¹ Paton, and Myres, , Geogr. Journal 1897, 44Google Scholar; Philippson, , Reisen u. Forschungen V 51Google Scholar, map, denies the presence of gneiss and marks andesite in this area.

² For the plotting of positions in the interior of the Myndos peninsula we have in general regarded Paton, and Myres', map (JHS XVI, pl. 11)Google Scholar as the most accurate; for the 200-metre contour we have attempted to combine Philippson's indications with those on the Turkish map, but refer the reader to Paton and Myres' map for a truer impression of the relief. Paton and Myres did not undertake an accurate survey to the east of the Myndos peninsula, and we have therefore followed the Turkish map in this part of our sketch map.

³ Paton, W. R. and Myres, J. L., ‘Carian Sites and inscriptions’, JHS XVI (1896), 188–271, pls. 9–11.CrossRefGoogle Scholar

⁴ By the generosity of Sir John Myres we were enabled to carry a copy of ‘Carian Sites and inscriptions’ with us in the field, and subsequently had his notes and correspondence with Paton at our disposal. We take this opportunity of mentioning also with gratitude those who have helped us in the field, notably Osman Bilgin and Ahmet Davas, and Mrs. J. M. Cook, Miss M. Bean, Mr. R. V. Nicholls, and Mr. W. C. Brice, who accompanied us on some of the shorter journeys and have helped with the illustrations in this article; also, in addition to those named below, Dr. M. Mitsos and Mr. D. M. Lewis for assistance with inscriptions in Athens, Mr. I. Kondis, who gave us access to Biliotti's field notebook of 1865, and Mr. B. Ashmole for allowing us to refer to unpublished objects from Halicarnassus in the British Museum.

⁵ Ps.-Scylax 98a; Strabo XIV 656 f.; Pliny, NH II 204Google Scholar; Arrian, Anab. I 23, 3Google Scholar; Vitruvius II 8, 10–14; Steph. Byz. s.v. ῾῾Αλικαρνασσὸς.᾿

⁶ By Artemisia's canal (Vitruvius loc. cit.), Alexander's τὰφρος ἀξιὀλογος (Diod. XVII 27, 6), the fosse of the knights in A.D. 1476 (cf. Ann. IV–V 317); cf. also the ποταμὸς of Ps.-Scylax loc. cit. (p. 89).

⁷ Vitruvius loc. cit. Cf. also the distinction of the ὰκραν τὴν ὲν τῆ νὴσω̣ and Salmacis in Arrian, loc. cit., where the Persians withdrew to these two forts.

⁸ For the castle see Newton, , Halicarnassus II 73 f.Google Scholar, I, pls. 32–38; AA 1919, 59 ff.; Ann. IV–V 290 ff., pl. 5.

⁹ The enthusiasm which he displays in his description and the singular irrelevance to the matter in hand (crude brick construction) are signs of an extraordinary personal interest, and almost suggest that the memory is his own.

¹⁰ The plan Fig. 2 is based on Admiralty Chart no. 1606 and Newton, , Halicarnassus I, pl. 1Google Scholar; some towers and jogs have been added in the wall circuit, the modern habitational network has been omitted, and legends have been altered to fit with our views.

¹¹ Reisen IV 36 f.

¹² Travels in Asia Minor II 32.

¹³ Reisen IV 33.

¹⁴ Halicarnassus II 312 ff. Cf. Ross, , Reisen IV 33.Google Scholar

¹⁵ Halicarnassus II 270.

¹⁶ For the connection of Hermaphroditus with the nymph Salmacis see Ovid, , Met. IV 285 ff.Google Scholar

¹⁷ The ruined tower now standing there is not ancient.

¹⁸ In one of Gell's field notebooks now in the possession of the British School at Athens (BSA XXVIII 115)Google Scholar: ‘at Budrun Bey's gate (inscription ibid. 126, no. 16) gave the Bey 3 okes of coffee, 2 loaves of Sugar. The fount Salmacis lost to the people being under water at the old decayed mole. L. of entrance. It boils up.’ The loss of the spring can perhaps be accounted for by the general subsidence of the west coast of Asia Minor which is manifest in many places and can be calculated at m. since classical times.

¹⁹ A good view of the arsenal promontory is given Ann. IV–V 291, fig. 9.

²⁰ 98a. We interpret this sequence as referring to the main harbour, the bay off Kumbahçe, and a canal connecting the two. For ποταμὸς in this sense cf. LS ⁹; there is no stream worth mention near Halicarnassus.

²¹ Cf. p. 87, n. 6.

²² See also Admiralty Chart 1606.

²³ As Ross, , Reisen IV, plan opp. p. 39.Google Scholar The sub montibus latens of Vitruvius is unintelligible; it is normally emended to sub moenibus latens. The now submerged walling might rather be the west boundary of Mausolus' palace.

²⁴ Newton, 269, noted some large blocks a little distance inland from the middle of Kumbahçe bay, and therefore assumed that the circuit continued on the same axis to the sea.

²⁵ Spratt apud Newton (279); ÖJh VI 101.

²⁶ Reported in AA 1913, 476.

²⁷ 15 × 12·;30 m. as against a norm of c. 8 m.

²⁸ Mr. P. E. Corbett, who has examined these fragments, kindly informs us that such krater handles are usually to be dated in the first quarter of the fourth century, but that they continue, with carelessly drawn decoration (as here), into the second quarter. The trapezoidal work of the tower also seems a pre-Hellenistic mode (cf. Scranton, Greek Walls 167 ff.).

²⁹ Arrian, , Anab. I 20–23Google Scholar; Diod. XVII 24–27.

³⁰ The proper form of this name is in doubt, cf. Head, HN ² 630; RE s.v.

³¹ Cf. also Curtius, , Hist. Alex. VIII 1, 36, V 2, 5.Google Scholar

³² Cf. also Strabo XIV 656, 635. The Mausoleum, however, remained undamaged, as also, apparently, the brick and marble palace of Mausolus (Vitruvius, loc. cit., Pliny, NH XXXV 172).Google Scholar

³³ The modern motor road keeps to the east side of this crest to avoid the steep descent into the valley.

³⁴ It is perhaps worth remark, though not in any way decisive, that in Arrian's account (I 23, 5) Alexander, on entering the city after the Persians' withdrawal, looked down on the ‘island’ fort and Salmacis (cf. p. 87). This is one of the passages which disproves the recurring misconception that the ‘island’ to which the Persians withdrew was Arconnesus (the modern Karaada).

³⁵ Diod. XVII 24, 1 implies that Alexander sent his engines and corn by sea to Halicarnassus. But Arrian (I 20, 1) states that Alexander at this point dismissed his fleet; the Persian triremes ὲφὼρμουν τῶ̣ λιμὲνι the Persians were able to send reinforcements by sea from Halicarnassus to Myndus, and at the end of the siege their fleet effected the withdrawal to Cos (Diod. XVII 27, 5). The supplies from Miletus may in fact have been shipped to Torba Bay and thence brought overland to Yokuşbaşi by the Mausolan road (p. 131).

³⁶ Newton marks a small gate here, from which a narrow mountain path leads up the glen to Gökçeler (p. 123, ancient Pedasa); the easier, though longer, route by Çirkan branches from the Myndus road.

³⁷ See the plan, Newton I, pl. 73, where the walling in the pylon may be partly conjectural.

³⁸ Turner in 1815 counted thirty-two rows of seats (Journal of a Tour in the Levant III 55).

³⁹ Halicarnassus II 276 f.

⁴⁰ Halicarnassus 319 ff., pl. 48.

⁴¹ The pedestal Plate 13 (d) is one of a pair, mirror twins, in blue limestone, dug up by Newton (324 f.) and now built into the walls of a house. H. 1·24, width 0·48, front to back on moulding 0·44.; plain on top. A fragment with a fillet ornamented with rosettes, perhaps from another such pedestal, is built into a neighbouring house. Newton (270 note e) reported another limestone block, with a shield in relief and a triglyph, at the konak on the waterfront.

^41a For the mosaics see Hinks, BMC Paintings 125 ff.

⁴² At house no. 10 in Eski Çeşme: a headless statue of a draped female figure; white marble; pres. H. 0·70 m. In the cellar of the primary school at Bodrum, from a house in Eski Çeşme: (a) the inscribed relief, p. 99, no. 4 (Plate 12 (c)), (b) statue of Marsyas in the round engaged against a tree-trunk (Plate 12 (d–e)); white marble, in one piece; socket in top of trunk near front edge (0·045 m. diam., 0·;05 m. deep); overall H. 0·92 m., H. of figure 0·67 m.; fingers of right hand missing, slight damage to feet and pubes. Traces of red paint on the body. The thong is slung over a fork in the bole of the tree. The eyelids are heavy and the right one droops; and the modelling of the torso is shallow. The figure is slightly dwarfish and, despite its presumed descent from the Pergamene tradition, mild and undramatic in aspect. For the types of the Hanging Marsyas. cf. Stuart Jones, Cat. Conservatori 165 ff. The counterfeit red paint (which is also found on another white marble example in Kos Museum) seems to blur the sharp distinction between ‘red’ and ‘white’ types of the Hanging Marsyas.

⁴³ Cf. Vatican Cat., Braccio nuovo I, pl. 21 no. 127. A cross-legged seated barbarian from Halicarnassus, AM XLV, pl. 4, 1Google Scholar; Ann. IV–V 273 fig. 1; Möbius, AM L 45, takes it to be a slave from the Mausoleum statues (cf. Lippold Die griech. Plastik 256 n. 10 and Buschor, Maussollos 39).

⁴⁴ Travels in Asia Minor II 34.

⁴⁵ As in the fifth-century inscription SIG ³ 46.

⁴⁶ Plate 14 (a) 5, fragment with glaze bands on a pale slipped surface, near Geometric; 7, lip of cup or kantharos in soft pale ware with red glaze stripe and running circle and tangent (?) decoration; 6, black-glazed kotyle foot of Attic type; Plate 13 (c) 2, interior of bowl with stamped ovolo and palmette of first half of fourth century. A few other bits of archaic striped ware were found.

⁴⁷ Plate 14 (a) 8, rim of Attic-type bell-krater with R.F. laurel spray. The spoil of the counterscarp of the castle is likely to have come from an already disturbed area, since both Artemisia the Younger and Alexander are said to have dug through the isthmus (cf. p. 87, n. 6).

⁴⁸ AA 1919, 61.

⁴⁹ Ann. IV–V 328, marked on pl. 5.

⁵⁰ SEG IV, no. 191 (assuming αὺλὴ to be correctly read by Maiuri); a photograph AM XLV, pl. 4, 2.

⁵¹ Fragment of cup with impressed ovolo and palmette, early fourth century.

⁵² Striped fragments, including an upright rim of an open bowl with decoration on a white slip outside and a white band on glaze inside (Plate 14 (a) 3); fragments of a small relief pithos (Plate 13 (e) 9) and plain crock with incised handle (Plate 13 (e) 7). The fifth-century lamp rim (Plate 14 (a) 2) was found on the peak at Kaplankalesi.

⁵³ SIG ³ 45 = Tod, GHI ² no. 25.

⁵⁴ Vitruvius II 8, 12.

⁵⁵ A fragment from the neck and handle of a jar found at the counterscarp resembles in fabric and form the prehistoric ware from Erenmezarlik (p. 118).

⁵⁶ Halicarnassus II 325 ff., I, pls. 46–7.

⁵⁷ II 124 f., 154 f. Mr. R. A. Higgins dates the head ibid. 124 to the late fifth century. At Kislelik also on the east of the town Newton opened sarcophagi (334 ff.) containing vases about which Mr. P. E. Corbett of the British Museum has kindly informed us as follows: ‘The vase mentioned by Newton, , Halicarnassus II 335Google Scholar, is the late Attic pelike E 428; the two vases on p. 336 are a second pelike F 14, and a light cup-kotyle, 57. 12–20. 226. The two pelikai are pretty well contemporary, to judge by their shape; E 428 must be of much the same date as Schefold, Untersuchungen, nos. 474 and 514, which are shown to be earlier than 350 B.C. by comparison with Olynthus XIII, pl. 64–65, 50. Though F 14 is much corroded the main scene can be recognised as a fight between a mounted Oriental and a Greek: the mounted figure may be compared with Olynthus XIII, pl. 45. The two figures on the reverse of F 14 seem contemporary with the cloaked youths of Olynthus XIII, pl. 38, no. 28. Thus the date of both figured vases is about 360 B.C. or perhaps a little after. The form and decoration of the cup-kotyle suggest a date in the third quarter of the century, rather than the second, though not too long after the middle of the century’. The pelike E 428 was found in a sarcophagus, together with a fifth-century silver tetrobol of Chios. The pelike F 14 and the cup-kotyle were found in another sarcophagus. These burials seem therefore to date from the first decades of the Mausolan city; they appear to lie outside the main east circuit, but it is not clear in Newton's description whether or not they would lie outside an assumed course of the ‘exterior wall’ in this sector (see above, p. 90).

⁵⁸ Mr. Higgins kindly informs us that the rude horseman BMC Terracottas 92 (B 118) and a number of other similar pieces found by Newton in the sepulchral chamber of the Mausoleum, site (Halicarnassus II 147)Google Scholar are to be dated somewhere in the sixth century B.C. For the Halicarnassus terracottas see now Higgins, , BMC Terracottas I (1954), 102.Google Scholar

⁵⁹ Ashmole, Festschrift A. Rumpfs 5 ff., pls. 1–2.

⁶⁰ AA 1919, 70, fig. 6; its position appears in Newton, Halicarnassus I, pl. 35, 1; see also Markham, A. H., Budrum Castle (1904), pl. 6.Google Scholar For archaic sculptures from Halicarnassus, cf. Lippold, Die griech. Plastik 66.

⁶¹ The capital has since been moved into the adjacent municipal garden.

⁶² Plate 13 (e) 9 (p. 93, n. 52), and Louvre, , Courby, , Vases grecs à reliefs 85, pl. 3aGoogle Scholar, Feytmans, , BCH LXXIV 162 (Halicarnassus)Google Scholar; JHS VIII 71, fig. 10 (Asarlik, together with fragments of sarcophagi with relief decoration, BMC Vases I 1, 213 ff., figs. 300–3); JHS VIII 79, fig. 26 (Mandrais); Plate 13 (e) 8 (Gökbel, p. 135). To judge by the meagre remains that we have seen the Halicarnassian relief pithoi seem to be smaller and thinner-walled than those of Rhodes, and Cnidus, (BCH LXXIV 135 ff.)Google Scholar and to be devoid of figured decoration.

⁶³ E.g. the early fifth century female figurine, Louvre CA 235, found in the digging of a well at Bodrum, now Mollard-Besques, , Cat. raisonné des figurines, B 338, pl. 36Google Scholar, and ibid. B 339.

⁶⁴ For a more detailed account of its history see Newton, Halicarnassus II, chapters I–II.

⁶⁵ Strabo XIV 656; Paus. II 30,9; Steph. Byz. s.v. ῾῾Αλικαρνασσὸς ᾿ Cf. CIG 2655; Michel no. 452. Also Argives, Vitr. I 8, 12; Mela I 16, 3; cf. also Paus. II 30, 10, where with the descent of the Heraclidae Dorians from Argos were received into citizenship at Troezen.

⁶⁶ Tacitus, A. IV 55.Google Scholar The claim, put forward in A.D. 26, was that the city had stood for 1200 years and never suffered from an earthquake. It is likely that the age of the city was computed on a chronological system linked to the Troica and that a high dating was adopted to impress the Senate; it does not therefore follow that the foundation was regarded as anterior to the descent of the Heraclidae.

⁶⁷ Strabo XIV 653, (after the death of Codrus) Κνὶδος μὲν γὰρ καὶ ῾Αλικαρνασὸς οὺδ᾿ὴν πω, ῾ὸδος δ᾿ ὴν καὶ Κῶς This admittedly conflicts with the tradition that Anthes left Troezen when the sons of Pelops came from Pisatis (cf. Strabo VIII 374), but is supported by Pausanias' πολλοῑς ἔτεσιν ὕστερον (II 30, 9). It is notable that Halicarnassus is not mentioned in Diodorus' account of the Dorian foundations in the Southern Sporades and Caria (V 53 ff.).

⁶⁸ Hdt VII 99: cf. also Suidas s.vv. ῾ ῾Ηρὸδ οτος ᾿,῾Πανὺασις ᾿

⁶⁹ Perhaps with the King's Peace in 386 B.C.

⁷⁰ Cf. Tarn, , Alexander II 218Google Scholar, where three ancient authors are incorrectly cited as adhering to his view.

⁷¹ Ptolemy attacked Halicarnassus in 309 B.C., but was compelled to raise the siege by Demetrius Poliorcetes (Plut. Demetr. 7).

⁷² SIG ³ 46.

⁷³ SIG ³ 45.

⁷⁴ Cf. Paton-Hicks, Inscriptions of Cos p. xvii. The reputed Argive settlement at Iasus is said by Polybius (XVI 12, 2 f.) to have taken reinforcements under the son of Neileus from Miletus to make good its losses in war with the Carians; and it is not unlikely that Halicarnassus also received Ionic immigrants; a legend (Parthenius XIV) shows a Halicarnassian youth of the royal family held as a hostage by the ruling house at Miletus.

⁷⁵ History of Greece III 275.

⁷⁶ JHS XVI 204.

⁷⁷ ATL I 538.

⁷⁸ Alexander II 218.

⁷⁹ This case is in fact more arguable if one takes into account the ancient testimonies that both the Mausoleum and the palace of Mausolus were standing intact in later times. Cf. p. 91, n. 32.

⁸⁰ SIG ³ 46, 2–3.

⁸¹ Cf. the elegiac dedication of Panamyes (p. 93).

⁸² Cf. BMC Coins, Caria, pl. 18, Head, HN ² 617 f. It is worth noting in this connection that the series of late archaic and classical terracotta figurines from Halicarnassus (nn. 58 and 63) are thoroughly Greek and seem on the present evidence to be products of a local fabric.

⁸³ SIG ³ 45 = Tod GHI ² no. 25.

⁸⁴ Strabo XIV 659. Cf. Newton II 30 f.

⁸⁵ Herodotus (V 118) names Pixodarus, son of Maussollus, of Cindya, who was married to a daughter of the Syennesis of Cilicia, c. 497 B.C. Pixodarus and Maussollus are two of the three Hecatomnid male names; the known male names in Artemisia's family are Pisindelis, Lygdamis and Apollonides. In Suidas' entry s.v. ῾Πὶγρης ᾿ there seems to be a confusion between the two Artemisias, since Pigres can hardly have been a brother of the later one, whereas the unqualified phrase Μαυσὼλου γυναικὸς can only refer to the wife of the famous Mausolus; the phrase τῆς ὲν τοῖς πολὲμοις διαφανοῦς could well be referred to either Artemisia; the younger one was noted for her capture of a Rhodian fleet in the harbour of Halicarnassus and capture of Rhodes itself, and the stratagem which led to the capture of Latmus (Polyaenus VIII 53 f.) must also be referred to her. Vitruvius (II 8, 11) says that Mausolus was born in Mylasa and transferred his residence to Halicarnassus on account of the advantages of the situation.

⁸⁶ Strabo XIV 657 f.; Athenian tribute lists (ATL I 522, Μὺνδιοι παοἀ Τὲρμερα cf. Phot. s.v. ῾Τερμὲρια ᾿); Ps.-Scylax 99; Mela I 85; Pliny, , NH V 107Google Scholar; Anon. Stad. 276–8.

⁸⁷ The identification has not been attested by inscriptions; but Paton, who had a residence at Gümüşlük, remarked the preponderance of Myndian issues among the coins found there (JHS XX 80). The mediaeval testimonies alone might in fact be decisive for the location.

⁸⁸ BSA XLVII 184.

⁸⁹ Pliny, , NH II 204.Google Scholar

⁹⁰ The plan Fig. 3 is drawn after Admiralty Chart no. 1531 of the year 1837, with some remodelling of the legends and slight additions. Remains which are now no longer visible have been retained on the plan.

⁹¹ Arrian I 20, 5–7.

⁹² Diod. XX 37, 1.

⁹³ Polyb. XVI 15, 4.

⁹⁴ Appian, , BC IV 65, 71 f.Google Scholar

⁹⁵ Halicarnassus II 577.

⁹⁶ Cf. Ann. IV–V 367, figs. 27–9.

⁹⁷ Newton's explanation (II 574) that this side is less naturally strong could only be applied to the lowermost end of this stretch.

⁹⁸ The small foundation near the north tip of the peninsula, which Spratt and Newton (II 575,I, pl. 83) took for a bath, is now under water but must have stood on terra firma in antiquity.

⁹⁹ We noted an epikranon with carved cross among the architectural pieces close at hand.

¹⁰⁰ The wall of squared blocks traversing the base of the isthmus, which Newton (II 578) noted as ancient, can hardly be so, since pieces of tile can be seen in the joints.

¹⁰¹ Spratt marks it as extending from the peak almost to the north tip, and Paton apparently described it as doing so in the 1880s (JHS VIII 66, fig. 2); Guidi's assertion that it starts from the north summit of the peninsula and descends towards the sea in a southerly direction (Ann. IV–V 365) seems to correspond, assuming the topographical directions to be inverted.

¹⁰² Cf. Ann. IV–V 366, figs. 25–6.

¹⁰³ Cf. BSA XLVII 172.

¹⁰⁴ The fragment from a plate with palmettes and rouletting, Plate 13 (c) 3, is of the late fourth century or rather later.

¹⁰⁵ Paus. II 30, 9.

¹⁰⁶ ATL I 348.

¹⁰⁷ Paton, and Myres, , JHS XVI 204.Google Scholar

¹⁰⁸ Arrian, , Anab. II 5, 7.Google Scholar

¹⁰⁹ SEG I 363.

¹¹⁰ Florus II 20, 5.

¹¹¹ Appian, , BC V 7.Google Scholar L. Robert infers from the scanty coinage that Myndus suffered a decline in the imperial period (BCH LX 201). On the cults and prosopography of Myndus, see Petrarca, , Rend. Lincei XII (1936), 259 ff.Google Scholar

¹¹² Robert's discussion in Coll. Froehn. 65–9 should be regarded as decisive.

¹¹³ We adopt Robert's lettering in Coll. Froehn. 85–6. The visible crest from D to B subtends an angle of Θυαγγ[ελεὺς] degrees when seen obliquely from Alâzeytin at a distance of c. 6 km.

¹¹⁴ Robert, Coll. Froehn. 85, observes mistakenly that the tetrapyrgon is linked to the west angle of the main perimeter by a single wall. He has no doubt reached this conclusion from a study of his own composite photograph, pl. XXVI, but the appearances are deceptive; in fact, a wall runs up to each of the eastern towers of the fort: see Fig. 4. The tetrapyrgon, as Mr. E. W. Marsden pointed out to us, is of unusually powerful construction and designed to resist artillery.

^114a The fragments in question were deposited by Miss Akarca in the Archaeological Museum in Istanbul; we are indebted both to her and to the Museum Director, B. Rüstern Duyuran, for permission to announce this discovery in advance of publication.

¹¹⁵ BMC Sculpt. I 1, 149 (B 319). dated c. 520 B.C.

¹¹⁶ See RE s.v. ‘Theangela’.

¹¹⁷ The strongly moulded projecting rim of this vase seems best matched by the profiles of early Hellenistic vases (e.g. Trendall, Paestan Pottery, pl. 38 below).

¹¹⁸ We have no information regarding the ‘large tumuli’ seen by Paton and Myres on a distant skyline ‘probably in the neighbourhood of Theangela’ (JHS XVI 198); we suspect they may have been the fortified peaks of Theangela itself.

¹¹⁹ The case has been argued, conclusively as we think, by Robert, Coll. Froehn. 82 n. 7 and in ATL I 551–2, as against Ruge in RE s.v. ‘Theangela’. We have nothing to add: the intermediate form Συαγγελ[ῆς ] καὶ ᾿Αμυνα[νδὴς ] seems to us almost decisive in itself.

¹²⁰ 446/5 B.C., Συὰγγελα δὲ καὶ Μὺνδον διεφὺαξε

¹²¹ When the two pay separately in 444/3, the tribute of each is half a talent.

¹²² Strabo XIII 611 = FGrH II B, no. 124, fr. 25; Πολὲμων

¹²³ See the list given by Robert, Coll. Froehn. 91–4; his no. 18 now ὲπὶ ὶερὲως Πολεὶτου τοῦ ᾿Ανδροσθὲνου IV 8, no. 248.

¹²⁴ Nos. 9 and 10 in Robert's list.

¹²⁵ Cf. ATL I 552. For a similar claim by Myndus, see above, p. 111.

¹²⁶ ὲκομὶσθη ὺπὸ Μιχαὴλ Βογιατзῆ ὲκ τοῦ φρουρὶου Θεαγγελὶας (῾Αλικαρνασσοῦ) This man is known as a priest of Poseidon Isthmius at Halicarnassus (CIG 2655).

¹²⁷ It is accepted by Ruge in RE s.v. ‘Theangela’ 1374 and by Robert, Coll. Froehn. 84–5.

¹²⁸ ‘Der Vermerk der neuen Erwerbungen n. 199; ὲκ τῆς ῾Αλικαρνασσοῦ lehrt, dass der Stein von der Ruinenstätte (sc. Etrim) stammt.’ In fact, as was seen above, this entry belongs not to 199 but to 196.

¹²⁹ The alternative is hardly attractive, namely to suppose that Μενεκρὰτης Μενεκρὰτου Θεαγγελαὺς means merely that the stone, though found at Etrim, was shipped to Athens from Bodrum.

¹³⁰ IG II² 8831, Κρητικὰ Χρονικὰ is there dated by Kirchner to the first century B.C. If this is correct, it is conclusive against the supposed second-century incorporation in Halicarnassus. In the Addenda, however, a second-century date is preferred for the epitaph: ‘s. I a. Theangela non iam sui iuris fuit. L. Robert per litt.’ If this argument is, as we believe, baseless, it will be possible to revert to Kirchner's original dating.

¹³¹ The plan Fig. 5 was fixed where possible by compass intersections from the corners of the citadel, but the steep convex slopes made a complete system of intersections impossible; some details of the intermediate perimeter and the position of the vaulted tomb have been added from a freehand plan drawn by Mr. R. V. Nicholls in 1949.

¹³² Halicarnassus II 580 ff.

¹³³ JHS VIII 81 f., XVI 203 f.

¹³⁴ Cf. Paton's, drawing, JHS VIII 64Google Scholar, fig. 1. The gate appears on the right of the photograph in Plate 17.

¹³⁵ At this corner a stretch of a roughly piled cross wall runs across the neck towards the citadel; it hardly seems to belong to the original design of the circuit.

¹³⁶ Plate 14 (b) 1, probably from the neck of an amphora with striped decoration, similar to seventh-century Chian; Plate 14 (b) 4, from a late B.F. chariot-scene skyphos, with legs of horses and a man; fragments of early fifth-century glazed kylikes.

¹³⁷ Plate 14 (b) 3, oenochoe rim with painted ovolo pattern; Plate 13 (c) 1, foot of black glazed bowl with glazed underside and grooved resting surface, stamped palmette and rouletting; fragments of black glazed kantharoi.

¹³⁸ Wine amphora fragments, black glazed kylikes, kotylai, and bolsals, etc., R.F. fragment Plate 14 (b) 2.

¹³⁹ Newton II 580 ff.

¹⁴⁰ JHS VIII 67 ff., 454 ff.; XVI 243 ff.

¹⁴¹ Desborough, Protogeometric Pottery 218 ff. For the gold and bronze equipment from these tombs see BMC Bronzes 8, BMC Jewellery 100 f.

¹⁴² Paton, and Myres, , JHS XVI 204Google Scholar, noted this tower (at their sketching station ‘A’).

¹⁴³ Cf. ibid. 204, 264.

¹⁴⁴ JHS XVI 206, fig. 7. It was visited by the Italian cruise, Ann. IV–V 363 f.

¹⁴⁵ A photograph JHS XVI, pl. 9, 1.

¹⁴⁶ The inscription no. 55 was also found at Geriş.

¹⁴⁷ JHS VIII 79.

¹⁴⁸ These are grey throughout, or tend to grey in the core; the shapes seem to be jars with vertical handles (as at Bozdağ, p. 118) and a bowl with incurving rim.

¹⁴⁹ JHS VIII 81.

¹⁵⁰ JHS XVI 206 f.

¹⁵¹ JHS VIII 78.

¹⁵² JHS XVI 203.

¹⁵³ It was drawn from memory after we had left the site in a thunderstorm.

¹⁵⁴ JHS XVI 203.

¹⁵⁵ JHS XIV 376 ff., with drawings figs. 2–3. The citadel on the summit was also seen by the Italian, cruise, Ann. IV–V 365, fig. 23.Google Scholar Our plan (Fig. 9) was laid out to scale, with intersections in the SE parts only and traverse for the rest of the circuit; an intersection on a very narrow base, which showed the site as distinctly broader in the middle part north of the tower, has been disregarded in our plan.

¹⁵⁶ JHS XIV 376. Guidi, , Ann. IV–V 365Google Scholar, remarked Christian paintings in one.

¹⁵⁷ JHS XIV 373. We noted ancient blocks built into the walls, but could not recognise any in situ.

¹⁵⁸ Ibid. 374, fig. 1.

¹⁵⁹ Ibid. VIII 81; XVI 202, with drawings figs. 4–5.

¹⁶⁰ Ann. IV–V 425 ff., with a sketch plan of the citadel fig. 30.

¹⁶¹ The plan of the citadel (which is overgrown with pinewood) is drawn mainly freehand; the lower circuit was plotted in by traverse and intersection on an insecure basis.

¹⁶² JHS XVI 216.

¹⁶³ JHS VIII 81 f. We also heard of one from a cottager here.

¹⁶⁴ Ann. IV–V 427 ff., figs. 32–34; cf. Maiuri, , Parola del Passato III (1948), 13 ff.Google Scholar; JHS XVI 247, fig. 22.

¹⁶⁵ Levi, , Κρητικὰ Χρονικά IV 177, n. 73Google Scholar; bronze fibulae and arms are mentioned among the few fragments of grave furniture recovered, Clara Rhodos I 124.

¹⁶⁶ Plate 13 (e) 1, fragment of large coarse vase with spiral or cable pattern on a cream slip; 2, fragment of slipped vase with stripes; 3, fragment of body with springing of belly handle, wavy line in glaze; 6, fragment of closed vase with curvilinear decoration: also fragments of skyphoi, one being of the Ionic Late Geometric form with nicked rim.

¹⁶⁷ JHS XVI 249 ff., figs. 26–30; the tumulus, figs. 26–28, is described and illustrated by Maiuri, , Ann. IV–V 429 ff.Google Scholar, figs. 35–38, who gives the diameter as 22 m. See below, pp. 166 f.

¹⁶⁸ JHS XVI 199 f., pl. 9, 3–6.

¹⁶⁹ Ann. IV–V 432 ff., figs. 39–46.

¹⁷⁰ Two fragments with painted concentric circles widely spaced, probably of the eighth or seventh century, in the Ashmolean Museum.

¹⁷¹ JHS XVI 200, fig. 2.

¹⁷² JHS XVI 248 ff.; below, pp. 166 f.

¹⁷³ Cf. JHS XVI 251, fig. 28.

¹⁷⁴ JHS XVI 198 ff.

¹⁷⁵ Ann. IV–V 439 f., fig. 47. See p. 166.

¹⁷⁶ JHS XVI 200. It appears from letters shown us by Sir John Myres that Paton could not definitely establish the position of this sanctuary but ascertained from a ‘retentive old dyer’ at Syme that it was located near Alâzeytin (apparently to the exclusion of Theangela); it is apparently this site that is referred to in BMC Terracottas 92, where pieces similar to the sixth-century horseman B 118 (cf. p.94, n. 58) are mentioned.

¹⁷⁷ JHS XVI 201.

¹⁷⁸ Or Karabağlar, cf. ATL I 498.

¹⁷⁹ BCH XII 282 f. Cf. the Pserima of Pliny, NH V 134Google Scholar; for mediaeval forms of the name see Tomaschek, Zur hist. Topographie v. Kleinasien 22. According to Segre, , Ann. XXII–XXIII 219 f.Google Scholar, Coan and Mausolan coins have been unearthed in the fields.

¹⁸⁰ Newton, , Halicarnassus II 579.Google Scholar

^180a The inscriptions of Karatoprak are related by Petrarca to the cult of Artemis Myndia (Rend. Lincei XII (1936), 260).Google Scholar

¹⁸¹ These have the form of simple boxes open at one end; one which we measured was 0·85 m. broad, 0·75 m. high, and 2·20 m. from front to back.

¹⁸² JHS XVI 204, Hill ‘F’.

¹⁸³ Ibid. 262.

¹⁸⁴ Cf. Newton II 588 f.; BCH XIV 120. This site, as L. Robert has remarked (Études epigr. et phil. 165 f.), is probably that of the important mediaeval fortress of Strobilos near Myndus (Tomaschek, Zur hist. Topographie v. Kleinasien im Mittelalter 38 f.); the name Strobilos implies a conical hill, and St. Willibald in the eighth century after Christ described this Strobilos as a city on or at a high mountain (T. Tobler, Descriptiones Terrae Sanctae 20 and 60). For the Christian inscriptions of Çifitkalesi see Grégoire, Inscr. chrét. nos. 232 f.

¹⁸⁵ Cf. Hamilton, , Researches in Asia Minor II 38Google Scholar; Newton, loc. cit.

¹⁸⁶ Cf. JHS XVI 203.

¹⁸⁷ Loc. cit.

¹⁸⁸ H. 0·42; breadth 0·38; thickness in centre 0·12 m.; the back is roughly worked off. It seems too rough for an oscillum.

¹⁸⁹ JHS XVI 202, pl. 11 (sketching stations Θ and Υ).

¹⁹⁰ JHS LXXIII 125, fig. 12.

¹⁹¹ Pullan gives the depth of the blocks in the Mausoleum core as one foot (Halicarnassus II 183). We did not revisit Myndus after seeing these quarries, and have no measurements for the greenstone blocks in the fortifications there; Guidi's figure of 0·45 m. (Ann. IV–V 368) applies to a tower with headers and stretchers which may be of a later date than the original system of fortification.

¹⁹² JHS VIII 78.

¹⁹³ At Azacik, , JHS XVI 206.Google Scholar

¹⁹⁴ Ibid. 207, pl. 11, sketching station ‘Q’.

¹⁹⁵ JHS XVI 261 f. Newton's inscription has alpha, not mu, in the first place, and the sigma square.

¹⁹⁶ JHS XVI 207 f. The plan ibid., fig. 8 is inaccurate in the marking of the north wall; the north face of the tower should according to our notes be shown as forming a right angle with the east face and so creating a bend in the wall here.

¹⁹⁷ Cf. Newton II 592 ff., tombs and perhaps a tower at Filkecek.

¹⁹⁸ Newton II 595; JHS XVI 208 f., 253 f., 262. Another rock tomb on the east side of the bay has a chamber 3·57 × 3·20 m. with single cubicles on the sides and two large cubicles at the back (that on the right, no doubt the owner's, having a niche for an inscription over the door).

¹⁹⁹ Newton II 595 ff.; JHS XIV 376 f., XVI 210; Ann. IV–V 363, figs. 20–22.

²⁰⁰ Prof. Haspels has examined samples, and reports that the coarse local ware is ancient (one fragment apparently Hellenistic), while the glazed ware is to be dated between the thirteenth and seventeenth centuries.

²⁰¹ Paton and Myres mark a Byzantine church inland here (JHS XVI, pl. 10) and tumuli at Torba bay (ibid. 254). A mausoleum was noted here, of which we hope to give a photograph in our concluding article on the Carian coast.

²⁰² JHS XVI 198 f.

²⁰³ Ibid. 194 (‘compound tumuli’ ibid. 249, 254).

²⁰⁴ Cf. Newton II 602, where both routes are remarked; Prokesch v. Osten in 1827 followed the coastal one (Denkwürdigkeiten 443 f.). The walking time Bodrum–Milâs is reckoned a good 12 hours (cf. Ross, , Reisen IV 38Google Scholar, Newton, II 602, 610).

²⁰⁵ Admiralty Chart 1546.

²⁰⁶ The remains here are too slight to correspond to the fortified site remarked by Paton on a hilltop here (JHS XVI 200). This site was discovered by Paton in 1893 after Myres had gone away, but not investigated. In a letter, which Sir John Myres has kindly allowed us to quote, Paton referred to it as a ‘mountain fortress on the hill just above Budrum, on the S.W., with walls of seemingly great extent’, and in a subsequent letter he spoke of it as ‘on the hill above H. Georgios [at Kumbahçe], Boudroum’. We reconnoitred the mountain ridge here without finding any trace of a fortified site other than that mentioned above, but examined a hilltop a few hundred metres to the east of the crest ringed by a rocky escarpment which so closely resembles a fortification that we were certain while approaching it that it was the site noticed by Paton.

²⁰⁷ For the type see p. 166. The name is for convenience of reference only.

²⁰⁸ The half closing of the windows is shown by the bonding to belong to the original construction; the windows are mostly c. 0·20 m. wide on the exterior, though varying from 0·17 m. (the stair light) to 0·35 m. (the adjacent window on the north). The top three steps of the staircase (each a single block) are visible, while the lower part of the flight is buried under the collapse of the roof; the risers are 0·20–0·25 m. high, and there is an incline of about 0·05 m. on the tread.

²⁰⁹ Cf. Alâzeytin, , Ann. IV–V 435Google Scholar, fig. 32 (the original arrangement of the doors of this house has been altered), Urun, (JHS XVI 202, fig. 3, main block on north)Google Scholar, the house or tomb at Farilya, (JHS XVI 253, fig. 30)Google Scholar, and a building at Etrim (p. 114).

²¹⁰ JHS XVI 200.

²¹¹ AM XII 224, figs. 1–2. Paton, and Myres, (JHS XVI 254, n. 4)Google Scholar were unable to find these remains.

²¹² Ibid. 331 f.

²¹³ Ann. IV–V 440 ff., fig. 48a–b.

²¹⁴ Ann. IV–V 442 ff., figs. 49–52; JHS XVI 255 f., fig. 31.

²¹⁵ Paton, and Doerpfeld, , AM XX 466 ff., pls. 12–13Google Scholar; JHS XVI 201; Ann. IV–V 449 ff., figs. 57–61.

²¹⁶ Ann. IV–V 445 ff., figs. 54–55. We do not know whether this is the same as Paton and Myres' site ‘on a precipitous hill (1300 ft.)’ overhanging the bay (JHS XVI 198), but it is clearly that laid down on Admiralty Chart 1604. Cf. Hula-Szanto, Bericht über eine Reise in Karien (SB Wiener Akad. CXXXII) 30.

²¹⁷ Cf. SCE II pls. 130 ff.

²¹⁸ Hula and Szanto saw substantial remains hereabouts (op. cit. 26 f.).

²¹⁹ Cf. Kiepert, R., FOA VIII 7.Google Scholar

²²⁰ This site is not the same as the tower above the sea marked on Admiralty Chart 1604 and mentioned by Maiuri, (Ann. IV–V 448 f., fig. 56)Google Scholar; cf. also JHS XVI 197. It is, however, that visited by Hula and Szanto (op. cit. 26).

²²¹ Ap. Newton II 627 ff.

²²² JHS XI 109 ff.

²²³ Ann. IV–V 386 ff.

²²⁴ AJA 1935,.341 ff.

²²⁵ Devambez, , AJA 1935, 344 ff.Google Scholar, now in Smyrna Museum. Two terracotta figurines of classical date in the Louvre (Mollard-Besques, , Cat. raisonné C 160–1, pl. 80Google Scholar) are reported to have come from Ceramus.

²²⁶ JHS XVI 197.

²²⁷ We understand that L. Robert in 1946 found a fragmentary inscription at Gökbel which may or may not be the same as our no. 67. To an enquiry by letter Professor Robert has not as yet replied; we therefore publish the inscription in case it is new, more especially as we understood it to have been unearthed in 1952. If we are in fact anticipating Professor Robert's publication, we offer our apologies.

²²⁸ Πισιδὶας is suspect. In ATL I 537, n. 5 it is proposed to read Μιλησὶας

²²⁹ This is in fact accepted in ATL I 538: see below n. 237. Mausolus did not merely attach the Lelegian towns to Halicarnassus as demes; he was concerned to man the great new city, and the bulk of the Lelegian population was certainly transferred there. The old sites ceased to be inhabited as townships, as is clear from the remains, though perhaps not all entirely deserted.

²³⁰ There were only eight Lelegian towns, but Pliny's towns are not stated to be Lelegian. Side, Medmassa, and Uranium might be claimed as non-Lelegian; Theangela (or at least Syangela) was Lelegian, but was left free by Mausolus.

²³¹ For the site at Gökçeler see above pp. 123 ff. and below pp. 149 ff.

²³² Cf. Jones, Cities of the Eastern Roman Provinces 32.

²³³ ATL I 536 says that ‘Alexander was incensed with Halicarnassus and indeed destroyed it after the siege; but he certainly restored it.’ We do not know on what authority this last statement is made. For the destruction, cf. p. 91 above.

²³⁴ Jones CERP 383, n. 7 suggests that ‘the missing sixth city is perhaps Termera’.

²³⁵ Normally, libera is either obelised or made to apply (contrary to practice) to Bargylia.

²³⁶ Whether the error in Pliny is due to his own carelessness or to a faulty manuscript tradition, we are not concerned to decide. The corruption, in all MSS. but one, of Theangela to Thagela may perhaps help to explain the mistake.

²³⁷ It is observed in ATL I 552, n. 2 that Pliny's authority for his six cities cannot have been Callisthenes, who would have said Syangela, not Theangela. The substitution of Termera for Theangela obviates this difficulty also. We note further that the editors of ATL are hard put to it (ibid. 538) to find names for the cities synoecised by Mausolus other than those in Pliny's list, and are driven to include Cindya and even Halicarnassus itself. We cannot agree that this last is implied by Strabo's τὰς ἒξ . . . εὶς μὶαν though the words might conceivably be so interpreted. Halicarnassus cannot have ranked as a Lelegian city in the fourth century. As for Cindya, the evidence suggests that it was absorbed by Bargylia, not by Halicarnassus; see Polybius XVI 12, where Artemis Cindyas is a goddess of Bargylia. This absorption is in fact accepted in ATL I 474, 503, 538; cf. Jones, CERP 50, 388.

²³⁸ This distance is in fact such that no inhabitant could be too far from the nearest city to ride or walk in with reasonable convenience to exercise his civic functions. Mausolus' cities—Myndus, Halicarnassus, Theangela, Bargylia–Cindya, Mylasa—are remarkably evenly distributed over the countryside.

²³⁹ Pliny, NH V 107Google Scholar: Myndos et ubi fuit Palaemyndus; Steph. Byz. s.v. ῾Μὺνδος ᾿πὸλις Καρὶας . . .ὲστι καὶ πὸλις Καρὶας ἂλλη Παλαιὰ Μὺνδος

²⁴⁰ We see no reason why the memory of a Palaemyndus should be preserved if its site was the same as that of the familiar city. For this reason (among others) we cannot agree with the suggestion in ATL I 522 that Old Myndus was on the peninsula which closes the harbour at Gümüşlük. For the much-quoted polygonal wall on this peninsula see above pp. 110 f.; even if it be accepted as of Lelegian date, its position is quite unsuited to the defence of a settlement on the peninsula itself. Running down the backbone of the hill, it is intelligible only as the extremity of a larger circuit such as that actually standing on the mainland. Its situation is exactly comparable to that of the wall over the western extension of the acropolis hill at Caunus; see JHS LXXIII (1953), 12, fig. 3.

²⁴¹ The site at Burgaz might also claim consideration; but the expression Μὺνδιοι παοἀ Τὲρμερα in the tribute lists is then less intelligible, with the town at Bozdağ intervening.

²⁴² Their non-Lelegian character was noted by Judeich, (AM XII (1887), 335)Google Scholar in his description of the walls; Robert, Coll. Froehn. 85 n. 2, quotes Judeich and adds: ‘il me semble que l'appareil des monuments de Theangela est celui des monuments Cariens de la presqu'île d'Halikarnasse … et d'Alazeitin’. This comment holds good for certain of the buildings inside the city, and in a modified degree for the circuit walls also.

²⁴³ Against Judeich's late Hellenistic dating Robert, op. cit. 85, notes that the walls were no doubt standing in the late fourth century, when the city withstood a siege by Eupolemus. In the treaty between Eupolemus and Theangela (Coll. Froehn. no. 52) it is provided that Eupolemus shall eventually take over τὴμ πὸλιν καὶ τὰς ἂκρας Robert, ibid. 81–6, understands ‘the city and the citadels’, identifying the latter with the two peaks A and B. This seems to us mistaken. Πὸλις of course, often denotes the lower, inhabited city as opposed to the fortified acropolis; but at Theangela, as Robert himself emphasises (ibid. 82), there was no such lower city; the whole city was evidently within the walls. There is therefore no point in mentioning the two peaks unless they were separately fortified as inner citadels, a question which Robert does not consider. The mere fact that the mountain rises to a double summit is obviously not enough. We saw no evidence of such fortification on A; on the west side of B there is a stretch of wall which might perhaps be so interpreted, and in fact a more recent examination (see p. 171 n. 359) shows that there was a definite inner circuit defending this peak. If B was so fortified, it was doubtless included among the ἂκραι but the points on the site which are most clearly denoted by the term are surely the forts on the subsidiary peaks C and D. The meaning is that no Theangelan garrison shall be maintained in the city after Eupolemus takes over, a point well deserving mention in the treaty; but to provide for his occupying the city, including the double mountain-top, is plainly futile. An exact parallel is afforded in Arrian's account of the siege of Halicarnassus (I 23, 3), where the defenders retire to the fortified posts at the extremities of the city: αὺτῶν δὲ οΙ μὲν ὲς τὴν ὰκραν τὴν ὲν τῆ νὴσω ὰπεΧὼρησαν οὶ δὲ ὲς τὴν Σαλμακὶδα ὰκραν οῦτω καλουμὲνην

²⁴⁴ In particular those illustrated by Robert in Coll. Froehn. pls. XXVII d, XXVIII h, and in RA 1935 II 162, no. 10, and the house described above, p. 114.

²⁴⁵ The absence of recognisable Roman remains of any kind suggests that under the Empire the inhabitants moved their living-quarters down to the plain—a very understandable proceeding, in view of the arduous ascent to the mountain-top.

²⁴⁶ Callisthenes, writing c. 330 B.C., still uses the form Syangela; but he is speaking of the Mausolan synoecism and would naturally use the name appropriate at that date: it would be incongruous to speak of Mausolus preserving Theangela. In any case (as may be seen in Turkey to-day) the new name would take time to become generally recognised. There is no epigraphical reference to Syangela that need be later than Mausolus: Σφαγγειλαῖος in an inscription of Oropus, (AE 1917, 231Google Scholar, Robert, Coll. Froehn. 94, no. 17) is dated to the fourth century, before 338. Conversely, the earliest epigraphical mention of Theangela is in the Eupolemus inscription, which is supposed to date c. 315 B.C.

²⁴⁷ For Pliny's, notice (NH V 107) see above pp. 143 f.Google Scholar

²⁴⁸ The site at Aspat (above, p. 129) shows no evidence of occupation before Christian times, and is not a serious rival candidate. The allusion in ATL I 522 to ‘the modern town of Assarlik’ is an error: Asarlik is the name given to the ancient site, which is now quite deserted.

²⁴⁹ For the evidence from pottery of occupation at the end of prehistoric times see p. 118.

²⁵⁰ He is mentioned again in Her. VII 98.

²⁵¹ See Mr. Threpsiades' communication in the current reports from Greece for the year 1953, esp. AJA 1954, 231, pl. 43, 2.

²⁵² We infer this from the termination -ικὸν which must apparently refer to something other, or more, than the city of Termera: cf. such inscriptions as συμμαΧικὸυ ᾿Ολυμπικὸν ὰρΧιερατικὸν

²⁵³ Or identical with the latter, if the later dating of the coin, now abandoned, is right.

²⁵⁴ The decline in prosperity from this time on, suggested by the sherds at Asarlik (above, pp. 117 f.), may have been a consequence of this break-up.

²⁵⁵ The union under Tymnes paid in the name of Γερμερῆςν Similarly, Syangela under Pigres pays generally as Συαγγελῆς

²⁵⁶ For Tymnes in the ‘Carian Syntely’ of 425 B.C. see below, pp. 162 f.

²⁵⁷ This is the view taken by Head, op. cit.

²⁵⁸ The explanation of the proverbial Τερμὲρια κακὰ seems to be uncertain. Suidas' account, which refers it to the sufferings due to the banditry practised by the Termerans, is apparently confirmed by Philip of Theangela and by Photius s.v., who attribute the practice to the eponymous founder Termerus. (The passages are quoted in FHG IV 475.) But a different explanation is given by Plutarch, Thes. II, who makes it equivalent to ‘being paid in one's own coin’ or ‘given a dose of one's own medicine’: ῾Ηρακλῆς τὸν Τὲρμερου συρρὴξας τὴν κεφαλὴν ὰπὲκτεινεν. ὰφ᾿ οὺδὴ τὸ Τερμὲριου κακὸν ὸνομασ- θῆναι λὲγουσι παὶων γὰρ, γὰρ, ὼς ὲοικε, κεφαλῆ τοὺς ὲντυγχὰνοντας ὸ Τὲρμερος ὰπὼλλυεν (Not merely ‘a misfortune one brings on oneself’, as LS⁹s.v., where, moreover, the more usual explanation is disregarded.)

²⁵⁹ I 175: ἢσαν δὲ Πηδασὲες οὶκὲοντες ὺπὲρ ῾Αλικαρνησσοῦ μεσὸγαιαν, τοῖς ι ὂκως τι μὲλλοι ἀνεπιτὴδεον ε̆σεσθαι αῦτοῖσὶ τε καὶ Τοῖσι περιοὶκοισι ὴ ὶρεὶη τῆς ᾿Αθηναὶης πὼγωνα μὲγαν ὶσΧει. τρὶς σφι τοῦτο ὲγὲνετο. οῦτοι τῶι περὶ Καρὶαν ὰνδρῶν μοῦνοὶ τε ἀντὲοΧου Χρὸνον ῾Αρπὰγω καὶ πρὴγματα παρὲσΧου πλεῖστα, ὸρος τειΧὶσαντες τῶ οῦνομὰ ἐστι Λὶδη Cf. VIII 104: οὶ δὲ Πηδασὲες

²⁶⁰ XIII 611, quoted above, p. 143.

²⁶¹ V 107, quoted above, p. 143.

²⁶² JHS XVI (1896), 215–6 no. 4 = CIG 2660 = SGDI 5731: cf. Robert, Ét. Anat. 440. Not, as ATL I 537, halfway between Gökçeler and Bitez.

²⁶³ Prokesch, , Denkwürdigkeiten III 441 (c. 1827)Google Scholar, gives the form Pedess, the Admiralty Chart Petasa.

²⁶⁴ Though Ruge in RE allows it only ‘a certain probability’. Earlier locations of this ‘Halicarnassian’ Pedasa at Etrim or at Karacahisar no longer need refutation, as these sites are securely assigned respectively to Theangela (above, p. 112) and Hydissus (Robert, , AJA XXXIX (1935), 339).Google Scholar

²⁶⁵ We take little account of the variant forms Pedasa, Pedason, Pidasa, which are to all appearances interchangeable. St. Byz. s.v. ῾Πὴγασα᾿ strangely insists that the name should be spelt with a gamma.

²⁶⁶ Cf. Milet I 3, 352 ff.

^266a A discussion of the site is promised by Robert, Coll. Froehn. 79, n. 3.

²⁶⁷ Her. VI 20: τἀ δὲ ὺπερὰκρια ἒδοσαν Καρσὶ Πηδας εῦσι ὲκτῆσθαι

²⁶⁸ The Persian motive was perhaps to divide and so weaken a people who had given them serious trouble in the past (Her. I 175: for V 121 see below).

²⁶⁹ Herodotus does not distinguish Carians and Lelegians (cf. I 171).

²⁷⁰ He might have been provoked by the Milesian intervention at Labraunda to turn against Miletus, but it seems to us quite clear from Herodotus' words that this is excluded. To take the road to the east would be to abandon the expedition with its object unachieved; the opposition was by no means yet suppressed, as the event showed.

²⁷¹ The total absence of running water in this region, combined with the mountainous nature of the country, would make a camping site practically impossible to find. We have assumed that the road taken by the Persians followed roughly the course of the modern chaussée; if instead they attempted the road along the coast from Siralik to Torba (above, p. 131), the success of the ambush is even easier to understand.

²⁷² Since Herodotus (VIII 104) calls the city Πὴδασα we should be inclined to read ὲπὶ <Πη> δὰσοισι in the text of V 121, but certainty is impossible.

²⁷³ With the proviso that the half-talent assessment in 425 may relate to Gökçeler.

²⁷⁴ Side was not assessed, nor apparently Telmissus.

²⁷⁵ Pigres, probably of Syangela, at Salamis (above, p. 114); Myndian ship in the Persian fleet (Her. V 33). For Madnasa, see below, p. 155. Termera also continued to pay her reduced tribute, and she too had a captain in the Persian fleet (Her. V 37).

²⁷⁶ Εῦρωμον δὲ καὶ Πὴδασα καὶ Βαργὺλια καὶ τὴν ᾿Ιασὲων πὸλιν (also Abydos, Thasos, Myrina, and Perinthus) ὲλευθὲρας ἀφεὶναι τὰς φρουρὰς ὲξ αὺτῶν μεταστησὰμενον Cf. Livy XXXIII 30.

²⁷⁷ Of the various forms attested—Telmissus, Telmessus, Telemessus—we have chosen the first as having the authority of the only strictly local inscription; but the choice has no significance, and in particular the form of the name can never, it seems, be used to distinguish the Carian and Lycian cities.

²⁷⁸ ATL I 554, ‘in situ near Pelen’ refers presumably to the reported provenience of the stone as learned by Paton: see below. We have not ourselves seen this stone.

²⁷⁹ In l. 15 of this inscription, ὰποκατὲστησε τῶι θεῶι τὴν Χὼραν ὲξῆς ῶστε [θ]υσὶας καὶ τιμὰς τῶι ᾿Απὸλλωνι συμβὲβηκεν ὲπιτελεῖσθαι we do not understand ὲξῆς and wonder if the true reading be not ὲξ ὴς (τὰ) ς τε [θ]υαὶας κτλ In l. 23, [ὰπο- δειλιῶν] is wrongly restored; [δεικνὺμενος] or something similar is required.

²⁸⁰ ATL I 554: once before Λὺκιοι καὶ συν(τελεὶς) once between Phaselis and Calynda. Telmessus, though of Lycian origin, seems to have held aloof from the rest of Lycia at least until the middle of the fourth century (Theopompus, , FGrH. 115, F 103Google Scholar), so might well be separately assessed.

²⁸¹ It is true that the Lycian Telmessus was near the border of Caria and Lycia; but this border was an important one, since Lycia was at this time still independent, whereas Caria was part of a Roman province—a province, moreover, governed by Cicero's brother, to whom the description is not improbably due. The mistake would be the more reprehensible since Telmessus is taken as an outstanding example of a faculty common to all the Carians. For Ruge's second reason for supposing Caria to be an error here, see below in connection with the Karadağ site.

²⁸² Or the later koinon, if Ruge be right in supposing Arrian's account to be of late origin.

²⁸³ Neither Cicero nor Arrian makes any mention of Apollo, for whom at Telmissus the two inscriptions and the coin are the only evidence. In Cicero's case this is natural, as he is concerned to give a rational explanation of the Carians' powers of divination: having these rich lands, in which many strange things are liable to germinate, they are familiar with the phenomena of nature. In general, it seems that the Telmissians' divinatory faculties made more impression in antiquity than the oracular powers of Apollo.

²⁸⁴ Paton, –Myres, , JHS XIV (1894), 373.CrossRefGoogle Scholar

²⁸⁵ It is accepted by Kiepert, by Ruge in RE, and in ATL I 554.

²⁸⁶ The site is now very thickly overgrown; more was perhaps visible sixty years ago.

²⁸⁷ See above, p. 123. This fort is, we imagine, that shown on Admiralty Chart 1546; Paton and Myres make no reference to it, and take the ruins on Karadağ to be the fort shown on the chart.

²⁸⁸ He did not, of course, see the stone at the church himself.

²⁸⁹ The identity of these names is universally accepted and hardly needs justification. Mednassa and Methnassa are in fact variant readings in Pliny.

²⁹⁰ For the connection of Madnasa with Caryanda see below, p. 158.

²⁹¹ See the descriptions of the two bays in Mediterranean Pilot (7th ed. 1941), pp. 313–14.

²⁹² For this information we are gratefully indebted to Mr. Aubrey Diller. It appears that Müller's, note in GGM I 73 is incorrect.Google Scholar

²⁹³ The change to λιμὴν involves the further alteration of ταὺτη e.g. ὸμὼνυμος ταὺτην ῶκουν Koraes; ὸμὼνυμος τοὺτῳ ὴν ᾢκουν Müller (Didot 1877); Meineke's Teubner text (following the Vatican epitome) omits ταὺτῃ ATL loc. cit. follows Meineke. Newton points out that all MSS. of Strabo and Steph. Byz. have λὶμνη (Halicarnassus II 599), and Chandler does not seem to have questioned the reading λὶμνη.

²⁹⁴ Corrected from Myridos.

²⁹⁵ MSS. Termera: see above, p. 144.

²⁹⁶ ATL loc. cit. accepts Jacoby's conjecture πὸλις καὶ λιμ <ὴν καὶ> νῆ <σος> ὸμὼνυμος

²⁹⁷ In the face of this evidence, it is strange that scholars (including even so good a judge as Robert, : see Rev. Phil. LXII (1936), 283)Google Scholar should so often have sought the classical Caryanda on the Carian mainland. R. Kiepert's suggestion in FOA VIII, Text 7, that the Caryandans may have had country houses and farms on the island, and that the name of the city was eventually transmitted to the island, seems an almost exact inversion of the actual course of events. ATL loc. cit. rightly insists on an island site, but leaves the evidence for a mainland site (Strabo, Mela, Pliny) in the air.

²⁹⁸ He calls it a city, urbs; this is no doubt a mere assumption on his part: he would know there was at one time a city: there is no trace elsewhere of a city of Caryanda in Hellenistic or later times. Incidentally, it is perhaps not certain that Mela mentions Caryanda at all; it would presumably be possible to read 〈N〉aruanda, corresponding to Pliny's Nariandos. Stephanus' account is confused: see below n. 300.

²⁹⁹ The only reason offered (to our knowledge) for changing to λιμὴν is the mention of a harbour by Ps.-Scylax (see Kramer's note ad loc.), but this is clearly wrong-headed: Strabo's λὶμνη is expressly distinguished from the island recorded by Ps.-Scylax.

³⁰⁰ Strabo, loc. cit., Mela, loc. cit., Pliny V 107. Stephanus' evidence is neither one thing nor the other; he seems to have taken his facts from a variety of sources, and he certainly did not realise that two sites were in question. He has consulted Hecataeus; λὶμνη ὸμὼνυμος is apparently from Strabo; πὸλις may be his own invention or may be from Ps.-Scylax or Mela; πλησὶον Μὺνδου καὶ Κῶ might be from Hecataeus, as suggested in ATL I 498, but need not be so; both names may be taken from the extant authorities quoted above.

³⁰¹ Cf. the story of Diogenes quoted above, p. 111.

³⁰² This settlement at Lower Göl has, of course, no connection with the Lelegian site on the hill above, but was centred on the low ground by the shore, where later a considerable Byzantine town grew up. The Lelegian city at Göl was identified with Caryanda by Paton and Myres on the strength chiefly of a coin of Caryanda found on the shore close by (JHS XIV 375 f.). This identification (which has met with considerable approval, e.g. from Head, HN ² 612, Robert, , Rev. Phil. 1936, 283Google Scholar, Coll. Froehn. 84 n. 2) is in our view out of the question; first, because an island site is absolutely demanded for the early Caryanda (so ATL loc. cit.), and second, because the city at Göl is clearly one of the synoecised Lelegian towns, whereas Caryanda was not among the eight names recorded, as we believe, by Callisthenes, and is shown to have survived the synoecism by its mention in Ps.-Scylax (see above, p. 145). The lake lends no support, being connected in the authorities only with the later Caryanda. The coin is too portable an object to be relied on in isolation, but may well have come across with the settlers from the island.

³⁰³ It is mentioned by the Turkish traveller Evliya Çelebi in the seventeenth century, who speaks of gardens and orchards and a considerable population dwelling around. We are indebted for this information to the Educational Officer at Bodrum, Bay Necati Çavdar; we have not ourselves seen this lake, nor do we know its exact location. It seems certain that there is no other lake whatsoever in the region in question except that at Göl.

³⁰⁴ BMC Cat. Caria lxv, 140, pl. XXIII.1; second or first century B.C. Obv. Head of Apollo. Rev. Lyre: ΝΕ ΑΠ ΟΛΙ ΜYΝ ΚΟΛΒΑ (cf. Head, HN ² 623). The reading MYN is said to be ‘not quite certain’, though it appears perfectly clear in the photograph. ΚΟΛΒΑ is supposed to represent a magistrate's name: Colbasa in Pisidia struck coins under the middle Empire, but it has never, so far as we know, been suggested that our coin may belong to it.

³⁰⁵ Unless the text be altered; see above, n. 296.

³⁰⁶ See ATL I. Caryanda is next to Madnasa in lists 3, 12, 13, 23, and next but one in list 5; only in list 12 is it anywhere near Myndus.

³⁰⁷ We take it that λιμὴν in Ps.-Scylax means no more than a decent anchorage, not necessarily harbour-works: for a periplus this would be the essential information; but he is not consistent in mentioning or omitting this item. Konel is a very rocky island, and not easy to land on except in the early morning calm; if it can be said to have a λιμὴν, the term virtually ceases to have any meaning.

³⁰⁸ On this point see below, p. 163.

³⁰⁹ SIG ³ 1044, 1. 18.

³¹⁰ The statement in JHS XIV 375 that ‘the ancient name of Tarandos was certainly Taramptos, for which there is no alternative site’ (our italics) seems strangely inaccurate; if Taramptos is not Tarandos, it may have been anywhere on the whole peninsula not unduly far from Halicarnassus. It may in fact have been no more than a farm.

³¹¹ Tára(m)p might easily, with the help of a simple Turkicism, have finally become Torba (‘Bag’); other forms quoted are Durvanda, and Trupada, (Med, Pilot ⁵ IV (1918), 366)Google Scholar, of which the second (?= Turpada, ‘Turnip Island’) could well be corrupted from Taramptos.

³¹² So indeed it is taken to be in ATL III 210 f. (‘by apotaxis from Bargylia’). We hesitate to attach significance to the ‘Karandakia Rocks’ south of Salihadasi and close to Siralik (Admira name has any genuine historical basis.

³¹³ JHS XVI 208; accepted by Kiepert.

³¹⁴ Her. 1. 173, where Termilae is recorded as the old name of the Lycians, without reference to Termera. Stephanus knows Τρεμὶλη as a name for Lycia.

³¹⁵ Except that Telmera is a variant reading for Termera in Pliny, , NH V 107Google Scholar. If a name must be found for Dirmil, Pliny's Nariandos is available, but there is no positive evidence for an identification. Nariandos seems to be otherwise unknown; for the Ναρυανδει̑ς in the neighbourhood of Stratoniceia see Robert, Ét. Anat. 569.

³¹⁶ The modern port is Küllük.

³¹⁷ The figure Κ′ is also wrong; as pointed out in GGM loc. cit. and in ATL I 506 n. 1, the correct figure 80 is recorded by Pausanias VIII 10, 4.

This passage of the Stadiasmus is full of difficulty. The MS. has (according to Müller's notes): 288, ὰπὸ Βαρβυλὶων εὶς ᾿Ιας ὸν στὰδιοι σκ᾿ (? measured by following the ancient coast all round the ‘Little Sea’; the direct crossing is barely 50 stades): 289, ἀπὸ ᾿Ιασοῦ ὲπ`ἀκρωτὴριον Ποσεὶδιον στὰδιοι ρκ᾿ (really about 220): 290, ὰπὸ ᾿Ιασοῦ εὶς τὴν ᾿Ακρὶταν στὰδιοι σμ`; 291, οὶκεῖται κατὲναντι Πὰσσαλα πηγὴ κτλ. As a connected passage this will not hang together. If Acrita is Agathemerus' Arcitis, the modern Arki east of Patmos, the figure 240 stades from Iasus is little more than half the true distance; Müller accordingly reads Ποσειδὶου for ᾿Ιασοῦ in 290, but believes that this section is a later addition. We strongly suspect the same of 291, for the following reasons. (1) If Passala the port of Mylasa was mentioned in the original text, it should have come between Bargylia and Iasus. (2) Notes of this kind are unusual in this part of the Stadiasmus, which consists in general of a mere list of places and distances. (3) If we suppose a reader's marginal note, illegible handwriting will explain the corruption. (4) Such a note will also account for the abbreviation σταδ.; elsewhere in this passage στὰδιοι is written in full. If 290 and 291 are omitted, the catalogue proceeds in straightforward fashion, Bargylia—Iasus—C. Poseidion—Panormus—Miletus. But even if added later, the note is not for that reason devoid of all value; it is a confirmation of Stephanus' note on Passala. In this case the corruption must apparently extend to πηγὴ, which can hardly be right; if a correction must be found, we are tempted to suggest <καὶ ναυ> τὴγια. Other explanations of the passage are possible, but hardly attractive. E.g. Passala might be identified with Pliny's island and placed in the neighbourhood of Acrita; the rest of 291 is then a confusion with Stephanus' Passala, the port of Mylasa. But the drawbacks to this are obvious.

³¹⁸ It is identified by Meineke, ad loc., most unconvincingly, with Patara in Lycia.

³¹⁹ Aemyndus is very likely to be corrupt; it has a peculiarly unconvincing look. We cannot, however, approve the suggestion there made to suppose a lacuna and read, e.g. Lamps〈imandus; inde sinus Iasicus et in continent〉e Myndus; it is surely unlikely that Pliny would interrupt his catalogue of islands to name a city on the mainland which he has already recorded in its proper place, unless for some special purpose. We see no such purpose here.

³²⁰ It is observed in ATL I 522 that no cape west of Petra can be described as opposite Kumburnu; we do not understand this, and suspect that ‘east of Petra’ was intended.

³²¹ Fenerburnu is regarded in the Mediterranean Pilot as the north-west entrance-point to the Cos Channel, Kocaburun as the north-eastern entrance point.

³²² Kiepert interpreted the name Astypalaea as ‘Erniedrigung’ (from the Phoenician), with reference to the low isthmus between the lofty cape and the high ground on the mainland; this description would apply equally to the cape south of Myndus but not (as Ruge supposed) to Sağiralaca, which is flat and featureless.

²²³ The relative positions are then more or less correct, but the compass-points are turned through an angle of about 90°.

³²⁴ P. 447, with pl. XXIII; cf. Meritt, Epigraphica Attica 119 ff.

³²⁵ Op. cit. 553, with map on p. 554.

³²⁶ ATL III 211 n. 77.

³²⁷ Of the upsilon in l. 112 nothing is now to be seen on the stone.

³²⁸ See above, p. 161. There are in fact no islands at the head of the gulf; Crusa is supposed in ATL to have been, like Passala, absorbed since antiquity by the alluvial advance of the coast-line. If Taramptos is Salihadasi, the largest island in the Iasian Gulf, and if Pliny's list extends to the head of this gulf, it is curious that Taramptos is not named in it.

³²⁹ whether or not our suggested location at Burgaz be correct; see above, p. 155.

³³⁰ Above, p. 161.

³³¹ See p. 147. Robert's ingenious suggestion for ll. 114–5, [Κ]υμν̩[ισσῆς π]α̩ρὰ Μ[ὺνδον] (Rev. Phil. LXII (1936), 282), would also suit our location of the syntely; but it is declared epigraphically impossible, on the grounds that l. 115 cannot have contained more than nine letters: part of the tenth letter, if it existed, should be visible on the stone. The earlier suggestion [Σ]ὺμπ̣[ρ̣α] (see ATL I 553), i.e. Symbra in western Lycia (see most recently JHS LXXIII (1953), 26), is naturally no longer maintained.

³³² The first preserved letter in l. 115 might be gamma or delta equally with alpha.

³³³ See above, p. 159.

³³⁴ For the single inscription in which Taramptos is named, see above, p. 159. There is no indication as to what sort of a place it was.

³³⁵ Nariandos is available, or another of Pliny's islands; Side may have been assessed for the first time. In I. 112 it seems that the bottom extremity of an upright stroke was formerly visible in the third place; [Πὰ]τͅ[αλος ] or perhaps [Να]ρͅ[ιανδῆς ] would meet the requirements. But such speculations are hardly profitable.

³³⁶ On Kiepert's map; approved in ATL loc. cit. Robert, Coll. Froehn. 84, n. 2 says the site is ‘entièrement incertain’, though he seems to consider Gökbel a possibility.

³³⁷ Cf. Paton and Myres' classification of the tombs, JHS XVI 242 ff.

³³⁸ F. P. Johnson has justly drawn attention to the startling resemblance (which cannot be explained solely by similar physical and geological conditions) between the Lelegian architecture here and the Dryopian constructions at Styra and on Okha, Mt. in Southern Euboea (AJA 1925, 398 ff.)Google Scholar; the Hellenistic date that he there suggests for the houses at Alâzeytin hardly seems tenable, however. In addition to similar wall construction, roofing, and doors the Styra dragon house also has loophole windows near the lower corners of the two long rooms (cf. the farmhouse near Bodrum, pp. 132 f.). There can be no doubt that these monuments on the Halicarnassus peninsula were among the ruins attributed by Strabo to the Leleges (VII 321, XIII 611); but Strabo's assertion that the Lelegian monuments were encountered all over Caria, and house ruins at least in the Milesia as well, does not entirely accord with the distinctive character of the architectural remains on the Halicarnassus peninsula. Cf. Paton, and Myres, , JHS XVI 268f.Google Scholar

³³⁹ Asarlik (p. 118), Gökçeler (p. 125), Burgaz (p. 120), and perhaps on the Kaplan Dağ. We cannot say whether the built chambers at Gürice (p. 121) and by the farmhouse SE of Bodrum (pp. 133 f.) were covered by mounds; those at Etrim (p. 114) seem not to have been.

³⁴⁰ Cf. Desborough, Protogeometric Pottery 218 ff. Bittel, Kleinasiatische Studien 70 f., Stresses the relationship with Mycenaean tholos tombs.

³⁴¹ JHS XVI 246 f.

³⁴² JHS VIII 81; Paton found fragments of marble and an Attic sherd, perhaps RF, inside the tomb. Cf. also the tomb at Etrim (Plate 16 (c)) now dated about the late fifth century (p. 113).

³⁴³ JHS VIII 73 f. Cf. JHS XVI 243 ff.

³⁴⁴ JHS VIII 78 fig. 26.

³⁴⁵ JHS VIII 248 ff. (a)–(e).

³⁴⁶ Ann. IV–V 354, fig. 9, identified by Guidi as a tomb of Lelego–Carian type in the necropolis on the seashore at Iasus, and ibid. 439, fig. 47 by Maiuri as a compound tumulus at the back (?) of the valley of the bay of ‘Gören-kuiu’ below Alâzeytin. Despite the superficial differences, we find it impossible to believe that the two illustrations have not come from one negative. The photograph shows what appears to be the thick part of the ring of a compound tumulus; and in Guidi's print (where the background is printed stronger) it is evidently situated on a hill slope high above a valley. Guidi's description bears no relation to the construction illustrated, and he is clearly mistaken in his identification of the photograph; Maiuri's location could fit the surroundings, and the position near Alâzeytin may therefore be accepted.

³⁴⁷ JHS XVI 251, n. 1.

³⁴⁸ Cf. the Carian series, JHS XVI 260, fig. 40, which has a more elaborate façade than any we have seen on the peninsula.

³⁴⁹ Cf. p. 131, n. 198; JHS XVI 262, figs. 43–4 (Farilya); ÖJh VI Beibl. 101 ff., figs. 24–31 (Halicarnassus).

³⁵⁰ Ann. IV–V 472.

³⁵¹ Besides the inscribed altars and base mentioned above we have seen a considerable number of uninscribed altars.

³⁵² Cf. Hamilton, , Researches II 43Google Scholar, Texier, , Déscription III 176, pl. 104 (Cnidus)Google Scholar; Antiquities of Ionia V 23, pls. 44–5, Ross, , Reisen III 73 (Lindus).Google Scholar

³⁵³ JHS VIII 72, figs. 14–15. Rock-cut tombs of this sort are not uncommon at Caunus and in Lycia.

³⁵⁴ To this general statement an exception must be made in the case of the site on Karadağ, which possesses no arable territory at all and has no modern counterpart; its occupants must have lived by pasture alone. The site at Göl, which we equate with Madnasa, commands a certain area of good land by the shore and lake, but it is impossible to doubt that its inhabitants, like their modern successors, were largely interested in the sea.

³⁵⁵ The little fort at Dirmil may also have served such a purpose.

³⁵⁶ Especially at Asarlik (pp. 117 f.) and Gürice (p. 121).

³⁵⁷ See pp. 110 f. and 114.

³⁵⁸ Ps.-Arist. Oecon. II 14–15; Polyaenus, , Strat. VII 23, 1.Google Scholar The tetrapyrgon at Theangela (pp. 113 f., Fig. 4), if a part of the original design, is an outstanding example of Mausolus' strategical foresight.

³⁵⁹ Bean and Miss A. Akarca spent a week at Etrim in September 1955. This visit made possible a more complete examination of the site of Theangela; and we hope to give a plan of the site in our concluding article on the West Carian coast.