^{page 128 note 2} Op. cit. 479–95.

^{page 128 note 3} See, for example, Butler-Barber, , The Elegies of Propertius (Oxford, 1933), xxv ff.Google Scholar; Camps, W. A., Propertius Elegies Book I (Cambridge, 1961), 6 f.Google Scholar

^{page 128 note 4} Kenney, , CR xix (1969), 298Google Scholar; Henry, , CPh lxvi (1971), 198Google Scholar; Courtney, , Phoenix xxiv (1970), 50.Google Scholar

^{page 128 note 5} The datable poems with their approximate dates are: i. 6, 30 B.c.; ii. 7, 28 B.c.; ii. 10, 25 B.c.; ii. 31, 28 B.c.; ii. 34, 26 B.c.; iii. 18, 23 B.c.; iv. 6, 16 B.c.; iv. 11, 16 B.c. For the historical allusions see Butler-Barber, loc. cit.

^{page 128 note 6} The percentages of polysyllabic pentameter endings in the different books are: i, 36·5; ii, 10·7; iii, 2·2; | iv, 1·3. The figures quoted by Williams (op. cit. 482) need slight correction; cf. Platnauer, M., Latin Elegiac Verse (Cambridge, 1951), 17.Google Scholar

^{page 129 note 1} Op. cit. 482 f.: ‘Propertius, when he collected his poems into books, kept more or less to the chronological order in which they had been composed.’

^{page 129 note 2} Williams, (op. cit. 481)Google Scholar lays some stress on the phrase ‘si tres sint pompa libelli’ (Prop. ii. 13. 25) as evidence for a preconceived plan, but this argument does not seem compelling.

^{page 129 note 3} Op. cit. 495.

^{page 129 note 4} See above, p. 128 n. 5.

^{page 130 note 1} See Cameron, A., CQ xviii (1968), 333.Google Scholar

^{page 130 note 2} For a discussion of the points raised in this paragraph see Wickham, E. C., Horace (Oxford, 1874), 19 ff.Google Scholar; Gow, J., Q. Horati Flacci Carmina (Cambridge, 1896), xvi ff.Google Scholar; Nisbet-Hubbard, , A Commentary on Horace: Odes Book I (Oxford, 1970), xxxv ff.Google Scholar

^{page 130 note 3} Horace (Oxford, 1972), 21.Google Scholar

^{page 130 note 4} The question of the ‘publication’ of short poems in the ancient world is notoriously difficult. It seems reasonable to distinguish four stages: (1) oral recitation, (2) private circulation of individual poems in writing, (3) private circulation of collected poems, (4) circulation of collected poems through booksellers. The relative importance of bookshops in Augustan Rome is uncertain; Horace, at least, did not intend his poems to be on sale there (Serm. i. 4. 71).Google Scholar

^{page 130 note 5} For example, i. 12, i. 24, and i. 29 appear to be later poems, while iii. 6 is relatively early.

^{page 130 note 6} For example, Book I begins with a parade of nine odes in different metres, Book II with a long series of alternating sapphics and alcaics, and Book III with six great political odes.

^{page 130 note 7} Wickham's discussion of the principles of arrangement (loc. cit.) is not very conclusive.

^{page 130 note 8} ‘Qui modo Nasonis fueramus quinque libelli, / tres sumus: hoc illi praetulit auctor opus.’

^{page 131 note 1} Thus iii. 9 is earlier than i. 14, iii. 1 is earlier than ii. 18.

^{page 131 note 2} Examples are i. 3 and ii. 4, i. 4 and ii. 5, i. 5 and iii. 7, i. 6 and ii. 19.

^{page 131 note 3} For a discussion of this point see Cameron, , op. cit. (above, p. 130 n. 1), 328 ff.Google Scholar

^{page 131 note 4} In this he may well have been influenced by the example of Propertius i–iii, as Williams, argues (op. cit. 514 ff.).Google Scholar

^{page 131 note 5} This is, in essence, Williams, 's conclusion (op. cit. 495).Google Scholar

^{page 132 note 1} The phrases are taken from Camps, W. A., Propertius Elegies Book IV (Cambridge, 1965), iGoogle Scholar, and Butler-Barber, , op. cit. xii f.Google Scholar

^{page 132 note 2} Cf. Williams, , op. cit. 482Google Scholar: ‘The assumption of a direct relationship between actual experience and poetry in Propertius is certainly false.’ Williams has some less dogmatic remarks on the whole question of the relationship between poetry and experience in his Horace (above, p. 130 n. 3), 1 f.

^{page 132 note 3} The word first occurs in the title of Martial xiv. 189; see Williams, , Tradition and Originality, 483.Google Scholar

^{page 132 note 4} Skutsch, , CPh lviii (1963), 238.Google Scholar

^{page 133 note 1} The correspondences are fully worked out by Otis, B., HSCPh lxx (1965), 1 ff.Google Scholar

^{page 134 note 1} See Courtney, E., Phoenix xxii (1968), 250 ff.CrossRefGoogle Scholar

^{page 134 note 2} See Skutsch, O., HSCPh lxxiii (1969), 153 ff.Google Scholar The total numbers of lines in the balancing pairs of Eclogues are: Ecl. i + ix, 150; Ecl. ii + viii, 181; Ecl. iii + vii, 181; Ecl. iv + vi, 149.

^{page 134 note 3} The manuscripts of Propertius are notoriously erratic in their division of the elegies. The traditional number of elegies in Book II, as fixed by the early editors, is 34.

^{page 134 note 4} Williams, (op. cit. 494)Google Scholar talks of ‘a complexity achieved by carefully chosen juxtapositions of elegies’, but it is difficult to discover any consistent plan for the book.

^{page 135 note 1} The discerning of patterns speedily becomes a subjective exercise. These fifteen poems seem to be organized as follows: 1, 2–3, 4, 5–6, 7, 8–9, 10, 11–12–13A, 13B, 14–15 (that is, a series of pairs or groups linked by a series of single poems).

^{page 135 note 2} For an attempt to find such an arrangement here see Woolley, A., BICS xiv (1967), 80 ff.Google Scholar

^{page 135 note 3} Cf. Baker, R. J., AJPh xc (1969), 333 ff.Google Scholar, and Courtney, E., Phoenix xxiv (1970), 48 ff.CrossRefGoogle Scholar

^{page 136 note 1} iii. 22. 42.

:Compare iii. 21. 1 with i. 6. 13, and iii. 24. 11 with i. 1. 27.

:Compare iii. 18. 3–4 with i. 11. 2–4, and iii. 18. 7–8 with i. 11. 29–30.

^{page 137 note 1} For example, in ii. 31 and iii. 4.

^{page 137 note 2} Op. cit. 493.

^{page 137 note 3} On iii. 20 see my forthcoming article in Mnemosyne.