliny may have misconstrued a Greek source which (referring to Cos and its demes) stated that Calydne was another deme (of Cos) and that to the south (of Cos) lay Nisyros. In NH iv. 71 the name of the single town is also given (Calymna, in qua oppidum Coos (vel sim.), where the meaning of Pliny's source may have been that Calymna was a deme of Cos). Fortunately the threefold division of Calymna under Coan rule seems to be so securely established on the epigraphical evidence that Pliny need not be used to support it.

²⁹⁹ Delatte, , Les Portulans grecs 306Google Scholar

³⁰⁰ This fortress is said to have been founded by survivors from the destruction of Rina in the ninth century and to have been abandoned later when the population removed to the kastro above the watershed in the southern valley. The kostro was fully inhabited until the population moved down to the khora after the Greek War of Independence (Newton, , Travels i. 296Google Scholar) and so cannot come into consideration here.

³⁰¹ Reisen ii. 101; Travels i. 317.

³⁰² Cf. above, p. 130, n. 296.

³⁰³ Ann. xxii–xxiii. 139.

³⁰⁴ Cf. the theatres of the Coan demes, as Isthmos and Halasarna.

³⁰⁵ Segre, , Ann. xxii–xxiii. 42Google Scholar, gives it to Panormus; but it may further be remarked that in the reorganization of the deme system the new name Panormus is not likely to have been given to a deme whose centre was so far inland.

³⁰⁶ Travels i. 285, 296. In 1841 Ross noted fifty to sixty houses and a leper station there; he did not give a name to the place (Reisen ii. 93). In the second half of the century it was growing steadily, as Newton predicted, into the domi nant place in the island.

³⁰⁷ See above, p. 129, n. 287.

³⁰⁸ CR i (1887) 80.

³⁰⁹ Newton was told that there was a place called ‘Potha’ on Telendos (op. cit. 315), but this can hardly affect the present issue. If Ποθ- in the name Pothaea is the Doric form of Προσθ-, the ancient deme-name may also be preserved in the western half of the southern valley, which in Ross's time was called τὰ ᾿Μπροσθά, (Reisen ii. 96)Google Scholar and is called by Reïses (op. cit. 25, 55); Segre also speaks of a vicus Brostà there (loc. cit.) and Newton of a cave called (op. cit. 316).

³¹⁰ Loc. cit.

³¹¹ e.g. op. cit., inscriptions nos. 7. 15; 32. 9; 79 A 6, 9, 15; 79 B 36. The mentions in no. 11.9 and no. 31.34 are obscure, and Segre remarks that his restorations are exempli gratia.

³¹² Segre's note (op. cit. 85) gives the impression that Herzog believed the and to refer to Calymna. But Herzog, (Klio ii. 319)Google Scholar took them to refer to Cos, and the to be Calymna and its adjacent islands (cf. the mention of and (= Cos) in the Coan decree relating to the same emergency (Segre, , Riv. Fil. lxi. 365)Google Scholar, and that of Halasarna, SIG ³ 569. 19). The Calymnian inscription (no. 64) then goes on to relate that the (Coan) admiral decided to meet the armada from Crete and the battle (in which Calymnians took part) was fought off Cape Laceter. Segre had in fact previously recognized that this was so (op. cit. 378), but he does not seem to have perceived its necessary implication for the objectives of the expedition; the Cretan fleet was not heading in the direction of Calymna and its islands, but steering a course round the other side of Cos, which must inevitably lead to the city of Cos at Skandarion. For the position of Cape Laceter see above, p. 126.

³¹³ Segre, Ann. xxii–xxiii, no. 79, p. 121. The Cnidian arbitration of the ensuing claim dates to the early years of the third century, and Segre dates the loan c. 360 B.C.; he remarks that it is quite unknown what this public work was which involved an expenditure beyond the resources of the state, but suggests that it may have been at the sanctuary of Apollo. So far as we are aware, there is no evidence of constructional work at the sanctuary in the fourth century; and we prefer to suppose that the new town-sites were the object.

³¹⁴ Cf. Reïses, op. cit. 10.

³¹⁵ Cf. FOA xii, p. 2, n. 1; ATL i. 513; BSA 1. 160 f. For the forms of the ethnic in the lists (read as ) cf. ATL i. 332 f. Steph. Byz. s.v. gives

³¹⁶ Cf. Paton's, remarks, CR viii. 375 ff.Google Scholar

³¹⁷ Die Insel Leros (1898). A description of the island, with some views and a copy of the Italian large-scale map, is also given by Bertonelli, F., L'Universo Nov. 1928, 1067–83.Google Scholar

³¹⁸ Athen. 655 c; Ael. NA iv. 42; Antonin. Liberal. 2; Phot.-Suid. s.v. On the finding of the two decrees, Ross, , Reisen ii. 121Google Scholar, Paton, op. cit. 376, Bürchner, , AM xxi. 33Google Scholar, Die Insel Leros 36 f. References in documents of Patmos (dated A.D. 1087 and later) and in the Greek portolan III (Delatte, op. cit. 306) guarantee that the name Partheni is a survival.

³¹⁹ Cf. Bürchner, , Die Insel Leros 15.Google Scholar

³²⁰ BSA xii. 172.

³²¹ In his book (1888), to which we have not had access; Ross mentioned these remains without having seen them (Reisen ii. 122).

³²² For other remains in the north of the island see Bürchner, op. cit. 16 ff.

³²³ Op. cit. 172 ff., figs. 17–18. There appears to have been considerable damage to the fortifications since their visit. We could not recognize the line of the curtain on the north side, but their plan shows it clearly. In front of the chapel of the Panagia on the summit there are remains, presumably of an earlier Christian church. The mosaic floor was in black, red, white, yellow, and grey; it had a border of ivy scroll, and geometrical panels within this. Cf. Ross, op. cit. 118.

³²⁴ For the kastro see Gerola, , Ann. ii. 61 ff.Google Scholar

³²⁵ Ross, op. cit. 119; Bürchner, op. cit. 32; Rehm, , Milet ii. 2, 25.Google Scholar

³²⁶ Cf. his article in 2, Dec. 1956.

³²⁷ These were unfortunately chiselled away by the occupying authorities.

³²⁸ Mr. Samarkos in his article cites the following from the priest's garden: (the last of these, which we saw, is in early Hellenistic stoichedon). Other inscriptions apparently from the region of the modern town are the base ἥρως, Ἀνδρίδης (Clara Rhodos ii. 235, no. 137Google Scholar, on the kostro) and another a late epitaph of one Tryphaena, and perhaps the epitaph Ann. i Notiz. 318. It is not clear whether the phrase cited above constitutes an epigraphical text.

³²⁹ AE 1915, 133 no. 12 (then built into a wall by the harbour of the island, where were traces of an ancient building).

³³⁰ Antiochus, son of Aeschines, being stephanephoros of Miletus in that year (Milet i. 3, no. 126).

³³¹ See below, p. 137, no. 4. Another phrourarch dedication, perhaps of 70/69 B.C., Sakkelion, , AE 1890, 221.Google Scholar

³³² Quoted by I. Paradeisos in the peculiar form The epithet is of course inappropriate in this context; our informant appeared to be confident of the reading of the name Artemis.

³³³ Haussoullier, , Rev. phil. 1902, 125ff.Google Scholar (non vidimus); Rehm, , Milet ii. 2, 22 f.Google Scholar

³³⁴ Ap. Strabo xiv. 635.

³³⁵ Lepsia at least must be included in the Milesian islands of which L. Robert has found mention in a decree of Miletus and another of Aptera, (Hellenica i. 113 ff.)Google Scholar; but Rehm's, doubts (Milet ii. 2, 20, and 26)Google Scholar and Robert's statement that Miletus laid her hand on Lepsia, Patmos, Leros, and Aigiale (in Amorgos) in Hellenistic, times (Ét. num. gr. 11)Google Scholar are incompatible with the literary and epigraphical evidence for Leros, for which see Bürchner, , Die Insel Leros 30 ff.Google Scholar

³³⁶ xiii. 611; cf. our discussion in BSA l. 143ff.

³³⁷ Cf. BSA l. 91 (with plan, fig. 2).

³³⁸ Milet iii. 2, 51. This dating is presumably based on the theory that Priene was refounded by the Athenians in the mid-fourth century B.C. at the time when they were on friendly terms with the arch-rebel Orontes, who may possibly at some time have gained possession of Ionia as far south as the Maeander. Hence the assumption that Orontes' reconciliation with Artaxerxes Ochos will have caused the suspension of work at Priene. The resemblance of the fortifications of Heracleia with those of Lysimachian Ephesus, which Krischen also remarks, is less striking.

³³⁹ For Pleistarchus cf. Robert, , Inscr. Sinuri 57 ff.Google Scholar

³⁴⁰ Thus Robert, with due caution, suggests Eupolemus as the author of the fortifications at Heracleia (loc. cit.); von Gerkan, , Griech. Städteanlagen 16, n. 2Google Scholar, and Magie, , Roman Rule in Asia Minor 918Google Scholar, suggest Lysimachus.

³⁴¹ Scranton's definition of them as ‘pseudo-isodomic’, coupled with his acceptance of Krischen's historical dating, seems to us misleading (Greek Walls 134, 181 f.), and we should rather not apply the term here. If, on the other hand, Scranton's classification of them is retained, it seems as though the commencing date of this ‘pseudo-isodomic, quarry to hammer face’ class must be raised. The treatment of the face is the normal classical one; and since on historical grounds the fortifications of Philippi also are more naturally to be dated in the middle of the fourth century, it seems in fact that mid-fourth-century dates actually preponderate in this class.

³⁴² Bean, , JHS lxxiii. 20, nos. 3–4.Google Scholar

³⁴³ Op. cit. 12.

³⁴⁴ Cf. BSA l. 108 ff., 145, n. 240.

³⁴⁵ Inland from this was the hyperakrian Pidasa, perhaps at Palamutlu Kale on a high summit of the Grion range.

³⁴⁶ For the testimonia for Pytheos see Hiller, , Inschr. v. Prime 204, no. 465.Google Scholar

³⁴⁷ Ps.-Arist., Oecon. ii. 1348 a.

³⁴⁸ For Eurome see especially Laumonier, , RA 1933, ii. 40 f.Google Scholar The history of the territory called the Euromos in Hellenistic times is one of flux (sympolities made and unmade, dispute over the cities in the Euromos, and over the possession of land there by Pidasians and Latmians). The use of the name to denote a region considerably greater than the territory of the city of Eurome might possibly be explained by an earlier, abortive synoecism there. The change of name from to if it is a ‘euphemistic’ change like that of Syagela–Theangela (for the extraordinary complexity of the problem see Robert, , Hellenica viii. 35 ff.)Google Scholar, would most naturally fit with Mausolus' hellenizing policy.

³⁴⁹ For the site, at Karacahisar, see Robert, , AJA xxxix. 339 f., pl. 39Google Scholar; the fortifications shown seem to be fourth-century or Hellenistic.

³⁵⁰ For references see BSA l. 135.

³⁵¹ Ibid.

³⁵² Diod. xv. 76, 2: Strabo xiv. 657. See our discussion of its problems above, pp. 120 ff.

³⁵³ In Diod. xvi. 7, 3 (where it is clear that Diodorus had no conception of the significance of Mausolus' participation in the war) the Coans are named among the revolting states; but the contemporary authorities speak as though only Chios, Rhodes, and Byzantium were in revolt (Isocr., Peace 16, Antid. 63; Dem. Rhod. Lib. 3). Cf. Judeich, , Kleinasiatische Studien 283, n. 1Google Scholar: ‘Wenn Diodor an der ersten Stelle die Koer mit unter den “abfallenden Bundesgenossen” aufzählt und 21, 1 noch einmal dieselbe Phrase aufnimmt, so wird man aus dieser einen Angabe schwerlich, wie das bisher gewöhnlich geschehen 1st (vgl. u. a. Schaefer Dem. I² 166 f.), auf die Zugehörigkeit von Kos zum zweiten attischen Seebund, über die wir sonst nichts wissen, schlieβen können. An der dritten Stelle 21, 2 werden von Diodor in Übereinstimmung mit Isokrates, Demosthenes und Trogus ganz richtig genannt nur ’

³⁵⁴ CMAI 1953. 411–13.

³⁵⁵ BCH lviii. 298. It is worth noting that Laumonier had already remarked the resemblance between the terraces of Labranda, , Alinda, , and Pergamon, (RA 1933, ii. 44).Google Scholar

³⁵⁶ Op. cit. 413. For Alinda see the plan, Trémaux, Exploration arch, en Asie min. Alinda, pl. 1 (cf. Perrot-Chipiez, Histoire de l'art v, fig. 224); Lebas, , Voyage arch. pl. 62.Google Scholar Cf. JHS xvi. 240, where Paton notes close resemblances with Çömlekçi Kalesi (Asardağ = Kildara?, above, p. 99); Laumonier, op. cit. 298, fig. 10. The Praxitelean Aphrodite at Alexandria ad Latmum (Steph. Byz. s.v. ) may possibly be regarded as a further indication of the status of Alinda in the fourth century.

³⁵⁷ Martin, R., L'Urbanisme dans la Grèce antique (1956), 149 f.Google Scholar

³⁵⁸ Op. cit. 148. The great platform, which Martin (following Newton) supposes to be the ‘Temple of Mars’ and which he has made the culminating point of his terraces, is not (as Vitruvius says of the temple) in summa arce media, but in the flat ground (cf. BSA l. 87 f.).

³⁵⁹ For the testimonia see BSA l. 147–9.

³⁶⁰ BSA l. 113, 146, pl. 16 c; and above p. 93, Fig. 8.

³⁶¹ The photograph by Mr. R. V. Nicholls. The position of the tomb is shown approximately in our plan, BSA l. 117 fig. 5, under the west end of the citadel.

³⁶² For the underground gallery and chambers on the south side of the Mausoleum see Newton, , Halicarnassus ii. 127 f., 147 f.Google Scholar; cf. also BSA l. 94 f.

³⁶³ The chamber tumulus at Burgaz (Paton, and Myres, , JHS xvi. 246 f.Google Scholar, cf. BSA l. 166) and perhaps also one at Gürice, (BSA l. 121)Google Scholar might be explained as dynastic if (as we have assumed on other grounds) the dynasty of Termera moved farther north in the peninsula during the second half of the fifth century (ibid. 148).

³⁶⁴ On this see Paton and Myres, op. cit. 196.

³⁶⁵ Cf. BSA l. 143 ff.

³⁶⁶ BMCoins Caria lxi f.; Head, HN ² 621.

³⁶⁷ See AJP 1937, 390 f.; ATL i. 156.

³⁶⁸ Contrast in col. i, lines 113–14 of the same list

³⁶⁹ Robert has drawn attention, in this connexion, to the name (gen.) in his fourth-century inscription of Kildara, (Hellenica viii. 14).Google Scholar

³⁷⁰ Meritt, and West, , Athenian Assessment 21, 71 f.Google Scholar; cf. ATL i. 114 fig. 161.

³⁷¹ Cf., e.g., (list 2), (list 1), (list 23). For Tymnessus see now Robert, , Hellenica x. 189 ff.Google Scholar

³⁷² e.g. Nikon (near Yenice, see above, p. 71) and Aridolis (tyrant of the Carian Alabanda, Hdt. vii. 195) may have been established as dynasts.