* This paper was given as a seminar to the Department of the Classics at Harvard in November 1998, exactly twenty years after I first spoke on the symposium at sea there. I am exceedingly grateful for the warm reception and the ensuing discussion, in particular to Kathleen Coleman and Albert Henrichs. One could wish no better forum to test one's ideas.

¹ Alpers, K., ‘Zwischen Athen, Abdera und Samos’, in M., Billerbeck and J., Schamp (edd.), Kainotomia: Colloquium Pavlos Tzermias (Freiburg, 1996), 19–55Google Scholar, fr. 7 with p. 51.

² Jacobsen, Eric, ‘Die Metamorphosen der Liebe und Friedrich Spees Trutznachtigall’, Del Kongelige Danske Videnskabernes Selskab: hist.-fil. Meddelelser 34, 3 (Kopenhagen, 1954), 105ff.Google ScholarVon Koppenfels, Werner, Esca et Hamus. Beitrage zu einer historischen Liebesmetaphorik (SBAW, phil.hist. Kl. 1973, 3: Munich, 1973)Google Scholar; these were cited by Alpers in a letter to me; Flury, P., Liebe und Liebespsrache bei Menander, Plautus und Terenz (Heidelberg, 1968), 87ff.Google Scholar Of older and useful collections of material, one can in particular mention Birt's pupil Hoeltzer, V., Depoesi amatoria a comicis atticis exculta, ab elegaicis imitatione expressa (diss. Marburg, 1899), 73ff.Google Scholar, whose examples are augmented by Preston, K., Studies in the Diction of the Sermo amatorius in Roman comedy (Chicago, 1916), 55ffGoogle Scholar; Murgatroyd, P., ‘Amatory Hunting, Fishing and Fowling’, Lalomus 43.2 (1984), 362–8Google Scholar collects examples without citing these. Flury, Liebe und Liebespsrache, 87 cites Menander frr. 570 and 656 Thierf. from monologues by lovesick youths for the metaphor of being hooked, but one cannot agree with him that there is any tragic parody.

³ For example, Fantham, E., Comparative Studies in Republican Latin Imagery (Toronto, 1972), 39ff.Google Scholar

⁴ Preston (n. 2), 55. I have not found anything of interest in either Moll, O., Dioskorides (diss. Zürich, 1920)Google Scholar, or Castri, M. B. Di, ‘Tra sfoggio erudito e fantasia descritta: un profilo di Dioscorida: 3 epigrammi erotici e scoptici II’, Athene e Roma n.s. XLII (1997), 51–73Google Scholar at 61–3 who deals laboriously with the poem under discussion.

⁵ R.E. VIII (1913) s.v. Hetaira 1346ff. (K. Schneider) with list of hetaira-names, including Korone, Drosis; H. Herter, ‘Die Soziologie der antiken Prostitution’, JbAC 3 (1960), 70–111.

⁶ Alpers (n. 1), 51 noting Aristainetus, Ep. 2.1.17 lines 22–4 Mazal.

⁷ Murgatroyd (n. 2), 363; see also the commentary on Cratinus fr. 231K-A.

⁸ Earlier Hesychius gives κ-3584 Κοραχνί … καì τò γυναικîον αἰϨοîον but this looks like a confused explanation of a comic joke of the ἐκκορεῖν type. The unexpected appearance of an Apollo Korax from Libya (BE 1996 no. 117) makes one realize how little one knows.

⁹ I suppose this is where Passow in his Greek-German Lexikon got the explanation of KopdrTcu as ‘vom Anklopfen der Bettler an die Thüren’. But he might have been thinking of the begging songs of the koronistai, widely attested in Greece; on these see Robertson, N., ‘Greek ritual begging in aid of women's fertility and childbirth’, TAPA 113 (1983), 143–69Google Scholar, esp. 142.

¹⁰ Athenaeus 8.359E. But they got their name from the object they carried, and other groups had a swallow.

¹¹ Pollux 7.111 citing Posidippus fr. 8K-A. It would lead too far to pursue all the possibilities, but obviously the use of κορώνη in marriage songs (PMG Carm. Pop. 35: LSJ Suppl. s.v. κορικορώνη), and words like κνσθοκορώνη suggest old word plays on κóρη and κορώνη, not immediately relevant here.

¹² H. Frisk, Griechisches Etymologisches Wörlerbudi s.v. κόραξαι … κοράξαι ‘wohl eig(entlich) “sich anhaken”’; Chaintraine's Dictionaire étymologique s.v. gives: ‘probablement issu de κόραξαι “crochet”’.

¹³ LSJ gives Polybius 1.22.3 as the earliest example. There was also a siege machine of the same type. For ‘iron hands’, see Diod. Sic. 17.44.4 used along with ‘crows’ in warfare.

¹⁴ J. Henderson, The Maculate Muse (Yale, 1975) gives Aristophanes, Equites 1285 and Nubes 978, but they bear no relation to the metaphors here; he deals with ςρόσοѕ = semen on p. 145.

¹⁵ P. 102, 6 Bekker.

¹⁶ This was observed also by Poehlmann, E. and Tichy, Eva, ‘Zur Herkunft und Bedeutung von Kollops’, Serta Indogermanica (Göttingen, 1982), 287–315Google Scholar at 294. This important discussion, which came to my attention late in the writing of this article, is referred to hereafter as Poehlmann-Tichy. I am glad to find it echoes my scepticism about much of our lexicographical material, which the authors discuss in detail, as well as the history of lyre-pegs; there are a few errors, but not such as to affect the usefulness of the conclusions.

¹⁷ Aristophanes, Tagenistae fr. 520K-A, where the cook gives a list of desirable things to cook; the list consists possibly of sexual double entendres and therefore Poehlmann-Tichy argue that we cannot be totally certain it was meant to make sense at a practical level. The fragment is also badly corrupted. Taillardat on Suetonius, Περἰ βλασϕημιῶν 3.67 compares scortum for the alleged semantic shift from ‘leather’ to ‘prostitute’.

¹⁸ Aristonicus, the Homeric commentator of Augustan times, is cited by the Et. Gudianum, s.v. κόλοψ, deriving from the scholia on the Odyssey, according by Poehlmann-Tichy, 290, n. 29. The gloss derives from the Et. Genuinum at this point, and the citation comes from Orion, as Klaus Alpers reminds me. Apollonius Sophistes and others also makes this same deduction about glue. In fact, ἐκολλόπωσε in Achaeus Tragicus fr. 22.1 will not mean ‘glued together’ (as LSJ alleges) but ‘pegged together’ like κολλοπίζω. We can await Martin de Leeuw's edition of the Antiatticist for a proper discussion.

¹⁹ Poehlmann-Tichy (n. 16), 294.1 had come to this conclusion independently and on different grounds.

²⁰ Ibid.

²¹ See Kassel-Austin on Eubulus fr. 10, and my brief comments on Aristophanes of Byzantium fr. 36. The comments of Tosi, R., ‘Osservazioni sul rapporto fra Aristofane di Bisanzio e L'antiatticista’ in MOUSA: Scritti in onore di Giusseppe Morelli (Bologna, 1997), 171–7Google Scholar at 176 appear to forget that the aim of Aristophanes' book was to show that some forms and meanings were genuinely classical, and that aspect is not now preserved in Aristophanes' discussion of κόλλοψ, which must necessarily be incomplete.

²² Poehlmann-Tichy (n. 16) attribute most of the subsequent Eustathian discussion to Diogenianus; Hunter, R. L., Eubulus: The Fragments (Cambridge, 1983)Google Scholar, fr. 11 does not attempt to separate the sources.

²³ As pointed out by Hunter (previous note).

²⁴ Since πάσαλος has in comedy and epigram the expected obscene sense of penis, it is perhaps surprising that that specific meaning is never alleged for κόλλοψ. On the other hand it is evident that words like ‘peg’ or ‘screw’ invite sexual innuendo in all languages, and it would seem that there is no obvious need for an explanation in terms of thick skin at all.

²⁵ Sch. Aristophanes Nub. 348; and the gloss must be from comedy: fr. 849K-A.

²⁶ Most recently Hummel, P., ‘Le Labeur et la Gräce’, RPh 70 (1996), 247–54.Google Scholar

²⁷ The first example of the equation is given by Aristophanes grammaticus, fr. 36 c. 200 B.C., by which time it was normal. The scholia minora on Homer, Od. 21.407, like Apollonius Sophistes' lexicon, gloss one word with the other. In Aristophanes and classical authors the word means a cake of some form.

²⁸ For ‘celebrate Aphrodisia’ as the correct meaning, see the examples in LSJ ποιέω A.II.3. For ‘to the accompaniment of, see LSJ s.v. ύπό B.4.

²⁹ The passages are adduced by Poehlmann-Tichy (n. 16), 290: Sch. Aristophanes, Vesp. 574, as well as Lucian and Homer scholia. κόλλοπες are pegs sticking out of a windlass, used as handles to turn it, like a capstan, in ps-Aristotle, Meek prob. 852bl 1 ff.

³⁰ Heron, Behpoiiku 9.11, cited by LSJ.

³¹ Eupolis fr. 55K-A with further references, and Herter (n. 5), 80 with a discussion of known prices. Theognis 560 first uses the metaphor but in a very general sense of a woman having ‘another harbour’ = ‘an affair’.

³² Zagagi, N., Tradition and Originality in Plautus = Hypomnemata 62 (Göttingen, 1980), 125CrossRefGoogle Scholar, n. 72. writes: ‘There is nothing to be gained by comparing Asin. 159 with Eup. fr. 48K [i.e. 55K-A] …: the latter may not refer at all to admission into a lupanar as Kock suggested … indeed as the fragment stands there is no means of knowing whether ellimenion is not meant in a purely literal sense.’ She does not mention lines 241 ff., as far as I see.

³³ Cf. A.P. 9.41.

³⁴ Menaechmi 344.

³⁵ Ussher, R. G., Aristophanes Ecclesiazusae (Oxford, 1973)Google Scholar, ad loc. gives the nautical parallels but misses the point, as does Murgatroyd, P., ‘The sea of love’, CQ 45 (1995), 7ff.CrossRefGoogle Scholar, who, while claiming to provide the first comprehensive and detailed survey of the topos, does not mention Eupolis fr. 55K-A (or indeed Kassel-Austin), and on this passage concludes with neither sense nor sensitivity: ‘there may even be allusion to the vagina in the harbour mouth of 1107’ (p. 11).

³⁶ Prop. 4.5.50; A.P. 5.159 attributed to Simonides; Diphilus fr. 42.22 with the note of Kassel-Austin, referring to Herondas 2, and Headlam ad loc.

³⁷ Priapus is the harbour god par excellence, according to several epigrams at the beginning of A.P. 10.

³⁸ Preston (n. 2 ), 49.

³⁹ Alcaeus, fr. 306 Voigt (mistakenly given as Alcman by Henderson [n. 14], 161); despite the difficulties, it should not be doubted that Alcaeus did deal with a prostitute-ship.

⁴⁰ I dealt with the interaction of sympotic erotic and other metaphorical ships in ‘Symposium at sea’, HSCPh 80 (1976), 161–170, incidentally with regard to Horace's Epode 7; that point is accepted, albeit grudgingly, by Cavarzere, A., Orazio: il Libro degli Epodi (Venice, 1992)Google Scholar, 173 but the concept and its implications remain unknown to Mankin, D., Horace: Epodes (Cambridge, 1995)Google Scholar and many others. Gargiulo, T., ‘Maree Vino nei Persiani: una conjettura a Timoleo fr. 791, 61–2 P.’, QUCC 54 (1996), 73–81Google Scholar gives a bibliography of the metaphor since 1976.

⁴¹ Zenobius 4.56; Hesychius K3585; Suidas κ2067; it is no t mentioned in Von Prittwitz Gaffron, E., Das Sprichwort im griechischen Epigramm (diss. Munich, 1911).Google Scholar

⁴² A s W. Bühler in his editions of Zenobius has made abundantly clear.

⁴³ GVI 1255, 1260, 1263, and other selected examples on pp. 370ff.; A.P. 7.322, 74, 695. 698; 11.364; 16.154, 178, etc. On the opening (Aufforderung zum Ansehen), see Citroni's comments on Martial, Ep. 1.24.1, and especially Grewing, F., Martial: Buch VI (Göttingen, 1997)CrossRefGoogle Scholar, on Ep. 6.73.5, who does not mention this example, because it does not fit the usual pattern. Di Castri (n. 4), 61 believes oddly that the use is paratragic, esp. Euripidean.

⁴⁴ Rouse, W. H. D., Greek Votive Offerings (Cambridge, 1902), 322ff.Google Scholar gives the inscriptional terminology, e.g. ἀνατίθημι δῶρον, ϰαρι⋯ριον The deification of the beloved is of course a hackneyed theme; examples and bibliography in Flury (n. 2), 94ff.

⁴⁵ Kleinknecht, H., ‘Zur Parodie des Gottmenschentums bei Aristophanes’, ARW 34 (1937), 294ff.Google Scholar, esp. 309, though Deubner, L., ‘Ein Stilprinzip hellenistischer Dichtkunst’, NJb 47 (1921), 365ff.Google Scholar is fashionably cited (Bulloch on Callimachus, Hymn 5, line 107, n. 2; Mineur on Callimachus, Hymn to Delos, line 153; Heath, ClAnt 7 [1988] 78; etc.). But even Deubner admitted that the ‘futuristic’ expression was part of an old hymnal topos; see the material in Führer, R., Form-problem Untersuchungen zu den Reden in der frühgriecnischen Lyrik (Munich, 1967), 57Google Scholar, 61, 77; and less clearly Nisbet-Hubbard on Horace, Odes 1,105, 188ff.

⁴⁶ Compare Radermacher's note on the futures at the end of the agon in his commentary on Aristophanes' Frösche², 341: ‘Euripides speaks as a prophet’.

⁴⁷ For the overall for m (‘if you do X …: but if you do Y …’), compare the oracle of the Pythian Apollo to Sybaris reported by Timaeus, 566 FGrH 50. Another example is given by Merkelbach, R. and Strauber, J., ‘Die Orakel des Apollon von Klaros’, EA 27 (1996), 1ff.Google Scholar, no. 11, lines 10,18: ‘if you do as I say … if you do not do as I say …’. There will also be found examples for the kenning-vocatives common in such oracles.

⁴⁸ IG XII 3, 1125, 1089; Smith, C., ‘Inscriptions from Melos’, JHS 17 (1897), 14CrossRefGoogle Scholar; Bosanquet, R. C., ‘Excavations of the British school at Melos: the hall of the Mystae’, JHS 18 (1898), 60–80.CrossRefGoogle Scholar

⁴⁹ Eisler, R., Orphische-dionysische Mysterien-gedanken in der christlichen Antike (Bibl. Warburg, 1922/3), 102ff.Google ScholarGeyer, A., Das Problem des Realitätsbezuges in der dionysischen Bildkunst der Kaiserzeit (Würzburg, 1977), 140–1Google Scholar with a long discussion. Moormann, E. M., ‘Imperial Roman mosaics at Leiden’, OMRO 71 (1991), 97–107Google Scholardeals especially with the complex excavation history; P. Asimakopoulou-Azaka, ‘Κατάλογος Ρωμαικών ψηϕιδωτών με ανθρώπινες μορϕές’, Hellenika 26 (1973), 216–54, no. 41. There is a brief discussion of kurtai in Kankeleit, A., Kaiserzeitliche Mosaiken in Griechenland, 2 vols (diss. Munich, 1994), 65–8Google Scholar, who thinks of a glass bottle-trap, no t really a practical possibility in antiquity. I a m grateful to D r Z. Welch and Professor K. M. D. Dunbabin for help with this bibliography.

⁵⁰ Blanchard-Lemée, M. et al., Sols de L'Afrique Romaine (Paris, 1995)Google Scholar, figs 81, 106, 108. In some versions ‘la Tunisie’ is substituted for ‘L'Afrique’ in the title.

⁵¹ See especially Alpers, K., Das allizistische Lexikon des Oros (Berlin, 1981)CrossRefGoogle Scholar, fr. 136 and the references there. This word raises thoughts of connections with Phorkys. Speculations about this, a fishing Dionysus, and much else can be found in R.E. XXII.1 (1953), s.v. Porkos, Porkeus (Radke). I am not convinced that a discussion of these matters would advance the argument here.

⁵² Diogenian 4.65 Schn.-Leutsch derived from Plato.

⁵³ Festus, p. 169M.

⁵⁴ Geyer (n. 49), 141.

⁵⁵ Horace, Odes 3.10, and cf. Athenaeus 426Bff. But at this period, the Roman fashion was to ask the boy to put hot or cold water into one's wine, and the old Greek fashion of mixing in a krater did not apply. ‘No water’ means that one would drink pure, undiluted wine: no bad thing with the fine wines of imperial times, and a compliment to Dionysus.

⁵⁶ Poinssot, C., ‘Quelques remarques sur les mosaiques de la Maison de Dionysos et d'Ulysse à Thugga (Tunisie)’, in La Mosaique gréco-romaine. Colloques internationaux du C.N.R.S. Paris 1963 (Paris, 1965), 219–30Google Scholar, figs 16–21. Blanchard-Lemée (n. 50), 119, fig. 79 gives an excellent two-page photograph.

⁵⁷ For water-into-wine as a result of Dionysus' passage, see Slater, W. and Steinhart, M., ‘Phineus as monoposiast’, JHS 117 (1997), 203–11.Google Scholar

I do not feel competent to explore here the iconography of Venus piscatrix (LIMC III.l, s.v. Eros 1006–7 and esp. 921, no. 842g), save to note that it goes back to the fourth century B.C. and that LIMC is mistaken to parallel Ovid, Ars. Am. 1.47–8 for an image of Venus with a fishing rod.