^{page 461 note 1} Aitia i, fr. i (Pfeiffer). The militant tone of the passage indicates that an orthodoxy is being challenged, and this is confirmed by the evidence from papyri of a flourishing epic tradition throughout the Hellenistic period. Apollonius’ Argonautica, with its reduced scale and careful diminution of much of the essentially heroic tone of the saga, especially the stature of Jason himself, certainly seems to have been influenced by Callimachean doctrine; but it is difficult to assess its typicality for the period as a whole.

^{page 461 note 2} It is still impossible to determine how continuous the different parts of Aitia in fact were; but there can be little doubt that, taken as a whole, it was not έν ἄєισµα διηνєκές but a loosely structured collection in the tradition of Hesiod's Eoiai and (later) Ovid's Fasti.

^{page 461 note 3} Cf. the of Boio (if this is the correct spelling of nis name), which was translated by Ovid's friend Aemilius Macer (Tr. 4. 10. 43–4), Nicander's five books of , the prose collection of by Eratosthenes and Parthenius’ (probably, like the extant , a mixture of prose and verse), Calvus's Io, and Cinna's Smyrna. Our knowledge of these depends on testimonia and fragments. Extant examples include Moschus’ Europa and Silenus’ song in Virgil's sixth Eclogue.

^{page 462 note 1} Cf. Slater, D. A., Ovid in the Metamorphoses (Cambridge, 1912)Google Scholar, and the essays by Alfonsi, L. and Hubaux, R. Crahay-J. in Ovidiana (ed. Herescu, N. I., Paris, 1958).Google Scholar

^{page 462 note 2} The definition rudis indigestaque moles is justified neither by the etymology of χάος nor by its use in Hesiod or other early Greek writers. It is specifically Stoic and can be traced back to Zeno; see Fränkel, H., Ovid, A Poet between Two Worlds (Berkeley, 1945), 208Google Scholar, and for the Stoic rejection of Void see Coleman, , Mnemos. 13 (1960), 34–8.CrossRefGoogle Scholar

^{page 462 note 3} Great stress is placed throughout the sermon on this doctrinaire vegerarianism, which ‘seems to have been a point of special ridicule in Roman literature’. See Segal, Charles, ‘Myth and Philosophy in the Meta-morphoses’, AJP xc (1969), 257 ff.Google Scholar, esp. 278–89, for a convincing refutation of the lofty philosophical interpretation of Pythagoras’ sermon.

^{page 462 note 4} This is not the only place in the sermon where Ovid could have made the link, if he had really been concerned to do so. For he often gives the changes in his stories a symbolic character; formas in noua corpora mutatas—the essential form abides, the material embodiment is altered (the language here is quite Aristotelian). Clytie pines in jealousy at Phoebus’ love for Leucothoe and becomes the heliotrope (4. 169–270), the daughters of Minyas who stay at home in the service of Pallas during the festival of Bacchus are changed into bats (4. 1–42, 389–415), and so forth. This could easily have been related to the version of metempsychosis that assigned destinations according to men's behaviour in this life. But quoslibet occupat artus spiritus (15. 166) from the mouth of Pythagoras deliberately rejects any such connection, and only the superficial resemblance to the theme of the poem, animam sic semper eandem / esse sed in uarias doceo migrare figuras (15. 171–2), is in fact taken up.

^{page 463 note 1} As St. Denis, E. de noted, ‘Le génie d'Ovide d'après le livre XV des Métamorphoses’, RÉL xviii [1940]Google Scholar, III. Ovid's own rationalistic attitude to metamorphosis and the figments of the mythological tradition generally is clearly revealed in Tr. 2. 64,4. 7. 11–20. In treating these fabulae in his earlier work the poet is inviting us to share his own suspension of disbelief and his delight in the fantasies of the creative imagination.

^{page 463 note 2} As Fränkel noted, op. cit., p. 75. However, even here the linear sequences are interrupted by flash-backs.

^{page 463 note 3} Many of the problems of chronological reconciliation belong to the tradition itself, for instance how to make Theseus and Hercules contemporary and yet preserve their associations with so many different stories. Ovid's attempts to cope with these difficulties, especially in Books 5 and 6, are well discussed by Grimal, P., ‘La chronologie légendaire dans les Mètamorphoses d'Ovide’, in Herescu's Ovidiana, 245–57.Google Scholar

^{page 464 note 1} A characteristic of many of the stories in the collection, as Wilkinson, L. P. observed, Ovid Recalled (Cambridge, 1956), 147.Google Scholar

^{page 464 note 2} Geographical groups were apparently employed by Nicander in his ‘Ετєροιούµєνα, though like the other pre-Ovidian poems on the subjects (see p. 461 n. 3) it seems to have been a catalogue work in the manner of Hesiod.

^{page 464 note 3} Grouping by thematic association has been considered in a general way, e.g. by Crump, M. M., The Epyllion from Theocritus to Ovid (Oxford, 1931), 195–242Google Scholar, Wilkinson, op. cit., 148, and Otis, Brooks, Ovid as an Epic Poet (Cambridge, 1966), 81–90Google Scholar, to establish large structural divisions in the work. Thus Wilkinson: (a) 1. 1–451 Prologue and Cosmogony, (b) 1. 452–6. 420 The Gods, (c) 6. 421–11. 193 Heroes and Heroines, (d) 11. 194–15. 870 Historical Personages. Otis: (a) 1–2 The Divine Comedy, (b) 3. 1–6. 400 The Avenging Gods, (c) 6. 401–11. 795 The Pathos of Love, (d) 12–15 Troy and Rome. The objections to this mode of analysis are twofold. First that the sections themselves do not form selfevident and compelling structural units; the disagreements as to where the divisions fall is in itself significant. Secondly, many of the individual stories in any given one of these sections could be more appropriately placed under the heading of one of the others.

^{page 464 note 4} An instance occurs in Callimachus’ fifth Hymn with Artemis and Actaeon inset within Pallas and Teiresias. Crump, op. cit., recognized the importance of the epyllion for the composition of Metamorphoses, but her discussion is marred by a failure to appreciate that it is this particular form of epyllion, as further exemplified in Moschus’ Europa or Catullus 64, and not the single-tale form represented by Theocritus, Id. 13, that provided the effective structural unit in a perpetuum carmen of this kind.

^{page 465 note 1} Wilkinson's admirable discussion of Ovid's gods (op. cit. 190–203) ought to have effectively put paid to the view, held by Heinze and others, that the poet was ‘clearly striving to sacrifice as little as possible of the divine sublimity’.

^{page 465 note 2} Strangely there is no mention of the traditional gadfly—tabanus or asilus. Io is pursued by an Erinys and tortured by stimulos in pectore caecos. H. Fränkel, op. cit., p. 211, plausibly suggests an echo here of Heroides 14. 85–109, where Io's sufferings are similarly psychological. Her wanderings enable Ovid to account for the Egyptian Isis; he seldom misses the chance to gather in incidental metamorphoses!

^{page 465 note 3} Tiresias in Book 3 is another instance.

^{page 465 note 4} The instances of verbal repetition from the preceding tale of Apollo and Daphne noted by Frãnkel, op. cit., p. 85, are surely too trivial and infrequent to create a sense of wearisome monotony capable of inducing slumber in any reader attentive enough to notice them.

^{page 465 note 5} It is instructive in this respect to compare Io here and Europa in 2. 833 ff. with the two stories as told by Moschus, where the association of the two in a single narrative poem brings out quite different aspects of both. A similar comparison can be made between Athena and Teiresias, with the inset Artemis and Actaeon, in Callimachus’ fifth Hymn and the more dissociated versions of the two tales here in 3. 138–252, 316–38, though the comparison is made less effective by the fact that Ovid adopts a different variant of the former.

^{page 465 note 6} The next instance is Apollo and Coronis (2. 531–632) enclosing Neptune and Corone (547–95).

^{page 466 note 1} The connection between the bay and poetry is admittedly alluded to very briefly (559), while its associations with Roman triumphs and Augustus are elaborated. At first sight this seems an intrusive piece of flattery to the emperor, a ‘bow to Augustan convention’. Yet can Ovid really have intended to honour Augustus—as distinct from seeming to honour him!—by thus introducing him into the context of a fanciful tale of metamorphosis and a god's frustrated amours, especially when that god is none other than the chosen imperial patron?

^{page 466 note 2} Similarly tenuous narrative links recur throughout the poem, e.g. 6. 412–23, which links the Anatolian group, beginning with Arachne (6. 1–145) and concluding with Pelops (401–11), to Tereus, Procne, and Philomela (433–721). At 15. 622, as L. P. Wilkinson has pointed out to me, there is in fact no narrative link at all. However a thematic link is established between Aesculapius, which begins here, and Hippolytus, the inset (497–546) to Egeria (488–550), by the fact that both are divine migrations to Italy. This would have justified placing it immediately after Egeria; but Ovid clearly thought the juxtaposition of Aesculapius aduena and Caesar, in urbe sua deus (745–6), more important at this stage of the Roman sequence.

^{page 466 note 3} What follows in pp. 466–9 is a more detailed discussion along the lines briefly sketched in CR n.s. xvii (1967), 48.Google Scholar

^{page 466 note 4} The fama … ueteris … rapinae was told at 5. 339 ff. Cross-references of this kind, which are common enough in the poem— the Heliades in 10. 91 and 2. 329–66 is another—also contribute to its perpetuitas.

^{page 466 note 5} The vast wealth of the traditional metamorphosis literature is frequently indicated by incidental references of this kind that are not elaborated in the poem. Elsewhere in Book 10 we find allusions to Attis (104–5), Icarus andErigone (450–1), and Proserpina's love for Menthe (728–30); and no account is given of how the Cerastae acquired their horns (223–3).

^{page 467 note 1} 207–8 allude ingeniously to the alternative fable of the hyacinthus, which Ovid is to treat specifically in Book 13.

^{page 468 note 1} For the symbolic character of petrification cf. the similar fates of the stubborn and jealous Aglaurus (2. 819–32) and of Niobe, so paralysed by grief that the only movement left was that of her tears (6. 303–12).

^{page 468 note 2} See further the sensitive discussion by Frankel, op. cit., 95.

^{page 468 note 3} As Otis points out (op. cit., 389 f.), Ovid has suppressed the brutish agalmatophilia of the original version of Pygmalion in order to heighten the contrast with Propoetides.

^{page 468 note 4} Like the soliloquies of Medea (7. 11–71), Scylla (8. 44–80), and Byblis (9. 487–516) this one reveals the same morbid interest and skill in exploring the psychopathology of women in love that Ovid had shown in Heroides.

^{page 468 note 5} Although Myrrha's conflict is not externalized, like Phaedra's in the two great scenes that Euripides devoted to her and the nurse, the decision here is clear-cut, and there is none of the pathos of that terrible final question øάρµακον; with which the dramatist reveals Phaedra's bewildered and desperate capitulation.

^{page 469 note 1} Leaving aside the possible incestuous implications of the line, it is an ironie mode of revenge on the goddess who had employed her son in the exquisitely cruel assault on Dido in Aeneid i, a passage that Ovid may have intended his readers to recall here.

^{page 469 note 2} A link here with the Cadmus group in Book 3, since Echion was one of the five survivors of the serpent's teeth sown by Cadmus.

^{page 470 note 1} The model of animal behaviour invoked by both Iphis (9. 731–4) and Myrrha (10. 322–8) brings out the point that homosexuality is more unnatural than incest. Myrrha's pathos is in the actual consummation of her desire, as Otis notes (op. cit, 226).

^{page 470 note 2} The thematic sequence of unnatural passion exhibited in Tereus, Scylla, Byblis, Myrrha, and Ceyx, analysed perceptively by Otis, op. cit., 205 ff., plays an important part in joining together the various groups to which each of these stories organically belongs.

^{page 471 note 1} For instance the Family of Cadmus (3. 1–4. 603) begins with the hero's arrival in Thebes and his slaughter of Mars’ sacred serpent and ends with the fulfilment of the resultant prophecy in the transformation of Cadmus and his wife Harmonia into serpents. The long inset within this story begins with a linear sequence Actaeon, Semele, Teiresias, Narcissus, and Echo (138 ff.). There follows a simple inset group, with Pentheus (511–733) enclosing Acoetes and the Tyrrhenians (582–691), and then a more elaborate one, with Minyades (4. 1–415) enclosing Pyramus and Thisbe, Clytie and Leucothoe, Salmacis and Hermaphroditus (55–388). The whole major inset ends with Ino (416–562). The closely knit group has a narrative link with the preceding Europa (2. 833–75), since Cadmus is the heroine's brother, and a thematic one with the group that follows, since its hero Perseus was conceived in miraculous circumstances closely parallel to those of Bacchus. The chronological difficulties obviated by the inset technique here are well discussed by Otis, op. cit., 130.

A briefer but no less complex structure is exhibited by the contest between the Pierides and the Muses (5. 294–678). The representative of the Pierides is given a single tale Typhoeus (319–31). Calliope begins her contribution with Ceres and Proserpina (341–571). This is interspersed with digressions, first on the punishment of Typhoeus (346–58) which completes the earlier tale, then Cyane, The Lizard (409–61), and finally Ascalaphus, Acheloides (539–63). There follow Arethusa (572–641), already begun in Ceres and Proserpina (487–508), and Triptolemus (642–61). The contest itself is part of an even longer group concerned with Pallas and the Muses, beginning with the goddess's visit to Helicon (5. 250 ff.), ending with her victorious contest with Arachne (6. 1–145), and ingeniously linked, as always, to what precedes and follows it.

^{page 472 note 1} Op. cit., 101; cf. Wilkinson, op. cit., 221: ‘It looks very much as if his delight in invention was flagging [sc. by Book 15], as well it might be by now, so that he made haste to his designed conclusion. But this decline really began with Book 12, in which we are aware of a conscious effort to raise the poem into a “higher” stram.’ As Segal observes, loc. cit. 257, the ‘historical’ section actually begins at 11. 194 with Laomedon's founding of Troy.

^{page 472 note 2} Thus Aeneas’ encounter with Anius in Delos (13. 631 ff.) leads to an account of the miraculous powers enjoyed by the King's daughters and their transformation into doves (650–74), the brief tale of Orion's daughters and the Coroni on the presentation cup (685–701). The Sicilian part of Aeneas’ journey (13. 705 ff.) is the occasion for the Scylla—Galatea sequence (13. 730— 74), enclosing Polyphemus and Acis (750–897) and Glaucus (917–65) and linked by the role of Circe in the transformation of Scylla herself to the Circe group later in Book 14. For the remoteness of the latter, with its emphasis on lust and violence, from the Roman-Augustan themes of Virgil's Aeneid see Segal, loc. cit., 270 ff.

^{page 472 note 3} That Ovid should be unwilling to challenge comparison with Virgil is not only out of character in a poet who was nimium amator ingenii sui but flatly contradicted by the evidence of Dido (Her. 7), Cacus (Fast. 1. 545–78), and Orpheus and Achaemenides in the present poem. He was perfectly capable of treating familiar subjects in his own way, and this is just what he has done with Aeneas’ journey.

^{page 473 note 1} The transition from science to mythology is more abrupt in Book i, where the Cosmogony is followed by Saturnia Regna, Concilium Deorum, Lycaon (the first fabulous transformation), Diluvium, Deucalion and Pyrrha, More abrupt but less significant; for the distinction between res prudentes and fabulae there reflects a stage in human thought that had de-mythologized the origins of the physical world but not the prehistory of the race. Even Lucretius’ anti-mythological evolutionism still allowed its picture of primitive human brutishness to be coloured by lingering hints of the Golden Age tradition (cf. De rer. nat. 5. 1390–1435).

^{page 473 note 2} That some of this at least would have been taken seriously in Antiquity is clear from Var. R. 3. 4. 1 and Virg. G. 4. 543 ff., Sext. Emp. P. 1. 11 and a number of passages in Pliny, e.g. Nat. 8. 105, 122, 137.

^{page 473 note 3} As was emphasized by Anderson, W. S., ‘Multiple Change in the Metamorphoses’, TAPA xciv (1963), 27Google Scholar, the absence of any reference to the concept of Roma aeterna, in contrast to Fast. 3. 72, is so remarkable as to be surely of calculated significance.

^{page 473 note 4} A point well made by Wilkinson, op. cit., 215, who nevertheless sees the doctrine of transmigration as ‘an imposing climax to the stories of metamorphosis’. As Segai observes, loc. cit. 280, the emphasis of the Sermon is less on metempsychosis than on vegetarianism—which was precisely the doctrine of the Pythagoreans that attracted most ridicule and satire in the Latin literary tradition.

^{page 473 note 5} Cf. Otis, op. cit., 302.

^{page 473 note 6} Neither Apollo nor Cybele, two of the most obvious migrating deities, are included here. But they are treated fully in Fasti. Indeed, the somewhat slender representation of specifically Latin or Italian fables is certainly to be explained by their presence —actual or intended at this time—in the poet's elegiac calendar. So Wilkinson, ‘The World of the Metamorphoses’, in Herescu, 's Ovidiana, 231–44, especially 240.Google Scholar

^{page 473 note 7} As in Fast. 3. 153; cf. Cic. Rep. 2. 28–30, Liv. 1. 18. 2, where the chronological error is castigated.

^{page 474 note 1} Here Segal, loc. cit. 288–9 strives unconvincingly for a serious and profound interpretation of the synthesis–a meeting of ‘Greek theory and Roman practicality, philosophy and actual government’.

^{page 474 note 2} Although Caeneus is already associated with the battle in Il. I. 264 ff., this metamorphosis appears first in Hellenistic literature, Ap. R. Arg. 1. 57–64. In Apollod. Epit. 1. <21>, 22 Lapiths and Centaurs is catalogued immediately before Caeneus. Although this may have inspired Ovid, it is the tonal contrast of the two tales and the effect produced by the inset-technique of throwing the emphasis on Caeneus that are important here.

^{page 474 note 3} Cf. the famous sculpture from the West end of the temple of Zeus at Olympia.

^{page 474 note 4} Cf. Horace's Centaurea … cum Lapithis rixa super mero debellata (Carm. 1. 18. 8–9) as a warning of the dangers of excessive drinking, which is of course the context in which Antinous introduces his reference to the tale in Od. 19. 293 ff.: . But Homer does not give it the extended heroic treatment, as Ovid does.

^{page 474 note 5} The tedious exaggeration and frigid epic manner of which Otis complains (op. cit., 283, 314) are thus revealing of the poet's intentions.

^{page 475 note 1} As Wilkinson (op. cit., p. 230) well observes; cf. also Otis, op. cit., p. 285.

^{page 475 note 2} It is part of the anti-epic character of this whole epic section that Ovid takes care to include Achilles’ craving for his rightful honour even after death, when his ghost rises up indignantly to demand the sacrifice of Polyxena (13. 441–8). The incident was traditional, but the contrast between the petulant ghost and the noble dignity of Polyxena (456–73) is very striking. For the similar function of the Polyxena episode in belittling the epic tradition cf. Cat. 64. 363–70, on which see now Bramble, J. C., ‘Structure and Ambiguity in Catullus LXIV, PCPS cxcvi (1970), esp. 25–6.Google Scholar

^{page 475 note 3} Otis correctly noted that ‘the net result of the three pieces [including Hecuba] thus arranged was to destroy the significance of the Trojan theme’ (op. cit., 282); but found this baffling, since he started from the assumption that the apparently intrusive ‘epic panels’ in this poem of fanciful caprice were meant to be taken, out of their context, as a serious contribution to the Augustan Myth.

^{page 476 note 1} See further Galinsky, G. K.'s discussion of Cipus in TAPA xcviii (1967), 181–91.Google Scholar

^{page 476 note 2} e.g. Virg. G. 4. 560–2, Aen. 6. 791–7; Hor. Carm. 1. 2, 3. 3, 4. 5.

^{page 476 note 3} The same kind of overtone is surely to be apprehended in Venus’ patronage of the gens Iulia. The epithet Cythereius applied to Aeneas (13. 625, 14. 584) belongs stylistically to the epic conventions of the patronymic, Laomedontius and Anchisiades in the Aeneid for instance. But it calls our attention not to the hero's epic heritage as a prince of the Trojan royal house but to the fact that he was his mother's son, Dido's lover. Nor is there after all so much to boast about in genealogical descent from a goddess who, in pleading for the life of Julius Caesar, couples her rescue of Aeneas from Diomede with that of the unheroic Paris from the Greek king that he had cuckolded (15. 805–6), and who throughout the poem has appeared as the inspiration of suffering and catastrophe to gods and men alike.

^{page 476 note 4} Aug. 68–9, 71.

^{page 476 note 5} Cf. the echo in parte tarmen meliore mei (875) of parte sui meliore uiget (g. 269), pars optima restitit illi (14. 604), from the apotheoses of Hercules and Aeneas.

^{page 476 note 6} Cf. Am. 1. 15. 33: cedant carminibus reges regumque triumphi and the passages from the Tristia cited by Segal, loc. cit. 291.

^{page 476 note 7} Cf. the echo of this same phrase in Virgil's announcement of his intention to write epic in G. 3. 8–9: temptanda uia est, qua me quoque possim / tollere humo uictorque uirum uolitare per ora.

^{page 477 note 1} Here Segal seems to hold back from the logical conclusion to his own preceding discussion of the poem (loc. cit. 292): ‘It may well be that Ovid feit it necessary or expedient to adopt a façade of Augustanism at the end of his work’, a façade that is no more than ‘a polite nod to the official attitudes over the resistant amorality of the tales which after all constitute the bulk of the poem. … The Augustan seriousness is offset at least in part by the continuing un-Augustan vividness of Circean passion.’ Rather it is drastically undercut both by the absurd exaggeration of its style and by its very introduction into a context of enchanting caprice.

^{page 477 note 2} As Rand, E. K. acutely observed a long time ago, Ovid and His Influence (Boston, 1925), 92.Google Scholar