¹ So, in Propertius, O. Skutsch, CP LVIII (1963), 238 f., A. Woolley, BICS XIV (1967), 80 ff., W. Nethercut, AJP LXXXIX (1968), 449 ff., Juhnke, H., Hermes XCIX (1971), 91 ff.Google Scholar, G. Wille, IBK XXII (Festschrift für Robert Muth, 1983), 598 ff., et al. But not even Homer has escaped.

² So e.g. Williams, G., Tradition and Originality in Roman Poetry (1968), 177Google Scholar, id., The Third Book of Horace's Odes (1969), 23: ‘Horace … collected the first six so-called ‘Roman Odes' at the beginning …’; Nisbet, R. G. M. and Hubbard, M., A Commentary on Horace, Odes Book II (1978), 6Google Scholar.

³ See Housman, , Letters, 197, 15 June 1922Google Scholar. It is notable, however, that until he knows for certain which poems are to be included, he wishes all to be printed purely according to metre, and only later transposed into the quite different arrangement which he will then be able to devise.

⁴ The contrasts prevent us from regarding the two as a single poem. On the other hand, it is most improbable that two originally unconnected poems were adapted to make a pair. Such a compromise is perhaps possible, though never preferable, elsewhere; it would not materially affect the case of this paper. My concern is not with biography but with intentions as manifested in the books: it matters less whether those intentions operated in the production of a first or of a final version of each poem. But, in this particular area, the importance of intentions of some kind is not be denied.

⁵ See, in particular, E. Burck, WS LXXIX (1966), 405 ff. He still gives more weight to numerical structures than I should care to; so too does Becker, C., Hermes LXXXIII (1955), 314Google Scholar ff., on the Eclogues.

⁶ A gallant attempt is made by Brooks Otis, HSCP LXX (1965), 1 ff.

⁷ ‘Minimal literary relevance’, Williams, The Third Book of Horace's Odes, 23; ‘trivialities’, Nisbet, and Hubbard, , A Commentary on Horace, Odes Book I (1970), XXIIIGoogle Scholar (in their second volume they are less forthright).

⁸ See e.g. the first paragraph of the Cambridge History of Classical Literature II (1982)Google Scholar: ‘In some respects … the literary life of Greece and Rome retained the characteristics of an oral culture … nearly all the books discussed in this history were written to be listened to.’

⁹ In the second edition of the Clouds, for readers, Aristophanes exceeds the limits of the stage: thus at the end the school is burnt down, and there are five speakers (contra the edition of K. J. Dover, pp. 266 f.).

¹⁰ Contrast Williams, , Tradition and Originality, 177Google Scholar.

¹¹ Ed. Fraenkel, , Kleine Beiträge II, 199Google Scholar ff.

¹² With Ribbeck, I divide II. 1 into ia and ib (47–78).

¹³ ‘Sed nos immensum spatiis confecimus aequor,/et iam tempus equum fumantia solvere colla.’

¹⁴ O. Skutsch defends nothing but the opening sed (HSCP LXXIX (1975), 230 f.). For this, the ἀλλἀ at Tyrt. fr. II. 1 and fr. 10. 15 would be a most doubtful parallel, even if fr. II (transmitted by Stobaeus) were known to be a complete poem, and if fr. 10 (despite Lycurgus) were known to be two. I suggest that fr. 10. 13 f. are interpolated.

¹⁵ e.g. A. Dieterich, Kl. Schr., 190 f., Grimal, P., Les Intentions de Properce et la composition du livre IV des «Elégies» (Coll. Latomus XII, 1953), 46,Google Scholar E. Burck, WS LXXIX (1966), 408, Macleod, C. W., Papers of the Liverpool Latin Seminar I (1976), 146Google Scholar.

¹⁶ W. A. Camps classifies 3, 5, 7, and 8 more tentatively: ‘a miscellany, but all in varying degrees related to the love theme which was the poet's earlier preoccupation’ (p. 3 of his edition). This smooth formulation in part conceals, in part ignores, the problems that are raised by poem I; so, I fear, does Camps's whole account of the book.

¹⁷ Hubbard, M., Propertius (1974), 116 f.Google Scholar

¹⁸ It is doubtful whether dies in 1. 69 (‘sacra diesque canam et cognomina prisca locorum’) gives evidence of an intention which would have been executed in the projected Aetia. dies takes its sense from the context, and denotes festi dies: it is almost a synonym for sacra. The aetiological poems do not in fact explain directly any festivals of the state (although the festival of the Bona Dea is dwelt on in 9). But the phrase, in its context, gives an adequate notion of the type of poem in question. Programmatic statements need not be rigorously precise if this suits the dramatic or rhetorical purpose of the writer (cf. Russell, D. A., Mnemosyne, Ser. IV, XXXIV (1981), 72CrossRefGoogle Scholar ff.). Propertius wishes to stress his patriotism and his proximity to Callimachus.

¹⁹ In the Rome of old ‘nulli cura fuit externos quaerere divos’ (IV. 1. 17).

²⁰ In Propers, Wege der Forschung CCXXXVII (1975), 163Google Scholar.

²¹ See Maxfield, V. A., The Military Decorations of the Roman Army (1981), 85Google Scholar.

²² See Syme, R., History in Ovid (1978), 187 f.Google Scholar

²³ Fam. XIV. 1–4 = 6–9 SB. Note especially the passion of XIV. 4. I (cf. Prop. IV. 3. 6), 2. 2; the tears of 3. I (cf. Prop. IV. 3. 4), and elsewhere; Terentia's piety in 4. I (cf. Prop. IV. 3. 57).

²⁴ These lines will be inspired by Tib. II. 5, as has long been recognized.

²⁵ putat recc. P's pudet, which Fedeli accepts, is clearly inappropriate. One might, however, have expected quis in place of non: might non have come from nomen above it?

²⁶ See especially 9. 3 f. (cf. I. 3 f.), 5 f., 19 f.; 4. 9–14, 73–8 (cf. I. 20 f.—note also Tarpeius pater I. 7); 10. 18 (cf. I. 10), 25 f., 27–30 (cf. I. 34 f. ?); 2. 59 f. (cf. I. 21, 5; in 2. 60 one should perhaps read for grata the parva suggested by a friend of Camps).

²⁷ See Platner-Ashby, , Topogr. Dict., 203, 269Google Scholar.

²⁸ See Plut., Rom. 17, Lloyd-Jones, H. and Parsons, P. J., Supplementum Hellenisticum (1983), no. 724CrossRefGoogle Scholar; RE IV A 2331 ff., Hubbard, M., Propertius, 119 f.Google Scholar

²⁹ WS LXXIX (1966), 422 f.; the contrast ha d been noted by Celentano, L., Univ. di Napoli, Annali della fac. di lett. e filos. V (1956), 55Google Scholar.

³⁰ Dieterich, Kl. Schr., 191, etc.

³¹ With Heinsius's avis, generally and rightly accepted. The ‘honoratis … aquis’ of the MSS is incredible.

³² The most important discussion of the poem is Macleod's, loc. cit. (n. 15), 141 ff.

³³ We have long forgotten the hospes of the opening, who, as has often been realized, is simply a bold extension of the device used in epigrams on objects seen (cf. Alcaeus, A. P. IX 588 = HE 106 ff., Theocr., Ep. 17 Gow). Catullus 4 exploits the same device. The first section of our poem does not read like a tour.

³⁴ Camps sees the reproof of Horos in I. 71 f as applying only to 87 f., which he places with Marcilius after 68, and which he interprets as a proposal to write ‘poetic prophecies’ (p. 62 of his edition). It would seem that these prophecies are actually to be uttered by other people (so that ‘dicere fata’ in 71 has an unexpected sense); and in any case the second line of the couplet cannot refer to them. I find it a very artificial notion that Horos can be understood to censure one part (one line) only of the programme, and to leave the rest intact. Besides, the tone of the couplet is quite out of place in the section 57–70—indeed, there is no niche for the lines anywhere in the poem.

³⁵ ‘at tu finge elegos’ (135) shows that epic is in question. I do not agree with Macleod that ‘the poem Propertius risked writing would have included, as lines 45–8 indicate, bellicose material’; and in fact poems 6 and 10 do treat of war.

³⁶ Macleod sufficiently refutes the contention that I. 71–150 form a new poem (F. H. Sandbach, CQ N.s. XII (1962), 264 ff.). If they did, the problems they raise would not be sensibly diminished.

³⁷ Anderson, R. D., Parsons, P. J., Nisbet, R. G. M., JRS LXIX (1979), 125Google Scholar ff.

³⁸ See, for example, Williams, , Tradition and Originality, 177 f.Google Scholar He allows that the interrelation of the two poems must constitute an exception to what ‘it is reasonable to assume’ about most Roman poems.

³⁹ This reticence is startling in a sphragis which opens as this one does. It is notable also that the poet has not named himself hitherto. His name appears in the other books eight times, and Catullus, Tibullus, Sulpicia, and Ovid in his Amores, allow or compel their own names to enter. Rothstein (in his first edition) and Butler and Barber suppose that the disclosure of the name might be left to the title of the book; but the excuse is palpably unconvincing.

⁴⁰ A friend quoted by Camps thinks that the poet means by the nymphs ordinary Roman girls. Propertius gives us no reason to think this, and lines 7–10 (especially 7 and 10) are scarcely to be reconciled with this hypothesis.

⁴¹ This picture is not found in Theocritus 13, Propertius' model. The Boreadae have a very different role in the story at A. R. I. 1300 ff.

⁴² Again, there is nothing of this in Theocritus or Apollonius.

⁴³ Compare 20. 7–10 with Ecl. 8. 6 f., 20. 36 with Ecl. 8. 37, 20. 45 with Ecl. 5. 59-and possibly 20. 49 f. with Ecl. 6. 43 f.

⁴⁴ Among the other effects in the book one may include the surprises of discovering that Cynthia will remain in Rome (8b), that Gallus is deeply in love (10), that a poem in this book is being spoken by an ancient door (16).

⁴⁵ The sins described in 64. 399 ff. are all at home in Catullus' own day: compare with 399 Lucr. III. 72 (‘sanguine civili’ 70), contrast Hes., Op. 184; compare with 401 f. the allegations at Cic., Clu. 27 f. and Sail., Cat. 15, contrast Hes., Op. 182; compare with 403 f. the allegation in Cat. 88, 89, and 90.

⁴⁶ It would require a curious view of the poet to suppose that he had already written 21 before he contemplated collecting his poems, but could not bear to leave it out, and therefore wrote 22 to explain it for the general public (cf. Williams, loc. cit. (n. 38)). On 22 the article of Leo, F., Ausgewählte Kl. Schr. II, 169 ff.Google Scholar, is still worth consulting, for it brings out effectively the strangeness of the poem. At his conclusion—that we have only the beginning of the poem—one shudders.

⁴⁷ Syme, , History in Ovid, 101 f.Google Scholar On the Aelii Galli see also Nisbet and Hubbard II, 223 f. The grounds for relating Propertius to the Aelii Galli are that the Postumus of III. 12, who is married to an Aelia Galla, may be the Propertius Postumus of ILS 914 (Rome). To the innocent outsider these grounds seem highly speculative.

⁴⁸ See Syme, , History in Ovid, 98Google Scholar (Ov., Tr. IV. 10. 47).

⁴⁹ Lyne, R. O. A. M., The Latin Love Poets (1980), 112 ff.Google Scholar

⁵⁰ cf. ZPE XLI (1981), 39.

⁵¹ Compare Pease in his edition of the De Divinatione I, 17, 20. On Quintus' career and writings, see Bailey, D. R. Shackleton, Cicero: Epistulae ad Quintum Fratrem et M. Brutum (1980), 3 ff.Google Scholar

⁵² The contention that it is two, made popular again by Skutsch (HSCP LXXIX (1975), 228 ff.), rests on shaky foundations. The mention of ‘tres libelli’ in II. 13. 25 may be met by the hypothesis—which could be supported on other grounds—that Books II and III were published together. The absence of Book I from the grammarians need have no significance in the light of the nature, and the paucity, of their citations from Propertius. The De Natura Deorum and the Ad Herennium are quoted a similar number of times by the grammarians, and their first books do not appear. For the nature of the quotations from Propertius, cf. E. P. Menes, CP LXXVIII (1983), 136 ff. It may also be noted that Book II. Ia–9 (supposedly the original Book I) is quoted less often than Book in or the original Book II: Skutsch illegitimately heaps together the figures for what on his hypothesis is two books.

⁵³ I am grateful to Mr. Jasper Griffin, Professor Hugh Lloyd-Jones, and Mr. R. B. Rutherford for their encouragement, to the Editorial Committee for their comments and advice.