1. An earlier version of this paper was presented to the University of Wales Classical Colloquium at Gregynog Hall in May 1977. I am grateful to my colleagues in the University for their comments on that occasion, and especially to Dr. Stephen Mitchell for his subsequent constructive criticism.

2. Nothing is known of Graecinus and little can be speculated, save that he was probably an intimate friend of Ovid's and the same Graecinus as the addressee of Ex Ponto 1.6, 2.6, and 4.9.

3. The Latin Love Elegy, 2nd edn. (London, 1969), p. 170.Google Scholar

4. For further discussion of the nature and objectives of literary criticism with special reference to classical writers see Wilkinson, L. P., Presidential Address to the Classical Association, 1972, esp. pp. 6–7Google Scholar, and my own remarks in Bulletin of the Council of University Classical Departments, 1978.Google Scholar

5. See Bonner, S. F., Education in Ancient Rome (London, 1977), esp. Ch. 16.Google Scholar

6. See Bonner, , op. cit., pp. 213–14Google Scholar; Kroll, W., Studien zum Verständnis der Römischen Literatur (Stuttgart, 1924), Ch. VIIGoogle Scholar; Williams, G., The Nature of Roman Poetry (Oxford, 1970)Google Scholar, Ch. IV (a clearer account than the expanded version in the same author's Tradition and Originality in Roman Poetry).

7. e.g. Polystratus, , Anth. Pal.Google Scholar 12.91; Philodemus, , Anth. Pal.Google Scholar 12.173; anonymous epigrams, Anth. Pal. 12.88 and 89 (for the dating see Gow-Page, , Hellenistic Epigrams (Cambridge, 1965), Vol. ii, pp. 559–60).Google Scholar

8. e.g. Strato, , Anth. Pal.Google Scholar 12.246; Silentiarius, Paulus, Anth. Pal. 5.232Google Scholar; Aristaenetus, , Epist. 2.11 (see below, n. 12).Google Scholar

9. I follow Housman, and Barber, (Sexti Properti Carmina, 2nd edn. (Oxford, 1960))Google Scholar in seeking to omit 11–12. Camps, W. A. (Propertius. Elegies Book II (Cambridge, 1967), p. 155)Google Scholar wishes to place the couplet after line 24. For other suggestions see Enk, P. J., Sex. Propertii Elegiarum Liber II (Leiden, 1962), pp. 285–6.Google Scholar

10. The Greek name could either conceal the identity of a real person or indicate an imaginary character; cf. Lynceus (Prop. 2.34.9).

11. Cf. Ovid, , Am.Google Scholar 2.4, which expands into a whole poem the single topic of attraction to an unlimited number of female types.

12. The similarity between the two poems has long been recognized: see especially Neumann, R., Qua ratione Ouidius in Amoribus scribendis Propertii Elegiis usus sit (Diss. Göttingen, 1919) pp. 79–85Google Scholar; cf. now DuQuesnay, I. M. le M. in Ovid, ed. Binns, J. W. (London, 1973), pp. 21–2Google Scholar; Morgan, Kathleen, Ovid's Art of Imitation: Propertius in the Amores, Mnem., Suppl. 47 (1977), pp. 50–2, 54–6Google Scholar. Mrs. Morgan does DuQuesnay an injustice in suggesting (op. cit., p. 3) that he fails to consider Ovid's reasons for ‘borrowing’ extensively from Propertius (see e.g. DuQuesnay, , op. cit., p. 21Google Scholar: ‘Ovid has given a new twist to Propertius' suggestion that two lovers are better than one …’), and she does Neumann an even greater injustice in claiming (op. cit., p. 50, n. 12) that ‘in Am. 2.10 he regards the primary influence as a letter of Aristaenetus (II.11) and Prop. 2.22A is a secondary influence.’ Neumann knew very well (as Mrs. Morgan seems not to know) that Aristaenetus wrote in the fifth century A.d., and all he is attempting to do is to refute, with justice, the view which sees a common source for the ‘dual love’ theme as used by both Ovid and Aristaenetus in a lost corpus of Alexandrian love elegy.

13. Anth. Pal. 12.173.4 οủκ οἶδ' οὐκ οδ' ἥν εὶпεῶ δεῖμε пοεινοτέρην ‘I do not know which I should say is the more desirable.’ Cf. Anon. Anth. Pal. 12.87.3–4.

14. Anth. Pal. 12.157.3 χεαίνει δὲ βαρὺς пνεύσας Пόος ‘I am storm-tossed by Desire's strong gale.’ Cf. id. Anth. Pal. 12.167.3; Aristaenetus, , Epist.Google Scholar 2.11; Ovid, , Am.Google Scholar 2.4.8.

15. Anth. Pal. 12.88.5 τμήξατ', ἐμοὶ τοῦ' ἡδύ ‘Cut me in two–I should relish that.’

16. Anth. Pal. 12.89.1. Κύпρι, τί μοι τρισσοὺς ἐ' ἕνα σκοпὸν ἤλασας ίούς; ‘Cypris, why have you shot three arrows at this one target?’

17. A name given to Venus because there was a temple of her worship on Mt. Eryx in north-west Sicily; see further Nisbet–Hubbard on Horace Od. 1.2.33.

18. Anth. Pal. 12.88.1–2 Δισσοί με τρύχοσι καταιγίςοντες ἔρωτες, / Εὔμαχε. ‘Two loves bear down on me like a storm and are wearing me away, Eumachos.’

19. The studied symmetry of lines 5–8 is well conveyed by Christopher Marlowe's translation of them (The Elegies of Ovid, c. 1597):

Both are well-favoured, both rich in array, Which is the loveliest, it is hard to say. This seems the fairest, so doth that to me, And this doth please me most, and so doth she.

20. See Otto, A., Die Spricbwörter und sprichwörtlichen Redensarten der Römer (Leipzig, 1890Google Scholar; reprinted Hildesheim, 1962), pp. 321–2.

21. See DuQuesnay, , op. cit. (n. 12), pp. 13–16Google Scholar; Barsby, J. A., Ovid: Amores Book I (Oxford, 1973), pp. 23–9.Google Scholar

22. See esp. Aphthonius ap. Spengel, L., Rhetores Graeci (Leipzig, 1854Google Scholar; reprinted 1966), Vol. ii, pp. 23–7 (see also Bonner, , op. cit. (n. 5), pp. 256–60)Google Scholar, Amores 1.9 may be seen thus: 1–2: the initial statement of the thesis and an immediate repetition of it; 3–30: a series of ‘reasons’ for its validity (τὸ τῆς αὶτίας) incorporating (23–4) a comparative illustration (пαραβολή), 31–2: an interim summing up; 33–40: a catalogue of exempla (пαραδείγματα);41–4: a citation of personal Ovidian, rather than ancient, testimony (μαρτυρία пαλαιῶν) to the truth of the maxim; 45–6: the conclusion (ὲпίλογος).

23. For a useful corrective to this extreme view see Higham, T. F., ‘Ovid and Rhetoric’, Ovidiana (Paris, 1958), pp. 32–48, esp. 47–8.Google Scholar

24. See Hofmann, J. B., Lateinische Umgangssprache, 3rd edn. (Heidelberg, 1958), p. 198.Google Scholar

25. It is just possible that Ovid's manner of introducing his addressee owes something to the rhetorical device of introducing a χρεία by way of a direct address to its originator (see Bonner, , op. cit: (n.5), p. 257)Google Scholar. If so, there is some humour in the poet's elevation of Graecinus' alleged remark to the status of a great man's ‘saying’, only to prove the falsity rather than the truth of it in what follows.

26. e.g. Mimnermus, fr. 1; Luck (loc. cit. (n. 3)) supplies further parallels.

27. For the former cf. Prop. 3.8.20; Ov. Her. 16.219; and for the latter cf. Cat. 68. 5–6; Prop. 4.7.6; Ov. Her. 19.158.

28. Kenney, E. J. (P. Ouidi Nasonis Amores ètc. (Oxford, 1961Google Scholar (corrected edn., 1965)) app. crit., p. 51) suggests that sunt is to be understood ἀпὸ κοινοû in line 23, i.e. graciles sunt, sed non sunt sine uiribus, artus, but I prefer to assume that Ovid first states a general truth, i.e. that slenderness of build does not necessarily indicate physical weakness (23), and then proceeds to claim that the truth holds good in his own case (24). Propertius does the same in reverse order at 2.22A.21–2.

29. The end Ovid envisages for himself was not in fact unknown; cf. Val. Max. 9.12.9. (Brandt, P., P. Ouidi Nasonis Amorum libri tres (Leipzig, 1911Google Scholar; reprinted, Hildesheim, 1963) supplies further references ad loc.). Ovid's sentiment is reminiscent of Propertius' at 1.6.25–8:

me sine, quem semper uoluit fortuna iacere, hanc animam extremae reddere nequitiae. multi longinquo periere in amore libenter, in quorum numero me quoque terra tegat.

Those lines, however, are positively sedate in comparison with Ovid's.

30. Cf. Tib. 1.1.1–6; 1.10.29–40; Prop. 3.5.1–6.

31. Conscious artistry in sentence-structure and word-order is again perceptible in Ovid's treatment of the exempla (31–4): a verb of identical inflection begins and ends each couplet (induat… emat, bibat… quaerat), and each pair of verbs is joined by a co-ordinating et.

32. e.g Tib. 1.1.59–66; Prop. 1.1.17–24, 1.19.

33. Prop. 1.7.24.

34. Cf. Prop. 4.1.64, a boast which Horace seems to mock at Epist. 2.2.100.

35. An expression which owes its inspiration to the range of wrestling imagery commonly used in erotic contexts; see Aristophanes, , Peace 894–9.Google Scholar

36. See Pichon, R., Index Verborum Amatoriorum (Paris 1902Google Scholar; reprinted Hildesheim, 1966) s.v. uoluptas, decipere, lasciuus, mouere, opus.

37. Lee, A. G. (Ovid's Amores (London, 1968))Google Scholar well translates, ‘Pleasure's a food that builds me up.’

38. For utilis cf. Ars. Am. 2.709–10. The single line utilis et forti corpore mane fui (28) does duty for all Propertius's mythological exempla at 2.22A.29–34 (cf. DuQuesnay, , op. cit. (n. 11), pp. 21–2).Google Scholar

39. Trist. 4.10.81–2.

40. For felix see CLE (ed. Buecheler) 394.1–2; 1085.4; for cum moriar, CLE 1237.16.

41. Aen. 6.669.

42. See Tränkle, H., Die Sprachkunst des Properz, Hermes Einzelschriften, 15 (1960), p. 151.Google Scholar

43. 33 instances in Lucretius, 35 in Virgil's Aeneid, c. 40 each in Lucan, Statius, and Seneca (tragedies), 19 in Valerius Flaccus, 101 in Silius Italicus, 35 in Ovid's Metamorphoses; contra, 16 in Ovid's elegiacs, 8 in Horace, 2 in Tibullus, 4 in Propertius.

44. It will be obvious from this discussion that I see in Amores 2.10 only a superficial resemblance to Propertius 2.22A. Kathleen Morgan, however, (op. cit. (n. 12), p. 54) argues for a high degree of structural parallelism between the two poems.

45. See Syme, R., The Roman Revolution (Oxford, 1939), Ch. XXIX.Google Scholar

46. It is perhaps for this reason that neither Lee, A. G. (in Critical Essays on Roman Literature: Elegy and Lyric, ed. Sullivan, J. P. (London, 1962), p. 179, n. 9)Google Scholar nor Barsby, J. A. (op. cit. (n. 21), p. 18Google Scholar, n. 8) includes it in his list of the most successful poems in Book 2.

47. I risk plagiarizing Professor Kenney, who makes a similar claim for Ovid's poems from exile (Proc. Cam. Phil. Soc. N.S. 11 (1965), 49).Google Scholar