^{page 320 note 1} For useful surveys with ample bibliography see Giomini, R., ‘Ricerce sulle due edizioni degli “Amores”’, Atti del Convegno Internazionale Ovidiano i (1959), 125 f.,Google Scholar and d'Elia, S., ‘Il problemo cronologico degli Amores’, Ovidiana: Recherches sur Ovide, ed. Herescu, N. (1958), 210 f.,Google Scholar to which add Luck, G., Die römische Liebeselegie (1961), at pp. 166 f.Google Scholar (a section missing from the 1959 English edition, Latin Love Elegy). The most recent editor, Lenz, F. W., Ovid: die Liebeselegien (1965), at p. 164, refrains from express- ing an opinion.Google Scholar

^{page 320 note 2} Wilkinson, L. P., Ovid Recalled (1955), 47.Google Scholar

^{page 321 note 1} The banality of this puto is surely deliberate: ‘I would have felt it, I suppose…’.

^{page 321 note 2} This sudden switch in Ovid's train of thought is much better brought out by Kenney's parenthesis after amore than by Munari's full stop or Lcnz's semi-colon.

^{page 321 note 3} Well characterized by Wilkinson, , op. cit., p. 49: see also below, p. 326.Google Scholar

^{page 322 note 1} TAPhA lxxvi (1945), 191 f.Google Scholar

^{page 322 note 2} Op. cit. (n. 1), 184 f. So also Harder and Marg in their translation of the Amores (second edition, 1962),Google Scholar and cf. too Wilamowitz, , Hellenistische Dichtung i (1924), 239 n. 1.Google Scholar

^{page 322 note 3} Wilkinson, , for example, op. cit., p. 52Google Scholar note, dismisses Oliver's arguments with a facile ‘ingenious rather than convincing’. They are accepted in toto by Thibault, J. C., The Mystery of Ovid's Exile (1964), 47.Google Scholar

^{page 323 note 1} On this point see most recently Sullivan, J. P., TAPhA xcii (1961), 522 f.Google Scholar L. Herrmann's identification of Corinna with Propertius' Lycinna (Atti [n. 1] ii, 307 f.) hardly requires to be taken seriously—not least because Herrmann has missed the obvious point that Lycinna too is presumably a metrically equivalent pseudonym.Google Scholar

^{page 323 note 2} In Critical Essays on Roman Literature: Elegy and Lyric, ed. Sullivan, J. P. (1962), 117;Google Scholar and see particularly now Brooks Otis in Haw. Studies lxx (1965), 1–44, especially PP. 29 f.Google Scholar

^{page 323 note 3} Well brought out by Brooks Otis, , TAPhA lxix (1938), 197 f.Google Scholar

^{page 324 note 1} On this interpretation see Richter, W., Gött. Gel. Anz. ccxv (1963), 187 f.Google Scholar

^{page 324 note 2} Cf. Butler and Barber's edition, pp. xxxi–ii.

^{page 324 note 3} As taken for granted by Munari, , third edition (1959), p. x n. 5.Google Scholar

^{page 324 note 4} Butler and Barber, p. xxv.

^{page 324 note 5} The same is also true of all the multibook poems of Claudian, as I shall be showing in a study of Claudian to be published shortly by the Clarendon Press.

^{page 324 note 6} Berl. Phil. Woch. xxxiii (1913), 1228–9.Google Scholar On the ‘hoc quoque’ see Fraenkel, Ed.Horace (1957), 137 n. 2.Google Scholar

^{page 324 note 7} Op. cit. 193 n. 13: cf. also Port, W., Philologus lxxxi (1926), 450.Google Scholar

^{page 325 note 1} See especially Otis, , op. cit. 197 f.Google Scholar Luck (p. 185) strangely claims that 3. 12 (as interpreted by him) is ‘Mass einzige Beispiel, in den der ältere Zeitgenosse bewusst parodiert wird’.

^{page 325 note 2} Op. cit. ch. ii.

^{page 325 note 3} In addition to Luck's own very sensible discussion, see Butler and Barber, pp. xxxvi f., and Day's, A. A.Origin of the Latin Love Elegy (1938).Google Scholar

^{page 325 note 4} In addition to the points discussed above, note the double -que in Errorque Furorque at line 35, normally confined to epic—and (of course) mock—epic (cf. Fraenkel, Ed., Elem. Plautini [1960,] pp. 199 f.). Obviously a deliberate touch.Google Scholar

^{page 326 note 1} Butler and Barber, p. xxvi.

^{page 326 note 2} Wheeler, A. L., AJPh xlvi (1925), 11 f.Google Scholar

^{page 326 note 3} See in general Williams, Gordon, JRS lii (1962), 28 f.Google Scholar

^{page 326 note 4} It so happens—and perhaps just as well for Ovid—that Augustus was absent from Rome between 22 and 18 B.C. (cf. Jones, A. H. M., Studies in Roman Government and Law [1960], 12).Google Scholar

^{page 327 note 1} Kenney, E. J., Ovidiana 208.Google Scholar

^{page 327 note 2} See Ullman, B. L., AJPh (1912), 162 f;Google ScholarOtis, B., ‘Horace and the Elegists’, TAPhA lxxvi (1945), 185.Google Scholar

^{page 329 note 1} See particularly Wifstrand, A., Studien zur griech. Anthologie (1926), 5–30:Google Scholar brief summary in Gow, and Page, , Greek Anthology: Hellenistic Epigrams i (1965), xviii.Google Scholar In the course of an important and very comprehensive review of Gow, and Page, in Gött. Gel. Anz. ccxix (1967), 23–61,Google Scholar Luck revives the now generally discredited view of an alphbetical arrangement for Meleager's Garland (the system apparently followed by Philip of Thessalonica for his). The few examples he adduces could never be enough to disprove the wholly convincing illustration of the other pattern by Wifstrand and Radinger, and can in any case be adequately accounted for by the simple consideration (cf. also Gow and Page, i. xviii n. 3) that Meleager juxtaposed poems on the same theme (often imitations of each other), thus greatly increasing the probability that consecutive epigrams would open with the same word or words. It may be that when confronted with this situation Meleager occasionally arranged them alphabetically, but only as one among the many other patterns that went to make up the skilful interweaving of his Garland (hence, of course, the name). Meleager's two consecutive addresses to Dawn (Anth. Pal. 5. 172–3)Google Scholar open (because this time the girl is sleeping with someone else). Alphabetical order; but this is not why Meleager so arranged them, nor, surely, would he have reversed the order had been interchanged. The traces of alphabetical order which Luck detects in Berl. Klass. Texte v. 1. 75 are hardly impressive.Google Scholar

^{page 329 note 2} Mattsson, A., Untersuchungen zur Epigrammsammlung des Agathias (Diss., Lund, 1942), ch. i.Google Scholar

^{page 329 note 3} Cf. Kroll, W., Studien zum Verständnis der römischen Literatur (1924), 228,Google ScholarBarwick, K., Philologus lxxxvii (1933), 63 f.Google Scholar

^{page 330 note 1} See Fraenkel, , Horace (1957), 112,Google Scholar and Skutsch, , Class. Phil. lviii (1963), 239, n. 6.Google Scholar I would follow Kenney (OCT p. x) in obtaining round numbers for the three books of the Am. by athetizing 3. 55 and dividing 2. 9 and 3. 11 (after L. Mueller) into two separate poems. Those who hold to the authenticity of 3. 5 (against MS. evidence as well as stylistic considerations) will have to reject the division of 3. n (so Lenz, pp. 18 and 233).

^{page 330 note 2} It should be noted, however, that Port (op. cit., p. 452, n. 219) draws precisely the opposite conclusion from the same phenomena.

^{page 330 note 3} So too does Luck, p. 196, though without going so far as to suggest that Ovid deliberately split such pairs between books for a special effect.

^{page 331 note 1} There seems little justification for the common view that the poem might equally be alluding to the final defeat of the Sugambri 8 B.C. (so, most recently, Lenz, , op. cit. [p. 320. n. 1], 183). The parallels from Propertius and Horace are very much against this.Google Scholar

^{page 331 note 2} H. Fränkel even goes so far as to calculate Ovid's rate of composition over this period (Ovid: a poet between two worlds [1945], 194). He estimates that ed. I will have contained about 85 poems since ed. 2 contains 50, and thus that Ovid will have ‘produced at the slow rate of 12 acceptable elegies a year’. But (a) it is hard to credit a facile composer like Ovid with such a remarkably slow rate, and (b) it is implausible to postulate such a steady uniform rate over a period of several years.Google Scholar

^{page 331 note 3} Goold, G. P. (Harv. Studies lxix [1965], 14) is surely correct to write Iason here (and elsewhere in Ovid) for Kenney's anomalous Iaso. See Goold's valuable discussion at pp. 9–14 for some very sensible remarks on orthography in general.Google Scholar

^{page 332 note 1} So, most recently, Lenz, , op. cit., p. 208,Google Scholar and Luck, p. 230, n. 5, despite the sensible remarks of d'Elia, with much of which I am in agreement. It is not relevant that ‘artes tu perlege nostras’ at Rem. 487 does refer to the Ars. For Rem. was indisputably written after Ars. If we knew on other grounds that Am. ii. 18Google Scholar was later than Ars, then we should have to construe it as referring to Ars also. The question is, if Am. ii. 18Google Scholar was written before Ars, could 1. 19 refer to the Amores ? To which the answer is surely ‘yes’.

^{page 332 note 2} Op. cit. 215: for another argument in favour of putting 2. 18 in the first edition, see Giomini, , Atti i. 132 f.Google Scholar

^{page 333 note 1} In the second edition this poem appears as I. 14, though it must presumably have occurred in one of the last books of the first edition. I may perhaps be deemed too fanciful if I suggest that this is one of the poems Ovid transferred from one book to another to make up a round number and even pattern in each book. Obviously it is most unlikely that Ovid simply discarded two whole books in toto. No doubt in general it will have been the earlier books he pruned most drastically, leaving the later books (written when he was nearly 30) comparatively intact. Thus it is natural that he should have transferred the odd late poem to an early book to even out matters.

I am grateful to Professors F. R. D. Goodyear, S. Commager and O. Skutsch and to Mr. E. J. Kenney for comments on a draft of this paper. Naturally they must not be taken to endorse my views.