² SEG xlii. 785; Duchêne (the editio princeps); Ph. Gauthier, Bull. Ep. 1993. 395; Graham, A. J., ‘The woman at the window: observations on the “Stele from the Harbour” of Thasos’, JHS 118 (1998), 22–40CrossRefGoogle Scholar.

³ On the problems of the authorship of the Hippocratic Corpus, see Smith, W. D., The Hippocratic Tradition (Ithaca and London, 1979)Google Scholar.

⁴ For a good description of the site before the modern excavations, see Baker-Penoyre, J. ff., ‘Thasos part II. Topography’, JHS 29 (1909), 202 48CrossRefGoogle Scholar. There are good, more recent, descriptions in Rech. i, 9–12 and Grandjean, 463–4.

⁵ Guide, 54–5.

⁶ See Grandjean, Y. and Salviat, F., BCH 119 (1995), 661–7CrossRefGoogle Scholar, fig. 6. The sherd came from a fill behind the blocks of the external face of the East wall of the NE bastion. New information was also recovered about the terrace of the Pythion, and the authors' plan (fig. 1) supersedes that by Baker-Penoyre (n. 4), 208 (fig. 1).

⁷ As by Duchêne, 71–4.

⁸ Guide, 39–40.

⁹ Duchêne, 91–3. The identification of the sanctuary was made by Launey, M., ‘Une dédicace thasienne à Hèracles’, BCH 58 (1934), 173–83CrossRefGoogle Scholar, at 181–3. The possibly complicating factor of a hypothetical second shrine of Herakles at Thasos, presumed to be in the vicinity of the agora, is discussed by Duchêne, who accepts the arguments of those who have rejected the theory of a second shrine. From the point of view of topography the issue is unimportant, since it is obvious that for the ancient Thasians there was one shrine that they called the sanctuary of Herakles (as this inscription itself confirms). H. Seyrig, the originator of the theory of a second shrine, already recognized this; see ‘Quatres cultes de Thasos’, BCH 51 (1927), 178–233CrossRefGoogle Scholar, at 191–2 n. 3.

¹⁰ Little of the wall has been preserved or discovered between the tower of Sotas in the SW and the closed harbour, but excavations have found a few traces, which allow its course to be conjectured; see Grandjean, 323 with n. 2, 324 and plan IX. Rather more is known of the wall to the NE of the closed harbour (Guide, 43–9), and it may be confidently assumed that the wall was continuous along the seaward side of the city.

¹¹ See Grandjean, 484 n.

¹² Duchêne, 101.

¹³ Ibid., 90–1. Cf. Grandjean, 426, where, using Duchêne's thesis of 1986, he cites the translation ‘de la falaise’.

¹⁴ Duchêne, 100.

¹⁵ Ibid.

¹⁶ Duchêne's discussion (90–1) is vitiated by his dependence on poetic sources, which do not provide suitable analogies for the meaning of the word in our inscription.

¹⁷ I rely on the texts which can be searched electronically.

¹⁸ Of ‘Byzantine’ date is Hist. Alex. Magni (recensio e), 40. 2: παράτινοςλιμνηςὄχθ̢̢η.

¹⁹ The article is often omitted in the addresses of patients in the Epidemics; see e.g. i, case 14, παράἭρηςἱρόν. So we should not see any significance in its omission here.

²⁰ The basic senses of ἀκτή, sea-coast or tongue of land projecting into the sea, are well illustrated by Herodotus; see Powell, J. E., A Lexicon to Herodolus (Cambridge, 1938Google Scholar; reprinted Hildesheim, 1960) s.vv. ἀκτή, Καλὴάκτή, Λευκὴ ἀκτή. Conze (17) already concluded that the akte referred to the flat beach or shore of the lower city.

²¹ SEG xix. 684, but, for the best edition, see IGSK Ephesos Ia. 23. For the date, see the convincing treatment of Antonius Albus' cursus by Eck, W., ‘Die Laufbahn des L. Antonius Albus, Suffektconsul unter Hadrian’, Epigraphische Studien, 9 (1972), 17–23Google Scholar.

²² I follow the useful translation in IGSK.

²³ ‘Philologie et réalités’, Journal des savants, 1962, 5–74, at 34–6.

²⁴ See Lanciani, R., The Ruins and Excavations of Ancient Rome (London, 1897), 527–8Google Scholar.

²⁵ See Wilberg, W., Forschungen in Ephesos iii (Vienna, 1923), 169–70Google Scholar and ÖJh 58 (1988)Google Scholar, ‘Grabungen 1987’, 8–9; 59 (1989), ‘Grabungen 1988’, 8–9; 60 (1990), ‘Grabungen 1989’, 31.

²⁶ Blackmail, D., ‘Ancient harbours in the Mediterranean’, Journ. Naut. Arch. II (1982), 79–104CrossRefGoogle Scholar, 185–211, does not mention the inscription, and states that at Ephesus the harbour area has been ‘defined but hardly excavated“ (88).

²⁷ At Piraeus there was a quay/jetty called Χῶμα see Garland, R., The Piraeus (London, 1987), 151, 218Google Scholar; but Hesychius’ definition of Χῶμα is ὕψωμαγῆὕψωμαγῆς, ὄχθη.

²⁸ These investigations are reported in BCH III (1987), 622–6Google Scholar; 112 (1988), 736–42; 113 (1989), 734–40; 114 (1990), 881–7; (1991) 712–20; 116 (1992), 721–6; 117 (1993), 647–52.

²⁹ The position of the mole is indicated on plans of Thasos, as Guide, fig. 4, and its remains have always been visible and can be seen in photographs, e.g. BCH 117 (1993), 648Google Scholar, fig. 1.

³⁰ Ibid., 647–52.

³¹ The cult at Paros is attested by Apollod. Bibl. iii. 15. 7, which shows that it had special characteristics, flutes and wreaths being forbidden in the worship; see O. Rubensohn, RE s.v. Paros (1949), 1845–6 and Berranger, D., Reckerches sur l'histoire et la prosopographie de Paros à l'époque archaïque (Clermont-Ferrand, 1992), 195–9Google Scholar. The latter follows (197) the over-clever attempt by Seyrig (n. 9), 184 n. 2, to distinguish the cult at Thasos from that at Paros, because the Graces on the famous relief from the Passage of the Theoroi carry wreaths, and it would be odd for deities to carry something forbidden to their worshippers. But they are carrying wreaths to welcome Hermes; cf. the perceptive interpretation by J. Pouilloux, ‘Une enigme thasienne: le passage des théores’, Thasiaca, 129–41, at 139; and it is against all probability and analogy for the Thasian cult not to be related to that at Paros.

³² Epid. i, case 14: παρά Ἥρης ἱρόν.

³³ Epid. iii, 12 cases, 1: παρά Γῆςἱρόν.

³⁴ Epid. i, case 5: παρὰἀρχηγέτην, ‘near the founder’ (W. H. S. Jones, Loeb). The preposition is the same as in the case of other deities; see, in addition to Hera and Ge, i. 21, παρὰ Διονύσιον. (In the same chapter παρ᾿Ἡρακλεῖ, or παρ᾿Ἡρακλείω̨—the reading is uncertain—have a different case, but Jones could find no difference of meaning, whether παρά is used with accusative, genitive, or dative [Epid. i, Loeb, Hippocrates, 145 n. 2]). The archegetes seems likely to be the hero to whom we have an inscribed dedication, ἥρωι Σω[τ]ίωνιἀρχηγέτηι (Rech. i. 127, pl. 36. 1), dated by letter forms to the last quarter of the fifth century. The stone was found in the excavations of the quarter by the Gate of Hermes; see Martin, R., ‘Un nouveau reglement de culte thasien’, BCH 64/5 (1940/1941), 163–200CrossRefGoogle Scholar, at 179 n. 3. This quarter was originally seen as a sanctuary, but that was denied by Grandjean (274 n. 2), who treats all the buildings as remains of habitations (283–90). The finding-place of the inscription is, therefore, unlikely to tell us anything about the location of the possible sanctuary. Deichgräber (13), although he refers to Pouilloux, unconvincingly suggested that the archegetes of the Epidemics might be Apollo, and the god's name might even have dropped out of the text. Other scholars have thought of Telesikles, the presumed oikistes of Thasos, and envisaged a shrine in or near the agora; see, e.g., Fredrich, 235, and R. Martin, ‘Thasos: quelques problemes de structure urbaine’, CRAI 1978, 182–97, at 189. A divine founder was already preferred by Conzc, 16. An archegetes also need not be the founder of the whole city, as the Hero Archegetes of Rhamnous shows. On the nature and cult of the archegetes, see Salapata, G., ‘Hero warriors from Corinth and Lakonia’, Hesp. 66 (1997), 245–60CrossRefGoogle Scholar, at 256–7, with the bibliography in n. 82.

³⁵ This is the earlier building, dated to between 525 and 490, which adjoins the building of Roman date identified as the sanctuary of foreign deities; see Grandjean, 316–17 (sondage D7, 2a). If the earlier building was a shrine (Grandjean has some reservations), it is unattributed. Of the other potential candidates, the sanctuary at Evraiokastro is not only extramural, but has been attributed, with some probability, to Demeter and Kore; see Grandjean, 277. A well built rectangular building of probable 5th-e. date, situated just outside the walls, in the right angle where the city wall turned SW to join the defences of the closed harbour, may have been a shrine, but there is epigraphical evidence which may show that it belonged to Soteira; see Martin, R., ‘Fouilles de Thasos’, BCH 68/9 (1944/1945), 147–50, 154–6Google Scholar. Several other buildings discovered in excavations were once identified as sanctuaries, but these identifications were rejected by Grandjean, 274 with n. 2.

³⁶ Duchêne, 93–4.

³⁷ Grandjean, 484–5, regards the association of the sanctuary of the Charites and the Passage proposed by Duchêne as very convincing (Grandjean had a copy of Duchênc's thesis; see 411 n. 1), and J.-Y. Marc accepts that the road of the sanctuary of the Charites is the great artery from the Passage to the Herakleion; see ‘L'Agora de Thasos’, in Ecole française d'Athenes, L'Espace grec (Paris, 1996), 105–13, at 105Google Scholar.

³⁸ See Blondé, F., Muller, A., and Mulliez, D., BCH 119 (1995), 681 7CrossRefGoogle Scholar, at 686 7. The dating was based on the style of the sculptured reliefs, but is not contradicted by the pottery found in the fill of the foundation trench.

³⁹ The term used in LIMC s.v. Charis, Charites 16 (E. Harrison). Such empty rectangular spaces, surrounded by a moulded border have also been found at Thasos (a) together with the relief of Silenos at the city gate, and (b) at the gate of Herakles and Dionysos, though the latter is known only from an impressionistic 19th-c. drawing; see Picard, Ch., Les portes sculptées. Les murailles, i (Études Thasiennes, viii. 1; Paris, 1962), 38, 69Google Scholar fig. 22. There has been much debate about their meaning. See, for example, Picard, ibid., and ‘Basrelief ionien archaïque de Thasos’, Fondation Piot. Monuments et mémoires, 20 (1913) 36–69, csp. 43–5. The suggestions include cultic niche, niche for offerings and symbolic door of naos/naiskos. The last is based on the fact that the surrounding moulding and shape of the rectangle are like those of Ionian doorways; See Picard, ibid., and M. Launey, Le Sanctuaire et le culte d'Héraklès à Thasos (Études thasiennes, i; Paris, 1944) 62–4. Picard thought he could exclude the door idea by pointing to the continuation of the moulding along the bottom, but he was over-confident when he dismissed those who have regarded the function of the Thasian niches as enigmatic (Les Portes sculptées, 52). It thus seems impossible to determine the exact significance of the ‘offering-niche’ in the relief of Apollo and the Nymphs.

⁴⁰ For the reliefs, see LIMC, loc. cit., and s.v. Apollo 716, Artemis 1166; Pouilloux (n. 31), fig. 1 (plan) and figs. 4, 5 (photographs of the reliefs). The question of the positioning of the reliefs in the Passage was long debated (ibid., 129 n. 5); but was finally settled by Salviat, ‘Lcs colonnes mitiales du catalogue des théores et les institutions thasiennes archaïques’, Thasiaca, 107–27, at 109–11, using their measurements and his reconstruction of the walls of the Passage. For the inscriptions, see IG xii. 8. 358 = Duchêne no. 20 = Sokolowski, LSCG 114.

⁴¹ See Guide, 50; cf. Pouilloux, Thasiaca, 135 n. 18. For the date, about the middle of the fourth century, see Garlan, Y., ‘Contribution à une étude stratigraphique de l'enceinte thasienne’, BCH 90 (1966), 586–611CrossRefGoogle Scholar, at 596–611.

⁴² Lazaridis, D., Thasos and its Peraia (Athens, 1971), 76Google Scholar para. 382.

⁴³ Ch. Picard, Mon. Piot (n. 39), 57; Les Portes sculptées (n. 39), 15.

⁴⁴ Id., Mon. Piot, loc. cit.; Les Portes sculptées, 38. His suggestion that little votives, perhaps including money, might have been deposited in the ‘offering-niches’ found in the gate of Silenos and that of Herakles and Dionysos, was only supported by the statement that he found many fragments of red-figure cups at the latter gate. But the ‘offering-niche’ at the Gate of Silenos is very shallow. Picard did not give actual or estimated original depth, but wrote that it was ‘peu profonde’ (‘Les fouilles de Thasos’, CRAI 1912, 193–235, at 205) and ‘guère profonde’ (Les Portes sculptées, 101). J. Boardman, however, accepts the idea that the niche was for offerings; see Greek Sculpture of the Archaic Period (London, 1978)Google Scholar, caption to pl. 223.

⁴⁵ IG xii, Supplementa 414 = Duchêne, no. 23 = Sokolowski, Suppl. 63.1

⁴⁶ See Picard, Ch., ‘Un rituel archaïque d'Héraclès thasien trouvé à Thasos’, BCH 47 (1923), 241–74CrossRefGoogle Scholar, at 241.

⁴⁷ See e.g. Seyrig (n. 9), 185–98, and Pouilloux, J., ‘L'Héraclès thasien’, REA 76 (1974), 305–16, at 314–15CrossRefGoogle Scholar.

⁴⁸ See J. Des Courtils and Pariente, A., ‘Problèmes topographiques et religieux à l'Hérakleion de Thasos’, in Étienne, Roland and Dinahet, M.-T. (eds). L'espace sacrificiel (Paris, 1991), 67–73Google Scholar.

⁴⁹ For this altar, see Roux, G., BCH 79 (1955), 354Google Scholar; Pouilloux, Thasiaca, 133.

⁵⁰ See Salviat, F., ‘Thasos. Monuments de l'agora, le “Passage”’, BCH 80 (1956), 418–20CrossRefGoogle Scholar, who saw continuity of cult from the earlier altar (eschara) to the later altar of the existing Passage. The recent excavations beneath the floor of the Passage, however, have shown that there was archaeologically an interruption between the use of the eschara and the later arrangement of the Passage; Blondé, Muller, Mulliez (n. 38), 686.

⁵¹ Fredrich, 233–4.

⁵² Mon. Plot, 20 (n. 39), 60.

⁵³ It has been suggested that the connection of the Passage with the theoroi is Seeondary, because the inscribing of the names of the theoroi on its walls only began in the fourth century; see Roux (n. 49), 359. But the association could well be older than the time when the lists began to be inscribed. The office of theoros existed long before the inscribing of the list.

⁵⁴ Ibid., with fig. 28.

⁵⁵ See Pouilloux, Thasiaca, 133; Rech. ii. 404. The reading was corrected by Chamoux, F., ‘L'île de Thasos et son histoire’, REG 72 (1959), 348–69, at 369CrossRefGoogle Scholar.

⁵⁶ See n. 49.

⁵⁷ Pouilloux, Thasiaca, 133.

⁵⁸ Ibid., 129–35.

⁵⁹ See e.g. G. Daux, ‘Les fouilles de l'École française d'Athènes à Thasos en 1954’, CRAI 1955, 469–79, at 476–7.

⁶⁰ See in general Stengel, P., Die griechischen Kultusaltertümer (Munich, 1920), 11–17Google Scholar.

⁶¹ Pouilloux, Thasiaca, 129.

⁶² Roux (n. 49), 359.

⁶³ Pouilloux, Thasiaca, 133–9; Martin (n. 34, CRAI), 182–9. I am most grateful to F. Blondé, A. Muller, and D. Mulliez for kindly sending me a copy of the preliminary report of their 1999 season, which contains a brief account of the discoveries which they take to prove that the Passage is on the site of an earlier city gate. All students of Thasos will eagerly look forward to the full publication of these very important excavations.

⁶⁴ e.g. Ch. Picard, Mon. Piot (n. 39), 69; Roux (n. 49); Grandjean, 484–5.

⁶⁵ Ibid, and plan XI. For more recent work, see F. Blondé, A. Muller and D. Mulliez, ‘Une nouvelle place publique à Thasos’, RA 1987, 25–39; BCH 117 (1993), 661–4Google Scholar; 118 (1994), 445–7; 119 (1995), 681–7; AEMTH 10B (1996Google Scholar, appeared 1998), 813–23.

⁶⁶ Cf. Pouilloux, Thasiaca, 133. Salviat called it ‘propylées’; ‘Dédicaces de magistrats à Thasos’, BCH 82 (1958), 319–28CrossRefGoogle Scholar, at 328.

⁶⁷ Pausanias (IX. 35. 3) says the three Charites were worshipped with mysteries at Athens ‘before the entry to the Acropolis’ (Pausanias(IX.35.3)). At i. 22. 8 he says that statues of Hermes Propylaios and the three Charites were ‘right at the very entry to the Acropolis’ (Κατὰ δὲτὴνἔσοδοναὐτηνἤδητὴνἐςἀκρόπολιν). On the basis of these statements R. E. Wycherley distinguished the sculptures, which he thought might have stood within the Propylaea, from the place of cult, which he thought ‘probably occupied the bit of space east of the Nike shrine’; see The Stones of Athens (Princeton, 1978), 131Google Scholar. In any case, there was a wealth of cults, statues and dedications both within and in the near vicinity of the Propylaea, as Wycherley shows, 127–34.

⁶⁸ The altar in the propylon is very similar to another found near the NE corner of the temple; see Yavis, C. G., Greek Altars (St Louis, 1949), 132Google Scholar (Section 52, nos. 1 and 2) and, for plans, Gabrici, E., ‘Il santuario della Malophoros a Selinunte’, Mon. Ant. 32 (1927)Google Scholar, pl. 1 and 2. Gabrici dated the propylon by its architecture to shortly before the fall of Selinous in 409 (87), and he dated the altar(s) by the building. Yavis, however, using, it seems, style and analogy, dates them to ‘probably … early sixth century’ (132). If he is right, an older altar was placed in the new propylon, but that is obviously possible, since the propylon is the only entry to the temenos (Gabrici, 75), so it had a predecessor of some sort, and there is no problem about continuity of cult.

⁶⁹ See n. 67 above.

⁷⁰ Harrison (n. 39), pp. 192–3.

⁷¹ Cf. Delorme, J., Gymnasion (Paris, 1960), 114Google Scholar, who thinks that the οἰκήματα appear to be shops (n. 5).

⁷² This text is cited by Duchêne, 94.

⁷³ Ibid., 79–85.

⁷⁴ Plato, Ap. 17 C; Hp. Mi. 368 B.

⁷⁵ SEG xxvi. 72. 5–6. cf. 46. R. Stroud argued that the money-changing tables were probably situated in the NW corner ol the Agora; see ‘An Athenian law on silver coinage’ Hesp. 43 (1974), 157 88, at 167CrossRefGoogle Scholar.

⁷⁶ There was also money-changing in the Piraeus; cf. SEG 26. 72. 41–6.

⁷⁷ SIG ³ 218. 9–10.

⁷⁸ Ps. Arist. Oec. 1346b 24–5. For the date, see B. A. van Gronmgen's discussion in Anstote. Le Second livre de l'économique (Leiden, 1933), 61Google Scholar.

⁷⁹ e.g. Lucian, Dial. D. 4 (24). 1; P. Oxy. viii. 1128. 14 (AD 173); i. 76–19 (AD 179): P. Flor. i. 5. r. 7 (AD 244/5); P. Lond. 1722. 20. 22 (late Byzantine) and many other examples.

⁸⁰ As F. Salviat, ‘Le vin de Thasos. Amphores, yin et sources écrites’, in. J.-Y Empereur and Y. Garlan (eds), Recherehes sur les amphores grecques, 145–95, at 154.

⁸¹ Ibid.

⁸² See Pantel, P. Schmitt, La Cité au banquet (Rome, 1992), 4Google Scholar.

⁸³ Ibid., 313–24.

⁸⁴ In that ease it may be misleading to translate συμπόσιον as ‘salle des banquets’, as Salviat (11. 80), 154.

⁸⁵ Theophr. Od. 51.

⁸⁶ For that term and building, cf. Miller, S. G., The Prytaneion (Berkeley and Los Angeles. 1978), 66Google Scholar.

⁸⁷ Schmitt Pantel (n. 82), 331.

⁸⁸ Duchêne, 96–7; Grandjean, 295–7.

⁸⁹ Schmitt Pantel (n. 82), 330–1.

⁹⁰ e.g. that this is the beginning of Thasian urbanism. or that the symposion at Thasos was an activity of the civic community, of people who belonged to a polis as citizens (ibid.). Duchêne, on the other hand, sees an aristocratic ethos here (98).

⁹¹ Od. 51.

⁹² IG xii. 8. 262, where the restoration πρυτα]νεῖον has been made in line 1. A different supplement, ἐπι]νεῖον, was tentatively suggested by Y. Grandjean and F. Salviat in their very radical reinterpretalion of this document, which they take to be an Athenian decree; see ‘Decret d'Athènes, restaurant la démocratic à Thasos in 407 av J. C: IG xii. 8. 262 complété’, BCH 112 (1988), 249–78CrossRefGoogle Scholar. (Their bold idea has not been accepted by some scholars; see SEG 38. 851 and the full arguments of Gauthier, Bull. Ep. 1989, 254, and of Tréheux, 384.)

⁹³ See Duchêne, 98–9; Salviat (n. 66), 327–8; S. G. Miller (n. 86), 234. Immediately outside the SW corner of the agora there is a large public building, whose function is unknown, and one conjecture is that it was the prytaneion; see AR 44 (1997–1978), 93Google Scholar. The earliest phase of construction is Hellenistic.

⁹⁴ Miller (n. 86), 29.

⁹⁵ Ibid., 29–30.

⁹⁶ Martin and Pouilloux (n. 63); Grandjean, 481 n. 1, points out that the idea is also found earlier.

⁹⁷ Martin, 188–9. for the character of the terrain, see Martin, 188, Pouilloux, 138 and Grandjean, 480.

⁹⁸ Grandjean, 297–8, 480–85.

⁹⁹ As Pouilloux, , ‘Inscriptions, topographic et monuments de Thasos’, REA 61 (1959), 273 99CrossRefGoogle Scholar, at 280 n. 2.

¹⁰⁰ See Blondé, Muller and Mulliez (n. 38), 687–8.

¹⁰¹ See Grandjean, 482.

¹⁰² Ibid., 477.

¹⁰³ Ibid.

¹⁰⁴ See n. 63.

¹⁰⁵ Is it possible that the eommercial agora known from the Epidemics (see below) was once the sole agora of the city?

¹⁰⁶ Duchêne, 102.

¹⁰⁷ This against J.-Y. Marc (n. 37), 105.

¹⁰⁸ e.g. IG xii. 8. 265. 6.

¹⁰⁹ IG xii. 8. 361.

¹¹⁰ As Martin (n. 34, CRAI) 189.

¹¹¹ As at Athens; see IG i³. 1087–90.

¹¹² Also at Athens; See Aeschin. iii. 176, Dem. xx. 158.

¹¹³ Duchêne, 101 4; See esp. his plan (102).

¹¹⁴ Ibid., 100.

¹¹⁵ Ibid., 96–7.

¹¹⁶ Sec Rocchi, G. Daverio, Frontiera e confini nella Grecia antica (Rome, 1988) no. 2, pp. 96–9Google Scholar; Plassart, A., ‘Orchomène d'Arcadie. Inscriptions (2)’, BCH 39 (1915), 53–134CrossRefGoogle Scholar. at 53–97; Buck, Greek Dialects, no. 20.

¹¹⁷ Daverio Rocchi's translation is fault) at this point.

¹¹⁸ Daverio Rocchi, no. 18, pp. 167–9. (2nd c. BC).

¹¹⁹ Conze, 16; cf. Duchêne, 99.

¹²⁰ (n. 67), 187.

¹²¹ One is reminded of the brilliant combination of named and numbered streets in William Perm's Philadelphia, which has been copied all over America. E. Greco, ‘Nomi di strade nelle città greche’, Κοινά. Miscellanea di studi archaeologici in onore di Piero Orlandini, ed. M. Castoldi (Milan, 1999) 223–9. appeared after my paper was complete.

¹²² See Duchêne's careful and detailed description, 21.

¹²³ ἀπὸδὲτο̃Ἡρακλέ[ος]ἱρο̃μέχριςθαλάσσηςτὴνὁδὸνταύτηνἐπιστάταικαθαιρόντωνκτλ.

¹²⁴ ἀπὸΧαρίτωνἱρο̃μέχριςτῶνοἰκημάτων.

¹²⁵ Duchêne, 99.

¹²⁶ Ibid.

¹²⁷ Katastasis is rendered ‘constitution’ by W. H. S.Jones; see the Locb Hippocrates i, 141. W. Smith says that it means ‘condition’ or ‘situation’; see Locb Hippocrates, ii, 7 n. 7. It is clear from the Epidemics that the author was thinking of the climatic conditions as they affected health.

¹²⁸ IG xii. 8. 277. 81 = Rech i. 26. col. 5, 26, IG xii. 8. 263. 2. The latter fixes Antiphon's theoria within the years of Thasos' revolt from Athens, which have usually been dated 411–407. Salviat has argued convincingly, however, that the revolt should be dated from the summer of 410 to the spring of 407, and regards the most probable date for Antiphon's theoria to be 410–409; see ‘Les archontes de Thasos,’ΠρακτικάτουΗΔιεθνούςΣυνεδρίουΕλληνικήςκαιΛατινικήςΕπιγραφικής(Athens, 1984), 233–58Google Scholar, esp. 254.

¹²⁹ ‘Die Epidemien und das Corpus Hippocraticum’, Abhand. Preuß. Akad. Wiss. Phil.Hist. Klasse, 1933, no. 3; see 25–74.

¹³⁰ Ibid., 74–5.

¹³¹ ὁκατάΜηδοσάδεωκώμης. This is the reading preferred by Littre, but not by W. Smith in the Loeb Hippocrates, vii. The evidence and issues were well discussed by A. Foesius in his edition of 1621, p. 1137.

¹³² As Smith, Loeb Hippocrates, vii. 10.

¹³³ Satyros, vi. 8. 29, and Gorgippos, vi. 8. 32.

¹³⁴ Deichgräber 22–33.

¹³⁵ e. g. cases 28–35 of vii are also described in v. 63, 98, 99, 62, 60, 61, 96, 97.

¹³⁶ Deichgräber, 24.

¹³⁷ These coincidences are given below in the Appendix. In this part of Deichgräber's work there are many minor inaccuracies, and, more seriously, some of his precise identifications are rather adventurous, because of the dominant homonymy at Thasos, where the same names recur from generation to generation. This also leads him to some questionable chronology. Thus on p. 18 he argues from coincidences of names that the archon list Rech. i. 34 records those in office c. 380–350, which is significantly earlier than Pouilloux' dating (Rech. i. 274), 345–315. And Pouilloux has numerous convincing prosopographical arguments.

¹³⁸ Conze, 17.

¹³⁹ Ibid.

¹⁴⁰ Chamoux (n. 54), 362, Conze, 17.

¹⁴¹ e.g. Salviat, , BCH 84 (1960), 312CrossRefGoogle Scholar and n. 4; Grandjean, 476.

¹⁴² The reading and spelling of the name is discussed in the Appendix below.

¹⁴³ See Isserlin, B. S. J., ‘The canal of Xerxes: facts and problems,’ BSA 86 (1991), 83–91Google Scholar, esp. 88.

¹⁴⁴ The otherwise useful entry in Liddell and Scott is marred by some errors. The first, general, rendering given, ‘any broad flat body or space’, seems to reflect Hesychius’ definition, τόποςπλατὺςκαὶμέγαςὑποθαλάττιος, while ignoring the important final word (‘under the sea’, ‘by the sea’?). Meaning 2, ‘flat beach’, is not justified by the citations. Meaning 4, ‘flat land liable to overflowing’, involves mistranslating in one cited passage (Polyb. x. 48. 7) and uncertainty in the exact sense in the other (Dionys. Per. 626). Finally, the two passages of Oppian, Halieutica i. 121 and v. 650, used to justify meaning 5, ‘level sea’, clearly refer to ledges of rock.

¹⁴⁵ See above n. 34.

¹⁴⁶ Fredrich, 229.

¹⁴⁷ These ravines are well indicated on the drawing by J. Perrin-Fayolle in Rech. i, pl. 1. 3.

¹⁴⁸ See Grandjean, 332, E6, 7 (plan VIII).

¹⁴⁹ See above p. 306.

¹⁵⁰ See ibid.

¹⁵¹ See Eust. Od. ii. 7; cf. Diog. Laert. ix. 114.

¹⁵² Paus. i. 1. 3; cf. Garland (n. 27), 152. Two agoras are advocated by Aristotle, Pol. vii. 1331b, 10–13, one which he calls τὴνἀναγκαίανἀγοράν(translated ‘the business agora’ by Rackham in the Loeb), and the other which he calls τὴν ἄνω, used for leisure pursuits (ἐνσχολάζειν).

¹⁵³ See Conze, 1 7 and pl. 2.

¹⁵⁴ See Fredrich, 230.

¹⁵⁵ Conze, 17.

¹⁵⁶ See Tréheux, J., ‘Une nouvelle voie thasienne’, BCH 79 (1955), 427–41CrossRefGoogle Scholar, fig. 1; Grandjean, 316–19 (D7, 2).

¹⁵⁷ Conze, 16–17.

¹⁵⁸ For the forms of these names, see Appendix below.

¹⁵⁹ I rely on LGPN i, ii, iiiA and Osborne, M. J. and Byrne, S. G., The Foreign Residents of Athens. An Annex to the Lexicon of Greek Personal Names: Attica (Studia Hellenistica 33; Leuven, 1996)Google Scholar.

¹⁶⁰ Deichgräber, 20.

¹⁶¹ Ibid., 13.

¹⁶² Grandjean, 487–8.

¹⁶³ For a full discussion, see Wycherley (n. 67), 237–45.

¹⁶⁴ Grandjean, 487.

¹⁶⁵ Ibid., 339. For the Turkish police station, see Launey (n. 39), fig. 1 opp. p. 13.

¹⁶⁶ See Graham (n. 2).

¹⁶⁷ Ibid. 38–9.

¹⁶⁸ See Garlan, Y., Vin et amphores de Thasos (École française d'Athènes, Sites et Monuments v, Paris, 1988), 28–30Google Scholar, cf. 25. The same conclusion had already been reached by Bon, A.-M. and Bon, A., Les Timbres amphoriques de Thasos (Études Thasiennes iv; Paris, 1957) 27Google Scholar. The new discoveries have shown that before c. 340 (or a little later; Garlan, 24) the stamps consisted of four elements, (i) ethnic, if present; (ii) name of eponymous magistrate, (iii) symbol and (iv) name of manufacturer. After c. 340 we have (i) ethnic, (ii) name of eponymous magistrate, (iii) symbol of manufacturer; see Garlan, 17–19.