¹ All references are to Pindarus, pars II, Fragmenta, Indices post B. Snell edidit H. Maehler (Leipzig, 1975; henceforth ‘Snell-Maehler’)Google Scholar, the text of which is followed, unless otherwise stated.

² Fr. 286 = schol. on A. Eum. 11. Schachter, Albert, Cults of Boeotia I, Acheloos to Hera, BICS Supp.38. 1 (1981), p. 57Google Scholar suggests that this fragment was part of a Pindaric ‘suite pythique’.

³ What survives of the title is: Π[ ]‥[ΑΙΣ ΕΙΣ ΔΕΛΟ]Ν, the first letter contributed by P.Oxy. 2440 fr. 1 line 8, the rest preserved in P.Oxy. 2442 fr. 14 line 3a. This conforms to the regular pattern of titles given to Pindar's Paeans by Hellenistic editors which specified (a) the performers (in the dative) and (b) the place of performance (εἰς + accusative). Cf. the titles of Pindar, Pa. VI, VIII in Snell–Maehler; also Simonides [?] PMG 519 fr. 35 line 12. It is likely that these titles were in many cases reliably deduced from the poems and that Paeans often specified the nationality and destination of the chorus.

⁴ The all-important word Ομρου together with the openings of the next two lines is preserved in P.Oxy. 841 fr. 17. Some early editors placed this fragment in (what are now) lines 23ff. Its correct position was determined by Lobel, (P.Oxy. 26, pp. 40–1Google Scholar, with pl. v), who recognised traces of the openings of the same lines in P.Oxy. 2442 fr. 14(b), the position of which in the poem was known. As for the lacuna in line 11, 1 would prefer Ομρου [μρου [μυ οὐ τρι] πτν…, first suggested by Newman, J. K., ‘Callimachus and Pindar’, ICS 10 (1985), 182Google Scholar to Lobel's Ομρου [δ μ τρι] πτ… The trouble with the latter is that Pindar does not use δ to introduce hanging circumstantial clauses. Grammar requires μ with an imperative, but the scope of οὐ is τριπ] τν κττ' μαξιτν/ἰντες, not the imperative.

⁵ The passages are P. 4.277ff., I. 3/4.55ff., N. 721ff.;. Fitch, E., ‘Pindar and Homer’, CPh. 19 (1924), 57ff.Google Scholar, argued that I. 3/4.55ff. and N. 7.21ff. could refer to the account of the quarrel between Odysseus and Ajax in the Aithiopis, but it seems more likely that Pindar is thinking of the favourable portrayal of Ajax in the Iliad.

⁶ 3.104.4–6: he attributes lines 146ff. and 168ff. to West, Homer. M. L. (‘Cynaethus' Hymn to Apollo’, CQ 25 [1975], 166)CrossRefGoogle Scholar suggests that belief in Homeric authorship of the Hymn is also indicated by Ar, . Birds 575Google Scholar: Ἰριν δ γ' “Ομηρος ἔφασκ' ἰκλην εὖναι τρρωνι πελειῃ which must allude to h. Horn. Ap. 114, where Iris and Eileithuia are compared to τρρωσι πελειοω, since this simile is not used of Iris in the Iliad or Odyssey. Since all of this evidence relates to the Delian part of the Hymn, it remains possible that the Pythian part was not regarded as the work of Homer in the 5th century. Perhaps it was thought to have been added by a rhapsode such at Cynaethus (cf. schol, . Nem. II.lc, 29.9ff.Dr.)Google Scholar. I note that Janko, R., Homer, Hesiod and the Hynat (Oxford, 1982)Google Scholar, has recently reasserted the view (contra West, op. cit.) that the Delian part of the Hymn is earlier than the Pythian part, concluding that the former dates from the 7th century (pp. 114–15) and the latter from 585 B.C. (p. 132).

⁷ Slater, W. J., Lexicon to Pindar (Berlin, 1969)CrossRefGoogle Scholar, s. ἴππο‘| ‘The First Person in Pindar’, HSCP 67 (1963), 155ffGoogle Scholar. The rule does not seem to apply to Pa. VI, though see Hoekstra, A., ‘The Absence of the Aeginetans’, Mn. 15 (1962), 9ffGoogle Scholar. Hamilton, R., Epinikion (De proprietatibus iitterarum, series practica 91), 113ff.Google Scholar, defends the implausible view that the first persons at Pa. 11.28 and Pa. IV.21ff. represent places speaking.

⁸ See Lefkowitz, M. R., ‘The First Person in Pindar’, HSCP 67 (1963), 155ff.Google Scholar The rule does not seem to apply to Pa. VI, though see Hoekstra, A., ‘The Absence of the Aeginetans’, Mn. 15 (1962), 9ff.Google Scholar R. Hamilton, Epinikion (De proprietatibus iitterarum, series practica 91), 113ff., defends the implausible view that the first persons at Pa. 11.28 and Pa. IV.21ff. represent places speaking.

⁹ Cf. in particular the contrast between new and old styles of dithyramb in fr. 70b. Hardie, A., ‘Horace Odes 1.37 and Pindar Dithyramb 2’, PLLS 1 (1976), 113ff.Google Scholar, thinks that fr. 70b makes a point about subject matter also.

¹⁰ Cf. Norden, E., Die antike Kunstprosa ⁵ i.33 n. 3Google Scholar, Aristides 45.8 with Höfler, A., Der Sarapishymnus der Aelius Aristides, Altertumswissenschaft, Tübingen Beiträge zur 27 (1935), 30Google Scholar.

¹¹ See Köhnken, A., Die Funktion des Mythos (Berlin, 1971), pp. 50ffCrossRefGoogle Scholar.

¹² For example N. 4.69, P. 11.38, P. 4.247, N. 6.53, O. 9.80. On the metaphor of the path of poetry, see Becker, O., Das Bilddes Weges, Einzelschriften, Hermes IV (Berlin, 1936), 58ffGoogle Scholar. (68ft. on Pindar), and Durante, M., Sulla Preistoria della Tradizione Poetica Greca, Pane seconda, Graeca, Incunabula 64 (Rome, 1976), 123ffGoogle Scholar.

¹³ See Van Groningen, B. A., Gnomon 35 (1963), 128Google Scholar; Richardson, N., ‘Pindar and Later Literary Criticism’, PLLS 5 (1985), 393ff.Google Scholar; Newman, , op. cit., p. 182Google Scholar. Callimachus, , Aitia 1.25ffGoogle Scholar. is as follows (Apollo speaks): πρς δ σε] κα τδ' ἄνωγα, τ μ πατους ἄμαξαι | τ στεβειν, τρων δ' ἴχνια μ καθ' μ|δφρον λ]ν μηδ' οἶμον ν πλατὺν, λλ κελεὺθους | τρπτο]υς, εἰ κα στεινοτρην λσεις (In addition, I bid you to tread where the wagons do not go, to drive your chariot neither along the even tracks of others nor along the broad paths, even if that way you drive a narrower path). In particular, τρπτο]υς may be presumed to pick up τρι]πτν in Pa. VIIb. 11 and τ μ πατουσιν ἄμαξαι to echo μαξιτν in the same line. We may reflect that it is particularly appropriate that Apollo should speak these lines if their model was a contrast (in a paean) between two different poetic treatments of his own birth. Pa. VIIb was also known by Plato, if Friedländer, ‘Plato Phaedrus 245a’, CPh 36 (1941), 51–2Google Scholar, was right to recognise an imitation of lines 19–20 at Phaedrus 245a: …ὅς ἄνευ μανας Μουσν π ποιητικς θὺρας φκητα' poetic programme covered both form and content. It is indicated, for example, by Ep. 27.1: Ἡσιδον τ τ' ἄεισμα κα τρπος, where τρπος could refer to form and ἄεισμα to content.

¹⁵ I take it as uncontentious that Callimachus' poetic programme covered both form and content. It is indicated, for example, by Ep. 27.1: ‘, where τρόπος could refer to form and ἂηισμα to content.

¹⁵ Alcaeus' Hymn to Apollo is not likely to have been a serious rival and in any case Himerius, , Or. 14.10f.Google Scholar, does not give the impression that it dealt with the birth at any length. If there was a hymn to Apollo attributed to Olen, no information about it has come down to us.

¹⁶ Sourvinou-Inwood, C., ‘The Myth of the Four Temples at Delphi’, CQ 29 (1979), 231ffCrossRefGoogle Scholar. She sees the four or possibly five temples in Pindar, , Pa. VIIIGoogle Scholar (the first of laurel, the second of bees’ wax and birds’ feathers, the third built by gods, the fourth built by heroes and possibly a fifth one built by men) as a combination of (i) a quasi-historical laurel temple, (ii) an expansion by Pindar of the idea in the Pythian part of the Hymn (lines 274–9) that gods, heroes and men co-operated on a single temple at Delphi and (iii) the invention of the second temple to make the sequence continuous.

¹⁷ ‘The daughter of Coeus’ is elsewhere Leto, , as at Pa. XII.13, h. Hom. Ap. 62Google Scholar, Alcaeus T, line 1, Aristonoos, , Paean 5–6Google Scholar (Powell, J. U., Collectanea Alexandrina [Oxford, 1925Google Scholar; henceforth ‘Powell’], p. 162), Κοιηῖς at Call. Del. 4.150, Κοιογενς at P. fr. 33d.3, Κοιογνεια at A. R. 2.710.

¹⁸ Snell's supplement in lines 43–4: ᾖ Δις οὐκ θλο[ισ' μβναι λχος | Κοου θυγτηρ π[ντονδ' ἒφυγεν must be on the right lines.

¹⁹ Of unknown length, but it must have been short if, as seems to be the case, ἒσσατο in line 40 is not part of it. It is unlikely to have begun with the start of the epode (line 41) since exact correspondence between start of speech and start of stanza is extremely rare: see Führer, R., Formproblem-Untersuchungen zu den Reden in der frühgriechischen Lyrik, Zetemata, 44 (Munich, 1967), 72Google Scholar.

²⁰ Cf. O. 1.52ff., O. 9.35ff., N. 9.33ff.

²¹ Cf.O. 1.52, N. 9.33.

²² Snell supplements: ἢπιστ μ[οι] δδο[ι]κα κσ[εβ λγειν φτις]| δ‥

²³ π μοι λγο | τοτον, στμα, ῥῖψον | πε τ γε λοιδορσαι θεος | χθρ σοφα, κα τ καυχσθαι παρ καιρν | μαναισιν ὑποκρκɛι (Mouth, cast this speech away from me, for insulting the gods is a hateful form of wisdom and inopportune boasting sounds in harmony with madness). On the analysis of this passage see Stinton, T. C. W., ‘Si credere dignum est’, PCPhS 202 (1976), 67ff.Google Scholar, Privitera, G. A., ‘Il criterio della pertinenza: Pind. Ol. 9.35–41’, RFIC 114 (1986), 48ffGoogle Scholar.

²⁴ Cf. Call, . Del. 53ff.Google Scholar, Arist. fr. 488 Rose, , Pliny, , HN 4.12.66Google Scholar. On the similar allusion in fr. 33c.6, see below.

²⁵ In line 50 the meaning of τς is probably ‘until‘, as suggested by Schmidt, V., ‘Zu Pindar’, Glotta 53 (1975), 39ffGoogle Scholar. Snell tries to make it work as a relative pronoun by supplementing ….Λατὼ Κρονδας π' σφαλεῖ πδῳ, which seems highly unlikely.

²⁶ Cf. O. 14.14:θεν κρατστου παῖδες, which refers to Zeus as father of the Charites.

²⁷ It is just possible that τοξοφρον…γνον refers to Apollo only, but the fact that Pindar consistently presents Apollo and Artemis as twins makes it almost certain that they are both meant here.

²⁸ Cf. the fairly deliberate closes of Pa. 1.5ff., Pa. II.102ff., Pa. V.43ff., Pa. VI.178ff., Pa. XIV.35ff.

²⁹ Cf. the scholion to line 33 (preserved in P.Oxy. 841 fr. 19): λχος π τν λοχεαν, from which editors restore λχος in the text.

³⁰ The coronis marks the end of the triad rather than the end of the poem in P.Oxy. 659 (Pindar, fr. 94a and b)Google Scholar, P.Lit.Lond. 46 (Odes of Bacchylides), P.Oxy. 2042 (Pindar, Ol. 2Google Scholar: coronis at lines 52–3, between the third and fourth triads). In P.Oxy. 841 there is a coronis at the end of every strophe of the monostrophic Pindar, Pa. V. See Stephen, Gwendolin M., ‘The Coronis’, Scriptorium 13 (1959), p. 5 n. 15CrossRefGoogle Scholar. Snell–Maehler are so confident that the coronis in P.Oxy. 2442 fr. 14 col. 2 line 20 marks the end of the poem that they print the lines that follow under Pa. VIIc, associating them (for no compelling reason) with P.Oxy. 841 fr. 21.

³¹ Lines 14–16: χαῖρε, μκαιρ' ὧ Λητοῖ, πε τκες γλα τκνα | Ἀπλλων τ' ἄνακτα κα Ἂρτεμιν ἰοχαιραν, | τν μν ν Ὀρτυγῃ, τν δ κραναῇ ν Δλῳ … Lines 14–18 of the Hymn should not be regarded as an interpolation: see Cassola, F., Inni Omerici (Verona, 1975), pp. 486–7Google Scholar, Miller, A., From Delos to Delphi, Mnemosyne Suppl. 93 (1986), pp. 17ffGoogle Scholar. Eichgrün, E., Kallimachos und Apollonios Rhodios (Berlin, 1961), p. 260 n. 38Google Scholar, suggests that in calling Delos, ‘Ortygia’ Pa. VIIbGoogle Scholar reacts against the tradition. Lycophron, (Alexandra 401)Google Scholar apparently describes Delos as ρτυγος πετρυμνης and a scholiast explains that Asteria transformed herself into a quail before a second transformation into an island (Scheer, E., Lycophronis Alexandra [Berlin, 1908], ii. 149, 24ff.)Google Scholar. There is an outside chance that this was mentioned earlier in Pa. VIIb.

³² The crucial point is that two small fragments of the papyrus roll from which Pa. XII comes (P.Oxy. 1792) contribute to Pindar, Pa. VI: fr. 15 (renumbered as fr. 60 by Lobel in Ox. Pap. 26, p. 17) = Pa. VI.128–31 and fr. 16 = Pa. VI.134–6.

³³ If it was a paean we would expect it to have specified the identity of the performers and the place of performance (see n. 3). It is likely enough that the place of performance is Delos. There is a reasonable chance that Να]ξθεν in line 6 indicates the identity of the chorus. Snell–Maehler print a title ΝΑΞΙΟΙΣ ΕΙΕ ΔΗΛΟΝ in brackets.

³⁴ I omit lines 21ff., which are very fragmentary. Snell–Maehler print five smaller fragments of P.Oxy. 1792 as ‘Pa. XII(a)–(e)’, but these as likely as not do not belong to the same poem as the main fragment.

³⁵ Zeus' role is minimal in Callimachus' Hymn to Delos also, where he does n o more than calm Hera's anger (line 259; see Mineur, W. H., Commentary on Callimachus' Hymn to Delos, Mnemosyne Suppl. 83 [1984] on this line and on line 195Google Scholar).

³⁶ For example Il. 11. 182—3: πατρ νδρν τε θεν τε | Ἴδης ν κορυφῇσι καθζετο πιδησσης, Il. 8.51–2: αὐτς δ' ν κορυφῇσι καθζετο κδεῖ γαων, | εἰσορων Τρώων τε πλιν κα νας Ἀχαιν.

³⁷ The supplement π[ρ]ονο|ᾳ requires a brief comment. In the editio princeps (Ox.Pap. 15 [1922], p. 57)Google Scholar Grenfell and Hunt restored the end of 1.11: φυλξ[αι χρ]νον, which was generally accepted until Lobel (Ox.Pap. 22 [1961], pp. 13–14 with pl. XIXGoogle Scholar) added a new fragment of the same papyrus containing the letters αιπ, which made some form of πρνοια inevitable. The dative singular with φυλξ[ᾳι is stylistically attractive, but it remains possible that a different case of πρνοια was in construction with another word or words following, for example πρνοιαν ἴχοντα (cf. , S.Ant. 283Google Scholar) or προνοας ἔξ (cf. Hdt. 1.120, IG i.² 115.11). For another possibility, see n. 41.

³⁸ Cf. προνοιαισι το πεπρωμνου at , A.Ag. 683Google Scholar, τς παλαιχτου προνοας at , S.Tr. 823Google Scholar (referring to divine πρνοια), θεᾳ προνοᾳ at , E.Ph. 636Google Scholar (which Fraenkel, E., (Aeschylus Agamemnon [Oxford, 1950], n, p. 330)Google Scholar takes as modelled on , A.Ag. 683Google Scholar), το θεου προνοη at Hdt. 3.108. The same word is used for reverence for the gods, e.g. at S. OC. 1180. See Martin, J.-P., Providentia Deorum (Collection de l'école française de Rome 61, 1982), 13ffGoogle Scholar.

³⁹ Cf. O. 9.49; λγοντι νν, which introduces the story of the flood and O. 7.54: φαντ δ' νθρώπων παλαια | ῥσιες …, which introduces the story of the early history of Rhodes. On such expressions, see Richardson, , op. cit., 395Google Scholar, Newman, J. K., Augustus and the New Poetry, Collection Latomus 88 (Brussels, 1967), p. 46Google Scholar.

⁴⁰ Bruneau, P., Recherches sur les cultes de Délos à l'époque hellénistique et à l'époque impériale (Bibliothéque des Écoles françaises d'Athenes et de Rome 217 [Paris, 1970]), pp. 222ffGoogle Scholar. presents evidence for the worship of Zeus an d Athene on Mt Cynthus in the so-called Kunthion, established at the start of Delian Independence (314 B.C.). That Athena was worshipped there from early times is proved by the archaic inscription from the Delion on Paros, , IG XII 5, 210(3)Google Scholar: Ἀθηναῃ Κυνθῃ. See Bruneau, p. 232; Verbruggen, H., Le Zeus Cretois (Paris, 1981), pp. 200ff.Google Scholar; Rubensohn, O., Das Delion von Paros (Wiesbaden, 1962), pp. 43ffGoogle Scholar. We have no direct evidence that Zeus was worshipped there in the classical period but it is a likely inference from the fact that Zeus and Athena were joint occupiers of the Kunthion during the period of Independence.

⁴¹ According to the Deliacus of Hyperides (Blass, , Hypehdes [Leipzig, 1917], 124ff.Google Scholar, Suda and Etym. M. s.Πρνοια Ἀθνα). Snell, , H. 90 (1962), p. 5Google Scholar, suggested that at Pa. VIII. 82ff., where Athena seems t o be instrumental in establishing a mechanism of prophecy at Delphi, Pindar was associating a function of Athena Πρνοια with the Delphic Athena Προνοα. There is no evidence for worship of Athena Πρνοια on Delos except a dubious statement in Macrob. Sat. 1.17.55; cf. Bruneau, , op. cit., p. 249Google Scholar. Still, there is a chance that in selecting the word π[ρ]οναοι|ᾳ at Pa. XII. 11 Pindar was influenced by the role of Athena Pronoia in Delian myth. There is even an outside possibility that line 11 should be supplemented in such a way as to incorporate the title of the goddess, e.g.: Π[ρ]οναοι|ᾳ σὺν Ἀθνᾳ.

⁴² As at O. 3.35, N. 9.4.

⁴³ E.g.: scholion on E. Hec. 458; h. in Apoll. 4–5 in Die griechischen Dichterfragmente der römischen Kaiserzeit, gesammelt und herausg. von. E. Heitsch, Abhandl. Akad. de r Wiss. in Gottingen, phil.–hist.Kl., 3.F, 49 (1961), p. 168.

⁴⁴ See n. 31. For ancient suggestions about the identity of Ortygia, see Allen, T. W., Halliday, W. R., Sykes, E. E., The Homeric Hymns ² (Oxford, 1936)Google Scholar, ad loc., Laager, J., Geburt und Kindheit des Gottes in der griechischen Mythologie (Winterthur, 1957), pp. 68ffGoogle Scholar.

⁴⁵ See conveniently RE 2.21–2.

⁴⁶ Apollod. 1.4.1; Serv. ad Verg. Ecl. 4.10, Aen. 3.73; Roscher, W., Ausführliche Lexicon der griechischen und römischen Mythologie (Leipzig, 1884–) I, 577–8Google Scholar. Artemis was born on the 6th of Thargelion (D. L. 2.44), Apollo on the 7th (id. 3.3).

⁴⁷ Laager, p. 79 with n. 1. The only 5th-century Athenian source he cites is E. IT. 1235ff.: εὓπαις Λατοȗς γνος|ὅν ποτε Δηλισιν|καρποφροις γυλοις|<ἔτικτε> χρυσοκμαν|ν κιθρᾳ σοφν …, but here the reference seems to be exclusively to Apollo, not to Apollo and Artemis.

⁴⁸ …]εφθεγξαντο suggests a refrain: cf. πφθεγμα as a technical term for a refrain, for example παιανικν πφθεγμα at Ath. 15.696f. According to Hephaestion, de poem. 7.3 (71.21 Consbr.) πιφθματικ differ from φᾁμνια in so far as the former contribute something to the sense while the latter are superfluous to it.

⁴⁹ At N. 7.1, O. 6.41ff; also Isyllus', Paean, lines 52ff, Powell, p. 134)Google Scholar. The fact that Eileithyia had a place in Delian cult may also have been a reason for retaining her (Olen's hymn to her: Paus. 8.21.3,9.27.2; cf. Call. Hymn to Delos line 257; the second group of Hyperborean maidens brought tribute for Eileithyia: Hdt. 4.33.5; a statue of Eileithyia from Delos at Athens: Paus. 1.18.5., Pingiatoglou, S., Eileithyia (Würzburg, 1981), p. 33Google Scholar; there was an Eileithyiaion in the sanctuary of Apollo at Delos, at least during the period of Independence: Pingiatoglou, pp. 33–6).

⁵⁰ Λητὼ δ' ννμρ τε κα ννα νκτας λπτοις, | ὠδνεσσι ππαρτο … She also suffers in Callimachus' Hymn to Delos (lines 63, 153, 210).

⁵¹ These fragments are assigned to the Hymn to Zeus on metrical grounds. This current orthodoxy on the Hymn to Zeus appears to be unknown to Stéfos, A., Apollon dans Pindare (Athens, 1975), pp. 19ffGoogle Scholar.

⁵² The stanza of which fr. 33c is the beginning went on for at least 6 lines more, and two lines are lost from the start of the stanza that fr. 33d belonged to.

⁵³ Meyer, H., Hymnische Stilelemente in der frühgriechischen Dichtung (Diss. Würzburg, 1933), 56ff.Google Scholar; Radt, S., Pindars Zweiter und Sechster Paian (Amsterdam, 1958), p. 103Google Scholar.

⁵⁴ See fr. 31 (p. 11 in Snell–Maehler); Snell, , The Discovery of Mind, tr. Rosenmeyer, T. G. (Oxford, 1953), p. 80Google Scholar( = Die Entdeckung des Geistes ⁴ [Göttingen, 1975], p. 88)Google Scholar.

⁵⁵ See fr. 30; Snell, op. cit., pp. 75ff. (= pp. 85ff.).

^* I would like to thank Maria Cannata Fera, Nicholas Richardson and Christiane Sourvinou-Inwood for their comments on earlier drafts of this paper