^{page 164 note 1} All references are to Fasti, in the Teubner edition of Alton, Wormell and Courtney (Leipzig, 1978), unless the context obviously suggests otherwise. I cite by editor's name only the editions of F. Bomer (2 vols, Heidelberg, 1957–8), J. G. Frazer (5 vols, London, 1929), C. Landi (rev. L. Castiglioni, Turin, 1960), H. Peter (Leipzig, 1907). The following books and articles are referred to by the author's name only:

Beard, M., ‘A Complex of Times: No More Sheep on Romulus' Birthday’, PCPS 33 (1987), 1–15Google Scholar.

Brink, C. O., Horace on Poetry ii, The ‘Ars Poetica’ (Cambridge, 1971)Google Scholar.

Horace on Poetry iii. Epistles Book II: the Letters to Augustus and Florus (Cambridge, 1982)Google Scholar.

Fantham, E., ‘Sexual Comedy in Ovid's Fasti: Sources and Motivation’, HSCP 87 (1983), 185–216Google Scholar.

Frécaut, J.-M., L'esprit el I'humour chez Ovide (Grenoble, 1972)Google Scholar.

Hinds, S., The Metamorphosis of Persephone: Ovid's and the Self-conscious Muse (Cambridge, 1987)Google Scholar.

Mckeown, J. C., ‘Fabula proposito nulla tegenda meo: Ovid's Fasti and Augustan politics’ in Woodman, A. J. and West, D. (edd.),Poetry and Politics in the Age of Augustus (Cambridge, 1984), pp. 169–187Google Scholar.

Peeters, F., Les Fastes d'Ovide: histoire du texte (Brussels, 1939)Google Scholar.

Porte, D., Vetiologie religieuse dans les Fastes d'Ovide (Paris, 1985)Google Scholar

Syme, R., History in Ovid (Oxford, 1978)Google Scholar.

Wallace-Hadrill, A., ‘Time for Augustus: Ovid, Augustus and the Fasti’ in M. Whitby. Hardie, P. and Whitby, M. (edd.). Homo Viator: Classical Essavs for John Bramble (Bristol, 1987), pp. 221–30Google Scholar.

Wilkinson, L. P., Ovid Recalled (Cambridge, 1955)Google Scholar.

Wimmel, W., Kallimachos in Rom (Wiesbaden, 1960)Google Scholar.

I treat Metamorphoses throughout as in almost every respect an earlier poem than Fasti.

² ‘The Fasti was modelled largely on Callimachus’ Aetia' is the confident assertion of McKeown (p. 178). This is in one sense clearly true, but it will emerge that I see Aetia less as a model here and more as a kind of mine whose resources of narrative technique could be exploited by Ovid's own ingenium. For the more obvious aetiological background see Peeters, pp. 54ff.

³ The relation of these to traditional hymns is stressed by Nisbet, and Hor, Hubbard on. Carm 1.10 (p. 127)Google Scholar.

⁴ Such instructions are ‘standard in cult-hymns’, Hopkinson, N. on Call. h. Dem. p. 78Google Scholar with further examples.

⁵ On Verrius Flaccus and the Fasti see now Wallace-Hadrill, , pp. 225ffGoogle Scholar., arguing for a less rigid distinction between poetry and politics than McKeown's. The recent material he assembles supports the positions of Winther, H., De Fastis Verri Flacci ab Ovidio adhibitis (Berlin, 1885)Google Scholar and Peeters, pp. 23 and 54 against the scepticism of Bömer i.22–3. On the Varronian background see McKeown, pp. 170f. (persuasive here), Bömer i.23–4 and Peeters, pp. 49–51 who uses Hülsen, Ch., Varronianae doctrinae quaenam in Ovidii Fastis vestigia exstent (Berlin, 1880)Google Scholar, which I have not seen. On the influence of Propertius' aetiological poems see Bömer i.25–6 and the excellent discussion in Hubbard, M., Propertius (London, 1974), pp. 121–34Google Scholar.

⁶ Cf. the objections of Porte, D. ‘Les Fastes d'Ovide et le sourcil latin’, Latomus 37 (1978), 851–73Google Scholar, with literature on earlier views. This ponderous approach, now amplified at great length (see ‘Porte’ in n. 1), seems to miss the point that aetiology is a functional device for giving narrative a new dimension of sophistication, and one which Ovid uses wittily and inventively.

⁷ Cf. Fantham's attempt (pp. 194ff.) to relate the comedy to Priapean humour. Fantham ignores Porte's article.

⁸ Cf. Holleman, A. W. J., ‘Ovid and the Lupercalia’, Historia 22 (1973), 260–8Google Scholar.

⁹ Cf. McKeown, pp. 177ff.

¹⁰ Cf. Beard, pp. 7f. Wallace-Hadrill (p. 228) largely goes along with Beard, but sees the problems arising from Ovid's undeniably comic tone in treating some ‘moral’ issues.

¹¹ Most sympathetically presented by Johnson, W. R., ‘The Desolation of the Fasti’ CJ 74 (1979–1980), 7–18Google Scholar; cf. Fantham, pp. 210–15 (‘the material was drying up’, p. 215).

¹² As is illustrated by, for example, U. Blank–Sangmeister (see n. 42 below) and S. D'Elia's appreciation of the poem's ironies (see n. 94 below).

¹³ Wallace–Hadrill, pp. 224ff. His advocacy of the Fasti Praenestini needs to be qualified by the point made by Porte, whom he entirely ignores: ' Les Pastes de Préneste ne sont pour lui qu'un point de départ, un rpéertoire de sujets qu'on peut modeler ensuite ⋯ son gré' (p. 42. Cf. her whole discussion of ‘Le choix d'Ovide’, pp. 39–124, and Bömer i. 22–3).

¹⁴ Wilkinson (pp. 266–9) sets out some examples of what he claims to be the loss to narrative integrity of Ovid's following the calendar arrangement, but he altogether misses the point I touch on in this paragraph and hope to explain more fully elsewhere, that a satisfactorily integrated narrative was sacrificed for ingenious and subtly ironic effects which could be obtained by exploiting the calendar order. This point is well put by M. Beard (p. 8), developing her claims for ‘paradigmatic meaning”: ‘It was precisely the Roman calendar's reliance on building up associations and images on a paradigmatic model outside any determining narrative that gave the individual festival a fluid meaning in relation to the others in the sequence.’ I shall be exploiting the idea of ‘paradigmatic’ allusion later.

¹⁵ The background to the ‘Homeric’ lines 119f. is fully presented by Austin on Aen. 6.625ff., and to the ‘Pindaric’ lines 123–4 by Kiessling-Heinze on Hor. Carm. 4.2.33 and Bömer, on Met. 10. 148–50Google Scholar.

¹⁶ OLD s.v. nomen 12.

¹⁷ ‘Allerdings ist die Gleichsetzung Ganymedes-Aquarius nicht allgemein anerkannt' concludes Bömer on 2.145, adjusting the rather misleading impression conveyed by Frazer's view (ii.315) that the equation was a ‘popular fancy’ and by Hyg. Astr. 2.29. A full range of possibilities is covered in Röscher's, Lexicon 6. 974–7Google Scholar.

¹⁸ On this point Bömer on 2.145 has nothing to add to Frazer's account (ii.315ff.) and both derive their information (with acknowledgement) from Ideler, J. L., ‘Über den astronomischen Theil der Fasti des Ovid’, Abhandl. der histor.-philol. Klasse der Königliche Akademie der Wissenschaften zu Berlin aus den Jahren 1822 und 1823 (Berlin, 1825), pp. 160fGoogle Scholar. That Columella (11.2.14) follows Ovid in recording the rise of Aquarius at its mid-point is at least as likely as that both use some (unknown) independent source. Columella's ‘mediae partes Aquarii oriuntur’ is simply a prose version of Fast. 2.145.

¹⁹ So also at Priapea 3.5–6 Buecheler, attributed to Ovid by the elder Seneca (Con. 1.2.22). The variants in the Greek version of the Ganymede story are documented by Gerber on Pindar, 01. 1.45 s.v. χρέоς. See als o Frazer on Apollod. Bibl. 3.12.2, Bömer on Met. 10.155ff. and Röscher's, Lexicon cited in n. 17Google Scholar above.

²⁰ On the comparison see Binder, G.Aeneas und Augustus: Interpretationen zum 8 Buch der Aeneis (Meisenheim am Glan, 1971), pp. 150ffGoogle Scholar., Gransden, K. (ed.), Virgil, Aeneid VIII (Cambridge, 1976), p. 16Google Scholar, Syme, R., The Roman Revolution (Oxford, 1939), pp. 306Google Scholar and 313. There is a full bibliography on the comparison in Brink iii. 39–42 (on Hor. Ep. 2.1.5).

²¹ As in Horace (see Brink ii.391), though Ovid's manipulation of the potential ambivalence is naturally more self–consciously mocking than Horace's, e.g. at 6.249ff.

²² E.g. ‘annalibus eruta priscis’ 1.7, 4.11, a phrase in which the associative, scholarly emphasis occurs in eruta ‘unearthe d by investigation', see TLL 5.2.845.33ff. as well as 844.77ff. given by Bömer. The Fasti usages should be included with those listed by Brink (iii.337) to illustrate the parallel Horatian use of eruere in connection with the poet-researcher. For other examples of supposed scholarly work by Ovid cf. 3.844 and the reflections on the origin of the name Agonalia at 1.317–32 which may at least in part derive from the alternatives in the Fasti Praenestini (Wallace-Hadrill, , p. 229)Google Scholar. Bömer (i.22ff. and ii on 1.7) takes this ‘antiquarian-researcher” pose far too seriously, as does Frazer (i.xi.xiii) who has Ovid gathering his material with the earnestness of the author of The Golden Bough ! Wilkinson, (pp. 264–8)Google Scholar is rightly more sceptical about the degree of serious dedication to scholarship involved. Callimachean research is, by contrast, precise about its sources (cf. frr. 75.53–6, 92.2–3 Pf.).

²³ We shall presently see that this example has a more complex side to it (see section V iii and nn. 61 and 62 below). Wilkinson (p. 250) quoting these and other examples fails to see that Ovid uses the invocation-technique not to ‘enliven his poem’ but as a literary device for defining the status both of the causa and of the poet who expounds it. Porte (pp. 30ff.) is equally innocent of the function of the technique.

²⁴ Cf. 145–8, 161–2, 165–6, 171–2, 175–6, 183–4, 189–90, 227–30, 255–8, 277–8.

²⁵ A list of earlier appearances to poets is given by Hor, Nisbet-Hubbard on. Carm. 2, p. 315Google Scholar, all involving Greek or Greco-Roman deities. The instruction given by the (Etrurian?) Vertumnus in Propertius 4.2 is not, of course, an epiphany, while Romulus' intervention at Hor. Sat. 1.10.32ff. is not in answer to any invocation of him in the poet's Graeci versiculi.

²⁶ Bömer on 1.93 notes the parallel with Aetia fr. 1.21f., but cf. also the approaching epiphany of Apollo in the second hymn.

²⁷ So emphasised by Ovid in 90 to make the combination as dramatic as it is unexpected: ‘nam tibi par nullum Graecia numen habet’. What concerns us is that Ovid believed Janus to be exclusively Roman; the wider possibilities now canvassed are given by Bömer on 1.89.

²⁸ Aen. 8.335–41. Virgil already has Carmentis as a prophetess (‘vatis fatidicae’ 340), so naturally Ovid gives her something to say which must command acceptance. Cf. Eden on Aen. 8.336ff. for the etymological link between her name and carmen (‘prophecy’) and Bömer on 1.462.

²⁹ Aen. 8.31–65; see Bomer on 5.635 and Rutledge, E. S., ‘Vergil and Ovid on the Tiber’, CJ 75 (1980), 301–4Google Scholar. The application of ‘Greek’ poetic invocation to the figures of early Roman mythology continues the technique used in the case of Janus and found again in that of Flora.

³⁰ This metaphor is found as early as 1.4 and is used as a sphragis at 2.863–4 and 3.789–90 (see Otto, , Die Sprichworter und sprichwörtlichen Redensarten der Römer [Leipzig, 1890]Google Scholar s.v. linter). Examples listed by Bömer on 1.4, to which could be added Georgics 2.41ff. and 4.116f., bear out the significance of the ship-motif equally for the didacticism of Fasti and Ars amaloria.

³¹ As it is by Beard, p. 2 and Porte, p. 17, following Frazer i.xiii–xiv. Wilkinson (pp. 266–8) and Bömer (i.29–30) more sensibly recall Wissowa's view that Ovid's knowledge of the details of rituals is extremely sketchy and often wrong in elementary ways. Fasti is a literary, not a liturgical, text and it is worth noting here Bulloch's, A. W. comment in his demonstration that Callimachus' fifth hymn was not used for practical liturgical purposes: ‘… the careful insertion of references to ceremonial particulars [has] to do not with realism, but verisimilitude. Indeed the very presence of such details betrays precisely the literary nature of our text.’ (Callimachus: the fifth hymn [Cambridge, 1985], p. 5Google Scholar; my italics).

³² On the version followed here, in which Remus is killed by Celer, see Bömer, i.26–8, 45 and ii. on 2.809. In his detailed analysis of the traditions Ogilvie (A Commentary on Livy 1–5 [Oxford, 1965], p. 54)Google Scholar rightly observes that Virgil suppresses the story of Remus’ murder by Romulus, as used in Ennius, Horace and Livy, to make Augustus' Romulean pretensions more commendable. Cf. Wagenvoort, H., ‘The Crime of Fratricide: the Figure of Romulus-Quirinus in the Political Struggle of the First Century B.C.’, Studies in Roman Literature, Culture and Religion (Leiden, 1956), pp. 161–83Google Scholar. As I explain, Ovid's purpose is in my view quite different.

³³ 4.852 ‘invito frater adempte, vale’ combines Catullus 68.20, 92‘ …misero frater adempte mihi’ with 101.10 ‘…frater, ave atque vale’. Bömer strangely misses this.

³⁴ In Livy (1.7.2–3) Remus acts deliberately (‘ludibrio fratris’ 7.2 –contrast Ovid's, ‘ignorans’ in 841)Google Scholar, and is killed by Romulus who speaks his famous taunt ‘increpitans’.

³⁵ Frécaut's attempt (p. 229) to show that the Ovidian Romulus here is a humorous cross-reference to Livy's Aemilius Paullus fails for want of persuasive evidence. The Romulus-Augustus association is more likely to have been in Ovid's mind.

³⁶ In fr. 86 Pf. explicitly, but a growing weight of evidence supports the view that in the first two books of Aetia the Muses in turn answered Callimachus' questions. It is no w possible to trace the intervention of Erato(SH 238.8 with editors' note ad loc.) as well as of Clio, who provides an accurate genealogy of the Graces when Callimachus had been depending on written sources (Schol. Flor. Pf. i, p. 13.3O ff., cf. fr. 43.56 Pf.) and of Calliope (fr. 7.22 Pf. and Pfeiffer's note on fr. 759). There is however no evidence that Callimachus' Muses disagree among themselves.

³⁷ For the extent of the metaphor in Ovid see Bömer on 4.10. The Callimachean reference is of course to the poetic carriage and journey of Aetia fr. 1.25–8 (Wimmel, , pp. 105–9)Google Scholar. Propertius had already used this motif at 3.11.9–11 and especially 2.10.1–2 where Stahl, H.-P. has recently argued for a greater independence from Callimachus (Propertius: Love and War; Individual and State under Augustus [Berkeley, 1985], pp. 156ff.)Google Scholar.

³⁸ The two latter are also supported by two sisters each (108). Clio and Calliope appear in a similar capacity in Aetia (see n. 36).

³⁹ Polyhymnia combines Roman Maiestas (see Bömer on 5.9: Porte, p. 217) with the gigantomachy; Urania tells a Romulean story (see Bömer on 5.55; Porte, p. 203) and Calliope the Greco-Roman story of Arcadian Evander (see Bömer on 5.80).

⁴⁰ L. 6.33. Calliope's Maia is treated as an alternative at 6.35 and also by Festus (p. 120.8ff. Lindsay), and is preferred by Censorinus, , De die natali 22.12 (p. 56.4 Sallmann) and Servius, in Georg. 1.43. Polyhymnia's Maiestas seems to be a suggestion unique to OvidGoogle Scholar.

⁴¹ That by choosing Elegia in preference to Tragoedia in Amores 3.1 Ovid might be supposed to be giving offence is suggested by his supplication to Tragoedia to avert her anger (67–8), which evidently proves successful (‘mota dedit veniam” 69).

⁴² The comical function of the iudicium which opens Fasti 6 is analysed differently by Blank-Sangmeister, U., ‘Ovid und die Aitiologie des Juni in Fast. VI. 1–100’, Latomus 42 (1983), 332–49Google Scholar, though she too makes the comparison with Paris (pp. 333ff.) and has her own penetrating discussion of the comic tone of the passage (pp. 348f.).

⁴³ Cf. 6.91–2. I take the ‘placidus dux’ here to be Augustus, but it does not affect my argument substantially if it refers to Tiberius, as ‘dux venerande’ must do in the account of the reconstruction of Concordia's temple at 1.637–50 (Bömer on 1.646). If Tiberius is the ‘dux’ here, Frazer must be right (ii.239–41) in taking the reference to be to Tiberius‘ rebuilding in A.D. 10, when the temple was re-dedicated to Concordia Augusta in honour of the reigning emperor (Suetonius, , Tib. 20Google Scholar; cf. Syme, p. 29). All the more reason why Ovid might be expected to pay special attention to what Concordia has to say, but in the end she fails to convince.

⁴⁴ Porte (p. 65) also connects the iudicia which open books five and six, but her view that Ovid's apparent perplexity at 5.108–10 ‘sert ici á díssiimuler un reéel embarras scientifique’ is hard to reconcile with the fact that Ovid has already formed his conclusion at 1.41 and repeats it at 5.427 and 6.88. As elsewhere, Porte here has much too limiting a view of the potential of Ovidian humour.

⁴⁵ That Ovid should play so freely with a variatio on a solemn invocatory topos like the Muses distances him from the earlier tradition, including Callimachus and the epigrammatists. For the Alexandrian treatment of the topos, largely followed by Horace, see Nisbet–Hubbard on Hor. Carm. 1.26.1 s.vv. Musis amicus and cf. the note on 1.24.3. The early significance of the Muses in Roman poetry is explained by Skutsch on Ennius, Ann. fr. 1 (pp. 144–7). The topos remains equally functional in Propertius (3.2.15–16; 3.3.37–8; cf. Camps on 4.6.12) and in Ovid's amatory poems (e.g. Am. 1.1.30; 3.8.23). As I argue above, Ovid's refusal to award the palm to Calliope in Fasti 5 gains in piquancy if we recall Calliope's outright victory over the daughter of Pierus in Met. 5.662ff.

⁴⁶ Cf. Fast. 2.6 with Bömer's note ad he. Luck on Trist. 3.2.5 and Fordyce on Catullus 50.2. Other Ovidian examples occur at Am. 3.1.27, Trist. 2.538 and 5.1.43.

⁴⁷ Even the normally unsmiling Frazer recognises the tone here as ‘playful’ (iii.164).

⁴⁸ Fast. 2 begins with a parallel literary manifesto in which the ‘ludic’ approach of the amatory poems is set aside for something allegedly very different (cf. 8: ‘ecquis ad haec illinc crederet esse viam?’). As with Fast. 4 the recusatio turns out to be deceptive: the declaration of the gulf between the two styles (8) is immediately (9) followed by a fresh application of the militia metaphor, allusively imported from amatory elegy (see Bömer on 2.9) to remind us that Ovid's ludic technique is still very much alive (cf. also 4.7 for a repetition of the device). The prelude to the fourth book contains a wider area (cf. 10) of allusion and is more obviously self-referential. Otherwise the overall effect is very much the same as in the prelude to the second book.

⁴⁹ Ovid's picturesque account of the bull an d heifer (100–4) is a much reduced adaptation of Georg. 3.209ff., in which the bull uses its feritas instead of laying it aside (103).

⁵⁰ The reference to DRN is emphasised by the Lucretian ‘blanda voluptas’ at 99 (see Bömer ad loc.), but can be seen particularly in the following:

Fasti 4 : DRN 1

93–4 : 4–5

96 : 7–8

99 : 18–20

100–4 :15–20

105–6 :14 (Martin³).

In view of these associations it is difficult to believe that the invocatory adjective alma in 4.1 does not echo ‘alma Venus’ at DRN 1.2 as Porte (art. cit. n. 6 above, pp. 867ff.) already suggests, though Bomer ad he. shows that alma is used in Fasti of other deities as well. For further associations between Fasti 4 and DRN cf. the Cybele passage at Fast. 4.179–372 with its equivalent at DRN 2.600–60. Bailey's commentary on the latter passage records the parallels in detail. P. Ferrarino's account of the relation (‘Laus Veneris: Fasti IV. 91–114’ in Herescu, N. I. (ed.), Ovidiana: recherches sur Ovide [Paris, 1958], pp. 301–16) lacks the support of detailed analysisGoogle Scholar.

⁵¹ Cf. Bömer on 5.339 for full references to the paraclausithyron in amatory elegy.

⁵² The parallels with Fast. 4.23–60 are given in the Korn–Ehwald–von Albrecht edition adloc. (ii.402). This point obviously depends on my declared preference for the priority of Met.

⁵³ Am. 3.13.31–6. From the first line of this poem we learn that Falisci was the birthplace of Ovid's wife. The appearance of Falisci in Fast. 4 (73–4) is quickly followed (79–80) by mention of Ovid's own birthplace, Sulmo (cf. Am. 2.16), and its derivation. That the two birthplaces should be so closely linked in a celebration of Venus may seem appropriate enough, though commentators seem to have missed it. It should be noted that the exilic querella for Sulmo, which was inserted at 81–4, is not a mere appendage to the mention of Ovid's birthplace but is especially ironic in view of the fact that both Sulmo and Falisci were themselves founded by exiles who fled in the opposite direction (73, 79), and gains added piquancy from a wider context which includes such prominent exiles as Evander (65), Ulysses (69), Antenor (75) and Aeneas (78).

⁵⁴ I here adapt Beard's idea that the Roman Fasti showed ‘reliance on building up associations and images on a paradigmatic model outside any determining narrative’ (p. 8).

⁵⁵ Livy 3.17.6 ‘Romule pater, tu mentem tuam … da stirpi tuae’ (Bömer's reference). Cf. also Aen. 8.400. For examples of mens as ‘determination’, ‘firmness of mind’ seeOLDs.v. 7, where the definitions hardly meet the requirements of some of the examples given.

⁵⁶ OLD s.v. 9, especially Georg 3.267 ‘mente m Venus ipsa dedit’, where mens refers to the strong sexual impulse of mares.

⁵⁷ A suitable parallel for the quick succession of events here occurs in the rape of Proserpina (Met. 5.395). See Bömer on 3.21.

⁵⁸ At 199–200 the story is put off until the treatment of the Consualia in August or December. Ovid may well have seen special possibilities in treating it there, since my view is that a narrative of the rape in Fast. 3 will add nothing to what we already know of Romulus‘ ‘patria mens’.

⁵⁹ An unusual word for Ovid to have used in such a context, according to Bomer adloc, who refers to TLL 6.454.23ff. But the adverb arises quite naturally from the frequent use offelix in the sense of fecundus, prole auctus (TLL 6.436.56–437.15), and Mars may well be using feliciter to point to Silvia's fecundity as the mother of twins. So, Hypsipyle informs Jason that she has now given birth to twins: ‘nunc etiam peperi …/ felix in numero quoque sum, prolemque gemellam/…dedi’ (heroid. 6.119, 121–2). There may also be an undercurrent here suggesting that Silvia was naively duped by Mars: Stahl (op. cit. n. 37 above, p. 181) reminds us that felix in Propertius can carry the undertone of ‘naive’, ‘simple’, and I note possible examples of this undertone also in Fast. 1.540; Am. 1.8.27, 2.11.30 and Trist. 5.1.30 where it adds a new dimension to the accepted sense.

⁶⁰ See section III above on 5.445–92.

⁶¹ ‘Die Ableitung ist nicht haltbar’ is Bömer's judgement (on 5.479). Porte (pp. 237f.), who also sees that this nonsense is Ovid's own invention, tries to make out a case for his ‘ingeniosite’ in inventing ‘cette derivation subtile’. Neither asks why Ovid should present a derivation which is so transparently false, or invent one when others were ready to hand. There are, of course, Hellenistic precedents for altering names in this way (cf. Dieg. IX, 12–14 to Call. Iamb. fr. 201 Pf.), but I know of none which exploits the topos as wittily as Ovid does here.

⁶² Porphyrio (on Hor. Ep. 2.2.209) has a similar derivation, doubtless copied from Fast. 5. No Varronian or Verrian etymology survives (the Varronian fragment from de vita populi Romani at Nonius, p. 197.Iff. Lindsay offers none).

⁶³ Mercury's shady reputation in Roman literature goes back to Plautus (Fantham, p. 187), and Ovid may well be exploiting a motif from earlier comedy. For the ‘invention’ of the temple-dedication to him recorded in Livy 2.27.5–7 and Val. Max. 9.3.6 see Ogilvie (op. cit. n. 32 above), p. 303.

⁶⁴ Ovid records the furtum, as well as the infamia Mercury gained by punishing the innocent Battus, at Met. 2.680–707. That Battus is there punished for his ‘periura pectora’ (2.705) is ironic in view of Mercury's own performance before his father (h. Horn. 4.368–86). Cf. Frazer on Apollod. Bibl. 3.10§2, Nisbet–Hubbard on Hor. Carm 1.10.7, Brown, N.o., Hermes the Thief (London, 1947)Google Scholar , Röscher, Lexicon i. 2342ff.

⁶⁵ There is more here than the humour which Frécaut (p. 287) sees. Jupiter smiles at the false oaths of lovers (Ars am. 1.633) because he has deceived so many of his lovers himself. Cf. also Fast. 5. 686 with Ars am. 1.634. Mercury's ‘knowing smile’ (Fast. 5.691) imitates that of his father, for which see h. Horn. 4.389.

⁶⁶ 1 cannot trace an earlier instance where the association is exploited in this way. At Aristophanes, Plutus 1155–8 these two attributes of Hermes are juxtaposed as part of a longer list, which is not quite the same thing.

⁶⁷ This was pointed out by Heinze, R., ‘Ovids elegische Erzählung’ (Bericht iiϋber die Verhandlungen der Sächsischen Akademie der Wissenschaften, Leipzig, 1919), p. 96.1Google Scholar. See Bömer on 5.694.

⁶⁸ For a comparison with the fuller version of the deception in Sophocles’ Ichneutae see Radermacher, L. (ed.), Der homerische Hermeshymnus (Vienna, 1931), pp. 183ffGoogle Scholar. and (more concisely) Page, D. L., Literary Papyri: Poetry (Lib, Loeb Class.., London, 1962), p. 28Google Scholar.

⁶⁹ For Castor cf. Nem. 10.112f. with Fast. 5.709–10 (and cf. Cypria fr. VI Allen); for Lynceus cf. Nem. 10.128–31 with Fast. 5.711–12; for Idas cf. Nem. 10.132 with Fast. 5.713.

⁷⁰ This hymn contains at 167–8 an example of the motif of the winds carrying off and rendering void the prayers of men, which Ovid has just used at 686. Theoc. 22.167f. should be added to the parallels collected there by Bomer.

⁷¹ Frr. VI–XI Allen. See Gow, , Theocritus ² (Cambridge, 1952), ii. 383Google Scholar.

⁷² These are summarised by Bömer on 5.694, where he rightly sees that the differences show ‘Ovids Technik in eigentümlichem Licht’.

⁷³ Ovid's narrative here is thus firmly in the Hellenistic tradition. See Call, Bulloch on. h. 5, p. 163Google Scholar, for ‘studied detachment’ in narrative and p. 177 for ‘confining of dramatic action to a minimum’.

⁷⁴ 212: оűтω ⋯u⋯δαРίδαɩς Πоλεμίζεμε⋯ оúκ έν έλαøρѿ.

⁷⁵ On the significance of ‘non exspectato volnere’ (710) see Bömer ii.334.

⁷⁶ One odd detail of the kind is that the location was free of trees (707). This seems to contradict directly Nem. 10.114δρυ⋯ς έν σтέλεχει ήμένтυς, where the scholiast traces the same detail to the Cypria (= fr. XI Allen). For such descriptive detail in narrative elegy see Otis, B., Ovid as an Epic Poet ¹ (Cambridge, 1970), p. 52Google Scholar.

⁷⁷ I disregard such Ovidian features as Castor's death at the hands of Lynceus (709–10), which simply reverses the situation in Theoc. 22.20Iff., or the odd detail recorded in the previous note, which can be seen as an oblique reference to Nem. 10.

⁷⁸ Frazer iv. 120: ‘Ovid is clearly mistaken in giving the name Aphidna to the scene of the combat … for Aphidna was in Attica’, with supporting evidence for the Laconian location for the fight in n. 4; Bömer ii.334: Grúnde fúr die Verlegung [ sc. des Kampfes] sind nicht bakannt.

⁷⁹ This seems the best explanation (Frazer iv. 119–20; Bömer ii.334, with supporting testimony for the ‘Helenaraub’). A Laconian Aphidna, for which the only evidence is Steph. Byz. 149.16 Meineke. is rightly ruled out by Bomer, loc. cit. Hyginus, Astron. 2.22 has a version (one of a pair) different from either known before, in which Castor is killed at Aphidna fighting against the Athenians. This is almost certainly copied by Avienus, Phaen. 372ff. and by the scholiast on Germanicus (p. 68 Breysig). In an excellent note in his edition of Avienus’, Phaen. (Coll. Bude, [Paris, 1981], pp. 196–8)Google Scholar, J. Soubiran sets out the difficulties an d probable affinity of the various stories, though the complexities he points to are quite ignored in the relevant notes in Boeuffle's, A. Le edition of Hyginus, Aslron. (Coll. Bude [Paris, 1983], p. 170)Google Scholar. Soubiran's observation that the first of Hyginus' two stories, which has Castor killed fighting the Athenians at Aphidna (2.22), is really a new version was not given adequate weight by Frazer or Bomer, who tend to identify it with the earlier versions. The story of Castor's death at the hands of Lynceus (or Idas), which is the second of Hyginus” alternatives an d which he places in Sparta. is altogether different from the other account by which Castor is or is not killed at Aphidna fighting the Athenians to rescue Helen. Hyginus clearly distinguishes the two stories and locates the Lynceus/Idas version in the Peloponnese. Ovid's version is found nowhere outside Fast. 5 and is explicable only as a ‘confused’ conflation of the Lynceus/Idas version with the scene of Helen's rescue (whether or not Castor was killed at Aphidna). It seems to me impossible to credit Ovid with so crude a conflation, with its nonsensical result, when he has obviously taken such pains to pare down his sources an d fashion his narrative to suit his poetic intention. Would Ovid make such a basic mistake when even Hyginus can get the distinction right? As I explain, 1 believe Ovid's ‘error’ to be Mercury's ‘error’, and to be an attempt by the god at ‘narrative deception’ of the reader. For a similar ‘confusion’ at Fast. 4.500 see Hinds, S., LCM 9 (1984). 79Google Scholar.

⁸⁰ The question of whether ‘Hyginus’ is the Palatine librarian C. Iulius Hyginus, Ovid's close friend (Suet, . Gram. 20)Google Scholar, has an obvious bearing on the relation between the Fasti and the Astronomica, but is still unresolved. Syme rejects the association (p. 217 n. 3). accepting the earlier judgements now summarised in Le Boeuffle's edition of Aslron. p. xxxiii. Le Boeuffle himself is much more sympathetic to identifying the two (pp. xxxi–xxxviii). I continue to treat Fasti and Aslron. as independently constructed sources for their common material.

⁸¹ On the use of recusatio see Wimmel, pp. 162ff. and especially 187ff. in connection with Hor. Carm. 1.6, an important parallel example.

⁸² Cf. Frazer on 6.481 an d the references given by Bömer on 6.475.

⁸³ Catullus 64.132–201 is an obvious source for much of the venom which Ariadne is here of course pouring out for the second time. Like Virgil (Aen. 4.305–330), Ovid avoids direct competition with Catullus by varying the situation in which the Catullan speech is exploited. Even in Heroides 10 there is no extended speech for Ariadne on the shore of Naxos.

⁸⁴ This point is made by Hollis, A. S. in his introductory note to the section (Ars am. 1 [Oxford, 1977], pp. 121–2)Google Scholar.

⁸⁵ Hinds, p. 40, following Merkel's, observation in the preface of his edition of Fasti (Berlin, 1841), p. cclvi. Porte (p. 165)Google Scholar also spots the link but makes little of it.

⁸⁶ Hinds sees Fast, and Met. as ‘contemporaneous ’ poems (p. 40; cf. 10–11: ‘They are in some sense what most Ovidian scholars have held them to be, simultaneous compositions’), and he must therefore take 4.418 as a ‘cross-reference’ (p. 40) to Met. and not as a backward glance. But recognosces would be an odd verb for a cross-reference. Ovid's three uses of the verb (cf. also Met. 11.61–2: ‘quae loca viderat ante / cuncta recognoscit’ and Fast. 1.7: ‘sacra recognosces annalibus eruta/wucis’) all refer to the recollection of what was known in the past (OLD s.v. 3). Nor is Hinds strictly consistent: on p. 127 there is a clear implication that the Fasti version is taken as having been composed before that in Met. 5.

⁸⁷ This verdict depends, of course, on the priority of the Met. 5 version.

⁸⁸ Cf. especially pp. 125ff. ‘It is hard not to see some significance in the fact that Met. 5.341 ff… is represented as being composed and sung by the Muse who, later to be confirmed more explicitly as the Muse of epic, is already something of a special patron of elevated poetry’ (p. 126).

⁸⁹ Cf. section IV above. Hinds' astonishing omission of reference to Calliope's speech in the certamen Musarum in Fast. 5 undermines his attempt to base a generic distinction on her presence in Met. 5. Hinds' distinction would, I think, rest more securely on developing the contrast between the Muse's victory in Met. 5 (663ff.) and her failure to repeat the success in Fast. 5 (cf. 107–10).

⁹⁰ The account of Aristaeus, Cyrene and Proteus (G. 4.315–558) is used for Fast. 1.363–80, but some of the material may be older than either version (see Bömer ad he). For the comparison of both poems to a ship see n. 30 above.

⁹¹ Cf. Call. h. 5.55–6, where Bulloch's note provides mor e parallels an d discussion of the generalised instances. Mr A. S. Hollis draws my attention to the specific instance of Callimachus' questioning of Theugenes (fr. 178 Pf., 20ff.). The device shows Ovid and Virgil writing expansively within a tradition and is not merely, as Frécaut has it, ‘piquer la curiosité du lecteur’ (p. 160 on Fast. 2.304).

⁹² On Call. h. 5.55–6, comparing fr. 612 Pf. áμáρтυρоν ούδν áείδω.

⁹³ Cf. Bömer i.29; in his note on 2.584 he fails to see the common functions of these ‘persönliche Erkundigungen’ in Ovid and those which Bulloch lists in Hellenistic poetry. The formulaic element in such attributions is illustrated by Hollis on Met. 8.720–1.

⁹⁴ Wilkinson (pp. 247–9) takes these examples of the ‘personal approach’ very literally, as does Bomer on 4.905, though the same approach in his note on 6.396 gets him into serious topographical difficulties. Ovid's encounters need not be autobiographical in any historical sense, and would hardly be worth recording in a poem like Fasti if they were; it is naive to believe that Ovid actually went about researching Fasti in the way they suggest. They are rather evidence of his working within a literary tradition. D'Elia, S. (Ovidio [Naples, 1959], pp. 360–70)Google Scholar more sensibly suggests a comparison with the colloquial tone of Horace's Satires and Epistles 1.

⁹⁵ Cf. Ars am. 2.357 ‘sed mora tuta brevis’, 653–4, 3.752, Rem. 83, 92 and Pichon, R., Index verborum amatoriorum (Hildesheim, 1966), p. 207Google Scholar s.v. mora. The ‘longa mora’ is by contrast a dangerous omen, cf. Heroid. 1.73–4.

⁹⁶ The rejection of arma by the amatory poet (Am. 1.1.1) is recalled here. It matters little to my argument if ‘voxque’ is read here with Burman, since ‘Venusque’ is sufficiently explicit. Palmer (1898,ad he.) needlessly impugns the authenticity of the couplet.

⁹⁷ Cf.DRN 5.517–25 and 1241–9, Georg. 1.84–93, Aetna 102–17 and 282–305, all of which are considerably more elaborate than the hellenistic equivalent we find at Call. h. 1.6–7, or at Arat. Phaen. 98–9.

⁹⁸ Fast, E.g.. 3.771–90, 839–8, 6.569–636. More examples can be found in Wilkinson, p. 265Google Scholar. If, as Wilkinson suggests, the reason is that ‘[Ovid] did not think deeply about what he read’, why should he bother to record multiple causae at all, when one would – as was the case in most instances – suffice? The parallels in other didactic poems suggest that Ovid knew clearly the effect he wanted to achieve, though he may also have been influenced by the appearance of simple alternative explanations for a festival (the Agonalia) in the Fasti Praenestini (see Wallace-Hadrill, , pp. 225–6Google Scholar and 229).

⁹⁹ ‘“Si credere dignum est”: Some Expressions of Disbelief in Euripides and Others’, PCPS 22 (1976), 60–89Google Scholar, especially pp. 63–5 which quote examples from Fasti. ‘Vix equidem credo’ occurs also at 2.203 (where cf. Bömer) and 551. See Fréecaut, pp. 162–4.

¹⁰⁰ Cf. Cic. Div. 2.98 with Pease ad he. (p. 518); Plut. Rom. 12.1.

¹⁰¹ It is just possible that an alternative reading at 807 ‘ipse locus causas vati facit’, which Frazer actually prints but which Landi–Castiglioni, Bomer and Alton–Wormell–Courtney all rightly reject in favour of ‘ipse locum casus vati facit’, is in fact a witty scribal gloss on the section 783–806 and emphasises this stretching of the poet's didactic powers: ‘the point reached in the month [i.e. the conjunction of the Parilia and Rome's birthday] itself provides [an abundance of] causae for the poet!’

¹⁰² Beard seems to me to make too much of the content of individual causae here, e.g. on p. 10 where she misses the literary significance of the section for the poet's authoritative persona and so fails to see how, fortified by his special gifts and status, the poet may treat the topic of Rome's birthday from an unconventional angle. Porte (p. 30) is equally insensitive here.

¹⁰³ Cf. section III, p. 171 above.