¹ See Friedlander, P., ‘Pattern of Sound and Atomistic Theory in Lucretius’, AJP 62(1941), 16–34;Google ScholarSnyder, J. M., Puns and Poetry in Lucretius′ De Rerum Natura(Amsterdam, 1980).Google Scholar

² Rouse, W. H. D. and Smith, M. F., edd., Lucretius, De Rerum Natura, LoebClassical Library (Cambridge, Mass, and London, rev. 1975).Google ScholarOther examples offered, besides 3.860 itself, are 1.651, 3.262 and 5.299. An investigation of how many other Lucretian tmeses are in some way appropriate to the sense might do worse than to start with the unanalysed catalogue in Munro, H. A. J., ed., T. Lucreti Cari De Rerum Natura Libri Sex(London, 4th ed. 1886), commentary on 1.452.Google Scholar

³ See TLLand OLDs.v. gregofor the belated entry of the simple verb into Latin (first at Statius. Ach.1.373) as a back-formation from aggregoand congrego.

⁴ For the relationship between se(also written sed)and sine,see Paulus, Fest.p. 336 M. sed pro sine inveniuntur posuisse antiqui;for a modern discussion, Ernout-Meillet s.v. sed, se, so.A few examples of seas an independent preposition are found‘ fossilized’ in archaic legal formulae (thus, in the XII Tables, sefraude);otherwise it is wholly superseded by sine,and survives only as a prefix.

⁵ These ‘other’, confused interpreters of physics, against whom Lucretius directs his polemic in 1.464ff. (and also, implicitly, in 1.462 nee...fatendumst),are identified byBailey, C., ed., Titi Lucreti Cari De Rerum Natura Libri Sex(Oxford, 1947),Google Scholarcommentary on 1.459–63, 464–82, as Stoics;Furley, D. J. in BICS 13(1966), 13–14, rejecting this identification, inclines to the view that the poet's target here need not be specific.Google Scholar

⁶ I would tentatively suggest that Lucretius underlines the lesson by actually including in his sentence a quiet pointer towards the ‘correct’ interpretation of theseinsegregari. sehere is an archaic equivalent of sine(n. 4 above): so is it mere coincidence that just above this sein our texts of Lucretius we encounter... the word sineconiunctum est id quod nusquam sinepermitiali discidio potis est seiungi sequegregari 1.451–2 Suggestive placement of related verbal elements in the same line, or in two successive lines, is a standard feature of Lucretian etymologising: see Snyder, op. cit. 67–8 with 90–108. An example at 2.344–5 is noted by me in The Metamorphosis of Persephone: Ovid and the Self-conscious Muse(Cambridge, 1987), 37.

⁷ For discussion of the ligna/ignespassage, see Snyder, op. cit., chapter 2; on the diversity of Lucretian word-play, see especially her chapter 5.

⁸ This short paper owes its origin to Robert Wardy, whose ‘DRNcontra Anaxagoram: Lucretius on what Atoms are not’ (seen by me in typescript; forthcoming in Classical Philology)is to my mind the most stimulating and suggestive article written on Lucretius for years. I am grateful to him, as also to Catherine Atherton, Glenn Most and Patricia Rosenmeyer, for especially helpful comments on an earlier draft.