² Unless otherwise stated, the text printed is that of H. Dörrie's 1971 Berlin edition.

³ ‘tu mihi sola places’ is found in all three passages cited: Hollis (Ars Am. loc. cit.) styles it ‘the typical declaration of the elegiac lover’.

⁴ Fischer, U., ‘Ignotum Hoc Aliis Ille Novavit Opus. Beobachtungen zur Darstellungskunst Ovids in den Heroides unter besonderer Berücksichtigung der Briefpaare her. 16 und 17 (Paris und Helena) und her. 20 und 21 (Acontius und Cydippe)’ (Diss., Berlin, 1969), p. 158Google Scholar.

⁵ See Costa, ad loc.

⁶ On the manuscript tradition of the Heroides, see Tarrant, R. J. in Reynolds, L. D. (ed.), Texts and Transmission (Oxford, 1983), pp. 268–73Google Scholar.

⁷ Platnauer, M. (Latin Elegiac Verse, p. 88)Google Scholar notes only five other instances, in none of which is a long open vowel elided.

⁸ CR 11 (1897), 429–30 = Diggle, J. and Goodyear, F. R. D. (eds.), The Classical Papers of A. E. Housman (3 vols., Cambridge, 1973), pp. 418–19Google Scholar.

⁹ Mus. Helv. 42 (1985), 57.

¹⁰ A note on the chronology of the conjectures of Palmer and Housman is perhaps in order. Palmer first published his conjecture in 1893 (CR 7 (1893), 101), Housman his in 1897, although Palmer was familiar with it some time before 1894, as he refers to it in the apparatus to Heroides 20 in Postgate's Corpus, published in that year. Housman presumably knew of Palmer's conjecture when he published his own, although he makes no reference to it. In his review of Palmer's edition (CR 12 (1898), 175 = Classical Papers, p. 475), he quotes lines 179–80 in a list of what he terms ‘conjectures intrinsically bad’, although his chief dispute is with Palmer's interpretation of ‘si…nee… / et’, and of Palmer's ‘certa salutis eris’ he says simply ‘So Palmer writes and so Planudes seems to read’. Housman's failure to discuss Palmer's conjecture more fully is both surprising and disappointing.

¹¹ In the absence of P, the fourteenth-century Greek translation of Planudes is an important witness. Palmer includes Planudes' translation in his edition of the Heroides and discusses its merits in his introduction.

¹² I am grateful to Professor E. J. Kenney for this suggestion.

¹³ I am indebted to Mr A. S. Hollis for this suggestion.

¹⁴ Cf. also Rem. 671, Amores 1.11.19; see further Kühner-Stegmann, vol. 2.2, 228–9.

¹⁵ For instances in Ovid where the MSS support qui against quis, cf. Fasti 3.791, Heroides 1.105, Met. 11.279 (there are six others). Only in Met. 1.248 does the manuscript evidence appear to favour quis, and perhaps here qui the reading of the recentiores, should be preferred.

¹⁶ We are faced with a similar dilemma regarding readings in Heroides 19.205–6:

‘si tibi non parcis, dilectae parce puellae,

quae numquam nisi te sospite sospes ero.’

Most MSS have erit, but editors prefer ero; I wonder whether erit might not be better.

¹⁷ The text of Heroides 21.15–146 is dependent on two fifteenth-century MSS. and three early printed editions. Our sole authority for lines 147–250 is the 1477 Parma edition, π: see Dörrie, H., ‘Untersuchungen zur Überlieferungs-geschichte von Ovids Epistulae Heroidum’, Nachr. d. Akad. Göttingen, Phil. Hist. Kl. (1960), 365ff., 377ffGoogle Scholar.

¹⁸ HSCP 74 (1970), 181–2.

¹⁹ The usage of parallel indicative clauses in place of alternative conditional clauses is perfectly regular in poetry, cf. Amores 2.2.19–22, ‘scripta leget secum; matrem misisse putato; / venerit ignotus: postmodo notus erit; / ibit ad affectam, quae non languebit, amicam: / visat, iudiciis aegra sit ilia tuis.’

²⁰ For the easy confusion of crinibus and cruribus in MSS, cf. Fasti 5.37.

²¹ See also the references in Ph. Bruneau, , Recherches sur les cultes de Délos à l'époque hellénistique (Paris, 1970), p. 547Google Scholar.

²² Although it is not known for certain what audience Plutarch is addressing, it is possible, given his connections with Rome, that he is writing for an educated Roman, as well as Greek, audience.

²³ Evidence regarding Roman attitudes to anklets is more limited, although some scholars believe that they were seen as vulgar and tasteless, e.g. Balsdon, J. P. V. D. (Roman Women p. 263)Google Scholar: ‘Jewels were worn…in the hair and – though not of course by respectable married matrons – in anklets’; cf. also Lowe, W. D. on Longus 1.5, Barini, C., Ornatus Muliebris (Turin, 1958), pp. 105–6Google Scholar. These views reflect modern rather than ancient prejudices, I think. The evidence cited in support of this view consists of two passages, Hor. Epist. 1.17.55–6 (‘nota refert meretricis acumina, saepe catellam / saepe periscelidem raptam sibi flentis’) and Petr. 67: ‘venit ergo galbino succincta cingillo, ita ut infra cerasina appareret tunica et periscelides tortae phaecasiaeque inauratae…eo deinde perventum est, ut Fortunata armillas suas crassissimis detraheret lacertis…ultimo etiam periscelides resolvit et reticulum aureum, quem ex obrussa esse dicebat.' The value of this evidence is dubious: if one is to condemn anklets on the basis of the passage from Petronius, then one must likewise condemn hair-nets and bracelets, which is obviously ridiculous. It is not anklets per se that are vulgar, but rather it is that Fortunata's own particular anklets (like all her jewellery) are vulgar, and the fact that she takes them off in public. We simply lack the evidence to make conclusive statements about the attitude of Romans to anklets.

²⁴ Cf. Met. 10.344: ‘osculaque admoveam’, where see Bömer.

²⁵ Cf. Fasti 4.851: ‘osculaque applicuit posito suprema feretro’.

²⁶ applicat is ascribed to Housman in Palmer's edition, although Housman does not appear to have published it in his own right; possibly he suggested it to Palmer in private correspondence.

²⁷ Wurzburger Jahrbucher für die Altertums-wissenschaft N.F. II (1976), 234–5Google Scholar.

²⁸ Erbse quotes Tib. 1.4.53–4; Ovid, Heroides 15.44, and Ars Am. 1.667, none of which provides a very good parallel.

²⁹ Cf. also Ter. Eun. 383–4: ‘quae nos nostramque adulescentiam / habent despicatam.’

³⁰ Cf. also Amores 2.3.15, Met. 8.262; see further TLL 6.3.2444.19ff.

³¹ The interpretation of ‘offensam habet’ offered here removes the clear statement that Cydippe is actually angry with the rival, and thereby eliminates the objection which has been made on grounds of sense to Diggle's conjecture (based on Bentley) in line 207: ‘tu mihi siqua foret, tu nostra iustius ira’. Cydippe is not actually angry with the rival, he only thinks she is – if she were really angry, then Acontius would be the more deserving focus of her anger.

³² CQ 29 (1979), 421.

³³ ‘cetera cura tua est’, cf. Call. fr. 75.40–1: λοιπόν, Ἀκοντίε, σεῖο μετελθεῖν / …ηνιδιην ⋯ς Διονυσιάδα.

³⁴ HSCP 74 (1970), 183.

³⁵ CP 70 (1975), 224.

³⁸ The one objection to this interpretation is that the ablative hoc is juxtaposed with the comparative plus but not construed with it, and this is rather awkward writing for Ovid.