^{page 60 note 2} The theme of the poem as a whole is closely paralleled by Anth. Pal. 5. 275 written by Paulus Silentiarius in the age of Justinian I. Scholars have disagreed as to whether Paulus is there imitating Propertius or whether both have a common source now lost in Hellenistic poetry, but general opinion seems to favour the second of these alternatives, and I am sure this is right: the differences between Paulus' and Propertius' poems seem to me most easily explained if we realise that Paulus is fairly accurately reproducing the Hellenistic motif whilst Propertius has introduced changes to suit his far more subtle purpose. 1. 3 bears thematic similarity also to two other Propertian poems (2. 29 A and 2. 29 B) and the likelihood that the basic motif is Hellenistic must be very great. Lines 31–3 of the present poem afford a self-contained and interesting illustration of how Propertius uses, but transforms, a borrowed idea.

^{page 60 note 3} ‘Sunt qui Propertium malint’ in: Critical Essays on Roman Literature; Elegy and Lyric, ed. Sullivan, J. P. (Routledge, 1962), pp. 107–48.Google Scholar

^{page 61 note 1} An obvious parallel suggests itself here: the moon wakes Cynthia from sleep, while Cynthia wakes Propertius from his dream. It may have been intended.

^{page 61 note 2} If A. W. Allen, for example, had properly grasped this, he would not have gained his impression that at the end of the poem Cynthia ‘returns in the reader's consciousness to a realm of timeless being and permanent reality, which she shares with the heroines’ (op. cit. p. 133).

^{page 61 note 3} See below p. 76.

^{page 62 note 1} Cf. statements of fidelity like 1. 19.

^{page 62 note 2} Cf. 1. 8A, 1. 11, 1. 12, 1. 15.

^{page 62 note 3} Cf. below p. 76.

^{page 62 note 4} Cf. 2. 6. 42 semper amica mihi, semper et uxor eris (and all 2. 6), 2. 16. 22, etc. The convention seems typically Propertian, but cf. too Tib. 1. 3. 83 ff. and K. F. Smith's note ad loc.

^{page 63 note 1} It is possible, but unlikely, that arma is supposed actually to indicate the offensive weapon of amatory war. The method of expression in these lines seems to me to be too oblique for such an allusion to be in place. On this see Edwards, M. W., T.A.P.A. XCII (1961), 138 Google Scholar.

^{page 63 note 2} Shackleton Bailey is certainly right to object to Enk's equation of tempto here with subigito. At the stage of the context in line 15 there is nothing to suggest such an explicit interpretation of the verb. Indeed the generality and vagueness of it seem particularly chosen to provide the progression and climax outlined above.

^{page 64 note 1} We would thus be seeing the poem through ‘Tychoist’ eyes. This is a more convincing method of approach than supposing, as Shackleton Bailey suggests, that Propertius could have ‘forgotten’ what he wrote in line 8 by the time he got to line 24.

^{page 64 note 2} This last alternative is I think, far from impossible. Parallels could be adduced for vague uses of manus, and the sort of position I am envisaging is something like the Vatican Ariadne's. Cf. Silentiarius, Paulus A.P. 5. 275. 2 Google Scholar περὶ κροτάφους πῆχυν ἑλιξαμένη.

^{page 64 note 3} C.Q. XLIII (1949), 23.Google Scholar

^{page 64 note 4} See further below, p. 72.

^{page 65 note 1} Only Butler in the Loeb translation, so far as I have seen, has taken it in this way.

^{page 65 note 2} This goes against the general opinion of the commentators.

^{page 65 note 3} Shackleton Bailey for one favours the emendation. Catull. 76. 9 omnia quae ingratae perierunt credita menti is always quoted by its supporters; the opening unit omniaque ingrato does however occur at Prop. 1. 17. 4.

^{page 65 note 4} This is the first time in surviving literature that sleep appears winged: see Cahen ad loc.

^{page 66 note 1} I like Curran, L. C.'s suggestion (Y.C.S. XIX (1966), 193)Google Scholar that the list of the first six lines might recall the Hesiodic Catalogue of Women.

^{page 66 note 2} There is a good discussion of Propertius' use of mythology in A. W. Allen, op. cit. pp. 130 ff.; we may contrast the practice of Tibullus. Cf. too Ovid's decorative, simple use of a comparison to three mythological heroines, which he explains, at Am. 1. 7. 13 ff.; note also Lee, A. G., ‘Tenerorum Lusor Amorum’ (in Critical Essays on Roman Literature: Elegy and Lyric, ed. Sullivan, J. P. (Routledge, 1962), pp. 149–79), pp. 154 ff.Google Scholar

^{page 66 note 3} See e.g. Rothstein on lines 1, 3, and 5.

^{page 66 note 4} It is possible that Propertius was the first Latin writer to use the forms Theseus, Cepheius, and Edonis.

^{page 66 note 5} See L. C. Curran, op. cit. pp. 192, 193 and also below p. 68.

^{page 67 note 1} Cf. M.W. Edwards, op. cit. pp. 140, 141. On p. 141, Edwards also acutely talks of ‘the suggestion of exile that lies behind the conventional metonymy Cnosia’; it seems brought out by the position of the metonymy in between desertis and litoribus.

^{page 67 note 2} Cf. for example L. C. Curran, op. cit. p. 197: ‘by using this word here [accubuit] he boldly fuses the moment of Perseus' discovery of Andromeda with the consummation of their marriage, ignoring the time Perseus had to spend in dealing with Andromeda's suitors and kinsmen’.

^{page 67 note 3} As commentators point out, this force seems to be at least suggested in the other Propertian uses of the word (though I am not really convinced for 4. 4. 67); but of course since this is the first occurrence of the word in Propertius, we do not have to allow for the active associations of the other instances.

^{page 67 note 4} We might compare the position of the Vatican Ariadne.

^{page 67 note 5} See Luck, G., The Latin Love Elegy (Methuen, 2nd ed. 1969), p. 122.Google Scholar

^{page 67 note 6} Propertius places his Bacchante in an unusual locale: see Rothstein on line 5. The almost pastoral-sounding scene seems to me to have been specially chosen for its gentle, attractive sound.

^{page 68 note 1} We can see how Propertius' conscious mind hedged this implication by blurring the role of Theseus: it was just the Thesea carina that left.

^{page 68 note 2} Thus the discrepancies between Cynthia's and Andromeda's situation, which have worried some commentators, are intentional and significant on a subtle level.

^{page 68 note 3} One aspect of Cynthia that the poet is not trying to stress, I think, is the depth of her sleep—though this has been the notion of several commentators. Perhaps the three figures of 1–6 were deep asleep, but Propertius does not stress this, as he does other features. In fact, as we learn from 31 ff., Cynthia herself is not presumably very fast asleep. Contrast too mollem … quietem (7), which, though not referring itself to light sleep, certainly does not seem designed to suggest great depth.

^{page 68 note 4} The emphasis of what I say about line 8 may be contrasted with L. C. Curran's remarks, op. cit. p. 195.

^{page 69 note 1} Cf. Prop. 2. 29. 40.

^{page 69 note 2} With ebria, there is a theoretical possibility of ‘transient ambiguity’ —we might take ebria momentarily as referring to Cynthia; and some scholars have done so: see M. W. Edwards, op. cit. p. 132. But I cannot see what possible point the poet could intend by this; indeed it seems to me ruinous to the tenor of the poem so far established. It is of course a fallacy that a poet will always exploit syntactical ambiguity.

^{page 69 note 3} The state of preservation of pre-Propertian literature makes what immediately follows subject to the obvious provisos.

^{page 69 note 4} Cf. Ov., Rem. 378 Google Scholar … iambus ∣ seu celer, extremum seu trahat ille pedem (the image of course being of someone limping).

^{page 70 note 1} The expression of this is noteworthy, including the archaic nondum etiam (see Tränkle, , Die Sprachkunst des Properz … (Wiesbaden, 1960), p. 46),Google Scholar and again a ‘poetic’ acc. graec. construction. Contrast the more simply expressed Lucr. 3. 527, which is possibly in his mind here.

^{page 70 note 2} On the interpretation of lines 15–16, see above p. 63.

^{page 70 note 3} Cf. above p. 64.

^{page 70 note 4} The force of the pluperfect has been missed: cf. e.g. Camps ad loc.: ‘Propertius often makes the pluperfect do duty for preterite or imperfect.’ The tense has I think here another function as well as that outlined above. For Propertius to say ‘I had not disturbed her’ seems to imply that something subsequently did. And of course we learn that that was the case (31 ff.). The pluperfect is thus implicitly anticipatory, and adds a structural cohesion to the whole.

^{page 71 note 1} For other examples of lexical ambiguity in Propertius, see Edwards, pp. 133 ff.

^{page 71 note 2} For similar effects see Edwards, pp. 139 ff.

^{page 71 note 3} It is worth mentioning that the topicality and vividness of this story was probably greater for a contemporary reader, since he would presumably have been familiar with Calvus' Io. Very possibly, if we possessed Calvus' depiction of Argus, the point of Propertius' comparison, which seems generally to have been at the most only vaguely grasped, would be clearer.

^{page 71 note 4} Rothstein's suggestion ad loc. that Argus has an erotic interest in Io is without foundation in any evidence, and misses the poetic point of the comparison.

^{page 71 note 5} See Roscher ii. 270 ff. The wording of line 20 suggests I think this interpretation, and it is clearly more suitable, since some connection between Io and Cynthia must be being envisaged. Propertius however is not consistent in his depiction of the transformed Io: see L. C. Curran, op. cit. p. 201 n. 17.

^{page 71 note 6} Cf. Camps's note on line 20. Other interpretations of ignotis (cf. e.g. Enk ad loc.) miss the point Propertius is trying to make in this comparison. Efforts to emend ignotis betray even more lack of comprehension of the force of the poetry.

^{page 71 note 7} The form Inachis occurs here for the first time in surviving Latin, and appears in Greek for the first time only at Moschus, Europa 44; see Buhler's note ad loc. ( Bühler, W., Die Europa des Moschos, Hermes Einzelschrift 13, Wiesbaden 1960).Google Scholar We note that Propertius maintains the Greek genitive.

^{page 72 note 1} All Propertius' reactions in the poem fit in well with the dramatic setting in which he is supposed to have drunk too much: he is now getting maudlin.

^{page 72 note 2} For discussion on the interpretation of these lines, see above p. 64.

^{page 72 note 3} See Enk's commentary on line 24.

^{page 72 note 4} Cf. Catull. 65. 19 missum sponsi furtivo munere malum, 68. 145 furtiva munuscula.

^{page 73 note 1} Cf. Nape, Ovid's in Am. 1. 11. 1 fGoogle Scholar. colligere incertos et in ordine ponere crines ∣ docta.

^{page 73 note 2} The closest parallel I can find is Tac. Agr. 39 quorum habitus et crines in captivorum speciem formarentur, which is rather different.

^{page 73 note 3} The compression in fact is such that I am not sure that the full sense duxti suspiria is entirely clear. It is normally, and probably correctly, taken to mean ‘give forth sighs’—i.e. on the exhaled breath: cf. Ov., Met. I. 656 Google Scholar alto tantum suspiria ducis ∣ pectore, 10. 402 suspiria duxit ab imo ∣ pectore (Ovid seems to have Propertius' phrase in mind), Verg., Aen. 2. 288 Google Scholar gemitus imo de pectore ducens, etc., but we will note the force of duco in these phrases is brought out by the indication of the ‘source whence’. And in view of the phrase aera sim. ducere (e.g. Cic., N.D. 2. 6 Google Scholar) meaning ‘breathe (in)’, I wonder if Propertius is not rather thinking of the drawing in of a deep audible breath; cf. Rothstein's paraphrase ‘ … wird der ruhige Schlaf doch durch einen besonders tiefen und schnellen Atemzug unterbrochen’.

^{page 74 note 1} Note e.g. the more colloquial verb for ‘bear’ portare: cf. Tränkle, op. cit. 6 f.; see, also, his remarks on this present line at 10 f., though I am not sure he is making a valid point.

^{page 74 note 2} Enk wonders whether the sense of levibus is ‘tenuibus et minimas etiam rimas penetrantibus’ or ‘blandis’, ‘mitibus’. Clearly, I think, in view of what is said above, the second is the important notion.

^{page 74 note 3} The use of praecurro is apparently original: see the commentators. For the interpretation of diversas, the difficulty of which seems to have been caused by Propertius' desire for an out-of-the-ordinary word, see conveniently Camps ad loc.

^{page 75 note 1} See Shackleton Bailey ad loc. For lumina of ‘rays’, cf. Lucr. 2. 162.

^{page 75 note 2} Propertius uses the phrase patefecit ocellos again at 2. 15.7.

^{page 75 note 3} There is perhaps an echo with radiis patefecit of Enn. Ann. 558 inde patefecit radiis rota Candida caelum (cf. too Ov., Met. 9. 795 Google Scholar postera lux radiis latum patefecerat orbem). If such an echo were struck in the mind of the reader, it would enhance the tone of elevation I think these lines have.

^{page 75 note 4} The nearest parallel that I can find for the use figo is Apul., Met. 10. 17 Google Scholar et primum me quidem mensam accumbere suffixo cubito … perdocuit, where the context seems to explain the rather mechanical verb. We may contrast more natural phrases like Verg., Aen. 4. 690 Google Scholar cubitoque adnixa.

^{page 75 note 5} I have mentioned at the beginning how we are meant to appreciate the lack of substance behind her accusations.

^{page 76 note 1} nox here has the association of a night of love, as often in the amatory poets.

^{page 76 note 2} Cf. Enk ‘mihi debitae’, Camps ‘the night that should have been spent with me’.

^{page 76 note 3} Cf. Rothstein ad loc.

^{page 76 note 4} Cf. above p. 62.

^{page 76 note 5} Instances of this force of languidus are cited by Rothstein ad loc.

^{page 76 note 6} Thus, simply, fessa; many commentators see it as meaning ‘when weary with spinning’.

^{page 76 note 7} It is seen as a variant for the conventional Threicius: see Enk ad loc.

^{page 76 note 8} saepe should be taken syntactically with moras: see e.g. Rothstein and Shackleton Bailey ad loc.

^{page 77 note 1} Cf. above p. 66.

^{page 77 note 2} Cf. above p. 67.

^{page 77 note 3} We might compare the varying effects that can be obtained by judicious use of the familiar ‘you’ form in modern European languages.

^{page 77 note 4} leviter has caused trouble with those (including some later MSS) who apparently do not understand Cynthia's abrupt change of tone and approach in 41, when she switches from bitter robust sarcasm to wheedling self-pity.

^{page 77 note 5} When we remember that this leviter itself was adapted to the mollitia attributed previously by Propertius to the idealised Cynthia (cf. above p. 64) and how Cynthia was woken, ironically, by mere levibus radiis (cf. p. 74), it seems clear that the use of mollis, levis, etc. throughout the poem is a planned idea, the notion being exploited at different times for different effects. See also below.

^{page 77 note 6} See above p. 68.

^{page 77 note 7} See above p. 75 for a discussion of the force of figo here.

^{page 77 note 8} And yet the suggestion to read nixa in 34 still receives attention.

^{page 77 note 9} We are of course reminded of the Homeric εὐνῇ ἐνὶ μαλακῇ but this is a neat example of how we should beware of assuming Propertius ever uses an epithet purely decoratively.

^{page 78 note 1} We notice how mollis and torus are juxtaposed in both lines 12 and 34, but in different syntactical relationships. In this way the transference that has been made is underlined. (In the ideal portion of the poem it is the people who have mollis applied to them; in 34 this is dramatically no longer the case.)