¹ Dörrie (1960) ii. 377–9. References to the text of the Heroides in this paper follow the standard numeration as used by (e.g.) Showerman-Goold, and not that of Dome (1971).

² See Bömer on Met. 8.595 ff.

³ Reeve, p. 337 n. 1; cf. Fischer, p. 74 and n. 1. On Reeve's expedient for reconciling this judgement with his condemnation of the passage see further below, p. 417 n. 73. 1 believe 21. 145–248 to be equally worthy of Ovid, but the state of the text towards the end of the epistle makes the point more difficult to sustain. Cf. also below, p. 428 n. 118.

⁴ Luck, G., Untersuchungen zur Textgeschicbte Ovids (Heidelberg, 1969), p. 21,Google Scholar cited with (qualified) approval by Reeve, p. 336 n. 7 (but cf.ibid., p. 337 n. 1). The dubious honour of coining this fallacious dictum appears to be ascribed to Rudolf Ehwald.

⁵ See Housman, A. E., CR 11 (1897), 426–7 = Classical Papers pp. 413–14, citing in addition Met. 5. 414, 9. 131–2, 10. 568–70, 11. 134–7, 263, F. 4. 597–8.Google Scholar

⁶ Cf. Pohlenz, M., ‘Die Abfassungszeit von Ovids Metamorphosen’, Hermes 48 (1913), 4–9;Google Scholar Latta, p. 119; Kraus, p. 56; Tracy, Valerie A., CJ 66 (1970-1), 330.Google Scholar

⁷ I am extremely grateful to Dr. Fischer for sending me a copy of her dissertation, and I only hope that my frank criticisms of what seem to me to be fundamental weaknesses in her case may not appear a churlish return for her courtesy. In view of the weight attached by Mr. Reeve and Professor Goold (cf. Goold, p. 483) to her arguments it seemed to me indispensable to examine them closely before they begin to be taken on trust and an orthodoxy is established by default. That is also why this paper is so long.

⁸ In quoting from the disputed passages I give the text of π (ignoring trivial printer's errors) unless I indicate otherwise. Only the edition of Dörrie (1971) makes it clear that n is our sole authority for the transmitted text.

⁹ Cf. also ex P. 3.7.24 ‘… seque semel uera scire perisse fide.’

¹⁰ See Jacobson, H., ‘Ennian influence in Heroides 16 and 17’, Phoenix 22 (1968), 299–303;CrossRefGoogle ScholarWhite, Diana G., ‘Ovid, Heroides 16. 45–6’, HSCP 74 (1970), 187–91.Google Scholar

¹¹ For this timeless sense of the perfect participle see Laughton, E., The Participle in Cicero (Oxford, 1964), pp. 2–3.Google Scholar

¹² Cf. Latta, pp. 87–8.

¹³ Simple phrases, of which Fischer (p. 73 n. 2, p. 114) makes extraordinarily heavy weather.

¹⁴ Cf. Palmer, pp. 436–7.

¹⁵ See Dörrie (1971) ad loc; Diggle, J., ‘Notes on the text of Ovid, Heroides’, CQ N.S. 17(1967), 139;C. Austin-M. D. Reeve, Notes on Sophocles, Ovid, ', Maia 22 (1970), 6; Fischer, pp. 75–6.CrossRefGoogle Scholar

¹⁶ A point of some delicacy for the poet to negotiate. Cf. on Ovid's treatment of the Judgement from Oenone's point of view Jacobson, pp. 193–4.

¹⁷ at Itali; cf. Fischer, pp. 76, 136; Reeve, p. 336 n. 1. The arguments of Asteroth (reported by Fischer, p. 76) and Latta (pp. 127–8, 160 n. 36) for deletion do not convince. Nor do I see any difficulty about taking ‘subeunt’ as historic present (Reeve, loc. cit.). On the relationship of this couplet with 17. 195–8 see below, §§5.2. 13–14.

¹⁸ fades π, corr. Heinsius, prob. Bentley. The construction is equivalent to ‘quid faceres si turn adesses?’, and the correction is absolutely necessary; cf. Goold, p. 479. The conviction of Palmer (p. 441) and Fischer (pp. 77, 137) that it makes no difference to the sense is mystifying; cf. Reeve, p. 337 n. 1. Fischer's objection to bic (Itali: bine π) in line 104 also falls to the ground if faceres is accepted.

¹⁹ Cf. also Tr. 3.4. 55–6, 4.2. 57; Scholte, A., Publii Ovidii Nasonis ex Ponto liber primus (Amersfurt, 1933), p. 158 (ad 1.8.34); Cic. Or. 101 ‘eloquentia ipsa, quam nullis nisi mentis oculis uidere possumus’. Cf. also Plat. Symp. 219 a 2–3 )… .Google Scholar

²⁰ Noticed by Dörrie (1960), ii. 360 n. 1.

²¹ We need no longer take seriously the possibility of a Renaissance origin of the passages.

²² There seems to be little to choose between ‘malis’ and ‘malo’. If I incline to the latter it is less for the reason that Homeric ships had only one mast (Palmer, pp. 441–2) than because the variation ‘antennas’ (pi.) … ‘malo’ (sing.) is more elegant than two plurals; cf. Hor. C. 1.14. 5–6, cited by Palmer, p. 442.

²³ Bomer on Met. 2.337, 5.262; on nec tamen ═ see below, p. 410 n. 48.

²⁴ ‘iubemur’ is the right tense but is open to the same objection as ‘iubebar’ on the score of vagueness: if it refers to Venus as the source of the sailing orders it is excessively elliptical, and it was not perhaps really for her to give them.

²⁵ As we might do by reading, for instance, ‘concepimus’ at line 125; cf. Met. 7.9, 17, 9.520, al.

²⁶ This is the correct text, a change of one letter (‘nemorosis’ for ‘-us’) from the reading of it. The vulgate ‘in mediis nemorosae’ is Francius's superfluous conjecture, on the model of A. A. 1.289 ‘forte sub umbrosis nemorosae uallibus Idae …’

²⁷ Cf. Latta, pp. 132–3.

²⁸ Negligence on somebody's part is all that ‘nudas’ can be made to prove; whoever wrote it did so for the same reason, that the familiar detail came so readily to mind that he quite forgot that there had been no previous allusion to it. Fischer's theory (p. 143) of two distinct versions of the story introduces quite unwarranted complications. If the double Heroides were written shortly before Ovid's departure into exile (cf. above, p. 396 n. 6), it is understandable that they were never carefully revised.

²⁹ Cf., for what they are worth, Alciphron 4.14(1.39). 4–6, Rufinus, A. P. 5.35, 36 (XI, XII Page).

³⁰ This is what may be called the basic version: see Dörrie (1971) ad loc. The expression is modelled on 17. 119–20.

³¹ Not 1–38; see below, §§5.3.8, 5.4.3.

³² Here too the rhetorical divisions of the speech do not coincide with the limits of the disputed passage; see preceding n.

³³ ‘Ne nescia pecces’ ═ (i) ‘so that you may be under no misapprehension through ignorance’-, (ii) ‘so that when you sin <, as I shall now show you that you must>, you will know why.’ Translators and commentators seem to have missed the play on the two senses of pecco: see OLD s.v. 1 and 3.

³⁴ Cf. Fischer, pp. 143–4.

³⁵ Fischer, p. 144.

³⁶ Hyginus 92; Apollod. Epit. 3.2; Cypria (Kinkel 17); Eur. Tro. 929–30, Hel. 27–9; Isoc. Hel. 41–2; Lucian, Dial, deorum 20.14; Colluthus 164–5.

³⁷ Cf. Latta, pp. 134–5.

³⁸ On this theory the naivete that the poet makes Paris attribute to Menelaus verges on idiocy. What on earth did they suppose he had come to Sparta for?

³⁹ Or not all at once; it is one of the delights of reading her letter to see how quickly (131 ff.) she acquiesces in what she has just affected to disbelieve.

⁴⁰ Fischer, pp. 62–3.

⁴¹ Fischer, pp. 145–6, Reeve, p. 335.

⁴² Cf. Birt, T., GGA (1882), 839–40; Latta, pp. 130–1. Both, however, weaken their respective interpretations by taking Helen's words to refer only to one passage of Paris' letter, Birt to 121–6, Latta to 47–50.Google Scholar

⁴³ This point is clearly and persuasively made by Latta (p. 130), but apparently ignored by Fischer (p. 146).

⁴⁴ The antecedent of ‘quos’ at 17.239 is generally taken to be the uates; so, for instance, Showerman-Goold, Prevost, Ripert. Contra Riley: ‘I fear, too, the presages of the prophets, which, they say, forewarns [sic] us that Ilion shall be burnt by Pelasgian flames.’ ‘Monitus’ is next to ‘quos’, and it is more pointed, one would think, to say that the words of the seers, rather than the seers themselves, foretell. On the other hand ‘monitus’ as subject of ‘praemonuisse’ is perhaps odd writing. Did the poet write ‘quibus’?: ‘I fear the warnings of the seers, by which men say they have foretold …’

⁴⁵ Fischer, pp. 116–17, repeating and amplifying the arguments of Asteroth.

⁴⁶ Fischer, p. 144.

⁴⁷ Another idiomatic phrase of which Fischer (p. 116 n. 4) makes very heavy weather. It gives pointed and perfect sense in the context.

⁴⁸ Cf. e.g. Her. 12.117, 16.373, 18.99; Bömer Met. 6.329, 7.453; Shackleton Bailey on Cic. Fam. 4.12 (253).3; Roby, ii. 460, §2206, citing inter al. Cic. Sen. 13. For ne tamen cf. e.g. Her. 20.129, 213, 21.231. For tamen ═ cf. above, p. 401 n. 23.

⁴⁹ One further point made by Fischer in this connection ought perhaps to be mentioned, since it is echoed by Reeve (p. 336 n. 5). She objects that 39–144 spoil a progression of emphasis in Her. 16 from past to future: ‘Dieses Prinzip wiirde durch die Verse 43 ff und 121 ff durchbrochen, da sie schon im ersten Teil des Briefes einen Hinweis auf die Zukunft geben warden’ (p. 140). In point of fact lines 149–62 (on Theseus) take the story back to a stage before the events hinted at in 15–36 (Venus' promise and the voyage to Sparta); then there follows another allusion to the Judgement (165–70), then reference to Helen's hoped-for reception in Troy (185–8), then events in the immediate past and the present (217–62), then Menelaus' absence (299–316), then again the anticipated triumphal reception at Troy (331–8); and these passages are interspersed with pleas, arguments, and exempla. It is true that the epistle ends (341 ff.), as Fischer notes (Ioc. cit.), with references to future events, chiefly by way of dramatic irony, but to call all this an ‘illusion of a narrative progression’ (Reeve, loc. cit.) seems to me greatly exaggerated. The excursus in 41 (see below, §§5.3.8, 5.4.3)-128 indeed ‘puts the clock a long way back’, but it is not equally true that ‘the prophecies in 49 and 123–4 put [it] a long way forward’ (Reeve, loc. cit.). They do admittedly in a sense carry the reader forward, but to the present; for the emphasis is less on the temporal implications of these vaticinations than on the actual state of Paris' feelings and his flippant interpretation of the apparently ominous words. It cannot be seriously maintained that they constitute a disruption of a carefully planned narrative structure. As to the digression in which they figure, I have already argued (above, §5.2.5) that towards the beginning of the epistle was the best-or at the very least a perfectly appropriate-place for it.

⁵⁰ Cf. Latta, pp. 123–4, 134–5.

⁵¹ Fischer (pp. 71–2, 80 n. 2) seems not to have fully grasped this point. See Latta, p. 154 n. 16.

⁵² Rem. 44 = Tr. 2.20 ‘uulnus opemque feret’ cannot be allowed as a real instance to the contrary.

⁵³ Cf. Fischer, p. 72, who however exaggerates in terming ‘nuntia fama tui’ ‘nichtssagend’.

⁵⁴ Fischer objects that durch die Änderung in v. 38 die Erklarung des “ante … animo vidi”, die darin liegt, das die “fama” von Helenas Schönheit kündete, beseitigt wird und an ihre Stelle eine neue, grössere Unwahr-scheinlichkeit tritt' (p. 72). This is, I think, mistaken. The correction still leaves fama as the agent responsible for both phenomena.

⁵⁵ ‘Crede sed” was introduced into the text by Daniel Heinsius (to assist the connection of thought between 144 and 145, it should be noted: see below, §5.3.7) ‘ex veteribus codicibus”; Dörrie (1971) reports it from Flor. Laur. 24, 8, a manuscript of the thirteenth century.

⁵⁶ Pace Reeve, p. 334 n. 5. His note appears to suggest that crede sed can be accommodated in the category (‘syntactical form”) crede mibi. The stumbling-block, however, as with credis et, is hoc; it is this which makes it difficult to classify either phrase as parenthetic-formulaic. There is “hoc quoque si credis” at Her. 18.121; but see below, p. 420 n. 81.

⁵⁷ It is of course a familiar folk-tale motif: Thompson, S., Motif-index of Folk Literature v (Copenhagen, 1957), 333, 111, 11.1.Google Scholar

⁵⁸ That must be the gist of lines 141–4. I do not challenge Reeve's view (p. 336 n. 3) that my own attempt at emendation (Kenney, pp. 179–81) does not dispose of all the difficulties; but I cannot agree with him that 143–4 ‘could … be deleted without damage to the sense’. Lines 141–2 only state that the fame of her beauty was everywhere; 143–4 make the clinching point that it was unrivalled, which is not the same thing. Reeve also objects that my “idea of separating nec tibi par usquam Pbrygiae from nomen habet is an unhappy one: the sense required by the antithesis of which it forms part is not “you have no equal in Phrygia” but “your fame has no equal in Phrygia” I should not go to the stake for my suggestion, but I do not think Reeve's argument invalidates it; for a parallel to the compendious phraseology ‘nee tibi par <est>’, sc. in the matter of fame, one need look no further afield than line 141 ‘magna quidem de te rumor praeconia fecit’, sc. in the matter of beauty. In each case the pronominal reference is expanded and clarified in what follows.

⁵⁹ Cf. Birt, , GGA (1882), 841.Google Scholar

⁶⁰ ‘… plump und ilbertrieben’ comments Fischer (p. 80), comparing Prop. 2.2. 13–14; but that does not prove it to be spurious. Paris' ‘quantum reminiscor’ too she finds ‘nicht eben geschmackvoll’ (p.81). But good taste was never Ovid's strongest suit; and is not a stroke of wit intended? The recollection of Venus' beauty, overwhelming at the time, deserts him at the first sight of Helen: 135 ‘ut uidi, obstipui’ eqs.

⁶¹ Cf. above, p. 411 n. 55. Daniel Heinsius seems to have been in no doubt as to the authenticity of 39–144; his note on ‘crede sed’ reads merely ‘magis arridet’.

⁶² Fischer's argument here (p. 138) over looks ‘quidem … sed’ at 131–4; as Reeve implies (p. 334 n. 4), it is difficult to see any difference. The sense is equivalent to a concessive clause: ‘Though Menelaus showed me everything, yet what I really wanted to see …’; ‘Great as was the fame of your beauty, yet it was outdone by the reality’. Cf. Bömer on Met. 6.679.

⁶³ Fischer, following Asteroth, objects to the connection given by ‘nee tamen’ as ‘logisch schief’ (p. 71). Cf. however above, p. 410 n. 48.

⁶⁴ On this kind of argument cf. von Wilamowitz-Moellendorff, U., Erinnerungen 1848–1914 (Leipzig, 1928), p. 222: ‘Aristoteles … dem das Buch [Ath. Pol.] abgesprochen war, nach der nicht selten be- folgten Methode: “ich habe mir's anders gedacht, also ist es nicht das richtige”.’Google Scholar

⁶⁵ Cf. Latta, p. 154 n. 17.

⁶⁶ Dörrie (1960), ii. 377–9.

⁶⁷ Fischer's scepticism on this head is un founded and appears to rest in part on a mis understanding of the historical facts. It is untrue that at the time of the presumed losses (and the loss of 21. 15–144 must have occurred not much later) the tradition of the Heroides was ‘ziemlich breit’ (Fischer, p. 133). It was only from the twelfth century onwards that Ovid's poems began to circulate widely: see Tafel, S., Die Uberlieferungsgescbichte von Ovids Carmina amatoria. Verfolgt bis zum 11. Jahrbundert (Tubingen, 1910), pp. 66–72. To account for the state of affairs here outlined we are required to postulate the existence of no more than three Carolingian manuscripts of the Heroides apart from the surviving P (see my stemma, Gnomon, loc. cit.); and the interval between the re-emergence of the Heroides from oblivion and the genesis of the manuscript which, lacking 16. 39–144, 21. 145–248, and 15–144, was the source of the entire main tradition must have been quite short.Google Scholar

⁶⁸ It is unrealistic to attempt greater precision; cf. Reeve's justified scepticism (p. 335 n. 2) about Luck's reconstruction of the archetype. Fischer's objection (pp. 82, 133) that to calculate the lost verses as a multiple of a typical Carolingian page takes no account of a possible lacuna after line 50 (postulated, incidentally, by Scaliger, not Bentley) and perhaps also another after 88, carries no weight. A moment's thought suffices to show that those losses have no necessary connection with the loss of the passages discussed here and could have occurred at any time in the previous history of the text.

⁶⁹ Fischer, pp. 81–2; 133 ‘Ein nur zufälliger, so verbluffend passender Textausfall ist kaum glaublich'; Reeve, p. 335, ‘unless there is something wrong with [the con nection of thought between 38 and 145], the accidental loss of 39–144 would be a coincidence so enormous as not to be worth contemplating.’ The point had of course been taken before, e.g. by Kraus, pp. 63, 66.

⁷⁰ Scaliger was almost certainly right to suspect that something has been lost after 50. It is strange, in this so ample treatment of the story, to find no mention of Paris' exposure and nurture by shepherds; and as it stands line 51 is very abrupt. The lost portion need not have amounted to more than half a dozen verses. Cf. Latta, p. 157 n. 26. The case for a lacuna after line 88 is less clear; see Fischer, p. 75. Cf. above, p. 418 n. 68.

⁷¹ Cf. Kraus, pp. 65–6. The comparison with the narrative insertions of the single letters, adduced by Kraus, , RE 18 (1942), 1926, in support of authenticity and disallowed by Fischer (pp. 147–9), seems to me largely irrelevant; the double epistles must be judged in their own right. Much of Fischer's argument is disabled by what seems to me this fallacy.Google Scholar

⁷² As argued by (e.g.) Pulbrook, M., Hermatb. 122 (1977), 43 n. 9.Google Scholar

⁷³ Evidently he used in the Metamorphoses material that could not conveniently be accommodated in the Heroides and that was too good to waste. Thus the storm which, as appears from the comparison of Musaeus with Virgil (G. 3. 259–62), must have been the centre-piece in his source for Hero and Leander, was transferred to the Ceyx and Alcyone episode in Met. 11 (see Pohlehz, Hermes 48 (1913), 7–8). Conversely it is equally plausible that good material which did not fit the final plan for the Metamorphoses might have been used in Her. 16–21.1 should not rule out of court the idea that 16.39–144 represent an authorial expansion of an originally briefer treatment; cf. Kraus, p. 66. I am not, however, convinced by Reeve's variation on this idea: ‘I dare say a zealous executor may have found this fragment among Ovid's papers and with additions of his own incorporated it into the only possible context. Such speculation I do not find congenial, and it would never persuade me to leave the whole of 39–144 in the text for fear of sacrificing a fragment that Ovid might have written. Not interfering with finished poems is surely more important than not sacrificing fragments, and I do not believe that Ep. 16, when Ovid finished it, included any part of 39–144’(p. 337 n. 1). This argument depends on the premiss that Her. 16. 1–38, 145–378 can be described as a ‘finished poem’, which I deny.

⁷⁴ ‘… durch sie wird ganz in der Weise, die wir auch in den Einzelbriefen gefunden haben, eine Perspektive auf den Ausgang der Geschichte eröffnet; in ihrer Missdeutung durch Paris liegt eine tragische Ironie, durch die die Verblendung der Leidenschaft ins Licht gesetzt wird’ (Kraus, p. 66; for reservations as to the relevance of the single epistles cf. above, p. 416 n. 71). On the immediately humorous effect of Paris' mis interpretation see above, §§5.1.2, 5.1.6.

⁷⁵ Rand, E. K., Ovid and his influence (London, 1926), p. 27.Google Scholar

⁷⁶ It would be of the essence of the strategy to ensure that the god or goddess chosen suited the circumstances and would be respected by the victim. Diana was ap propriate in Cydippe's case because the trick was worked in her sanctuary and women used to swear by her.

⁷⁷ Kenney, , Philol. 111 (1967), 225–6 ═CrossRefGoogle ScholarArion 9 (1970), 400–2.Google Scholar

⁷⁸ The distinction between cum + indie. and cum + subj. is not always as rigid as some standard authorities (e.g. H.–S. pp. 624–5) would lead one to suppose: cf. Roby, p. 321, §§1731 (Plautus and Terence), 1735 (Cicero: esp. Ad fam. 6.14.1): K.–S. ii. 349. A possible alternative is to take the perfects as frequentative: ‘As often as I say all this, as often as I reject your suit, as often as I plead my case to my own satisfaction, <yet each time I end up>, I must admit, fearing … and suspecting …’; but the close parallel with the Amores passage makes this interpretation less likely than the other.

⁷⁹ At line 9 the words ‘utque cupis credi’ refer to and modify the sense of the whole passage; the verse only restates and varies what has just preceded. At 31 ff. the questions with which she begins her letter proper (‘ergo te propter …’ marking the start of the argument) colour what follows: her account of her attacks of illness is, so to say, apodotic to the unspoken protasis ‘if it is as you say …’. The words of line 53 ‘causa latet’ show that at this point of the letter she still does not accept implicitly what Acontius had told her at 20. 109 ff. Cf. Bopp, p. 100.

⁸⁰ TLL s.v. 1509. 31–6 also cites Phaedrus 3.13.9 (doubtful) and Martial 5.64.1 (irrelevant).

⁸¹ Not all the candidates for deletion listed by Goold, p. 479, deserve it, but others could be added, e.g. 4. 4–5 (Peters), 10. 85–6 (Bentley). A clear case in the double Heroides would seem to be 18. 119–21: see Housman, , CR 11 (1897), 427–8 ═ Classical Papers, pp. 414–15.Google Scholar

⁸² Not by Merkel (sic Showerman-Goold), who indeed omitted both our passages in his edition.

⁸³ Perhaps it is not totally indefensible: if Hymenaeus is figured, more Ovidiano, as both hymn and personified god, then it might be apt for him ‘to rise joyfully’. However, the conceit, though not in itself too far-fetched for Ovid, would be abruptly introduced; and though Latin can say clamortollitur, it does not appear that either surgere or consurgere is attested in such a sense. Perhaps there has been inopportune scribal reminiscence of Catull. 62.1.

⁸⁴ Kenney, pp. 182–3.1 am flattered by Dcirrie's ascription of this conjecture to Bentley. I owe it to myself to reclaim it, for it is printed by Showerman-Goold without any ascription at all.

⁸⁵ Reeve (loc. cit.) also points to 110 ‘insidias legi, magne poeta, tuas’, suggesting that 181 is ‘an over-ambitious adaptation’ of that verse.

⁸⁶ nil turn iurauimus Flor. Laur. 36, 2: nil nos i. Bentley: sed nil i. Palmer.

⁸⁷ It is printed by Showerman–Goold.

⁸⁸ Cf. Prevost's expanded version: ‘que cette joie perverse te soit voluptueuse’.

⁸⁹ Cf. Fischer, p. 191, rightly insisting that ‘ista’ ought to mean ‘this/that of yours’. Showerman–Goold's version of ‘ista uoluntas’ as ‘that state of my will’ is not very satisfactory; it entails accepting the explanation of van Lennep, approved by Palmer, that uoluntas, which generally bears a good sense when used absolutely ═ ‘disposition’, here takes its colour from ‘offensam’ in the preceding verse. That seems forced; and the difficulty of ‘ista’ remains.

⁹⁰ Editors, including most recently and unfortunately Showerman–Goold, attribute such a feeling to her in the concluding couplet by printing the unnecessary correction ‘coniungere’ (Aid.¹) for π's ‘contingere’. See Kenney, pp. 183–4, Fischer, p. 196 (approving).

⁹¹ The worst damage to the end of the epistle may have been done in antiquity. I do not think this very likely, but the possibility cannot be ruled out.

⁹² It is not in fact clear to me that this is really preferable to Francius's ‘uelut’.

⁹³ Cf. Bopp, p. 143.

⁹⁴ Fischer predictably calls ‘tamen’ ‘sinnlos’ (p. 194); cf. above, p. 401 n. 23, p. 410 n. 48.

⁹⁵ Pace Fischer (p. 194), this can be in dicated by a comma just as well as by a colon (Dörrie), perhaps better. The fact that ‘ope qua’ eqs. looks forward and not back is clear from the syntax. Anxious punctuation may occasionally be needed to save an incoherent writer from himself, but this situation does not often arise with Ovid. True hyperbaton is another matter.

⁹⁶ The objection of Fischer (p. 202) and Reeve (loc. cit.) to treating Her. 13.93 as a parallel is also disabled by this consideration.

⁹⁷ If the poet did indeed write ‘nescioqua’ at 21.233 he may have intended sarcasm: ‘the god complains that faith has “somehow” been breached. “Somehow”, indeed!’– Acontius, that is to say, knows very well about things of this kind that happen ‘somehow’.

⁹⁸ For ‘desensisse’ Heinsius' ‘de te sensisse’ seems to me superior to ‘de me sensisse’ (Aid.¹), which modern editors prefer.

⁹⁹ ‘ … the only remedy I can devise for 235–8 [233–6] is to scrap all four lines and start again’ (Reeve, p. 336 n. 7).

¹⁰⁰ Reading (e.g.) ‘numina’ with Dilthey.

¹⁰¹ Who is stillon guard: 21. 17–24.

¹⁰² Beeson, C. H., Lupus of Ferrieres as Scribe and Text Critic (Cambridge, Mass., 1930), pp. 21–2.Google Scholar

¹⁰³ Jacobson, pp. 337–8.

¹⁰⁴ Not thatCallimachus' Acontius is all that dynamic a figure; but such as he is he dominates the story, and the other characters exist chiefly as foils to him: Bopp, pp. 18–19, 26.

¹⁰⁵ Frr. 67–75 Pf.; Kenney, , philol. 111 (1967), 223 ═ Arion 9 (9170), 398–9; Bopp, pp. 9–11, 19, 96, 160–2.CrossRefGoogle Scholar

¹⁰⁶ I hope to enlarge on this when op portunity permits. Meanwhile it is enough to point out that Musaeus' characterization of Hero is flagrantly inconsistent and there fore almost certainly his own invention. Behind his and Ovid's very different treatments there may perhaps be glimpsed a hint of the presentation of Hero in the Hellenistic original (whose existence, pace Gelzer, I see no reason to doubt) as a type of the frustrated housebound female familiar to us in Horace's Neobule: Her. 19. 5–16 ~ Mus. 187–93, Agathias, A. P. 5.297. This, however, is a far cry from Ovid's delicate and sympathetic portrait.

¹⁰⁷ On the implications of all this I may refer to the sensitive analysis of Bopp, pp. 45 ff., esp. 45–53, 96–101.

¹⁰⁸ This important distinction is over looked by both Bopp (p. 144) and Fischer (pp. 199–200).

¹⁰⁹ An argument in favour of keeping ‘nunc’ at line 233; cf. above, §6.1.12.

¹¹⁰ Fischer apparently fails to grasp this point and so oddly misinterprets Kraus, referring (p. 200 n. 2) to an ‘Annnahme’ on his part.

¹¹¹ rapta Kenney, pp. 181–2: cauta codd.: coepta Dilthey.

¹¹² ‘Ky. subjektiviert die Orakelbefragung dadurch, dass sie deren günstigen Ausgang als weiteres giinstiges Zeichen für A. bucht: sie wird fiir sie ein Moment mehr, sich von diesem jungen Mann, der es sogar versteht, die Götter auf seine Seite zu bringen, ge-winnen zu lassen’ (Bopp, p. 163). The ‘downgrading’ of the role of the oracle and the replacement of Cydippe's father by her mother were among the modifications which were in a manner forced on Ovid by his ‘elegization’ of the story as he found it in Callimachus: his transformation, that is, of the Callimachean relationship (if such it can be called) of the two lovers into a relationship, recognizably elegiac in spite of the bizarre features inherent in the original plot, of wooing and surrender. AH this is admirably argued by Bopp; see esp. pp. 144, 162–3.

¹¹³ See Vessey, D. W. T. C., ‘Notes on Ovid, Heroides 9’ CQ N.S. 19 (1969), 349–61; contra Jacobson, pp. 228–42.CrossRefGoogle Scholar

¹¹⁴ Cf. Jacobson, p. 229.

¹¹⁵ The intervention does not, as with Cydippe, affect Helen's decision; its function is to contribute to the dramatic irony of 245–56.

¹¹⁶ Cf. Bopp, p. 45.

¹¹⁷ As Bopp (p. 80 n. 1) has duly remarked; cf. id., p. 79 n. 1, pp. 116–17. On Callimachus' indifference to the way in which Acontius is supposed to have heard of the outcome of the appeal to Delphi cf. Bopp, p. 18.

¹¹⁸ This paper is quite long enough already, and I do not propose to consider here Fischer's criticisms of the disputed verses on artistic grounds. The gravamen of her objections consists in the two heads:

(a) that ‘Sie bringen gedanklich nichts Neues; die Anordnung der einzelnen Punkte erscheint kunstlos, z.T. aus Gedankenver-bindungen im Brief des Acontius und im ersten Teil des Cydippebriefes übernommen’;

(b) that ‘In dem uns vorliegenden zweiten Teil des Briefes tritt uns eine andere Cydippe entgegen als im ersten Teil. Es liegt ein Bruch in der Gestaltung vor’ (pp. 221–2). In this admittedly subjective area I can only say that I find Bopp's interpretation of the epistle entirely persuasive, and I doubt if I could put the case better than she has done.

¹¹⁹ Cf. Bopp, pp. 102, 133.

¹²⁰ It does not have to do all the work. The real force of the expression resides in the structure, a tricolon crescendo articulated by the thrice repeated ‘cum’ (cf. Bopp, p. 133). This is highly Ovidian writing: cf. (e.g.) Am. 3.9. 21–2, quoted by Wilkinson, L. P., Golden Latin Artistry (Cambridge, 1963), pp. 203–4.Google Scholar

¹²¹ As allowed by Fischer, p. 187. This obvious point receives further support, if support be needed, from the fate of 21. 15–144, attested only in a handful of witnesses, amounting perhaps, after Reeve's eliminatio, to no more than three: .

¹²² The simplest reconstruction along these lines, so far as I can see, involves postulating, what is not inherently improbable, the survival into Carolingian times of two ancient copies of the Heroides, one containing and one lacking the disputed passages.

¹²³ The possibility that the passages, if spurious, are by two different hands does not seem to me worth pursuing. In that case we should have to imagine two forgers of equal technical competence; and unless each hit independently on the idea of including one instance of the peculiar use of nee noticed above (§3.1) as an Ovidian hallmark, then one must have imitated, not only his model, but also the pseudo-model produced by his colleague. The point is worth making because it throws into relief the remarkable courage of the postulated single forger in essaying to complete two such very different epistles, in tone and feeling, as those of Paris and Cydippe.

¹²⁴ Relevant comparisons are with Consolatio Liviae, Nux, Her. 15 (Ep. Sappbus), and Am. 3.5 (Somnium), none of which is by Ovid. Possibly Her. 9 should be added to the list, but of that I am far less certain. The Halieutica, it should perhaps be expressly stated, is irrelevant. On the generally low technical standards in the field see Speyer, W., Die literarische Fdlschung im heidnischen und cbristlichen Altertum. Ein Versuch ihrer Deutung (Munich, 1971), p. 113.Google Scholar

¹²⁵ Relevant comparisons here are more difficult to find. The ‘Helen-episode’ of Aeneid 2 comes immediately to mind; but it is very much shorter and offers a much higher concentration of features identifiable as anomalous: cf. Murgia, C. E., Calif. Stud. Class. Ant. 4 (1971), 213. On the general improbability of the idea that an interpolator, working back from Her. 17, could have produced 16.39–144 see Latta, pp. 132–6.Google Scholar