¹ The fragment is preserved by Aelian, N.A. 12.9. I give the text of Kassel, R. and Austin, C., Poetae Comici Graeci, iii.2: Aristophanes (Berlin, 1984), pp. 47–8, fr. 29.Google Scholar

² Bergk, T., Aristophanis Fragmenta (Fragmenta Comicorum Graecorum, 2.2) (Berlin, 1847), p. 67Google Scholar: ‘Haec autem ad modum oraculi praecipiuntur mulieri, ut senis robur excitet.’

³ The infinitive is printed, e.g., in the editions of Kock (1880–8), Hall–Geldart (1920) and Edmonds (1957), all of whom follow Bergk's interpretation of the fragment as an oracle. The Loeb translation of Scholfield, A. F., Aelian: On the Characteristics of Animals, iii (Cambridge MA, 1959), p. 25Google Scholar provides a good illustration of this approach: ‘Give the old man's loins a thorough shaking, as the wagtail does, and work a powerful spell.’ Bergk and Kock give several examples from comic authors of an infinitive used as an imperative, including one instance in a comic oracle (Eq. 1039).

⁴ Blaydes, F. H. M., Aristophanis deperditarum comoediarum fragmenta (Hal. Sax., 1847), pp. 14–15Google Scholar, printed τελ⋯ει (but see below n. 26) and suggested that Amphiaraus is the old man seeking the oracle and that the subject of the second verb is the action prescribed by the first verb: ‘Lumbum autem senis (sc. Amphiarai) tanquam cincli (pennas caudinas) a summo exagita seu concute: namque id facit potentem incantationem.’ In this century Taillardat, J., Les images d'Aristophane: Études de langue et de style (Paris, 1962), pp. 106–7Google Scholar, defended the MS. reading, and translated as follows: ‘Comme une bergeronnette, remue le bout de ta croupe pour aider le vieillard. Amphiaréôs rendra l'incantation efficace.’ He is followed by Henderson, J., The Maculate Muse: Obscene Language in Attic Comedy (New Haven, 1975), p. 179Google Scholar. Kassel–Austin, op. cit. (n. 1), print the MS. reading τελ⋯ει. The switch from the second-person imperative to the third-person indicative (future or present) is awkward but preferable (according to Taillardat and Kassel–Austin), since the action described by τελ⋯ει ⋯παοιδ⋯ν is more suitably undertaken by a divinity.

⁵ Aside from the disagreement over the identity of the old man and the subject of the second verb (both mentioned above in note 4) I cite, e.g., two other minor discrepancies: (1) Bergk, op. cit. (n. 2), emends the final word in the first line to the nominative (κ⋯γκλος), a change challenged by Kock (‘nam non suam senex, sed mulier senis coxam quassare iubetur’) and resisted by most editors; and (2) Henderson, op. cit. (n. 4), 179, combines Bergk's κ⋯γκλος with an emendation proposed by Taillardat (⋯νδρ⋯ πρεσβ⋯τῃ) and translates the first part of the verse as follows: ‘Wiggle your ass like a wagtail against the old man.’

⁶ The following abbreviations will be used throughout for the basic corpora of ancient Greek magical texts

DT = Audollent, A., Defixionum Tabellae (Paris, 1904).Google Scholar

DTA = Wünsch, R., Defixionum Tabellae Atticae, Appendix to Inscriptiones Graecae, iii (Berlin, 1897).Google Scholar

Suppl. Mag. = Daniel, R. W. and Maltomini, F.Supplementum Magicum, i (Papyrologica Coloniensia, 16.1) (Cologne, 1990).Google Scholar

PGM ² = Preisendanz, K. and Henrichs, A., Papyri Graecae Magicae: Die griechischen Zauberpapyri² (Stuttgart, 1973–1974).Google Scholar

The numbers following these abbreviations refer to the number of the texts in the collection, unless indicated otherwise.

⁷ Brashear, W. M., ‘Ein Berliner Zauberpapyrus’, ZPE 33 (1979), 261–78Google Scholar; Maltomini, F., ‘P. Berol. 21243 (formulario magico): due nuove letture’, ZPE 74 (1988), 247–8Google Scholar; and Janko, R., ‘Berlin Magical Papyri 21243: A Conjecture’, ZPE 72 (1988), 293Google Scholar. Prof. Maltomini has kindly provided me with a copy of his new edition of the text (it will appear in the second volume of Suppl. Mag. as text no. 72), which I print here with one exception: I prefer to retain Brashear's ϕιλεῖν in the penultimate line to Maltomini's ϕιλ⋯ν.

⁸ There are, in fact, indications both in much earlier Greek myth and wedding ritual, and in Neo-Assyrian magical ritual as well, that the particular type of ritual employed here (the throwing or presentation of apples) was exceedingly old; see Faraone, C. A., ‘Aphrodite's KEΣTOΣ and Apples for Atalanta: Aphrodisiacs in Early Greek Myth and Ritual’, Phoenix 44 (1990), 233–6.CrossRefGoogle Scholar

⁹ All of the parallels that I discuss below are listed by Brashear, op. cit. (n. 7), 268, whose comments are in turn cited by Kassel–Austin, op. cit. (n. 1), 48.1 have ignored other obviously related expressions that appear in magical texts, such as, e.g.: τ⋯λει τ⋯ν προγνωστικ⋯ν (PGM² iii.194); τ⋯λει τελε⋯αν τελετ⋯ν (xii.306); τ⋯λει τ⋯ μ⋯γα ϕυλακτ⋯ριον (iv.1690); τ⋯λεσον τ⋯ ⋯νγεγραμμ⋯να (xixa.16); or ἔλθετε κα⋯ τελειώσατ⋯ μοι τ⋯ν πραγματε⋯αν τα⋯την (DT 38.14–15). As for the adjective ⋯γαθ⋯ς with ⋯παοιδ⋯, see the boast at Euripides’ Cyclops 646: ⋯λλ᾽ οἶδ᾽ ⋯πῳδ⋯ν Ὀρϕ⋯ως ⋯γαθ⋯ν π⋯νυ.

¹⁰ PGM ² xx.4: τ[⋯λεσον τε]λ⋯αν ⋯παοιδ⋯ν. Maas, P., ‘The Philinna Papyrus’, JHS 62 (1942), 33–8CrossRefGoogle Scholar, reconstructed the text by joining P. Berol. 7504 and P. Amherst ii, col. II(A). For the most recent commentary, see Suppl. Hellenisticum no. 900. The spell in question is usually assumed to be hexametrical, because its fragments seem to be of that metre and because the two other charms preserved in the papyrus are hexametrical, leading to the plausible hypothesis that this particular handbook was a collection devoted solely to hexametrical spells.

¹¹ PGM ² iv.294–5: τελ⋯σατ⋯ μοι τ⋯ν τελε⋯αν ⋯παοιδ⋯ν.

¹² P.Oxy 412.14, a curious papyrus fragment of the third-century A.d. author Julius Africanus, who quotes a wildly expanded version of the necromancy in Odyssey 11, in which Odysseus asks the rivers, earth and those who punish oath-breakers to fulfil his charm. The lines seem to be a transplanted version of Il. 3.278–80 (part of the oath scene between the Trojans and the Greeks), in which the second hemistich of line 280 (ϕυλ⋯σσετε δ᾽ ὅρκια π⋯στα has been replaced with a variation of the coda under discussion: τελε⋯ετε δ᾽ ἄμμιν ⋯οιδ⋯ν. A few lines later (line 25) Anubis, Hermes and Zeus are invoked and commanded: κρη⋯νατε τ⋯νδ᾽ ⋯παοιδ⋯ν. For text and general discussion, see Wünsch, R., ‘Der Zaubersang in der Nekuia Homers’, ARW 12 (1909), 2–19Google Scholar and Vieillefond, J.-R., Les ‘Cestes’ de Julius Africanus: Étude sur l'ensemble des fragments avec édition, traduction et commentaire (Florence, 1970), pp. 277–91.Google Scholar

¹³ Brashear, op. cit. (n. 7), 263, col. i.26–7 (π⋯τνια θε⋯,…τ⋯λεσ⋯ν μ[οι] τελ⋯αν ⋯παοιδ⋯ν) and col. ii.8 (δ⋯σ[ποι]να Ἶσι, τ⋯λει τελ⋯αν ⋯παγοιδ⋯ν, and 264, col. ii.25 (Kυπρογ⋯νεια τ⋯λει τελ⋯αν ⋯παοιδ⋯ν). See Brashear's comments ad loc. for the probable identity of the π⋯τνια θε⋯ at col. i.26 as Isis. Although these three spells are all very fragmentary, each time the phrase appears, it is marked with a peculiar ‘Schlangelinie’ and followed by an ek thesis or indentation of the following line – clear indications that in each instance these are the final words of the incantation (see Brashear, ibid. 262 for discussion).

¹⁴ Maltomini, F., ‘P. Mon. Gr. Inv. 216: Formulario magico’, in Carlini, A. et al. Papiri letterari greci (Biblioteca degli studi classici e orientali, 13) (Pisa, 1978), pp. 237–66Google Scholar, no. 34–republished in Carlini, A. (ed.), Papiri letterari greci della Bayerische Staatsbibliothek di Monaco di Baviera (Stuttgart, 1986), pp. 40–55Google Scholar, no. 28 –, argues persuasively (p. 247) for a join between two small papyrus fragments of a first-century B.C. magical handbook that yields τελ]⋯αν ⋯πα ọιδ⋯ν, after which the papyrus is blank, an indication that this phrase marks the end of the spell. The spell itself is fragmentary, but (as Maltomini notes) the mention of alienating someone from her husband and other vocabulary points inescapably to an erotic spell. Prof. Maltomini has kindly sent me his new edition of P.S.A. Athen. 70, an erotic charm of the first century A.d. (or perhaps, as Carlini suggested per litteras to Maltomini, late first century B.c.), which will appear as text no. 73 in the forthcoming second volume of Suppl. Mag. He restores the final line as follows: [τελε⋯]αν ⋯πạ[οιδ⋯ν].

¹⁵ The four late-antique examples are: PGM ² iv.2939 (σὺ δ⋯, Kυπρογ⋯νεια θε⋯, τ⋯λει τελ⋯αν ⋯παοιδ⋯ν, which is printed with one minor change on p. 261 of the second volume of PGM ² as part of a hexametrical hymn to Aphrodite reconstructed by Heitsch: {σὺ δ⋯} Kυπρογ⋯νεια θε⋯, ‹σὺ› τ⋯λει τελ⋯αν ⋯παοιδ⋯ν); PGM ² vii.992 (καλ⋯ς μο[ι τ⋯λει τα⋯την τ⋯ν ⋯πα]οιδ⋯ν); an erotic defixio, Suppl. Mag. 45.32 (τελῖτε τελ⋯αν τ⋯ν ⋯παϋδ⋯ν); and the final line of an erotic spell in a new Munich papyrus soon to be published by W. Brashear in Studien zur altägyptischen Kultur (…]ε τελ⋯αν τ⋯⋯ṿ [⋯παϋ]δ⋯ṿ).

¹⁶ Brasheari op. cit. (n. 7), ad loc, notes both of these parallels without comment.

¹⁷ Cameron, A., ‘Sappho's Prayer to Aphrodite’, HTR 32 (1939), 8–10CrossRefGoogle Scholar, Segal, C., ‘Eros and Incantation: Sappho and Oral Poetry’, Arethusa 7 (1974), 148–9Google Scholar, and Burnett, A.P., Three Archaic Poets: Archilochus, Alcaeus, Sappho (London, 1983), pp. 254–5Google Scholar. Cameron also points out that Aphrodite's claim to lead people to love even when they are unwilling (line 24: κωὺκ ⋯θ⋯λοισα) appears in the hymn to Aphrodite, op. cit. (n. 15), embedded in an elaborate erotic spell (PGM ² iv.2935: οὺκ ⋯θ⋯λοντα).

¹⁸ In the magical incantations spoken by the human practitioner, the imperative or optative appears (as one would expect) instead of the divinely assured future tense employed by Aphrodite in Sappho's poem. Cameron, op. cit. (n. 17), 8, cited the long list of such pairs in PGM ² iv.1510–20: εἰ κ⋯θηται, μ⋯ καθ⋯σθω, εἰ λαλεῖ πρ⋯ς τινα, μ⋯ λαλε⋯τω, εἰ ⋯μβλ⋯πει τιν⋯, μ⋯ ⋯μβλεπ⋯τω, εἰ προσ⋯ρχετα⋯ τινι, μ⋯ προσερχ⋯σθω, εἰ περιπατεῖ μ⋯ περιπατε⋯τω, εἰ π⋯νει, μ⋯ πιν⋯τω, εἰ ⋯σθ⋯ει, μ⋯ ⋯σθι⋯τω, εἰ καταϕιλεῖ τινα, μ⋯ καταϕιλε⋯τω, εἰ τ⋯ρπετα⋯ τινι ⋯δονῇ μ⋯ τερπ⋯σθω, εἰ κοιμ⋯ται, μ⋯ κοιμ⋯σθω, ⋯λλ᾽ ⋯μ⋯ μ⋯νον τ⋯ν δεῖνα κατ⋯ νο⋯ν ⋯χ⋯τω. Other shorter examples include: PGM ² iv.2740–2 (εἰ δ⋯ τινα ἄλλον ἒχουσ᾽ ⋯ν κ⋯λποις κατ⋯κειται, κεῖνον ⋯πωσ⋯σθω); Suppl. Mag. 45.49–50 (εἰ δ⋯ κα⋯ ἓτερο[ν] ἒχει ⋯ν κ⋯λποις, ⋯κεῖνον μ⋯ν ὑπερθ⋯σθω); and P.S.A. Athen. 70 as restored by Maltomini, F., ‘Osservazion al testo di alcuni papiri magici greci I’, in Pintaudi, R. (ed.), Miscellanea Papyrologica (Papyrologica Florentina, 7) (Florence, 1980), p. 172Google Scholar (⋯⋯ν καθε⋯δῃ μ⋯ [καθευδ⋯τω, ⋯⋯ν ϕ⋯γῃ μ⋯]ϕαγ⋯τω, ⋯⋯ν π⋯ν[ῃ μ⋯ πιν⋯τω). The last example is the most important parallel as it dates to the first-century A.d. or earlier and it ends with the traditional coda (see note 14 above). Giacomelli, A., ‘The Justice of Aphrodite in Sappho Fr. 1’, TAPA 110 (1980), 135–42Google Scholar, argues that Aphrodite's words do not refer to a specific individual (and thus do not constitute a promise), but that they are simply the statement of a natural law of erotic relationships, i.e. that all who are sought shall some day be seeking. Drawing on some homoerotic parallels in the Palatine Anthology (all of which deal with the unavoidable passage in status from erômenos to erastês) she suggests that the absence of specific names or pronouns in the penultimate stanza of Sappho's poem is a deliberate device of the poet to stress the universal nature of Aphrodite's pronouncements. The parallels, however, from the magical spells quoted above indicate that such incantations are similarly generic and vaguely worded; thus no pronouns or names are used to indicate the victim (e.g. εἰ κ⋯θηται, μ⋯ καθ⋯σθω) and indefinite pronouns (not names) are used to indicate the potential rivals for her affections (e.g. εἰ λαλεῖ πρ⋯ς τινα, μ⋯ λαλε⋯τω). For this peculiarity of magical formulae, see Graf, F., ‘Prayer in Magic and Religious Ritual’, in Faraone, C. A. and Obbink, D. (eds.), Magika Hiera: Ancient Greek Magic and Religion (Oxford, 1991), p. 190.Google Scholar

¹⁹ Segal, op. cit. (n. 17), 158 n. 16, notes how the placement of the caesura after the fifth syllable of lines 21 and 23 emphasizes the repetition of ταχ⋯ως. The doubling ἣδη, ἣδη, ταχ⋯, ταχ⋯ is extremely common in magical incantations, especially erotic ones, see, e.g.: PGM ² i.262; iii.35, 85 and 123; iv.973, 1593 and 2037; vii.248, 330, 410 and 993; viii.52 and 84; x.50; and xliii.27. One also finds ταχ⋯ως (PGM ² i. 107; iv.72, 384 and 1265) and τ⋯χιστα (PGM ² iv.2619, 2742, 2757 and 2782).

²⁰ The phrase κατατρ[⋯]χω, αὐτ⋯ς δ⋯ ϕε⋯γει appears in one of the erotic incantations in the Berlin spell that ends with the traditional coda (Brashear, op. cit. [n. 7], col. ii.9–25). Winkler, J. J., ‘The Constraints of Eros’Google Scholar, in Faraone and Obbink, op. cit. (n. 18), 239 n. 55, suggests that this spell seems to employ the same language of pursuit and flight used by Sappho, another indication, perhaps, that Sappho is reflecting a traditional erotic spell. The phrase, however, appears in a very fragmentary part of the papyrus and its wider context is unknown. See also Levi, P., ‘The Prose Style of the Magical Papyri’, in Proceedings of the XIV International Congress of Papyrologists (London, 1975), 215 n. 21Google Scholar, who suggests that Aphrodite's question in line 20 of Sappho's poem (τ⋯ς σ᾽ …⋯δικ⋯ει;) recalls complaints like that expressed in a Cnidian defixio of Hellenistic date (DT 2): ⋯δ⋯κημαι γ⋯ρ, Δ⋯σποινα Δ⋯ματερ, a rhetorical strategy common to a type of Greek imprecation known as a ‘judicial prayer’; see Versnel, H. S., ‘Beyond Cursing: The Appeal to Justice in Judicial Prayers’Google Scholar, in Faraone and Obbink (n. 18), 65–9, who discusses the Cnidian text and three curses of similar date from Attica, including, e.g., DTA 98 (ϕ⋯λη Γ⋯, βο⋯θει μοι. ⋯δικο⋯μενος γ⋯ρ…).

²¹ Or Isis, her Egyptian counterpart from the late Hellenistic period onward; see Brashear, op. cit. n. 7), 268–9, for a brief discussion of the later syncretism of Isis and Aphrodite and some select bibliography.

²² The usual procedure in magical and other prayers seems to be to ask the deity in the beginning to come and listen to the prayer or spell, and then request at the end that she bring to fulfilment what she has just heard. See, e.g., an elaborate hexametrical hymn to Selene that begins with ⋯λθ⋯ μοι…εὐμεν⋯ῃ δ᾽ ⋯π⋯κουσον ⋯μ⋯ν ἱερ⋯ν ⋯παοιδ⋯ν (PGM² iv.2786–8) and ends with κα⋯ μοι τ⋯δε πρ⋯γμα πο⋯ησον (2871–2).

²³ See, e.g., the Berlin papyrus published by Brashear, op. cit. (n. 7), col. i.27: τ⋯λεσ⋯ν μοι τελ⋯αν ⋯παοιδ⋯ν, or PGM ² vii.992: μο[ι τ⋯λει τα⋯την τ⋯ν ⋯πα]οιδ⋯ν.

²⁴ See above, notes 10–15. I disqualify the two fragmentary spells discussed in n. 14 (P. Mon. Gr. 216 and P.S.A. Athen. 70) and the new Munich papyrus (quoted in n. 15), where in each case lacunae make it impossible to know the construction of the verb. I am, however, in complete agreement with the editors of these spells (Maltomini and Brashear, respectively, ad loc.) who assume that some form of imperative was employed.

²⁵ For Bergk's emendation and its influence, see above, note 3. Kassel and Austin, op. cit. (n. 1), 48, report the two objections of Wilamowitz (see below, n. 26): (1) that Aristophanes would not have used the uncontracted form τελ⋯ειν; and (2) that the use of a present infinitive after an aorist imperative was impossible. The first point is decisive since even if we argue that Aristophanes is mimicking some archaic usage peculiar to traditional hexametrical compositions, we run into the problem that in hexametric verse the form τελ⋯ειν appears to be used only to indicate the asigmatic future, e.g.: ἒμελλεν ⋯κτελ⋯ειν (Od. 10.26–7) or ⋯λπ⋯μενοι τελ⋯ειν (Hes. fr. 204.85 M–W).

²⁶ Blaydes, who printed τελ⋯ει in his 1847 edition of the fragments, op. cit. (n. 4), much later in his life (Adversaria in comicorum graecorum fragmenta, ii [Hal. Sax., 1896], p. 55Google Scholar) suggested τ⋯λεσον, an emendation approved by Wilamowitz, , Griechische Verskunst (Darmstadt, 1921), p. 349 n. 1Google Scholar. The form τ⋯λεσον can in fact be defended by its appearance in the magical tradition discussed above, beginning with Sappho's prayer to Aphrodite and then again (in a hexametrical context) in two of the earliest magical papyri, PGM ² xx (n. 10) and the Berlin papyrus (n. 7), which both date to the first century B.c.

²⁷ Here, too, the magical tradition discussed above provides support. The present imperative is by far the most popular and it is attested early both in conventional prayers (e.g. Clytemnestra's appeal to Zeus quoted above) and in the magical tradition, appearing thrice in the first-century B.C. Berlin papyrus (n. 7) and twice in late-antique spells: PGM ² vii.992 (n. 14); and PGM ² iv.2939, the final line of a fairly polished hexametrical hymn to Aphrodite (n. 15): Kυπρογ⋯νεια θε⋯, ‹σ⋯› τ⋯λει τελ⋯αν ⋯παοιδ⋯ν. We can in fact more easily reconcile the present imperative to the MS. readings (τελ⋯ει and τελεῖ) than we can the aorist τ⋯λεσον. E. Courtney and R. Janko both suggest (per litteras) the addition of σ⋯ (as in the hexameter quoted above from PGM ² iv.2939) to restore the metre: ⋯νδρ⋯ς πρεσβ⋯του. ‹σὺ› τ⋯λει δ᾽ ⋯γαθ⋯ν ⋯παοιδ⋯ν.

²⁸ PGM ² xixa.50. The spirited translation is that of O'Neil, E. and Kotansky, R. in Betz, H. D. (ed.), The Greek Magical Papyri in Translation, i (Chicago, 1986), p. 257Google Scholar. For similar examples of such exhortations to attack, see, e.g.: PGM ² iv. 1540–1 (κατ⋯καυσον τ⋯ν ⋯νκ⋯ϕαλον, ἔκκαυσον κα⋯ ἒκστρεψον αὐτῇς τ⋯ σπλ⋯γχνα ); PGM ² iv.2767 (ϕλ⋯ξον ⋯κοιμ⋯τῳ πυρ⋯ τ⋯ν ψυχ⋯ν); PGM ² xvi (κα⋯σον τ⋯ν καρδ⋯αν); PGM ² O[stracon] 2.29–30 (κα⋯σον, π⋯ρωσον τ⋯ν ψυχ⋯ν); Suppl. Mag. 42.37–8 (βασαν⋯σατε αὐτ⋯ς τ⋯ σ⋯μα); 45.31–2 (κα⋯σατε αὐτ⋯ς τ⋯ μ⋯λη, τ⋯ ἧπαρ, τ⋯ γυνεκῖον σ⋯μα), DT 51 (κατακα⋯νετε…ψυχ⋯ν κ⋯ τ⋯ν καρδ⋯αν); and DT 271.12–14 (⋯γαγεῖν κα⋯ ζε⋯ξαι τ⋯ν Oὐρβαν⋯ν…βασανιζ⋯μενον).

²⁹ See, e.g., some of the examples quoted above in note 27.

³⁰ See Taillardat, op. cit. (n. 4), and Henderson, op. cit. (n. 4).

³¹ See LSJ s.v. σεισοπυγ⋯ς and Thompson, D. W., A Glossary of Greek Birds² (Oxford, 1936), pp. 140–1Google Scholar, who quotes most of later sources in full and in the original.

³² See, e.g., Kaimakis, D., Die Kyraniden (Beiträge zur klassichen Philologie, 76) (Meisenheim am Glan, 1976), pp. 63–4Google Scholar (κινα⋯διον used in love potions and erotic amulets) and 229 (σεισοπυγ⋯ς used in a love potion).

³³ At Plut. 883–5 the ‘Just Man’ disregards the threat of Carion the sycophant explaining that he is wearing a special ring, to which Carion replies: ⋯λλ᾽ οὐκ ἒνεστι “συκοϕ⋯ντον δ⋯γματος” a reference to magical rings bearing inscriptions such as σκορπ⋯ον δ⋯γματος; see Bonner, C., Studies in Magical Amulets (Ann Arbor, 1950), pp. 4–5Google Scholar, and Kotansky, R., ‘Incantations and Prayers for Salvation on Inscribed Greek Amulets’Google Scholar, in Faraone and Obbink, op. cit. (n. 18), pp. 110–11.

³⁴ At Acharn. 1128–31, when Lamachus orders his slave to pour oil on his shield and then claims that he sees legal problems in the future for Dicaeopolis, he is imitating a form of prognostication performed by gazing into mirrors or still pools of liquid, a practice known elsewhere from Pausanias and the magical papyri; see: the scholia ad loc.; Halliday, W. R., Greek Divination (London, 1913), p. 153Google Scholar; Delatte, A., La catoptromancie grecque et ses dérivés (Liège, 1932), pp. 133–5Google Scholar; and Dodds, E. R., The Ancient Concept of Progress and Other Essays (Oxford, 1973), p. 186 n. 4.Google Scholar

³⁵ I thank J. Henderson for this insight. The traditional suspicion that the fragment is part of an oracle arises naturally enough from its metre and from the fact that there existed near the border with Boeotia a famous oracular shrine of Amphiaraus, after whom the play is named. The authority cited by, e.g., Kassel and Austin, op. cit. (n. 1), 48, is Pausanias 1.34.4, whose testimony is ambiguous at best. He describes the oracular shrine as a place where people spent the night in hopes of getting oracular dreams, not pronouncements. In fact, he specifically dismisses as spurious the hexametrical oracles of Amphiaraus collected by Iophon of Cnosus (otherwise unknown and therefore unable to be dated; see Bouché-Leclercq, A., Histoire de la divination dans l'antiquité, ii [Paris, 1879], p. 225)Google Scholar. But even if Amphiaraus' shrine did not actually produce hexametrical oracles we cannot rule out the possibility that the fragment under discussion was put into the mouth of a manifestly fraudulent oracle-monger, a frequent butt of Aristophanic jokes; see Smith, N., ‘Diviners and Divination in Aristophanic Comedy’, CA 8 (1989), 141–7.Google Scholar

³⁶ I should like to thank W. M. Brashear, E. Courtney, J. Henderson, R. Janko, R. Kaster, R. Kotansky, F. Maltomini and A. Rossius and Z. Stewart for their comments on earlier versions of this paper. Naturally the defects which remain are my own. I owe additional gratitude to Bill Brashear and Franco Maltomini for allowing me access to the preliminary texts of the papyri discussed in notes 7, 14 and 15. This paper was researched and written at the Center for Hellenic Studies during the academic year 1991–2. I owe many thanks to Z. and D. Stewart, the other junior fellows, and the staff for making my stay a particularly enjoyable and productive one.