¹ The lack of reference to the twelve other letters in the collection has cast doubt over their authenticity. Hinds, S. E., ‘Medea in Ovid: scenes from the life of an intertextual heroine', MD 30 (1993), 9–47Google Scholar at 30–4, however, persuasively demonstrates the inadequacies of using Amores 2.18 to draw positive conclusions about the unmentioned letters’ authenticity.

² See Hinds (n. 1), 32, who follows Palmer's similar observation.

³ Cf. Casali, S., P. Ovidii Nasonis Heroidum Epistula IX(Florence, 1995), 12–13.Google Scholar R. J. Tarrant writes, ‘Extended argument with persuasion as its object is central to the Heroides as to no other work of Ovid; perhaps not coincidentally, the futility of those arguments is nowhere else as obvious’ (‘Ovid and the failure of rhetoric’, in D. C. Innes, H. Hine, and C. Pelling [edd.], Ethics and Rhetoric [Oxford, 1995], 63–74 at 67). The argument is often conducted in a plaintive tone, as the frequency of queror and querela (in the single letters) indicates:Her. 1.8, 1.69–70, 2.1–2, 2.8, 2.26, 3.5–6, 5.73, 6.17, 7.29–30, 8.68, 9.2, 13.158, 14.67, 14.110; the words also appear in the Epistula Sapphus at 15.71,15.136.

⁴ But cf. what Helen says at Her. 17.193 (although she also develops Medea's betrayal at greater length [17.229–33]). At Prop. 1.15.17–20 Hypsipyle is also mentioned as the lover abandoned by Jason.

⁵ see Verducci, F., Ovid's Toyshop of the Heart (Princeton, 1985), 66.Google Scholar

⁶ Hinds (n. 1), 44.

⁷ Damon, C., ‘Poem division, paired poems, and Amores 2.9 and 3.11', TAPhA 120 (1990), 269–B–A sequencing (5–6–7) to Propertian pairing within a book (1.7–9) to Ovidian pairing in separate books (1.4–2.5 and 2.19–3.4) is an important heuristic precedent for considering Heroides 6 and 12 a pair.Google Scholar

⁹ Jacobson, H., Ovid's Heroides (Princeton, 1974), 108.Google ScholarBarchiesi, A., P. Ovidii Nasonis Epistulae Heroidum I-III (Florence, 1992), 21Google Scholar discusses the tendentiousness of Hypsipyle's portrayal of Medea to Jason.

¹⁰ Knox, P., Ovid: Heroides, Select Epistles (Cambridge, 1995), 24–5.Google Scholar

¹¹ Hinds (n. 1), 27: ‘Hypsipyle's letter, in fact, is more about Medea than it is about Hypsipyle herself.’ Cf. Verducci (n. 5), 62: ‘Only Ovid exploited the erotic potential merely suggested in Apollonius, and only Ovid made any connection at all between Jason's casual treatment of Hypsipyle and his rejection years later of his wife, Medea, for his new bride, the princess Creusa. Thus, it is apparent that however otiose in its rehearsal of the stale details of Medea's career Hypsipyle's malediction might seem to be, it would still strike the reader as surprising, and as a surprisingly unfamiliar, daring, unprecedented alteration of the tradition, simply by virtue of the fact that it is Hypsipyle who speaks.’

¹² Jacobson (n. 9), 102.

¹³ Ibid., 113 in reference to Palmer's interpretation of Her. 12.121–2.

¹⁴ Ibid., 94.

¹⁵ Ibid., 95.

¹⁶ Knox, following Peters, deletes the couplet 6.139–40 in which Hypsipyle explicitly mentions the Lemnian women's crime; his reasons are persuasive to me. Retaining the two verses, as Goold does, amounts to another important difference between Heroides 6 and Apollonius, in which Hypsipyle attempts to conceal the Lemniadum facinus (Arg. 1.793–833). Jacobson (n. 9), 96–7, 106–8 provides a more comprehensive account of similarities and differences between Heroides 6 and Apollonius.

¹⁷ Knox observes dangers of postulating influences in the absence of complete textual records (‘Phaethon in Ovid and Nonnus’, CQ 38 [1988], 536–51 and ‘Ariadne on the rocks: influences on Ovid, Her. 10’, in P Knox and C. Foss [edd.], Style and Tradition [Stuttgart and Leipzig, 1998], 72–83). Nevertheless, calling Apollonius’ Argonautica ‘canonical’ is fair with reference to Medea's story. M. Leigh, ‘Ovid, Heroides 6.1–2’, CQ (1997), 605–7 at 607 makes a useful point: ‘[The approach suggested] may be said to privilege two accounts which survive at the expense of the multiplicity of ancient versions of this legend which do not…. However, closer examination suggests that [Euripides’ Medea and Apollonius’ Argonautica] are indeed the crucial texts.’.

¹⁸ Dido makes a similar point to Aeneas at Her. 7.133–8. See also Knox (n. 10), ad loc.

¹⁹ At Apoll. Rh. Arg. 4.162–3 Jason takes the fleece at Medea's bidding. The Ovidian dedi corresponds with Apollonius’ κoúρης κεκλoμ⋯νης.

²⁰ The verse recalls the incipit of Catullus 76, but does so perversely: Catullus recalls his good deeds to demonstrate the good things he deserves; Medea recalls her kind services to emphasize the punishments Jason should endure.

²¹ In Ennius’ Medea Exul Medea also uses first-person verbs (frs. 274–5 Vahlen; but see Jocelyn, H. D., The Tragedies of Ennius [Cambridge, 1967], 350Google Scholar on the authenticity of these lines): ‘non commemoro quod draconis saeui sopiui impetum, / non quod domui uim taurorum et segetis armatae manus’.

²² For Medea's fame as an enchantress, cf. Pind. Pyth. 4.233; Theoc. Id. 2.15–16; Apoll. Rh. Arg. 3.26, 3.528fF., and 4.1677; Euphorion 14.3 Powell; Hor. /. 4.9–14, 5.61–6, C. 2.13.8; Prop. 2.1.54, 2.4.7, 3.11.9; Tib. 2.4.55–60. At Met. 7.135–8 Medea doubts the efficacy of her gramina, only to celebrate her magic powers at Met. 7.199–206; Hypsipyle uses language similar to Met. 7.199ff. to caricature Medea at Her. 6.86–93.

²³ That Eur. Med. 469ff. and Apoll. Rh. Arg. 4.355ff. lurk in the background is a given; it is quite possible that Ovid's Medea (e.g. fr. 1 Owen), among other productions, does as well. Met. 7.56, where Medea herself says, ‘magna sequar: titulum seruatae pubis Achiuae’, appears to take into account Hypsipyle's words here. Knox's comment (n. 10), 190–1, on tauros (Her. 6.97) fails to appreciate the especial relevance of Heroides 12 in the light of 6.97–104: ‘apparently the fire-breathing bulls of Aeetes, whom Medea did not yoke, though she did make it possible for Jason to harness them by protecting him from the flames. O. is more interested in equipping Hypsipyle with a rhetorical point than in mythographical precision.’ This comment indirectly reveals the kind of loss that ensues when one excludes Heroides 12 from the collection (as Knox, ‘Ovid's Medea and the authenticity of Heroides 12', HSCP [1986], 207–23 recommends doing— although he acknowledges in his recent commentary that his Harvard Studies judgement is not shared by many: ‘Despite occasional protests… Ovidian authorship of Her. 1–14 is assumed by most modern scholars’ [pp. 7–8]).

²⁴ Apoll. Rh. Arg. 3.1380ff. Although Jason does engage the spartoi, he does so under Medea's instruction: he knows it is safe. With that said, Jason's heroism is far too complex to be judged only in terms of Homeric βíη. The bibliography here is vast; for some recent studies of this perennial issue, see Hunter, R.; “‘Short on heroics”: Jason in the Argonautica’, CQ 38 (1988), 436–53CrossRefGoogle Scholar; Dyck, A. R., ‘On the way home from Colchis to Corinth: Medea in book 4 of the Argonautica’, Hermes 117 (1989), 455–70Google Scholar; Clauss, J. J., The Best of the Argonauts (Berkeley, 1993)Google Scholar; Hunter, R., The Argonautica of Apollonius: Literary Studies (Cambridge, 1993)CrossRefGoogle Scholar; and now Holmberg, I. E., ‘Mñτις and gender in Apollonius Rhodius’ Argonautica’, TAPhA 128 (1998), 135–59.Google Scholar

²⁵ A further echo of Catullus 76?

²⁶ In this instance Hypsipyle fits the Ovidian conceit, militat omnis amans. Her. 6.79–82 also resonate against the Euripidean speech (Med. 536fF.) in which Jason kindly reminds Medea of the benefits she has accrued by living in a civilized, Greek polis.

²⁷ Verducci (n. 5), 57.

²⁸ Ibid., 66: ‘From the vantage point of [Heroides 6] the true irony of Hypsipyle's letter emerges with shocking clarity: it is she, and not Medea, who is the true Medea of Ovid's collection.’

²⁹ Jacobson (n. 9), 104.

³⁰ Hinds (n. 1), 34–43 and Barchiesi, A., ‘Future reflexive: two modes of allusion and Ovid's Heroides’, HSCPh 95 (1993), 333–65Google Scholar at 343–5 have separately demonstrated how the conclusion of Medea's letter displays signs of generic consciousness: Medea is about to enter Euripides’ tragedy. Yet whereas Medea hesitates to become cothurnata, Hypsipyle not only is happy to allude to (and predict [sc. 6.153fF.]) Medea's buskined existence, but even suggests (6.151) she would rise to such enormity herself.

³¹ Verducci (n. 5), 58–9.

³² Leigh (n. 17), 605–7 notes how the beginning of Hypsipyle's letter displays this quality.

³³ At Apoll. Rh. Arg. 4.438ff. Medea prays that the Furies drive Jason from his country. Virgil's Dido, another persistent parallel for Ovid's heroides (and author of the next letter!), draws similar curses on Aeneas’ head at Aen. 4.615ff.

³⁴ Jason and Medea's discussion of Ariadne at Apoll. Rh. Arg. 3.1074–1101 is a happy parallel.

³⁵ Kennedy, D., ‘The epistolary mode and the first of Ovid's Heroides', CQ 34 (1984), 413–22CrossRefGoogle Scholar at 414. G. D. Williams, ‘Ovid's Canace: dramatic irony in Heroides 11’ CQ (1992), 201–9 at 201 and Leigh (n. 17), 605 emphasize the same point. Fowler, D. P., ‘First thoughts on closure: problems and prospects’, MD 22 (1989), 75–122Google Scholar at 100–1 contains a brief discussion of the complexity that intertextuality introduces to closure.

³⁶ Rosati, G., ‘Il parto maledetto di Medea (Ovidio, Her. 6, 156s.)’, MD 20–1 (1988), 305–9Google Scholar at 309 writes, ‘… il tragico storico di Medea (in una disposizione cronologicamente conseguente dei suoi momenti più drammatici) è compiutamente delineato nella maledizione di Ipsipile’.

³⁷ Williams (n. 35), 201; cf. Casali, S., ‘Tragic irony in Ovid, Heroides 9 and 11’, CQ 45 (1995), 505–11.CrossRefGoogle Scholar

³⁸ Many of the heroides note the tears they shed over their letters (e.g. Briseis at Her. 3.4, Phaedra at Her. 4.175–6, Dido at Her. 7.185–6, Ariadne at Her. 10.149–50, Laodamia at Her. 13.113–14); Ovid himself does so at Am. 2.18.22 and, importantly, at Trist. 1.1.14. Kauffman, L., Discourses of Desire: Gender, Genre, and Epistolary Fictions (Ithaca, NY, 1986), 36Google Scholar describes how tears can function in epistolary discourse as an irrefutable communication with the power to materialize words.

³⁹ Miller, J. F., ‘Ovidian allusion and the vocabulary of memory’, MD 30 (1993), 153–64Google Scholar discusses a number of such annotations. See also Wills, J., Repetition in Latin Poetry (Oxford, 1996), 30–1.Google ScholarHunter, R., Apollonius of Rhodes: Argonautica Book III (Cambridge, 1989), 200Google Scholar notes the interplay (through memory and time) of Her. 12.67–9 with Arg. 3.927–31.

⁴⁰ Cf. also Ajax and Ulysses in Metamorphoses 13.

⁴¹ Medea also hesitates to describe the murder at Apoll. Rh. Arg. 4.735–7, where she tells her story to Circe.

⁴² Hypsipyle refers to the infamous hands of Medea (6.128), whereas Medea's own, the agent of so much evil, cannot recount its deeds (12.115). Medea's inability to describe the details of Absyrtus’ murder is a sign of Ovidian generic play. See Barchiesi (n. 30), 343–5.

⁴³ C. Guillén, ‘On the edge of literariness: the writing of letters’, Comp. Lit. Stud. 31.1 (1994), 1–24 at 5: ‘There is hardly an act in our daily experience, rooted in life itself, that is as likely as the writing of a letter to propel us toward inventiveness and the interpretation and transformation of fact.’

⁴⁴ Dido's curses against Aeneas and his descendants (Aen. 4.622–9) have similar literary ramifications: in a less direct manner they occasion the composition of Ennius’ Annales. I owe this point to Professor Barchiesi.

⁴⁵ Apollonius is quite clear that his Medea lacks control over some important decisions in her life once Hera and Aphrodite have conspired to direct it (Arg. 3.25ff.).

⁴⁶ One could make the case that Hypsipyle's letter offers Jason precisely what he seeks in marrying Creusa, and so could be relevant to a later life. However, Euripides'Hypsipyle would suggest this was never a possibility.

⁴⁷ See note 17 supra.

⁴⁸ In our texts Heroides 12 also begins like Heroides 6. (Whether or not we have the beginnings as Ovid wrote them is for now an irresolvable issue.) Medea starts abruptly, with no salutation, as she recalls the time she gave Jason in Colchis. Hypsipyle, too, begins her letter without greeting as she accuses Jason of his failure to write her. Heroides 12's beginning also implies that Jason has yet to give Medea the time for a conversation (hence the striking delay of the standard Medea opening); this implication brings with it the concomitant need for a letter (not a drama). I owe this point to Dr Heyworth.

⁴⁹ Unlike Jason, Propertius will not abandon his mistress. He fears in 1.6 Cynthia will say (1.6.18), ‘et nihil infido durius esse uiro', when (1.6.9–10) ‘minatur / quae solet ingrato tristis arnica uiro’ (cf. Her. 12.21). Jason must be hard as a Virgilian oak: he endures querimoniae from two heroides, while Propertius says right off (1.6.11), ‘his ego non horam possum durare querelis’. Propertius offers an instance of the success of ‘Ovidian’ rhetoric (contra Tarrant [n. 3]).

⁵⁰ Medea, however, vacillates at moments in her letter and asks to be taken back (12.193fF.). On such occasions Ovid follows Argonautica 4, in which Medea reinstates herself at an earlier point in her relationship with Jason. In Euripides’ (and Ovid's?) tragedy Medea addresses Jason only as she plots her revenge.

⁵¹ See Barchiesi (n. 30), 343–5 and Hinds (n. 1), 41–3 for the import of this line in relation to Prop. 2.34.66.