### Cyril Bailey (ed.), Titi Lvcreti Cari: De Rerum Natura: Libri Sex, Vol. 2: Commentary, Books I–III

Find Location in text

# pg 794BOOK II

In the First Book Lucr. has laid the foundations of the atomic system and has shown that the universe consists of an infinite number of atoms, small, indivisible, eternal particles, moving in space infinite in extent. In the Second Book he proceeds to consider deductions from these principles, many of which are of great importance for the Books which are to follow. The correspondence between Lucr. and Epicurus' Letter to Herodotus is not here so close as in the First Book, and a number of topics are included which no doubt had their place in the Μεγάλη‎ 'Επιτομή‎, which was probably Lucr.'s chief authority. After a proem on the blessings of philosophy (1–61), which is interesting as containing Lucr.'s most explicit references to the moral theory of Epicurus, he proceeds at once to the motion of the atoms, which is dealt with very briefly in the Letter (§ 43), and in the course of it introduces the famous theory of the 'swerve' of the atoms (216–93). Passing to a new point he shows that the atoms have many different shapes, but not an infinite number, and that the number of atoms of any given shape is infinite (478–580); here he agrees with the Letter, § 42. He deals also with the effects of difference of atomic shape on sensation (381–477) and the variety of their combinations and the effects on the species of things created (581–729). Still treating of the atoms he argues that they are without secondary qualities (730–1022) and in particular that they are devoid of sensation (865–990); this section prepares the way for the account of the soul in Book iii. Finally, leaving the atoms and turning to a quite different topic, which is dealt with in the Letter, § 45, he shows that there is an infinite number of worlds (mundi, κόσμοι‎), and describes the process of their formation and destruction. This section appears to have its place here as a preparation for Book v, and constitutes an argument for the theory that in the original draft of the poem v immediately succeeded ii. It may be noticed that twice in the Book (167–83, 1090–1104) Lucr. digresses to refute the theological view of creation, which he had attacked so severely in the introduction to Book i.

The main contents of the Book may be tabulated as follows, but the argument is complicated and further subdivisions will be found necessary in the commentary:

Proem: The blessings of philosophy: 1–61.

1. A. The Motion of the Atoms: 62–332.

1. (a) Introduction. 62–79.

2. (b) The incessant movement of the atoms and their union in compounds. 80–141.

3. pg 795(c) The velocity of their motion. 142–64.

4. (d) Digression against the theological view. 167–83.

5. (e) The universal motion downwards due to weight. 184–215.

6. (f) The swerve of the atoms. 216–93.

7. (g) Appendixes:

1. 1. The permanence of matter and motion. 294–307.

2. 2. The motion of the atoms within a body at rest. 308–32.

2. B. The Shapes of the Atoms and their Effects: 333–729.

1. (a) The variety of shapes and their effects on sensation. 333–477.

2. (b) The number of shapes and of the atoms of each shape. 478–580.

3. (c) The variety of atomic combinations and the differences within species. 581–729.

3. C. The Atoms are without Secondary Qualities: 730–1022.

1. (a) Colour. 730–841.

2. (b) Other secondary qualities. 842–64.

3. (c) Sensation. 865–990.

4. (d) Summary. 991–1022.

4. D. The Infinite Worlds, their Formation and Destruction: 1023–1174.

1. (а) Introduction. 1023–1147.

2. (b) Proofs. 1048–89.

3. (c) Digression against the theological view. 1090–1104.

4. (d) The growth and decay of worlds. 1105–74.

## Proem: the Blessings of Philosophy. 1–61.

1. It is pleasant to watch from the land the distress of those in a storm at sea; not that anyone's suffering is a source of joy, but because it is pleasant to contemplate evils from which one is free. 5. So too it is good to watch a battle in which you have no part. 7. But the greatest joy is to dwell on the heights of philosophy and to look down on the wanderings and struggles of others, as they strive with one another for wealth and power. 14. How blind men are, how dark and dangerous their little lives! They do not understand that all that human nature demands is that the body should be free from pain and the mind from anxiety and fear.

20. (a) The body needs only what can take away pain and give it enjoyment. Nature has no need of banquets at night to the accompaniment of music in a gorgeous palace, if man can lie on the grass by a stream, resting and eating simple food, especially when the sun shines and flowers are blooming. 34. Nor again does fever depart more quickly from the rich man's couch than from the pg 796poor man's blanket. 37. For the body, then, wealth, birth, and glory are of no profit.

39. (b) Nor are they for the mind, unless when you gaze upon your legions manœuvring in sham fight in the Campus, your superstitions flee and your fears of death depart. 47. But it is not so; care and fear have no respect for the weapons of war, the power of kings, or the display of wealth. 53. Philosophy alone can dispel the darkness which hangs over life. 55. Like children in a dark room we fear things no more to be dreaded than what children foresee, 59 and our terrors must be dispersed not by the light of day but by the knowledge of nature and its workings.

The proem to the Second Book stands rather apart from those prefixed to the other Books. It does not like the proems of i and iv speak of Lucr.'s own work, nor does it like those of i, iii, v, and vi contain a direct eulogy of Epicurus. Its interest and importance lie in the emergence more explicitly than in any other part of the poem of the moral theory; see Prol. IV, §§ 10–14.

The ultimate goal of the Epicurean moral theory was 'pleasure' (ἡδονή‎), but that did not mean the 'kinetic' pleasures of bodily and mental excitement, such as were recommended by Aristippus and the Cyrenaics, but the 'catastematic' pleasure of freedom from pain and care; the 'kinetic' pleasures were to be avoided, because they often produced pain as their consequence.1 The aim then must be a state of equilibrium in which all pain was removed and nothing introduced which could be a source of future pain. And when pain was removed, a 'limit' (πέρας‎) was reached beyond which pleasure could not be increased, but only 'varied' or 'embroidered' (ποικίλλεται‎).2 This idea is developed with regard both to the body and the mind.3 The body must, as far as possible, be relieved from the pains which attack it (ἀπονία‎), and this can be done by simple treatment and remedies, and for its enjoyment it does not require the elaboration (ποικίλματα‎) of luxury, but merely the relief from the 'want' caused by hunger and other desires. Bodily pleasure is best secured by the simple life, and wealth and power can add nothing to its happiness. Similarly what the mind needs is to be freed from its characteristic pains, namely its fears, so that it too can attain to the tranquillity (ἀταραξία‎) which is its equilibrium. The greatest of the fears which beset the mind are those of superstition—the fear of the intervention of the gods in the affairs of this pg 797life and of the punishment of the soul hereafter. These fears, again, cannot be expelled by wealth or power, but only by the study of philosophy, which can reveal the true causes of things and so remove the false fears.

These ideas are clearly set out by Lucr. with his characteristic elaboration and ornament from 16–61. He considers separately body and mind, and in treating of the body deals separately with its positive pleasures (20–33) and the banishment of pain and disease (34–6). In treating of the mind he both shows that wealth and authority cannot exorcize its fears and claims for philosophy the power to do this in words repeated from the First Book.

There remain the introductory lines (1–13) which to almost all readers have an unpleasant taste of egoism and even of cruelty. The Epicurean philosopher, secure in his own independence, gazing on the troubles and struggles of his fellow-men is an almost cynical picture; Bacon referred to it ironically as 'Lucretian pleasure'. Nor can it be wholly defended, for it is true that Epicurus' hedonism was essentially individualistic; the Epicurean must be freed from the pains of body and mind, and it would no doubt enhance his sense of pleasure to observe the contrast in the lives of others. Perhaps the only pleas which could be made in extenuation are that in practice the Epicurean, like the founder himself, showed a large degree of kindness to others (εὐγνωμοσύνη‎), and that it was the aim of Lucr. to make converts, so that as many men as possible might share the Epicurean tranquillity. We may remember too the strong Epicurean injunction against taking part in public affairs: λάθε βιώσας‎ was to be carried out literally.

1. Suave mari magno …: the sentiment was a commonplace in Greek and Latin literature; cf. e.g. Archippus (Kock, fr. 43), ὡς ἡδὺ τὴν θάλατταν ἀπὸ τῆς γῆς ὁρᾶν, ὦ μῆτέρ, ἐστι μὴ πλέοντα μηδαμοῦ‎ and Cic. Ad. Att. ii. 7. 4 cupio istorum naufragia ex terra intueri. Notice that Lucr., as though conscious of its harshness, endeavours to tone it down in 3–4.

suave: the omission of est magnost is unnecessary; cf. ii. 5 suave again, neque enim mirum ii. 87, nec mirum ii. 338, vi. 130, quid mirum v. 1238, and see Prol. V b, § 13 (a).

mari magno: 'on the open sea', a local abl. The epithet magnum is traditional; cf. Liv. Andr. Odyss. 23 arvaque Neptuni et mare magnum; Ennius, Ann. 445 indu mari magno fluctus extollere certant. For the form mari see i. 286 n.

2. mangnum alterius spectare laborem: so Virg. Georg. i. 158 heu magnum alterius frustra spectabis acervum.

laborem 'struggles'; it is not necessary to follow Merrill in giving it the unusual sense of 'danger'.

pg 798

3. non quia … quemquamst: in classical prose non quia and non quo are followed by the subj. when they introduce a false reason, which is also denied as a fact: Lucr. observes this rule with non quo ii. 336=723, non quo multa parum simili sint praedita forma, 692 non quo multa parum communis littera currat, vi. 71 non quo violari summa deum vis possit; non quia occurs only here with indic.

4. careas: the idiomatic 2nd pers. sing. of the subj., 'the reader', 'one'; see i. 327 n. and Prol. V b, § 10 (a). carere is more commonly used of a good thing which is lacking than of a bad thing of which one is free.

5–6. The transposition of these two lines, made by Avancius, has been universally followed. The lines are an anticipation of the idea in 40 ff., except that here it is a real battle, not a sham fight.

5. suave: for the repetition from 1 Bignone compares Theocr. i. 1–7.

6. tua: referring to the idiomatic 2nd pers. in careas 4.

7. sed nil dulcius est …: the idea recurs e.g. in Ciris 14 si me iam summa sapientia pangeret arce … unde hominum errores longe lateque per orbem despicere atque humiles possem contemnere curas and other passages collected in Merrill's note.

bene … serena: note the accumulation of epithets, munita, edita, serena with templa. Most editors follow Munro in maintaining that doctrina sapientum must be taken with munita, when edita becomes an awkward interruption. It is impossible to construe doctrina sapientum with edita, as editus in the sense of 'lofty' appears only to be used absolutely, but Giussani may be right in taking munita and edita together 'built up on high'.

8. templa: not 'high places' (Merrill), but in the vague Lucretian sense 'quarters', 'dwellings' (see i. 120 n.).

serena: certainly with templa and not abl. fem. with doctrina.

9. passim goes in sense with errare and not with videre.

10. viam … vitae: 'a road through life', 'a way of life' such as can only be found in philosophy.

11. certare ingenio …: for the general idea of the life of ambition, condemned by Epicurus, see v. 1129–35 and iii. 995–1002 (Sisyphus), and Epic. Κ‎.Δ‎. vii ἕνδοξοι καὶ περίβλεπτοὶ τινες ἐβουλήθσαν γενέσθαι, τὴν ἐξ ἀνθρώπων ἀσφάλειαν οὔτω νομίζοντες περιποιήσεσθαι‎.

12–13 recur in the prooem to iii (62–3).

12. labore: plabore OQ and in iii. 62 tabore, both clearly mistakes for labore. Diels argues here that the reading of OQ represents paborel, and that the original text was pavore. But against that is the sense and the evidence of iii. 62.

13. rertumque potiri: used of political power. Cf. ii. 50 and iii. 1027 rerumque potentes.

14. For the form of the exclamation cf. v. 1194 o genus infelix humanum.

pg 799

15. The life of one who does not know the true aim of life according to Epicurus is in tenebris because he cannot see the light of truth and in periclis because in consequence his conduct may be wrong.

16. hoc aevi quodcumquest: 'this little span of years'; quodcumque has a diminutive effect.

16–19. nonne videre … metuque: difficult and much vexed lines. Translate with the text given: 'To think that they do not see that human nature cries out for nothing more for itself than that pain may be kept far sundered from the body, and that, withdrawn from care and fear, she may enjoy in mind the sense of pleasure.'

16. nonne videre: for this idiomatic use of an exclamatory infin. with -ne cf. Ter. Andr. 253 tantamne rem tam neglegenter agere? Cic. Cluent. 6. 15 nonne timuisse? and other examples collected by Lachmann. The idiom is common and, as Lachmann saw, is spoiled by videre est of the Itali, which was adopted by some of the earlier editors and by Diels; nor is there anything to be said for the reading of AB, nonne videtis, adopted by Marullus. The subject of videre is a vague plural, sc. homines, derived from hominum in 14, not, as Lachmann and Giussani think, hominem, which it would be difficult to supply.

17. sibi: with latrare 'cries out for itself', not, as Giussani would take it, with videre.

naturam: 'human nature', as in 23, of which corporea natura (20) 'the nature of the body' is a part and is contrasted with animus in 39.

latrare 'to bark for', 'to yearn for'; cf. Paul.–Fest. 121: latrare Ennius pro poscere posuit; Enn. Ann. 584 animus cum pectore latrat. Ennius' phrase is founded on Hom. Od. xx. 13 κραδίη δέ οἱ ἔνδον ὑλάκτει‎, imitated again in Hor. Sat. ii. 2. 18 latrantem stomachum. In view of Festus' explanation Giussani must be wrong in taking it to mean that 'nature barks at man' sc. continually enforces on him. Like Lachmann he does not approve of the restriction of natura in the sense of 'human nature', and for this reason adopts this interpretation and takes sibi with videre, hominem being assumed as subject. For natura in the restrictive sense cf. ii. 944 grandior ictus quam patitur natura (sc. animantis) and ii. 369 quod natura (sc. agnorum) reposcit, and of course many places where there is a limiting genitive such as animi, mentis, etc.

utqui: see i. 755 n. Ernout thinks that here the suffix -qui denotes 'le peu d'exigence de la nature', sc. 'that in some way or other'. This is perhaps fanciful.

18. seiunctus … absit: 'should be parted and absent', a Lucretian redundance.

mente … metuque (19). The reading of OQ can be retained: natura is the subject of fruatur and semota agrees with it. mente stands by itself 'in mind'; cf. Cic. De Fin. iii. 37 quis est qui … nulla animo adjiciatur voluptate? iucundo sensu is governed by fruatur and cura … metuque by semota. The accumulation of pg 800ablatives is awkward and might have been altered in revision, but cf. i. 473, ii. 623. Marullus attempted to mend the construction by reading mensque, which has been accepted by many editors, Lachmann by reading menti', taken with iucundo sensu, 'a pleasant sensation in the mind', a queer expression. Diels, following Bergk, believes that mente stands for mentis and that mentis is nominative (see Addenda). Lachmann dealt drastically with the whole passage and read nisi ut, cui (Avancius) corpore seiunctus dolor absit, menti' fruatur iucundo sensu, cura semotu' (Lambinus) metuque, 'that the man, from whose body pain is parted and absent, may enjoy a pleasant sensation in his mind, freed from care and fear'. He was followed by no one except Bernays. I agree with Munro that in spite of its awkwardness the text of OQ should be retained.

20–33. The general sense of these lines is clear. Pleasure on the Epicurean theory consists in the removal of pain due to want, and when once pain has been removed, pleasure cannot be increased in quantity but only varied in quality: Κ‎.Δ‎. xviii (quoted on p. 796, n. 2). These 'varieties' of food and living Epicurus calls τὰ ἐν διαίτῃ ποικίλματα‎ (Frag. Vat. lxix), which Lucr. here translates delicias. Robin (on l. 23) is right in maintaining that there is also reference here to Epicurus' triple division of desires (Ep. ad Men., § 127, Κ‎.Δ‎. xxix, xxx) into those which are natural and necessary, natural and unnecessary, and unnatural and unnecessary. The satisfaction of the first class is essential for life, for it is that which removes pain (demant … dolorem); on the other hand, desires of the third class which are idle (κεναί‎) must never be satisfied. But ποικίλματα‎ of the second class may be occasionally permitted (Ep. ad Men., § 130), because an occasional indulgence enables us to enjoy more the simple diet on which we ordinarily live. So Epicurus writes to a friend (fr. 39 C. B.) 'send me some preserved cheese, that when I like I may have a feast.'

This doctrine Lucr. has in mind here: but few things are necessary to remove pain (20, 21; cf. Ep. ad Men., § 130 τὸ μὲν φυσικὸν πᾶν εὐπόριστόν ἐστι, τὸ δὲ κενὸν δυσπόριστον‎). At the other extreme all the elaborations of the banquets in great houses must be eschewed, but we may indulge in the simple pleasures (non magnis opibus 31) of a picnic in a pleasant place. The Epicurean doctrine is elaborated with Lucr.'s usual poetic vision.

Generally speaking, the sense is clear and the text is sound, but editors have differed very widely in the punctuation of the lines and in particular as they take 22 with what precedes or with what follows.

With the punctuation in the text I take 22 with what precedes and translate thus: 'We see then that but few things are needed at all for the nature of the body, such as may remove pain, yes and pg 801can also supply many delights; nor does human nature itself from time to time need anything more acceptable, if there are not elaborate banquets, but (cum tamen 29) we can have a frugal meal in a pleasant country place.' The clauses quae demant … and delicias … possint are then parallel, the subject of possint is pauca which is then contrasted with multas, 'a few things can both remove pain and give many pleasures'. Lucr. deliberately allows the use of simple ποικίλματα‎. neque continues the sentence in its normal sense of 'nor', interdum refers to this occasional simple indulgence, and gratius is the object of requirit 'something more delightful' than the simple picnic.

In so doubtful a passage some other views must be considered.

1. Lachmann's view is substantially that taken above, except that he places a full stop after possint, as do Bernays and Diels. I have substituted the semicolon in order to make the use of neque more natural. Giussani also has full stop after possint, but places comma after interdum: 'it is delightful sometimes, and nature does not feel a loss if', etc. This is possible, but gratius interdum is otiose and requirit is left without an object. Ernout places comma at dolorem, and full stop at possint, but regards quae … dolorem as giving the subject of possint and explaining pauca: 'for our bodily nature we can see that we have need of but few things—such as are enough to remove pain—in order that (uti) they may also supply many delights; nature herself asks for nothing more agreeable', etc. This may be right and does not differ substantially from my view, but I think it gets the emphasis wrong. It is the idea that only a little is needed to remove pain which should be most prominent.

2. Munro, uniting 22 with what follows, takes a quite different view. He places full stop at dolorem and comma at interdum, takes uti as concessive in sense, and neque = the simple non, as it is sometimes in early Latin and in Lucr. at iii. 730 and vi. 1214. 'Even though they may be able to supply many delights pleasantly from time to time, nature herself does not feel a lack if', etc. The subject of possint is then the simulacra, etc., anticipated from 24 ff. (or it might be vague sc. homines, cf. 16), and the deliciae of 22 will be not the simple permissible 'delights' but the elaborate and luxurious pleasures of the third class. Lucr. so uses the word in v. 1450, where delicias is explained by carmina, picturas, etc., but this is not decisive for, if deliciae =ποικίλματα‎, it may equally be used for the simpler or the more luxurious 'variations'. Munro's construction is undoubtedly awkward, especially in the concessive uti and the use of neque = non. Brieger follows Munro, except that he places a comma at possint and takes 23 as Giussani does; Mèrrill too, adopting Munro's view in general, regards neque as 'pendens', i.e. anticipating another neque which does not occur.

pg 802

3. Martin places a semicolon after omnino and commas at dolorem and interdum. I presume that he takes quae … dolorem as the subject of possint and neque … requirit as parenthetical.

It is not easy to make up one's mind between these many possibilities, but I feel sure that it is more in accordance with Epicurus' thought to take 22 with what precedes, as in my own view and the others in 1.

20. ergo: a general inference, 'and so'.

corpoream ad naturam: 'to satisfy the bodily nature', a slightly different idea from that which would have been expressed by the more normal dat. The body needs the removal of pain (21) and sickness (34), and simple pleasures occasionally (23).

pauca … esse opus: see i. 1051 n. and Prol. V b, § 2. 1.

21. quae demant cumque: 'just such as may remove'. Lucr. is fond of this use of quicumque in places where a simple relative might have sufficed. Cf. iv. 247, 281, 737, etc.

22. substernere: 'to spread for our use'. Cf. Cato, R.R. 37. 2 eam (segetem) ovibus substernito. Munro notes that it has much the same force as the simple sternere, but that sub has the meaning it has in subministrare, etc.

23. natura: 'human nature', as in 17.

ipsa probably 'human nature by itself', when not stimulated by habits of luxury.

24. aurea … iuvenum simulacra: such as those described by Homer, Od. vii. 100 χρύσειοι δ‎' ἄρα κοῦροι ἐϋδμήτων ἐπὶ βωμῶν ἔστασαν, αἰθομένας δαΐδας μετὰ χερσὶν ἔχοντες, φαίνοντες νύκτας κατὰ δώματα δαιτυμόνεσσι‎; cf. Od. ix. 5 ff. Lucr. no doubt has these passages in mind, but it is probable that wealthy houses in his own day had candelabra of this kind.

25. An unusually close rendering of Homer's words.

igniferas: a typical Lucretian compound, recurring in v. 458 ignifer … ignis, 498 aether ignifer, vi. 379 igniferi fulminis.

27. fulgēt: the e lengthened before a vowel in arsi, and probably with a recollection of its original length (= fulgeit); see Prol. VI, § 12 (b). The only other example in Lucr. is v. 1049 ut scirēt animoque videret (where Lachmann reads scirent). It is more frequent in Virgil, who was doubtless influenced by Greek practice. Macrobius vi. 2. 5, possibly through a semi-conscious objection to the lengthening, quotes the line with fulgens … renidens, sc. est from sunt in 24, which is unlikely to be what Lucr. wrote. Lachmann objected both on the ground of the lengthening and also because Lucr. elsewhere uses forms from fulgĕre, not fulgēre; Brieger adds the further objection that fulget would not be used of the sheen of silver, but only of gold (cf. fulvus). For these reasons they read fulgenti, taking it with auro with a 'postponed' -que. But fulgēre occurs in vi. 213, fulget in v. 768; though the third conj. forms are more frequent in pg 803Lucr.: fulgĕre occurs in v. 1095, vi. 165 and fulgit in vi. 160, 174, 214, 218. As regards Brieger's objection, Lucr. uses the verb with vague words like lumine v. 708, nitore v. 768, both of the moon's light, and in ii. 500 has fulgens purpura and in v. 785 florida fulserunt viridanti prata colore; Cicero, Cat. ii. 3. 5 also writes fulgent purpura. These passages suggest that, though the word was naturally used of a golden or yellow colour, it could have a more extended use. Moreover, with Lachmann's emendation the 'postponed' -que is very awkward and it would be difficult to know that fulgenti did not go with argento. fulget should certainly be retained.

28. citharae must be taken as a dat. 'resound to the lyre', but the construction is very unusual, as Ernout points out. Macrobius twice quotes the line, once (vi. 2. 5) with citharam, once (vi. 4. 21) with cithara, and Lucr. himself has the acc. in iv. 546 reboat … bombum. In Hor. Od. iii. 10. 5 audis quo strepitu … nemus inter pulchra satum tecta remugiat ventis? which is quoted in support of the dat., ventis is in fact abl. The only alternative would be to regard citharae as nom. with reboant trans. 'make the roof resound', but that is unparalleled and it is better to accept the dat. on the analogy of respondere.

laqueata aurataque should be taken together 'with gilded panels'.

aurataque: Lachmann objected to the use of the word so close to auro (27) and emended to ornataque, which is very weak: Bernays 'improved' this to arquataque 'arched'—a word used by Lucr. in iv. 333 in the sense 'with the jaundice'. Lucr. clearly felt no objection to such close repetitions of a word or an idea: cf. 1 and 2 of this book magno … magnum also 48–9 metus … metuunt, i. 719–20 undis and many other instances.

templa: here in the unusual sense of 'rafters', 'cross-beams'; cf. Festus 367 templum significat … tignum quod in aedificio transversum ponitur and so in Vitruvius iv. 2 supra cantherios et templa debet esse conlocatum (see i. 120 n.). Ernout thinks that 'rafters' could not be laqueata and could not 'resound' and prefers to take laqueata … templa as equivalent to laquearia. Macrobius in both passages referred to above quotes the line with tecta: the variation between citharam and cithara shows that he is quoting from memory and tecta may therefore be disregarded.

29–33. Repeated with a few alterations in v. 1392–6 and imitated by Augustan writers.

29. cum tamen: 'although in spite of the absence of all these luxuries'. For the use of cum tamen in concessive sense with indic. cf. i. 565 n., i. 825 (= ii. 690), ii. 859, iii. 107, 645, vi. 140, and iv. 718, v. 16 with subj. with no apparent difference of meaning. See Prol. V b, § 9 (b).

31. non magnis opibus: 'at no great cost'; the Epicurean ποικίλματα‎ must be inexpensive.

corpora curant: cf. v. 939 pg 804curabant corpora; also Virg. Aen. iii. 511 and elsewhere. Servius on Georg. iv. 187 notes that the phrase implies food and washing.

32. tempestas: 'the weather'; cf. i. 178 n.

anni tempora: 'the season of the year'.

33. viridantis floribus herbas: cf. v. 785 florida fulserunt viridanti prata colore.

34. nec calidae citius …: a new point. Not only can simple living give as great pleasures as wealth and luxury, but wealth can do no more than poverty to cure disease. Lucr. returns to the idea of 20 and 21 pauca … quae demant cumque dolorem.

35. textilibus picturis: 'woven pictures', sc. embroideries. Cf. v. 1427 nos nil laedit veste carere purpurea atque auro signisque ingentibus apta and Virg. Aen. iii. 483 fert picturatas auri subtegmine vestes.

36. plebeia … veste: 'under a common coverlet'. Cf. v. 1499 dum plebeia tamen sit (vestis) quae defendere possit, and for vestis in this sense Ov. Met. viii. 658 vilisque vetusque vestis erat, lecto non indignanda saligno.

37 ff. Lucr. now passes to the second point introduced in 16–19, the needs of the soul, and treats it by a continuation of the ideas most recently in his mind. Just as riches are of no avail to cure our body (nostro in corpore), so they can do nothing to dispel the evils of the soul (animo quoque 39), its cares and fears, unless indeed we are prepared to believe that the great commander, looking on his army, is freed from all the anxieties and fears produced by religio. And then comes his own answer (53): no, it is only philosophy which can banish idle fears. The argument is not direct, but is dictated by the line of Lucr.'s thought in the first half of the paragraph.

37. nostro in corpore: 'in man's body': Lucr.'s frequent use of nos and noster in a general sense. Cf. i. 883 nostro quae corpore aluntur.

gazae … nobilitas … gloria regni: pick up the ideas of 11–13.

38. proficiunt: a medical word, used of successful methods of treatment; cf. Plin. N.H. xx. 23, 98 (radix anethi) vel in febribus proficit, Hor. Ep. ii. 2. 149 herba proficiente nihil curarier.

gloria regni: for the idea cf. v. 1129–30.

39. quod superest: 'for the rest', 'on the other side of the picture'; cf. i. 50 n.

animo quoque: opposed to nostro in corpore 37.

putandum: without est: cf. i. 111 n. and see Prol. V b, § 13 (a).

40. si non: 'unless indeed' = nisi.

tuas: W. R. Inge (C.R. liv. 188) suggests duas of two opposing sides, taking the reference as being to Memmius and not to the reader in general; a very improbable correction.

campi: 'the plain', cf. ii. 6 per campos instructa, 323 praeterea magnae legiones cum loca cursu camporum complent. Here at any rate the reference is probably to the Campus Martius, pg 805where such military reviews would be held. campus is used in this special sense in Cic. Ad Fam. vii. 30. 1 in campo certe non fuisti, Hor. Od. iii. 1. 11 descendat in campum petitor, Sat. i. 6. 126 fugio campum, and elsewhere. In 58 b.c. Caesar remained for three months in the Campus before he left for Gaul and was attacked by Memmius for doing so. Possibly the recollection of Caesar's army was in Lucr.'s mind. Bignone has a different view of the scene (see Addenda).

41. Diels expunges this line as a doublet of 43 a. The sense will then run more smoothly, but I think it is more probable that 43 a is a doublet of 41; see n. on 43 a.

fervěre: the obvious correction of l 31 for fruere O, eruere Q. The form fervěre is archaic and recurs in vi. 442; Lucr. also has fervunt in iv. 608, and does not employ forms from the 2nd conj. Lucilius 259 fervěre, ne longum vē: ego hoc lectoribus tradam shows that 3rd conj. infin. was regular in early Latin. For the variation of verbs between 2nd and 3rd conj. see notes on i. 300 tuimur, ii. 27 fulget, iii. 917 torrat, and Prol. V a, § 12. fervere is frequently used of the bustle of a crowd by Virgil and others: e.g. Aen. iv. 407 opere omnis semita fervet.

videas: the idiomatic 2nd pers. subj. 'one sees'; see n. on i. 327 and Prol. V b, § 10 (a).

belli simulacra cientis: as in ii. 324, 'waging a sham fight'; so also Virgil, Aen. v. 585 pugnaeque cient simulacra and 674 belli simulacra ciebat.

42 [43]–43 [42]. These two lines are written in uncials, like the capitula, in O and G; Q omits them and marks a lacuna of 3 lines. In O lines omitted by the first scribe are not infrequently inserted later in uncials by the rubricator who inserted the capitula; e.g. in this book 94, 502, 508, 710, 809, 887, 962, 1012, 1028, 1112. In view of Q's lacuna it looks as though these two lines in the archetype had some mark which caused the scribes to think that they were spurious or corrupt. Both lines are clearly corrupted in the middle and I believe that they should be transposed (see n. on 42 [43]). It is, however, most convenient to deal with their difficulties in the MS. order.

43 [42] epicuri is manifestly corrupt. Lachmann, comparing v. 1228 cum validis pariter legionibus atque elephantis, read magnisque elephantis, but though elephants are appropriate to the invading army of Pyrrhus, which Lucr. seems to have had in mind in v. 1228, they are most unsuitable here at a review in the Campus. Bernays thought that epicuri might be due to a gloss ἐπικούροις‎, explaining subsidiis in the sense of 'reinforcements', and read hastatis; but such a gloss would explain obscurum per obscurius, if indeed subsidiis was obscure at all. Several editors have seen that the natural completion of the line would be the mention of cavalry as well as of reinforcements. Bouterwek's equitatus and Reid's equitum vi were in the right direction as far as sense goes, but would not account pg 806for the corruption. By far the best conjecture is Munro's brilliant et ecum vi: he assumes that it might have been written et ecū vi and then the corruption to epicuri was easy, p and t being constantly confused in these MSS.; cf. tariter O in the next line. The form of the expression is supported by ii. 264 equorum vim cupidam and ii. 326 virum vi. Lucr. does not elsewhere use the gen. plur. ecum (or equom Brieger) and the phrase does read a little like an emendation, but with some hesitation it may be accepted.

42 a [43] itastạṭuas O; itastuas O1; itasiuas G; statuas Itali. The dots under at in O, the regular sign for letters that are to be omitted, were probably copied conscientiously from the archetype. While the corrector of O obeyed what he believed to be their intention, the Itali made the obvious correction statuas. The reading of the archetype might have arisen, as Diels suggests, from isttavas, i.e. istatuas = statuas with a vulgar prothetic i; more probably, I think, statuas was written, the dots below a and t were inadvertent, and ita was then written above to indicate that statuas was right. In the same way non est was inserted to correct dies in Cic. Phil. vi. 1. 3 hodierno autem dies non est pene scio que = hodierno autem die nescio quae; so possibly in ii. 356 non quit = 'it is not quit' (see n. there). There can be little doubt that statuas is the right reading, but it is not easy to construe. Robinson Ellis (J.P. xiv. 90) comparing Shakespeare, Henry IV, Pt. I, Act iv, Sc. i, 'glittering in golden coats, like images', would take it as a substantive: the army seen from a distance looks like a collection of statues; this is most improbable, and it is certainly the verb. Diels by expunging 41 makes it the verb after si non in 40; this would mend the construction, but spoil the sense which must be 'unless, when you see your legions drawn up …, your fears vanish' (44). Munro puts the construction right by reading ornatasque, but with the lines in the MS. order no second verb is needed. I accept his correction, but transpose 42 and 43; the second verb then comes in naturally: 'you see your troops preparing for mock warfare and then you draw them up for fight, supported by the auxiliaries and cavalry'.

The MS. corruptions point so clearly to statuas that other conjectures may be disregarded, as they pay no attention to them, supposing them to arise from a mere stopgap. Lachmann, again following v. 1228, conjectured validas, which is weak. Bernays read pariter, supposing that it had dropped out by haplography; M. J. Boyd (C.R. lii. 119) suggested paribus, comparing Virg. Aen. xii. 344 paribusque ornaverat armis. Both these conjectures would give excellent sense, but they do not account for the MS. readings. Merrill's porro would obviate the necessity for ornatas 〈que〉, but is open to the same objection. For Martin's conjecture stlattas see n. on 43 a.

pg 807

43 a. This line is cited by Nonius (503. 28) as from Lucr. ii. Munro inserts it after 46, where it interrupts the sense less, but, seeing that Q marks the loss of 3 lines here, this, if any, would seem to be the place for it; classem will then be used in the old military sense of 'regiment' and refer to legiones in 40. But it badly interrupts the sequence, and the argument from the 3-line space in Q cannot be pressed; after ii. 841 Q similarly leaves space for 3 lines, corresponding to the 2-line capitulum in O. It is most probably a misquotation of 41 and should not be inserted in the text at all. Martin places the line between 42 and 43, takes classem in its normal sense of 'fleet' and for itastuas in 43 reads stlattas, a word explained by Festus (312 M., p. 411 Lindsay) as meaning 'a broad kind of ship' (= lata) and cited by Martin from Carm. Epigr. iii. 2294. 19 (Lommatzsch). The conjecture is as improbable in Lucr. as Diels's tonguit in ii. 356, nor is it likely that he was contemplating a naval as well as a military review. Moreover, as Boyd points out, stlattae might be ornatae armis, but not pariter animatae. For Bignone's view see Addenda.

44. timefactae: a real participle, 'alarmed', and not, as Merrill takes it, adjectival, 'craven'. Elsewhere it is cited only from Cic. De Off. ii. 24 timefacta libertas.

religiones: 'superstitious fears'. See i. 109 n.

45. effugiunt animo: 'flee from your mind' (cf. iv. 41 Acherunte … effugere), in a different sense from effugere with acc. 'to escape'.

pavidae: O, as a predicative adj. is stronger than pavide Q, which was adopted by Lachmann and most modern editors. Merrill, Diels, and Martin have pavidae.

46. pectus: a necessary correction of Lambinus for tempus OQ and adopted by all editors except Martin, who supports tempus by vi. 1192 ad supremum denique tempus 'your last hour'; I presume that he means here 'leave your time empty', but it is a very odd expression.

47. ludibria: 'trivial'; the substantive is used adjectivally.

48. sequaces: 'following', 'dogging'. For the thought cf. Hor. Od. iii. 1. 40 post equitem sedet atra cura and iii. 16. 17 crescentem sequitur cura pecuniam.

49–50. sonitus … tela picks up 40–3.

reges … potentis corresponds to gloria regni 38. For rerumque potentis see n. on ii. 13 rerumque potiri.

51. fulgorem … ab auro similarly refers to 24–8. ab auro goes closely with the substantive, ab denoting the origin. Cf. Virg. Georg. ii. 243 dulcesque a fontibus undae and many other parallels collected by Munro. In Lucr. himself we may compare i. 719 Ionium glaucis aspergit virus ab undis, though there the construction is assisted by the verb; similarly with e i. 1085 magnasque e montibus undas.

52. vestis splendorem purpureai corresponds to 35. Though Lucr. uses the -ai gen. frequently with substantives, it is noted that pg 808he employs it with only three adjectives, here, iii. 693 gelidai, iv. 537 nigrai.

53. omni' … rationi': the suppressed s twice in one line is unusual. omni' is predicative, 'this is altogether the power of reason'. Cf. i. 377 id falsa totum ratione receptum est, but it is not necessary in either place to follow Munro in regarding the adj. as taking the place of an adv.

54 is a rather clumsy and awkward link of connexion: see note on 55–8.

laboret: 'struggles' as in i. 849 and elsewhere. Note the repetition of tenebris (tenebras) four times in six lines.

55–8 recur in iii. 87–90 and vi. 35–8, being in all three passages followed by the three lines hunc igitur … ratioque taken from i. 146–8. Giussani maintains that they are most in place in vi and that they were first written for that place and then imported by the poet into the third book, where they fit the context fairly well, and lastly here, where 54 was written to patch up an otherwise remote connexion. This may be so, and it should certainly not be assumed that lines repeated were necessarily written first for the place at which they first appear in the poem. See the discussion on i. 921–50. The thought is a commonplace; cf. Plato, Phaedo 77 e ἴσως ἔνι τις καὶ ἐν ἡμῖν παῖς, ὅστις τὰ τοιαῦτα φοβεῖται‎ and Arist. De Insomniis iii. 462 a ἐνίοις γὰρ τῶν νεωτέρων καὶ πάμπαν διαβλέπουσιν, ἐὰν ᾖ σκότος, φαίνεται εἴδωλα πολλὰ κινούμενα, ὥστ‎ʼ ἐγκαλύπτεσθαι πολλάκις φοβουμένους‎. Seneca refers to to these verses in Ep. 110. 6.

55. pueri: 'children' as i. 939.

56. nos in luce: opposed to pueri … in tenebris.

59–61 = i. 146–8; see notes there.

## A. The Motion of the Atoms. 62–332.

### (a) Introduction. 62–79.

62. I will now explain the motions by which the atoms create and dissolve compound things, the force which causes this, and the velocity of their motion through the void. 66. For matter is not a solid mass since we see things diminish in bulk and the lapse of time causes their dissolution; yet all the while the sum of things remains intact. 72. The particles at their departure lessen the things they leave, but increase those they join; some things grow old, others bloom, and the atoms never rest. 75. The sum of things is constantly renewed and living creatures live by mutual exchange. 77. Species wax and wane; in a short space the generations change and hand on the torch of life to others.

In the first five lines Lucr. sets out the syllabus of the first main section of the book (80–332), which deals with the motion of the atoms pg 809and its effects in the production and destruction of compound things. It has three main divisions. With the making and unmaking of things by atomic motion (62–3) he deals in 80–141, and there is reason to think that he may have treated the subject more fully later (see 163 n.). Of the causes of motion, gravitas is dealt with incidentally in 84 and 184–215, and the clinamen 216–93; both of these result in ictūs (85) which are the immediate cause of the formation and dissolution of things. The velocity of atomic movement is treated in 142–63. It must be noted that this is not a complete syllabus of the book, such as is prefixed to the other books, as it makes no mention of the second, third, and fourth main sections on the forms of the atoms and their effects in combination (333–729), on the secondary qualities (730–1022), and on the formation and destruction of worlds (1023–1174). It must be regarded simply as the argument of the first section.

This summary gives Lucr. the chance of linking up the discussion on which he is embarking with his proofs of the existence of the void in i. 329–97. There (335–45) the motion of things was adduced as the first proof of the existence of void. Here the idea is carried one step farther; the destruction of things, which we see going on before our eyes, is a proof of atomic motion, for it is produced by the departure of atoms from compounds (67–70). This leads to another thought: destruction is balanced by creation and that too is due to the motion of the atoms, which when they depart from one thing go to give increase to another; they never stay permanently fixed in any one body (72–5). So in our world we see this exemplified in living creatures; individuals and even races perish and hand on their life one to another (76–9). And coupled with these ideas of the ceaseless atomic motion and the continual creation and destruction which result from it is the further notion that the totality of things does not suffer increase or diminution (71); yet its constituents are ever changing places so that it assumes new forms (75). The sequence of ideas is complicated and one is interwoven with the other. As so often, we see Lucr. here not working on a logical sequence of thought, but on a mental picture, the various elements of which are in his mind at once, and each finds expression as it comes uppermost. (Giussani has an interesting analysis of these lines, though it is in some respects too subtle.)

62. nunc age … expediam: a form of introduction repeated in vi. 495–7 and adopted by Virgil in Georg. iv. 149. Lucr. has nunc age in i. 265, etc., and nunc … expediam in iv. 633–4, vi. 639–41, 680–2, 1090–3.

genitalia … corpora: 'the atoms which are the cause of creation'. See i. 58 n. For genitalia with objective gen. see ii. 548. Notice genitalia, … gignant, genitas to reinforce the idea.

pg 810

63. res varias: 'the different kinds of things' because, as is seen in 100–11, the density and therefore the character of things differs according as the atoms are in close or loose combination, or in other words, according to the amount of void between them.

resolvant: 'undo' the combinations and set the individual atoms free again to join another combination (73). So i. 57 quove eadem rursum natura perempta resolvat. In this sense Lucr. also uses retexere 'to unravel' the atomic mesh in i. 529 and v. 267 and 389.

64. qua vi facere id cogantur: 'by what force they are compelled to act thus', i.e. to move and form and dissolve things. There are really three elements in this vis: gravitas which is the original cause of the downward movement of the atoms in space, the clinamen which is the cause of the meeting of the atoms, and the ictus which results and which by its angle and its severity causes the varieties of atomic formation (see introductory note).

ollis: for the archaic form see i. 672 n.

65. mobilitas: 'rate of motion', a rather strange word probably employed by Lucr. owing to the scansional impossibility of either of the more obvious words velocitas and celeritas.

mobilitas … meandi: 'the rate of motion of their moving' is a characteristic Lucretian expansion. For meare of atomic movement see i. 428, ii. 158, etc.

66. tu … memento: the editors note the dictatorial tone of this instruction to Memmius (or the reader); it is the philosopher speaking to his pupil. Cf. iii. 135 tu cetera percipe dicta, also iv. 931 and vi. 920.

praebere memento: 'remember to show yourself', for the prolate infin. see Prol. V b, § 12 (a).

67–8. non … materies: sc. matter is not absolutely continuous, but is interrupted by void. So at the outset of the discussion on the void i. 329 nec tamen undique corporea stipata tenentur omnia natura. On the other hand (i. 610), the atoms themselves minimis stipata cohaerent partibus arte; there is no void between the minimae partes.

68. quamque videmus: the obvious correction by Q1 for quamquidemus OQ.

69. fluere: 'waste away' with the idea of the atomic flux from things. Cf. ii. 1128 fluere atque recedere corpora rebus, 1139 rarefacta fluendo.

70. subducere: vetustatem is the subject, and the object is an acc. omnia derived from omnia nom. in 69.

ex oculisque: for the postponed -que cf. e terraque i. 187 and n. there.

71. summa … 75 rerum summa: there is considerable difficulty in the use of these terms, though Robin (in Ernout) alone among commentators has perceived it. summa, strictly speaking, should mean 'the universe', the totality of atoms and space, and it alone in fact remains incolumis, because nothing can be added to it pg 811or taken from it. So in v. 361 the summarum summa is contrasted in this respect with mundus (364) and haec rerum summa (368) 'our world'. But Lucr. cannot be thinking of the universe in this sense here because (a) he uses all through words implying our experience or observation, videmus (68), cernimus (69), videatur (71), which means here 'is seen', not 'seems'; (b) novatur in this sense is impossible of the universe. His illustration all through is from our world of experience, which is an aggregate of compound things, mortales, gentes, saecla, etc., but it has a wider application: his mind passes from our world to other worlds and so to the sum of other worlds and the space of the intermundia, in which the free atoms are always moving. In fact once again it is not logic but a picture which is in his mind, that of the innumerable worlds of compound things conceived together as a whole, a 'universe of things'. This may be said to be incolumis because its total content is unchangeable, but it is 'renewed' because the atomic motion is always forming and destroying new compounds within it. See n. on i. 235.

72. cuique: sc. rei, 'each compound thing' The dat. is 'Lucretian' and not unlike genitalia corpora rebus in i. 58.

73. augmine donant = augent just as augmina sumunt ii. 188 = augentur. For augmen in Lucr. see i. 435 n.

74. cogunt: the subject is of course corpora, 'the atoms' (72), and illa and haec are objects.

75. nec remorantur ibi: 'and they do not stay permanently there', i.e. in the concilia which they form, haec and illa.

rerum summa: see n. on 71. It must be the equivalent of summa there, though with a more concrete sense 'the total of things', i.e. of matter (see i. 235 n.), and could not mean either 'our world' or 'a world', because no perishable world semper novatur.

76. mortales: here 'living creatures', not, as it usually is in Latin, 'human beings'. So in ii. 1153, v. 791, 805 mortalia saecla 'races of living creatures'. Lucr. is here narrowing his thought again to experience in this world.

mutua: adverbial, 'live by mutual exchange', in origin an internal acc.; so again fungi mutua: iii. 801, iv. 947, and with other verbs: iv. 301, v. 1100, vi. 1084. The expression here, as Giussani points out, is not quite exact: living creatures receive their life from one set of creatures and hand it over to another.

77. gentes: not, as Merrill takes it, 'nations', but 'races' of animals as well. Lucr. does not mean, as Robin points out, that the numbers of different species increase or decrease absolutely, but, as he explains in ii. 532, they become rare in one part of the world and plentiful in another.

78. saecla: here probably 'generations', as is shown by the thought of the next line: mutantur being 'change with one another'. See i. 20 n.

pg 812

79. The metaphor of the λαμπαδηφορία‎ is not uncommon in classical writers, e.g. Plato, Laws vi. 776 b καθάπερ λαμπάδα τὸν βίον παραδίδοντας ἄλλοις ἐξ ἄλλων‎, Varro, R.R. iii. 16. 9 nunc cursu lampada tibi trado. Whatever it may be in other places, here the idea must be that of a team-race in which the members of a team pass on the torch to one another.

## (b) The incessant movement of the atoms and their union in compounds of varying density. 80–141.

80. To suppose that the atoms can stay still and so produce the motions in things is a mistake. For since they are wandering in the void, they must be propelled either by their own weight or by the blow of another atom. 85. When they have met, they leap apart in different directions, since they are hard and solid and there is nothing behind to stop them. 89. To realize the incessant movement of all of them, you must remember that the universe has no bottom where the atoms might come to rest, but that, as I have shown, it stretches out without limit in every direction. 95. Therefore the first-bodies never come to rest in the infinite void, but move incessantly in every direction. 98. Some after they meet leap apart at great distances, others at short intervals. 100. (1). Those which in closer union separate only at short distances, entangled by their interlocking shapes, make hard things like rock and iron. 105. (2). Others which separate and meet again with long intervals between them make air and the bright light of the sun. 109. (3). And many more wander independently through the void, those which have been rejected from union and are unable to enter a body and link their motions with others.

112. We can see a picture of this atomic motion in a familiar sight. When the rays of the sun penetrate into a dark room, we see many small particles moving together in the light; 118 they seem to be engaging in an unending warfare, assailing one another incessantly in squadrons, driven about by countless meetings and partings. 121. From this comparison we may understand the constant movement of the atoms in the void; a little thing can give us a pattern and a concept of the great. 125. More than this, the motion of the particles in the sunbeams is a proof of secret motions going on beneath the ken of our senses. 129. You can see many of them propelled by unseen blows changing their course and moving back again in every direction; these deflections are due to the movement of the atoms. 133. First the atoms move themselves; then the small nuclei, which come nearest in force to the atoms, are stirred by the secret blows of the atoms and in tum set in motion larger nuclei. 138. So motion passes up the scale from the atoms and gradually becomes perceptible to us, so that even the particles which we can pg 813see in the sunlight are set moving, yet we cannot see the blows which start them.

This first section on the movement of the atoms is of supreme importance, for it lays the foundations of the whole conception of the Epicurean kinetics. Epicurus treated of the subject in the Letter to Herodotus (§§ 43, 44, and again §§ 61, 62), but though Lucr. from time to time reproduces expressions from those sections, he is not here following the Letter closely, and in particular makes no mention of certain subtle problems with which Epicurus deals concerning the relation of the movement of the atoms to the movement of the compound bodies which they compose. The whole question of the Epicurean kinetica has been treated with great ingenuity (though sometimes with over-subtlety) by Giussani in Studi Lucreziani, pp. 97–124, and by Bignone, Epicuro, Appendix, 225 ff. I have tried to put together the main ideas and discuss some of the difficulties in The Greek Atomists and Epicurus, Part II, chap. v, pp. 310–38. Here it is sufficient to attempt to elucidate the argument of Lucr.

The main idea which runs through the whole section is that the atoms, whether 'free' in the void, or enclosed in compound bodies, are always in motion.1 The form which the argument takes is best understood if it is realized that Lucr., following Epicurus, is correcting two mistaken or insufficient views in the traditional atomism of Democritus. Democritus had held2 (1) that the 'free' atoms were 'naturally' always in motion in all directions in the void, and (2) that in a compound body, the component atoms being interlocked with one another suspended their own 'natural' motions and adopted that of the whole compound body.

As regards the first statement Epicurus did not deny it—he fully agreed with it—but he endeavoured to get behind it to the causes of this ceaseless motion in all directions. He held that there were two; firstly, that all atoms fall 'downwards' in the void owing to their own weight, for unlike Democritus, who thought that weight was only the resistance of bulk to the motion of the whirl, Epicurus regarded it as an essential property of the atom. This was the original cause of motion, and the diversity of the motions of the atoms was due to the blows caused by their collisions with one another, which changed their direction, these blows being due ultimately to the 'swerve' (παρέγκλισις‎, clinamen). The result is exactly that described by Democritus, but Epicurus has explained how it arises. This notion Lucr. reproduces shortly in 84–5: 'the pg 814reason why the atoms move in all directions (vagantur) in the void is either their weight which is carrying them downwards or the blows which are moving them in other directions; for when they meet they dash apart in varying directions'. He assumes, then, here the two causes which he is going to discuss later, weight in 184–215 and the swerve 216–93.

Secondly, Epicurus held as against Democritus that the atoms never 'stop' or 'rest' or in any way lose or diminish their original motion. The 'free' atoms in the void are everlastingly moving from one collision to another; the collisions only change the direction of their motion, but do not in any way impede or diminish it. Similarly in a compound body the atoms are still in constant motion in all directions; the only difference is that in the compound the component atoms so combine their motions one with another as to harmonize (consociare … motus 111) and the motion of the whole body is the result of this harmonization. Nor is there any difference in this respect between hard and dense bodies, like rock and iron, and the loosest textures, such as air or light, except that in the former the 'trajects' performed by the atoms between one collision and another are much shorter than in the latter. In every body there is some void in which the atoms move. But though the motion of the atoms is unaffected by their entrance into a compound body, that of the body is affected by their motions. For the movement of the atoms in all directions acts as a retardation to the whole body and causes its movement to be slower; the ἀντικοπαί‎ of the individual atoms are the cause of an internal ἀντικοπή‎ in the whole body, which may retard it just as might its collision with another body outside. Just as the atoms by their union create the thing, so by their motions they determine the motion of the thing.

It is this second series of ideas which Lucr. principally sets himself to explain and illustrate in this paragraph. In the first two lines he enunciates the main theory that the atoms never stay still and that it is by their motions that they create the 'new' motions in compound bodies. After a few lines on the causes of their diversified motion (see above) he comes back to the consequences of their motions and meetings. First (89–99) is the vivid picture of the everlasting movement of the 'free' atoms in the void and their clashes and collisions, with the additional note of the infinity of the void which is the guarantee of the eternity of movement. Then he makes a careful distinction between the three possible results of meeting: some atoms may enter a close compound in which they can only perform short 'trajects' (100–4), some may unite in looser compounds at long intervals and perform long 'trajects' (105–8), while others may be rejected from all compounds because they cannot fit in their movements with the whole, and so remain free atoms moving in the pg 815void (109–12). Certain details and difficulties in the conceptions are dealt with in the notes.

Then follows the ingenious and beautiful image of the motes in the sunbeam which by their diverse movements, resembling the movements of troops in a battle, give us a picture of the motions of the atoms (113–31). And finally he shows that these movements of the motes are not only an illustration, but the outcome of atomic movements. In the smallest nucleus ἀντικοπή‎ causes some retardation of the whole, which increases as the mass of the particle grows larger, until at last a body is formed which is not only large enough, but moves slowly enough to be, like the motes, perceptible to the senses.

Epicurus' kinetics must be recognized as one of the most penetrating and ingenious parts of his theory. Alike in the idea of the ceaseless motion of the atoms and in that of the constant internal atomic vibration in things he anticipated the results of some of the most subtle investigations of modern times. Munro and Masson quote interesting parallels from Clerk Maxwell and other modern physicists. The comparison with the motes is also one of Lucr.'s most careful and finest bits of writing.

80. cessare: 'to be at rest' quiescere, and also 'to stop' sistere. The atoms never were stationary nor have they ever stopped their motion, nor ever will. The two ideas are really the same: cessare, as Giussani puts it, is at once 'star fermo' and 'fermarsi'.

81. cessando … motus 'and by remaining still produce the motions in things' is slightly inexact. The belief of Democritus was that in a compound thing the atoms lost or suspended their own motion and adopted that of the thing; Epicurus' view was that they continued their own motions and by so doing produced in combination the motion of the thing. It would have been easier to express it by saying 'and in producing the motion of things lose their own motion'. How exactly the various motions of the component atoms united to form that of the compound is not clear; Giussani in his essay goes beyond the evidence in endeavouring to explain it.

rerum … motus: 'the motions of things', i.e. of compound bodies, res in its normal Lucretian sense; they are 'new' because they only come into existence with the formation of the compound. Robin finds difficulty in the expression and takes it to mean 'physical' or 'sensible' movements, but I do not grasp his difficulty.

82. For the expression cf. ii. 229 avius a vera longe ratione recedit: also i. 880 and ii. 740; for vera ratio see i. 51 n.

vagaris … vagantur 83; a good example of the close repetition of a word in different senses.

83. per inane vagantur: 'wander', i.e. move in all directions pg 816'through the void'. The phrase is repeated in 105 and 109.

cuncta necessest … alterius: see introductory note. These are not of course alternative causes of motion, for weight alone is the original cause, but alternative causes of motion in various directions. They may either move 'downwards' through gravitas or in some other direction through a blow; ferri then is important 'to be carried along', 'moved in one direction or another'.

84. gravitate: 'their heaviness' βαρύτης‎, a quality, not quite the same as pondus, their actual weight, though ponderibus propriis is used in the same sense in ii. 218. Cf. i. 1076.

85. forte: 'from time to time'. Cf. ii. 630 Phrygias inter si forte catervas ludunt; gravitas as the cause of motion downwards is always active, the blows which drive the atoms in other directions are intermittent.

nam: the logic is a little compressed: we can be sure that the blow is one of the causes, because we know that when atoms meet they move asunder in various directions, whereas gravitas could only cause movement downwards. Giussani has a more elaborate explanation.

⟨cum⟩ cita: OQ are defective; Itali have concita, but a conjunction is needed as well, and Wakefield's cum (quom) has universally been accepted. cita is probably a participle 'as they move', as Munro takes it, comparing i. 1001 ex infinito cita corpora materiai, and not, as Giussani supposes, adj. 'swiftly', combining with saepe to emphasize the incessant motion of the atoms. So iv. 1215 semina cum Veneris stimulis excita per artus obvia conflixit conspirans mutuus ardor.

86. conflixere: intrans. as in vi. 373 hic quoque confligunt hiemes and iii. 833 ad confligendum venientibus undique Poenis; see Prol. V b, § 14 (b).

diversa: emphatic 'in different directions'; it is this which proves the occurrence of a blow.

87–8. neque enim … obstet: these lines give the reasons why the atoms leap asunder in all directions from a collision, firstly they are completely hard and solid, so that none of the energy of the blow can be expended in setting up motion of internal particles, as it would be in the case of a compound body; and secondly they are moving in the void and there is no obstacle to their movement in any direction (until they meet another atom). Epicurus puts the idea more fully in Ep. ad Hdt., § 44, reversing the order of the reasons': ἥ τε γὰρ τοῦ κενοῦ φύσις ἡ διορίζουσα ἐκάστην αὐτὴν τοῦτο παρασκευάζει, τὴν ὑπέρεισιν‎ ('resistance') οὐχ οἴα τε οὖσα ποιεῖσθαι‎· ἥ τε στερεότης ἡ ὑπάρχουσα αὐταῖς κατὰ τὴν σύγκρουσιν τὸν ἀποπαλμὸν ποιεῖ ἐφ‎' ὅποσον ἄν ἡ περιπλοκή‎ ('their interlocking in the compound') τὴν ἀποκατάστασιν‎ ('separation') ἐκ τῆς συγκρούσεως διδῷ‎.

87. neque mirum: without est, as in ii. 338, v. 1238, vi. 130; see ii. 1 n.

88. ponderibus solidis: 'in their solid masses'. Cf. i. 987. All pg 817through the atomic theory mass and weight are closely connected in idea.

a tergo ibus obstet: 'stands in their way in the rear', i.e. behind them as they are moving 'backwards', so to say, from a collision. OQ have a tergibus, but Vossius's conjecture has been accepted by most recent editors. ibus is a bye-form of the dat. plur. of is. It occurs in Plautus but with i long, Curc. 506, Mil. 74, Truc. 106. The short i in Lucr. may be due to analogy with quibus, as Ernout suggests, or possibly Lucr.'s form is derived directly from ĭs, and Plautus' form represents iibus (cf. iis). ĭbus is introduced again by conjecture by Lachmann in iv. 934 and vi. 755, by Munro in vi. 1012; see notes there and Prol. V a, § 8. Diels and Martin return to the MS. reading a tergibus, but, as Ernout points out, there is no other instance of the use of the plural in this phrase, and whereas tergum, -i and tergus, tergoris occur regularly, there is no authority for a form tergus, -ūs of the 4th decl., such as would be implied in tergibus. For Diels's view see Addenda.

89–94. The second reason for the diverse movement of the atoms is the infinity of space. If space were finite, they would all congregate at the bottom (i. 984 ff.); as it is infinite, they go on and continue their collisions and movements.

89. iactari: 'being tossed about', in their collisions and separations; so again in ii. 122 and 548. It is a more violent form of the idea in vagantur 83 and emphasizes the collisions; cf. iaculando in ii. 1108.

90. pervideas: an emphatic form of videas: so pervideamus i. 956, and pervideas i. 1117. Perhaps it has the idea of seeing through the superficial actions of phenomena to the atomic movements behind.

totius … in summa: 'in the sum of the whole', i.e. in the universe; see i. 993. totius acts here as a substantive; in i. 988 and vi. 650 summai totius it is adj. with summa as substantive. The fullest form is seen in vi. 679 summam summai totius omnem.

imum … in summa: probably an intentional play upon words; for summa see i. 235 n.

91. habere … consistant: sc. habere corpora prima ubi consistant. A characteristic Lucretian involution; see i. 15 ita capta lepore te sequitur cupide quo quamque inducere pergas and Prol. V b, § 14 (a).

92. quoniam spatium …: notice the cumulative repetition of the idea of infinity sine fine modoque … immensum … in cunctas undique partes. Diels, following Bonnet and Bockemüller, reads sine fine modoque omitting est, so as to bring sine fine modoque inside the acc. and inf. construction, but the independent statement is in no way unnatural.

94. ostendi: sc. i. 988–1007.

95–9. Once again the argument is much compressed. In 95–7 Lucr. is still thinking of the 'free' atoms tossed about in the void with their collisions and clashings and separations. Then suddenly in pg 81898–9 and without any introduction he passes to the union of the atoms in concilia in which, although the atoms still preserve their motions in every direction, they have yet gathered in a compound body which has its own limits and harmonious motions; the compound body is more than an arbitrary delimitation of a certain number of atoms in a certain extent of space, like the stars in a constellation. One would certainly have expected the idea to be introduced more definitely; but once again Lucr. is describing a picture in his mind rather than following out an argument. These atomic concilia Lucr. divides into two classes, those in which the atoms still leap back at (comparatively) large intervals, and so form air and light (107, 108) and those where they are more closely massed together, where their 'trajects' are small, 50 that they make hard things like rock and iron (103).

These lines correspond roughly to Ep. ad Hdt., § 43 κινοῦνταί τε συνεχῶς αἱ ἄτομοι τὸν αἰῶνα καὶ αἴ μέν‎ (something is probably lost here) εἰς μακρὰν ἀπ‎' ἀλλήλων διιστάμεναι, αἵ δὲ αὖ τὸν παλμὸν‎ ('recoiling', 'vibration') ἴσχουσιν, ὅταν τύχωσι τῇ περιπλοκῇ κεκλιμέναι ἤ στεγαζόμεναι παρὰ τῶν πλεκτικῶν‎. But both Giussani and Ernout notice that there are differences between the notions of Epicurus and Lucr. Epicurus seems, like Lucr., to distinguish between two classes of compounds, those in which the atoms move apart at considerable distances (I cannot agree with Giussani that αἴ μέν‎ … διιστάμεναι‎ alludes to the 'free' atoms), but divides the second class into those which are kept together by their own interlacing and those which move inside an 'interlacing' case. The latter will then presumably be liquids, which stand between air or light and the solid bodies. Lucr. does not make this distinction here, but the idea of the 'case' (στεγάζον‎, vas) plays an important part in his discussion of the nature of the soul in iii. 323–49, 425–44, where he compares the function of the body in enclosing the soul to that of the vessel which contains water. See notes there and my commentary on Epicurus, Ep. ad Hdt., § 43.

95. nulla: the error of OQ multa is corrected in L.

96. reddita: 'assigned', 'granted'. Cf. ii. 142 quae mobilitas sit reddita materiai corporibus and many other places in Lucr.

per inane profundum: see i. 1108 n.

97. vario: 'in different directions', corresponds to diversa 86.

exercita: 'driven on'; cf. 120 conciliis et discidiis exercita crebris; and iv. 862, 1128.

98. partim: acts as nom. and corresponds to pars in the next line; see n. on i. 483.

confulta OQ is kept by most recent editors. It is, as is shown by 101, a mere synonym for convecta and denotes the atomic clash; it has no suggestion, as Munro supposes, of the smooth and fine atoms of air, etc., which rest 'cushioned' against pg 819one another. conflicta of some of the Italian MSS. is impossible as the participle of confligere in a middle sense (86), Heinsius's compulsa is feeble and palaeographically improbable, as is Lambinus's contusa. consulta of the Verona edition is palaeographically better, but a pass. part. of consilire is impossible.

99. ab ictu: 'from the blow', 'from their collision'; prob. local, not instrumental 'by the blow' or temporal 'after the blow'. But see Prol. V b, § 15 (a).

100–11 is an elaboration of the ideas sketched in 95–9. Lucr. distinguishes between three possible modes of behaviour on the part of the atoms. (1) Some enter close-packed concilia, where they interlock and interlace and can only rebound at short intervals, and so form the hard bodies; (2) others enter less dense concilia, where they still rebound at long intervals and so form air, light, etc.; (3) others cannot adapt themselves either in shape or in motion to a concilium and so continue to wander as 'free' atoms through the void. The general sense is clear, but the description of the second class is obscure and the text possibly corrupt.

100. conciliatu: the act of forming a concilium; cf. ii. 134.

101. exiguis intervallis goes both with convecta and resultant: 'meet and leap back at short intervals'.

102. indupedita … perplexis … figuris: this is the περιπλοκή‎ of Epicurus. Lucr. explains in ii. 445–6 that the atoms which can interlock in this close union are those which are 'hooked' or 'branchy' in shape (hamatis … et quasi ramosis). The ideas of the effects of differently shaped atoms on the density or rarity of the compounds which they form was enunciated by Leucippus and worked out at great length by Democritus. Epicurus took over this tradition of atomism, but it is not clear how he reconciled it with his own idea of the incessant movement of atoms; it would be difficult for atoms closely locked by 'hooks' and 'branches' to recoil from one another even at short distances. I doubt whether he or Lucr. ever really faced this difficulty; the two conceptions were placed, as here, side by side. For the expression cf. ii. 459 at non esse tamen perplexis indupedita.

103. radices: 'roots' continues the idea of the entanglement of the atoms: the 'roots' of stone are the small atomic nuclei of the character of stone which again unite to form a piece of stone.

fera ferri corpora: almost a Lucretian pun. Cf. in a more pointed phrase Tib. i. 10. 2 quam ferus et vere ferreus ille fuit. Perhaps translate 'the brute bulk of iron'.

104. ⟨de⟩ is the necessary addition of l 31.

105–6. These lines have caused great difficulty and no fully satisfactory solution seems yet to have been found. paucula Itali, is the obvious correction for paucuia OQ. The older nineteenth-century editors had no stop at horum (104) and took 105 with what pg 820precedes: 'the close-knit atoms make rock and iron and a few other things of this sort, which go on (porro) wandering through the great void'. Munro has no note in his commentary and appears to find no difficulty. But particles of rock and iron do not 'wander through the great void', and the parallels of 83 and 109 make it clear that magnum per inane vagantur is a phrase applied to atoms and not to compounds: further, paucula is unnecessary after cetera, which would include all other solids. Brieger, perceiving these difficulties, placed a full stop at paucula and supposed a lacuna after 105 which he would fill with praedita corporibus mage levibus atque rotundis. But this does not avoid the difficulty of cetera … paucula and creates a new difficulty in the cetera of 106, which comes in very awkwardly. Hoerschelmann read parvola for paucula in 105 and corpora for cetera in 106. This constitutes a large change and is unsatisfactory, for there is no reason why the atoms constituting air and light should be 'small'. Giussani would expunge 105 as a variant on 109. More recent editors are agreed in placing a full stop at horum and retaining the MS. text. The difficulty then is the meaning of paucula … cetera. It could with a stretch be taken to mean 'a few others', 'a few of the rest': but can the atoms, which constitute the whole extent of air and light as well, be said to be 'few' in comparison with those which make the solids? Only, as Robin sees, in respect to a given measure of space: in a cubic foot of air there would be fewer atoms than in a cubic foot of iron. paucula then would come almost to mean 'rare, dispersed, isolated', as Giussani would take it, comparing two passages where paucus seems to approach this sense, iii. 278 corporibus quia de parvis paucisque creatast and iv. 71 quanto minus indupediri pauca queunt et quae sunt prima fronte locata: see notes there. The only way of getting over these difficulties which has yet been proposed is that suggested by Merrill in his later edition of the text (1917), namely to transpose paucula and cetera. The two latter classes of atoms are then opposed to the first which unite in the close concilia, and cetera is divided by apposition into paucula and multaque praeterea (cf. the subdivision of corpora into partim … partim in i. 483–4): 'of the rest which (do not enter these close unions but) pursue their wanderings through the void, a few do form concilia but with large intervals, and many others continue to wander as free atoms'. The accidental transposition of two words of similar scansion at the beginning of two successive lines is not improbable and would be to some extent paralleled by the transposition of voluntas and voluptas in ii. 257–8 and possibly of cortus and pontus in i. 271 and 276, and of conterimus and conficimus in ii. 1161–2. There is still a slight difficulty in paucula, but Lucr. is not thinking of the sum total of atoms which behave in different ways, but has before his mind the picture of a certain portion of space pg 821occupied either by many atoms in close union or a few in a looser concilium.

105. porro … vagantur must be taken together 'go on wandering in the void' instead of entering the close compounds. In a sense, as Giussani points out, all atoms whether 'free' or in concilia 'wander in the void', but Lucr. clearly uses vagari of those performing long 'trajects' between collisions whether in loose compounds or as free atoms.

106. dissiliunt longe longeque recursant: it is the addition longeque recursant which distinguishes the atoms in the loose compound. The 'free' atoms dissiliunt longe after their collisions, but then do not meet again.

107. in magnis intervallis repeats the idea of the previous line; the intervals are of course only 'great' in comparison with the tiny 'trajects' in the close compounds.

108. sufficiunt: 'supply', 'provide'. Cf. iii. 704 aliam naturam sufficit ex se, vi. 479 altas sufficiunt nubis, and Virg. Georg. ii. 423 ipsa satis tellus … sufficit umorem.

109. multaque praeterea …: are the 'free' atoms; and so Lucr. works back again to the picture in 85–8 and 95–7.

110. conciliis rerum: 'the unions which make things'.

nec usquam … recepta: 'and could not in any compound (usquam) be admitted and unite their motions too'. This contains an important point of theory. An atom to be admitted into a compound body must not only be the right size and shape, but its motions must be adaptable to those of the compound. Munro wrongly takes it 'and even though admitted (etiam recepta) have not been able to adapt their motions', but this is not good atomic doctrine. etiam probably goes, as translated above, with consociare … motus; Ernout prefers to take it with potuere 'nor were they able either', which is weaker.

112–24. This famous illustration of the behaviour of the atoms from the movements of motes in a sunbeam, which Lucr. has so beautifully worked out in detail, appears to have been traditional in atomism. Arist. De Anima i. 2. 404 a attributes it to Leucippus and Democritus: οἷον ἐν τῷ ἀέρι τὰ καλούμενα ξύσματα ἅ φαίνεται ἐν ταῖς διὰ τῶν θυρίδων ἀκτίσιν‎, and Lactantius (De Ira Dei 10. 9) says of Leucippus: haec, inquit, per inane irrequietis motibus volitant et hue atque illuc feruntur, sicut pulveris minutias videmus in sole, cum per fenestram radios ac lumen immiserit. Cf. also Stob. Ecl. i. 41. 43 where the same image is said to have been used by Epicurus to illustrate the movements of the soul atoms in the body.

112. memoro Itali;

memoror OQ. uti memoro can only mean, as in iv. 749 haec fieri ut memoro, facile hinc cognoscere possis (cf. vi. 1031 hic tibi quem memoro), 'as I say' equivalent to ut dico (iv. 1207, cf. quod dico ii. 870), an idiomatic equivalent in many languages pg 822for 'as I have said'. See Prol. V b, § 9 (d). But Lucr. has not yet referred to the motes in the sunbeam. Munro tries to get over the difficulty by regarding ut memoro as equivalent to qualem or quam memoro, comparing Plaut. Amph. 738 recte dicit, ut commeminit, where, as Ernout points out, ut has its normal sense of 'as'. Ernout himself supposes a constructio ad sensum: quod res ita fiunt ut memoro, simulacrum …, 'an illustration that things happen as I have said', but this is very unnatural. In a very interesting section on Lucr.'s habit of visualization and the way in which pictures in his mind lead him to anticipate their expression in words Büchner (Beobachtungen über Vers und Gedankengang bei Lukrez, pp. 26–8) thinks that this is an instance where the vision of the motes has been so firmly in his mind while he was describing their motions that he can say illogically but naturally 'as I said'. This is made the more probable by the occurrence of the illustration in the Atomists and Epicurus, which would have suggested it to his mind. See Prol. VII, § 33. I am inclined to accept this interpretation; but it is just worth noting, as Lachmann recalls, that memoror, which is the reading of OQ, is used by the translators of the Bible in the sense of 'recall', 'remember'. Is it possible that memoror is right and that it means 'as I recall', 'now I come to think of it'? rēī: scanned as spondee; see n. i. 688.

simulacrum et imago are pure synonyms and cannot be distinguished. The correction to the singular simulacrum is necessary.

113. semper is perhaps an exaggeration and does not imply more than 'often'.

instat: 'is present'. He means, as J. S. Mill might have said, that it is a 'permanent possibility of sensation'.

114. contemplator enim: a new introductory phrase used in vi. 189 and in Virg. Georg. i. 187 contemplator item.

cum … cumque should be taken together = 'whenever' on the analogy of quandocumque, quotienscumque. Lucr. is fond of the unnecessary addition of cumque to relatives and frequently with this tmesis: cf. ii. 21 quae demant cumque dolorem, vi. 85 qua de causa cumque, 738 quae sint loca cumque, and more closely parallel to the present instance v. 583 ut est oris extremis cumque notata. Munro and Merrill take cumque = et cum and regard solis lumina and inserti radii as parallel nominatives, fundunt being intransitive in sense: 'when the light of the sun and when the admitted rays pour through the dark room'. Apart from the grammatical difficulties involved Lucr., as Ernout points out, regards light as an emission from rays, as just below 117 radiorum lumine in ipso; cf. iv. 375, v. 287. lumina must therefore be acc. and fundunt trans. with inserti radii as its subject.

115. Cf. Virgil's imitation Aen. iii. 152 se plena per insertas fundebat luna fenestras.

opaca domorum: 'a dark room', the usual Lucretian use of n. plur. and gen.; see i. 86 n. and Prol. V b, § 4. 2.

pg 823

116. multa minuta modis multis: the quadruple alliteration is notable even in Lucr.

per inane: strictly of course the motes are moving in air, not in void, but Lucr. no doubt uses the word to enforce the parallel with the atoms which magnum per inane vagantur.

117. misceri: 'mingled in confusion', a word which, oddly enough, Lucr. never uses of the atoms.

118. proelia pugnas edere: so with the same asyndeton in iv. 1009. Some of the Italian MSS. have pugnasque which was adopted by Marullus, but Lucr. has only one hypermetrical elision v. 849 debere ut, and then not with -que, which is frequently so used by Virgil. Lachmann has an interesting note on the history of this scansional licence, but agrees with all subsequent editors in deciding against it here.

119. turmatim: 'in squadrons', a military term used by both Caesar and Livy; in Lucr. we may compare ii. 1172 viritim. For the form see Prol. VII, § 6.

dare pausam: 'to make a stop'; the idiomatic poetical use of dare = facere. See i. 288 n., and cf. ii. 187 dent … fraudem, v. 442 motus … dare, 1329 dabant … ruinas; dare pausam is used by Ennius and other early writers. For pausa see i. 747 n.

120. conciliis et discidiis: the two technical terms for the union and dissolution of atoms in a compound; as in the use of inane 116 Lucr. wishes constantly to recall the phenomena of which the motes are an illustration; so exercita recalls 97 varioque exercita motu.

122. iactari: cf. 89 n.

123. dumtaxat: 'at least', 'in some degree'. dumtaxat is an old legal word, originally a clause 'as far as touches' (taxat being the subjunctive of *taxere from the root tag- in tangere). Hence it obtains two distinct meanings: (a) 'so far, if not farther', 'at any rate', e.g. Hor. A.P. 23 denique sit quidvis simplex dumtaxat et unum; (b) 'so far and no farther', 'at most', 'only', as in iii. 377 with Lachmann's punctuation rara dissita sunt, dumtaxat. Here it is used in the former sense.

magnarum parva: purposely put in juxtaposition to emphasize the contrast.

124. exemplare: said to be the only existence of a substantive with the -ar termination which retains the final -e. Lachmann quotes an interesting scholium on Lucan iv. 563 hic exemplar posuit, Lucretius exemplare, nos in usu exemplum dicimus. See Prol. V a, § 1.

notitiai: notities in Lucr. is a translation of Epicurus' word πρόληψις‎, a 'general concept', formed in the mind, like a composite photograph, out of repeated sense-images, and therefore an 'anticipation' by which fresh images can be identified or a new object made. Its full technical sense is seen in v. 182 and 1047, and less completely in iv. 476, 479. Here, as in v. 124, it is used in the more general sense of 'concept', 'idea', and in ii. 745 in the more functional sense of 'comprehension'. See Prol. IV, § 4. For the pg 824variation in this word, as in materies, between 1st and 5th decl., see Prol. V a, § 2.

125–41. A further point: the movement of the motes in the sunbeam is not only an illustration but the outcome of the invisible movements of the atoms; see introductory note.

125. hoc: abl. going with quod 127, 'for this reason it is even more right that you should notice these bodies …'

par est: cf. i. 189, 458.

126. turbare: used absolutely 'to make a disturbance', not, as Munro says = turbari. So again in ii. 438, iii. 493, v. 502, 504, vi. 370; see Prol. V b, § 14 (b).

videntur: 'are seen to', opposed to the invisible motions of the atoms.

127. materiai: almost equivalent to principiorum.

128. clandestinos caecosque: the double adj. emphasizes the importance of the point. Note caecos … caecis 129, a characteristic Lucretian repetition.

129. plagis … caecis: the 'invisible blows' given to the compound body by the movements of the atoms which compose them.

130. retroque repulsa reverti: the alliteration assists the notion of rapid movement given by the dactylic rhythm.

132. omnibus: not abl. with principiis but dat.

error: literal and physical 'motion': 'this motion comes to all the motes from the atoms'.

133. prima: almost adverbial and corresponding to inde 134.

per se: sc. by their own atomic movement.

134. 'Those bodies which are composed of a small aggregate of atoms'. For conciliatu cf. ii. 100; a certain correction of conciliata OQ.

135. 'And are, as it were, nearest to the force of the atoms'; an important point. The atoms moving at incredible speed have tremendous force of impact in their collisions; the small nuclei, though retarded by internal ἀντικοπή‎, are next to the atoms in the force of their blows and so are able to rouse to motion the 'slightly larger' nuclei (137). This idea of the force of impact of the atoms recurs in iv. 193, where Lucr. explains that the simulacra can traverse immense distances in an instant of time quod parvola causa est procul a tergo quae provehat atque propellat. Cf. also vi. 340–2.

136. illorum: sc. principiorum.

137. proporro: 'in their turn', continuing the chain of blows. proporro is a certain correction by Turnebus of porro OQ: earlier restorations need not be considered. The word is peculiar to Lucr. and occurs again in ii. 979, iii. 275, 281, iv. 890, v. 312. See Munro's n. on v. 312.

138. 'The motion rises from the atoms and gradually issues to our senses'; the motions of the atoms are below the ken of our senses; cf. ii. 312 omnis enim longe nostris ab sensibus infra primorum natura pg 825iacet, and so the motions of the motes may be said to 'rise' into our ken. For exit paulatim nostros ad sensus cf. Ep. ad Hdt., § 62 πυκνὸν ἀντικόπτουσιν ἔως ἄν ὑπὸ τὴν αἵσθησιν τὸ συνεχὲς τῆς φορᾶς γίνηται‎. Just as the atoms gather together in larger and larger combinations till they compose a body which is visible owing to its size, so their motions become slower and slower through ἀντικοπή‎ until they are slow enough to be perceived.

140. illa quoque: sc. the motes.

141. nec quibus …: the relative clause is tacked on, as so often in Lucr., with a slight anacoluthon in construction. Cf. 87–8 durissima quae sint ponderibus solidis neque quicquam a tergo ibus obstet.

aperte: otiose but it makes a fine assonance with apparet. After 141 Giussani inserts 308–22. They would no doubt be more logically in place here, but it is rash to rearrange an unfinished poem.

## (c) The velocity of their motion. 142–64.

142. It is not difficult to realize the speed of the atoms. Firstly consider how, when dawn renews the light upon the earth and the birds sing upon the trees, the sun suddenly appears and bathes the world in his rays. 150. Yet the sun's light does not travel through the void, but is delayed because it has to cut its way through the air; nor do the particles of light travel singly but in complex union, so that they are retarded within as well as without. 157. But the atoms in their solid singleness travelling through void without hindrance, each of them a single whole moving in one direction, must far exceed the sun's light in speed and travel many times the distance in the time that the sun's rays spread over the sky.

Lucr. proceeds to consider, as he promised in 65, the rate at which the atoms move. Epicurus in Ep. ad Hdt., § 461 (where he is possibly referring not to the atoms but to the εἴδωλα‎ of sight) says, 'their passage through the void, when it takes place without meeting any bodies which might collide, accomplishes every comprehensible distance in an inconceivably short time. For it is collision and its absence which take the outward appearance of slowness and quickness.' Lucr. expounds this idea here: quickness and slowness of movement in things are determined by the amount of ἀντικοπή‎, whether external and internal. A moving object may be delayed either by colliding with other external objects or by the internal collisions and consequent movements in all directions of the atoms which compose it. The fastest moving thing of which we have pg 826cognizance is light. But it is moving not in void but through air, which causes external ἀντικοπή‎, and though it is a very 'loose' compound (105–8), it yet is a compound and is therefore retarded by internal ἀντικοπή‎ as well. The atoms on the other hand between their collisions are moving through void and are themselves solid and whole with their inseparable parts (i. 599–614); there is therefore no ἀντικοπή‎ internal or external to delay them and their pace must far exceed that of light. In fact, as Epicurus says, they travel over 'every comprehensible distance in an inconceivably short time'. The argument follows on closely from the ideas in the end of the last section and gives Lucr. the opportunity for a marvellous description of dawn and sunrise. It is probable that further arguments followed to correspond to primum in 144, which have been lost in the lacuna after 164: see notes there.

142. quae mobilitas sit reddita: cf. ii. 64 quaeque sit ollis reddita mobilitas magnum per inane meandi.

144. primum: 'in the first place', one of Lucr.'s regular introductions to a series of arguments: i. 742, 931, iii. 94, etc. Editors, seeing that there was no other argument to follow and noting Virgil's imitation Aen. iv. 584 et iam prima novo spargebat lumine terras … aurora, have taken primum with cum 'when first'. But in ii. 1030 principio is similarly used with no deinde or praeterea to follow, and in the present instance it may well be that there was a second argument in the passage lost after 164.

145. variae volucres: 'the birds after their kinds', see i. 589 n.

146. aera per tenerum: cf. i. 207 aeris in teneras … auras.

liquidis … vocibus: cf. v. 1379 liquidas avium voces.

148. convestire: the metaphor is probably derived from Cicero's Aratea, where it occurs several times. It passes on to Virgil, Aen. vi. 640 largior hic campos aether et lumine vestit purpureo.

150. vapor: 'heat', as always in Lucr., who regards it as a concrete body.

151. meat: in spite of the double subject, vapor and lumen serenum, the verb remains in the singular as the two ideas are the same to Lucr., i.e. the particles which form the sun's light also make his heat.

152. aerias … undas: 'while it, as it were, beats aside the waves of air' and so cleaves a path. The indic. is vouched for by the reading of OQ. Munro, following Lambinus and Lachmann, reads diverberet, holding that dum = donec 'until'. But that is not the idea. The light does not move slowly until it cleaves a path and then fast again: it has to be cleaving the air all the time. quasi goes with diverberat, apologizing for the metaphor, though in i. 222 the verb is used without apology.

153. singillatim: 'individually', 'separately', as in iv. 89, 105, pg 827a typical Lucretian word; see Prol. VII, § 6.

corpuscula … vaporis: 'the particles of heat', sc. the atomic nuclei which unite to form the sun's rays. Just as each one of these is formed of atoms which produce internal ἀντικοπή‎, so the particles themselves do not move singillatim, but are united in the complex structure of the ray, colliding with each other and producing further ἀντικοπή‎. corpuscula is used 5 times by Lucr., usually, as here, of the (compound) particles of things, but in ii. 529 of the atoms themselves.

154. complexa … conque globata: words which are most applicable to the denser compounds, cf. 102 indupedita suis perplexis ipsa figuris, are here applied to the looser structure of light. complexa used in a passive sense 'interwoven' as in v. 922. It is the equivalent of perplexis in ii. 102, but no doubt chosen for the assonance with conglobata.

155. inter se retrahuntur: 'are dragged back by one another' by the internal ἀντικοπή‎, as opposed to extra officiuntur, the retardation by the air.

156. officiuntur: the editors notice that this is the only use by Lucr. of the personal passive of a verb which governs the dat.

157. solida primordia simplicitate takes us back, as it is intended to do, to the discussion of the 'least parts' of the atom in i. 599 ff. Cf. esp. 609 sunt igitur solida primordia simplicitate. The atom with its inseparable parts and no admixture of void can have no internal ἀντικοπή‎.

158–9. cum … foris is in opposition to non per inane meat vacuum (151) and ipsa … unum to nec singillatim … conque globata (153–4).

158. remoratur: 'checks', 'delays'. The word has been corrupted both in O and Q and the divergence of their readings suggests a blot or erasure in the archetype, but Pontanus's correction must be right. remorantur is used intransitively in ii. 75.

159. ipsa suis e partibus unum: 'they being themselves each one of them a single whole of their (inseparable) parts', a very compressed expression. For this sense of unum cf. i. 604 alterius quoniamst ipsum pars, primaque et una, and for suis e partibus, iii. 545 contracta suis e partibus (sc. anima). So Ernout, Diels, and Martin. Munro, followed by Giussani, takes the phrase in the same sense, but thinks it is necessary to read the plural una agreeing grammatically with ipsa 'being themselves single wholes'. For una plural Giussani compares ii. 919 animalia sint mortalibus una eademque and iii. 616 unis sedibus, and, after Munro, Cic. Pro Flacco 63 unis moribus. On the whole I think the singular is tenable. It is anyhow impossible that unum should be an emphatic anticipation of unum … locum in the next line, for though such repetition is frequent in Lucr., there is no reason to emphasize unum locum.

pg 828

160. unum in quem coepere locum: 'the one direction in which they have begun to move', either downwards owing to gravitas or in some other direction owing to an ictus. Munro is wrong in confining it to the former, and Giussani needlessly asks how atomic motion can be said to have a beginning; coepere refers merely to the start of the particular direction in which the atom is at the moment moving.

conixa: 'pressing on', a certain correction for conexa OQ; the whole point is that the atoms are not 'connected'.

161. praecellere: 'to surpass'; here only in Lucr.

163–4. Repeated in iv. 207–8 with lumina for fulgura.

163. multiplex: 'many times as great', = πολλαπλάσιος‎.

164. pervulgant: 'spread over', 'fill'; cf. ii. 346 et quae pervulgant nemora avia pervolitantes. Munro well compares i. 3 quae terras frugiferentis concelebras.

fulgura: a fine variant for lumina which he has used in 162.

165–6 have no connexion in sense with what precedes, and all modern editors (except Bernays, who transposes 167 before 165–6 and reads persectati for persectari), have agreed with Pontanus that a considerable passage—perhaps a whole leaf—has been lost after 164, of which these two lines are the conclusion. But there is considerable divergence as to what the lacuna contained. Lachmann and Munro believe that it was something which led up to the anti-theological argument of 167–83 and therefore include the lines in the bracket which they continue to 183: see below. Lachmann imagines a statement that many philosophers think that the world was created by the will of the gods and find it easier to refer generally to this 'rather than to follow up the movements of every atom'. Munro thinks that the gods were the subject of persectari and videant and supposes a statement like that in Cic. De Div. ii. 105 negant id esse alienum maiestate deorum, scilicet causas omnium introspicere, ut videant quid cuique conducat. But, as Giussani points out, contra haec in 167 is fatal to the idea that the previous lines had anything to do with the anti-theological argument, which must begin with the emphatic at quidam contra haec. He would construct the lost passage quite differently: in the first place he supposes a second argument (and possibly a third) for the velocity of atomic motion introduced by deinde or some other particle corresponding to primum in 144. Then he argues that Lucr. has not fulfilled the promise in 62 of showing how the atoms by their motions res varias gignant genitasque resolvant (98–108 cannot be regarded as a sufficient treatment, esp. as it has nothing about atomic dissolution). 165–6 might well be the conclusion of this exposition: the minds of men (or some such subject) can understand the general principles of atomic concilium and discidium but they cannot 'follow up the movements of each individual atom, in order to see in what way each one of them pg 829behaves'. This view, which is adopted by Diels, seems to me a far more probable explanation of the lost passage, nor can I accept Munro's suggestion that the passage was in fact never included in the poem; Lucr. could not have composed 165–6 without a context.

163. singula quaeque: 'each individual atom', not, as Giussani takes it, 'each atom of every species'. The question of the different kinds of atoms does not arise till 332.

166. videant: the obvious correction of deant OQ. ui was lost by haplography after ut.

## (d) A digression: the world was not divinely created for men. 167–83.

165. Some hold that without divine agency nature could not arrange the seasons and their fruits so suitably for the use of men, nor all the gifts which pleasure, the guide of life, leads them to accept, including the joy of love, which saves the race from perishing. 174. But this view is far from the truth. Even without knowledge of the atoms, from climate and weather and many other things I could show that the world is not divinely created; it has so many faults, as I will show later. Now we must go on with the atomic motions.

Lucr. pauses in the course of his exposition of the motion of the atoms to answer the view of the Stoics and the champions of religion that the world was divinely created for the pleasure of men. His argument is brief and dogmatic; the world is so imperfect that it cannot be a divine creation. This argument is expounded in v. 195–234, a passage which he seems to have in mind in 182, and which is linked to this by the repetition of 177–81 in v. 195–9. The argument is traditional.

Lachmann and Munro bracket these lines together with 165–6, which on their view belong to the same argument, on the ground that they interrupt the orderly exposition of the motion of the atoms and that 183 is impossible side by side with 184. They think that they were written by the poet—after the passage in Book v—but were not intended to come in here and were never adjusted to their place.

More recent editors have removed the brackets, in my opinion rightly. Lucr. likes from time to time to break up a long argument with a digression which does not demand such strenuous thinking (i. 921–50 is an obvious example). In particular he indulges from time to time in outbursts like this against the theological point of view; v. 110–234 is the most conspicuous example (cf. also ii. 1090–1104 and vi. 379–422) and it is to be noted that that passage too is bracketed by Lachmann. Moreover, if, as Giussani thinks, the lacuna before 165 contained an exposition of the ways in which the atoms gathered together to form things, the anti-theological protest pg 830in this place would be natural and to the point. Nor, though there is some awkwardness, is 183 impossible with 184; 183 is the return to the general subject after the digression, and 184 the introduction to the special part of it which the poet is now going to treat.

These lines must certainly be taken in connexion with v. 195–234. It is, I think, probable that they were written after the passage in Book v and inserted here. But they are not inappropriate here, especially on Giussani's view of the contents of the lacuna, which has altered the whole aspect of the problem.

167. quidam: those who uphold the theological view of the world, including no doubt the adherents of popular religion, but in particular the Stoics, whose teleology, as Robin points out, was essentially anthropocentric; it was for man's sake that the world was created by the 'world-spirit'. This is just the view which Lucr. opposes here.

ignari materia. 'ignorant of matter', i.e. of its laws and behaviour. A very odd gen., the nearest parallel to which in Lucr. is iv. 436 maris ignaris. See Prol. V b, § 4. 3. Ignarus is elsewhere constructed in Lucr. with an indirect question as in v. 88, vi. 64, ignari quid queat esse, v. 998 ignaros quid vulnera vellent. Moreover, though materiai occurs over forty times in the poem, and materiae three times, the gen. is always preceded or followed by a defining substantive, most frequently by corpora or copia. This led Hoerschelmann to suspect that a line was lost after 167 beginning with corpora and supplying an acc. and inf. construction for ignari. He himself suggested corpora sponte sua volitare invicta per aevum, but accepting Giussani's view of the lacuna after 164, it is more likely to have been of the form corpora sponte sua caelum terrasque creare. It is further to be noted that the MSS. at the end of 168 have reddi, which suggests an acc. and inf. construction with materiai naturam taken together; but it does not seem possible to make sense on these lines, and in fact natura is not elsewhere combined with materiai. On the whole, then, it seems best to accept the strange ignari materiai.

168. credunt: OQ have reddi, but a main verb meaning 'think' or 'believe' is required. On the analogy of i. 154 = vi. 91 fieri divino numine rentur Manilius read rentur here and has been followed by most editors. Diels and Martin follow Brieger in reading reddunt on the analogy of 179 = v. 197 aliisque ex rebus reddere multis. But reddere there does not mean to 'think', which is the meaning required here, but 'to give an account of', 'to prove'. This is shown by the only other place where reddere occurs alone in this sense i. 566 possint tamen omnia reddi … quo pacto fiant (which is exactly parallel in sense to non poterit ratio reddi just below in 572); reddere in both places is a short form of rationem reddere, which Lucr. uses ten times in the poem. I therefore prefer Pontanus's credunt, which is palaeo-pg 831graphically more probable than rentur: it is, moreover, the kind of variation which Lucr. likes to make in parallel phrases.

169. 'So suitably to the needs of men', admoderate being constructed with the dat. It is a ἅπαξ λεγόμενον‎, necessitated perhaps, as Merrill suggests, by the impossibility of accommodate; Lucr. does not use the adv. apte, which would have been the obvious solution, though aptus adj. occurs frequently.

171. et iam cetera: 'yes, and the other things as well'. iam strengthens et; indeed this must be the origin of etiam.

quae is the object of adire and dia voluptas the subject of suadet.

172. ipsaque deducit: breaks off in a Lucretian anacoluthon. deducit, the metaphor from leading the bride home, appropriate here in view of the thought of 173. Note an elaborate alliteration here, d.d.d.d. and d.v.d.v.

dia voluptas: 'holy pleasure'; for dia see n. on i. 22. voluptas here of course the Epicurean moral motive, as in i. 1; see Prol. IV, § 10.

173. blanditur … propagent: 'woos them to propagate': a paratactic construction: for blandiri in this sense Munro compares Liv. xxi. 1. 4. Hannibalem … blandientem patri Hamilcari, ut duceretur in Hispaniam.

saecla: 'races' as usual.

174. quorum: sc. hominum, implied in humanum.

176. a vera … ratione: the 'true reasoning' is, as frequently, the philosophy of Epicurus; see n. on i. 54.

177–81 = v. 195–9, except quodsi iam for nam quamvis, paratam for creatam (180), mundi for rerum (181), and a possible variation at the end of 181, small changes which indicate that Lucr. intended the passage in both places. See Prol. VII, § 30.

177. quamvis … ignorem: 'even though I were ignorant' (though I am not) as is shown by v. 195 quodsi iam … ignorem.

178. caeli rationibus: 'the behaviour of the sky', climate and weather, as is shown by the context in Book v. 195–9. Cf. also v. 1183 caeli rationes ordine certo et varia annorum cernebant tempora verti. For ratio see i. 54 n. Lucr. answers his opponents on their own ground, see 169–70, hence ipsis; even from what the Stoics claim as the ground of their belief in divine creation.

ausim: for the form cf. v. 196 and ausis ii. 982, iv. 508, v. 730, vi. 412. Ernout (Morph., § 238) regards them as 'subjunctive-optative forms', like siem, sim. See Prol. V a, § 11.

179. reddere: 'account for', 'prove'; see n. on 168.

180. divinitus 'by divine agency'; cf. i. 150 n.

181. tanta stat: restored by Lachmann from v. 199. But OQ have quamquam, which suggests that there may have been some variation of the phrase here. Many suggestions have been made, but none are convincing. The best are Postgate's quanta stat praedita culpa! (adopted by Merrill) but it is a rather unnatural exclamation, and pg 832the old correction of Pontanus, which was for long the vulgate, quae tanta est, is perhaps better; Diels's quae quam stat is doubtful Latin. For stat cf. Virg. Aen. xii. 407 iam pulvere caelum stare vident, and Lucr. i. 563 et finita simul generatim tempora rebus stare.

182–3. These lines are added for the sake of connexion, 182 to anticipate the fuller treatment of the question in Book v, 183 to bring the reader back after the digression: see introductory note.

## (e) The universal motion downwards due to weight. 184–215.

184. Nothing can by its own force tend upwards. Fire is only an apparent exception. It is true that it comes into existence and grows upwards, just as do crops and trees, yet all things that have weight move downwards as far as they can. 191. When flames leap up to the roof of a building, that is not their natural motion, but is due to upward pressure; just as when blood spirts upwards from the body. 196. See how water drives up a log; the more it is pressed downwards, the more vigorously it is shot back till it is more than half out of the water. 201. All these things in void naturally fall downwards. So flames under pressure leap upwards, though the natural tendency of their weight is to drive them downwards. 206. We often see shooting stars moving in any direction, the stars themselves fall towards the earth (at their setting), the sun's rays are scattered over the earth in all directions, thunderbolts move athwart the rain, lightnings flash from the clouds from all sides. Fire then often falls on earth.

In 84–5 Lucr. mentioned the two causes of the incessant movement of the atoms in all directions, their weight (gravitas) which caused a natural downward movement and blows (ictus), which deflected that movement in other directions. For the time he assumed these two causes (see introductory note to 80–141). Now he returns to them, in this paragraph to the motion downwards through weight, in the next section (216–93) to the blows which result from the atomic swerve.

It is noticeable that in this paragraph he does not deal explicitly with the motion of the atoms downwards, but with the universal tendency of all things that have weight to move downwards, in spite of apparent instances to the contrary. The reason no doubt is that Lucr. is led aside from his main point in order to refute his old enemies, the Stoics. As has already been noticed (see introductory note to i. 1083–1113) they held that though bodies of earth and water tend to the centre, those of air and fire tend upwards away from the centre. It is this doctrine which Lucretius is here concerned to refute, and he uses in effect two arguments: (1) 191–205. That in all cases of upward motion, the natural tendency of bodies is to pg 833move downwards, but they are propelled upwards by some extraneous force. This is aptly illustrated by the case of the log which, however much submerged in water, is yet shot up again. (2) In 206–15 he appeals to our experience of the phenomena of the sky; we are all familiar with instances in which fiery bodies do move downwards, e.g. stars at their setting and the sun's rays, or in slanting directions across the sky, e.g. shooting stars, lightning, and meteors (fulmina). His arguments are less satisfactory than usual, and in the first he contents himself with a reiterated dogmatic assertion of his own view. Controversy seems often to have this effect on him, witness his arguments against his predecessors in the middle of Book i and that against the Stoics at its end.

The argument here naturally raises the question what can be meant by 'upwards' and 'downwards' in infinite space. It would seem, especially in view of his emphatic assertion in i. 1070 that infinite space can have no centre, that the difficulty must have been present to Lucr.'s mind. Yet he never deals with it. Epicurus, however, has a strange section on the subject in the Letter to Herodotus (§ 60). We cannot, he argues, predicate upward and downward motion in infinite space with reference to a highest and lowest point, an absolute top and bottom, for such do not of course exist, but we can with reference to ourselves or to any point in space of which we choose to think. The motion from our feet to our head, however prolonged, is to us motion upwards and the opposite motion downwards. This solution is of course mathematically absurd (inepte excogitatum, as Brieger calls it), but to an Epicurean it is not ridiculous. Firstly, it is founded on αἴσθησις‎; we know what we mean by upward and downward, and by prolonging the conceptions to infinity we get the idea we need. Secondly, the Epicurean conception of space is always based on a geocentric idea of the world, and space is conceived as radiating out in all directions from the earth itself; therefore it is consistent with this notion that the terrestrial conceptions of 'up' and 'down' should be applied to space in general. As Giussani says (Studi Lucreziani, p. 169), the contradiction is inherent in the conception of space itself, at once infinite and relative.

184. nunc locus est: an introductory formula imitated by Virgil, Georg. ii. 177.

185. sua vi: 'of its own force', sc. by its own natural tendency; things can only move upwards through external compulsion.

186. sursum = subversum; see i. 1058 n. Lucr. uses sursum 15 times as adv., sursus twice, in 188 and in vi. 718. The latter is strictly the nom. masc. of the participle, but it has become so completely adverbial that it can be combined with versus.

187. dent … fraudem: 'deceive' = fraudent, cf. dare pausam 119 n. For the phrase cf. vi. 187 ne tibi sit frudi.

pg 834

188. sursus … versus: see 186 n.

augmina sumunt = augentur; see n. on ii. 73.

189. The plants and trees are mentioned here as an example of upward movement, just as they are in i. 1092–3, because the Stoics believed their growth to be due to internal fire moving them upwards; see n. there.

190. pondera … cuncta: 'all things with weight', equivalent in sense to 'all things owing to their weight'. Giussani suggests that pondere should be read, but that is excluded by the parallel line 205, though there pondera is used in a slightly different sense.

quantum in se est: 'as far as rests with them' apart from external force; equivalent to sua vi in 185.

191. tecta domorum: 'the roof of the house'; the n. plur. and gen. here have their full force; see Prol. V b, § 4. 2.

192. degustant: 'lick'; the poets usually employ lambere in this metaphorical sense: e.g. Virg. Aen. ii. 684 lambere flamma comas and possibly Lucr. v. 396 lambens cuncta perussit.

tigna trabesque: here probably in a technical sense, trabes being the main beams and tigna the rafters let into them, as Munro explains. Cf. Caesar, Bell. Civ. ii. 9. 2 quoted by Ernout: supra eum locum duo tigna transversa iniecerunt … supraque ea tigna … transversas trabes iniecerunt. So again in vi. 241. Below in 196 the same words are used in a quite general sense, 'logs and beams'.

193. sponte sua … subiecta: 'that they do this of their own accord being shot up without (external) force'. Subiecta has been suspected by most editors on account of the meaning of the word and the false concord with ignes (191), and Lambinus substituted subigente 'without force driving them upwards'; Brieger's subeunte is much weaker. But the MS. text is defended by Postgate in J. P. xxiv. 134, who in support of the meaning quotes Virg. Gẹorg. iv. 385 ter flamma ad summum tecti subiecta reluxit, which looks like an imitation. The neuter might be explained as referring back to flammarum corpora in 187 or perhaps better as one of Lucr.'s loose references. Cf. e.g. 201 haec, iii. 558 corporis atque animi vivata potestas inter se coniuncta valent. Diels and Martin also accept subiecta. See Prol. 5 b, § 7. 1.

194. Notice the form of the comparison: the blood spirts upwards because of the pressure of the blood behind it in the body striving to escape.

quod genus: an internal acc. frequently used adverbially by Lucr. in the senses 'just as' and 'for example', e.g. iii. 266, 276. Cf. the uses of omne genus (see i. 1026 n.) and hoc genus vi. 917, and see Prol. V b, § 3. 2. Lucr. also uses quod genus est in the same way and Lachmann, holding that est should always be present, here changes e to est and makes corresponding changes in iii. 431 and vi. 1058. But later editors have not accepted this arbitrary pg 835principle, and Munro points out that quod genus is frequent in ad Her. but never found with est.

sanguis … cruorem: Merrill and Ernout distinguish between sanguis, the blood in the veins, and cruor, the gore which flows from them; cf. Tac. Ann. xii. 47 ubi sanguis artus extremos suffuderit, levi ictu cruorem eliciunt. It is probably true that cruor would not be used for the blood in the body, but sanguis is used for both: e.g. ii. 354 sanguinis exspirans calidum de pectore flumen.

195. alte emphasizes the upward motion.

197. respuat: 'rejects', 'repels'; cf. vi. 1054 of the magnet respuit ab se atque per aes agitat.

alte derecta: 'straight down', go closely together, derecta agrees with tigna trabesque, but in an adverbial sense. For the meaning 'perpendicular' cf. Caes. Bell. Gall. vii. 23 trabes derectae … in solo collocantur. Lachmann read deiecta taking it with alte 'thrown from a height', but that is pointless. The MSS. have altu, which the Itali correct to altum, used adverbially, a rather strained construction, which might, however, be what Lucr. wrote.

198. multi: 'many of us' together, a queer little addition to emphasize the pressure on the log and to bring out the picture in Lucr.'s mind.

199. revomit: 'spews it out' (cf. respuat 197), an almost certain emendation of Pontanus for the weak removet of OQ, which is nevertheless retained by Martin.

200. plus … parte: prob. not plus iusta parte 'more than you would expect' or, as Munro explains it, 'more than it would if left alone', though Lucr. uses plus parte aequa in iv. 1231, but in the mathematical sense 'more than a half. Cf. i. 617 dimidiai partis pars. When partes is used with a numeral, a denominator one greater must be supplied: duae partes equals 'two-thirds', tres partes 'three-quarters', and so on. So pars alone means 'a half; cf. Germanicus, Arat. 588 Bootes in terras abit et noctis plus parte relinquit. The fuller expression is plus media parte as in Ovid, Metam, iii. 43 media plus parte leves erectus in auras.

201. haec … cuncta: a vague general reference to the flames, the gore, and the logs. So in iii. 506 haec refers to vis animi atque animai (499).

quantum est in se: notice the slight variation from quantum in se est 190.

202. vacuum per inane: again in a loose sense as in 115 per inane: Lucr. is thinking specially of the logs which would fall down in the air. deorsum: as in iii. 1016, but d$eo¯$rsum in 190, 205, and elsewhere.

203. debent flammae quoque: all editors have accepted Marullus' transposition for the metrically awkward quoque debent flammae of OQ. Cf. i. 290 sic igitur debent venti quoque flamina ferri.

204. expressae: 'driven upwards' is emphatic; lit. squeezed out, which is just the idea of the σοῦς‎ of Democritus.

quamquam … pg 836pugnent: Lucr. uses quamquam with subj. again in ii. 277 quamquam vis extera multos pellat. There seems to be no distinction of meaning; syntax had not yet become rigid. Cicero according to Ernout reserves the subj. after quamquam for cases where the verb is also potential.

205. pondera here, in contrast with 190, must mean 'their weights', not heavy things in general. Note that Q reads pondere, which may well be right.

in sest … deducere: Marullus for inest … ducere: cf. 190. Lachmann prefers est in se … ducere, cf. 201, but deducere is needed for the sense to reinforce deorsum.

deducere pugnent: an unusual prolate infin.; see Prol. V b, § 12 (a).

206–15. Here Lucr. starts his second argument: not only is movement upwards the result of external force, but as a matter of fact fire and light do move downwards or in slanting directions.

206–7. Imitated by Virgil, Georg. i. 365 saepe etiam stellas vento impendente videbis praecipites caelo labi, noctisque per umbras flammarum longos a tergo ducere tractus.

206. nocturnasque faces caeli = shooting stars. Cf. v. 1191 noctivagaeque faces caeli flammaeque volantes.

sublime: internal acc. used adverbially as in iv. 133, vi. 97; cf. proclive ii. 455 and see Prol. V b, § 3. 2.

208. natura: sc. the natural working of atomic laws.

209. in terram seems the natural correction of in terra OQ, but see 212 in terras. If there is a distinction in meaning, it probably is, as Giussani suggests, that here the earth is regarded as an astronomical body. Martin reads in terras here, which may be right.

stellas et sidera: Munro distinguishes them as 'stars and constellations', but Lucr.'s usual word for constellations is signa and the expression is probably merely tautological. The reference must be to the setting of the stars and not, as Merrill seems to suppose, to shooting stars, which have been 'dealt with in 206–8.

210. 〈caeli〉 Bernays's supplement for a defective line, is generally accepted. Cf. Cic. Arat. 297 caeli de vertice. The Itali read summo and Marullus suggested aetherio.

omnis … in partis: sc. not only downwards but sideways.

211. conserit: 'sows' the fields; an unusual metaphor helped out by arva. It is used metaphorically by Catullus 64. 208 caeca mentem caligine Theseus consitus.

212. A rather stiff and unnecessary verse, summing up the two previous lines and emphasizing the downward movement of the sun's rays.

vergitur: middle, an unparalleled use. The active is usually intransitive in sense.

213. transversos per imbris: 'athwart the rain', a transferred epithet; see n. on i. 10, and Prol. VII, § 12. For its direct use cf. iv. 22 transversum ferre, vi. 190 transversa per auras: so also pg 837vi. 719 adverso flumine, 725 fluctibus adversis, ii. 217 rectum per inane, vi. 689 rectis faucibus.

214. abrupti nubibus: cf. Ep. ad Pyth., §101 where among various causes of lightning Epicurus has κατ‎ʼ ἐκπιασμόν, θλίψεως τῶν νεφῶν γινομένης‎.

215. concursant: not, as in iii. 395 to 'run together' = concurrere, but 'run about', almost equivalent to discurrunt; Giussani takes it 'follow after one another', but this is less probable.

cadit … vulgo: a rather abrupt summing up of the paragraph, which hardly follows from what he has said.

## (f) The swerve of the atoms. 216–93.

216. As the atoms are falling straight down through the void owing to their weight, at undetermined times and places they swerve a little with just the smallest change of direction. 221. If it were not so, all would go on falling like rain-drops through the infinite void, there would be no collisions and no blows, and nature could have created nothing.

225. It is a mistake to suppose that the heavier atoms, moving straight through the void, fall down on the lighter and so cause the blows which set up the creative motions. 230. It is true that through water or through air increase of weight causes increase of speed in the fall, because water and air cannot offer an equal resistance to all bodies, but yield to the heavier more quickly. 235. But void cannot at any place or time resist anything, but must from its very nature yield; therefore all things, of equal or unequal weights, must fall equally fast through void. 240. So the heavier can never catch up the lighter, nor of themselves cause the blows which make the various motions which lead to creation. 243. Therefore the atoms must swerve a little, but only the very least, for we must not suppose a (normal) slanting motion, which facts would contradict. 246. For it is obvious that bodies with weight cannot, as such, move sideways in their fall, at least perceptibly. But no one could possibly perceive that the atoms never swerve at all from the straight line.

251. Again, if every motion is the outcome of the last in an endless chain and the atoms do not by swerving initiate a motion to break the law of destiny, and prevent an infinite causal sequence, how can living things have free will? 257. How can there be wrested from fate that will by which we move as our pleasure bids, and swerve in our movements at no definite time and in no definite direction, but just as our mind suggests? 261. (1) For there can be no doubt that it is the will which starts the motion, which then passes through the limbs. See how race-horses when the barriers are suddenly removed cannot leap forward as quickly as their mind desires. For pg 838the whole mass of matter over all their bodies has to be stirred, so that in every limb it can respond to the mind's desire. 269. It is clear then that the start of motion comes from the act of will in the mind and thence spreads through all the body and the limbs. 272. (2) Again, it is quite different when we move under compulsion from someone else. For then it is clear that the whole body is propelled against our will, until the will has checked it limb by limb. 277. It is obvious then that, though we are often compelled to move unwillingly by some external force, yet there is something in our breasts which can resist. 281. At its bidding sometimes the whole mass of the body is compelled to change its direction, and after an external impulse it is checked and stopped. 284. Therefore there must be in the atoms too this same power, another cause besides blows and weight, from which we derive our power of will; for nothing can be created of nothing. 288. For weight prevents every motion from being due to the external force of blows. But what prevents the mind from having an internal necessity in all actions, and doing all under compulsion, is the tiny swerve of the atoms in undefined directions at uncertain times.

A long section on an important point. The three paragraphs must be considered together, for they are all part of an intricate argument. Lucr. has explained (by implication) in the last paragraph the downward fall of the atoms in the void owing to weight, which is the ultimate cause of all motion. He has now to consider the cause of the blows (ictus, plagae) which set the atoms moving in all directions, and so lead to the formation of nuclei and the creation of things. He rightly (225–42) rejects the idea that the heavier atoms catch up the lighter in their fall and so cause the necessary meetings; for, as he argues, although we see heavier bodies falling faster than light ones, that is because they are falling through air (or water), which although it is only a loose atomic texture (natura tenvis 232), yet offers resistance to falling bodies, and a greater resistance, the lighter the falling body. But void can offer no resistance at all to any body, therefore all will fall at an equal rate. (This argument has always been recognized as one of the most felicitous pieces of a priori reasoning on the part of the Epicureans.) 1

The cause of the blows must then be sought elsewhere, and Lucr. finds it in an occasional unpredictable 'swerve' of the atoms from the absolutely straight line. If we imagine them falling in parallel perpendicular lines, the effect of the swerve will be that the swerving pg 839atom will come into collision with an atom in the next line, which it will set moving in a slanting direction until it collides with another in the third line, while it rebounds itself, hits an atom in its own line, and sets it moving slantwise towards the line on the other side. In this way the swerve of a single atom might in time set up collisions throughout the whole number. Indeed some critics have supposed that this single swerve was what Lucr. had in mind, but it is obvious from the language of the whole section that the swerve is constantly repeated and may be made by any atom at any moment.

This famous theory of the swerve (παρέγκισις‎, clinamen) is not mentioned in any extant work of Epicurus (though it is possible that it was referred to in a passage now lost in § 43 of the Letter to Herodotus), but it is dealt with at some length in the inscription of Diogenes of Oenoanda1 and is referred to by Cicero,2 Plutarch, and other ancient commentators. Even in antiquity it was ridiculed; Cicero3 calls it a 'puerile invention' and says elsewhere4 that 'to invent such a theory was more disgraceful than to be unable to defend his views', for it is a breach of the natural law that nothing happens without a cause (see l. 287). The earlier modern critics mostly echoed these opinions, but of recent years, largely in view of a tendency to similar views in physics and biology, the 'swerve' has been taken more seriously and is not regarded as 'absurditas' or 'insania', as Marullus called it. For it has another and more important function in the Epicurean system besides the mechanical production of the atomic clash.

Lucr., following his regular proceeding, argues (251–93) in favour of his theory from experience, but in a most unexpected quarter. He passes from the primitive movements of the free atoms in the void to the most delicate and complex of all atomic structures, the mind of man, and maintains that the fact of man's free will proves the existence of the swerve of the atoms (284–93). It was on this question of free will that Epicurus most conspicuously broke away pg 840from Democritus and the Atomists, who set up necessity (ἀνάγκη‎) as the governing principle of the universe and adopted a rigid determinism. Epicurus, anxious above all things for his moral system, saw that he must secure free will as an exception to the universal determinism. 'It were better', he says in the Letter to Menoeceus, 'to follow the myths about the gods than to become a slave to the "destiny" of the natural philosophers: for the former suggests a hope of placating the gods by worship, whereas the latter involves a necessity which knows no placation.'1 Diogenes of Oenoanda2 brings the two notions closely together imagining a discussion with a follower of Democritus who denied that there was any 'free movement' in the atoms; 'Epicurus', he says, 'brought it to the light, showing from the facts of experience that it was a swerving movement. And the chief proof is this, that if destiny is believed in, all exhortation and rebuke is annulled', i.e. it is no use telling a man what he ought to do or ought not to do, unless he is free to do it. It is interesting that in this account the swerve is spoken of as a free movement of the atoms. Cicero3 too speaks as if the desire to preserve free will was the main motive of the idea of freedom in the movement of the atoms. The same sequence of thought is seen in Lucr. himself (284–7); the free will of man shows that there is an element of freedom even in the atoms.4 What in them must be regarded as a kind of spontaneity becomes in the conscious and complex concilia of the mind an act of free will. Throughout nature, then, there is this element of spontaneity, which at 'uncertain times and uncertain places' can break through rigid necessity and show itself as something wilful. Guyau5 was probably right, and finds support in Plutarch,6 in supposing that Epicurus attributed to the swerve of the atoms that 'chance' in nature, pg 841which he sometimes regarded as a parallel cause of phenomena beside necessity: the casus1 which Lucr. sets side by side with vis or the foedera naturai.

All these ideas are brought together in a very illuminating reference in Philodemus'2 work on 'Methods of Inference'. In a passage in which he is laying emphasis on the Epicurean principle of insisting on 'non-contradiction' (οὐκ ἀντιμαρτύρησις‎) for hypotheses concerning the unseen (ἄδηλα‎) he says 'it is not sufficient in order to prove the tiny swerves of the atoms to adduce chance and our free will, but it must also be shown that nothing in phenomena contradicts either one or the other'. Here we have Lucr.'s procedure explained; his proof of the swerve is the existence of free will in man, and in 249–50 he insists that nothing in perception is against such a tiny swerve. Moreover, Philodemus' insertion of 'chance' as the alternative to free will lends colour, as Plutarch's reference does, to Guyau's suggestion that in the processes of inanimate nature the swerve is the cause of the element of spontaneous chance.

Seen in this larger light the idea of the swerve does not seem so absurd. It is an arbitrary assumption, but not a puerile expedient, for it is an assertion of freedom in a world of necessity (quod fati foedera rumpat 254), which is an essential element in Lucr.'s whole outlook on the world and on life. See also Addenda.

[It is worth noting that certain recent theories are not unlike that of Epicurus. Modern thinkers are forced back on the parallel question at what point in the scale of creation consciousness begins, and there are those who have asserted that it must be traced right back to the atoms. W. K. Clifford (Essays and Remains, vol. ii, pp. 61, 85) argued that 'in order to save continuity in our belief, we are obliged to assume that along with every motion of matter, whether organic or inorganic, there is some fact which corresponds to the mental fact in ourselves', and elsewhere 'a moving molecule of inorganic matter does not possess mind or consciousness, but it possesses a small piece of mind-stuff'. Haeckel (The Riddle of the Universe) postulates what he calls a 'psycho-plasm', and even speaks of 'atoms with souls'. J. S. Huxley (Essays of a Biologist, p. 243) says that 'we must believe that not only living matter, but all matter, is associated with something of the same general description as mind in higher animals'. Samuel Butler (Note-Books, PP. 72–3) says that 'no objection can lie to our supposing potential or elementary volition or consciousness to exist in atoms.… By pg 842giving them free will we do no more than those who make them bound to obey fixed laws.… The little element of individual caprice (supposing we start with free will), or (supposing we start with necessity) the little element of stiff-neckedness, both of which elements we find everywhere in nature, these are the things that prevent even the most reliable things from being absolutely reliable. … Atoms have a mind as much smaller and less complex than ours as their bodies are smaller and less complex.' Compton (The Freedom of Man) puts the physical and psychical problems together and argues for free will on the ground that as the result of our scientific investigations into the behaviour of the atom and the electron 'natural law' has to give place to 'uncertainty' and to 'statistical probability'.

Though these far-reaching speculations are not universally accepted, the physical movement of the atom is now recognized as something very like Lucr.'s 'swerve'. Sir James Jeans in the Presidential Address to the British Association in 1934 says in describing Bohr's theory: 'The electron did not move continuously through space and time but jumped, and its jumps were not governed by the laws of mechanics, but … by the laws of probability. Of 1,000 identical atoms, 100 might make the jump, while the other 900 did not. Before the jumps occurred, there was nothing to show which atoms were going to jump.']

216. te cognoscere avemus: Lucr. is said to be unique in constructing avere with acc. and inf.

217. corpora: here 'the atoms', though, as Giussani suggests, Lucr. may have supposed that compound bodies too had the power of an occasional swerve. This would accord well with the idea that the swerve was the cause of chance in inanimate nature; see introductory note.

rectum per inane should be taken together 'straight through the void', with the same transference of the epithet as in transversos … per imbris in ii. 213. Cf. also iv. 550 rectoque foras emittimus ore, vi. 689 rectis … faucibus eicit. This is better than to regard rectum as an adv. and take it with feruntur.

218. ponderibus propriis should certainly be taken with feruntur; the atoms fall straight through the void owing to their own weight, the weight which is their property. It is the equivalent of pondera, quantum in sest ii. 205.

incerto tempore … incertisque locis: the swerve does not occur at fixed intervals either of time or of space; it is spontaneous and unpredictable. Some atoms may not swerve at all, others presumably, though Lucr. does not say so, more than once.

ferme emphasizes incerto 'at quite uncertain times'. Cf. v. 242 haec eadem ferme mortalia cernimus esse; vi. 10, 324. According pg 843to Ernout it should be regarded as standing for ferime, a superlative of fere. Here it is a certain restoration in the Itali for firme OQ.

219. spatio depellere paulum: 'swerve a little from their course', equivalent to recta regione viarum declinare 249. Recent editors agree that there is no reason to doubt the MSS. reading. spatium, though so frequently used by Lucr. in the technical sense of 'space', is here the common metaphor from the race-course and depellere is used intransitively, according to Lucr.'s frequent habit; see Prol. V b, § 14 (b). Cicero has in a metaphorical sense a striking parallel to both usages: De Amic. 12. 40 deflexit iam aliquantum de spatio curriculoque consuetudo maiorum. The older editors doubted the intrans, use of depellere. Marullus's decedere is feeble and Avancius's se pellere also weak. Munro inserted se after propriis in the previous line, which places the reflexive pronoun very far from the verb; cf. his view of ii. 250. Lachmann invented the word decellere and disliking also the use of spatio read incertisque loci spatiis decellere 'swerve at uncertain intervals of space', comparing 163 multiplexque loci spatium. Hoerschelmann, adopting decellere, proposed spatio decellere paulo 'swerve by a small interval'. But all these changes are gratuitous.

220. tantum quod: 'only so much as'. It should be regarded, as Ernout says, as an acc. of the type of quod genus, mirum quantum, etc. He quotes a good parallel from Suet. Aug. 98 navis quae tantum quod adpulerat 'had only just come to shore'.

momen mutatum: 'a change of direction', 'an oscillation', prob. here, as Munro says, a metaphor from the ῥοπή‎ of the balance: elsewhere in ii. 1169, iii. 144, 188, 189, vi. 474 Lucr. uses the word in the sense of 'movement'. For his fondness for substantives with the termination -men see Prol. VII, § 2.

possis: the idiomatic second person 'one might call'; see i. 327 n. and Prol. V b, § 10 (a).

221. declinare: 'swerve' as in 250; cf. inclinare 243, and clinamen 292.

222. imbris uti guttae: the comparison lies in the parallel lines; rain does not necessarily fall in downward lines.

per inane profundum: see i. 1108 n.

223. offensus: 'collision' and plaga the 'blow' that results from it. offensus is Lucr.'s equivalent for offensio, which will not scan. He uses it again in ii. 438, iv. 359, vi. 333.

224. principiis: dat., almost a 'Lucretian' dat. for principiorum. Cf. i. 58 n.

225–50. Lucr. deals with the view that the heavier atoms might catch up the lighter in the downward fall; see introductory note.

225. aliquis: editors have been inclined to say that Lucr. has Democritus in mind, but it is very doubtful whether he assigned weight as an absolute property of the atoms; the evidence seems pg 844rather to show that it was something which they acquired though their bulk in resistance to the 'whirl'. (See my Greek Atomists and Epicurus, pp. 128 ff.) Further, when Lucr. definitely differs from Democritus he is inclined to say so, as in iii. 370. It is more probable that aliquis is quite general in reference: 'anyone who holds this theory is mistaken'.

potesse: see i. 665 n.

226. feruntur: the indic. may, I think, be retained, as by all recent editors except Ernout. It is undoubtedly part of the or. obl. and in normal Latin would be subj., but Lucr. is not rigid about such dependent clauses. Cf. i. 1058 et quae pondera sunt sub terris, where the indic. might be defended on the ground that Lucr. would admit the existence of things with weight beneath the earth (see n. there). But v. 629 seems to supply a very close parallel: quanto demissior eius cursus abest procul a caelo terrisque propinquat, tanto posse minus cum signis tendere cursum, where emendation to subj. is impossible. Ernout holds that ferantur is necessary here and supposes the indic. to be due to the influence of 217. See Prol. V b, § 9.

227. incidere: here and in 240 with dat., in iv. 568 with acc., in vi. 145 with in and acc.

228. genitalis … motus: 'the motions which result in creation'. Cf. i. 58 genitalia corpora rebus, ii. 62 genitalia materiai corpora.

reddere: 'produce'. Cf. ii. 890 vitalem reddere sensum.

230. per aquas: note that Lucr. here, as in ii. 196–205, regards motion through air and through water as of the same nature: they are both loose compounds.

231. pro ponderibus: 'in proportion to their weight'.

232. tenvis: nom. with natura; cf. i. 281 mollis aquae natura and n. there. For the disyllabic form tenvis with consonantal u see i. 875 n.

233. aeris haud: a certain correction of Niccoli's for badly corrupted readings in OQ.

234. exsuperata: the neuter in reference to the two naturae of water and air. Cf. iii. 559 where the neuter refers to two potestates, and see Prol. V b, § 7. 1. OQ have exsuperate and, as Munro says, Niccoli may be right in preferring exsuperatae.

236–7. Repeated almost verbally from i. 1079–80.

236. subsistere: 'resist', 'sustain' itself beneath their attack. rēi disyllabic; see i. 688 n.

238. inane quietum: the void is motionless in contrast to the ever-moving atoms, and being motionless cannot offer resistance. The epithet is poetical.

239. aeque ponderibus non aequis: 'it yields equally to bodies unequal in weight'. Cf. i. 1074 ff. omnis enim locus ac spatium … concederedebetaeque ponderibus. For the idea see Ep. ad Hdt., § 61, quoted in n. 1, p. 838.

pg 845

240. levioribus incidere: see 227 n.

241. per se: 'of their own power', i.e. without the aid of the clinamen.

242. qui: certainly nom. plur. and not abl. 'by which', though this is combined with a plural antecedent in v. 233.

varient: 'vary', i.e. change the direction of.

243–50. The summing up of the argument, but with the proviso that the swerve must be only the very least, lest the experience of the senses that all bodies naturally fall downwards should be contradicted.

243. inclinare: a variant for declinare in 221, 250. So Cic. in speaking of the swerve in De Nat. Deor. i. 26. 73 uses inclinatio.

244. nec plus quam minimum: this insistence on the slightness of the swerve is of course important; see Plut. de Soll. Animal. 7 (quoted in n. 6, p. 840) ἐπὶ τοὐλάχιστον‎ and Cic. De Fato 10. 22 intervallo minimo—id appellat ἐλάχιστον‎ The conceṗtion of the minimum plays an important part in the Epicurean theory, and we have met it already in the minimae partes of the atom (i. 599 ff.) and in the conception of the minimum of time in connexion with the motion of the atoms; see introductory note to ii. 142–64, p. 825, n. 1; and on the whole idea of the minimum Bignone, Epicuro, Appendix, pp. 228 if.

ne fingere … videamur: not quite, as Giussani takes it, 'the swerve is so small that we cannot call it an oblique direction', but rather 'lest we should seem to be supposing a normal slanting motion' in contradistinction to the universal movement downwards. We must not introduce the idea of a second normal motion for the atoms, but only that of an occasional slight spontaneous deviation from the regular motion.

245. res vera: 'the fact of experience' which tells us that all natural motion is downwards.

247. in sest: for inest OQ l 31 as in 205.

248. quod: 'as far as you can perceive' by the senses, quod = quoad, as in quod potero, etc.

249. nil is not the subject of declinare but an adverbial acc. 'that they, the atoms, do not ever swerve at all'.

recta is Niccoli's certain addition to a defective line in the MSS. and is accepted by all editors except Lachmann, who appears not to realize that regione here means 'direction' not 'spot' and reads nulla 'that no atom swerves from any spot on its journey' with a redundant negative. But for regio in this sense in Lucr. cf. iv. 514 normaque si fallax rectis regionibus exit and 1272 eicit enim sulcum recta regione viaque; see n. on i. 64.

250. declinare … sese: with much hesitation I am inclined to think that the MSS. reading may be retained. declinare is intrans. in ii. 221, but in ii. 259 we have declinamus … motus trans. declinare pg 846sese seems possible and would give the same variation as is found with insinuare intrans, in i. 409, iii. 485, v. 73, vi. 89, 385, 778, but with se i. 116, ii. 436. The distance between the infin. and the reflexive pronoun is awkward but not impossible in Lucr.; if Munro's insertion of se in 218 were accepted, there would be a close parallel. But the two difficulties make the reading doubtful. Of emendations the best are either Winckelmann's poscat (better palaeographically than Lachmann's praestet) for possit 'who could claim that he saw', but the expression is unnatural, or Giussani's sensu (better than Bernays's sensus, for it is nonsense to ask 'which of the senses?') for sese: 'who could perceive by sensation?', but sensu is otiose. Nencini's suesse is feeble. Giussani rightly notes that the argument of these two lines is dictated by the rule of the Canonice that where the senses cannot give positive evidence we may accept a theory against which there is no evidence (οὐθὲν ἀντιμαρτυρεῖ‎ Ep. ad Hdt., §§ 47, 48, etc.). This principle is applied in ii. 867–70 and frequently in the astronomical section of Book v; see Prol. IV, § 9.

251. 'If all motion is for ever linked in a chain.' Motion is the cause of change and therefore of all occurrences and actions; if it is linked in an everlasting chain of which there is no breach, the result is a rigid determinism.

motus is the correction of the Itali for motu OQ. I accept it but fill the vacant space in the next line with motu, supposing the two words to have been interchanged, like voluntas and voluptas in 257–8. The sense of 251 is then complete in itself and 252 is an independent clause. Martin alone among editors keeps motu here and reads motus in 252: this makes a very involved sentence si omnis novus motus semper conectitur et exoritur vetere motu. For the sense of conectitur Smith aptly quotes the saying attributed to Chrysippus by Aul. Gell. (vii. 2. 7) quamquam ita sit ut ratione quadam necessaria et principali coacta atque conexa sint fato omnia.

252. The deficient line was filled with semper of the Itali by all editors till Martin, who, as explained above, reads motus keeping motu above. I think his general idea is on the right lines, semper being a weak repetition of semper in 251. But the sense is greatly improved if motu be read here: 'And a new motion arises from the old in a determined series.' The occurrence of the two words in the same place in successive lines will account both for the change and the loss. The repetition motus … motu … motus in three following lines is not un-Lucretian; cf. 307 motus, 309 motu, 311 motus.

ordine certo: opposed to incerto tempore ferme incertisque locis 218. For the phrase cf. v. 732, 1183, 1439, and notes there.

253. motus principium go together, 'a start of motion'. The swerve is not of course a start of motion in the sense that it can move a stationary body, but it is a start of motion in a new direction; pg 847so plaga is coupled with gravitas in ii. 84–5 as one of the causes of motion; see ii. 83 n.

254. quod fati foedera rumpat: 'which could break the law of fate', i.e. of the Democritean ἀνάγκη‎. As Giussani points out, the foedera fati are not the same as Lucr.'s own foedera naturae, for the latter would include the clinamen; fatum is εἰμαρμένη‎. The clinamen is natural and is no breach of the 'laws of nature'.

256. libera is separated at a great distance from voluntas with which it goes, but it is the important word and must be emphasized.

257. voluntas … 258. voluptas: Lambinus; OQ have voluptas … voluntas. It is generally agreed that libera voluptas is an improbable expression. The natural expression is libera voluntas 'a free act of will' which is 'wrested from determinism'. So in 258 quo ducit quemque voluptas is closely paralleled by ii. 172 ipsaque deducit dux vitae dia voluptas. Pleasure gives the incentive and will determines the action. The alternative, adopted recently by Ernout and Diels, is with Lachmann to keep voluntas in 258 and read potestas in 257. This is supported by 286 unde haec est nobis innata potestas and perhaps by iv. 877 passus proferre queamus, cum volumus, though that might be quoted on the other side. But voluntas is the motive force in 261, 270, and 276; and palaeographically the transposition of two almost, identical words at the ends of the lines is much more likely than the loss of potestas.

257. per terras: 'all over the world', cf. ii. 608, etc.

259. declinamus … motus: 'we turn our motions in one direction or another'; Lucr. is here intentionally using of the act of free will the expressions applied to the swerve of the atoms. So nec tempore certo nec regione loci certa repeats incerto tempore ferme incertisque locis 218 and is repeated of the atoms in 293. The transitive use here is some support for declinare … sese in 250.

tulit: cf. iii. 44 si fert ita forte voluntas.

261–71. This is the first of two arguments Lucr. adduces to support his view that there must be free will, which in sentient beings is the start of motion. The fact that between the desire to move and the actual movement there is an interval of time shows that the motion is originated in the mind by the will and then permeates through the body till the limbs move.

261. dubio procul: an appeal to experience: we all know that it is our will which determines our motion and then sets the limbs moving.

his rebus: one of Lucr.'s vague references to the thought in his mind, here primarily the motus of walking (258–9), but also to other acts of will.

sua cuique voluntas: 'his own will gives to every animal the start of motion, etc.'

262. principium: sc. motus principium, as in 253.

rigantur: 'flow through the limbs'; a favourite metaphor of Lucr.'s, from the pg 848channel which irrigates the crops. Cf. iv. 907 somnus per membra quietem irriget, and of light iv. 203 caelumque rigare, cf. v. 282, 594.

263–4. Imitated by Virgil in Georg. iii. 103 nonne vides cum praecipiti certamine campum corripuere ruuntque effusi carcere currus?

patefactis … carceribus: at the start of a race the chariots were arranged behind barriers, which were drawn away. So again iv. 990 carceribus patefactis.

263. tempore puncto: a strange expression for 'in a moment', which recurs in ii. 456, 1006, iv. 214, but nowhere outside Lucr.; so also vi. 230 puncto … in tempore. In the more normal expressions temporis puncto (i. 1109) and temporis in puncto (iv. 164, 193) puncto is clearly a substantive 'in a point of time', but in temporē puncto and puncto in tempore, puncto looks grammatically like a participle and the phrases are much more difficult to explain. Munro regards puncto as participle 'while time is pricked', i.e. marked down; Merrill (C.Q. xviii. 42) takes puncto as substantive 'time made into a point', i.e. 'atomized time'; Ernout thinks it is an adjective in a substantival sense 'in time considered as a point'. He is, I think, right in saying that in iv. 214 iamne vides igitur quam puncto tempore imago, quam makes Munro's explanation impossible, and that in vi. 230 et liquidum puncto facit aes in tempore et aurum, puncto must go adjectivally with tempore. These two passages are also fatal to the natural proposal to read tempori' puncto, for it is unlikely for the s to be suppressed and then the i elided. The phrase should probably be accepted as an idiom of which no precise grammatical account can be given. Cf. tanto quique iii. 700, v. 343. See Prol. VII, § 13.

264. equorum vim: like the Greek idiom withβία‎. Cf. ii. 326 virum vi and, with Munro's emendation, ii. 42 ecum vi. equorum is the certain restoration in the Brescia edition for quorum OQ.

265. de subito: cf. iii. 643. The coupling of preposition and adverb is archaic; cf. the more usual derepente, and see Ernout's note.

267. conciri OQ is established against conquiri O1, adopted by Munro, firstly by concita, a Lucretian repetition, and secondly by the meaning. The materiai copia has to be 'roused' but not 'sought out'. For the 4th conj. form see Prol. V a, § 12.

268. conixa: 'making an effort', for the body too has sensation (see Book iii) and can join in the effort of the mind. Brieger supports conexa OQ, but that is a word always used by Lucr. in reference to the 'interweaving' of the atoms.

269. corde: explained by Munro and Giussani as 'the seat of the animus'. But the animus is in the pectus (iii. 140) and cor elsewhere appears to be nearly equivalent to animus: e.g. iv. 44 id licet hinc quamvis hebeti cognoscere corde (repeated in v. 882), v. 1456 alid ex alio clarescere corde videbant, v. 1107 ingenio qui praestabant et corde pg 849vigebant. So here a corde and ex animi voluntate are not different in meaning.

270. id: again a vague Lucretian reference to all that has preceded, like haec in 201. Cf. i. 392 tum putat id fieri, 667 ex nulla facere id si parte reparcent. Editors, seeing that the only direct reference is initum motus in 269, have explained that initum suggested initium and so caused a neuter pronoun. Others wish to introduce an unparalleled abl. voluntati (voluntatei Housman) and omit id.

272–83. The second argument really contains two points. The main idea is that voluntas not only, as has been shown, can start motion in things, but also, if they have been impelled by an external force, voluntas can slow down and stop the motion. Secondly, that the external push starts the whole body moving at one moment, whereas voluntas only gradually stops it, as it communicates its will per membra per artus, a proceeding exactly similar to the process of starting described above. It is a careful and subtle piece of writing.

272. nec similest ut cum: 'nor is it like as when', a curious construction, but Lucr. is fond of this kind of expression. Cf. v. 1065 longe alio sonitu rabie restricta minantur et cum, and also 1070–1 and 1077.

ictu is explained by viribus … coactu in the next line.

273. coactu: a typical Lucretian formation found here only in the poem, but elsewhere (only in the abl.) in Cicero and Caesar.

274. materiem totius corporis omnem: i.e. the whole body is moved at once in contrast to the action of voluntas described in 266–8; so also per membra 276, per membra per artus 282.

275. perspicuumst nobis: Itali have perspicuum est nobis, whence Lachmann corrected the strange reading of OQ perspicuum nobisst. Lachmann counts 9 other instances in which st is in the MSS. attached to the wrong word.

276. refrenavit: 'has put the brake on', 'restrained' as one would a galloping horse. In 283 we have the whole process of the original slowing of the motion and the final stop.

277–83. Giussani, arguing that 'no one could read these lines without a sense of their dislocation', places 281–3 before 277–80. All other editors have succeeded in performing this feat, and it is clear that 282–3 which contain reference both to the starting of motion by voluntas and its stopping should come at the end of the two arguments.

277. quamquam … pellat: for quamquam with subj. see ii. 204 n. The earlier editors read pellit … cogit … rapit, but pallat OU and cogat point clearly to the subjunctive.

multos: Giussani is probably right in thinking that here the picture in Lucr.'s mind is that of a jostling crowd, carrying people along without their wishing it. Otherwise it might mean 'an external force often drives men'. No pg 850other editor seems to have found any difficulty in what is surely a strange expression.

278. pellat: Avancius for pallat OU now universally accepted; see n. on 277.

279. rapi after cogat: again Avancius's correction for rapit OU, which is only possible if pellit and cogit be read in the previous line.

in pectore nostro: which is the seat of the animus; see iii. 140 and n. on corde 269.

280. obstareque: for -que attached to short e see i. 134 n.

281. arbitrium: here only in Lucr.; by other authors an act of will or choice is often described as liberum arbitrium.

282–3 give the two sides of the argument: free will can both start the body moving in a new direction, and stop it when it has been externally propelled; in both cases the action is gradual, as the 'order' given by the mind spreads per membra per artus.

283. refrenatur … residit: 'is checked and stopped': see 276 refrenavit and n. there.

284–93. These concluding lines are by no means easy to interpret, though no editor except Giussani seems to have perceived their difficulty. They should be divided into two parts. 284–7 are the conclusion of the section on the clinamen. First of all he argues directly from the free will of which we are conscious in our own actions to a similar freedom of movement in the atoms. This is exactly the form in which Cicero, De Fato 23, quoted on p. 840, n. 3, represents Epicurus as arguing. Secondly, he definitely represents the clinamen as a third cause of motion (285) besides the pondus and plagae postulated in 85; this too is confirmed by Cic. (De Fato 10. 22) who speaks of the swerve as tertius quidam motus extra pondus et plagam. This assertion seems strange, seeing that the clinamen itself is the original cause of the plagae, but in the actual history of an atom few of the plagae it received would be the direct result of a swerve and far more the result of the rebound of another atom from a previous collision. The statement is perhaps illogical, but Lucr. has the picture of the atoms falling in their natural motion downward, which is continually interrupted by collisions, and from time to time by a spontaneous swerve. It must be remembered that there never was historically, so to speak, a 'pre-clinamen', or 'pre-blow' era. The last six lines are the conclusion of the whole section on motion from 62 onwards, and Lucr. goes back in thought, as Giussani pointed out, to the original opposition to the views of Democritus (see introductory note to 80–141). Democritus had been content to say that the atoms were all 'naturally' in motion in all directions, meeting and clashing. Epicurus, as has been seen, tried to go behind this idea and to discover the causes of this motion and the resulting 'blows' (plagae). The first he found in the weight pg 851of the atoms which caused their natural motion downwards (85, 184–215): he can therefore assert that pondus prevents plagae from being the sole cause of motion and creation (288). And secondly, there is the clinamen, which is the cause of motion in other directions and so of the plagae themselves, which he here represents in its other and more important capacity, as saving the mind and will of man from rigid determinism. (For mens in 289 see n. ad loc.)

284. in seminibus quoque: Lucr. argues from the experience of the freedom of the will back to the atoms 'there must be in the seeds too (as well as in man's mind) another cause of motion'.

idem: 'the same thing', explained by esse aliam

285. aliam … causam: 'a third cause': cf. v. 536 aliam naturam.

286. motibus: again a 'Lucretian dat.', 'another cause of motion': cf. v. 509 motibus astrorum nunc quae sit causa canamus and see Prol. V b, § 5. 1.

287. Tohte and others wished to expunge this line on the ground that Lucr.'s principle nihil ex nihilo is only applicable to the creation of matter, as in i. 150. But there is no reason why he should not apply it to causes as well. It is in fact an important point in the reasoning: one must not assume free will in the mind of man, unless one can find some cause for it far back in the constitution of things.

288. omnia fiant: more than 'all things are created'; rather 'all things take place', equivalent in sense to Lucr.'s normal use of gerantur.

289. externa quasi vi: sc. by the blows; pondus is an internal cause, arising out of the constitution of the atoms.

mens is Lambinus's correction for res OQ, almost universally accepted. Martin restores res but does not explain how he takes it, except by reference to Cic. Nat. Deor. i. 25. 69 (quoted in n. 2, p. 839), which does not seem to support it. It might mean either 'the actual facts of the case' or 'the compound body itself', but with either of these meanings 290 and 291 are unintelligible. Though the correction is a large one, it must, I think, be accepted: mens may have been written mēs.

necessum: here only so completely a substantive that it has an adj. with it.

290. intestinum: 'internal', as opposed to external compulsion; in iv. 118 it is used substantivally 'an intestine'.

291. devicta quasi: 'as it were defeated' in a struggle to obtain its freedom, a highly metaphorical expression. Brieger, arguing that in fact there had been no such struggle, preferred to read devincta. But cf. i. 1082 medii cuppedine victae and n. there.

quăsī: there appears to be no other instance of a final long i in this word. Ernout thinks the lengthening is due to its place in the arsis of the principal caesura of the line (cf. ferē ii. 370 and iv. 935). The MSS. quaei points to quasei, the spelling marking the length, pg 852and presumably the i was originally long, but shortened in an iambic word, as in mihĭ, tibĭ, etc. See Prol. VI, § 12 (b). Lachmann thinking the scansion impossible and wishing for an object to ferre patique inserted id after quasi, and Munro hoc. But, as Giussani points out, the line becomes then an otiose enlargement of 289–90.

ferre patique: absolutely, 'to suffer and endure' as a slave to necessity.

292. id facit: once more the emphatic statement that the swerve of the atoms is the natural cause of the act of free will.

clinamen: the famous word, after which the theory is named, here only. It is formed, according to Lucr.'s habit, with the suffix -men. See Prol. VII, § 2.

## (g) Two appendixes to this section. 294–332.

### 1. The permanence of matter and motion. 294–307.

294. Matter was never more densely packed, nor more loosely than it is now: for it cannot either be increased or decreased. 296. From this it follows that the movement of the atoms too is as it always has been and ever will be; 300 therefore the things created will be formed and grow and flourish under the same conditions, just as nature has appointed for each of them. 303. There is no force which can alter the sum total of things; there is nothing outside the universe into which any kind of matter can escape, nor whence some new force could enter the world and change the nature and the motion of things.

This passage links up the idea of the permanence of matter, contained in the two propositions in the First Book that nothing is created out of nothing (159–214) and nothing utterly perishes (215–64), and with the new notion of the motions of the atoms and the consequent creation of things, which has been explained in this first section of Book ii. It is based—and in particular 303–7—on the principles enunciated by Epicurus in the Letter to Herodotus:1 'the universe always was such as it is now, and will always be the same. For there is nothing into which it changes: for outside the universe there is nothing which could come into it and change it.' These principles Lucr. elaborates and draws inferences from them. The addition or loss of matter to the universe would have the effect of making the sum of matter either denser or rarer; this would alter the nature of the motions of matter, for it would increase the number of concilia with either small or great intervals between their component atoms; and this change of motion would in its turn change the character of the things formed by concilia or at least pg 853change the proportions between one kind and another of such concilia. But the sum of matter can never be either greater or less: therefore none of these consequent changes can happen; there will always be the same relative density and rarity of matter; motions will therefore always preserve the same proportions, and the relative number of various things will remain such as nature has ordained.

The last five lines seem to come in oddly at this point and to repeat what has already been said, and Giussani supposes that they were originally written for a place just after the proof of the indestructibility of matter in Book i. 264, which would correspond to the place in which Epicurus' statement is found in the Letter to Herodotus. But they are necessary here. In the previous argument Lucr. has assumed the permanence of matter, on which it all depends, basing it no doubt on the proofs given in Book i. Now he wishes to add the further proof, of which he avails himself again in iii. 816 ff. and v. 361 ff., that there is nothing outside the universe into which matter might escape (deperit 296) or from which new matter might come (adaugescit). It should be noted that in 306–7, as both Giussani and Bignone (Epicuro, Appendix III, p. 253) have rightly seen, a third idea is introduced, that new matter would have the further effect, as explained above, of changing the atomic motions and therefore the character of things created. This links these lines closely on to 294–302 and shows that they are vital in their place here. This idea of change through a new force entering the universe is further just the idea which Epicurus has put rather obscurely in the passage already quoted.

294. stipata magis: 'more densely packed', because the atoms being greater in number would have on the average to perform shorter 'trajects' between their collisions. Cf. i. 329 nec tamen undique corporea stipata tenentur omnia natura.

295. maioribus intervallis: 'with larger intervals' between atom and atom, in this the opposite of stipata magis.

296. adaugescit: transitive, 'nor does anything come to swell its bulk'. The verb is used only here in Lucr., but it is used in Cic. Arat. Prognost. 181 stridor … ortus adaugescit scopulorum saepe repulsus. Lucr. is, however, fond of formations from this root and has adaucta particip. in ii. 564 and vi. 508, adauctu in ii. 1122, and adaugmen in vi. 614.

deperit inde: 'is lost from it', i.e. the copia materiai.

298. ante acta aetate: 'in ages past', i. 234 n.

299. post haec OQ should be retained against posthac ed. Junt., which Lucr. does not use.

300. consuerint is probably generic, not, as Giussani thinks, causal: 'the sort of things which have been created'. Brieger pg 854unnecessarily changes to the indic. cons$ue¯$runt.

eadem condicione: 'under the same conditions' of atomic motion and therefore of creation.

301. vique valebunt: 'and will flourish in their strength': note the four stages, birth, existence, growth, and flourishing, vi valebunt represents what he calls in i. 564 aevi flos. Cf. v. 1112 viresque vigebant. OQ have uiquo, of which this is a much better correction than the old vulgate inque valebunt, of which Lachmann says 'vix Latine'.

302. quantum … naturai: not 'during the time which nature allows to each', but more vaguely 'in so far as the laws of atomic motion allow growth and flourishing to each'. Some things reach a higher state of development than others.

303–7. With these lines should be compared iii. 816–18 repeated in v. 361–3:

• sicut summarum summa est aeterna, neque extra
• quis locus est quo diffugiant neque corpora sunt quae
• possint incidere et valida dissolvere plaga.

There, however, the idea is different from that of the present passage in two respects; Lucr. is there thinking of the destruction of the universe, not of the permanence of matter, and consequently the entrance of atoms from without is represented as a blow which might destroy things and not a force which might change them.

303. rerum summam: not, as it sometimes is, the universe, but, as in i. 1008, the 'sum total of things', i.e. matter and the things which the movements of matter create. See n. on i. 235.

commutare: notice here the idea of 'change' which is developed in 306–7; see introductory note.

304. genus ullum materiai: 'any class of matter' is a strange expression, though no commentator appears to have noticed it. It apparently suggests that some classes of atomic nuclei or concilia or even of independent atoms would be more likely to 'escape' than others. Or is it merely a loose expression for pars ulla?

305. omni … omne: 'the universe', atoms and void together.

〈extra〉 is Munro's supplement for a defective line, which is suggested both by i. 963 nunc extra summam quoniam nil esse fatendum and by iii. 816 (see n. on 302–7) neque extra quis locus est. The sense too requires it and it is far better than other proposals: nequerursusin omne Marullus, which is weak and not improved by Diels's nequerursusin omnest; neque in omneseorsum⟩ Lachmann; ⟨immensoex omni Polle; all the words suggested are otiose.

306. nova vis appears, as Giussani says, to have the double sense of 'supply' (cf. vis materiai ii. 544) and 'force'. The entrance of new atoms from without the universe would at once be an addition to matter and also, as explained above, a new force which would pg 855change the motions of things. In English 'forces' would perhaps represent it. If a choice must be made between the two meanings, probably that of 'force' should be chosen, as this is the main idea of these lines and it would correspond to ulla … vis in 303.

irrumpere: intrans.; cf. erumpere v. 596 and 952, and see Prol. V b, § 14 (b).

307. 'Change the nature of things and deflect the motions': the actual order would be the reverse: the nova vis would alter the atomic motions and so change the nature of created things.

## 2. Atomic movement within a body at rest. 308–32.

308. We need not wonder why, when all the atoms are in motion, yet the whole seems to be at rest, except when some individual body moves. 312. The atoms are far below the ken of our senses and therefore hide their motions from us too. 315. Even among visible things distant objects often conceal their movements. A flock of sheep grazing on the hill-side move forward as they feed and the lambs play and butt; yet from a distance all this is blurred—a spot of white on a green hill. 323. So when legions are manœuvring in the plains and the sheen of their arms rises to heaven and gleams on the ground, the earth trembles with their tread and the hills echo to their shouts, while the cavalry wheel around and cross the fields shaking the ground with their rush; 331 yet there is a place high up on the mountains whence all seems to be at rest—a bright spot on the plain.

After the completion of the account of atomic motion Lucr. thinks well to answer a possible objection, that in spite of the incessant atomic motion which he has described the world as a whole seems to be at rest, except when some individual body moves. He answers it without much difficulty, pointing out that the atoms are far below the ken of the senses and that it is therefore impossible to see their motions, any more than it is sometimes possible to detect the movements of individuals in a crowd seen from a distance, indeed the crowd itself seems but a spot of colour at rest; see Prol. IV, § 7. This gives rise to two beautiful Lucretian pictures of the flock of sheep and the army, and he brings out in them both the extent of the real motion and the apparent rest. The passage goes closely in thought with the illustration from the motes in the sunbeam (112–41), and Giussani places these lines after 141. The transference rests again on a modern desire for logical sequence in Lucr.'s thought; the lines would fit well no doubt in that position, but, as far as we can tell, the idea occurred to Lucr. at this point in his writing. Possibly in a revision of the poem he might have transferred it himself, but it is not the business of the modern critic to reconstruct the poem for him.

pg 856

309 is a halting line and the conclusion sint in motu is more awkwardly spondaic than any other line in the poem. Lachmann does not ease it by suggesting that in motu may be regarded as a single word.

310. 'Yet the whole seems to stand wholly at rest': summa … summā, a play on words, such as Lucr. loves; see n. on i. 336 and Prol. VII, § 25. summa here not quite 'the world', but the whole scene within our view.

311. proprio … corpore: 'with its individual body', i.e. the body of a whole compound thing.

dat … motus here = movetur, in i. 819 = movet; cf. iii. 560 vitalis edere motus.

312. longe … infra: lit. 'far below beginning from our senses', i.e. if the point where our senses ceased to be able to perceive be taken as starting-point, one would have to go far down in the scale before one came to the atoms. For the idea cf. i. 600, ii. 128, and 138–9, and for the expression iv. 111 primordia tantum sunt infra nostros sensus.…

314. cernere iam nequeas: 'you can no longer see them', cf. i. 600 quod nostri cernere sensus iam nequeunt.

nequeas: idiomatic 2nd pers. subj.; see n. on i. 327.

surpere = surripere, a contraction used in other parts of the verb both by Plautus and by Horace; cf. surgere and see examples quoted by Ernout. It looks as if it were a popular form.

316. spatio … locorum: 'the distance between the places'.

diducta: 'separated', i.e. from the spectator.

317. nam introduces the example as in ii. 352 nam saepe ante deum delubra.…

laeta: 'rich', 'fruitful'; see i. 14 n.

318. lanigerae: a typical Lucretian compound with the root ger-. So again i. 887, ii. 662, v. 866, vi. 1237. Cf. squamigeri i. 372, 378.

reptant: the editors notice that this form well suggests the slow advance of the flock as it grazes, as does the rhythm of. 317.

319. gemmantes rore recenti: 'bejewelled with fresh dew', a fine Lucretian picture; so again v. 461 gemmantis rore per herbas.

320. coruscant: 'butt'; in Juv. xii. 6 it is transitive frontem-que coruscat. Notice that the lambs give the quicker and more violent motion as opposed to the slow movement of the grazing sheep.

322. velut: Lachmann for veluti OQ on the ground that veluti and uti are never elided.

323–32. The second example, that of the manœuvring legions, is worked out with even greater elaboration. It has clear reminiscences of the description in ii. 41–3 and the scene may be imagined as taking place again in the Campus Martius. If so, the montes of 327 and 331 will presumably be the hills of Rome. But this notion ought not to be pressed, and camporum 324 (campi 40) is slightly against it.

pg 857

324. belli simulacra cientes: as in ii. 41.

325–7. A Homeric reminiscence: Il. xix. 362 αἴγλη δ‎ʼ οὐρανὸν ἷκε, γέλασσε δὲ πᾶσα περὶ χθὼν χαλκοῦ ὑπὸ στεροπῆς‎· ὑπὸ δὲ κτύπος ὤρνυτο ποσσὶν ἀνδρῶν‎. Cf. Il. ii. 457 ὡς τῶν ἐρχομένων ἀπὸ χαλκοῦ θεσπεσίοιο αἴγλη παμφανόωσα δι‎ʼ αἰθέρος οὐρανὸν ἷκεν‎, and Od. xiv. 267 πλῆτο δὲ πᾶν πέδιον πεζῶν τε καὶ ἵππων χαλκοῦ τε στεροπῆς‎. Imitated again by Virgil, Georg. ii. 281 fluctuat omnis aere renidenti tellus.

325. ubi: OQ, which may be retained with Diels and Martin, a semicolon being placed at campos 330. It is then a Lucretian anacoluthon, cf. 342–7. Marullus's correction ibi was accepted by previous modern editors: it mends the construction, but is feeble in itself, and the description should run on.

326. subter: adv. 'beneath' sc. their feet.

virum vi: cf. ii. 264 equorum vim and ii. 42 with Munro's emendation ecum vi.

327. excitur: for the 4th conj. form see Prol. V a, § 12.

clamore-que: -que attached to ĕ cf. i. 134 n.

328. sidera mundi: cf. i. 788; the expression is rather exaggerated here.

329–30. Notice this careful description of the cavalry: they are skirmishing about on the edges first (circumvolitant) and then they charge through the middle of the ground.

330. tramittunt: O. There seems no sufficient reason to alter to the more normal transmittunt Q: the MSS. vary in the compounds of trans. The acc. campos is due to tra-, not to mittunt.

impete: Lucr. uses this ablatival form 13 times and the gen. impetis once in vi. 327. But impetus always in the nom.; and impetibus abl. in i. 293, which may come from either form. Ernout thinks that the forms are invented on the analogy of praepes on scansional grounds. See Prol. V a, § 2.

331.unde〉: a necessary and obvious addition of Niccoli.

332. fulgor: 'a sheen'. Cf. candor 322. Giussani notes that it corresponds as the last word of the description with fulgor at the beginning of 325.

## B. The Variety of Atomic Shapes and Their Effects. 333–729.

This second main section of the book is long and rambling and contains several digressions. It deals with the variety of shapes (σχήματα‎) in the atoms and the results of this variety in compound bodies, especially the effects on sensation. It corresponds roughly to a paragraph in the Letter to Herodotus1 in which Epicurus pg 858argues that there must be an 'incomprehensible number of' (ἀπερίληπτα‎) variations in the shapes of the atoms, for it would not be possible to account for the great variety of things, if they were derived from atoms having a limited number of shapes. On the other hand, the varieties of atomic shape cannot be infinite (ἄπειρα‎) because, as he argues1 later on, after a point variety of shape can only be obtained by increase in size and, if such varieties were infinite, some atoms would be so large as to be visible. Yet the number of atoms of each shape is infinite. These principles are explicitly dealt with by Lucr. in ii. 478–521 and 522–80, but they are in his mind throughout the section, e.g. 333–41. They are a tacit protest against Democritus, who had held that the variety of atomic shapes was infinite and was quite prepared to say that some atoms were 'very large'. On the other hand, the whole section owes a great deal to Democritus,2 who had worked out with great elaboration the effects on sensation of the varieties of shape in the atoms. With these conceptions Lucr. deals principally in 381–477.

## (a) The variety of shapes and their effects on sensation. 333–477.

### 1. The atoms vary in shape. 333–80.

333. The atoms have an immense number of different shapes. Of any given shape there are an infinite number of atoms, but not all atoms are absolutely alike.

338. (1) Seeing that the atoms are infinite in number, they are bound not all to be alike in shape and form. 342. (2) Again, take the human race, or beasts, or fishes, or birds: even they within their species differ individually one from another. 349. Otherwise a mother could not know her child or the child its mother, as we see they do. 352. When a calf has been sacrificed at the altar, its mother will search for it everywhere, filling the air with her laments and often returning to the stalls in her quest. 361. Fields and streams she loves cannot distract her mind or solve her care, and no other calves can satisfy her as a substitute for her own. 367. Again, kids and lambs know their mothers and run to drink their milk. 371. There are differences between one grain and another and between the shells on the seashore. 377. In the same way the atoms, since they are formed by nature and not made by hand to one fixed form, must vary in their shapes.

In this paragraph Lucr. enunciates the general principle that all pg 859atoms are not alike, but that there are many varieties of shape and many atoms of each variety. This contention is supported by two arguments, the first an inference from the infinite number of atoms, the second an appeal to experience in the difference of individuals within species. His arguments are not as happy as usual: the first is merely dogmatic: it would be quite possible for all the infinite atoms to be the same in shape; Epicurus' argument (see introductory note) that if they were, they could not produce the immense variety in things is far more cogent. The second argument—an appeal to the analogy of experience—if taken strictly, proves, as Giussani points out, too much. For the natural deduction would be that each individual atom within the various species of shape would differ slightly from every other, which is by no means what Lucr. wishes to show. The paragraph must stand as a general argument for immense variety of shape; and it has certainly given us some of Lucr.'s finest descriptive lines. It may also be noted that the difference in individual animals and things is at once an illustration and also the outcome of the difference of shape in the atoms. In this respect the paragraph is like the description of the motes in the sunbeam (ii. 112–41).

333. exordia rerum: a 'scansional' variation of primordia as in iii. 31 and iv. 45 (where cunctarum … formis is repeated with variis for longe), iii. 380, etc. But it is not a technical term and is used elsewhere in the more general sense of 'beginnings': i. 149, ii. 1062, v. 331, 430, 677.

334. Giussani and Diels bracket this line and hold that it is interpolated from iii. 32 (= iv. 46). They note that it is otiose here, that qualia sint, appropriate in the summary in the other two places, is inapposite here, as it describes the contents of this section, and that the repeated quam (334, 335) in asyndeton is grammatically awkward. Giussani thinks that the poet may have noted it in the margin after writing the later passages. But, though it is a little clumsy, it should be retained, the variation longe) (variis being an indication of genuineness.

qualia sint is explained in what follows, and the asyndeton is not difficult for Lucr.

335. multigenis: found only here; but Lucr. is fond of such formations. Cf. i. 865, etc. alienigenis, ii. 741 caecigeni, 1106 primigenum, v. 437 omnigenis, though in the last instance the text is doubtful; see n. there.

336–7. 'Not that only a few are endowed with a like form, but that they are not all universally alike one to another'; this is a general enunciation of what Lucr. explains in detail in ii. 478–521 and 522–68, that there are not an infinite but only an immense number of varieties of shape in the atoms and that within each shape the number of atoms is infinite.

non quo … sint, sed quia … pg 860constant: this is the regular construction, to which Munro quotes parallels from many authors, non quo taking the subjunctive when the statement is denied both as a reason and as a fact, the indic. being of course used in the affirmative clause. For Lucr.'s usage see ii. 3 n. These lines are repeated as they stand in ii. 723–4 and with the necessary variation in ii. 692–4.

multa … parum = pauca, 'only a few', as in ii. 692, 723, and vi. 779, 1066.

337. vulgo: 'in general', 'everywhere'.

constant: the Itali rightly restore the plural from constat OQ; in 724 OQ have constant and in 694 O has constet and Q constant. Lachmann read constent in all three places, but there can be no possible reason for the subj. and he has not been followed by any recent editor.

338. nec mirum: without est; cf. ii. 87 n. and see Prol. V b, § 13 (a).

340. omnibus is constructed both with pari and with simili. prorsum goes with omnibus omnia, as Ernout holds, 'not absolutely all like all', not with pari, as Merrill thinks.

341. filo: filum in a transferred sense would appear to be a metaphor from weaving, not from the simple thread. Hence it is used by Lucr. in iv. 88 meaning 'texture', in v. 572, 581, 589 'size' (of the sun and moon), and here 'shape', being the equivalent of figura.

342. praeterea: OQ have praetere for which the corrector in O has made the obvious restoration. Cf. vi. 990[991]. But the editors have objected on two grounds, one the awkwardness of the anacoluthon praeterea genus humanum … quorum unum quidvis generatim sumere perge; the other that praeterea has no antecedent, such as the ordinary principio. The Juntine editor sought to mend the anacoluthon by reading horum in 347, which does not make it easier, nor does Winckelmann's praetereast. Lachmann more boldly read parturiunt for praeterea and Munro, abandoning his first conjecture praestat rem, read praeter eat 'let there pass in review', 'as at the sollemnis transvectio of the knights' (this seems odd as applied to the cattle and the fishes!). The anacoluthon is vouched for by the exactly parallel passage in iv. 123 ff. praeterea quaecumque … panaces absinthia taetra … quorum unum quidvis leviter si forte duobus.… More recent editors accept the anacoluthon, but assume a lacuna before praeterea, Giussani supposing that it contained an answer to an objection that there are many things exactly alike: 'yes, in things of man's manufacture (cf. 378–9) but not in nature', and Diels that the poet showed that all classes and individuals in nature differed in shape. But I have little doubt (in spite of the overlogical objection of Giussani) that praeterea connects the a priori argument of 338–41 with the a posteriori proof from experience. The MSS. reading should therefore be retained and neither emendations nor a lacuna are required. See Prol. V b, § 2. 3.

343. squamigerum pecudes: cf. squamigerum genus i. 162 n.: pg 861squamigerum is gen. plur. pecus is used in a wide sense. Cf. Plaut. Rud. 942 squamoso pecu and Hor. Od. i. 2. 7 omne cum Proteus pecus egit altos visere montis.

armenta: 'flocks', the tame beasts as opposed to the ferae, is Bentley's emendation for arbusta OQ, which has universally been accepted, arbusta cannot be right, because in all the rest of the passage Lucr. is speaking of birds and animals.

344. laetantia: 'gladdening' Munro, 'pleasing' Merrill, is probably a mere synonym for laeta, unless it can mean 'luxuriant'. Cf. laetas segetes Virg. Georg. i. 1. Note the repetition laeta … laetantia.

345. concelebrant: 'fill', 'haunt'. Cf. i. 4 n.

346. pervulgant: 'throng'. Cf. ii. 164 n.

347. 'Go on taking each one of these class by class.' quidvis Lachmann's correction for quodvis; cf. iv. 126.

sumere perge of the continuous action of working through the series. For generatim see i. 20 n.

348. tamen refers to the general sense of the previous line: 'though you take them class by class, you will yet find that individuals differ'.

350–1. quod posse … cluere: 'but we see that they can and that all these things are known apart from one another no less than men'.

351. minus atque: constructed on the analogy of aliter atque for the more strictly correct minus quam; cf. Virg. Aen. iii. 561 hand minus ac iussi faciunt.

nota: the neut. plur. refers not to the mater and the proles of 349–50, but more generally to the plural implied in unum quidvis 347.

cluere: here not quite equivalent to esse but 'appear', 'show themselves to be'.

352–66. This famous description of the cow looking for her lost calf is Lucr. at his best. It shows how the picture in his mind gets the better of him, since much of the detail is irrelevant to the argument, the point of which is contained in 364–6; no other calf could satisfy the mother. It also unfortunately contains several difficult problems of text and interpretation. Ovid imitated the passage closely in Fasti iv. 459 ff. and, as his lines have been used in the criticism of Lucr.'s lines, they must be quoted in full:

• ut vitulo mugit sua mater ab ubere rapto,
• et quaerit fetus per nemus omne suos …
• 463inde puellaris nacta est (Ceres) vestigia plantae,
• et pressam noto pondere vidit humum …
• 481quacumque ingreditur, miseris loca cuncta querellis
• implet.

There are also reminiscences of the passage in Virg. Ecl. viii. 85 ff., Georg. iii. 520 ff., Tib. i. 1. 31, and Stat. Theb. vi. 189 ff.

352. nam saepe: introducing the example, as in ii. 317.

ante pg 862deum … delubra: ritually correct, the sacrifice was made in front of but not in the shrine. Servius on Aen. ii. 225 says that some authorities define delubrum as locus ante templum.

decora: sc. probably with sculpture.

353. turicremas: a fine compound adj. used only here by Lucr. but imitated by Virgil in Aen. iv. 453 turicremis cum dona imponeret aris.

354. Virg. Aen. ix. 414 volvitur ille vomens calidum de pectore flumen.

sanguinis here of 'gore'; see 195 n.

356. quaerit: there is a serious corruption here, O having non quit, Q oinquit, and G oinquid. The corrector of Q read linquit which is feeble, as the vestigia must be those of the calf not of the mother; the same objection holds against Polle's urguet and Merrill's concit. The required sense is better given by Lachmann's noscit or Brieger's novit, but they are palaeographically very far from the MSS. Mooney (C.R. xxi. 171), comparing Ovid's line inde puellaris nacta est vestigia plantae, proposed nancit (nanquit) in the sense of nanciscitur, comparing a fut. form nanciam found in a fragment of Gaius Gracchus; this suggestion is adopted by Martin, but the word is too odd. Still odder, though palaeographically more probable, is Diels's tonguit which he supports by the statement of Festus (p. 489 Lindsay) that tongere was used in the sense of noscere, as in Ennius' phrase alii rhetorica tongent. I suggested in C.R. xvi. 330 that quaerit might be the solution and I still think it more probable than the other suggestions. The sense requires 'looks for' rather than 'finds' or 'recognizes', as in the second line of the Fasti passage: et quaerit fetus per nemus omne suos; so also in Virgil's imitation of this passage Ecl. viii. 85 cum fessa iuvencum per nemora atque altos quaerendo bucula lucos … procumbit. An abbreviation or a blot might have produced quit and then non might have been inserted either to complete the metre or more probably as a note that quit was wrong; cf. Cic. Phil. vi. 1. 3 hodierno autem dies non est … i.e. dies should be die; cf. note on itastạṭuas ii. 42 [43]. The readings of Q and G would then represent further corruptions.

357. convisens: 'searching', almost 'scrutinizing'. Cf. i. 145 res quibus occultas penitus convisere possis.

358. completque querellis: cf. Virg. Aen. viii. 215 discessu mugire boves atque omne querellis impleri nemus.

359. frondiferum: a compound already found in Naevius; see i. 18, 256, and Prol. VII, § 1.

adsistens: 'taking up her stand', rather more than merely 'stopping'. Cf. i. 965 nec refert quibus adsistas regionibus eius. This is the obvious correction of adsittens OQ, and there seems no good reason to reject it. Munro prefers absistens 'desisting from the search', which the cow clearly does not do, Brieger nemu' subsistens, and Lachmann assiduis (written pg 863adsidueis) which is unnecessarily far away. Everett's adsitiens shows a lack of humour.

crebra: neut. plur. adverbially rather than nom. fem.

revisit ad stabulum: for the construction cf. v. 636 ad hanc quia signa revisunt and vi. 1239 fugitabant visere ad aegros.

360. perfixa: 'pierced', 'stabbed'; cf. iii. 305 telis perfixa pavoris and vi. 392 pectore perfixo. It is a peculiarly Lucretian word.

361. vigentes: Q has vientes and Macrobius quotes it as virentes: but vigere is a favourite word of Lucr.'s in many senses.

362. illa: 'the streams she knows so well'. O1 has ulla, which was the old vulgate, but it is much weaker in sense.

summis labentia ripis: cf. Hor. Epod. ii. 25 labuntur altis interim ripis aquae and Od. i. 2. 18 vagus et sinistra labitur ripa.

363. subitamque avertere curam: the editors object that the cow has for long been looking for her calf and her care can hardly be described as 'sudden'. Of emendations Lachmann's solitam is to say the least unfortunate, and so was Munro's suggestion sumptam; the cow did not voluntarily take this care upon herself. Bernays's dubiam 'anxious' is palaeographically unsatisfactory, and the best emendation, if one is to be made, is Brieger's subito: the cow cannot 'of a sudden' turn aside her care. Munro kept subitam and explained it as equivalent to quae subiit, the care 'which has come upon her', pointing to the similar use of the pass. participle of compounds of ire, praeteritus in common use, obitus quoted from Laevius, and peritus from Claud. Quadr. Lucr. himself has similar uses, such as senectus 'old' iii. 272, v. 886, 896, and impensa 'overhanging' vi. 491. But though this is no doubt the step by which subitus came to mean 'sudden', it is impossible that in this one place in Latin literature Lucr. meant it to be understood in its earlier sense. It may surely be retained in the sense of the 'sudden recurrent pang'. The cow is tempted for the moment by the grass and the streams to forget her loss and then 'suddenly' it comes over her again (see C.R. xvi. 330).

364. vitulorum aliae species is not a mere periphrasis for alii vituli, but calls attention to the important point of their shape and appearance; the cow knows that her own calf is different from them all. This is the leading point in the description.

365. derivare: lit. 'to lead it into other channels'. Ernout well quotes Ter. Phormio 323 in me omnem iram derivem senis.

curaque levare: 'to relieve of her care'; the same construction in Hor. Sat. ii. 5. 99 curaque levarit, though the direct acc. is used elsewhere.

366. proprium: 'special to herself' or 'special to the calf'; reinforced by notum.

367–70. Lucr. follows up the example of the cow knowing her calf with other examples of offspring knowing their mother.

pg 864

368. cornigeras: another compound adj. Cf. squamigerum 343.

petulci: 'butting'. Cf. Virg. Georg. iv. 10 oves haedique petulci.

369. balantum pecudes: cf. vi. 1132 pigris balantibus and Virg. Georg. i. 272 balantumque gregem. Q has balatum which is supported by Itali. It could be taken with agni … petulci as gen. 'the bleating of the lamb', but pecudes = 'the mothers' would be strange, the change to the 'mothers' would be inconsistent with 370 and with the general tenor of these lines; see n. above on 367–70.

quod natura reposcit: sc. 'as nature demands', just as in ii. 17 and 23 natura is used for 'human nature'; see notes there; cf. also ii. 945 n.

370. ferē: this is its normal quantity in classical Latin; see Prol. VI, § 12 (b). fere occurs 7 times in Lucr. and its meaning is elusive. In iv. 935 fere res omnes and vi 288 tota fere … tempestas it is limitative 'almost all'; so in iv. 962 quisque fere 'almost everybody': in v. 1414, vi. 683, and vi. 1197 it is 'usually', and here it almost means 'always'.

ubera lactis: here lit. 'udders of milk', in i. 887 ubere lactis more metaphorically 'richness of milk'.

371. quodvis frumentum: the construction is not clear, but it is better to take it as a 'suspended' nom. (or acc.), 'in the case of any kind of cereal', rather than to 'supply' sumere perge from 347.

372. quidque suo genere: 'each in its own class'. Lachmann, thinking that quidque was awkward after quodvis, read quique, the archaic form of quoque abl. The attraction would then be like that in Virg. Ecl. vii. 54 sua quaeque sub arbore poma, and other instances collected by Lachmann and Munro. But if quodvis frumentum be taken separately (see n. on 371), the difficulty does not arise, and, as Ernout points out, in all cases where the attraction takes place suus, etc., precedes quisque, etc., and where it does not, the normal grammatical construction is used. I agree with Ernout and Martin in retaining quidque (Ernout writes quicque).

373. distantia: 'difference', as in ii. 525 and iv. 636. ii. 525 distantia … formarum suggests that formis is here a 'Lucretian' dative; see Prol. V b, § 5. 1.

374–6. Finally Lucr. turns to inanimate nature and shows that even there there is a difference between one example of a species and another.

374. parili: cf. i. 1067 n. noctes parilis agitare diebus. (pari ratione occurs in ii. 516.) It does not differ in meaning from simili ratione in 377.

375. pingere telluris gremium: 'painting the bosom of the earth', a picturesque Lucretian phrase, gremium and mollibus suggest the gentle lapping of small waves.

376. pavit: 'beats on'. So Cic. De Div. ii. 17. 34 speaking of the sacred chickens in the auspices says quia … necesse est aliquid ex ore cadere et terram pavire, terripavium … dictum est. Similarly pg 865pavimentum is that which you beat on with your feet. Nonius quotes the line with lavit, which was adopted by Marullus and earlier editors.

377–80. The summing up of the paragraph: the inference is not only from the examples cited from experience, but also from the a priori reason given in 338–41. Notice that he introduces here the contrast with the manufactured articles made by man, which is a new and striking point. Giussani regards this as impossibly abrupt and therefore an argument for the lacuna after 341, in which he supposes a reference to this point. Munro argues that it proves that praeterea is impossible in 342.

379. unius ad certam formam: 'to the fixed shape of a single pattern'. Strictly of course even such manufactured objects have slight variations.

unĭus: so scanned again in iii. 1013; elsewhere 5 times unīus.

380. quaedam: 'certain species' differ from other species; Winckelmann following Wakefield read quadam which is weak.

## 2. Different effects produced by different atomic shapes (381–477).

### (α‎) Difference in capacity of penetration (381–97).

381. (1) On this theory we can see why lightning can penetrate where earthly fires cannot, because it is composed of smaller atomic shapes and so can pass through apertures which the flames of torches, etc., cannot penetrate. 388. (2) Moreover light can pass through horn but rain cannot, because its atoms are smaller. 391. (3) So wine flows easily through a sieve, but oil more slowly, because it is composed of atoms which are either larger or more hooked and intertwined, so that they cannot so easily be separated and pass through the individual holes of the sieve.

Lucr. has shown in the preceding paragraph that the atoms differ in shape. He now proceeds to consider the difference of effect which is thus caused in compounds. Here he takes first their difference in capacity to penetrate; some things can pass where others cannot, some are quicker than others in passing through. This is because compounds which pass more easily are composed of smaller atoms or atoms less intricately interwoven owing to their shape, so that they can more easily separate and pass through individually. The paragraph seems simple and straightforward, but really involves several considerable difficulties.

1. Lucr. undoubtedly speaks all through of individual atoms which differ in size or shape. He calls them figurae (385, see n. there), corpora (390)—both of which are ambiguous terms—but also elementa (393), and finally primordia (396), terms only applicable to atoms. But can Epicurus in this context—we have unfortunately pg 866no parallel passage extant—have been thinking of atoms and not rather of atomic nuclei or particles? It is impossible in the instance of the oil passing slowly through the sieve (391–7) that he can have held that each atom would have required a separate hole in the sieve to pass through, for in that case it would have been so large as to be perceptible. This would have been quite possible for Democritus, who was prepared to say that some atoms were 'large', but not for Epicurus. Or again—and this is a more serious objection—if a compound were to split up into its component atoms., it would cease to be the 'thing' any more: it would not be 'lightning' or 'oil' that passed though, but individual atoms of different sizes or shapes suited to make lightning or oil. I agree with Giussani, who has discussed this question in his notes and in Stud. Lucr. 78–84, that there is confusion in Lucr.'s description here, and that he probably misunderstood—not for the only time—his Epicurean authority.

2. Though the main subject of this section is the shape of the atoms, Lucr. introduces in each example as an alternative reason, their size (385, 389, 393), and indeed in the second instance—that of the light from the lantern—it is given as the only cause. Robin in a learned note on 385 points out that to speak of size as separate from shape in this way is a contradiction of the true doctrine of Atomism, which regards size and shape as inseparable. I do not feel this difficulty so strongly, but both it and the introduction of size at all will disappear if it is realized that Lucr. here should have been speaking not of atoms but of atomic nuclei, whose size is determined by the number and shape and density of the component atoms.

3. The description in this and many other places of the atoms as 'hooked' or 'closely interlaced' (394) raises the question how such close mechanical connexion between atoms can be consistent with Epicurus' doctrine of their constant movement even inside compounds at atomic speed from collision to collision. If they are interwoven and fastened together by 'hooks' and 'eyes', such free motion is impossible. It might perhaps be argued that even when fastened together, they retain a certain vibration (παλμός‎) or 'jiggling motion', if one may so express it. Again the conception was easy for Democritus, who thought that in compounds the atoms lost their atomic motion and adopted that of the compound. Here I suspect that Epicurus took over the ideas and phraseology of the Atomists without sufficient consideration of their applicability to his own theories. There are similar lapses in his astronomy.

4. This paragraph follows rather abruptly on the preceding paragraph, and one would have expected some general introductory lines in which Lucr. explained that he was about to consider the pg 867effects of the differences in atomic shape before he proceeded to give examples. Brieger and Giussani accordingly assumed a lacuna. But in the unfinished state of the poem it is safer not to suppose that a passage has been lost, but to imagine that Lucr. wrote the series of examples as they occurred to him and might have added introductory verses in revision. He hurries on as the mental pictures occur to him.

381. animi ratione: 'by reasoning of the mind' as opposed to the evidence of the senses, which could not give information on this point. Cf. i. 425, 448, iv. 384. Editors thinking the expression queer in this place and, wishing to supply a connexion with the preceding paragraph, have emended: Lambinus 〈iam〉 animi (perhaps the best correction), tali Lachmann, parili Bernays (cf. 374), simili (cf. 377) Shackle (C.R. xxxv. 156). Ernout and Martin retain animi, rightly: see introductory n., § 4.

exsolvere: 'to explain', here only in this sense, but cf. v. 773 qua fieri quicquid posset ratione resolvi.

382. penetralior: cf. i. 494 penetrale frigus, 535 penetralem ignem; but here only in the comparative.

383. fluat: with Ernout and Martin I retain the MSS. fluat, fluere is regularly used by Lucr. of emanations: of cold iv. 260, of smell iv. 218, vi. 924, etc., of the simulacra of vision iv. 144, etc. Most editors adopt the ingenious but unnecessary emendation of Faber fuat, a form which Lucr. uses in iv. 637.

385. magis goes with parvis 'smaller'; cf. 393 maioribus est elementis.

figuris is not quite a synonym for atoms, but rather 'atoms considered from the point of view of their shape'. Cf. vi. 770 in terra cuiusque modi rerum esse figuras. See Prol. VII, § 10 (a).

386. transire foramina: 'to pass through holes' or 'pores'. Here Robin thinks that there is a reference to the idea of emanations fitting into pores of the right size (πόροι σύμμετροι‎) which was probably invented by Alcmaeon, and plays so large a part in the theories of Empedocles. In view of 397, where the foramina are undoubtedly the holes in the sieve, it seems simpler to take it in a less technical sense.

387. hǐc: as in ii. 1066, iv. 921, vi. 9. The short i is regular in early Latin, and Lucr.'s usage may be regarded as an archaism: see Prol. VI, § 22.

ortus … creatus is regarded by Lachmann as 'inelegant' and he therefore reads ortu as he does again in vi. 909, but he has not been followed; the abl. would surely be still more awkward.

388. cornum: 'a horn lantern' as in Plaut. Amph. 341 tu qui Volcanum in cornu conclusum geris. The picture here is of a man taking a lantern out into the rain. The form cornum as from a neuter of 2nd decl.: Lucr. has the normal forms cornua in ii. 619, v. 1034, and cornibus v. 1325. For such variations in declension see pg 868i. 191 n. and Prol. V a, § 2. Robin has noticed a close parallel to these lines in Empedocles (Diels B. 84. 1–6).

389. quare? nisi: this rhetorical interjection is unparalleled in Lucr.

390. corpora: 'atoms', but the idea is much more intelligible of atomic nuclei (see introductory n., § 1).

almus: Q1 for alimus OQ. Marullus's amnis is ingenious but could not be right.

391. quamvis = quantumvis, 'as fast as you like'.

vina: the plural occurs again in vi. 231, there, as here, metri gratia.

392. Note the contrast between the rapid dactyl and the spondaic slowness of the rest of the line. See Prol. VI, § 7.

393. maioribus … elementis must be 'larger atoms' (see introductory n., § 1). Robin in view of the Atomic doctrine that shape and size are inseparable (see introductory n., § 2) would suppose that the larger atoms here must be of a different shape: e.g. oil is composed either of more hooked and interwoven atoms or of larger round and smooth atoms. But I doubt if Lucr. was troubled by such subtleties.

394. hamatis: 'hooked' so that they would fit into the corresponding 'eyes' of other atoms; Democritus'ἀγκιστρώδη‎. Giussani partly on the ground of a statement in Plut. Plac. i. 3. 27 (Us. 270) that Democritus did not recognize such forms in the atoms as they would be brittle (εὔθραυστα‎), but more to economize the number of shapes, thinks they were only 'branched'. But 'branched' atoms are separately described in 428.

perque plicatis: -que is postponed, as inter se must go with per … plicatis not with hamatis; cf. ii. 1050 res ipsaque, vi. 1085 illa huiusque.

395–7. 'And so it happens that the atoms cannot be so suddenly separated from one another and severally pass through the several holes of each.' But what is the meaning of cuiusque? Giussani takes it to mean 'of each atom', as though each atom had a special hole of its own in the sieve. But this is a very strange idea, and it surely means 'of each sieve', i.e. of the sieve on each occasion when the experiment is tried. This I take to be Munro's interpretation when he says 'the several openings of any particular strainer'. Bruno conjectured coli usque to get a clearer reference; but it is difficult to assign any real meaning to usque.

### (β‎) 1. Differences to taste. 398–407.

398. Honey and milk are pleasant to the taste, but wormwood and centaury twist the mouth awry. 402. Things which are pleasing to taste are made of smooth and round atoms, while the bitter and rough are closely wrought of hooked atoms and so tear their way into our senses and burst into our body.

Lucr. proceeds to consider differences made to sensation by the pg 869shapes of the atoms in compound bodies. All this doctrine is found in the Atomic School and details were worked out at great length by Democritus. Lucr. considers first differences of taste and reaches the general conclusion that sweet things are composed of smooth and round atoms, bitter and unpleasant things of hooked or jagged atoms. There is no reason here to think that there is confusion with atomic nuclei, for though it would be those which would actually come into contact with the tongue and palate, yet the shape of the atoms composing them would give them their character and contour. The whole passage should be compared with iv. 615–72 where Lucr. deals again and at greater length with the sense of taste.

399. tractentur: 'are turned', 'rolled' in the mouth. Cf. iv. 623 suaviter omnia tractant umida linguai circum sudantia templa.

400. taetra absinthi natura: cf. i. 936 (= iv. 11), iv. 124 absinthia taetra.

401. centauri: cf. iv. 125 tristia centaurea. The present form seems to come from a nom. centaurium, formed on the analogy of absinthium, which is Greek.

pertorquent: plur. of the two naturae; cf. iii. 558 corporis atque animi vivata potestas inter se coniuncta valent; or because the real subject is absinthium et centaurium. Lachmann points out that strict grammar after huc accedit uti would suggest pertorqueat, but Lucretian usage would require the plural. pertorquere occurs here alone, the simple torquere being used by Virgil in his imitation Georg. ii. 246 ora tristia temptantum sensu torquebit amaror.

402. rutundis appears to be the invariable spelling in O and most frequently in Q.

404. qu$ae⌣$ amara: Lucr. not infrequently shortens a long vowel or a diphthong in hiatus in thesis: ii. 617 quĭ in or as, iv. 1061 sĭ abest, v. 7 sĭ ut ipsa, v. 74 quĭ in orbi, vi. 716 quĭ etesi$ae⌣$ esse, vi. 796 sĭ odoratast. Correspondingly, with words ending in -m, instead of elision he scans the syllable short in ii. 681 cŭm odore, iii. 394 quăm in his, iii. 1082 dŭm abest, vi. 276 cŭm eo. The usage is frequent in Plautus and occurs in Lucilius. By Virgil it is used sparingly and usually with Greek words, e.g. Ecl. ii. 65 ǒ Alexi, Ecl. vi. 44 Hylǎ omne sonaret. Horace's si mĕ amas (Sat. i. 9. 38), and cocto nǔm adest (Sat. ii. 2. 28) are more like Lucr.'s instances. See Prol. VI, § 13.

406. rescindere: 'to tear open'; cf. Virg. Georg. iii. 453 rescindere … ulceris os.

nostris sensibus: dat. 'of disadvantage'; sensibus almost equivalent to 'sense-organs'. Cf. i. 303 n.

407. perrumpere: 'to burst apart': a violent word used again in iv. 1148 validos Veneris perrumpere nodos.

### (β‎) 2. Differences to the other senses: summary. 408–43.

408. Moreover, all differences of pleasant and unpleasant sensation pg 870are due to differences of shape. 410. (1) Sound. The grating sound of the saw is due to atoms less smooth than those of the tunes of musicians. 414. (2) Smell. The atoms which enter the nostrils from burning carcasses are unlike those from a saffron-strewn stage and an incense-burning altar. 418. (3) Colour. Beautiful colours which delight the eyes have atoms of different shape from those of foul colours which prick the eyes and cause tears. 422. In general every form which soothes the senses contains smooth atoms, one which is harsh and painful has rough atoms. 426. There are some atoms which are neither smooth nor hooked, but have projecting angles; these tickle the senses rather than hurt them; in this class we may place wine lees and elecampane. 431. That fire and frost, which both prick the senses, have different jagged conformations, is clear from the touch of each of them. 434. For indeed it is touch which is bodily sensation, both when an external body enters it, or when something within hurts or gives pleasure, or when a body remaining outside disturbs the seeds within and causes confusion to the senses, as you can prove by striking any part of the body with your hand. 442. Therefore the forms of the atoms must differ widely, when they can produce such different sensations.

This paragraph divides itself naturally into two parts. In the first (408–21) Lucr. applies to the other senses what he has said in the preceding section about taste; pleasant sensations are due to smoothness of the atoms, unpleasant to roughness. It must be remembered, however, that there is a difference in the process of these sensations. Taste is effected by the direct contact of the tongue and palate with the object (see iv. 615–72). With the other senses the connexion is indirect and is accomplished by means of effluences coming from the object to the percipient. This Lucr. explains at length in the Fourth Book: sight (46–268), sound (524–614), smell (673–705). It is therefore strictly speaking the effluence, with which the sense-organ has contact, which must be composed of atoms of varying shape, but as it consists of particles coming off from the object, it is evidence of the conformation of the whole object.

In the second part (422–43) Lucr. sums up what he has said of the various senses in a general conclusion. All pleasant sensation is due to smooth atomic shapes, all unpleasant sensation to rough atomic shapes. He adds as a kind of appendix that there are intermediate shapes, not 'hooked' but 'angular', which produce a half-pleasing tickling sensation: for his illustration he returns here to taste. Passing to a more direct sensation than any he has hitherto dealt with he notices that the immediate sensations of hot and cold both imply a spiky (dentata) shape, but the shapes must differ in pg 871hot bodies from those in cold. This mention of direct contact leads him to insist on the supreme importance of touch for sensations: direct in the case of hot and cold and the like, direct too but specialized in taste, indirect and effected through the effluences in the other senses; see Prol. IV, § 3. Touch may be due to three causes, the entry of an external body, disturbance arising within, or disturbance caused by an external body which remains outside. In the last two lines he returns to the general idea of variety in atomic shape. The most important point in this part, which goes right back to Democritus, is the insistence on the necessity of touch for sensation, a necessity inherent in any purely material system.

408. postremo, which usually introduces the last of a series, seems a little strange here. It introduces the second half of the section on the effect of difference of atomic shape on the senses (β‎ 2). It is not, as editors have thought, inconsistent with denique (444), as that introduces the third section of the whole discussion of the effects of atomic shapes. It is therefore unnecessary with Giussani to suppose that it should be taken in a 'generalizing' sense, 'indeed', as denique is sometimes used.

bona … tactu: 'good and bad to the senses in touch'; sensibus is dat. and tactu abl. (or perhaps, as Giussani takes it, the supine, if indeed there is any real difference between abl. of tactus subst. and abl. of supine). This is better than to construe with Munro, bona sensibus et mala tactu, regarding tactu as dat. Cf. Ennius, Ann. 240 malaque et bona dictu and Virgil's imitation Georg. iii. 416 mala tactu vipera. tactu anticipates 434–41.

409. perfecta: not simply = adfecta (Giussani) but rather 'formed', 'shaped'.

figura: not the shape of the thing, but of the atoms which compose it, or of the atomic nuclei which constitute the effluence; there is a certain brachylogy here.

411. horrorem: the quality which makes your flesh creep or your hair stand on end: stridorem qui facit horrescere pilos, as Ernout says.

elementis: definitely 'atoms', not 'particles', as in 393.

412. Notice the composition of the line almost entirely of Greek words; cf. ii. 505 et cycnea mele Phoebeaque daedala chordis. Munro happily cites Quint. xii. 10. 33 itaque tanto est sermo Graecus Latino iucundior ut nostri poetae, quotiens dulce carmen esse voluerint, illorum id nominibus exornent. But, as Ernout points out, music was a Greek art, and there was very little vocabulary for it in Latin. For musaea cf. i. 934 (=iv. 9) musaeo contingens cuncta lepore and i. 947 quasi musaeo dulci contingere meile; for mele cf. ii. 505 and for organici iii. 132 ad organicos alto delatum Heliconi. For Lucr.'suse of Greek words see Prol. VII, § 9.

413. expergefacta figurant: 'rouse and shape'; for expergefacta cf. Hor. Od. ii. 10. 18 cithara tacentem suscitat Musam. figurant certainly pg 872has reference to the effluence or simulacra of sound which the player shapes, as it were, as they come off the chords.

415. cadavera is acc. and torrent transitive, the subject being vague. The cadavera are presumably the carcasses of beasts, not human corpses being cremated.

416. et cum goes back to neu simili 413: 'different from when': cf. v. 1066, 1071, 1082, and see n. on ii. 272.

scaena … perfusa: for the sprinkling of the stage with saffron water, cf. Hor. Ep. ii. 1. 79 recte necne crocum floresque perambulet Attae fabula and Ovid, Ars Am. i. 104 nec fuerant liquido pulpita rubra croco.

Cilici: the abl. form in -i is used for metrical purposes: cf. i. 286 imbri and n. there.

recens: adverbial acc. acting as adverb. Cf. vi. 791 nocturnumque recens extinctum lumen, and see Prol. V b, § 3.

417. araque … odores: as there was no thymele in the Roman theatre, this illustration was probably taken bodily from Epicurus.

Panchaeos: 'from Panchaea', an apparently mythical island off Arabia, here first mentioned in Latin.

419. oculos … pascere: cf. i. 36 pascit amore avidos inhians in te, dea, visus. So Terence, Phorm. 85 restabat aliud nil nisi oculos pascere and Sen. Ep. 58. 25 oculos …, ut dici solet, pascit.

420. lacrimareque: for -que attached to ĕ cf. i. 134 n.

421. diri: OQ have di. Many suggestions have been made, of which the best are diri Lachmann, which has been generally accepted, and foedi, Q1, which Martin prefers; but it is a rather senseless repetition of foeda.

422. figura, the restoration of Schneidewin, for videntur OQ, obviously a dittography from the previous line, has been generally accepted by editors from Lachmann onwards; it refers to the shape of the particles which constitute the effluence, cf. 409 and figurant in 413 (see n. there). The earlier reading was that of the Juntine, causa iuvatque, but causa is not the word Lucr. would use here; Brieger prefers the weak quaeque iuvat res; both of these involve the alteration of cumque, which is certainly genuine. Nor is Martin's tibi res (cf. i. 773) attractive.

423. haud … levore: aut … leviore OQ, a further sign that the scribe's attention was wandering here.

principiali … levore: 'smoothness in the atoms', principialis as the adj. of principia; the adj. is not found outside Lucr., but is used in a different sense in v. 246 principiale … tempus 'a time of beginning'.

424. quaecumque: sc. figura.

425. materiae squalore: 'roughness in the atoms', corresponding to principiali … levore 423. For squalor in this sense cf. squalida in ii. 467, 469.

repertast: probably 'has been invented', sc. by nature, rather than 'has been discovered to be'.

426. iam: 'when you come to them in the series', as Munro ex-pg 873plains. Ernout would take it with nec but the order is against it: Lucr. could have written nec iam.

427. flexis mucronibus unca: 'hooked with bent points': these are the hamata of ii. 394 and 405. The MSS. have uncaque: -que may be due to quae in the next line, if Martin's ut quae be accepted there.

428. angellis … prostantibus: 'with angles jutting out'. Here we have the exact description of what Giussani would take the hamata to be; see n. on 394.

⟨ut quae⟩: Martin's proposal for filling the vacant foot in the MSS. seems to me rather better than Munro's utqui (cf. i. 755); 'the sort of atoms which could' is more the sense required than 'so that they can', which is a little too teleological. Either of them is better than the other suggestions, et quae Itali, quaeque Lachmann, unde Bernays.

429. titillare: 'to tickle'. Munro aptly quotes Cic. De Nat. Deor. i. 40. 113 has leviores ducis voluptates quibus quasi titillatio (Epicuri enim hoc verbum est) adhibetur sensibus. 'Epicurus' word' was γαργαλίζειν‎ (fr. 411 Us.).

possint O must be accepted against Q's possunt, unless we read et quae or quaeque in 428.

430. fecula: 'wine lees' or 'tartar'. Cf. Hor. Sat. ii. 8. 9 faecula Coa.

inulae: 'elecampane'. Cf. Hor. Sat. ii. 2. 44 acidas … inulas. These are both 'tart' tastes rather than bitter.

432. dentata: 'toothed' or 'jagged'. The atoms composing both hot and cold are 'toothed' but in different ways. The neuter dentata covers ignis and pruina; cf. ii. 201 n. and 234.

433. tactus uterque = tactus utriusque; cf. Cic. Ad Att. xv. 1. 3 se autem utraque arma metuere and other passages cited by Munro.

434. The repetition of tactus and the exclamation lay emphasis on the supremacy of touch as the foundation of all sensation; see introductory note.

pro divum numina sancta: this appeal to the gods sounds oddly in the mouth of Lucr. the Epicurean. But it is purely conventional and poetical: as Faber says 'poetice, non philosophice'. Cf. v. 1156 etsi fallit enim divum genus humanumque.

435–9. Strictly speaking there are four possible forms of the tactus which causes sensation: (1) an external body which either remains outside or (2) enters the body, (3) an internal atomic disturbance which either remains within the body or (4) issues from it. But Lucr. has not arranged his statement so logically and it is not quite certain how it is to be taken. Munro and Giussani are, I think, right in holding that Lucr. makes three heads: (1) the external body coming in (vel cum res extera sese insinuat), (2) disturbance within the body either painful or pleasurable (vel cum laedit … per Veneris res), (3) the blow of an external body which causes internal atomic disturbance. In this case aut 438 should strictly be vel, parallel to vel in 435 and 436, but, possibly owing to the attraction of aut in 437, Lucr. substitutes aut in the last clause. Cf. Ovid, Metam. xv. 601–2 pg 874vel, si dignus erit, gravibus vincite catenis, aut finite metum fatalis morte tyranni. Merrill to save the grammar supposes two main divisions marked by vel, the second having three subdivisions (1) laedit, (2) iuvat, (3) ex offensu cum turbant, but though the last refers to internal disturbance, the external object as cause marks it off as quite different. Lachmann's atque for aut (438) with full stop at the end of 437 and comma after semina makes nonsense of the passage, which he has not understood.

435. sese insinuat: insinuare here transitive as in i. 116, vi. 860.

436. natast: 'has arisen'. Giussani notes that a stomach ache would be a good example of such painful disturbance arising within the body.

438. aut: see n. on 435–9.

ex offensu: 'through a collision' or 'blow' without, which causes internal dislocation. A blow on any limb causes pain not only in that limb. For offensus see ii. 223 n.

turbant: intrans.; see Prol. V b, § 14 (b).

439. Confunduntque⟩: on the view taken above of these lines cum in 438 must be parallel to cum in 435 and 436 and therefore Marullus's -que is necessary. With Lachmann's atque in 437, cum … semina may be taken as a temporal clause, and confundunt as the apodosis; but this view is otherwise untenable.

inter se: construe with concita; 'roused one by another'.

440. quamvis: acc. fem, of quivis, as in ii. 944.

441. ferias atque experiare: almost a hendiadys 'strike by way of experiment'; once again the idiomatic 2nd sing. subj.; see i. 327 n. and Prol. V b, § 10 (a).

442–3. The conclusion of the whole section since 398.

443. edere: 'produce', 'make manifest' the various sensations. For edere cf. ii. 816 variis formis variantis edere tactus and v. 722 variantis edere formas.

### (γ‎) Differences in texture and their effects on sensation. 444–77.

444. (1) Hard and compact things must be composed of hooked atoms closely linked: such are diamonds, which can withstand all blows, and rocks and iron and the brass sockets on doors. 451. (2) Liquid things are composed of smooth and round atoms; they are like a handful of poppy-seed whose round balls do not cling together and run easily downhill. 456. (3) Things easily dissipated, like smoke, clouds, and flame, though not entirely composed of smooth and round atoms, yet have not their atoms closely interwoven; 460 they can prick and penetrate, yet do not cling together, so that it is clear that they are made of sharp atoms, not interlaced. 464. (4) Liquid things which are yet bitter, like sea-water, have smooth round atoms inasmuch as they are liquid, yet also rough, but not hooked together: they must be round and rough so as to move easily and pg 875yet be pungent. 471. To show that in the sea rough atoms are mixed with smooth, mark how the liquid will flow away into the ground and leave the salt particles above, caught through their roughness in the ground.

Lucr. now passes to the third class of effects produced by atomic shape, those which affect the texture and density of compounds, and so produce corresponding effects on sensation. He has already shown (100–8) that in hard things the atoms are more densely packed than in such loose compounds as air and light. Now he adds that it is not merely a question of the intervals between atom and atom, but of their shape. Hard things are composed of hooked, 'branchy' atoms which can interlock and hold together more easily, liquid things of atoms which are smooth so that they cannot catch on to one another, and round so that they are individually more mobile. It is noticeable that he chooses here to place in opposition to the hard things, not the loosest textures of all, air and light, as in the earlier passage, but liquids, which are intermediate in character and have a more coherent body. He then passes on to two special cases in which atoms are mixed; evanescent things like smoke, etc., where there are some smooth atoms, some pointed, so that they can prick, but are yet not interlaced, and salt water, where again there are the smooth round atoms which produce the liquid and also the rough atoms which cause the pungent taste. This analysis he enforces by the reminder that in the salt-pan the water flows away and leaves behind the salt, so that the two kinds of atoms are segregated. The general ideas of this paragraph are simple, but it unfortunately contains several serious difficulties of text and interpretation.

444. durata is here merely equivalent to dura; cf. 449 duri robora ferri. In vi. 969 durata calore it has the full participial force of 'hardened'.

spissa = densa. Cf. vi. 127 spisso cava corpore circum.

446. ramosis: 'branchy'. so that they can easily interlock. In v. 1096 it is used in its literal sense of a tree, in vi. 133 of the 'branches' of clouds.

alte compacta: 'close-knit to its depths', not merely on the surface. For compacta cf. v. 880 ex alienigenis membris compacta, 919 compactaque membra animantum.

447. adamantina saxa: 'diamonds'. Pliny testifies to their hardness in N.H. xxxvii. 57 and uses the phrase respuentes ictus, which looks like a reminiscence of ictus contemnere sueta.

448. prima acie: 'in the front rank', a military metaphor.

ictus contemnere: cf. v. 379 validas aevi contemnere viris.

449. silices: 'blocks of stone'; cf. i. 571, v. 313. Munro thinks of the large paving-stones in a Roman street.

duri robora ferri: 'the strength of hard iron'. Cf. i. 882 robore … saxi.

450. aeraque … vociferantur: the exact sense of this phrase is pg 876not easy to determine. The Roman door consisted of two valvae, which turned on pivots (cardines) set into sockets in the floor and lintel; they might further be secured by bars (pessuli) fitting into sockets. The older editors took the line to refer to the cardo, but claustra is an unsuitable word for that kind of socket. Munro refers it to the pessuli, aera being the bolts and claustra 'staples' or 'sockets'. But claustra elsewhere, e.g. Aen. ii. 259 laxat claustra Sinon, appears to mean the bolt; aera being a quite general word may well refer to the sockets. Translate 'the bronze sockets which scream as they struggle against the bolts'.

restantia = resistentia 'struggling with'. This Munro takes to mean 'refusing to part from'; more probably perhaps 'struggling against', when the door is pushed or shaken by the wind. In any case cf. i. no ratio nulla est restandi.

451. e: Lachmann for ex OQ, as the MSS. never elsewhere have ex in this phrase (cf. 458, 466) or indeed before l.

452. flǔvido: but 464 flūvida, 466 flūvidus. So līquidus i. 349, iii. 427, and līquida ii. 452, iv. 1259, but lĭquidis, etc., elsewhere. In both cases the short vowel is the natural quantity, but Lucr. appears to allow himself the licence of lengthening the vowel for scansional reasons when the last syllable is short; in vi. 349 he has lĭquidus. This variation of quantity may be taken as a sign of archaism; in the earlier stages of scansion by quantity prosody was sometimes uncertain. See Prol. VI, § 20.

453–5. These lines are very difficult and have provoked much controversy and many different solutions. 453 was expunged by Lambinus and by Munro, who says it is 'quite out of place'; but it has not the look of an interpolation. 454 was placed by Goebel and Giussani before 453 and after 455 by Brieger.

It is well to begin by establishing the meaning of certain of the key words in Lucr. (1) glomeramen is used in ii. 686 dissimiles igitur formae glomeramen in unum conveniunt in the sense of a mass of atoms joining together (Epicurus' ἄθροισμα‎). This is the natural meaning of the word, which is supported by the uses of the verb: glomerari in iii. 541 is used of the anima gathering together in a dying body, glomerata in iii. 497 of the semina vocis emitted by the mouth, and in iv. 871 of vaporis corpora gathered in the stomach. glomeramen then would mean the atomic nucleus, the particle formed of atoms which joins with similar particles to form a thing, though it is not necessary, with Giussani and Masson, to give it the technical sense of 'molecule'. But the word has a quite different sense in v. 726 aufert luciferam partem glomeraminis atque pilai, where it means the 'round ball' of the moon. It seems clear that (perhaps as Giussani suggests, through confusion with globus) it had also for Lucr. the sense of a 'round object'. An intermediate instance pg 877of a word from the same root is seen in i. 360 in lanae glomere where it seems to mean a 'round ball' of wool. (2) papaver is only used in one other place by Lucr.: iii. 196 papaveris … altus acervus, where it certainly means 'poppy-seed'. (3) haustus is used in i. 412 largos haustus e fontibu' magnis of a draught of water, and in v. 1069 teneros imitantur dentibus haustus of the act of swallowing.

The views taken of the passage may be divided according to the meanings that they assign to papaveris haustus.

(1) If it means, as older editors assumed, 'a draught of poppy juice', then in 455 procursus, the older vulgate, must be read (unless perculsus could be taken as a substantive) and the lines must be regarded as an example of what was stated in 451–2: 'for even a draught of poppy-juice (the most sticky liquid imaginable) flows as easily as a draught of water; for its single particles are not closely bound together and its course downhill (proclive) is as swift (as that of water)'. But apart from the necessity of emendation and the meaning of papaver, the example seems gratuitous and almost meaningless; nor is it true that poppy-juice would move as fast as water.

(2) If papaver means 'poppy-seed', there are two possibilities, but in both we have not an example of the behaviour of liquids, but, in Lucr.'s manner, a proof from experience:

(a) With Giussani we may suppose that it is a reference to a children's game played in Italy apparently at the present day. A handful of poppy-seed is taken in the hand and brought near the open mouth and then swallowed at a gulp, when it runs down the throat. In order to preserve the meaning 'molecules' for glomeramina he places 454 before 453, takes haustus as 'swallowing' (cf. v. 1069) and keeps perculsus as a participle with the suggestion that it might be a substantive. The translation would then run something like: 'for their single molecules are not closely bound together: for the swallowing of poppy-seed is as easy as that of water, and when propelled (by the hand) it runs down the throat just as easily'. But the translation is forced, there is no evidence of this children's game in antiquity, and if Lucr. had intended to refer to anything so obscure, he would surely have given some further explanation.

(b) I am sure that Giussani is right that it is a proof from experience and that papaver is poppy-seed. But I take papaveris haustus to mean 'a handful of poppy-seed'; there is no parallel in Lucr., but in Ovid, Metam, xiii. 526 haustus arenae is 'a handful of sand', and in Stat. Theb. x. 427 pulveris haustus 'a handful of dust'. We then have the behaviour of water, composed of round atoms, compared to that of the handful of round poppy-seeds. In 453 haustus will be used in a double sense with papaveris and aquarum, and similarly glomeramina referring primarily to the round seeds will also suggest pg 878the round nuclei of water: 'for a handful of poppy-seed is as easy in moving as a draught of water; for its round particles are not closely knit together and given a blow it runs as swiftly downhill'. The passage would be made easier by Brieger's transference of 454 after 455, for then glomeramina could have its double application to the water-particles and the poppy-seeds.

453. quod was altered to Haupt's quasi by most modern editors, who quoted parallels for the construction item quasi, but did not believe in item … quod, quod has, however, been restored by Diels and Martin; cf. Stolz-Schmalz4, § 302, p. 542.

454. The implication is not of course that the draught of water separates into individual drops comparable to the individual poppy-seeds, but that it is not so interlaced as to prevent swift movement.

glomeramen: for the form see Prol. VII, § 2.

455. perculsus: cf. vi. 311 cum vementi perculit ictu. Elsewhere, i. 13 and 261, the verb is used in a metaphorical sense.

item: sc. quod aquarum as in 453.

proclive: 'downhill'. Neuter used adverbially as is the comparative proclivius in ii. 792 and iii. 311 (proclivis occurs in vi. 728); cf. sublime in 206, iv. 133, vi. 97, and recens ii. 416 n. See Prol. V b, § 3. 2.

volubilis: cf. iii. 190 also of water quippe volubilibus parvisque creata figuris.

456–63. are transferred by Hoerschelmann, followed by Brieger and Giussani, after 477, in order to bring together the two passages dealing with liquids, 451–5 and 464–77, and to establish the order of treatment, solids (444–50), liquids, and gases or 'airy things' (456–63). But this is purely arbitrary and the logic of the existing order is quite satisfactory. Lucr. deals with the two simple cases first, solids and liquids, and then with two cases (smoke, etc.) and salt water, where there is an admixture of atoms of different shapes (see introductory note).

456. postremo: again unexpected, as in 408, but it introduces the two complex cases after the simple ones.

puncto tempore: see ii. 263 n.

457. diffugere: a very carefully chosen word; they 'separate and vanish'.

fumum, nebulas, flammas: in acc. in agreement with omnia in 454 and governed by necessest. In the subsequent discussion the clouds seem to pass out of Lucr.'s mind, and the smoke and flames are in his thoughts, esp. in 460.

458. omnia: 'altogether', 'in their entirety'; cf. i. 377 id falsa totum ratione receptumst, and n. there. It is unnecessary with Muretus, whom Lachmann and Munro follow, to read omnibu'.

459. perplexis indupedita: cf. ii. 102 indupedita suis perplexis ipsa figuris. The atoms of smoke, etc., though they are pointed (acutis 463), cannot be interlaced tightly or the mobility of the smoke would be impossible. See 454 n.

pg 879

460. pungere uti possint corpus: as smoke does when it gets in one's eyes, and fire when it burns one, penetrareque saxa, as fire does. Cf. i. 491 dissiliuntque fero ferventia saxa vapore and i. 535 penetralem ignem. With these parallels and Lucr.'s general conception that heat (and still more, flame) is a concrete body, I cannot see why the editors have thought it necessary to 'emend'. Lachmann reads sese making penetrare trans. like insinuare in ii. 436 or declinare in ii. 250. Munro suggested laxa 'rare in texture', which is adopted by Diels, and later vesca 'biting', comparing i. 326. Mr. B. B. Shefton of Oriel College has suggested to me sensus, comparing ii. 407 perrumpere sensus: the meaning is appropriate, but it is palaeographically too far from saxa. Giussani, thinking that penetrare is not a strong enough word for the atomic disruption produced by the flame, substitutes terebrare. But there is nothing here of the forcing of passages in the rock which do not already exist. We may notice in this line that there is nothing corresponding to nebulas 457; see n. there.

penetrareque: -que again attached to short e; see i. 134 n.

461–3. quodcumque videmus … esse elementis: these lines and especially sensibu' sedatum have not yet been satisfactorily explained or emended. Munro keeps the MS. text and explains 'whatever we see the senses have been able to allay': i.e. the eyes can ultimately overcome the smart of the smoke. He compares ii. 956 ingentis plagae sedare tumultus. This interpretation, which is adopted by Ernout, is just possible, but it is forced and would concentrate attention on the one point pungere corpus, neglecting diffugere and penetrare saxa. The obvious correction palaeographically is that of the Itali sensibus esse datum. Nencini adopts this, reading quae cumque videmus sensibus esse datum facile ut cognoscere fossis … 'we see that it is granted that you can easily recognize by the senses that all these', etc. This is very ingenious, but the construction is complicated and cumque is left without any real meaning. Brieger too, followed by Giussani, adopts sensibus esse datum, assuming a line lost between 461 and 462, which he would fill ventis differri rapidis nostrisque veneno. This line is prompted by two other conjectures ventis esse datum Bernays, and venenumst sensibu' sed rarum Lachmann, which gives good sense, but is very far from the MSS. Of other conjectures may be mentioned Martin's sensibus dentatum, based on 432, and the strange conjecture of Diels quod utrumque videmus sentibus (Faber) esse datum: 'both of which properties (sc. the capacity to penetrate but not stick together) we see are given to brambles'. Housman conjectured senti ibus, 'whatever of prickliness', but senti is very improbable. Of all these proposals Munro's retention and explanation of the MS. reading is best, but it is not convincing and the two words must be left obelized.

pg 880

462. facile ut cognoscere possis is, unless Nencini's view is adopted, the main clause, and quodcumque videmus … is the acc. subject of esse.

463. acutis: 'sharp', 'pointed' so that they can prick, without interlacing as they would if they were hamata.

464. flūvida: see n. on 452.

465. sudor … maris: 'the sweat of the sea', i.e. the salt water. sudor, as Ernout says, is the equivalent of liquor (ii. 390) with the added notion of bitterness. The passage quoted from Empedocles in which the sea is described as ἵδρως γῆς‎ is not parallel in meaning though it may have suggested the phrase to Lucr.

debet: OQ. Ernout would keep it as it stands supposing that esse can be 'supplied', as est is in phrases such as nec mirum ii. 77, 338, quid mirum v. 1238. But this is extremely harsh, and in all parallel passages in the poem debet has a verb with it; v. 666 nec tamen illud in his rebus mirabile debet esse seems decisive. I therefore follow Bernays and other editors in supposing a line lost; Diels suggests esse tibi, veram si cognoris rationem, which might be the general sense. The alternative is to emend debet. The Itali have habendum (cf. iv. 256 minime mirabile habendumst), Lachmann reads habebis and Munro habeto. But these are large changes and would not account for debet.

466. flūvidus: see n. on 452. The earlier editions read fluvidum, supposing that quod was relative pronoun; but the scansion is impossible, and quod is of course conjunctive, 'as for its (the sudor maris) being fluid'.

467. et squalida sunt illis admixta doloris corpora: 'and many rough bodies, causing pain, are mixed with them'. This is Bernays's correction for e levibus atque rutundi admixta … in OQ, which is clearly a dittography of the last words of 466. Lachmann proposed et levibu' sunt illis admixta, but there is no reason to suppose that any part of the repeated words is right and 469 points to squalida here. Bernays's correction is better, though in both these versions doloris corpora is queer Latin; the nearest parallel in Lucr. is cladem … pericli v. 369. In order to avoid this Munro read squalida multa creant admixta doloris corpora, taking doloris as acc. plur. But apart from the form doloris for dolores, which Ernout considers impossible, the plural is unnatural and we should have expected dolorem. Diels's squalida quae sunt causa admixta doloris corpora would explain the gen. but is too involved, and Martin's levibus sunt hamata … does not improve matters. On the whole Bernays's suggestion seems best.

squalida: 'rough'. Cf. ii. 425 non aliquo sine materiae squalore.

468. necessumst: Lachmann for the defective necessu of OQ. It is impossible to choose between this and necessust (Diels, Martin) which Lucr. also uses at the end of the line, as in 710 and 725 of this pg 881Book. Ernout would write necessum or necessus, but elsewhere the word has est with it, though necesse is used alone in iii. 543; see Prol. V b, § 13 (a).

469. globosa: here only in Lucr., though globus is frequently used in Book v of the 'ball' of sun or moon and once iv. 119 of the heart: cordis globus. It is here a convenient synonym for rutunda.

constent: the early editions made the necessary correction for constet OQ.

471. quo, Pontanus for quod OQ, a necessary change.

472. unde: 'from which', i.e. from the mixture of rough and smooth atoms.

Neptuni: here and in vi. 1076 Lucr. avails himself of the poetic licence of using Neptunus for mare, which he deprecates in ii. 655.

473–7. These lines can be taken as they stand, if in 477 with Munro we change quo to quom or with Marullus (and Giussani) possint to possunt: 'there is a way of sundering them and seeing how, apart from the rest, the fresh water, when it trickles many a time through the earth, flows into a trench and loses its harshness; for it leaves behind up above the first-beginnings of its sickly saltness, since the rough particles can more readily stick in the earth' (or, reading quo … possunt, 'the more the rough particles can stick in the earth'). umor dulcis fluat is then used, as Giussani points out, in a double sense, 'the fresh water flows apart' and 'the water flows fresh'; the subject of percolatur is not logically umor dulcis but only umor, so that then dulcis is used predicatively. The difficulty is very slight, but in order to avoid it Lachmann transposed 476 before 474. But this does not make percolatur easier and leaves 475, 477 in awkward juxtaposition: 'so that it flows into a trench and is softened, so that the harsh atoms can remain in the earth'. Robin quotes interesting parallels in Greek for the separation of the fresh water and the salt in sea-water.

475. percolatur: 'it passes as through a sieve'. So v. 269, vi. 635 percolatur enim virus.

foveam is presumably the 'salt-pan' in which the salt is collected.

476. viri: 'the tainted water', i.e. salt water; so in i. 719 virus ab undis and in v. 269, vi. 635 quoted on 475.

477. quom … possint (Munro) is on the whole the better correction for quo … possint OQ than Marullus's quo … possunt. The clause cannot be final as the MSS. would make it.

## (b) The number of atomic shapes and of atoms of each shape. 478–580.

1. The number of different atomic shapes is not infinite. 478–521.

478. It follows from what has been said that the number of atomic shapes is limited. 481. For (1) otherwise, some atoms must be of pg 882infinite size. For within the limits of a single small size, shape cannot vary much. 485. Suppose an atom to consist of three, or a few more, minimae partes. Arrange these parts in all possible ways, changing top and bottom, right and left, observing what shape each arrangement gives. 491. If after that you wish to vary the shapes, you must add further parts, and more again, if you desire still further variety. New forms imply increase of bulk. 496. It is therefore impossible to believe that the atoms differ in shape without limit, lest some should be of vast bulk, which I have shown to be impossible. 500. (2) (If there were infinite atomic forms,) all that is now most beautiful in colour would be far surpassed by new colours. 504. So too lovely smells and tastes and the finest melodies would be outdone: something better would always be discovered. 508. So also would all that is worst to smell, hearing, sight, and taste be surpassed by something worse. 512. Since this is not so, but there is a limit both to good and bad, there must be a limit to the shapes of the atoms. 515. So again there is a limited scale of heat and cold; all the intermediate degrees of warmth lie within definite limits. 519. They differ only to a certain extent and are limited on the one hand by flames, on the other by frost.

After his long demonstration of the effects of the difference in the shapes of the atoms, Lucr. comes on to the two points dealt with by Epicurus in §§ 42 and 56 of the Letter to Herodotus (quoted in general note on 333–729). The two points hang closely together: (1) that the number of different atomic shapes is limited, (2) that the number of atoms having each of the shapes is infinite. In this paragraph Lucr. deals with the first statement and advances two arguments, one a priori, the second a posteriori based on our experience. In the first he follows Epicurus' arguments that after a certain point variation in shape can only be produced by increase in size, and that if such variation were infinite, some atoms would be very large and therefore perceptible to the senses. Epicurus was here refuting Democritus, who held both that some atoms might be very large, indeed 'the size of a world',1 and also that the number of atomic shapes was infinite.2 The two propositions clearly hang together and Epicurus, on the ground that in experience we cannot see atoms, denied them both. The atoms are far below the ken of the senses and the number of atomic shapes is 'incomprehensible' (ἀπερίληπτα‎) but not infinite (ἄπειρα‎). This argument Lucr. states clearly and elaborates ingeniously from the doctrine of the minimae partes (i. 599–634). Brieger (Neue Jahrb. cxi (1875), p. 630) worked this out clearly. If we suppose that the minimae partes (which are all pg 883the same size and shape) were square in form, then with three only two shapes are possible and . With four there are five possibilities in the same plane and two more in two planes. To get further varieties other minimae partes must be added. No doubt to obtain varieties of shape which could be described as 'incomprehensible' in number, we must suppose some atoms with a large number of partes, but still they would be far below the ken of the senses.

The second is an argument from experience and is based on the preceding demonstration of the effects on sensation of the different atomic forms. As it is, they can produce a large number of effects, both pleasing and displeasing, to taste, sight, smell, and hearing. But there is a limit to these, which could not be the case, if varieties of atomic shape were infinite in number. Similarly with heat and cold: there is here again a limit at both ends which cannot be surpassed. Our experience then leads to the same conclusion that the number of variations of atomic shape cannot be infinite. The argument here is straightforward and may be compared with that in i. 584–98.

Before this paragraph Brieger and Giussani mark a considerable lacuna, which they suppose to have contained the proof that there is a limit to the size of the atoms. This would be the natural place for such a proof, and there are expressions in the paragraph itself which almost seem to demand it: 478 quod quoniam docui, pergam conectere rem quae ex hoc apta fidem ducat would be far clearer if it followed such a demonstration, though it can just be taken to mean 'since I have demonstrated the variety of shapes, I must deal with the closely connected point that such shapes are not infinite in number'. But 481 rursum … auctu and still more 498–9 ne quaedam cogas … supra quod docui non posse probari explicitly demand a previous proof that the atoms are limited in size. The only passage which could be quoted as supplying this is that on the minimae partes and especially i. 615–34. But even there the proof is at most implied. There is therefore much to be said for the supposition of a lacuna. However, in the unfinished state of the poem it is rash to assume that a passage has been lost; Lucr. may never have written it and it is safer merely to say that the argument requires it.

478–9. quod quoniam docui … ex hoc apta fidem ducat: see introductory note. What makes it very difficult to refer this to the preceding paragraph is fidem ducat, which should mean 'derives its guarantee from', as it does in ii. 522–3, which are repeated verbatim from here. It is impossible to infer the limitation of atomic shapes from the account of their effects, which might equally well lead to pg 884the conclusion that they were infinite in number. Nor can quod and hoc refer to anything so remote as i. 599 ff., though Büchner (Beobachtungen, p. 6, n. 2) mentions this as a passage where Lucr.'s thought is 'suspended' and refers to something 'far back'. It is hard to resist the impression that Lucr. at least intended to write a paragraph proving the limited size of the atoms.

479. ex hoc apta: so in a literal sense iv. 829 bracchia tum porro validis ex apta lacertis.

480. 'Vary in a limited system of shapes';

ratio is the regular succession of shapes which results from the rearrangement and addition of minimae partes; cf. 519 ergo finita distant ratione creata. Note variare intrans. as in 484, etc.

481. quod si non ita sit: 'suppose it were not so'. Lucr.'s idiomatic use of the pres. subj. where the unfulfilled imperfect might be expected. Cf. i. 968 si iam finitum constituatur omne quod est spatium and n. there, and see Prol. V b, § 10 (b).

rursum: see intro ductory note; there is nothing for it to refer to.

482. auctu: not merely bulk but 'additional bulk'; cf. augmen 495. auctus is used in the same sense in ii. 1121, v. 846, 1171, vi. 168, 327.

483. 'For then in one and the same small compass of any atom you will …' OQ have in eadem unac uiusvis in brevitate, where the repeated in is impossible and Martin does not justify it by placing a comma after eadem. Lachmann objected too to eadem una without a copula, standing for the more usual una et eadem. He therefore wrote namque $ea¯$dem unius cuiusvis in brevitate, which makes a double change and produces a very harsh rhythm. Munro read in eodem (sc. semine), which is no better than Martin's attempt to preserve the MS. reading. The simplest correction is Brieger's iam for the second in, iam being used as in 481 'in these circumstances', 'then'. Lachmann notes that eadem una would be natural, if there had been any previous mention of brevitas, an additional argument for the missing section. brevitas occurs here only in Lucr.

484. corporis: the singular used of a single atom; see 490, 495, and cf. i. 600 corporis illius quod nostri cernere sensus iam nequeunt.

variare: within a single atom of course shape could not vary, for it is immutable. Lucr. means that in a series of atoms all of the same size, i.e. containing the same number of minimae partes, the shapes cannot vary much. This is clearly explained by what follows.

485. minimis e partibus: cf. i. 610 quae minimis stipata cohaerent partibus arte. The proof here depends entirely on the doctrine there.

486. paulo pluribus: to make the διαφοραί‎ of shape even ἀπερίληπτοι‎, there would have to be a good many more minimae partes in some atoms.

488. 'Placing them at top and bottom, transferring right to left.' It may be said that this is Lucr.'s equivalent of Leucippus' con-pg 885ception of the τροπή‎ of an atom, its position in relation to itself; the conception is modified by the idea of the minimae partes.

489. omnimodis: 'in every way'; see i. 683 n. and Prol. V b, § 6.

489–90. quam quisque … corporis eius: 'what form of outline in the whole atom each arrangement gives', i.e. what atomic shape is produced by each rearrangement of the minimae partes.

490. formai speciem: a tautologous amplification; cf. iv. 69 formai … figuram.

corporis: again the single atom.

491. quod superest: 'beyond that', 'if you want to go on': similar uses in i. 50, ii. 39, 546.

492. addendum partis alias: the verbal construction of the gerund; see i. in n. and Prol. V b, § 13 (b).

493. alias ut postulet ordo: 'that the (new) arrangement will require further minimae partes'.

494. etiam: 'further', 'beyond what you have already done'.

495. 'so then increase in the bulk of the atom follows on newness of shape.' An important generalization which contains the gist of all the controversy between Epicurus and Democritus; see introductory note.

augmen like auctus in 482. For the word cf. i. 435 n.

498. maximitate: an invention of Lucr. because magnitudine will not scan, but also no doubt to give a greater impression of size. It is elsewhere only found in Arnobius, who often imitates Lucr. In other places Lucr. gets over the difficulty by the use of filum, etc. See Prol. VII, § 4.

499. supra quod iam docui: the last and most direct reference to the missing proof that the atoms are limited in size.

500. The second a posteriori argument is introduced rather abruptly. Brieger and Giussani mark a lacuna of several lines in which they suppose that Lucr. stated generally that if varieties of atomic form were infinite, there would be infinite variety in their effects. But this is unnecessary, and we have already noticed a similar abruptness in ii. 381 (see introductory note to paragraph) where Lucr. proceeds at once to his examples.

iam is emphatic in the same sense as in 481 and 483—'if this were the case', 'if atomic shapes were infinite in number'.

barbaricae: the only occurrence of the word in Lucr. Merrill assumes that he means 'Phrygian', but it may be any kind of oriental garment.

Meliboea: from Meliboea, a Thessalian (Thessalico 501) town, beneath Pelion and Ossa.

501. Thessalico concharum … colore: transference of the epithet; see i. 10 n. and Prol. VII, § 12.

tacta Oudendorp, was adopted by Lachmann and other editors for tecta OQ; cf. vi. 1188 croci contacta colore. It is more probable than the other suggestions tincta ed. Junt., and infecta Winckelmann (cf. inficiunt iv. 80), which is preferred by Ernout on the ground that Lucr. elsewhere uses contingere not tangere in this sense.

pg 886

502. aurea pavonum … saecla: introduced with a typical Lucretian asyndeton after -que in 500: we have an exact parallel in i. 455–6 servitium contra paupertas divitiaeque libertas bellum concordia. Editors have adopted various expedients to avoid it. Lambinus read et after colore in 501. Munro and Giussani assume a line lost before 502 which Munro fills et quos ostendunt in solis luce colores. Lachmann more violently places a comma after aurea (not colore) and takes it with purpura, reading imitata for imbuta, referring the whole idea to embroidery; Bernays read caudaque pavonum … caeca ('blinding' for saecla). But none of these devices is necessary and the passage can be read quite happily as an asyndeton.

ridenti is a necessary correction for rident, as is novo for nova in the next line.

503. saecla: 'races', the usual Lucretian periphrasis. Burmann's 'emendation' pepla would not avoid the difficulty and seems only to arise from the unwillingness, which also beset Lachmann, to mention real peacocks, which are in fact more probable than embroidered peacocks.

504. Lucr. extends to smell and taste and in the next line to sound what he had said more fully about sight. iacerent must of course be supplied again.

smyrnae: 'myrrh' = murra. Similarly in Greek there is the doublet σμύρνη‎ and μύρρα‎, the former being the more common. It was used both as incense and as an unguent.

505. Composed entirely of Greek words except et and -que. Cf. ii. 412 n., and see Prol. VII, § 9.

Phoebea: probably denotes lyre-music, the lyre being Phoebus' instrument.

daedala chordis: probably go together, 'varied on the strings'.

507. 'For one thing after another would arise, each superior to the last.'

508–11. All that is worst to the senses would also be surpassed.

509. in meliores: sc. progredi, corresponding to and understood from cedere retro.

513. sed: Lachmann's necessary addition to a defective line.

utrimque: 'at either end', good and bad.

summam: here and in 518 the whole scale or range within which the varieties move.

515. hiemum usque is the emendation suggested by Munro in his notes and adopted by Diels and Martin for hiemisque OQ, with which -que has no sense. Editors have agreed on usque: the plural of hiems is unusual, but is used by Lucr. himself in vi. 373 hic quoque confligunt hiemes aestatibus acres. Merrill's brumae usque is therefore gratuitous. finitumst and remensumst are then used without an expressed subject, 'there is a limited distance from flame to the cold frosts of winter'. This is no doubt harsh, but not harsher than iv. 812 proinde quasi omni tempore semotum fuerit longeque remotum. Editors thinking that there must be an explicit subject have adopted either the suggestion of Lambinus spatium usque or Lachmann's pg 887iter usque: either would make the construction easier, but both are palaeographically improbable.

516. pari ratione: 'on the like system'; ratio as in 480 and 519.

517. omnis … calor ac frigus: 'all the degrees of heat and cold' as well as the tepid temperatures between them. There is really no difficulty in the expression and the 'corrections' of the Itali finis, Lambinus finit, and Lachmann ambit are entirely unnecessary.

518. interutrasque: 'in between', an idiomatic adverb formed on the analogy of alias, foras, etc. It is used in six other places by Lucr., and in all Lachmann prefers to write interutraque; but such repeated corruption is improbable and there seems no reason to doubt the form.

519. finita … ratione: cf. 480 n. and 516.

520. ancipiti: 'at either end'. Cf. vi. 168 ancifiti ferro 'a double-edged axe'.

mucroni: the metaphor is from the point of the stilus which makes a stroke or dot to end the series. For the form of the abl. in -i, cf. tripodi i. 739 and n. there, and Prol. V a, § 1. Here, as in i. 739, the unusual form is not required by scansion, as it is with Cilici 416, but it should be retained.

521. infesta: 'barred by', 'beset by'. Cf. Mela iii. 44 (quoted by Munro) tellus infesta frigoribus and Cic. De Prov. Cons. 2. 4 via … excursionibus barbarorum infesta. In v. 760 loca flammis infesta it is used in its more usual sense of 'hostile to'. It is Lambinus's correction for infessa OQ and is to be preferred to infensa of Itali.

## 2. The number of atoms of each shape is infinite. 522–68.

522. It follows from what has been said that the atoms of any given shape are infinite. 525. For since the number of shapes is finite, either those of each shape are infinite or the sum total of atoms is finite, which, as has been shown, is not the case, but rather the atoms for ever maintain the total of things by prolonging the succession of their blows. 532. It is true that some animals are rare in some parts of the world, but they are common elsewhere and so complete the sum; elephants, for instance, are common in India but rare with us. 541. Even grant that there might be some unique thing, yet unless there is infinite material for its creation, it could not be born nor grow. 547. For supposing the atoms requisite to create a unique thing were limited in number, how could they meet in the great turmoil of matter? 551. Just as in a shipwreck when the sea tosses asunder all the ship's parts and throws them on distant shores to warn mortals to shun the snares of the sea and not to trust its alluring calm, 560 so if you suppose any class of atoms to be limited, the tides of matter will toss them asunder through all time, so that they could never meet in combination, nor stay together, nor receive increase. 565. Yet it is clear that things are pg 888both created and grow. There must, then, be infinite atoms of every class from which all things are supplied.

This second proposition is at once the corollary of the first and an inference from it. It is an amplification of the simple statement of Epicurus that 'in each configuration the atoms that are alike are quite infinite in number'.1 This is clearly expressed in 523–5 and Lucr. adduces an a priori proof. Since the number of shapes is limited, the number of atoms in each shape must be infinite; otherwise the total number of atoms would be limited, and that he has already shown not to be the case (i. 1009–51); rather there are infinite atoms in an infinite void which from infinite time past have kept up the succession of clashes and blows which result in the creation of things. At this point we might expect an a posteriori proof derived from the immense number of created things of every kind. Lucr. does indeed give this argument but in the form of an answer to a possible objection. It might be said that since the number of certain classes of created things is comparatively small, we might infer that the classes of atoms required to produce them are limited in number. This objection Lucr. meets with the Epicurean principle of ἰσονομία‎, 'equal distribution' or 'equilibrium'.2 Its basic idea is that since the atomic meetings are regulated by law (or by chance), it must be (or the odds are) that an equal number of things of each class will be created. This principle Lucr. enunciates in its two possible aspects, law and chance, in ii. 1048–76, and it plays a great part in the astronomical section of Book v. Its application here is that if things are rare in one part of our world, then they will be found to be common in another, so that the balance is maintained. (In the Fifth Book, where he is harder put to it, he has to add, 'if not in our world, then in some other'.) Finally, Lucr. makes the utmost concession to his opponent and supposes that there is some unique thing created. Still there must be infinite atoms of the right kind to create it: for, if they were limited in number, in the clash and turmoil of atoms they would never be able to meet and form the creature; nor could the creature grow, for the right atoms to sustain it would not come near it: 'the concilium could never be formed, nor if formed could it remain in combination, nor receive additions and so grow' (563–4). The creation and growth of things, he argues in conclusion (565–8), prove that there are infinite numbers of atoms in each of the atomic shapes. The argument is rather technical, and depends, as it so often does in Lucr., on the acceptance of the Epicurean system in general.

pg 889

522–8. Lachmann brackets these lines, considering them to be a second attempt at an exordium to the paragraph, which originally began at 529 (with protinus for versibus). But 529 ff. do not sufficiently explain the subject of the paragraph and 522–8 are essential as an introduction: their similarity in form with 478–80 links the two paragraphs closely together (see introductory note to 478–521). Of subsequent editors none but the faithful Bernays accepted Lachmann's exclusion.

522–3 = 478–9. Lucr. rarely links paragraphs together with such formality: it is a proof of their close connexion in his mind.

525. cluere: almost equivalent to esse here; perhaps 'are reckoned', distantia: see ii. 373 n.

527. summam material: 'the sum total of matter', i.e. the total number of atoms; cf. ipsa … rerum summa in i. 1008, and here in 530, in a slightly different sense; see n. on i. 235.

528. id quod non esse probari: i.e. in i. 1008–51.

529–31. These verses were rejected by Susemihl (Neue Jahrb. cxxxiii (1886), p. 777, n. 1) and by Brieger (Phil. xxiv. 449), the latter supposing that their right place was after i. 1013. But they would not fit in there and such a transposition from one Book to another is extremely improbable. They must certainly be retained.

529. versibus ostendens: OQ have versibus ostendam. All editors have felt the awkwardness of the connexion with what precedes and, as a minor difficulty, the use of versibus without an epithet. The majority have retained ostendam, though only Martin without any alteration or explanation. Lachmann, having bracketed 522–8, drastically substitutes protinus for versibus, as Marullus had previously suggested nunc vero; these proposals are far too violent. Marullus later supposed a lacuna before the line, which he would fill with quod quoniam docui, nunc suaviloquis age paucis; so Diels, suggesting nunc vero, ratio quod vera docet, tibi paucis comparing vi. 82 multa tamen restant et sunt ornanda politis versibus, which is not a close parallel. As to the use of versibus alone, it occurs in i. 416 quam tibi de quavis una re versibus omnis argumentorum sit copia missa per auris, though not in this emphatic position at the beginning of the line. But the real objection to all these proposals is that 529–31 refer not to what Lucr. is going to prove but to what he has proved in i. 1008–51 and, as regards the succession of blows, in ii. 83 ff. and 216 ff. This Brieger saw and proposed a lacuna before 529 and ostendi for ostendam. Munro's solution is better; to remove the stop after probavi 528 and read ostendens: 529–31 are then, what one would expect them to be, an elaboration of id quod non esse probavi; versibus will also be easier with a participle than with a main verb. The corruption of -ens to -am is unusual, but Munro's suggestion is accepted by Giussani and Ernout, and with some pg 890hesitation I have adopted what seems to me the best solution of a difficult passage.

corpuscula: cf. ii. 153 n. Here it is used of atoms, not, as there, of atomic nuclei.

529–31. 'The bodies of matter gathering from infinite space continue to (usque) hold together the totality of things by means of the succession of blows, which goes on and on in all directions.'

530. ex infinito: here probably of space, as Giussani takes it, not of time, as Munro believes. Cf. i. 997, where there is the same ambiguity, but again it probably refers to space. Robin's attempt to see here a reference to 'the three unities', of matter, space, and time, seems to me ill-founded: usque and continuato are not sufficient to suggest the infinity of time.

summam rerum: 'the sum of things created', not quite an equivalent of summam materiai in 527, which refers to matter both within and without created things.

tenere: 'maintain', 'keep up'.

531. undique: all through infinite space the blows are going on, one resulting from another, to and from all directions.

protelo: protelum (from pro and tendere) is literally the line of oxen harnessed 'tandem' to the plough, as in Lucilius 248 quem neque Lucanis oriundi montibu' tauri ducere protelo validis cervicibu' possent, and 435 hunc iuga mulorum protelo ducere centum non possunt. Lucr. uses the same metaphor again in iv. 190 quasi protelo stimulatur fulgere fulgur. Donatus (ad Ter. Phorm. 213) suggests, probably wrongly, that Lucr. understood the metaphor to be from the continuous throwing of weapons.

continuato: the verb occurs only here in Lucr.

532. Lucr.'s answer to the possible objection is introduced rather abruptly, but he undoubtedly had in his mind all through this passage the close connexion between atomic shapes and the forms of created things (see introductory note). Notice for instance the close similarity of phraseology between this passage about the forms of animals and 560 ff., where he sums up his conclusions on the atoms of various shapes, animalia quaedam 532 = primordia quaedam 560 and genere in eo 535 = genere in quovis 567.

533. fecundamque minus is Lambinus's correction of fecundam-que magis OQ. It certainly simplifies the passage as this line then becomes an amplification of 532. But it is palaeographically a large change and magis might just be retained. Lucr. may have intended to say: 'for since you see animals more rare or nature more prolific in them, in other parts of the world there are plenty (or very few)', and then in the elaboration of the first part of his answer he forgot the second. The same difficulty occurs in iv. 1225, where the MS. reading magis can be rather more easily defended.

534. regione locoque alio: cf. iv. 786 cum praesertim aliis eadem in regione locoque; here terrisque remotis is a further tautological amplification. For regio see i. 64 n.

pg 891

535. numerum: the total, the right number of such animals as is required to preserve ἰσονομία‎.

536. sicut: for this form as against sicuti OQ see a learned note of Lachmann's ad loc.

537. anguimanus: a magnificent epithet, which Lucr. uses again of elephants in v. 1303. The only instance of an adj. with 4th decl. termination except in nom. or acc. sing.; see Prol. V a, § 1. For the idea see Cic. De Nat. Deor. ii. 47. 122 manus etiam data elephantis and a strange passage quoted by Ernout from Isid. Orig. xii. 2.14 rostrum autem proboscida dicitur, quoniam illo pabulum ori admovet, et est angui similis, vallo munitus eburneo. The last clause is clearly a reminiscence of 538, though it does not throw much light on it.

538. vallo munitur eburno: there seems no other reference in the classics to this strange legend of India's ivory wall. Pliny, Nat. Hist. viii. 31 sed tamen in extremis Africae, qua confinis Aethiopiae est, postium vicem in domiciliis praebere, saepesque in iis et pecorum stabulis pro palis elephantorum dentibus fieri Polybius tradidit, suggests an ivory paling for individual houses; and Isidorus, quoted on 537, speaks of the 'ivory wall' of the individual elephant. Giussani supposes that there was no definite legend, but a poetical exaggeration of the defence of India against Alexander by means of the elephants in the native armies. But the lines suggest something more distinctly legendary.

539. ferarum vis: cf. ii. 264 equorum vim, 326 virum vi, and with Munro's emendation ecum vi in 42.

540. exempla: 'specimens'.

541. Lucr. passes to the extreme case, in which a thing is not only rare, but absolutely unique, and argues that even so, unless the atoms of the right shape to compose it are infinite in number, they could not meet to form and nourish it.

sed tamen … concedam: cf. iv. 473 et tamen hoc quoque uti concedam scire.

quamlibet esto unica: 'let it be as unique as you will'. quamlibet goes with unica in sense, but its combination here with imper. illustrates the process by which it and quamvis came to act as conjunctions; see n. on quamvis i. 346.

542. nativo corpore: the expression recurs in ii. 1088, v. 60, 238, 241, but always with the meaning 'a body that has a birth' and in conjunction with mortalis. Here it has a rather different meaning, akin, as Giussani says, to proprio, 'alone with the body with which it was born', sc. a body of its own, unlike others. The nearest parallel is perhaps ii. 311 si quid proprio dat corpore motus.

543.nonsitinorbi Q1 for sit orbi O orbi Q. As the corrector may have had authority in a blurred text for his reading, it seems slightly preferable to Lachmann's ⟨nulla⟩ sit orbi, which has commended itself to most recent editors except Martin. The pg 892meaning is in any case the same. For orbi abl. see i. 286 and n. there.

544. vis material: 'the supply of matter', cf. ii. 306 nova vis.

545. progigni … concepta: the terms are appropriate to a living animal, which Lucr. has primarily in mind, though the argument is equally applicable to inanimate things.

546. quod superest: 'for the rest', 'after that'. Cf. ii. 491.

547–50. Robin appropriately quotes Diogenes of Oenoanda fr. xx. 6. 9 πῶς ἀπογεννήσωσι τὰ πράγματα χωρὶς ἀλλήλων‎; … εἰ γὰρ ἦσαν πεπερασμέναι οὐκ ἠδύναντ‎ʼ ἄν‎.

547. sumam hoc quoque uti: 'that I may assume this too', i.e. he will concede not only the existence of a unique thing, but also that the atoms required to create it are limited. The MSS. have the nonsense text sumant oculi, and Munro's correction is strongly supported by iv. 473 (quoted in n. on 541). Of the many other suggestions the only two which merit consideration are Winckel-mann's sumantur uti 'that they may be assumed to', a personal construction like that in i. 566 possint tamen omnia reddi … quo pacto fiant, and Martin's sumant alii, where however alii is weak. Palaeographically neither proposal accounts so well for oculi. Lachmann's si manticuler 'if I were surreptitiously to assume' is a good example of his ingenuity and perversity; for it gives precisely the opposite meaning to what is wanted. This criticism applies to some extent to sumam or sumantur, for this is really not an assumption but a concession on the poet's part: cf. id quoque uti concedam 541. If it were not for the close parallel of iv. 473 I should be tempted to suggest sumas quoque uti of the imaginary opponent, which would receive some support from 560 sic tibi si … constitues.

548. iactari: cf. ii. 89.

unius genitalia rei: 'required to create that single thing'. For genitalia with objective gen. see ii. 62.

550. pelago: a fine metaphor which suggests the following comparison with a shipwreck.

turba … aliena: 'turmoil of alien atoms', i.e. atoms of different kinds, unsuitable for the formation of the 'one unique thing'. For alienus in this sense cf. i. 181 alienis partibus anni and the use of alienigenis in i. 865, 869, 874.

551. rationem: 'means' of coming together. Cf. i. 110 nunc ratio nulla est restandi: see n. on ratio i. 54.

552. naufragiis … coortis: a rather strange expression on the analogy of tempestas coorta est. coortus is a favourite word with Lucr., but to alter with Marullus to ventisque coortis would spoil the alliteration: multis too is needed if the wreckage is to be scattered on 'all shores' 555.

553. cavernas: 'the ribs' of the ships, Q1 for MSS. caverna. cavernas was defended by Robinson Ellis (C.R. xi. 205) against Lambinus's generally accepted guberna. He quotes Servius, ad Aen. ii. 19 pg 893alii fustes curvos navium … cavernas appellarunt, and Cic. De Or. iii. 180 quid tam in navigio necessarium quam latera, quam cavernae, quam prora, quam puppis, quam antennae, quam vela, quam mali? Elsewhere in Lucr. the word is used of the vaults of the sky iv. 171, 391, vi. 252, or the hollows of the earth vi. 597, 683.

554. proram: O writes prorem, which Diels prefers.

tonsas: 'oars', a frequent poetic use; e.g. Virg. Aen. vii. 28, x. 299.

555. aplustra: well defined by Mayor on Iuv. x. 136 as an ornament which rose on the stern in the form of a crest or wing or fish's tail and carried a pole with coloured ribbons. Lucr. uses the word again in iv. 437; cf. Cic. Arat., fr. 2 navibus absumptis fluitantia quaerere aplustra; they may both be imitating Ennius.

556. indicium edant … ut vitare velint (558): cf. Hor. Sat. i. 4. 110 magnum documentum ne patriam rem perdere quis velit.

559. placidi pellacia ponti: one of Lucr.'s great alliterative phrases which he uses again in v. 1004.

560. tibi: 'ethic' dat., not with constitues.

quaedam: sc. the primordia of the shape required to make any particular thing: corresponding to animalia quaedam 532; see n. there.

561. aevum … per omnem: masc. as in iii. 605 non omnem possit durare per aevum; elsewhere in Lucr. the gender is uncertain, but there is no place in which it is bound to be neuter.

562. aestus diversi materiai: 'the currents of matter moving this way and that', the idea of the stormy sea is continued.

563–4. Note the three stages, in concilium … coire, remorari in concilio, and crescere adaucta. For both the two latter processes an additional supply of the right atoms would be required.

adaucta: 'increased over and above' the atoms required originally to make it. Cf. ii. 296 nam neque adaugescit quicquam neque deperit inde and n. there.

565–6. The evidence of experience is against the supposition: things are created and they do grow.

567–8. The conclusion not merely to the argument about the unica res, but to the whole paragraph.

567. genere in quovis corresponds to 535 genere in eo; see n. on 532. The genus is the 'class' of atoms of any one shape.

568. palam est: 'it is clear', 'obvious'; palam is treated as a neuter adj. Cf. palam fieri 565 and v. 1157 perpetuo tamen id fore clam diffidere debet.

## 3. Equilibrium of the forces of creation and destruction. 569–80.

569. Neither can the motions of destruction be for ever supreme, nor can those of creation and growth. 575. So an equal war is waged between them for all time. Now here and now there the forces of life are victorious or vanquished. 577. With the funeral is mingled pg 894the wail of new-born children. There is never a night or day which does not hear the lament of mourners mixed with the children's cries.

This paragraph is an appendix to the last section, just as 294–307 is to the section on atomic motion. And indeed it is closely connected with it in thought. There Lucr. shows that since the density of dispersion of atoms in the void is always the same, so must be also their motions and therefore the processes of creation. Here he has added to this the further idea that some atomic motions lead not to the creation of things, but to their dissolution. But again, seeing that the sum total of motions in all directions is the same, the creative motions will, on the whole view, be balanced by those of destruction. In some parts of the universe, for instance, when a new world is created or an old world dissolved, one or other set of forces may for the time prevail. But in the whole there is balance. It is clear then that here again we have an example of the principle of ἰσονομία‎, which appeared in 532–40,1 and it is the same balance of atomic motions which is responsible both for the equilibrium of the forces of creation and destruction, and also on a lesser scale for the equal distribution of animals in our world. The principle will be found again in ii. 1084–6 in another context.

In the last six lines Lucr. illustrates ἰσονομία‎ from the life of human beings. Every day children are born at the moment when their elders die: funerals are matched by birth-ceremonies.

Robin points out that this paragraph has a close kinship with Plat. Phaedo 72 b–d.

569. motus … exitiales: 'the motions which lead to destruction'; cf. vi. 566 exitiale aliquod tempus.

itaque: the reference cannot be to the previous paragraph, as this is not an inference from the infinite number of atoms of each shape, but rather to the general picture of the great pelagus of moving atoms, which Lucr. has before his mind. It is one of his 'suspended' thoughts.

570. salutem: 'life', 'existence'. Cf. ii. 863 fundamenta quibus nitatur summa salutis and iii. 125 ex aequo fulcire salutem.

571. genitales auctificique: both adjectives with rerum; the forces which produce things and cause them to grow. For the gen. see ii. 548. auctifici is a ἅπ. λεγ‎. and a typical Lucretian compound adj.

573. aequo … certamine: the contest is equal, neither side wins. So of the contest between sun and sea v. 392 tantum spirantes aequo certamine bellum magnis inter se de rebus cernere certant.

575. vitalia rerum: not, as Munro takes it, for primordia vitalia rerum, for the atoms in themselves are neither vitalia nor exitialia, pg 895but only their motions; nor quite, as Ernout suggests, equivalent to res vitales, like opaca domorum ii. 115, etc. vitalia stands by itself 'the vital forces of nature'. It is almost 'the forces which give life to things', like rerum genitales auctificique motus 571. See Prol. V b, § 4. 2.

577. visentes: the MSS. have visentis as in i. 808 animantis for the nom. But this can hardly be taken as evidence of a grammatical form.

579. vagitibus aegris: 'the sickly cries of the child': Lucr. is comparing the happy event of birth with the sorrows of death, and for the moment, as he remembers the sad wail of the child, the poet, as Giussani says, gets the better of the philosopher. O has aegris, Q aeris, but it spoils a fine touch to alter with Wakefield and Brieger to aegros, or with Purmann and Diels to acris, going with ploratus.

580. comites: cf. iii. 290 comes formidinis aura, vi. 1158 anxius angor … comes.

## (c) The variety of atomic combinations and the differences within species. 581–729.

This long and rather discursive section completes the discussion of atomic forms and their effects. Starting from the general proposition that the greater the diversity of atomic shapes in the composition of a compound, the greater its powers (586–8), Lucr. refers to the earth, which has in it the seeds of all things and is therefore the supreme example of a compound containing all atomic shapes (589–99). This leads to the fine digression on the worship of the earth as the Great Mother (600–60), which gives occasion for a statement as to the life of the gods (646–51). Returning then to the main subject Lucr. passes on to the difference of individuals within species (661–99) and argues that all these slight variations and modifications are due to difference of atomic shapes. Finally he maintains that not every combination is possible and that things must remain true to their species (700–29).

But, as Giussani has shown, the logical divisions of the section do not quite correspond to its paragraphs—or at least there is a cross-division. He argues for the variety of atomic shapes on two different grounds. First from the diversity of their products, and this section runs from 589 to 679; then more briefly from the variety of their qualities (680–5). And in all this the paragraph on the Great Mother is, as Giussani says, a kind of dedicatory hymn. This cross-division, though perhaps hardly conscious, should be borne in mind.

## 1. The mixture of atomic shapes and the resulting powers. 581–99.

581. It must be borne in mind that nothing in perceptible nature consists of only one class of atoms; in all things they are mixed. 586. And the more powers a thing has, the greater the variety of pg 896atoms of which it is composed. 589. The earth contains in it the seeds of water, which make the sea, and of fire, which smoulders below the surface or bursts out like Etna. 594. It can produce crops and trees for man's benefit and streams and pastures for the beasts. 598. Rightly is it called the mother of gods, the mother of beasts, and the parent of our human life.

The statements in this paragraph seem a little arbitrary and Lucr. does not adduce any immediate arguments for them. But it is not difficult to see how they arise from his general conception of the atoms and their behaviour. A compound consisting of atoms of only one class—notice that genus (584, 588) has become the technical word for atoms of the same shape—would have the least possible variety of combination and movement and would therefore be almost incapable of action. On the other hand, one with a large number of classes would have all sorts of means of combination between the atoms and all sorts of movements caused by the different kinds of deflexions after collision: it could therefore perform many different types of action and produce many different kinds of things. The earth—we must always remember Lucr.'s wholly geocentric conception—has therefore in it the seeds of all its various products.

In connexion with the first statement it is worth remembering Schömann's view (Opusc. iv. 341, see Masson, Lucr. Ep. and Poet. 273) that the gods are the sole example of beings created of only one class of atoms. There is no evidence for this view, and even if it were not a contradiction of Lucr.'s statement here (for it might be held that the gods were not among the things in promptu quorum natura videtur), it would at least follow that the gods had no vires or potestates, which is not a view that even Lucr. would have held.

We may notice again how in the last two lines of the paragraph Lucr. according to his frequent habit shows how science and popular belief may be reconciled. The earth containing in it the seeds of all things is in a sense the Mother of all things; popular belief has personified her in this character. The personification is false (ii. 652–4), but the idea has a philosophic basis. And so he passes to the great description of the worship of the Earth Mother.

581. obsignatum: 'sealed', 'certified', a metaphor from a legal document. So Ovid, Her. xiii. 66, perhaps recalling this passage, signatum memori pectore nomen habe. It is certainly imitated by Arnobius, Nat. ii. 6 obsignatum memoria continetis.

582. memori … mente: a periphrasis for the unmetrical memoria, as frequently in the poets.

583. nil esse … quorum: for nil esse eorum … quorum. Cf. i. 883 aliquid nostro quae corpore aluntur.

in promptu: 'manifest', sc. perceptible to the senses, cf. iii. 185, vi. 941.

584. genere: 'class', i.e. group of atoms of one shape: so 567 and 588.

pg 897

585. semine: here a collective singular, equivalent to principiis.

586. quodcumque: a necessary correction of Lachmann's for quaecumque OQ, which is universally accepted. Purmann suggested quo quicque which would make magis easier.

vis: acc. plur., cf. v. 1033, used also for the nom. in iii. 265 sed quasi multae vis unius corporis exstant. It is the older form, vires being formed by analogy (Ernout); see Prol. V a, § 1.

588. genera: see 584 n.

589. principio is not picked up till saepe itaque (661), where we get further examples of the difference made in compounds by formation from different classes of atoms. Or, if it be taken to introduce the series of examples of difference of product, it is answered by denique (680), where he passes to differences of quality: see introductory note.

590. volventes frigora: a typical Lucretian expression; volventes frigidam aquam would be the prose equivalent.

591. habet: sc. corpora prima from 589.

593. ex imis: 'from the depths'. Giussani is right in retaining the MS. reading here and supposing that there is a contrast between the places where succensa … ardent sola terrae, 'fire is smouldering just below the surface', and the fire of the volcanoes which comes from deep within the earth. Ernout objects that the expression ex imis is unparalleled, but we may compare v. 163 ab imo evertere summa and for the thought Virgil, Aen. iii. 577 fundoque exaestuat imo. All other editors have accepted the emendation of Avancius eximiis, 'Etna rages with conspicuous fire', but the sense is weak. In 607 this correction is necessary, but it does not follow that it should be made here; the corruption in 607 may be due to a recollection of the correct ex imis here.

impetus Aetnae: 'bursting Etna': another characteristic Lucretian periphrasis. Cf. vi. 281 gravis ignis impetus. Etna seems to have made a great impression on Lucr.'s mind: see i. 722 and again vi. 669. He does not mention Vesuvius, which is said not to have been very active before the great eruption in a.d. 69.

594–5. Compare and contrast ii. 169–70. It is the earth by its own atomic processes which supplies men with the kindly fruits, and not the gods.

596. So in ii. 875 vertunt se fluvii frondes et pabula laeta in pecudes and compare the herbas and latices of i. 385–6. In all these places the fluvii are appropriate as the drink of the beasts, and here Lucr. has already said (590) that the fontes are the direct product of the earth.

597. montivago: cf. i. 404 montivagae … ferai and ii. 1081 montivagum genus … ferarum.

598. magna deum mater: Magna Mater was the official title of pg 898Cybele in the Roman cult, which Lucr. is about to describe. Strictly as an Epicurean he could not say that the earth was the 'mother of the gods', for their bodies are created and they live in the intermundia. The expression recurs in ii. 659; cf. also v. 795 linquitur ut merito maternum nomen adepta terra sit, e terra quoniam sunt cuncta creata.

599. nostri … corporis: a rather strange phrase, to emphasize the distinction between the human race and the animals.

una: 'alone', but possibly in the sense of una eademque or unā = simul, cf. iii. 901 (Maas).

## 2. The Earth Mother and her cult. 600–60.

600. Greek poets have told us that the Earth Mother was enthroned in a car drawn by lions to show that the earth was suspended in the air; the lions were yoked to the car to show that even the wildest beasts must be tamed by the care of their parents. 606. The mural crown on her head signifies her protection of cities. 608. Wearing it, the image of the Mother is carried through many lands, and many different peoples call her the Mother of Ida and give her Phrygians for companions, because they say that crops were first grown in Phrygia. 614. The mutilated priests they add to show that those who have violated the mother's deity and shown ingratitude to parents are unworthy to have progeny. 618. They beat drums and cymbals and blow horns and pipes and carry the knives which symbolize their madness, to terrify the ungrateful and the impious mob with the goddess's divinity. 624. So when she moves through great cities blessing men in silence, they strew her path with coins and scatter roses on the goddess and her followers. 629. Warriors in arms, called Curetes, leap rhythmically and stained with blood shake the crests on their heads. 633. They call to mind the Curetes of Dicte, who are said once to have drowned the cries of the baby Iuppiter in Crete, boys dancing round a boy, armed and clashing their cymbals, lest Saturn should seize and devour him and cause the Mother an unending grief. 640. This is why they follow the Mother in arms, or else it is to show that the goddess enjoins them to defend their country and protect their parents. 644. All this, though it is beautifully set out, is far from the truth. 646. For the gods enjoy their immortality in peace far removed from our affairs. Free from all pain and danger, mighty in their own strength and needing no help from men, they are neither pleased by service nor moved by anger. 652. The earth has never any feeling, but because it possesses the atoms of many things it sends its manifold produce to the light. 655. So if men choose to call the sea Neptune and corn Ceres and wine Bacchus, they may speak of the earth as Mother of the gods, provided they do not taint their minds with false superstition.

pg 899This passage is a digression, for, except that it is all an amplification of the conclusion of the preceding paragraph, that the earth deserves the title of the Great Mother in virtue of its possession of the seeds of all things known upon earth, it does not bear on the general argument of this section. It is nevertheless of great importance in relation to Lucr.'s attitude towards the gods and to religion, and three main points may be noted. (1) 646–51 (= i. 45–9) contain the most explicit statement in the poem about the gods. Their life is lived in untroubled peace (in the interspaces between worlds, as we learn from other sources),1 where they know no pain or fear or care; they have perfect ἀπονία‎ and perfect ἀταραξία‎; they have no concern for the affairs of this world; they are not moved by man's prayers nor stirred to anger by his sins. This is the cardinal point in Epicurus' religious view, and it was to establish that belief and so to free man from fear of the gods that Lucr. wrote his poem. (2) The earth and all that is in it is not merely devoid of any divinity, but is entirely without sense (652, cf. v. 122–5, where he makes a similar affirmation about the stars); it is a collocation of atoms working by their own laws. Not only, then, are the beliefs of popular religion false, but equally the pantheism of the Stoics, who believed that the earth was permeated by the divine anima mundi. But (3) the myths of popular religion may to some extent be reconciled with philosophy, if they are interpreted as allegorical expositions of physical or moral truths. Detail by detail Lucr. works out his allegorical interpretation of the conception of Cybele and her worship. And in the last lines of the paragraph (655–60) he sums up his position: we may use these allegorical names for earthly things if we will, provided that we do not allow ourselves to be contaminated by 'religio', by the belief that these 'deities' are persons who can be influenced by the prayers or actions of men. For this conclusion we may compare v. 1194–1203, where Lucr. teaches the same doctrine and once again shows his interest in the details of ritual.

The individual points in the ritual of the Great Mother must be dealt with as they are met in the text. But certain general considerations should be had in mind throughout what is in fact a rather complex and difficult description.2 The ancient worship of Cybele, the Magna Mater, originated in Phrygia. It was there conducted by Phrygian priests, known as the Corybantes: most, if not all of them, were mutilated and known in that connexion as galli. Closely related to their mutilation was the legend of the self-pg 900mutilation of Attis, the beloved of Cybele, and the dramatic representation of his story was part of the celebration of Cybele's ritual. The worship was orgiastic and involved ritual processions, such as Lucr. describes, in which the image of the goddess was carried in her lion-drawn car, while round it the priests made loud music and danced wild and bloody dances.

Besides this Phrygian cult, a great female goddess was worshipped in Crete in much the same manner by priests known as Curetes, who, if not so wildly orgiastic in their ceremonies, yet danced and leapt and made loud music. The Cretan goddess too had associated with her a young male god, and when the Greeks came to know the cult, they identified the great goddess with Rhea and her son with Zeus; the Curetes were now supposed to be leaping and dancing, as Lucr. tells us, to hide from Kronos (Saturnus) the cries of the infant Zeus.

At some early period, as navigation and commerce established relations between the nations of the Mediterranean and the Aegean, these two cults became confused and even to some extent amalgamated. The likeness of the conception of the Great Mother was doubtless the basis of the assimilation, which was assisted by the coincidence of the names of Mount Ida in Phrygia and Mount Ida in Crete—each a seat of the worship of its own Earth goddess. The influence seems to have been mutual, but partial. The Curetes of Crete came to be known as Corybantes (Eur. Bacch. 120–5 and Ov. Fast. iv. 210, see n. on 629–31), and though there is no evidence (apart from a possible interpretation of Lucr. ii. 629–31) that the priests of the Phrygian goddess were known as Curetes, yet the rhythmical dance of the Curetes seems to have been incorporated into the ceremonial of the Phrygian goddess. And in the Greek and Latin poets the names of Cybele and Rhea were almost interchangeable.

In 202 b.c. by the command of the Sibylline Books the ancient black stone, which was the symbol of the Great Mother, was brought to Rome from Pessinus and her cult established as part of the State religion.1 The Megalesia were celebrated in March, reproducing the wild processions and their accompaniments. But the Roman State, fearful of the emotional effect of the cult, made the regulation that no Roman should take part in it; it must still be performed by the Phrygian priests. It is probable, as Giussani sur-pg 901mises, that outside Rome the cult was also performed in the Italian towns by itinerant bands of worshippers. Lucr. probably has in mind both the celebrations at Rome and those up and down Italy, magnas … per urbis (624), and it may be that it was more in the latter than in Rome itself that the bystanders took the active part described in 624–8.

600–7. For the general description in these lines compare Virgil's imitated and abbreviated version: Aen. x. 253 alma parens Idaea deum, cui Dindyma cordi, turrigeraeque urbes, biiugique ad frena leones.

600. Cf. vi. 754 Graium ut cecinere poetae. The double epithet veteres … docti is awkward, but probably docti poetae coalesces into a single word, and veteres is added. But is docti here a merely conventional epithet, or has it a more definite meaning? In other words, is Lucr. merely thinking of poets like Homer, Pindar, Sophocles, Euripides, etc., who referred to Cybele and her ritual, or has he in mind some Greek poet or poets who offered an allegorical explanation of the ritual? docentes 602 is in favour of the latter view, and the coincidence of allegorical explanations given by Varro, Ovid, and Lucr. (see notes following) suggests that they may have been drawing on a common Greek source.

601. After 600 Q leaves a space of two lines, which has been interpreted as suggesting a lacuna; the editors have also noticed that sedibus without an epithet is abrupt. Lambinus emended to sublimem; more modern editors have made other changes, but none seems worth recording. Lachmann, noting the space in Q, proposed to fill it with the line magnifice divam ex ipsis penetralibu' (sc. sedibus) vectam, Diels with far more probability, as far as the sense goes, suggests montibus egressam Phrygiae celsisque solere (sc. sedibus). For there can be little doubt that sedibus refers to a throne on which the goddess sat in the car. Fr. Préchac (Rev. Num., 1932, 119) has discussed this passage in connexion with coins of M. Volteius, which can

safely be dated anterior to Lucr., on which the Magna Mater, recognizable by her mural crown, is represented as seated on a chair in a car drawn by two lions. Gifanius long ago interpreted sedibus in curru as sella curuli (which almost certainly stands for sella currulis), the seat in the chariot occupied by magistrates in processions; cf. Liv. 28. 9. 15 iret alter consul sublimis curru. In support of this view may be quoted the words of Varro (see n. 1, p. 900) pg 902quod sedens fingatur, and the evidence of Ovid, who twice in his description (Fast. iv. 185 and 345) speaks of the goddess as ipsa sedens. It is also strongly confirmed by 603; they placed the goddess on a throne in the car to show that the earth was suspended in air, not propped by another earth. Grammatically there is no difficulty about sedibus without an epithet; it might be compared with versibus in ii. 529, or more probably, as Préchac explains, in curru = curulibus acts as the epithet; cf. ii. 51 fulgorem … ab auro (see n. there). Nor is there any need to suppose a lacuna; the blank in Q may be accounted for, as Préchac suggests, because the scribe of Q, which has not the capitulum De Matre Magna after 597, may have been doubtful where it should come and left space for the rubricator in both places. This suggestion is confirmed because Q begins 601 with edibus, s being added in the margin, as though for the capital letter which would be inserted after the capitulum by the rubricator, just as he begins 598 with uare, leaving space for q. I therefore retain the MS. reading as it stands without a lacuna: so Martin.

agitare: 'driving'; cf. Virg. Georg. iii. 18 centum quadriiugos agitabo ad flumina currus.

602. aeris in spatio … tellurem: for this allegorical explanation cf. Varro quoted in n. 1, p. 900. Lucr.'s own explanation of the suspension of the earth in air is given in v. 534–63.

603. 'And that earth cannot rest on earth'; therefore the goddess must not be represented on the ground, nor even in a car, but on a throne in a car: see 601 n. For the phrase cf. i. 1057 ipsum si quicquam posse in se sistere credis.

604–5. Cf. Ovid, Fast. iv. 215 'cur huic genus acre leonum praebent insolitas ad iuga curva iubas?' … desieram. coepit: 'feritas mollita per illam creditur; id curru testificata suo est'. Varro's explanation is slightly different.

605. molliri: Itali, a certain correction for moliri, confirmed by Ovid's mollita.

606–7. Cf. Varro, l.c. quod turris in capite, oppida; Ovid, Fast. iv. 219 at cur turrifera caput est onerata corona? an primis turres urbibus illa dedit? The 'mural crown', said to be imitated from Cybele's, was the reward for the soldier who first scaled the walls of a besieged city. For caput summum cf. v. 1138 capitis summi praeclarum insigne, where it is used in a less literal sense.

607. eximiis munita locis: 'embattled on glorious places': the epithet is transferred from the cities to the deity; see Prol. VII, § 12. Giussani aptly recalls the many Italian fortified towns perched on a hill-top.

eximiis is here a certain correction for ex imis; see 593 n.

608. per magnas … terras: 'through the wide earth' (Merrill), pg 903the plural being used as in ii. 215. More probably 'through great countries' with reference to the variae gentes (610) and the magnae … urbes (624) in which the Great Mother was worshipped.

609. horrifice: 'with terrifying effect'; cf. 632 terrificas. One can imagine the sensation which the orgiastic procession must have made on the minds of Roman spectators, accustomed to a much more sober and restrained ritual.

610. variae gentes: 'different nations'; Lucr. is presumably thinking of Phrygians, Cretans, and Romans, among whom this same cult was observed.

antiquo more sacrorum: 'keeping up the ancient rites'.

611. Idaeam … matrem: sc. the Mother from Mt. Ida in Phrygia, but as has been said in the introductory note, confusion with Mt. Ida in Crete helped towards the identification of the two cults. Lucr. is here insisting on the Phrygian cult as the origin of the worship. Mater Idaea was the official title of Cybele: Liv. xxix. 10. 5, quoting the Sibylline Books, si mater Idaea a Pessinunte Romam advecta foret; cf. other passages quoted in Munro's note.

612–13. The association of Cybele with the crops suggests the influence of the Demeter legend, with which also there was some confusion. For the idea that the corn-crops were first produced in Phrygia Merrill quotes the story in Hdt. ii. 2 of Psammetichus' discovery that the Phrygian word for 'bread' was the oldest.

612. ex illis finibus: sc. Phrygiae.

613. coepisse creari: in classical prose coeptas esse would have been written with a pass, infin. This irregularity occurs elsewhere and it does not help to say with Munro that creari = nasci. Elsewhere Lucr. is scrupulous about this construction and even uses queatur in i. 1045 and potestur in iii. 1010.

614–15. Varro (l.c.) again gives a different explanation: quod gallos huic deae ut servirent fecerunt significat qui semine indigeant terram sequi oportere; in ea quippe omnia reperiri.

614. gallos: the self-mutilated priests. There can be little doubt that they were in origin the priests of Attis who mutilated themselves in imitation of their hero (cf. Ovid, Fast. iv. 221 ff.), and then, when the cult of Attis became assimilated with that of Cybele, she took over the galli, retaining at the same time her own Corybantes, who were probably not mutilated. The origin of the name is obscure, but it is probably not due either to the 'Gauls' of Galatia or to the river Gallus in Phrygia, both of which explanations were current in antiquity: see Thesaurus ad voc.

numen … inventi sint: the reason Lucr. gives is not clear. Matris is certainly Cybele, as is shown by numen, but how had they violated her numen? By their ingratitude to their parents, says Munro, and the same idea seems to be suggested by 641–3. The good man will take up arms for his country pg 904and defend his parents; if he does not, he is not fit to have children himself. But Giussani is also probably right in seeing an allusion to the story of Attis, who spurned the love of Cybele, violavit numen Matris, and so was afflicted with madness, during which he mutilated himself; but this will not explain ingrati genitoribus. The former moral explanation must have been more prominent in his mind.

sint: Lachmann's correction for sunt OQ is necessitated by violarint, but his alteration of the MS. order to sint inventi is gratuitous.

617. quĭ in: cf. qu$ae⌣$ amara ii. 404 n. and Prol. VI, § 13 (a).

618–20: cf. Cat. lxiv. 261 plangebant aliae proceris tympana palmis, aut tereti tenuis tinnitus aere ciebant; multis raucisonos efflabant cornua bombos barbaraque horribili stridebat tibia cantu, and lxiii. 9 typanum, tubam Cybelles, tua, mater, initia, quatiensque terga tauri teneris cava digitis; Ovid, Fast. iv. 183 ibunt semimares et inania tympana tundent, and Met. iv. 29 impulsaque tympana palmis concavaque aera sonant. The coincidences with Lucr. in all these passages are obvious and it is noticeable that Ovid is at pains to keep up the alliteration which runs through Lucr.'s description; see Prol. VII, § 18 and Addenda.

tenta: 'taut', of the hide stretched on the drum.

palmis: the tympanum was played with the flat of the hand; cf. Cat. lxiv. 261 and Ovid, Met. iv. 29 quoted above.

cymbala circum: so Virg. Georg. iv. 64 tinnitusque cie et Matris quate cymbala circum, an obvious imitation in a very different context (the methods of causing bees to swarm).

619. raucisono: cf. v. 1084 of birds raucisonos cantus and raucisonos … bombos in Cat. lxiv. 263.

620. Phrygio … numero: the tibia was the Phrygian instrument, always opposed to the Greek lyre. The phrase may perhaps be technical 'in the Phrygian mode', as in v. 1409 numerum servare genus is 'to keep to one class of rhythm'.

621. tela: probably the ἅρπη‎, the Asiatic sickle or pruning-knife, which would be used for the emasculation (violenti … furoris).

praeportant: so in Cic. Arat. 430 Scorpios infestus praeportans flebile acumen. prae seems to have a menacing effect.

622. ingratos animos and impia pectora no doubt refer back to 614–15.

623. numine OQ should be kept; 'to terrify them with fear by the power of the goddess'. Editors, thinking that the two ablatives metu and numine are awkward, have accepted the suggestion of Havercamp numini', but see i. 473, ii. 18 n. The two genitives numini' divae would be hardly less awkward. Diels retains numine as a gen.

624. magnas invecta per urbis: 'borne into one great city after another'; cf. Virg. Aen. vi. 784 qualis Berecyntia mater invehitur curru Phrygias turrita per urbes, and Ovid, Fast. iv. 345 ipsa sedens plaustro porta est invecta Capena. The use of invecta and especially pg 905the passage in Ovid suggest that the procession would start outside a city and enter through a gate. For magnas per urbes cf. 608 per magnas … terras; Lucr. is here probably thinking of the itinerant worshippers passing from one city to another in Italy.

625. 'In silence blesses mortals with dumb salutation'; in reference to the image of Cybele in the chariot, or in Rome itself possibly to the actual black stone. There is no doubt an intentional contrast with the noise made by the worshippers, but it is not necessary with Giussani to regard this peculiarly solemn line as ironical; Lucr. is throwing himself into the mood of the worshippers and bystanders.

munificat: a ἅπ. λεγ‎.

626. sternunt iter omne viarum: cf. v. 739 Flora quibus mater praespargens ante viai cuncta … opplet and v. 1124 iter infestum fecere viai. iter omne viarum is a certain correction of OQ's corruption, for for which the earlier editions made poor substitutes.

627. largifica stipe ditantes: stips 'a small coin', so almost 'offertory', is a word particularly associated with religious ceremonies such as the stips of Ianus given on 1 January. It is specially connected with the offerings to the Magna Mater; cf. Cic. De Leg. ii. 9. 22 praeter Idaeae matris famulos eosque iustis diebus ne quis stipem cogito. Ovid, Fast. iv. 350 gives an historical, but probably spurious, reason for this permission: 'dic', inquam, 'parva cur stipe quaerat opes'. 'contulit aes populus, de quo delubra Metellus fecit', ait; 'dandae mos stipis inde manet.' Among the followers of the goddess were certain μητραγύρται‎, 'collectors', who were permitted on specified occasions to collect the offertory.

largifica: another ἅπ. λεγ‎. and a good Lucretian compound.

ninguntque rosarum floribus: 'they snow with roses', an unparalleled personal (and, if it governs iter omne, transitive) use of the verb. Ernout quotes for the verb from Pacuvius (Non. 507. 29) sagittis nivit, plumbo et saxis grandinat. In vi 736 Lucr. has the substantive ningues, a nasalized form standing to nives as ninguit does to nivit.

629–43. This description of the armed dance and its relation to the dance of the Curetes in Crete comes in oddly here after the picture of the conduct of the bystanders, which might seem to be the natural conclusion of the account of the procession. Giussani is inclined to regard 624–8 as a later addition, it being impossible to take that view with regard to 629–43 owing to the close connexion of 643 and 644. Taking the passage as it stands, it is not quite easy, partly through the uncertainty of the text and partly from our ignorance of the proceedings, to understand exactly what Lucr. is saying. It would appear that in the course of the procession 'an armed band'—whether those already described or a special group of worshippers—danced a rhythmical dance, cutting themselves with knives, and decorated with plumes nodding from their helmets, and pg 906that this band specially recalled the Curetes of Crete, who were said to have guarded the young Zeus.

629–31. Curetas … ludunt: the main textual difficulty of the passage is concerned with these lines. The MS. text is Curetas nomine Grai quos memorant Phrygios inter se forte catervas (O, catenas Q, a mere mistake) ludunt. It is clear that the main verb going with hic armata manus is referunt 633, and that not only quos memorant is subordinate, but that ludunt and exsultant should also be in a subordinate clause. Lachmann, supposing that catervas at the end of 630 was a dittography of the last word of 628 (cf. ii. 422), read quod armis, thus supplying in quod the conjunction to govern ludunt. This Munro describes as 'a certain correction, the sentence requiring the conjunction quod or quia, the sense armis', and it has been adopted by all subsequent editors before Diels. We then get the sense 'here an armed band, which the Greeks call by the name of the Phrygian Curetes, because from time to time (forte) they sport among themselves in arms and dance rhythmically … recall the Curetes of Dicte'. This makes sense, but its main difficulty lies in the statement that the Greeks called this armed band the 'Phrygian Curetes'. There seems to be no reason why they should, and while there is evidence that the name Corybantes was applied to the Cretan worshippers (see Eur. Bacch. 120–5 and Ov. Fast. iv. 210 hoc Curetes habent, hoc Corybantes opus), there is no mention anywhere of the application of the name Curetes to the Phrygians. Diels's solution appears to me (as to Martin) more probable. He retains catervas, accepts the suggestion of Frerichs Phrygias for Phrygios, places a comma after memorant, and reads si (sei) for se. We then get a much more probable statement: 'here an armed band, whom the Greeks call Curetes, whenever (si forte) they dance and leap rhythmically among the Phrygian throng, recall the Curetes of Dicte'. In this case we may suppose a special band who accompanied the Corybantes and galli and at times during the procession performed this armed dance which, originating in Crete and associated there in legend with the protection of the infant Zeus, had been incorporated in the Phrygian procession, but retained the mark of its origin in the preservation of the name Curetes for the dancers. The passage thus gains greatly in lucidity and directness of expression and makes the intrusion of 629 ff. much less objectionable, and for the repetition of catervas after so short an interval there are many parallels in Lucr.: see Prol. VII, §§ 21, 22.

631. in numerum: 'in time'; the dance was to a regular rhythm. For the phrase cf. ii. 637, iv. 769, 788, and in v. 1401 extra numerum.

sanguine laeti is on the whole the best emendation for sanguine flaeti OQ; they bespattered themselves with blood and were excited by it. Of the alternatives pleti is improbable, also pleni (Lane, pg 907Harvard Studies, ix. 16), and Bentley's sanguinolenti is too far from the MSS.

632. numine: 'with nodding'. It is true that this literal sense is almost unknown in Latin and that Varro, Ling. Lat. vii. 85, says, numen dicunt esse imperium dictum a nutu, but in Cat. lxiv. 204 annuit invicto caelestum numine rector there must at least be an allusion to the original sense and Lucr. approaches it again in iv. 179 in quem quaeque locum diverso numine tendunt. Lachmann refuses to admit this sense and changes in both places to momine. But seeing that the MSS. have in several places rightly kept momine, it would be odd that in these two they should have corrupted it. This line is repeated in v. 1315 with the substitution of undique for numine.

633. Dictaeos: of Mount Dicte in Crete, in a cave of which the infant Zeus was said to have been guarded from Kronos. Cf. Virg. Georg. iv. 151 canoros Curetum sonitus crepitantiaque aera secutae Dictaeo caeli regem pavere sub antro, and Ovid, Fast. iv. 207 ardua iam dudum resonat tinnitibus Ide, tutus ut infanti vagiat ore puer.

illum: 'that fabled cry'. Cf. ii. 362 fluminaque illa and i. 82 illa religio.

635. pueri: sc. the Curetes, the word being intended to suggest κοῦροι‎. The recently discovered hymn of the Curetes shows that it is addressed to the μέγας κοῦρος‎. It is probable that they were in origin the worshippers of the κοῦρος‎, as the galli were of Attis, and were similarly attached later to the Great Mother of Crete.

pernice: for this form of the abl. cf. simplice i. 1013 n., and see Prol. V a, § 1.

chorea: 'dance' = χορεία‎, a poetic word first used by Lucr. and Catullus.

636. After 635 the MSS. have the line armat et in numerum pernice chorea, obviously conflated from the two lines on either side of it.

637. aeribus aera: cf. Ovid, Fast. iv. 184 aeraque tinnitus aere repulsa dabunt.

638. malis mandaret: cf. Accius twice quoted by Cicero, De Or. iii. 217 and Tusc. iv. 77 ipsus hortatur me frater ut meos malis miser mandarem natos; the best MSS. of the Tusc. have mandarem but those of De Or. manderem, which seems the more natural word; cf. Sen. Thyest. 779 artusque mandit ore funesto suos. Does mandaret mean 'entrust to his jaws', which is a very unnatural expression, or is it from a first-conjugation form of mandere? There are similar conjugation variations in Lucr., but not between the 1st and the 3rd; yet it is tempting to suppose that Lucr. meant 'champ', 'devour'.

639. aeternum … sub pectore vulnus: cf. i. 34 n. matri not only as the Earth Mother, but because she was in the legend the mother of Zeus.

641. aut quia …: after the legendary explanation of the dance comes the moral and allegorical explanation. The idea of defence pg 908of parents has been already suggested by 614–15 and 622 and to this is added defence of country. It is a far-fetched moral from the performance of the Curetes.

praedicere: 'to enjoin' here and not, as usual, 'to prophesy'.

643. praesidioque … decorique parentibus esse: cf. iii. 897 non poteris factis florentibus esse tuisque praesidium. Was Horace consciously recalling this line in Od. i. 1. 2 o et praesidium et dulce decus meum? There are other Horatian imitations of Lucr.: see e.g. iii. 938 n. For the jingle părent … parentibus see Prol. VII, § 24.

644. What is the reference of quae? is it to the supposed precept of the goddess to defend country and parents? or is it more widely to the whole ceremonial just described? The use of disposta in i. 52 mea dona tibi studio disposta fideli suggests the former and it fits better with 645; it is a beautiful thought that the goddess should so instruct mankind, but it cannot be true, for the gods have no care for human affairs.

646–51. This famous description of the life of the gods occurs also in the MSS. in Book i. 44–9. After much hesitation I have retained the passage in Book i. See comment and notes there, also Prol. VII, § 30.

652–4. Lucr. here returns to the main idea resumed from 598–9 that the earth, though it has no feeling or divinity, may be allegorically called the Mother, because it contains the seeds of all things. It looks like the concluding passage of the paragraphs and if it were, would provide an immediate antecedent for itaque 661. Munro accordingly transfers these lines after 660 (680) and Giussani brackets 655–60 as a later addition. But it may well be that the poet first finished off what he had to say about the earth, then added some lines about similar popular personifications and then returned to the earth at the end. The case is a little like that of i. 208–14.

652. caret … sensu: cf. v. 122–5 and see introductory note to this paragraph.

653. potitur primordia: for construction of potior with acc. cf. iii. 1038 sceptra potitus, iv. 760 quem … terra potitast. In ii. 13 rerumque potiri it takes gen. Lucr. has similar variations between acc. and abl. with fruor. See Prol. V b, § 3. For the 3rd conj. form potĭtur see Prol. V a, § 12.

655. Lucr. is probably thinking here not only of popular and poetic usage, but, as Robin suggests, of the serious attempt of the Stoics to regard the names of gods as the personification of things, esp. earth, sea, and sky. See Cic. De Nat. Deor. i. 15. 40 idemque (Chrysippus) disputat aethera esse eum quem homines Iovem appellarent, quique aer per maria manaret eum esse Neptunum, terramque eam esse quae Ceres diceretur, similique ratione persequitur vocabula pg 909reliquorum deorum. Lucr. himself speaks in the popular sense of Neptune in ii. 472 and vi. 1076, and of Bacchus in iii. 221.

656. constituet: Lachmann, in order to preserve parallelism with mavult 657, changes to constituit, but Lucr. is not rigid in such matters.

abuti: 'to misuse'.

657. vocamen: a Lucretian formation with his favourite termination in -men; see i. 435 n.

659. deum matrem: as in 598 and with the same laxity of expression.

ipse: probably not simply picking up siquis 655, or with animum (Munro) = animum suum, but, as Ernout thinks, with vera re; he must not allow his real self to be contaminated in mind by these allegorical personifications.

660 (680). The line was transferred by Lachmann from 680, where it makes no sense, to this place where it is badly needed. On his calculation it was the last line of p. 73 of the archetype and might have been put in there by the scribe, when he noticed that he had omitted it.

religione … turpi: 'by degrading superstition', for religio see i. 63 n.

contingere parcat: for the prolate infin. see Prol. V b, § 12 (a).

parcat is oddly corrupted in O and Q.

## 3. The atomic shapes in the earth account for species, qualities, etc. 661–99.

661. (1) So it is that sheep, horses, and oxen pasturing in the same field and drinking the same stream under the same sky are different in kind and preserve the character of their parents. So different is the matter in each kind of grass and in every stream. 669. (2) So again each living thing is made up of bones, blood, veins, flesh, sinews, heat, and moisture, which differ owing to the different atoms which constitute them. 673 (3) All things which can be ignited have in common in their bodies the elements which can produce fire and light, sparks and ashes. 677. So you will find that all things conceal in their bodies the seeds of many things and many atomic shapes. 680. (4) Once again, there are things which have many qualities, colour, taste, and smell. They too must be made of different atoms: for since smell, colour, and taste touch the senses in different ways, they must vary in their atomic shapes. 686. Different forms then gather together in one compound and things are made of mingled seeds. 688. Even so in my verses you see many letters common to many words, yet verses and words contain different letters; not that there are only a few common to all, or that no two words are composed of the same letters, but all are not alike to all. 695. So though there are many atoms common to many things, yet the whole is in each case different; men, crops, and trees are rightly said to consist of different atoms.

pg 910This is a paragraph which has many links with different parts of the poem: it 'looks before and after'. Primarily it is a resumption of the discussion broken at 660 by the description of the processions of Cybele and amplifies the argument there. The earth, which contains the seeds of all things, can produce many different species of animals (661), can create all the different parts which unite to form a single body (669), can place the seeds of fire in things (673), and endow things with various qualities which appeal to our senses (680). All this is an elaboration of 581–99, the last being an additional point. But there are other links as well. The section about the difference of species goes back in thought to the first discussion of the differences of atomic shape in 333–80, and in the lines on heredity (665–6) anticipates the further discussion in iii. 741–7. The lines on the varying qualities of things (680–5) look back to the passage on the different effects on sensation of different atomic shapes (408–43) and onwards to the discussion on secondary qualities which forms the third main section of this book (730–1022). In 688–99 Lucr. emphasizes the general conception with the familiar illustration of letters in words, and in 692–4 introduces the new idea that although not all combinations are possible, yet any given class of atoms can enter several different kinds of created things. This idea is picked up again in the next paragraph (see introductory note there).

There is much, too, both in phraseology and in thought in this paragraph which carries one back to the discussion of the theory of Anaxagoras (i. 830–920), and, as Robin points out, the essential difference between the conceptions of Anaxagoras and Epicurus is here well brought out. Anaxagoras believed every existent thing to be an aggregate of an infinite number of particles qualitatively different from one another in substance; Epicurus held things to be constituted of a large, but finite, number of homogeneous particles, quantitatively different from one another in size, shape, and movement. It is clear from his reminiscences in phraseology and especially from the repetition of the analogy of the letters in words (cf. i. 912–14) that Lucr. had in mind here his argument against Anaxagoras.

661. itaque: the editors complain that there is no immediate reference for itaque, and Munro in order to secure it transferred 652–4 after 660 (see n. there). But it refers to the thought that has been in Lucr.'s mind since 599 and is another case of 'suspended reference'. Cf. iii. 106 saepe itaque … in exactly the same way. See Prol. VII, § 33.

campo … caeli (663) … aquai (684): 'earth, sky, and water'; notice that Lucr. introduces all three elements to account for the nature and growth of the animals. Cf. i. 2–3. It is possibly to secure completeness that he deals with the element of fire in 673–6.

pg 911

662. lanigerae: see ii. 318 n.

duellica: an archaic form; so duellum iv. 968. It is quoted also from Plautus more than once. Compare the variations duo and bis, bonus and duonus. Ernout notes that in antiquity the horse was essentially the animal of war, the ox being used for ploughing, etc.

663. buceriae: so bucera saecla v. 866, vi. 1237, from βούκερως‎. A gratuitous use of a Greek word, rare in Latin outside Lucr.

greges is here feminine, as also in Lucilius. Nonius 228. 22 notes this peculiarity of Lucr.

$eo¯$dem: disyllable. Cf. i. 306 n., and see Prol. V a, § 8, and VI, § 15.

sitim: for this correct form of the acc. from -i stems see Prol. V a, § 1.

664. sedantes: rightly restored by Q1 for sedentes.

665. retinentque parentum: again a necessary restoration by the Itali for retinente parente.

666. generatim: 'each after their kind', as in i. 20, etc.

668. dissimilis ratio: 'difference': a periphrasis where there is no substantive that will scan. In i. 653 and iii. 318 Lucr. uses variantia.

669. quamvis: pronoun, of course, here: 'anyone'.

ex omnibus unam: unam is Lambinus's necessary correction for una. But what does ex omnibus unam mean? Munro takes it 'any one of all the countless living creatures'. But, as Giussani points out, there is no particular reason for emphasizing the individual and he may be right in taking it to mean 'one composed of all its parts', ex omnibus being an anticipation of the catalogue in the next line. Cf. ii. 159 ipsa suis ex partibus unum.

670. Note that calor, like umor, is included as a concrete part of the animal, just as much as ossa, cruor, etc. See n. on vapore i. 491.

671. porro: Giussani, noting that in 669 and 673 porro introduces a new example in each case, objects to its use here, where it does not, and therefore corrects to formis. But he takes distantia to refer to the difference of the various parts of the body; if, however, we take it, as the next line seems to indicate, to refer to the difference of atomic formation within each of these parts, porro is right: 'each moreover of these parts differs internally in the shape of its component atoms'.

673–6 should be compared with the argument against Anaxagoras i. 870–2, 891–2, 897–912. Robin notices that the idea of the seeds which can create fire is that of Epicurus in Ep. ad Pyth., § 102 πυρὸς ἀποτελεστικῶν ἀτόμων‎.

674. condunt (Munro) is probably the best emendation of the MSS. traduntur. Q1 has the obvious correction tradunt, which Diels, followed by Martin, retains, referring to iv. 1222 quae patribus patres tradunt. There, however, Lucr. is referring to the transmission of hereditary qualities. The word would be appropriate in connexion with 665–6, but not here of the seeds of fire contained in wood, etc. Lachmann's celant and Bernays's cludunt would give good sense, but pg 912condunt is nearer to the MSS. The passive -ur termination may be due to cremantur in the previous line.

677. cetera: sc. all the other products of the earth.

mentis ratione: cf. animi ratione ii. 381 n. The senses here cannot give evidence and the mind must reason.

678. igitur: inferential in the apodosis 'then'. For a somewhat similar use see i. 419.

680. denique, as Giussani has pointed out, not merely introduces this as a further proof in this paragraph, but also refers back to principio in 589. Up till now Lucr. has been speaking of the difference in the concrete products of the earth, now he turns to the existence of different qualities in things as a further proof that the earth contains the seeds of all things.

681. cŭm odore: cf. qu$ae⌣$ amara ii. 404 and n. there, and see Prol. VI, § 13 (b).

in primis pleraque poma: I accept with hesitation Bruno's emendation poma for dona: 'the majority of fruits' are then offered as an example of things uniting taste, colour, and smell. They are a good instance, but the clause is a little abrupt. The MSS. text could only be kept if we suppose a line lost, explaining dona. Munro suggests quis accensa solent fumare altaria divom and compares iv. 1237 adolentque altaria donis and vi. 752 fumant altaria donis, but incense and the entrails of animals would not be a good example for Lucr.'s purpose. Lachmann, noting the use of donare for to 'endow with' in ii. 73, v. 1095 and 1215, proposed in privis pluraque dona: 'in special things there are other gifts'. But what other gifts? Sound is the only remaining 'secondary quality' and it would not fit well into the context here. Other suggestions have been made which do not require notice.

683. nidor: usually used of the unpleasant smell of burning, cf. vi. 791 acri nidore offendit naris. Here in a general sense as the equivalent of odor; cf. the closely parallel lines vi. 986–7 nam penetrare alio sonitus alioque saporem cernimus e sucis, alio nidoris odores.

penetrat … in artus: i.e. each of these qualities is perceived by its own sense-organ.

fucus … fucus (684): 'colour', a synonym for color as in ii. 745 nullo circumlita fuco and iv. 84 fucum mittunt. This correction is required by the sense in the place of sucus … sucus OQ, which is Lucr.'s equivalent for sapor as in ii. 845, vi. 987 (quoted above) and in many other places.

684.sorsum: rightly repeated by Gerard Vossius, as in iv. 494 sorsus item sapor oris habet vim, sorsus odores nascuntur, sorsum sonitus.

685. sensibus: again with the meaning 'sense-organs'; see i. 303 n.

primis … figuris: 'in the shapes of their atoms', equivalent to figuris principiorum; or it may possibly be, as Ernout suggests, a mere periphrasis for principiis. In any case cf. vi. 774–5 propter pg 913dissimilem naturam dissimilisque texturas inter sese primasque figuras. Lachmann wrongly changed to privis.

686. glomeramen: 'collection', 'group', 'mass', but without any technical sense of 'molecule'. See n. on ii. 454.

688–99. Giussani places after 724; see discussion there.

688–94. Lucr. has already used this comparison with the letters in words in i. 197, 823 ff., and 912 ff., and these lines are largely composed of those used in i. 823 ff. together with quotations from ii. 336–7.

688–90 = i. 823–5, 691 being varied from i. 826 to fit the context. See notes there.

692–4 are based on ii. 336–7, the first two lines being a variation on 336 and the last line repeated from 337.

692. currat: 'runs through', not unlike its use in v. 1373 olearum caerula distinguens inter plaga currere posset.

693. isdem, Pontanus's correction for idem, OQ, is much better than Lachmann's aut nulli … idem, for which he has to extract the nominative versus.

694. constant: again Lachmann substitutes constent from O1. See n. on ii. 337.

695–8. A complicated sentence: 'so in other things likewise since there are first-beginnings common to many things, yet notwithstanding they can exist with sums different from one another'. The nom. res for possunt is extracted from rebus in 695.

696. verum: the MSS. have rerum which Martin keeps with commas after item and sint, though it is hard to see how he takes it: the atoms cannot consist of different sum-totals of things! Pontanus's verum, going with tamen, is the simplest correction, better than Brieger's eorum (sc. principiorum), going with summa, with res as subject of possunt. But it is quite possible that rerum may be a mere dittography from rerum at the beginning of the line, or due to the very frequent occurrence of primordia rerum at the end of lines; in which case longe of Itali may well be right, or even Merrill's mixta. The choice probably lies between verum and longe.

698. aliis: sc. principiis.

## 4. Not all atomic combinations are possible. 700–29.

700. All seeds cannot be united in all kinds of combinations. For then there would be monstrous creations like the Centaurs, tree-men, Scylla and Charybdis, and the Chimaera all over the world. 707. But, as it is, all things are sprung from a definite seed and as they grow keep to their own species. 710. This must be so since from all foods the appropriate particles are dispersed through the body and enter into the appropriate motions. Alien matter is rejected, and in unseen ways, too, many particles leave the body, which could not pg 914form connexions or join in the vital movements. 718. This applies not only to living creatures, but to all things. For just as all created things differ from one another, each must have a different configuration of atoms. Not that but a few are similar in shape, but all are not alike to all. 725. Further, since the atoms are different, so must also be their intervals, paths, methods of connexion, weight, blows, meetings, and motions, which distinguish not only living creatures from one another, but also earth and sky and sea.

This paragraph is a modification or limitation of the previous paragraph and indeed of the general ideas of the section, and it is again closely connected with other parts of the poem. Hitherto Lucr. has reached the general idea that to make any one thing, especially a living thing, atoms of many different classes must be combined. This might give rise to the idea that any classes of atoms might so enter into combination. But now he sets a limit. Not all such combinations of atomic classes are possible. For if so we might have the classes of atoms appropriate to creating a man combining with those appropriate to a beast or a tree or a fish, and monstrous creatures, such as those told of in legend, might result (cf. v. 837–924). But it has been seen long ago (i. 159–214) that every class of thing has its own seeds, which not only unite to create it, but, joining it later, cause its growth. This can be seen in connexion with food. Of what is eaten such portions as can naturally combine with the body and its members are retained, form connexion with the existing parts, and fall into the appropriate motions; those which cannot do so are rejected, either openly in the excretions of the body or unseen in respiration and other ways. This conception is further elaborated at the end of this Book (1122–43). Lucr. applies it further (718 ff.) not only to living creatures, but to inanimate things as well. Different classes of atoms imply further differences in all the features of a combination, the intervals between the atoms, the methods of their combination, the blows, and the resulting meetings and motions. All this holds good in inanimate things as well as in living creatures and hence come the differences between even the great divisions of the world, earth, sea, and sky. They are composed of different atomic combinations behaving in different ways.

There is one more point involved in this section to which Giussani has drawn attention. Though not all combinations are possible, this does not mean that any given class of atoms can only enter into one class of created things. Many of the classes which create man, for instance, will also take their place in the formation of a beast, or even a tree or a rock. If man, for example, is made of classes abc a cow might be composed of bcd, a tree of ade, a rock of bfg, and so on; there are no exclusively man- or tree- or rock-particles. This is pg 915expressed shortly in the two lines 723–4. It would be far more explicit if, with Giussani, we transferred 688–99, to a position after 724. For we should then get the theory emphasized and treated at considerable length. But, as Giussani himself points out, 725–9 will then be awkwardly separated from what precedes with which they have a close connexion, for the earth, sea, and sky of the last two lines are the examples of the inanimate things first introduced in 718. It is better to leave the lines where they are; they fit well enough and we must suppose that in 692–6 Lucr., as so often, was anticipating a thought which was slightly more in place in the paragraph to follow and then picked it up again in 723–4.

700. omnimodis: see i. 683 n. and Prol. V b, § 6.

701. portenta: 'monstrosities'. In v. 839 ff. Lucr. describes the portenta which nature actually did create: androgynum, interutrasque nec utrum, utrimque remotum, orba pedum partim, manuum viduata vicissim, etc.

702. semiferas hominum species: Lucr. is thinking of the Centaurs, cf. v. 878, where he gives other physiological reasons why such monsters were impossible. For the word semiferas cf. iv. 586 Pan pinea semiferi capitis velamina quassans.

altos … corpore vivo: perhaps Lucr. is thinking of such metamorphoses as those of Daphne, Philemon and Baucis, or the sisters of Phaëthon.

703. egigni: a ἅπ. λεγ‎. The nearest approach to it in Lucr. would be vi. 761, if Lachmann's emendation effiant were accepted for e fiant; see n. there.

704. terrestria membra marinis: Scylla, expressly mentioned in the parallel passage v. 893.

705. Chimaeras: v. 905 prima leo, postrema draco, media ipsa, Chimaera ore foras acrem flaret de corpore flammam.

706. omniparens: a compound not known before Lucr.; it recurs in v. 259, again used of the earth omniparens eadem rerum commune sepulcrum; similarly in Virg. Aen. vi. 595 terrae omniparentis.

707. Repeated from i. 188.

708. Cf. i. 189 paulatim crescunt, ut par est semine certo, crescentes-que genus servant: the two passages are closely allied; cf. also i. 169 at nunc seminibus quia certis quaeque creantur.

709. conservare genus crescentia: the point is important. Not only must things be originally composed of the right classes of atoms, but only atoms of those classes will be able to join the compound and so cause growth.

710. necessust: as in ii. 725, iv. 1006, and vi. 206. Lucr. appears to write necessest, necessumst, and necessust indifferently. The form necessus occurs also in Plautus and in inscriptions, and the contraction necessust is on the analogy of tempust, opust, etc.; cf. probably iv. 961 intust. Munro would change in all instances to necessumst.

pg 916

712. convenientis efficiunt motus: 'perform' or 'produce' the appropriate motions. Cf. i. 1030 ut semel in motus coniectast convenientis.

714. multaque … plagis: 'and many bodies (sc. corpora) flee in invisible bodies from the body smitten by blows'; i.e. many atoms escape from the living body in the particles of respiration, perspiration, etc. The triple repetition (implied) of corpus in different senses is awkward, but this must be the meaning.

715. percita plagis: cf. i. 1025 ex infinito vexantur percita plagis.

716. quoquam: 'to any part' of the body.

intus: cf. intus in artus 711. This is Lachmann's correction for MSS. inte; it is far better than Brieger's inde or Bernays's in se, which is indeed hard to understand.

717. vitalis motus consentire: 'to feel the vital motions in conjunction with the atoms already in the body', i.e. to join with them in the vital motions. This must be the meaning and not, as Giussani would take it, 'to feel in harmony with the vital motions', which is bad grammar and bad Epicureanism. For vitalis motus cf. ii. 941 nec congressa modo vitalis convenienti contulit inter se motus. consentire is trans. here and governs motus just as imitari does. Diels suggests vitali … motu, but imitari precludes it.

719. The MSS. have legibus his quaedam ratio disterminat omnis, which is nonsense. Corrections have been made by almost every editor. Most (except recently Martin) are agreed that quaedam must be wrong and that ea or better eadem is required; in that case Bernays's hisce will explain the origin of quaedam; the form hisce recurs in vi. 647 and hasce in iii. 35, iv. 549, vi. 211; see Prol. V a, § 8 (b). disterminat occurs in Cicero's Aratea 94 quas intervallum binas disterminat unum and there is therefore a prima-facie case for supposing that Lucr. used it. The only methods of keeping both eadem and disterminat would be to read for omnis either (1) omne (Bernays) which Lucr. elsewhere only uses in the sense of 'the universe', or (2) omnia with Lachmann (comparing Virg. Aen. vi. 33 quin protinus omnia perlegerent oculis), a synizesis which Lucr. does not elsewhere employ, or (3) keeping omnis, to mark a lacuna after the line with Brieger, which would begin with the word res, or (4) to suppose that omnis is used as the equivalent of omnis res. This last seems just possible in view of i. 190 crescentesque genus servant. Apart from that the choice lies between ea res ratio disterminat omnis (Munro) and eadem ratio res terminat omnis (Lambinus). On the whole the latter is probably to be preferred: the syllable dis might, as Diels suggests, have come from dissimilis below it in the next line, and though Cicero has disterminat, Lucr. himself twice uses terminat in i. 1000 and i. 1012.

720. tota natura: 'in their entire character'; much as natura is used in ii. 17 for 'human nature'.

pg 917

721. quamque: a certain correction for cumque OQ.

723–4. Intentionally resembling 692–4, so as to recall the doctrine there taught. After these two lines Giussani inserts 688–99, but see introductory note.

724. constant: Lachmann constent as in ii. 337 (where see n.) and 694.

725–9. This is a new point; see introductory note.

726. vias: the course of the atoms inside the complex of the compound; Epicurus' πόροι‎.

pondera: weights will vary according to the intervalla: see i. 358–69.

728. terras ac mare … caelumque: the great divisions of the world are kept apart by their differences of texture, etc., and so ultimately by the different classes of atoms which go to compose them.

729. retentant = retinent 'keep apart', 'distinguish'.

## C. The Atoms are without Secondary Qualities. 730–1022.

The 'primary' qualities of the atoms, size, shape, and weight, have been proved or implied in the argument hitherto, size in the doctrine of the minimae partes (i. 599–634, 749–52, ii. 481–99), shape as the cause of differences to perception (ii. 333–729), and weight as the cause of the universal motion downwards (ii. 184–215). These are the only properties of the atoms, their coniuncta (i. 449 ff.). The things perceptible to the senses possess also certain changeable qualities (eventa) such as colour, sound, taste, smell, heat, or cold, and in the case of animate beings, sensation. None of these belong to the atoms of which things are composed, but are the resultants of the shape, position, and motions of the atoms.

This section is thus at once a completion in a negative sense of the description of the atoms which has occupied the larger part of the first two Books and also a continuation of the immediately preceding argument, for these 'secondary qualities' of things are a conspicuous example of the effects produced by differences of atomic shape, combined with difference in position and movement. The section corresponds in general effect and in part of its argument (notably 749–52) to § 54 of the Letter to Herodotus:1 'Moreover, we must suppose that the atoms do not possess any of the qualities belonging to perceptible things, except shape, weight, and size and all that necessarily goes with shape (e.g. the possession of inseparable pg 918parts). For every quality (sc. of phenomena) changes; but the atoms do not change at all, since there must needs be something which remains solid and indissoluble at the dissolution of compounds, which can cause changes; not changes into the non-existent or from the non-existent, but changes effected by the shifting of position ⟨of some particles⟩, and by the addition or departure of others.'

It is characteristic of Lucr. that what Epicurus puts in an abstract form and in application to all 'secondary qualities', he states in reference to the individual qualities separately and largely in a concrete form with a wealth of illustration. He takes colour as his chief example and deals with it at length (730–841), for he always finds the sensation of sight the most fruitful for his purpose (cf. iv. 26–216). The other 'qualities', heat, sound, taste, and smell, are dealt with more perfunctorily (842–64), and Lucr. then treats, again at length, of sensation (865–990). Sensation is not quite on the same plane as the other qualities, for it is an active power of animate creatures and not a passive quality perceived by others. But Lucr. places it here because it is in his view, just as much as the other qualities, a product of the size, shape, position, and movement of the atoms (894–6), and it is emphasized because it is the necessary preliminary to the account of the soul given in Book iii.

It should be remembered all through that Epicurus and Lucr. did not, like Democritus, regard the 'secondary qualities' as purely subjective, but as the real qualities or 'accidents' of concilia; they have an existence independent of their perception.

## (a) Colour. 730–841.

After an introductory paragraph (730–48), Lucr. proceeds to a series of seven arguments showing that the atoms are without colour. Of these the first two (748–94) are based on the a priori conception of the nature of the atom, the first (749–56) following closely the argument of Epicurus (Ep. ad Hdt., § 54) and the second resting on the idea of the differences of sensation caused by differences in the shape and position of the atoms.1 The last five are of a more miscellaneous character and are derived from the experience of sensation and phenomena.

## 1. Introduction. 730–47.

731. You must not suppose that white things are made of white atoms, black of black, and so on, the atoms having the same colour as the things. The atoms have no colour either like or unlike that of things. 739. This does not imply that the mind can have no cognizance of colourless atoms, for blind persons can know by touch pg 919things which for them have never possessed colour, and all of us in the dark can perceive colourless objects.

730. nunc age … percipe: an emphatic formula for the purpose of introducing a new and important point.

meo dulci quaesita labore: cf. iii. 419 conquisita diu dulcique reperta labore.

731. albis ex alba: for the order of the words cf. i. 841 ignibus ex ignis, and for ex similarly postponed ii. 791 variis ex and other examples quoted on i. 841 n.; see Prol. V b, § 15 (b).

732. ante oculos quae candida cernis: 'which you see shining white before your eyes'; Lucr. intentionally emphasizes the perception of colour in compound things in opposition to the mental conception of atoms without colour.

733. nigrant: the main verb is elsewhere unknown in the neutral sense 'to be black', though the participle nigrans is frequent in Virgil and elsewhere. Cf. the intransitive use of variare in ii. 480. Accius and Pacuvius use nigreo.

734. nive = neve 'nor' is the reading of O and should be retained, cf. iii. 286 ni for ne; the form is quoted from early Latin writers, such as Lucilius and Pacuvius, and is not infrequently found in inscriptions: see notes in Lachmann and Munro.

que sunt imbuta colore OQ: the obvious correction of que is quae; indeed the mistake is so common in the MSS. that it hardly amounts to a correction. This involves the acceptance of colorem from L: 'and do not suppose that the things which are steeped in any other colour possess this because …'. imbuta is nowhere else constructed with an internal acc. of a substantive, but with a pronoun it occurs in Tac. Hist. v. 5 nec quicquam prius imbuuntur quam contemnere deos, and Munro quotes in support Virg. Georg. iii. 307 vellera … Tyrios incocta colores. The alternative is to read quo … colore with Lachmann: 'and do not suppose that they possess any other colour with which they are steeped because …', but there is then no subject for sunt imbuta and gerere, which badly need it, and hunc becomes superfluous; Winckelmann (Philol. xxv. 68) went farther and read alio quovis … colore, but that does not help the construction. To avoid the difficulty of imbuta with acc. Lambinus read induta, but the sense is wrong: things do not 'don' a colour which they did not before possess. But it has some support from the parallel velata in ii. 797.

735. gerere = habere.

hunc: sc. colorem.

736. sint: subj. in virtual or. obl.; it is a false reason.

eius with consimili 'like to it', sc. the colour of the compound things.

738. par rebus: a 'compendious' comparison, 'like to the colour of things'.

neque denique dispar: 'nor indeed unlike it', anticipates the discussion (in ii. 776 ff.) of the possibility that things might be composed of atoms of many different colours.

pg 920

739. in quae corpora: sc. 'bodies of this sort', i.e. without colour.

740. animi iniectus: 'a projection of the mind', so 'an act of comprehension', no doubt a translation Epicurus' very obscure expression ἐπιβολὴ τῆς διανοίας‎. The mind, in Epicurus' view, being an aggregate of soul-atoms, its thoughts are caused, just as are the sensations of the senses, by films coming from objects outside; it can also be moved by the προλήψεις‎ 'general concepts', which it has stored up within itself (see n. on notitiam ii. 745). But some films, e.g. those coming from the gods, are so subtle that they cannot be perceived even by the mind, except by an act of attention (ἐπιβολή‎); other things like the atoms and the void cannot from their nature send off films and can only be grasped by the 'projection of the mind' on to an image which it creates for itself (usually by the combination of προλήψεις‎). The objection, then, here will be that the mind could not 'visualize' by such an act of projection atoms without colour. To this Lucr. retorts with the obvious examples of the blind, who have never known colour, and of persons in the dark, who are enabled to perceive or visualize things by the sense of touch; see Prol. IV, § 6. For fuller discussions of ἐπιβολὴ τῆς διανοίας‎ see Giussani, Stud. Lucr., pp. 176–82, my Greek Atomists and Epicurus, pp. 259–74 and Appendix III, 559–76, and Robin ad loc. (who does not seem to me to distinguish clearly enough between ἐπιβολὴ τῆς διανοίας‎ and ἐπιβολὴ τῶν αἰσθητηρίων‎. For the recurrence of the idea in Lucr. see ii. 1047 animi iactus liber and, if the text be right, ii. 1080 animalibus inice mentem and the notes at those places. Cf. too Cic. De Nat. Deor. i. 20. 54 immensam … magnitudinem regionum, in quam se iniciens animus et intendens late long eque peregrinatur.

741. caecigeni: 'those born blind', a ἅπ. λεγ‎. and a fine Lucretian compound.

741–2. The corrections lumina and dispexere made in the Itali for numina and despexere are necessary: dispicio is the regular word for seeing things in the dark or through a mist.

743. 'From the beginning of their existence which are not associated with colour.' Strictly speaking, the objects which the blind perceive by feeling have colour, but not for them; ex ineunte aevo goes in sense with caecigeni, sc. 'which the blind have from their birth known as colourless'. Bentley followed by Lachmann transposed this line after 748 to fill up the lacuna. But ex ineunte aevo is impossible there, for the atoms are eternal and never had a 'beginning of existence'.

745. verti in notitiam: 'enter our comprehension'; cf. vi. 1171 vertere in utilitatem. For notities in Lucr. see ii. 124 n. and Prol. IV, § 4. Here it is used in the functional sense of 'comprehension', 'the grasping of a concept'.

circumlita: 'painted with': note the pg 921variants coniuncta (743), circumlita, tincta (747) 'for the same idea. fuco = colore as in ii. 683.

746. nos ipsi: 'we ourselves', all of us who are not blind.

747. sentimus: 'feel', almost 'recognize'.

## Proofs. 1. Atoms which changed their colour would not be permanent. 748–56.

748. I will now prove ⟨that the atoms are without colour⟩. The colour of things is always changing, but the atoms cannot change; there must be an unchangeable foundation, lest all things should be destroyed. 752. For change is the death of what existed before. We must not assign colour to the atoms, lest that lead to universal destruction.

Lucr. is here closely following, almost translating, the sentence in Ep. ad Hdt., § 54, quoted in the introductory note (p. 917 n. 1). ποιότης γὰρ‎ … ἀδιάλυτον‎. At the outset of the poem (i. 215–64) he argued that things are not utterly destroyed but only resolved into their component seeds, and these seeds were subsequently shown to be the eternal and unchangeable atoms. But colour is constantly changing and that which has colour is liable to change; colour, then, must not be attributed to the atoms, or they will be lost as the permanent foundations of things. The argument is abstract and is based on the fundamental principles of Atomism.

748. The new paragraph should begin here, as Lucr. passes from the introductory considerations to the proofs. Martin keeps the text as it stands without any lacuna, supposing a contrast between fieri and esse: 'since I have shown that this happens (sc. the existence of apparently colourless things for blind men and persons in the dark) I will now show that it is the case (with the actually colourless atoms)'. But this is impossibly compressed and allusive. Bentley and Lachmann transferred 743 here, but the subject (primordia or its equivalent) must be expressed and ex ineunte aevo is impossible in this context (see n. on 743). Editors since Munro have assumed a lacuna. Munro himself supplies corpora, quae nullo constant coniuncta colore, but that is not quite the sense required: Lucr. is not going to prove that there are colourless bodies, but that the atoms are colourless; the lost line was probably more like corpora prima omni semper privata colore. Diels's supplement nullum in principiis exsistere posse colorem is nonsense: he has forgotten the existence of esse in 748. For vinco, 'I prove', cf. vi. 498 vincam: here it is used in the idiomatic sense 'I have proved'; cf. ii. 112 n. uti memoro, 870 quod dico, iv. 1207 ut dico. See Prol. V b, §9 (a).

749. omnis enim color omnino mutatur et omnis: most editors since Lachmann have adopted the reading of l 31 in omnis 'for every pg 922colour whatever changes into all colours'. But (a) this is not true, (b) mutatur in omnis is scarcely Latin, (c) quod facere in 750 is then very awkward, for quod should grammatically mean mutari in omnis colores and not simply mutari, (d) there should be an intermediate step in the argument: 'things which change their colour, change themselves'. Brieger, followed by Giussani, retains et omnis and supplies res, dum alium capiunt, mutari ipsae quoque debent. This gives the right sense, but, apart from the awkwardness of dum alium capiunt sc. colorem, omnis as nom. plur. is improbable (see Prol. V a, §§ 1. i and 6). I suggest that the missing line was something like res sese mutat, mutat quaecumque colorem.

750. quod facere: with the supplement suggested after 749 this means, as it should mean, mutari or sese mutare. This is strongly supported by i. 789 where quod facere = inter se mutare.

750–4 are repeated from i. 789–93, where they are used against the supposition of the elements passing into one another and so changing.

753–4. Lucr.'s axiom, used already in i. 670–1 and i. 792–3 and again in iii. 519–20. Its use here, where there can be no question of Heraclitus' theory of the flux, is against Pascal's theory that it represents a principle of Heraclitus; see i. 690 n.

755. contingas: lit. 'dye', 'steep', so 'contaminate'; the colour metaphor is continued. Cf. i. 934 musaeo contingens cuncta lepore and vi. 1188 croci contacta colore.

756. Repeated from i. 673; again in ii. 864.

## 2. Change of colour is better explained by colourless rather than coloured atoms. 757–94.

757. If atoms are colourless but varied in shape, so that they can produce varieties of colour by changes in their order, position, and motions, 763 it is easy to explain how black things might become white, as the sea does when churned by wind. 768. For a black thing can become white when its matter is disturbed and the atoms changed in order or increased or decreased in number.

772. But (a) if the sea were made of blue atoms only, it could never become white; for however much you rearrange blue things, they can never become white. 776. And (b) if the atoms which make the uniform colour of the sea are of all sorts of different colours, in the same way as a square figure may be composed of shapes of all sorts, then 780 (α‎) you ought to see the different colours in the one colour of the whole, as you see the different shapes in the square, 784 (β‎) the different shapes in the square do not affect the squareness of its outline, but a mixture of different colours would prevent the uniformity of the colour of the whole, 788 (γ‎) this supposition would destroy the idea from which we started, that white things are made pg 923of white atoms, etc. Indeed it is much easier to suppose that white things are composed of atoms without colour than of black or any other coloured atoms.

The argument of this section is complicated and more polemical than it appears at first sight. Having in the first argument given his a priori proof that the atoms must be colourless, if they are to be permanent, Lucr. goes on to apply this idea to the phenomena of the change of colour in things. The main argument of the paragraph may be summed up in the statement which Plutarch1 attributes to Epicurus: 'colour is not inherent in the nature of the atoms but is produced by definite arrangements and positions in relation to sight'. Lucr. starts (757–71) with the enunciation of this view. Colour was on Epicurus' theory a quality in things which was produced by the arrangement and position of the (colourless) atoms on the surface of things, and was conveyed to the sense of the percipient by the film which coming off the surface, yet preserved the atomic conformation of the original object. A change of colour occurred when an external blow (usually the incidence of light) or some other cause brought about a change in the position or order of the component atoms, as when the wind striking the waves caused a rearrangement which changes their colour from blue to white.

Lucr. then considers the alternative view, that the atoms themselves have colour. There are here clearly two possibilities, either that the atoms in any given thing are all of one colour (cf. 731–3), or that they are of many different colours. In the first case change of colour is impossible: no rearrangement of atoms could make a thing composed wholly of black atoms become white. The second possibility is more complicated in thought and was probably the theory of Anaxagoras.2 All things contain within themselves atoms of all colours, but the colour of the whole is that of the atoms 'of which it has most in it'. When these atoms are by a blow scattered from the surface and others are gathered together, the object then changes colour. Lucr.'s objections are three in number and are all concerned not with the phenomenon of changing colour, but with the impossibility of accepting the presupposition of a uniformly coloured thing having in reality a surface of many colours. His first two arguments are based on a comparison with a square figure composed of figures of many different shapes (a 'jig-saw puzzle' is the easiest illustration). In the first place (780–4), just as we can see pg 924the component shapes inside the square figure, so we ought to be able to see the patches of different colour within the uniform colour of the whole (as one might on a close view of a 'pointilliste' picture). We may compare the argument against Anaxagoras' general theory in i. 875–96. Secondly (784–7), though the internal figures do not in any way destroy the external outline of the square, because it is only a question of its boundaries, a variety of colours would destroy the uniformity of colour over the whole superficies. Thirdly (788–91), the composition of a single colour out of a variety of colours is a contradiction of the assumption from which the argument started (731–3) that things are composed of atoms of the same colour with themselves. There is considerable doubt whether 788–94 should constitute a separate paragraph, but Giussani is probably right in attaching them to the preceding lines as a third reason against the multi-colour view, not now as an argument in substance, but as a logical objection. See notes ad loc.

757. coloris … natura: merely a Lucretian periphrasis for color.

758. et: adversative 'and yet'; cf. i. 658 n.

variis … formis: is important. If the atoms were all of one shape, then their transposition could make no difference to the general effect on sensation: it is their variation that makes different colours and leads by transposition to a change of colour.

759. omne genus: most editors follow Lachmann in writing this as two words forming an internal accusative; see n. on i. 1026 and Prol. V b, § 3. Here OQ have omnigenus as one word, omnigenos O1 and It.

760–2. Repeated except for the first three words from i. 817–19. Seeing that Lucr. is there arguing against Anaxagoras, the occurrence of the lines here suggests that he has again Anaxagoras in mind. The substance of the lines recurs in ii. 894–6, where once more Lucr. is probably thinking of Anaxagoras.

760. propterea … quod: 'for this reason … that', a certain correction for praeterea OQ.

semina quaeque: not 'each of the individual seeds' but 'each of the different shapes of seeds'. Cf. i. 578 quaeque … corpora 'bodies of each kind' and n. there.

762. The motion of the atoms is not an immediate cause of colour, as it is of sensation (ii. 941), but because it causes the transposition of atoms it is the indirect cause of the change of colour.

763. extemplo: 'straight off', 'at once'; cf. facile extemplo iv. 742, vi. 898, and ii. 892–3 non … extemplo me gigni dicere sensus. extemplo is a certain correction of exemplo OQ.

possis: the regular idiomatic 2nd pers. pres. subj. (see i. 327 n. and Prol. V b, § 10 (a)), but here it influences fuerint and possint in the subordinate clauses.

765. marmoreo … candore: 'of marble whiteness', a poetical use of marmoreus repeated in 775. Cf. Arnobius Nat. 4. 22 ulnarum nivei pg 925marmoreique candors. candenti marmore 767 is a bold transposition of substantive and adjective.

766–7. candenti marmore: see previous n. marmor is usually found of the smooth surface of the sea: here of the marbled appearance of the breaking wave. We may compare Ruskin's youthful description of a breaking wave quoted in Frondes Agrestes, Section iv, § 31 n. 'One moment a flint cave, the next a marble pillar, the next a fading cloud.'

768. saepe: 'occasionally', almost equivalent to 'for instance'.

769. permixta: 'mingled up', 'disturbed'.

770. principiis: a 'Lucretian' dat. for gen. 'the order of the first-beginnings'; see Prol. V b, § 5. 1.

addita demptaque quaedam: note this; the change of colour may be produced or assisted by the coming or departure of atoms from the surface as well as by change of order and position. In the case of the waves the wind might remove or add particles.

772–5. See introductory note; this is the first possibility with coloured atoms; they might all be of the colour of the whole. Then, Lucr. argues, there could be no change of colour.

774. perturbes: the indefinite 2nd pers. subj. again, as in 763 possis; see Prol. V b, § 10 (a).

775. migrare: 'to pass into'; cf. iii. 757 traiciuntur enim partes atque ordine migrant, v. 830 omnia migrant, and even closer Ovid, Met. xv. 172 in varias … migrare figuras.

776 ff. The second possibility that the atoms are of several colours: see introductory note.

776. alio atque alio: 'one after another' emphasizes the succession of the different colours in order.

777. nitorem: 'brilliance', here more than a mere synonym for colorem.

779. quiddam quadratum unaque figura: 'something of a single shape, for instance, a square'. figuras of OQ is a mere mistake, and Marullus's correction figura est breaks the sequence of thought.

780. conveniebat: 'it would be appropriate', 'it would follow'; the idiomatic imperfect indic.; cf. i. 881 n., and see Prol. V b, § 9 (c). ut for uti OQ: corrected by Lachmann; see ii. 322 velut for veluti and n. there.

782. nitore = colore, with no special sense as in 777.

783. colores L for calores, a mistake of OQ.

784. praeterea: not a new argument in the main series, but the second argument against the supposition of multicolour atoms in 776.

785. extra: 'on the outside', 'in its outline': Lachmann's ex his is gratuitous and misses the point.

787. nitore here = colore, as in 782.

pg 926

788–94. These lines are usually printed as a separate paragraph, as though the argument was one in the main series of proofs that the atoms are colourless. But it is clear that they deal still with the idea of multicoloured atoms introduced in 776 and constitute the third argument against it: the idea is a contradiction of the reason which led Anaxagoras and others to suppose that atoms had colour, namely, that if one divides up any object into small pieces, they are all of the same colour as the whole (731–3). But in order to account for change of colour it was necessary to assume that things were made of atoms of many different colours (776 ff.); and these two notions are inconsistent. The difficulty is in fact exactly that which besets the main theory of Anaxagoras and with which Lucr. dealt in i. 875–96. Housman (Journ. Phil. xxv. 237) argues on the same lines. Brieger supposes a lacuna after 787 in which the causa of 789 was explained. But it is easily assumed from 731–3, the thought of which has been 'suspended' in Lucr.'s mind ever since and which indeed is quite sufficiently reproduced in 790–1. In 792–4 Lucr. comes back to a restatement of his own view, as he so often does at the end of an argument.

788–91. 'Then, further, the reason which leads us on and entices us sometimes to assign colours to the first-beginnings of things (namely that things are apparently constituted of particles of like colour to themselves), falls to the ground, since white things (on this hypothesis) are not made of white, nor those which are seen black of black, but of diverse colours.' The argument is not one of substance, but purely logical. Anaxagoras and his followers are inconsistent with themselves, as indeed they were in their main theory as Lucr. understood it (see i. 881 and notes).

788. inlicit ut: a certain emendation for inlicitu Q, which was fantastically 'corrected' by the Italian scribes and critics. inlicit contains the idea of deception, which is not contained in ducit; cf. iv. 1145 cavereque ne inliciaris.

789. nonnumquam: sc. we are inclined to be led astray by this reason; cf. 768 saepe.

790. occidit: 'perishes', 'falls to the ground'.

quoniam … creantur: 'since on their assumption'. Cf. ii. 817 quoniam non certis certa figuris est natura coloris.

non goes with creantur, and not, as some editors have tried to take it, with alba, creantur: OQ's creatur is a mere mistake.

791. nec quae …: 'and things that appear black are not created of black particles, but of multicoloured', nec = et non and the negative goes with creantur understood.

variis ex: the emphatic position at the end of the line gives the words weight, as the important point in the sentence. For the place of the preposition cf. i. 841 n., and see Prol. V b, § 15 (b).

pg 927

792. proclivius: 'more easily'. Cf. ii. 455 et perculsus item proclive volubilis exstat (in a more literal sense) and n. there.

794. qui contra pugnet et obstet: generic 'any other colour such as is opposed to white'.

## 3. Colour implies light which cannot affect the atoms. 795–809.

795. Since colour cannot exist without light and the atoms never appear in the light, they cannot have colour. 799. Moreover, colour changes according as light strikes straight or slantwise. The feathers on a dove's neck seem sometimes red, sometimes blue mixed with green; 808 the peacock's tail changes when it meets the light full. Since these colours are produced by the striking of light, they could not exist without it.

The argument here attaches itself to the previous argument (741–7) about the recognition of objects by blind men and persons in the dark; they can feel the shape of an object, but not see its colour. But it is itself of primary importance, since it establishes the statement that light is the necessary condition of the existence of colour. So Epicurus as quoted by Plutarch1 says: 'indeed without this part (sc. light) I do not know how it is possible to say that objects in the dark have colour', and Philodemus2 similarly remarks: 'things in the dark have not colour, but they are bodies'.

Lucr.'s theory appears to be that it is the beating of the rays of light on an object which causes the atoms on the surface to arrange themselves so as to show their colour, or else that it reveals the colour which is potentially latent in their superficial colour. On this point his statement is obscure, but it is clear that he held that an alteration in the direction of incidence of the light caused a rearrangement of the component atoms, which produced a change of colour (808). The original statement that there can be no colour without light is thus reinforced by the variation of colour with the variation of the light-blows.

796. neque in lucem exsistunt: 'do not issue into the light'. But what exactly is Lucr.'s meaning? It is most natural to take it to mean that the atoms owing to their smallness are invisible and therefore a fortiori cannot have colour. Giussani objects that, even though we cannot see them, the atoms are nevertheless subject to the light and the light does play upon them, but because it is impossible to alter the arrangement of the parts in the atom, the action of the light cannot produce colour in the atoms as it does in concilia, however small: he would therefore take the meaning to be 'are not pg 928subject to the action of light'. Ernout thinks that both meanings are implied; 'the atoms remain always invisible and escape the action of the light'. I am inclined to think that the distinction is too subtle for Lucr. and that all he meant was that the atoms are ἄδηλα‎. The phrase, if pressed, approximates more to the former meaning, that the atoms are invisible, for Giussani's interpretation, though it may be good Epicurean doctrine, can hardly be got out of the words as they stand.

797. velata: 'clothed with', a very picturesque metaphor, which is some support for induta in 734.

798. Sc. the atoms, being ἄδηλα‎, are in perpetual darkness. This idea seems to tell against Giussani's interpretation of 796, though he argues to the contrary.

799. quin ipso: not merely equivalent, as Munro holds, to quin etiam: the same light which causes colour also causes change of colour.

propterea quod: 'according as'.

800. refulget: Lachmann writes refulgit, believing that Lucr. used only the forms from the 3rd conj. But seeing his frequent variation between conjugations it is better to follow the MSS. (see ii. 27 n.).

801–7. The changing of colour 'on the burnished dove', etc., seems to have been a commonplace problem of the schools, like the bent oar, etc., which Lucr. discusses in iv. 436 ff. Cf. Cic. Acad. Pr. ii. 7. 19 nec vero hoc loco exspectandum est dum de remo inflexo aut de collo columbae respondeam. So again in 25, 79 tu autem te negas infracto remo neque columbae collo commoveri, and Sen. N.Q. i. 7. 2.

802. cervices OQ; frequently used in the plur. Munro states that it is the back of the neck and collum, the whole round. Brieger followed by Giussani corrects to cervicemst on the ground that est is needed and that Lucr. uses the singular in i. 35 and vi. 744. But Lucr. frequently omits est in subordinate clauses (e.g. i. 111) and general usage elsewhere is sufficient authority for cervices.

coronat: 'makes a ring round'. For the postponed circum cf. i. 937 n., and see Prol. V b, §15 (b).

803. rubra: a certain correction of rubro O robro Q.

pyropo: Plin. N.H. xxxiv. 94 states that pyropus was a compound of one part of gold and four parts of bronze, but it appears also to have been applied to a precious stone, probably garnet, and in view of zmaragdos (and caeruleum) that is more probably Lucr.'s meaning here. Cf. Ovid, Met. ii. 2 flammasque imitante pyropo, which is, however, not decisive in either sense.

804. quodam sensu: 'under certain conditions of perception', 'in a certain aspect'; cf. iv. 448 quodam sensu fit uti videantur omnia quae tuimur fieri tum bina tuendo. Brieger objects that the change is made not by a different mode of perception but by a different blow pg 929of light (cf. quodam … ictu 808); but the abl. is not instrumental, but one of circumstance.

805. caeruleum OQ: 'blue' or perhaps more definitely 'lapis lazuli': with the red of the pyropus and the green of the zmaragdi, a blue colour is required here. Wakefield, comparing an undoubted imitation of Lucr. in Serenus Sammonicus 952 coralium vero si collo nectere malis nec dubites illis veros miscere zmaragdos, emended to curalium 'red coral', which was adopted by Munro, etc. But the repetition of 'red' weakens the passage, and recent editors, Merrill, Ernout, Diels, and Martin, have all rightly returned to the MSS. caeruleum.

miscerĕ zmaragdos: for the short e before the double consonant cf. i. 372 n., and see Prol. VI, § 11. With this word the licence was generally permitted in Latin and is repeated by Lucr. in iv. 1126 viridi cum luce zmaragdi. The spelling with initial zm has the better authority; so Zmyrna in the MSS. of Livy.

806. larga: largo OQ. lux was undoubtedly used masc. in earlier Latin, e.g. luci claro Plaut. Aul. 748, cum primo luce Ter. Ad. 841, but there is no authority for it elsewhere in the MSS. of Lucr., and the correction of Q1 should be accepted.

807. obversa: 'as it turns round to (or "in") the light'. Cf. Virg. Aen. iii. 549 cornua velatarum obvertimus antennarum, Ovid, Met. iii. 676 obstantis dum vult obvertere ramos, and more closely in Lucr. himself iv. 166 speculum quocumque obvertimus.

808. This line points clearly to Giussani's theory that the colour is created (gignuntur) by the blow of light which brings about a rearrangement of the atoms.

809. posse: OQ's posset is again a mere mistake.

## 4. The sensation of colour is effected by contact, which implies form but not colour. 810–16.

810. The sensation of different colours is produced by the blows (of different films) on the eye. Touch or contact depends on form not on colour. 815. Colour is not, then, necessary for the atoms, which vary their blows by difference in shape and arrangement.

Giussani, whom recent editors have followed, is right in making this a new paragraph: it is in reality distinct from the previous argument. It is also rather obscure in expression, but may be explained thus. All sensation is ultimately referable to touch: sight is caused (cf. iv. 26 ff.) by simulacra coming off the surface of objects and falling into the eye, difference of colour by differences in the shape and arrangement of atoms in the simulacrum. But for contact all that matters is shape; colour is indifferent, as we know when we touch things in the dark. In order, then, to produce the sensations of different colours, all that the atoms which compose pg 930the film need have is shape. From this point of view it does not matter whether they have colour or not. The argument is ingenious, and though it does not prove that the atoms have not colour, it shows that they need not in order to produce the sensation of colour.

810. This idea of the sensation of colour being due to a blow on the pupil of the eye has been generally recognized as a remarkable anticipation by Lucr. of modern ideas. But it must be remembered that it is closely connected with the notion of the simulacrum, which is too crude to be now acceptable.

811. sentire colorem: 'to perceive colour'; cf. i. 298 sentimus odores, iv. 617 sucum sentimus.

812. et cetera: sc. ceteros colores, a vague neuter; cf. iii. 390 culices et cetera.

814. sint: a correction in the Aldine for sunt OQ; the indirect question requires the subj.

apta: 'connected with', 'associated with'. Cf. v. 537 coniunctam atque uniter aptam.

815. colores should be accepted for colore OQ. Lucr. uses both nom. (acc. in or. obl.) and abl. with opus est (see Prol. V b, § 2 (a)), but Nonius quotes the verse with colores, and the plural is required by the plurals in the next line.

816. formis: not only the shapes of the individual atoms, but their arrangement in concilia; corresponding both to figuris 817 and formamenta 819.

edere: 'produce', 'cause'; cf. ii. 443 varios quae possint edere sensus.

## 5. With multicolour atoms, the colours of species should vary. 817–25.

817. Since definite colours are not attached to definite atomic shapes, and all kinds of atomic configurations may be found in things of every colour, why are there not examples of every colour in each species of creature? 822. There should be white crows or black swans or swans of any other colour.

A very condensed and consequently obscure paragraph. Lucr. returns more definitely to the Anaxagorean theory that the atoms have colour. In the preceding section (810–16) he has shown that colour is caused by, but not associated with, shape. It is not assumed by the supporters of the rival theory, nor could it be granted by Lucr. himself, that atoms of a certain shape (figura) had a certain colour, e.g. that triangular atoms were always blue. For instance, both snow and marble are white, but the atoms composing the former are smooth and round, those of the latter jagged and rough. Nor again did the Anaxagoreans hold that certain configurations of particles (formamenta) were peculiar to any particular colour, and though this would come nearer to Lucr.'s pg 931view, he would hold that there were many different atomic arrangements which could produce blue—variis formis variantis edere tactus. If that were so, and the atoms were themselves coloured, then there is no reason why in any given species of thing there should not be every variety of colour, white crows and black swans. The atoms and the atomic configurations which make a swan might be of any colour. It is a dialectical point, and Lucr. does not suggest his own answer which would probably be that the arrangements of atoms necessary to create a given species can when struck by light only produce a limited range of colours.

The argument goes back to the thought of 777–94, which has since been 'suspended' in Lucr.'s mind, and Brieger for this reason places these lines after 794. But, as has been seen above, they are also closely connected with 810–16.

817–19. quoniam … nitore 'since on their view …' like ii. 790 ex albis quoniam non alba creantur, though here Lucr. would agree with the view: see introductory note.

817. figuris: the shapes of the individual atoms.

819. formamenta: not, as Giussani apparently takes it, a synonym of figurae, but 'the conformation of a group of atoms' in a concilium = σχηματισμός‎, as Robin suggests. The word is used only by Lucr. and his imitator Arnobius. It is a typical Lucretian formation, cf. iv. 556 formaturam, and see Prol. VII, § 4.

820. ex illis: probably ex formamentis, not ex figuris.

821. omne genus: see i. 1026 n.

genere: 'species' of created things: his examples are from living things, but the argument would apply to inanimate objects as well.

822. conveniebat: cf. i. 881 n., and see Prol. V b, § 9 (c).

volantis as Giussani points out, is not an otiose epithet, but helps out the poetic idea of iactare.

823. iactare: 'throw off'. Cf. ii. 846 nec iaciunt ullum proprium de corpore odorem and iv. 95 membrana coloris cum iacitur.

824. Cf. Juv. vi. 165 rara avis in terris nigroque simillima cycno. The discovery of black swans in Australia is unfortunate for Lucr.'s example.

825. 'Or of any other colour, single or varied', e.g. brown, or brown with yellow stripes or spots. -que is then disjunctive, as in ii. 830 purpura poeniceusque color = sive purpureus seu poeniceus. Munro would take varioque to be merely expletive of alio quovis, 'or of any other different colour'; cf. ii. 783 dissimilis longe inter se variosque colores. This is less probable.

## 6. The smaller the pieces, the less colour there is. 826–33.

826. The smaller the pieces into which a thing is separated, the more is its colour seen to disappear. A piece of red cloth torn up pg 932thread by thread loses all its purple or scarlet colour. 832. It is clear, then, that particles would lose all colour before the atoms were reached.

In the two last paragraphs Lucr. turns from the technical and argumentative form of proof to direct experience. We notice when we pluck asunder a piece of brightly coloured stuff that, when it is in quite small pieces, the colour loses its brightness and at last seems almost to be lost. But the smallest piece visible to the eye is still many times greater than the atoms (cf. ii. 312 omnis enim longe nostris ab sensibus infra primorum natura iacet); we may therefore infer that the tiny pieces will lose all colour before they reach the stopping-point of the atoms. Much more then are the atoms themselves devoid of colour.

826. quin etiam: this emphatic form of introduction may be intended to call attention to a new argument on quite different lines.

827. magis, hoc magis: O has this correctly, but as the first two words are written in an erasure, it probably had beneath the same perversion of the order as appears in Q.

828. evanescere paulatim: so in a different context v. 535 evanescere paulatim et decrescere pondus.

stingui: cf. i. 485 n. nulla potest vis stinguere. Here it probably has the idea of the colour losing its brightness.

829. austrum: the brilliant conjecture of Wakefield has been almost universally adopted since. The MS. reading aurum is impossible; gold cannot be 'plucked apart' (discerpitur) nor does aurum harmonize with the next lines. Conjectures like Lachmann's aurea (with purpura), Bernays's usu, and Winckelmann's ut iam (with cumst distractus in 831), are feeble expedients to get over the difficulty. The word is usually spelt ostrum (as Wakefield wrote it); cf. ii. 35 ostroque rubenti, but the vowel change is the same as in Claudius and Clodius, Paullus and Polla, caurus and corus, plaudo and (ex)plodo, etc. ostrum is the general term for red cloth and purpura (dark red) and poeniceus (scarlet) are varieties of it.

830. 'Purple cloth or scarlet, far the brightest of the colours', sc. probably within the range red, though it well may be 'of all colours'. The combination of the concrete substantive purpura with the more abstract poeniceus color is strange, but is accounted for by the absence of a word meaning 'scarlet cloth'. Giussani would take the phrase as a hendiadys, 'scarlet, the brightest of all purple cloths', but this is unnatural. Erncut notes that the form poeniceus, as against puniceus v. 941 is technical, 'the Phoenician colour'. The -que with poeniceus is disjunctive: 'purple or scarlet': see n. on ii. 825.

831. filatim: ἅπ. λεγ‎. invented by Lucr. with his favourite adverbial termination: cf. i. 20 generatim and n. there, and see Prol. VII, § 6.

pg 933distractum est: agreeing with austrum in 829. 'When the red cloth is torn asunder thread by thread, all the colour is lost.'

disperditur: Lachmann denies the existence of the passive of perdere and reads dispergitur. But perditur is found in Prosper Epigr. verbis Perderis ipse tuis, in the Vulgate of Proverbs ii. 22 impii de terra perdentur, and it may be the right reading in Hor. Sat. ii. 6. 59. Further, dispergitur gives the wrong sense: the colour is not 'scattered' but 'lost'.

832–3. Note the carefulness of this expression: there are many grades of size between the minimum visibile and the minimum of existence, the atom (cf. i. 599 ff.). Colour will disappear before you get down to the minimum of existence.

832. efflare: a picturesque word. Cf. v. 652 suos efflavit languidus ignis.

## 7. Things may be without colour just as much as without sound or smell. 834–41.

834. There are things which do not emit sound or smell; they are not therefore assigned to them. Thus since we cannot see all things, there are some without colour. 840. The mind can recognize them as much as it can things which lack the other qualities.

This last argument is again an appeal to experience and it is quite straightforward. If we admit the existence of things without sound or smell, why not admit the existence of others (the atoms) without colour? Note that 837 makes visibility the test of possession of colour; things invisible cannot have colour. Giussani again (see n. on 796) rejects this simple interpretation and states the argument thus: 'since the absence of any perception of smell is taken as evidence of the existence of bodies without smell, so the invisibility of some bodies should be proof of the existence of bodies without colour, which is the first condition of visibility'. But again this interpretation is too subtle: Lucr. would have said roundly that ἅδηλα‎ cannot have colour.

835. mittere: the simple verb for the more expressive emittere; cf. v. 1028 sonitus natura subegit mittere and several instances in the account of simulacra in the early part of iv.

836. attribuas: 'assign'. Cf. iii. 241 quarta quoque his igitur quaedam natura necessest attribuatur, where perhaps it has more the sense of 'adding', as in i. 681 atque alia attribui and ii. 925 quid opus fuit attribui id quod detrahitur?

838. orba: simply 'without'; cf. v. 840 orba pedum partim. Note the succession of synonyms for 'without' in these lines, sine and remota (839), privata (841), spoliata (842), secreta (843), sterila and ieiuna (845): a good instance of the wealth of Lucr.'s resources.

pg 934

839. remota: cf. v. 125 vitali motu sensuque remotum.

840. nec for ne OQ from Festus.

haec: sc. the quaedam of 838, not of 839: the atoms.

841. notare Lachmann rightly for notaque OQ.

## (b) The atoms are without other secondary qualities. 842–64.

842. The atoms are not only without colour, but also without heat or cold, sound, taste, and smell. 847. Just as in making scents or unguents you look for a scentless oil to be the foundation, which will not contaminate the odours, 854 so the atoms cannot communicate any secondary qualities of their own to things, since they cannot throw off any effluences. 859. All such things are perishable and must be kept apart from the atoms, if they are to be the everlasting foundation of things and save them from universal destruction.

Lucr. has argued at length that the atoms are devoid of colour; in this paragraph he summarily extends his conclusion to the other secondary qualities. It must be noted that they are treated as res, concrete things which can as it were be added to the matter which is the foundation of compound bodies; like the drops of perfume which are added to the foundation of scentless oil in the preparation of an unguent. This is the easier for Lucr., as he believes each of the secondary qualities to be caused by a concrete emanation, colour by the simulacra, sound and scent by emitted particles (iv. 524–614, 673–705), taste by the juices which issue from food (iv. 615–72). For this reason he can group with them heat and cold, which are similarly for him concrete bodies which penetrate into things; see i. 491 n. and i. 534 neque item manabile frigus nec penetralem ignem.

The comparison with the unguent goes in reality a little deeper than appears; Plato1 uses it in a similar context. Just as in it the materia prima, as Giussani calls it, is itself an inodorous oil, to which the perfumes are added, so in compound things, the atoms, which are the materia prima, are without secondary qualities, but these are acquired by the atoms united in concilia. The reason is (856) that compound bodies can emit the films, effluences, etc., which convey the colour, etc., but the atoms being solid and indestructible cannot. All these perishable things then, Lucr. adds in a very difficult passage (859–64, see notes ad loc.)—both the qualities and the effluences which cause them—must be kept far apart from the atoms, if they are to remain as the imperishable foundation of things, which preserves the universe.

There is a certain confusion right through the paragraph between pg 935the qualities and the effluences which convey them to the senses, which was natural to Lucr., and will explain the somewhat strange diction of the argument.

843. manere is not simply equivalent to esse (Munro), but means rather 'are permanently'.

secreta teporis: a gen. of separation: cf. i. 1041 aversa viai and n. there, also Prol. V b, § 4. 1. In i. 194 Lucr. constructs secreta with abl.; the cases are interchangeable in his mind.

teporis … frigoris … calidi vaporis represent three degrees of heat: cold, temperate heat, and great heat; cf. ii. 517 omnis enim calor ac frigus mediique tepores, and 858 calidum tepidumque vaporem. Cf. vi. 1163–5 nec nimio cuiquam posses ardore tueri corporis in summo summam fervescere partem, sed potius tepidum manibus proponere tactum. So tepet vi. 862.

845. sterila: from sterilus, sterilis being the usual form in classical Latin, as in iv. 1235 sterili, 1240, 1251 steriles; a Lucretian variation in declension. Cf. i. 340 sublima and n. there; also Prol. V a, § 2. The form sterilus is recorded by Festus (Lindsay, p. 419, M. 463. 1). It is here constructed with abl., more usually with gen., e.g. Tac. Hist. i. 3 virtutum sterile saeculum. Similarly ieiuna with abl. is not quoted elsewhere: it takes gen. in Cic. Or. 30. 106 ieiunas huius multiplicis … orationis auris civitatis accepimus. The unusual word ieiuna has been corrupted in the MSS.

suco: 'taste', identified with the juices that cause it; cf. iii. 223, iv. 615, 617, etc.

feruntur = sunt Munro, who translates it, as most editors do, 'are said to be'. But it would be an odd synonym here, for it is not a matter of general opinion that the atoms are without sound or taste. It more likely means 'are carried on', referring to the atomic motion; cf. ii. 160, 217, etc. So Giussani holds that it expresses their 'peculiar mode of existence'.

846. iaciunt … odorem illustrates the confusion between the quality and the effluence which causes it (see introductory note); strictly speaking it is the particles which convey the smell which are thrown off. For the phrase cf. ii. 823 album … iactare colorem.

proprium OQ: 'nor do they throw off from their body any peculiar scent of their own'. Cf. 855 non adhibere suum gignundis rebus odorem. But apart from the slight cacophony of ullum proprium, de corpore is otiose, and the correction of the Juntine edition, proprio, may well be right. The atoms cannot throw off any scent from their own body, because it is solid and without void (cf. 856); proprio de corpore taken together will then be equivalent to suum in 855. For the expression cf. ii. 311 siquid proprio dat corpore motus, where it is used of the body of a whole compound as opposed to that of the individual atoms.

847–53. For the point of this comparison see introductory note.

847. amaracini: 'marjoram'; again in iv. 1179 and vi. 973: usually pg 936in the shorter form amaracus: cf. Cat. lxi. 7 suave olentis amaraci and Virg. Aen. i. 693 mollis amaracus; also Plin. N.H. xxi. 61.

blandum: 'soothing', 'attractive'; perhaps 'delicious' is the best word.

stactae: 'myrrh', an odorous resin which oozes (στάζει‎) from the plant. Cf. Plin. N.H. xii. 68 sudant sponte, priusquam incidantur, stacten dictam. Cf. ibid. xiii. 17 and Plaut. Curc. 100 tu mihi stacte.

848. nardi: 'nard', prepared from a leaf: cf. Plin. N.H. xii. 42.

florem, then, cannot be literally 'the flower', but is used metaphorically 'the fragrance of nard', cf. iii. 221 Bacchi … flos.

850. quod licet: 'in so far as it is possible'; for this idiomatic use of quod in Lucr. cf. ii. 248 quod cernere possis. There is no need to follow Lambinus in correcting to quoad.

possis: the idiomatic 2nd pers. subj., cf. i. 327 n. and Prol. V b, § 10 (a); Lambinus changes gratuitously to potis es. For the combination of words and construction cf. Ovid, Ars Am. iii. 761 aptius est deceatque magis and Cic. De Am. i. 1 quoad possem ac liceret.

inolentis: a ἅπ. λεγ‎.

851. auram: 'breath', 'smell'. Cf. Virg. Georg. iv. 417 spiravit crinibus aura and Mart. iii. 65. 2 de Corycio quae venit aura croco.

853. concoctos: lit. 'boiled together'.

suo: with viro, not with corpore; Giussani wishes to take it with both.

contractans: 'touching', 'infecting'; as the line is quoted in some MSS. of Priscian; OQ have contractas and other MSS. contractos. Nonius 188. 7 quotes the line oddly as concoctosque suo servare et perdere viro. For contractans cf. vi. 854 sol nudum contractans corpus aquai.

viro: 'contamination', which it would have if it were not inolens.

854. propter eandemrem: 'for the same reason' (Lachmann) is a better correction of propter (propterea Q) eandem than the reading of Marullus propterea tandem; it is hard to assign any real meaning to tandem.

debent: 'they are bound to'; from their nature they cannot help it. Giussani emphasizes the importance of this word.

855. suum … odorem: 'smell of their own', like suo … viro 853 and proprium de corpore odorem 846. See n. there.

gignundis rebus: 'for the creation of things'. For the idiomatic use of the dat. of the gerund cf. i. 24 scribendis versibus, and see Prol. V b, § 5. 2.

856. quoniam nil ab se mittere possunt is important; the atoms being solid wholes with no admixture of void cannot send off effluences, as compound bodies can, which could convey secondary qualities. Cf. i. 613 unde neque avelli quicquam neque deminui iam concedit natura reservans semina rebus. There is again some confusion between the qualities and the effluences which cause them.

857. saporem … quemquam: the pronoun is used adjectivally; cf. ii. 234 calor … quisquam, 875 quemquam … sensum, iv. 689 haud quisquam … eorum (sc. odorum). The usage is archaic; cf. Sctum. de Bacan., 1. 10 neque vir neque mulier quisquam.

pg 937

858. calidum tepidumque vaporem: 'great and moderate heat'; see n. on teporis 843.

859–61. are extremely difficult lines, though their difficulty seems to have escaped the older editors.

1. Taking them as they stand, the meaning appears to be: the atoms cannot supply to compound things smell or sound or taste or cold or heat (854–8), 'nor the rest, which, since nevertheless they are such that they are mortal, the pliant being made of soft body, the brittle of crumbling, and the hollow of rarefied body, must all be kept apart from the atoms'. The reference, then, appears to be to the secondary qualities, quae referring to the previous list. In view of Lucr.'s confusion in this paragraph, already noticed, between the qualities and the effluences which cause them, the general sense is just possible. But there remain very serious difficulties in detail. tamen may perhaps with Munro be explained by reference to something implied, 'which, however they differ from one another, are yet all such that …', though this is forced. But what does cetera mean? He has enumerated all the secondary qualities (except colour, which does not enter into discussion in this paragraph) and there are no others to add to the list. Above all, what is the relevance of 860? Even if it is admitted that Lucr.is thinking of the concrete effluences, this description of them is very unnatural, in spite of Munro's attempt to place them in their categories (lenta fire and heat, fragosa ice (!) and cold, cava flavour, sound, smell).

2. Howard would supply corpora as the subject of constent: 'qualities which are such that the bodies they constitute are mortal'. This makes 860 a little more applicable, but still irrelevant, and grammatically it is difficult.

3. Brieger writes cetera, quae comitant, tamen ut mortalia constent, 'and the other qualities, which accompany things, yet in such a way that the things are mortal, the pliant being composed of soft body', etc. But this does not make the construction much easier and still leaves 860 irrelevant.

4. Giussani may well be right in supposing a considerable lacuna before 859, in which Lucr. explained that only things which contained void within them could throw off effluences 'such as … and the rest; but since these things are nevertheless such as to be mortal, etc. … they must be kept distinct from the atoms', i.e. we must not confuse atoms with the compound bodies which they compose, or suppose that they are like them in structure and properties. This would give good sense and get over the difficulties, but it is a large supposition.

On the whole it seems best to keep the passage as it stands and explain as in 1, but it is far from satisfactory.

859. cum … sunt: 'since they are …'; for cum in causal sense with indic. cf. i. 565 n., and see Prol. V b, § 9. 1 (b).

pg 938

860. molli Turnebus for mollia OQ, clearly right.

fragosa = fragilia, here only in this sense, but it corresponds to Lucr.'s use of fragor in the sense of 'breaking'. See i. 747 neque pausam stare fragori and n. there.

862. subiungere … fundamenta: 'to place the foundations underneath', 'to lay the foundations for'; the two words combine in one metaphor from building.

863. summa salutis: 'the sum of life', like summa rerum, summa belli, etc.; for salus in this sense cf. ii. 570 in aeternum sepelire salutem and iii. 324 ipsaque corporis est custos et causa salutis.

864 = i. 673, ii. 756.

## (c) The atoms are without sensation. 865–990.

Lucr. passes from the 'secondary qualities' of things to the problem of sensation. Sensation for him is not a possession of the atoms, but just like heat, smell, colour, etc., a quality acquired by certain compound bodies. But whereas in the production of the 'secondary qualities' the position and arrangement of the atoms is the main determining factor, in the production of sensation more stress is laid on their motions (885, 896, etc.). If the right kinds of atoms are placed in the right position and arrangements and perform the right motions, then in the compound body sensation results. Lucr. thus endeavours to place sensation among the 'secondary qualities', but there is in reality the great difference that whereas colour, smell, etc., are due to atomic arrangements in a compound, which produce an objective effect on a sentient being, sensation is a subjective and active experience in the body itself which possesses it. His solution in this case appears more crude and less satisfactory, but the same difficulty attaches to any attempt to explain sensation or consciousness as a mere function of material bodies.

The nine arguments which Lucr. adduces on this point divide themselves into three sections. In the first 865–930, which contains five arguments, he attempts to establish his own position and to refute that of those who hold that a sentient being is composed of particles which are themselves sentient. He is therefore arguing not so much against those who, like Plato, held that the ψυχή‎ was a spiritual and non-material existence, as against Anaxagoras. This is shown not only in certain parts of the argument (e.g. 902–6), but in its general resemblance to the refutation of Anaxagoras in i. 830–920.

The whole section is of course closely connected with and preparatory to the argument of Book iii. For, if it can be shown that a sentient body can be created out of non-sentient particles, it is far easier to show that the soul loses its consciousness at death, when the particles which created it are dissolved.

pg 939It might be suggested that Lucr.'s argument here is inconsistent with his earlier theory (ii. 216–93) that free will is due to the 'swerving' of the individual atoms. But the 'swerve' is in his conception purely mechanical and does not involve consciousness or sentience on the part of the atoms. In this it differs from certain modern theories, referred to in the introduction to 216–93, in that they do suppose something like a sentient atom.

## First two arguments, based on experience. 865–85.

865. Things which have sensation are yet composed of insentient atoms.

(1) 867. Phenomena offer no objection to this view, but rather lead us to think that living creatures can be created out of insensible things. Witness the creation of worms from mud after heavy rain.

(2) 874. There are other similar transformations in nature. Water and vegetable matter turn into the flesh of beasts and that, when eaten, changes into the bodies of men, which in their turn may feed beasts and birds. 879. Nature, then, changes food into living body and so produces the sensation of animals, just as wood may be changed into flame. 883. All depends on the arrangement and combinations of the atoms and their mutual motions.

Lucr. quotes two positive instances in which change in the internal arrangement of the atoms in a compound apparently gives rise to the creation of a sentient object. The argument is cruder than usual, and it is hardly necessary to point out that worms are not created by a readjustment of the particles of earth; they appear after rain if they were there before; and similarly that though food eaten may become nerves and sinews and flesh and even brain-matter, it cannot of itself produce sensation. Nevertheless the supposed creation of worms becomes a cardinal argument with Lucr. and he returns to it again in 898 and 928 and in iii. 719 and v. 797. The second argument is closely connected with the previous argument against Anaxagoras in i. 875 ff., and here and there resembles it in phraseology.

866. insensilibus: 'without sensation'; so again in 870 and 888. Like sensilis (888, 893, 895, 902) it is a word found only in Lucr. and late imitators like Arnobius.

867–70. The argument here is important and is derived directly from Epicurus' Canonice; see Prol. IV, § 9. In investigating the nature of 'things unseen' (ἄδηλα‎) sensation (αἴσθησις‎), though it cannot give immediate evidence, is yet still supreme. For (1) we must regard phenomena as giving us positive 'signs' (σημεῖα‎) with regard to the nature of the unseen (869), (2) we must not accept any hypothesis which is contradicted (ἀντιμαρτυρεῖται‎) or not supported pg 940(ἐπιμαρτυρεῖται‎) by phenomena (867–8). See Ep. ad Hdt., § 38 ἔτι τε κατὰ τὰς αἰσθήσεις δεῖ πάντα τηρεῖν‎ …ὅπως ἂν καὶ τὸ προσμένον καὶ τὸ ἄδηλον ἕχωμεν οἶς σημειωσόμεθα‎, D.L. x. 34 τὴν δὲ δόξαν καὶ ὑπόληψιν‎ (supposition) λέγουσιν, ἀληθῆ τέ φασι καὶ ψευδῆ‎· ἂν μὲν γὰρ ἐπιμαρτύρηται ἢ μὴ ἀντιμαρτύρηται, ἀληθῆ εἶναι‎· ἐὰν δὲ μὴ ἐπιμαρτύρηται ἢ ἀντιμαρτύρηται, ψευδῆ τυγχάνειν‎, and many other passages. Lucr. refers to the doctrine in i. 422–5 communis … sensus; cui nisi prima fides fundata valebit, haud erit occultis de rebus quo referentes confirmare animi quicquam ratione queamus (see n. there). He frequently acts on this principle, e.g. i. 265–328, ii. 67–70, 112–24, 308–32. Robin has a valuable note on these lines.

867. manifesta: the MSS. here have manufesta, but it is elsewhere always spelt with i; so in v. 481 maxuma.

868 is a repetition of the meaning of the previous line; in promptu cognita quae sunt is an amplification of manifesta.

869. manu ducunt: 'lead us by the hand' (χειραγωγεῖν‎), cf. Virg. Aen. iii. 372 ipse manu multo suspensum numine ducit in the literal sense, and for the metaphor Lucr. iii. 361–2 contra cum sensus ducat eorum; sensus enim trahit atque acies detrudit ad ipsas.

870. quod dico: 'as I have said'; cf. ii. 112 uti memoro, 747 quod … vinco fieri, iv. 1207 ut dico. The present tense is idiomatic. See Prol. V b, § 9. 1 (d). The parenthesis here points the contrast between insensilibus and animalia.

871. This belief in the spontaneous creation of worms, insects, etc., from mud is attributed to Archelaus of Athens and was held even by Aristotle as a γένεσις αὐτόματος‎. As Robin points out it is connected with the wider belief in the creation of all life from a primeval slime (ἰλύς‎) or clay; it may perhaps be regarded as a surviving relic of that idea. We may note the v alliteration, here perhaps intended to have an unpleasant effect.

872. putorem: used in this sense again of the muddiness of earth after rain in ii. 929 and in vi. 1101 ubi putorem umida nactast intempestivis pluviisque et solibus icta.

873. intempestivis: as in the parallel lines 929 and vi. 1102, of rain. The word should mean 'unseasonable', but in these places it seems rather to have the meaning of 'immoderate'.

874. The construction is continued from videmus in 865.

cunctas: as Giussani points out, the observation is quite general, but in its application Lucr. confines himself to the conversion of food into the bodies of animals.

875. fluvii frondes: 'streams and leaves', an asyndeton: sc. the drink and the food of cattle. The MSS. have fluvii in frondes which Martin retains, comparing i. 250–6, 'the water of rivers changes into grass and grass into the bodies of cattle'. This may be what Lucr. intended, but the general subject of the argument, the conversion pg 941of food into body, and the close parallel of ii. 596–7 unde etiam fluvios frondis et pabula laeta montivago generi possit praebere ferarum are in favour of Lambinus's omission of in, which is adopted by all other subsequent editors. Cf. also i. 885–7 in the argument against Anaxagoras.

877. naturam is equivalent to se in 875.

878. pennipotentum: a fine Lucretian compound, used again in v. 789, but not by any other writer. For the form cf. i. 33 armipotens.

879. omnis … cibos: 'all kinds of foods': note the emphatic repetition in sensus … omnis 880 and omnia 882.

880. hinc refers not to corpora viva but to cibos; it is, on Lucr.'s view, from the atomic rearrangement of the appropriate atoms of food that the sense in the body (anima) is created.

881–2. The parallel of the creation of fire out of wood goes back to the argument against Anaxagoras in i. 871–2, 891–2, 897–914, and is an additional reason for supposing that Lucr. has Anaxagoras in mind here (see introductory note). Giussani suggests that Lucr. is thinking of the Stoic belief derived from Heraclitus that fire is the basic element in the soul, but this is far-fetched and improbable.

881. alia longe: for the combination see v. 1030 non alia longe ratione and v. 1070 longe alio pacto.

atque: Q has adque here and elsewhere, but the form is never found in O; see Munro's note on its use.

882. in: a necessary addition by l 31.

omnia: sc. 'all the wood'; not strictly true, for the ashes remain.

883–5 repeat almost exactly i. 907–10 in the argument against Anaxagoras; another link: it is the shape, the order, the position, and the motions of the atoms which lead to the creation of different things. The principle is to be of great importance through the rest of this section: cf. 894–6.

884. quaeque: not each individual atom, but 'each kind of atom', as in i. 578, ii. 760, etc.

## Arguments 3 to 5, against the view that the atoms have sensation. 886–930.

The paragraph is obscure and contains many difficulties: it will be best to deal with the three arguments separately.

(3) 886. We are inclined to disbelieve that the sensible can be created out of the insensible, because we do not ordinarily see a mixture of stone, wood, earth, etc., producing life and sensation. 891. But I did not say that sensation was created immediately from all bodies which can produce sensible things, but that all depends on the size, shape, motion, order, and position of the particles. 897. These we cannot see in wood and earth, but when they are soddened by rain, they do produce worms, because the pg 942particles are moved from their former positions into those which create living things.

The proofs of the previous section were positive, instances in nature which support Lucr.'s view of the creation of the sensible out of the insensible. In this paragraph he turns to the negative side and the refutation of opponents. The third argument is comparatively straightforward and is directed against objectors to his own view. It is perfectly true that we do not ordinarily see instances of the creation of the sentient beings out of insensate matter. That is because it is not enough to have the appropriate matter present, but all depends on the arrangement, etc., of the component atoms. These we can never perceive, so that we cannot tell beforehand whether on any given occasion wood, etc., will produce a living being, but it is certain that when the structure of earth is altered by rain, worms are created. The argument is in fact an elaboration of (2).

Giussani thinks that there must be a reference to human experiments made with the intention of creating sentient beings out of inorganic matter, but this seems quite unnecessary for the understanding of the passage, and too modern in conception. Robin raises the question whom Lucr. is arguing against, and thinks it is the Stoics with their doctrine of the all-pervading πνεῦμα‎ in various tensions, some of which cause life and sensation. But it is surely more natural here, as all through this section, to suppose that he has in view Anaxagoras, who would say that sentient beings must be composed of 'particles like themselves' (ὁμοιομερῆ‎), i.e. sentient particles.

886. animum quod percutit ipsum: 'which strikes the very mind', which is the seat of sensation and therefore should know its own origin; a piece of Lucretian irony, helped out by the use of sensus in the next line. Others place a comma after percutit, taking ipsum either (with another comma after it) with id 'that very thing', which is weak, or as the object of movet, picking up and repeating animum. This is possible, but an unnecessary amplification.

percutit: cf. i. 922 acri percussit thyrso laudis spes magna meum cor.

887. sensus: 'feelings', 'opinions', almost equivalent to sententias, unusual in Lucr., but frequent in Cicero; cf. Virg. Aen. xii. 914 tum pectore sensus vertuntur varii.

varios: 'hesitating', 'wavering' from one side to the other.

888. sensile: the positive, 'that which has sense', opposed to the negative insensilia.

gigni: quoted by Priscian as nasci.

889. lapides: 'stones', 'rocks': Lucr. is giving the thought of the objector; these hard things do not produce sentient beings. ligna et terra corresponds to lignis glaebisque in 897, and they are the inorganic materials out of which Lucr. would suppose small animals might come when they were soddened by rain. But lapides is pg 943certainly odd, as he could not imagine the creation of worms out of rock. Giussani therefore proposed to read latices 'water' for lapides, which would correspond to imbribus in 873 and imbris in 898. But he is largely influenced by his own idea of experimental mixtures of water and earth (see introductory note), and the change is unnecessary if we take lapides as the ironical addition of the objector.

890. reddere: 'produce'; cf. ii. 228 quae possint genitalis reddere motus, etc.

891. rebus: the strange readings of O fedus and Q sedus must be due to an obscurity in the archetype.

892–3. 'That I did not say that sensation is created straight off out of absolutely all the things which create sensible things', i.e. it is not sufficient merely to mix up earth, wood, etc.; in a great many cases even after the falling of rain no organic creature will be produced. The construction is ex omnibus rebus quaecumque creant sensilia. For extemplo cf. ii. 763 perfacile extemplo rationem reddere possis. L has sensile for sensilia, whence Naugerius read sensile et extemplo with comma after res. This is adopted by Lachmann and Munro; the singular sensile would be paralleled in 932, but the meaning of 892 'out of all elements which produce things' is off the point.

894. quantula: 'how small'; the soul is composed of very small particles; cf. iii. 179 principio esse aio persubtilem atque minutis perquam corporibus factum constare.

897. 'None of which things can we see in wood and sods'; i.e. the atoms are far below the range of our perception, so that we cannot tell whether the right ones are there in the right order. Munro translates 'none of which conditions we find in woods and sods'. This cannot be right, for Lucr. would have said esse videmus, if he meant this, and it is not true, for some of the conditions, e.g. atoms of the right shape, must often be present, and on Lucr.'s own showing after rain sometimes all the required conditions are there.

quarum nil rerum: a combination of nihil and res which is a step towards Lucr.'s use of res picked up by neuter. Munro quotes Caesar, Bell. Gall. iii. 4. 3 quarum rerum … fieri nihil poterat and several other examples from the same author.

898. putrefacta: 'crumbled', 'made sodden'; cf. putor in 872 and 929.

899. vermiculos: the diminutive form is used again in iii. 728 vermiculorum.

900. antiquis ex ordinibus: 'from their original order'. For antiquus in this sense cf. vi. 871 rursus in antiquas redeunt primordia sedis and v. 86.

nova re: 'the newcomer', sc. the rain.

901. ita ut debent: 'in the way in which animals are bound to be born'. ut is comparative and the phrase is equivalent, as Giussani pg 944says, to ea ratione qua debent. Munro unnecessarily takes it as consecutive in sense, but 'idiomatic', and quotes parallels.

(4) 902–6. The argument is obscure and the text uncertain. With the text as given, assuming in the lacuna something like hi proprii sensus mortalia semina reddunt, we may analyse as follows:

902. Those who think that the sensible can be created out of sensible particles, which in tum owe their sensation to other sensible particles ⟨make the seeds of sensation mortal⟩, since they make them soft. 904. For in our experience sensation is always associated with nerves, etc., which are soft and therefore mortal.

After the third argument in which he answers an objection to his own view, Lucr. proceeds to two arguments in which he attacks the opposite view, that things which have sensation are composed of atoms with sensation. With the present text he will then be more directly attacking Anaxagoras: a sentient being on his theory would be composed of sentient particles, and those of smaller sentient particles, and so on. Lucr. replies that sense is always associated with the soft parts of the body and particles like them could not be immortal. It is in fact a resumption in this particular context of his charge against Anaxagoras in i. 847 adde quod imbecilla magis primordia fingit. This view is in effect that of Giussani, though he would assume a lacuna of more than one line, and it involves only the change of sueti 903 to suetis. It is adopted by Diels, who would write in the lacuna principiis (going with aliis), sumunt mortalia semina rerum, and by Martin.

902–5. Other views must be considered. OQ have

• deinde ex sensilibus qui sensile posse creari
• constituunt porro ex aliis sentire sueti
• mollia cum faciunt.

This cannot be construed, and the earlier editors tried to make sense by emendation. Lambinus suggested suetis, but did not correct the rest. Lachmann made four changes, ea for ex, seminibus for sensilibus, suetis for sueti, and iam for cum. But ea 'the particles' is vague, iam is weak, and the general sense not satisfactory. Bernays on the same lines read conficiunt for cum faciunt. Christ and Munro independently postulated a lacuna and Munro, keeping sueti, supposed in the lacuna ipsi sensilibus, mortalia semina reddunt, translating 'They who hold that the sensible can be produced out of sensible elements, accustomed thus to derive their own sense from elements ⟨which are sensible⟩ in their turn, ⟨thus render their own seeds mortal,⟩ when they make them soft.' But the parenthesis porro … sensilibus is then aimless, the point of porro is missed, and it is not easy to see against whom the argument is directed. Other editors, accepting the lacuna, fill it up variously, reading either sueti or suetis in 903. pg 945Giussani seems to me clearly right in assuming an argument against Anaxagoras.

905. quaecumque: 'all of which', in the loose sense in which Lucr. uses this word; cf. ii. 21 n. quae demant cumque dolorem. It is unnecessary, with Giussani, to write quae cumque and interpret 'which we always see', still more to read cuncta with Lachmann or cuique with Munro, who translates 'in everything'.

906. The construction probably is mollia consistere, mortali corpore creta, 'to be soft, created of a mortal body', a Lucretian involution. Giussani rather perversely takes it to be mollia creta mortali consistere corpore, though that would be a more logical statement.

(5) 907. But suppose such particles could be eternal; they must have sensation either like that of a part of the body, or like that of a whole living creature. 910. But (a) the parts of the body have no sensation independent of that of the whole, 914 (b) if they are like whole living creatures, then each of them must have the same sensations as the whole body, so as to harmonize with it. 917. But (α‎) they could not then be immortal, being living things; 920 (β‎) even if they could, they could only in union make a confused mass of sentient beings, not a living whole, any more than men could in union with beasts; 924 (γ‎) if in the body (?) they lose their own sensation and acquire another, why attribute it to them at all? 926 (δ‎) after all, since we see eggs turn into chickens and worms produced from sodden earth, we know that sensation can be created out of the insensible.

This is a compressed and crabbed piece of argument, complicated by difficulties of text, in which Lucr. successively concedes points to his opponents and shows that their position is still untenable. In (4) he argued that sensible particles would be 'soft' and therefore could not be immortal. He now waives that point and asks what kind of sense they will possess; will it first be like that of a part of a whole body? To this he replies (910–13) that the parts of the body do not feel separately; they can only feel because they are parts of a sentient whole; the sensation of a blow on the hand is not the hand's feeling but my feeling (it must be remembered that in iii Lucr. explains that the parts of the body feel because there is present in them a portion of the anima which is in close connexion with the central animus). In that case, then, the individual particles, having only this derivative power of feeling, could not have sensation of their own. Lucr. then considers the other alternative (914); has each of the particles a feeling like that of a whole and independent living creature? In that case (915–16), in order to make a harmony of sensation in the body, each of them must have independently any sensation which the whole body has. To this he makes pg 946two replies. The first (917–19) is parenthetical and not very serious; if they are such independent living things, they must be mortal like other living things and could not be indestructible primordia (cf. 902–6). Secondly (920–3) sensible atoms of this sort being put together in our body could only make a jumble of sentient beings, each having independent sensation, but could not make a whole living creature, which had sensation as a single whole. Next (924–6) he supposes his opponents to say 'yes, the sensible particles are independent sentient beings, but when they unite in the body of a living creature, they lose their independent sensations, and take on that of a sentient whole'. Then, answers Lucr., why attribute this independent sensation to them at all, if they are only to lose it, when their sensation becomes effective? Finally (926–30) he falls back once more on the apparent examples in nature of the creation of the sentient from the insensible, adding to the worms the creation of the chicken from the egg. The argument progresses step by step and is throughout based on atomic principles which become clearer in Book iii, for which this section is a preparation.

907. haec: sc. the sensible particles, which Lucr.'s opponents assume, the sensilia of 903.

908. partis: 'of a part or member of the body' such as the manus of 912.

909. 'Or be thought to be (parts) like to whole living creatures.' With some hesitation I now follow recent editors, Ernout, Diels, and Martin, in retaining similis OQ. This is strongly supported by 914, but in view of haec 906 we should expect similia: similes must be explained by the assumption of partes (nom. plur.) from partis in the previous line. Lachmann proposed simili (sc. sensu), 'or be thought to have a sensation like that of whole living beings', a compendious comparison, which is awkward and inconsistent with 914, where the particles are directly compared to whole sentient beings.

911. A very difficult line.' OQ have namque alios (alius Q) sensus membrorum respuit omnis, which cannot be translated. Hermann retained respuit, comparing vi. 68 quae nisi respuis ex animo and Cic. Pro Caec. 19. 56 hominum prudentium concilium … respuit hanc defensionem, and read nam ratio for namq̄ alios, 'for reason resists all sensation in the limbs'. But it is not 'reason' which is here concerned, but the immediate experience of sensation. For the same reason Martin's namque animus will not do, and it is palaeographically improbable. Polle suggested nam sensus, supposing namque alios to be a gloss: 'for sensation rejects all sensation in the limbs'. But as against all these proposals Lucr. could not have argued that there was no sensation in the limbs, for all through Book iii he asserts that there is sensation in the limbs owing to the presence of the anima. Lachmann rightly saw that respuit must stand for respicit, but his own pg 947namque alio 'for all the sensation in the limbs looks elsewhere', sc. is dependent on the body as a whole, is too vague, nor is it possible with Diels to retain namque alios 'for all sensation in the limbs looks to other sensations'. Giussani suggested namque animum 'all sensation in the limbs depends on the mind', but that is not how Lucr. would have put it, and it is palaeographically too remote, as Giussani saw himself. In Class. Rev. xxiv. 120 I suggested namque ad nos, which I would retain; 'for all sensation in the limbs depends on us as a whole'. Lucr. not infrequently uses nos in this vague sense 'the person', as he does in the next line a nobis, and in sentimus in 915 [923], and ad nos might easily produce alios. (In Martin's app. crit. this suggestion is assigned to Bentley, but I can find no trace of it and it is probably a misprint.)

912. secreta: 'cut off from us': the hand if severed from the body loses its sensation.

915 [923]. The line is retained in its place by Martin, but it is hard to see what sense can be made of it there. Bernays placed it here, Lachmann after the next line, Lambinus wished to eject it. It seems to fit best here, where it will be explained by the following line: 'each of the particles must have the same feeling as the whole if they are to be in harmony with the living sensation of the body.' Giussani says rightly that 917 [916] follows most naturally on 914; this is a small instance of Lucretian 'suspension' of thought, 915 [923] and 916 [915] are in effect a parenthesis.

916. consentire: 'harmonize with', 'share in the feelings of'; cf. ii. 717 vitalis motus consentire, and in a slightly different sense iii. 153, 801, where, as Ernout says, it is equivalent to συμπάσχειν‎.

917. qui: 'how', as in i. 168, 765, ii. 989 and elsewhere.

919. animalia sint (Lachmann) is generally accepted for animalibus OQ.

mortalibus una eademque: 'one and the same as mortal things'. Ernout holds that mortalibus is not here a dat., but abl. as though after a comparative, and compares Plaut. Amph. 293 nullust hoc metuculosus aeque.

920. ut possint: concessive, 'granted that they can'.

coetu: 'union', 'coming together' in a general sense; cf. i. 666 ignis in coetu stingui mutareque corpus.

921. vulgum: 'a mob': for the form as acc. masc. cf. aevum acc. masc. in ii. 561, iii. 605.

922. ut nequeant: 'as they would be unable'. The subj. is quite natural and there is no reason to alter with Lachmann to nequeunt.

homines armenta feraeque: asyndeton with -que added to the last member; cf. i. 455 n.

924. corpore must, I think, mean 'from their body', though it is tempting to take it with Giussani 'when in the body', sc. of the living creature to which they now belong.

pg 948

925. attribui: 'to be assigned to them'; cf. ii. 614 gallos attribuunt.

926. quo fugimus ante: 'to which we had recourse before' (i.e. in 871, 898), in the sense usually expressed by confugere. This is Lachmann's correction for quod fugimus OQ (which Hoerschelmann interpreted as quod supra omisi, which is nonsense), and though it is not quite the sense expected, for Lucr. did not previously regard the example of the worms as a last resort, it is the only correction which is palaeographically probable, except Martin's quod fudimus, which he apparently takes to mean 'as I said', comparing iii. 1033 animam moribundo corpore fudit, which is quite different in sense. Other suggestions, e.g. vidimus Purmann, diximus Giussani, sumpsimus Merrill, would give a more natural sense, but are too remote from fugimus.

927. quatenus: 'since' = quandoquidem.

928. alituum: the first use of this form of the gen. plur., which later poets found valuable: Lucr. uses it again in v. 1039. See Prol. V a, § 2.

effervere Marullus's correction for offervere is universally accepted; offervere is meaningless. For the 3rd conj. form effĕrvere cf. ii. 41 fervĕre and n. there, and see Prol. V a, § 12.

terram … cum … cepit is Marullus's correction for terram … quam … cepit OQ. The phrase is then clearly parallel to 872 putorem cum sibi nacta est intempestivis ex imbribus umida tellus. quam could be retained if with Martin we read terra and take it with effervere 'swarm out from the ground'; this may be right.

929. intempestivos … putor: see notes on 872–3.

930. Sums up the conclusion of the whole paragraph.

## Arguments 6–8 against the view that non-sentient atoms become sentient. 931–72.

The first five arguments are, as has been seen, directed against the view held in particular by Anaxagoras, that sentient beings are composed of particles which are in themselves always sentient. In the next three arguments Lucr. turns to another view, that insentient particles acquire sense by a process of change resembling birth. This theory would admit, as Lucr. would wish, that the particles were originally non-sentient, but differs from his view in thinking that they individually acquire sense when they go to compose a sentient being. His answer depends on the main idea that change and birth both require a concilium, a union of atoms, and could not take place in the single atoms which are not concilia. This idea is worked out in three different ways introduced by principio (937), praeterea (944), and praeterea (963), which must be considered separately below.

The argument is again obscure and is made more difficult by our ignorance of the exact theory against which Lucr. is arguing. pg 949There can be little doubt that he has now passed in mind from Anaxagoras to the Stoics, who were materialists but believed that sensation and consciousness could be produced by a greater 'tension' of the πνεῦμα‎ which is in all things. The editors usually refer to a passage in Plutarch1 where he quotes the opinion of Chrysippus that the embryo in the womb is like a plant in nature, but immediately after birth the πνεῦμα‎, cooled and restrained by the air, changes and becomes animal. This passage is really inapplicable, because it conceives a change not in individual particles but in the whole embryo, and moreover it is clear from 933 aliquo tanquam partu that Lucr. is not thinking of actual birth, but of a process of change resembling birth. But the argument does suggest this Stoic idea, which was probably applied to individual particles and was refuted in the Epicurean source on which Lucr. was drawing for this section. (Both Giussani and Ernout have useful notes on this section, and Giussani has shown clearly how the whole passage from 933–90 hangs together and is concerned with this theory, which is different from that previously dealt with.) The three arguments may now be taken separately.

## (6) 933–43.

933. To the theory that sensation can be produced from the non-sentient by a change or by a process resembling birth, it may be replied that neither birth nor change can take place without a previous union. 937. In the first place no body can have sensation before the living creature is born, because the matter which composes it is dispersed in air and water and earth and earthly things, 941 and has never met and united in an appropriate way to produce the vital motions by which the senses are kindled to keep watch over each living thing.

In this section Lucr. enunciates the new theory which he is going to refute and states the underlying objection that change or a process of birth implies previous union in a concilium. This he here supports by pointing out that sensation never exists before birth, because the individual formative particles may be anywhere in earth or sea or sky, and can only produce sensation when they unite and perform the proper motions. This idea is straightforward and follows directly from his previous arguments.

931. aliquis will then be some Stoic writer, or possibly in view of the fut. dicet one of his own followers.

dumtaxat: see ii. 123 n.; here it means 'at any rate'.

932. ex Lambinus is a better correction of ea OQ than a Q1, pg 950both palaeographically and grammatically; cf. ii. 888 ex insensilibus, 892 ex omnibus, 902 ex sensilibus.

non sensu go together 'that which has not sensation'.

sensum Lambinus must be read for sensus OQ, unless a plural verb is introduced at the end of the next line: see n. there.

mutabilitate: 'by a change', the actual process, and not, as the formation of the word suggests, liability to it.

933. Probably does not, as Giussani takes it, suggest an alternative process to that indicated by mutabilitate, but is explanatory of it. The process of change in the particles is like that of the process of birth. The creation of sensation in them is, as might be said nowadays, an epiphenomenon.

quo proditur extra 'by which it is thrust out' is the most satisfactory emendation of quod proditum extra OQ, and proditur is found in some of the Italian MSS. It involves the change of sensus 932 to sensum. Martin, following Bernays and Giussani, keeps quod taking it as a conjunction 'or else because it is thrust out by some kind of birth', but the order is awkward and it makes this a definite alternative to mutabilitate. Lachmann, keeping sensus in 932, read quod protinus extent, but protinus has little meaning and extra has the look of genuineness. For this reason Munro's quod proditus extet is also improbable.

935–6. Both birth and change imply an internal motion of parts among themselves. This can only take place in a compound body and not in an atom, whose parts are inseparable and therefore incapable of such motion (i. 597 ff.).

935. concilio: 'an act of union', here both in the atomic sense and metaphorically in the sexual sense suggested by the metaphor in partu.

936. conciliatu: 'a union', picking up concilio, but in a more concrete sense, almost 'a compound body', as in i. 575, ii. 100, 134. Goebel's nisi conciliatum is unnecessary, Christ's nisi concutiatur misses the point entirely.

937. principio: introducing, as usual, the first argument and corresponding to praeterea (944) and praeterea (963). Lachmann, not seeing the connexion of the following paragraphs with this, took it as equivalent to praecipue.

938. ipsam: a good correction in l 31.

naturam animantis = animans, the normal Lucretian periphrasis.

940. terraque creatis: 'the things sprung from earth', e.g. plants, trees and, on the Lucretian theory, animals. Cf. i. 1086 et quasi terreno quae corpore contineantur. Lachmann, again misunderstanding, proposed aethraque creatis.

941. vitalis convenientis: the accumulation of the double epithet with motus is awkward, but cf. 955 reliqui motus vitales; vitalis motus is shown to be right by its repetition in 948 and 955, and for motus convenientis see i. 1030 and ii. 712. If a change is to be made Lam-pg 951binus's convenienti with modo is more probable than Goebel's vitali. For modo adv. cf. ii. 1135 and iv. 1181; it apparently means 'now', 'yet'; see ii. 1135 n.

942. omnituentes: a characteristic compound. There is a Lucretian pun with omnituentes and tuentur (943) 'the all-seeing senses see to the safety of each living thing'. See Prol. VII, § 25. Changes proposed in 943 miss the point altogether.

943. sensus: here meaning 'the five senses', sight, hearing, etc.

## (7) 944–63.

944. A living creature is prostrated by a blow and the sensation of body and mind is disturbed, because the positions of the atoms are changed and their motions checked; 949. at last the shock through all the limbs loosens the living bonds of the soul and drives it out through the pores in the body. 952. What more can a blow do than dissipate and dissolve? 954. Sometimes, if the blow is less violent, the vital motions win; they assuage the disturbance caused by the blow, recall all the parts to their proper movements, frustrate the motion of death and rekindle the senses. 960. How else could creatures gather again their minds and return to life from the threshold of death rather than go on to the goal?

A paragraph whose meaning and connexion with what precedes is not clear, because Lucr. has not, as he does in the next paragraph, added the inference. The effect of a blow which deprives a creature of life and sensation is, he argues, due to the dispersion of the atoms from their original positions and the impeding of their proper motions, which ultimately drive the soul from the body. (But such a blow can only affect a concilium and it could not deprive individual atoms of sensation, if they had it, for they are not concilia.) Sometimes with a lesser blow the vital motions can regain the supremacy, the mind (animus) be gathered together again in its place, and the vital principle (anima) restored to the limbs, and the living being then recovers. (Again such a process is not possible except in a concilium and could not be performed in the individual atom.) The arguments here are applicable to either view, that the atoms possess sensation, or acquire it, but more particularly to the latter.

The justification for adding the inference is 967–72, where it is made explicit. But the whole argument depends on a view of sensation which is essentially Epicurean and is largely derived from Book iii, whose language this paragraph frequently anticipates. Lucr. again, as has often been noticed before, argues from his own basis and assumes his own conclusions.

944. quamvis: of course fem. acc. of quivis, as in ii. 440.

945. natura: 'its nature'; cf. ii. 17 nil aliud sibi naturam latrare, where it is 'human nature', ii. 369 quod natura reposcit, 'the nature of pg 952the lambs'. The word in Lucr. takes its exact meaning from the context.

946. corporis atque animi … sensus: sc. the anima dispersed through the body and the animus or mens situated in the breast—an anticipation of Book iii.

947–51. Note the very careful description of the stages: the position of the atoms is upset; their motions are checked; then the soul-clusters in the body are dissolved; and finally the soul is driven out of the body.

948. penitus: 'deep within', 'to the core'.

950. vitalis animae nodos: 'the vital bonds of the soul', i.e. the atomic interlacings which connect the soul-atoms with the body-atoms in the limbs; cf. vi. 356 dissoluunt nodos omnis, 878 nodosque relaxat.

951. dispersam: sc. animam.

caulas: the 'pores' or 'apertures' in the body; a Lucretian use of the word, which occurs in seven other places in the poem.

eiecit is the orthography in the MSS. here: the classical forms are eiecit or eicit, not eiicit: see Munro's note and compare traiecere iii. 513.

952–3. A blow can disperse and dissolve, but it could not do more, e.g. deprive sensible atoms of their sensation.

952. praeterea = amplius 'more'.

954. minus acriter together, 'less violently'.

955. vincere … vincere: Lucr. is fond of such emphatic repetitions, especially at these places in the line; cf. iii. 12 aurea dicta, aurea, iv. 789 mollia membra movere, mollia, v. 950 umida saxa, umida saxa, vi. 528 omnia, prorsum omnia, vi. 1167 flagrabat … flagrabat; the practice is adopted by Virgil in Aen. ii. 405 lumina frustra, lumina. See Prol. VII, § 21.

956. sedare: 'assuage', as in ii. 462 sensibu' sedatum, if the text is right there.

957. quicquid = quidque, 'each part', as frequently.

meatus: 'movements', a synonym for motus.

958. dominantem: 'getting the upper hand'.

959. accendere sensus: cf. 943 accensi sensus.

960. qua re: not 'why?' but 'in what (other) way?' probably best written, as by Munro, as two separate words.

leti limine: so again in vi. 1157 and 1208. In iii. 681 vitae cum limen inimus of the beginning of life. Modelled on the Homeric ἐπὶ γήραος οὐδῷ‎ Il. xxii. 60, etc.

961. possint: the subject is vague, animantes or possibly sensus. It is no advantage to alter with Lachmann to possit.

collecta mente: 'gathering together the mind'; it is the animus in the breast which is dominantior ad vitam quam vis animai (iii. 397), and therefore, unless it can be reintegrated, recovery is impossible. For pg 953collecta cf. iii. 925 cum correptus homo ex somno, se colligit ipse and Cic. Tusc. iv. 36. 78 quid est autem se ipsum colligere nisi dissipatas animi partis rursum in suum locum cogere?—a very close general parallel.

962. 'Instead of going on to the point to which it had almost run and passing away.'

decursum 'run to the end of a race' as in iii. 1042 decurso lumine vitae, iv. 1196 spatium decurrere amoris, and Plaut. Stich. 81 decurso aetatis spatio.

abire: here first used absolutely, according to Ernout, of passing away in death.

## (8) 963–72.

963. Pain arises when the particles of matter in the limbs are disturbed and shaken in their places; and when they return to their position, pleasure follows. 967. Therefore the atoms themselves cannot feel pain or pleasure, for they are not themselves composed of atoms, by whose rearrangement they might be disturbed, or feel pleasure. They cannot, then, be endowed with sense.

This section may be described as transitional. It hangs on closely to the two preceding proofs in that it still deals with the subject of the concilium. Pain and pleasure are forms of sensus, the two forms indeed which lie at the base of the Epicurean moral system. They are like the violent dislocations of the concilium described in the previous argument, which result in complete or threatened extinction of life in that they are due to some vis which disturbs the atomic complex. (A fuller account of the process has been given in ii. 434–41, where three possible forms of vis are enumerated.) But there is the difference that whereas the ictus of 944–62 drives the atoms out of their place and resolves their nodos, the vis which causes pleasure and pain attacks them suis … in sedibus (965); it disturbs them, but does not move them quite away from their natural abode, so that they can return again to their place (locum 966), and then pleasure follows. Here Lucr. makes his inference explicit and shows that the individual atoms cannot have this form of sensus, because they are not compound bodies formed of separable and movable particles.

But the argument is, as Giussani has shown, also a return from the proof based on the concilium to the general thesis that the atoms are insentient, and this is enforced in the concluding line. Once again the argument is based on strictly Epicurean conceptions and anticipates Book iii.

963. praeterea: the third link in the argument about concilium. Lachmann, who misunderstands this passage altogether, reads propterea, thinking that this argument was an inference from the preceding.

964. vi quadam: the vis may, as we learn from ii. 434, be either pg 954something which enters from without or arises within, or remains outside. Giussani in his explanation of this passage (n. on 931–72) takes account only of the first possibility.

vi … viscera viva: the alliteration has no apparent significance here.

965. Cf. v. 162 sollicitare suis ulla vi ex sedibus.

suis in sedibus: important, see introductory note.

intus reinforces suis in sedibus; cf. iii. 171 ossibus ac nervis disclusis intus adacta.

966. blanda voluptas: 'enticing pleasure', as in iv. 1263, v. 178.

968. ex se: 'arising from themselves', because there can be no internal rearrangement.

969. sunt ex: 'consist of', = constant ex.

970. quorum … laborent: 'so that by a change in their motions they might feel pain'. quorum is in effect equivalent to ut eorum, as in iv. 116. For novitate cf. iii. 150 cum cetera pars animai per membra atque artus nulla novitate cietur: so caeli novitate vi. 1103. In ii. 1040, v. 930 it is used in the more original sense of 'newness'. For laborent cf. i. 849, and iii. 176 quoniam telis ictuque laborat.

971. fructum … dulcedinis: cf. v. 1410 capiunt dulcedini' fructum.

## (9) 973–90.

973. If in order to explain sensation in all living creatures we must suppose sensible particles, what about the special particles composing men? 976. They must laugh and cry and argue about the composition of things and their own 'first-beginnings', since they, which are like the whole men they compose, must in their tum be composed of smaller particles like themselves. 983. Indeed everything that speaks and laughs and thinks must be made of particles which do the same. 985. But this is madness: a man can laugh and think and speak, though he is not composed of particles which do the same. 989. Why then can we not say that things which have sense are composed of insentient particles?

In this last argument Lucr. leaves the Stoic theory and comes back again to Anaxagoras from whom he started, using the same reductio ad absurdum which has already appeared in i. 919–20 (possibly added there after this passage was written). If the theory of ὁμοιομέρεια‎ is to be pressed, then man must have not only the sentient particles which he has in common with the animals, but over and above these he must have special 'seeds' which can talk and laugh and cry, as he does. The argument is not very serious, and Lucr. is not justified in drawing from this absurdity the general conclusion (989–90) that sentient things may be composed of insentient particles. But he loves to conclude a long argument with an ironical jeu d'esprit, as in i. 919–20, iii. 776–83.

973. uti: is not concessive 'although', but final, the uti clause being subordinate in sense to the si clause; 'if, in order that all living pg 955things may feel, we must …'

animalia: including men, who are specially designated as mortalibus in 980.

975. propritim: 'specially', 'peculiarly', a ἅπ. λεγ‎. with Lucr.'s favourite termination -im: see n. on generatim i. 20, and Prol. VII, § 6.

de quibus auctumst: is rightly explained by Brieger 'of which the human race is composed in addition'. Here and in 986 Lachmann, followed by Bernays, Munro, and Diels, reads de quibu' factumst and ex ridentibu' factus, on the ground that auctus, auctum ought to be constructed with the simple abl. not with de or ex. But they failed to observe the special sense here, which also explains the construction; it does not merely mean 'increased by', but 'made in addition out of', and the analogy of the usual construction of factus is followed in order to secure this meaning.

976–7. Cf. i. 919–20

• fiet uti risu tremulo concussa cachinnent
• et lacrimis salsis umectent ora genasque,

a typical instance of Lucr.'s habit of making small variations in his repetitions.

978. de rerum mixtura: 'about the mixture of particles in things'.

dicere callent: for the prolate infin. cf. ii. 66 tu te dictis praebere memento, and see Prol. V b, § 12 (a).

979. sibi proporro … quaerunt: cf. v. 312 quaerere proporro sibi. The present line supports the genuineness of these words in that very difficult line.

980. totis mortalibus assimulata: cf. ii. 914 totis animalibus assimulentur; mortalibus is substituted here with special reference to human beings; see n. on 973.

982. alia: undoubtedly the right correction of alii O1, ali OQV.

983. quippe sequar: 'I will push here the argument'; cf. i. 980 hoc pacte sequar. It is followed by the rather loose construction ut sit.

ridereque: -que attached to ě; see i. 134 n.

984. sapere: 'to think', as again in 987 and in iii. 145 idque sibi per se solum sapit.

985. delira: 'insane'; cf. i. 698 quod mihi cum vanum tum delirum esse videtur.

furiosa: 'raving mad'. Smith well compares Gaius, Inst. iii. 106 furiosus nullum negotium gerere potest, quia non intellegit quid agat.

986. auctus: see 975 n.

987. doctis … dictis: the jingle occurs already in Ennius.

989. qui: 'how?'

990. undique = omnino, as in ii. 916.

## (d) Summary and conclusion on the secondary qualities and sensation. 991–1022.

991. We are all sprung from heavenly seed. The sky is the father pg 956of us all, from whom mother earth, receiving the raindrops, conceives and brings forth crops and plants, men and beasts, since she provides for them their sustenance. 998. Truly she deserves the name of mother. 999. Thereafter what came from earth returns to earth, and what was sent down from the sky returns to the sky. 1002. Death does not destroy matter, but only dissolves its unions and by forming fresh combinations makes things change their forms and colour and acquire sensation and yield them up again instantaneously. 1007. All depends on the arrangements and combinations and motions of the same atoms. 1010. Neither the secondary qualities on the surface nor the sensation that is created and perishes are the eternal properties of the atoms. 1013. So in my verses all depends on the combination and order of the letters. The same letters indicate sky, sea, earth, rivers, sun, crops and plants, and living creatures; they are not all the same, though the greater part are, but they differ in their position. 1019. So in things, when the meetings, motions, order, position, and shapes of the atoms are changed, things too are altered.

This paragraph is the summary of all that Lucr. has said about the secondary qualities and sensation from 730 onwards. From the strictly Epicurean view his own doctrine is stated in 1002–12, which are a summary of the doctrine of concilia and of their acquisition and subsequent loss of secondary qualities and sensation. To this he has prefixed a fine poetical passage in which he identifies his view with the traditional explanation of the old myth of the sky-father and the earth-mother. The idea of the earth-mother has already appeared in the poem in i. 250–61, ii. 598–9, but that of the sky-father, which is referred to in i. 250, is elaborated here for the first time. There can be little doubt that Lucr. is here closely following and almost translating a passage in the Chrysippus of Euripides (fr. 839), which is always said to be founded on the teaching of Anaxagoras, the friend and master of Euripides:

• Γαῖα μεγίστη καὶ Διὸς αἰθήρ‎,
• ὁ μὲν ἀνθρώπων καὶ θεῶν γενέτωρ‎,
• ἡ δ‎ʼ ὑγροβόλους σταγόνας νοτίας
• παραδεξαμένη τίκτει θνητούς‎,
• τίκτει δὲ βορὰν φῦλά τε θηρῶν‎,
• ὅθεν οὐκ ἀδίκως‎ | μήτηρ πάντων νενόμισται‎.
• χωρεῖ δ‎ʼ ὀπίσω
• τὰ μὲν ἐκ γαίας φύντ‎ʼ εἰς γαῖαν
• τὰ δ‎ʼ ἀπ‎ʼ αἰθερίου βλάστοντα γονῆς
• εἰς οὐράνιον πάλιν ἦλθε πόλον‎·
• θνῄσκει δ‎ʼ οὐδὲν τῶν γιγνομένων‎,
• διακρινόμενον δ‎ʼ ἄλλο πρὸς ἄλλου
• μορφὴν ἑτέραν ἀπέδειξεν‎.

pg 957The same doctrine is seen in Euripides' Melanippe (fr. 484):

• ὡς οὐρανός τε γαῖά τ‎ʼ ἦν μορφὴ μία‎·
• ἐπεὶ δ‎ʼ ἐχωρίσθησαν ἀλλήλων δίχα‎,
• τίκτουσι πάντα κἀνέδωκαν εἰς φάος
• δένδρη, πετηνά, θῆρας, οὕς θ‎ʼ ἅλμη τρέφει
• γένος τε θνητῶν‎:

and the conception of the return of the elements to their own place is found in Supplices 532–4:

• ὅθεν δ‎ʼ ἕκαστον ἐς τὸ σῶμ‎ʼ ἀφίκετο‎,
• ἐνταῦθ‎ʼ ἀπελθεῖν, πνεῦμα μὲν πρὸς αἰθέρα
• τὸ σῶμα δ‎ʼ ἐς γῆν‎.

These ideas in Euripides are probably rightly attributed to Anaxagoras, his teacher, but, as Ernout points out, that is no reason for supposing that they were the invention of Anaxagoras. He quotes for instance from Empedocles (Diels B. 8):

• οὐδέ τις οὐλομένου θανάτοιο τελευτή‎,
• ἀλλὰ μόνον μῖξίς τε διάλλαξίς τε μιγέντων
• ἐστί‎.

Another passage of Empedocles (Diels B. 23), in which he compares the process of the formation of things to the mixing of colours by painters, may well have suggested to Lucr., or to Epicurus before him, the parallel of the letters in words. Nor was the idea previously unknown to the Romans; Ennius in his Epicharmus (fr. 48 Vahlen) says of Ops mater

• terris gentis omnis peperit et resumit denuo.

The use of this poetic allegory by Lucr. shows both his fondness for allegorical explanation (cf. ii. 600–60) and a desire which he exhibits more rarely to agree with his opponents when he can. But a certain stretching of these traditional ideas has to be permitted to adapt them to the Epicurean point of view. The conception of mother earth is entirely consonant with Lucr.'s ideas, and indeed in v. 805–15 we meet an even more literal idea of the birth of living things from the earth than here, where her function is mainly to supply their food (996). The part played by the rain from heaven, too, though it does not figure largely in Lucr. (cf., however, i. 250–1) is not inconsistent with Epicurean notions and may have commended itself to Lucr. here, because the thought of the creation of worms from the rain-sodden earth was so much in his mind (871–3, 898–9, 928–9). But the idea of the return of the elements to their places (999–1001) after death meant something very different to Lucr. from what it did to Euripides. For the latter, as is clear from the pg 958Supplices passage, it was the body which returned to earth, and the spirit (πνεῦμα‎) to the sky; this would have been an easier doctrine for a Stoic than for an Epicurean. Yet Lucr. might say without stretching the idea too far that the elements of heat, wind, and air which with the quarta natura compose the soul (iii. 231–57) were derived rather from sky than from earth and after the dissolution of the soul might return to the upper regions (cf. aere in ii. 940).

The important point for Lucr. in this physical explanation of myth is of course its reinforcement of his doctrine that sentient beings are not composed of sentient atoms, but are compounds of insentient particles derived from sky and earth. This he states in his own language in 1002–12. Death is material dissolution, not destruction; the death of one thing means the union of its particles in fresh combinations, with new shape, new colour, and new sensation, which they again lose at their dissolution; all depends, he repeats in familiar lines, on the positions, arrangements, and order of the atoms, and so, he adds in a new formula, secondary qualities and sensation are not the permanent properties of the atoms, but, he might have added, the properties or accidents (coniuncta, eventa) of compound bodies. And then follows the familiar illustration from the letters in words, repeated here in phrases already used and less aptly fitted than elsewhere to the context. Editors are probably right in thinking this was a patchwork passage, hastily composed by Lucr. as he was seized with the thought of using his illustration once again.

991. ŏr$iu¯$ndī: a trisyllable, in which the i is not consonantalized, as in ābiĕtĕ, pāriĕtĕ, etc., familiar in the Augustan poets, but is totally suppressed. Lachmann, in a very learned note, shows that this use is not infrequent in the comic poets, and is sometimes found elsewhere. His doctrine was attacked by later scholars, but L. Müller, De re metrica, p. 249, substantiated dŏmĭniă in Lucilius, ŏpĕr$iu¯$ntur in Laevius, and mĕl$iu¯$s in Varro. Lucr. never elsewhere employs this licence, but here it can only be accepted. For the sense of oriundi 'sprung from' cf. Ennius, Ann. 113 o sanguen dis oriundum.

992–3. Cf. i. 250–1 and see introductory note. The idea here is reproduced in yet another fragment of Euripides (898): ἐρᾷ μὲι ὄμβρου Γαῖ‎ʼ, ὅταν ξηρὸν πέδον‎ | ἄκαρπον αὐχμῷ νοτίδος ἐνδέως ἔχῃ‎· | ἐρᾷ δ‎ʼ ὁ σεμνὸς Οὐρανός, πληρούμενος‎ | ὄμβρου, πεσεῖν εἰς γαῖαν Ἀφροδίτης ὕπο‎. | ὅταν δὲ συμμιχθῆτον εἰς ταὐτὸν δύο‎, | φύουσιν ἡμῖν πάντα, καὶ τρέφουσ‎ʼ ἅμα‎, | δι‎ʼ ὧν βρότειον ζῇ τε καὶ θάλλει γένος‎. For the echo mater … terra see Prol. VII, § 24.

994–5. There is a question as to the punctuation of these lines. Most editors place a comma after genus humanum, taking it with parit in 994. Brieger, followed by Giussani, objects to the coupling of men with the crops and plants rather than with the beasts, and pg 959places commas after laeta and parit (995). But this involves an awkward asyndeton with omnia saecla ferarum and the traditional punctuation is better.

996. pabula cum praebet: explaining parit, but not very satisfactorily. Contrast v. 805 ff. where the earth parit mortalia saecla in a more literal sense; and see introductory note. It must be remembered that to Lucr. the processes of birth and growth, being both that of the aggregation of atoms, were essentially the same; see introductory note to i. 159–214.

omnes: sc. genus humanum et saecla ferarum, a good instance of Lucretian reference κατὰ σύνεσιν‎; see Prol. V b, § 7.

997. dulcem vitam: possibly, as Ernout suggests, a recollection of Homer's γλυκὺς αἰών‎.

998. The phrase is repeated, though on slightly different grounds, in v. 795, 821.

999–1001. See introductory note.

999. de where ex would be expected; cf. i. 788 de terra in sidera mundi, v. 794 de salsis exisse lacunis. de and ex are almost interchangeable in Lucr.; see a useful note of Merrill's on i. 283.

terra: but terras in 1000 in exactly the same sense, for metrical reasons.

1000. missumst as quoted by Lactantius, for missus OQ.

1001. rellatum: compounded with red-; see i. 228 n. rĕlatus in v. 686.

1002–3. See the lines of Empedocles quoted in the introductory note and a fragment of Anaxagoras himself (Diels B. 17): τὸ δὲ γίνεσθαι καὶ ἀπόλλυσθαι οὐκ νομίζουσιν οἱ Ἕλληνες‎· οὐδὲν γὰρ χρῆμα γίνεται οὐδὲ ἀπόλλυται ἀλλ‎ʼ ἀπὸ ἐόντων χρημάτων σιμμίσγεταί τε καὶ διακρίνεται‎.

1004. inde: either 'then' or 'from this material'; more probably the former.

coniungit et efficit: mors (1002) is still the subject; the death of one thing causes a new combination and the change of form and colour, etc. The idea is quite Lucretian, cf. the often repeated lines i. 792–3 nam quodeumque suis mutatum finibus exit, continuo hoc mors est illius quod fuit ante and the idea in iii. 964–7. Giussani quotes the proverb mors tua vita mea. Lachmann and Munro, objecting both to mors as subject and to the paratactic efficit … convertant, emended, Lachmann to coniungitur et fit ut, Munro to coniungit et effit ut, justifying effit by effiant, which is by no means certain, in vi. 761.

1005. res ita: Marullus, to obviate the parataxis, read res ut; again unnecessary.

1006. puncto tempore: 'in an instant of time', sc. at death. For the difficult phrase see ii. 263 n.

reddant: 'give up', 'lose' their sensation.

pg 960

1007–9. Cf. i. 817–19 (against Anaxagoras), 908–10, ii. 760–2, 894–6, a cardinal Epicurean doctrine.

1007. eadem: 'these same' first-beginnings, emphatic. A certain correction of earum OQ.

1010. penes residere: 'to remain permanently in the possession of'. penes, which is connected with penitus, contrasts with in summis (1011), residere with fluitare, nasci, and perire.

potesse: cf. i. 665 n.

1011. quod in summis … perire: Brieger is probably right in saying that quod in summis fluitare videmus rebus is the secondary qualities and in particular colour, with which Lucr. dealt at the greatest length in ii. 730–841 and which he has just recalled in 1005; while (quod) interdum nasci subitoque perire is sensation, the subject of 865–990 recalled in 1006. Giussani would prefer to take the whole clause quod … perire together as referring to phenomena as opposed to the internal ἄδηλον‎, the atoms. In any case Lachmann's substitution of cunctis for summis shows a misapprehension of the passage.

1012. interdum nasci subitoque perire corresponds to capiant sensus et puncto tempore reddant 1006. Lucr. is writing very carefully here.

1013–18. The illustration from the letters in words has occurred before in i. 823–6, 912–14, etc. The present passage is something of a cento from other lines, but there is no reason to suspect it or to bracket 1013–22 with Brieger and Giussani as a later addition by the poet. Munro excludes 1015–16 as out of place here, and Lachmann brackets from 1013 to 1104, considering the passage on the many worlds as a later addition; see introductory note to 1023–1104. In the lines as they stand there is just that combination and modification which Lucr. loves in a repeated passage.

1013 = i. 823, 1014 = ii. 884, 1015–16 = i. 820–1, with the characteristic and necessary change of constituunt to significant.

1017. Cf. ii. 458 si minus omnia sunt e levibus atque rutundis, at non …; which supports Lachmann's change of sint OQ to sunt. But the subj. is possible in the sense 'even if they were not all to be the same'.

1018. discrepitant: a word first found in Lucr. and used by him again in iii. 803 and vi. 1105. See Prol. VII, § 5. 'In i. 582 he has discrepat, impersonal, in the same sense.

1020 = i. 685 and recalls the beginning of ii. 727. 1020 in the MSS. intervalla, vias, conexus, pondera, plagas = ii. 726, and cannot be right here as the accusatives have no construction. It probably came from a marginal quotation showing the likeness of 1021 to 727.

## D. The Infinite Worlds, their Formation and Destruction. 1023–1174.

This, the fourth and last main section of the book, deals, as Robin pg 961has pointed out (n. on 1048–74), with a question which was much discussed among the Greek philosophers. Parmenides, Empedocles, and probably Anaxagoras, followed later by Plato and Aristotle and by the Stoics, insisting on the necessity for unity, held that there was but one world; on the other hand, Anaximander, Archelaus, and Diogenes stood for the plurality of worlds. The latter opinion was naturally adopted by the Atomists and by Epicurus,1 but there was again a difference. Leucippus (D.L. ix. 31, Diels A. 1), thinking that the atoms were in origin together in a mass of material, held that from time to time a portion of them fell off into 'a great void' (μέγα κενόν‎) and there formed a world. Epicurus with his view of the ceaseless movement of atoms in space could not thus conceive either of a mass of atoms or of a completely empty portion of space; he held rather that in any part of space a group of atoms might fall into the combinations and movements which would form a world. This need not even take place in a μετακόσμιον‎ but might happen within another and disintegrating world. This opinion Lucr. of course follows.

This section has no very close connexion with what precedes, but just as the previous section on sensation prepares the way for Book iii so this leads up to the subject of Book v. If the first draft of the poem was Books i, ii, v (see Prol. II, §§ 7–13), the collocation would be clear and obvious. It is written fully in the manner of the conclusions of the other books, a great conception set out in great language. Indeed its vision is not unlike the description of the infinite universe with which Book i ends (921–1113) and gives a clearer picture of its content. Lachmann brackets from 1023 to 1104 as a later and irrelevant addition of the poet. But 1105 would not follow 1012 at all happily and the discussion of the infinite worlds is an essential preparation for Book v.

## (a) Introduction. 1023–47.

1023. Now attend to a new and great theme: all great truths are difficult to believe at first, yet none is so great and wonderful but that in time it ceases to cause wonder. 1030. Think of the blue sky and all that is in it, stars, moon, and sun; if these were shown to men for the first time, what could there be more wonderful or incredible? 1038. Yet now we are weary of the sight and no one deigns to look up at the sky. 1040. Therefore do not in fear reject the theory, but weigh it carefully and, if it be true, yield to it; if false, fight against pg 962it. 1044. I propose to ask what there is in the infinite space outside our world, into which our mind can of its own will project itself.

The great theme is introduced by a solemn preface, calling attention to its importance, just as the conclusion of Book i is by 921–50. The idea that familiarity diminishes wonder is a commonplace, and Cicero1 has a closely parallel passage in which he connects it, as does Lucr., with the wonders of the heavens. It is interesting that there the reflection is put into the mouth of a Stoic.

1023. Cf. i. 921 nunc age quod superest cognosce et clarius audi, and for the form of expression i. 50 animumque sagacem semotum a curis adhibe veram ad rationem.

1024. vementer is required by scansion and is shown by Lachmann to be the classical form; vehemens may be due to a false analogy with vehere. Diels and Martin curiously write vehementer.

nova res: note the insistence on the novelty of his view; cf. 1025 nova species, and 1040 novitate … ipsa.

1025. accidere: accedere OQ may be an archaic form: they have the same form in v. 609.

1027. difficilis magis = difficilior ad credendum: in iv. 889 facilest factu Lucr. prefers the supine construction.

1029. quod non paulatim … omnes: lit. 'wonder at which all would not lessen by degrees'. quod mirarier is apparently the object of minuant. If so, this is the only instance in Lucr. of infin. used substantivally as the object, though there are several in which it is used as the subject; see Prol. V b, § 12 (b). Munro compares Ter. Andr. 392 nec tu ea minueris quae facis and Hec. 616 non minuam meum consilium, but though these illustrate the meaning, they do not affect the awkwardness of the construction. A closer parallel is found in a Pompeian graffito, C.I.L. iv et gelidae cursu minuerunt quaerere silvam. It may be that minuant is here equivalent to mittant 'cease' and that mirarier is a prolate infin.; cf. iv. 471 mittam contendere causam, iv. 690, vi. 1056. Lachmann emends to mittant, but that is unnecessary; minuant may be taken in either of the two suggested ways and should be retained.

mirarier is a generally accepted correction for miralier O1, miraliter OQV.

1030. principio … colorem … nitorem; … omnia quae (nom.); a typical Lucretian anacoluthon, cf. i. 445–8, and more closely ii. 342–7 genus humanum mutaeque natantes … quorum unum quidvis. See Prol. V b, § 2. The earlier modern editors, objecting to the anacoluthon and the absence of any second argument to follow principio, pg 963emended, Lachmann to percipito, a very unnatural word, and Bernays to suspicito comparing suspicere in 1039. But there is no real difficulty in the anacoluthon though accusatives picked up by nom. are more remarkable than the reverse process in the parallel passages; for the absence of a second argument cf. ii. 144 primum. Munro gratuitously supposes a line lost after 1029 of the form cuius, uti memoro, permulta exempla videmus, the accusatives of 1030–2 being then constructed after videmus.

1031. cohibet Lachmann must be right, the stars are what the sky 'holds in itself'. With cohibent (OQ) sidera will be the subject; but what do the stars contain?

palantia sidera: not necessarily 'the planets', πλάνητες ἀστέρες‎, as the editors explain, but all the stars, for according to Lucr. they all move in the turbo; cf. v. 644 quae volvunt magnos in magnis orbibus annos.

1033–5. si essent. si sint … poterat: an odd sequence of tenses for a classical author, but not impossible in Lucr. Indeed it may be held that there is a real difference of meaning; 'if they had been now for the first time, if they were at some future time to be, presented to men, what more wonderful could be conceived?' essent by itself is weak and obiecta repente probably goes both with essent and with sint. Editors have tried to correct the 'solecism': Lachmann and others adopt Orelli's extent for essent, but poterat implies an imperf. or pluperf. subj. in the protasis; Munro reads si nunc for si sint in 1034—an awkward construction—and Diels visunda for si sint, removing the comma after essent. This improves the construction but is palaeographically improbable and in itself awkward; i. 727 and ii. 577, which he quotes, do not justify visunda obiecta.

1035. poterat: the idiomatic imperf. indic. in unfulfilled apodosis: cf. v. 18, 1051, and n. there and see Prol. V b,. § 9 (c).

1036. gentes: a synonym for mortales (1033) or homines.

1038. quam: probably exclamatory; 'how true it is that no one …' i. 104, vi. 1080 quam multa, iv. 1203 quam saepe, vi. 801 quam facile are all much easier because of the following adj. or adv., but Cic. Att. ix. 11. 2 quam ille non probare mihi quidem visus est! quam illum νεανίαν‎ … timere is a good parallel. The alternative is to place comma or semicolon after fuisset 1037 and to take quam as acc. referring to species governed by suspicere, in caeli lucida templa being then an amplification (Ellis, J. Phil. iii. 261): 'a sight which no one now deigns to look at, gazing up into the sky', or with Giussani to take quam as object of videndi. But both these proposals are awkward and weaken the sense.

nemo, fessus stands naturally for nemo, cum fessus sit.

satiate: an alternative form to satietate, found in ante- and post-classical Latin. Cf. v. 39 satiatem, and see Prol. V a, § 2, and VII, § 4.

pg 964

1039. suspicere … dignatur: for the infin. cf. iii. 1045 indignabere obire, and see Prol. V b, § 12 (a).

1041. exspuere: cf. vi. 68 quae nisi respuis ex animo.

1043. dede manus … accingere: both metaphors from fighting. For the former cf. ii. 1129 manus dandum est.

1044–7. Munro oddly wishes to bracket these lines as a later addition by the poet. But it is essential after this long preface that he should state what his intended subject is.

1044. summa loci: 'the totality of space'; see i. 235 n.

1046. usque: with quo, 'the place right up to which'.

1047. animi iactus liber: 'the unfettered projection of the mind', Epicurus' ἐπιβολὴ τῆς διανοίας‎; cf. ii. 740 animi iniectus and n. there, and Prol. IV, § 6. But here it is not directed to a particular object, but is 'free' to roam over infinite space; cf. Cic. De Nat. Deor. i. 20. 45 quoted on 739.

## (b) Proofs. i. 1048–66.

1048. There is, as I have already shown and as nature shows, no limit to the universe. 1052. Therefore since space extends to infinity and seeds innumerable fly through the void in eternal motion, it is improbable that this one world should have been created and that all the matter outside it should produce nothing. 1058. For this world itself was made naturally when atoms of their own accord, after clashing and meeting many times with no result, at last brought together those which could be the foundations of a great world, of earth, sea, sky, and living creatures. 1064. So once again it must be admitted that there are similar aggregations of matter elsewhere, like our world which is held in the embrace of the sky.

This proof and the next must be taken together in close conjunction: they might seem at first sight to show little difference, but a closer examination shows that in this paragraph the efficient cause of the formation of a world is 'nature', 'probability', 'chance' (τύχη‎), whereas in the second it is 'force' or 'necessity', ἀνάγκη‎. Here we have an instance of the most important of all Epicurus' divergences from the Atomists. For Democritus the one ultimate cause was 'necessity'; a rigid determinism went right through his system. But Epicurus, largely for the sake of his ethical system, wished to preserve (see introductory note to ii. 216–90) the element of chance (τύχη‎), or, as Lucr. seems more usually to have regarded it, of 'spontaneity', something originating in the nature of the atoms themselves. This becomes clear when we compare the two accounts of the creation of a world. For Democritus' view the most explicit statement is that of Diogenes:1 'he said that all things come to pass pg 965by necessity, since the whirl is the cause of the creation of all things, and the whirl he calls necessity'. A more explicit1 account states that the atoms 'moved about, as it happened, in the void, and automatically collided with one another owing to an undetermined movement … and thus produced the world and all that is in it—or rather the innumerable worlds'. In this latter account there might seem to be an admission of chance as a concomitant cause with necessity, but it is probable that by 'automatically' Democritus meant 'in accordance with the laws of their own being',2 and that his conception of chance was of that which, though due to necessity, is beyond our capacity to predict; his argument is directed against the view that the world was made by design. For Epicurus, on the other hand, chance was an independent operative force, due probably to the 'swerve' of the atoms, which from time to time brought things to pass in defiance of the law of necessity, principium quoddam quod fati foedera rumpat ii. 254 (see introductory note to ii. 216–93).3 It would seem that Epicurus' considered opinion4 was that the original motion of the free atoms in the void was disorderly (ἄτακτος‎) and that chance brought about the formation of them into a cosmic whirl, but that when once an ordered world (κόσμος‎) was formed, 'necessity' or 'the laws of nature' took command, though the swerve might still bring about exceptions. If that is so, Lucr. in these two paragraphs seems to have been oblivious of this distinction, for he clearly places 'automatism' or 'chance' (sponte sua forte 1059) and 'force' (vis 1072) side by side as two concomitant causes of the formation of a world, just as in vi. 31 (seu casu seu vi) he suggests chance and force as equal causes of the evil in the world. In the present paragraph, then, Lucr. insists on the spontaneous action of the atoms as the cause of creation, whereas in the next he returns to the Democritean conception of necessity.

1048–9. Note the completeness of the description; the universe is boundless in every direction, from either side, up and down. For in cunctas … partis cf. i. 966 in omnis … partis, and 1007 in cunctas undique partis.

1049. ex: where classical writers would have had ab. See ii. 999 n., pg 966and Merrill's note on i. 283.

⟨supra⟩ subterque: Lachmann's certain correction for the deficient superque of OQ.

1050. uti docui: sc. i. 958–83.

res ipsaque: notice the postponed -que, not infrequent in Lucr.; cf. ii. 394 hamatis inter se perque plicatis, and n. there.

1051. vociferatur: 'proclaims aloud'; cf. i. 731 carmina … vociferantur.

elucet: 'shines clear'; the intransitive verb after docui and vociferatur is unexpected. Ernout suggests that Lucr. may have intended the verb to be transitive here, 'shows clearly'.

natura profundi: 'the nature of the deep', profundum being in effect equivalent to vacui or spatii; cf. i. 1002 est igitur natura loci spatiumque profundi (cf. v. 370), vi. 485 summamque profundi.

1053. undique versum: 'in all directions'; cf. ii. 188 sursus … versus.

1054. innumero numero: a jingle of which Lucr. is fond: it recurs in iii. 779, and cf. ii. 1086 numero … innumerali and vi. 485 innumerabilem … numerum.

summaque profunda: 'in the sum-total of the void', cf. vi. 485 summamque profundi; see n. on i. 235. Merrill strangely takes profundă as nom. 'vast in their number'; this is quite contrary to Lucretian usage.

1055. aeterno percita motu: 'driven on by eternal motion'; so again in iii. 33.

1056. hunc unum: 'only this one world'.

1057. foris: 'outside the limits of this world', 'in outer space'.

1058–61. Note the accumulation of words by which Lucr. enforces the spontaneous or casual movements of the atoms, natura, sponte sua, forte, temere incassum frustraque, coluerunt, coniecta.

1058–63. A complicated and much-discussed sentence; 'above all since this world of ours was so made by nature, and the seeds of things themselves of their own accord, jostling from time to time by chance, driven together in many ways, rashly, idly, and in vain, at last those united, which, suddenly cast together might become ever and anon the beginnings of great things, of earth and sea and sky, and the race of living creatures'. The construction is irregular and ipsa semina is picked up in a 'distributive' sense by ea; the meaning is 'as all the seeds of things jostled together, at last those came together which were suitable for making a world'. Cf. ii. 105–6 cetera … paucula.

1058. hic: 'our world', the hunc unum of 1057.

natura: 'by the action of their own nature', opposed alike to divine intervention and to the rigid 'necessity' of Democritus. Lucr. probably had in mind the clinamen as the ultimate cause of the meeting of atoms.

et ipsa OQ, which there is no reason to alter: the description goes on in a paratactic manner. Lachmann, in order to have a caesura in the next line, inserted ut (= quemadmodum 'as') between offensando pg 967and semina and Munro, rightly rejecting that proposal, yet influenced by the idea of a subordinate clause, gratuitously read ut for et here. ipsa goes closely with sponte sua as in ii. 1092 ipsa sua per se sponte and 1158 sponte sua … ipsa; cf. iii. 1041, iv. 736, v. 871, 1146, vi. 1020, and see Prol. V b, § 8. Giussani, not recognizing this characteristic Lucretian accumulation and wishing for further emphasis on hic, reads factus et ipse.

1059. The line is strictly without caesura, but a quasi-caesura is caused by the division between the preposition and verb of of-fensando. Cf. vi. 197 complerunt, magno in-dignantur murmure clausi, iii. 612 dissolui. quod si im-mortalis nostra foret mens, 715 haud erit ut merito im-mortalis possit haberi, v. 165 desiperest. quid enim im-mortalibus atque beatis, and possibly iii. 258 nunc ea quo pacto in-ter sese mixta quibusque and vi. 1067 quae memorare queam in-ter se singlariter apta. See Prol. VI, § 3. Lucr. seems to have been particularly conscious of the division in im-mortalis. Lachmann's insertion of ut after offensando (see 1058 n.) is therefore metrically unnecessary.

1060. temere incassum frustra: a marvellous accumulation of synonyms, repeated in v. 1002.

1061–3 are repeated with variations in v. 429–31 tandem conveniant ea quae convecta repente magnarum rerum fiunt exordia saepe, terrai maris et caeli generisque animantum.

1061. cōluerunt Lachmann, for OQ's colerunt. coluerunt = coaluerunt 'they have come together'; Lachmann has a learned note justifying the synizesis and comparing it with other instances. In Lucr. we have vi. 1068 colescere (coolescere OQ), vi. 491 coperiant, v. 342 coperuisse (cooperuisse OQ). This correction is better than L's colarunt 'they were sifted out', which is adopted by Munro, who quotes Ep. ad Hdt., § 73 πάντων τούτων ἐκ συστροφῶν ἰδίων ἀποκεκριμένων‎ and—a closer parallel—D.L. ix. 31 (of Leucippus, Diels Leuc. A. 1) ὥσπερ διαττώμενα‎.

ea picks up ipsa (1058) in a limiting sense Those of them'; see n. on 1058–63.

coniecta OQ Thrown together' may certainly be kept and is confirmed by conicere ii. 1073 and coniecta 1074. Lachmann in v. 429 rightly emended conventa to convecta and thinks that Lucr. must have written the same word here. In fact the change is characteristic. Brieger's concreta is gratuitous and not the right sense; the word here should express the preliminary clash.

1062. magnarum rerum: sc. 'worlds'.

exordia: a certain correction for ex ordine OQ.

semper: "always', sc. on all occasions when the right atoms are brought together, but in v. 430 saepe, sc. this collocation of the right atoms frequently happens. Both are right and the change characteristic. Lachmann again requiring identity in the lines alters in v. 430 to semper.

pg 968

1065. congressus materiai: 'gatherings of matter'; so again in v. 67.

1066. hic with short ĭ; see ii. 387 n.

avido complexu: 'in its eager embrace': the epithet is purely picturesque. It is used again of aether in v. 470 omnia sic avido complexu cetera saepsit.

## 2. 1067–76.

1067. Moreover, when there is abundant matter and space, and no thwarting force, things are bound to be created. 1070. There is an innumerable sum of atoms, and the forces of nature retain the same power to bring the atoms together in other places as they have here. 1074. Therefore there must be other worlds in other parts of space and different races of men and beasts.

This is not, as Munro believes, 'a mere variation of the last paragraph'. As Giussani rightly sees, whereas the emphasis in the last paragraph was on chance and spontaneity, here it is on force and necessity; note debent 1069, vis 1072; see introductory note to 1048–66. Notice too that the argument of the paragraph is syllogistic: whenever there is sufficient matter, ample space, and the necessary force, there will be atomic movement and the creation of things. These three conditions are fulfilled in space outside our world. Therefore other worlds will be created. Robin further notices a new point in the last three lines. Whereas in 1066 the deduction was the formation of other worlds 'like ours', here the emphasis is on worlds different (alios 1075) from ours and containing different (varias 1076) races of men and animals. He compares Ep. ad Hdt., § 45 οἵ θ‎' ὅμοιοι‎ (κόσμοι‎) τούτῳ καὶ οἱ ἀνόμοιοι‎ … ὅσοι τοιοῦτοι‎ … ὅσοι διάφοροι τούτοις‎. As Merrill notes, the idea in 1066 that these other worlds are inhabited by men and beasts is also a new point. It is vouched for in Ep. ad Hdt., § 74 οὐδὲ γὰρ ἂν ἀποδείξειεν οὐδείς, ὡς‎ ⟨ἐν‎⟩ μέν τῷ τοιούτῳ καὶ οὐκ ἂν ἐμπεριελήφθη τὰ τοιαῦτα σπέρματα, ἐξ ὧν ζῷά τε καὶ φυτὰ καὶ τὰ λοιπὰ πάντα‎ ⟨τὰ‎⟩ θεωρούμενα συνίσταται, ἐν δὲ τῷ τοιούτῳ οὐκ ἂν ἐδυνήθη‎. Robin in his note shows that this view was also held by Anaxagoras.

1067–9. The statement is here quite general and forms the major premiss.

1067. multa: for the purposes of a union to form things in any given place 'much' matter is sufficient; it need not be infinite.

1068. nec res nec causa: 'neither matter nor force'; res repeats materies … multa, and causa adds the third requisite.

1069. geri debent … et confieri res: res is used slightly differently with the two verbs; processes must be carried on and compound things created.

1070. et followed by ⟨que⟩ 1072 is an unusual sequence, to which pg 969Lachmann objects. He would read ex, which makes an almost impossible construction.

seminibus: the usual 'Lucretian' dative, where the gen. will not scan; see i. 58 n. and Prol. V b, § 5.

1071. After this line Brieger, followed by Giussani, supposed a lacuna, in which the second of the requisites enumerated above, namely space, should be mentioned. But this is demanding a super-Lucretian logic of Lucr., and in point of fact it is implied in in loca quaeque 1073.

1072. vis eadem natura manet OQ, of which in spite of Lachmann's objection Marullus's visqueeademetnatura manet is far the best correction; vis is required by the argument to represent causa in 1068. Lachmann himself reads quis (dat.), and Diels, following Postgate, hisque in the same sense, but this destroys the parallelism with 1067–9. Similarly Brieger's sique weakens the argument, for natura by itself is not here a sufficient representative of causa.

1073. quaeque queat 1 31 for the manifestly corrupt quaequeat O, quaequead Q.

quaeque goes with semina, not loca: 'each class of seeds into its own place'.

1075. A Lucretian interweaving: the construction is alios terrarum orbis in aliis partibus.

1076. varias: probably not a mere variant for alias, but implying difference.

## 3. 1077–89.

1077. Add to this that in the universe nothing is created or grows which is unique: all belong to some class of which there are many examples. 1080. Consider the animals; it is true of the beasts, of men, of fish and birds. Therefore sky, sea, earth, sun and moon, and the other features of a world are not unique but there are countless numbers of them, 1086. for they too have a beginning and an end just like all things on earth.

After the two a priori arguments in the previous paragraphs, Lucr. passes after his manner to an a posteriori argument, an analogy drawn from the things of experience. Here we meet again with the Epicurean notion of ἰσονομία‎, the equal distribution of things (cf. ii. 522–46). It is founded on arguments closely resembling those of the previous paragraphs: if there are everywhere matter and space, and the laws of nature (or the workings of chance) are uniform, then there will be an equal number of things of each class created, though (cf. ii. 532–40) they are not necessarily equally distributed. This is true, he says, of all classes and species of living things which we know. Now the great constituent parts of a world, sky, sea, earth, etc., are like the species we know in that they have a birth and a death, are created and dissolved. They must, then, be like pg 970all other created species and cannot be unique. The argument at the end of the paragraph is rather compressed.

1077. in summa: 'in the sum-total of things', i.e. the universe; see n. on i. 235.

una, unica … unica solaque: note the emphatic repetition of the idea.

nulla … una: an emphatic negative; cf. iii. 263 nil ut secernier unum.

1078. gignatur … crescat: both the birth of things and their growth are important for the argument; it is the same coming together of atoms which causes a thing to grow after birth until, as we learn in ii. 1131 ff., it is no longer able to take in more than it loses.

1079. alicuiu' siet: a certain correction by Gronovius of alioquoiuis siet O, aliquoiuis Q.

saecli: 'race', 'class', equivalent to genus.

permultaque … genere: again the repetition of the same idea adds emphasis to it.

1080. in primis animalibus indice mente invenies: OQ. This might be retained, as it is by Martin, in the sense 'among animals first you will find with the mind to guide you'. in primis animalibus must then be taken, as Munro explains it, as equivalent to primum in animalibus: he compares iv. 468 invenies primis ab sensibus esse creatam notitiam veri, but that is not the same; it means 'arises first from the senses'. For indice mente Martin compares ii. 677–8 cetera consimili mentis ratione peragrans invenies, but that is really a mental process, whereas here reasoning is not concerned; it is a matter of observation. The traditional emendation is Gronovius's inclute Memmi, cf. v. 8. But there seems no reason for the intrusion of Memmius at this point, and it does not get over the difficulty of in primis animalibus. Lipsius made the brilliant suggestion inice mentem, 'turn your mind to the animals', ἐπίβαλλε τὴν διάνοιαν‎, another reference to the Epicurean ἐπιβολὴ τῆς διανοίας‎, as in ii. 740 animi iniectus (see n. there) and 1047 animi iactus liber; see Prol. IV, § 6. This has been adopted by most editors since, and it has the additional advantage that in primis can be taken in its normal sense 'in the first place', animalibus being the normal dat. after inice. No other suggestions except Tohte's adice mentem, which was halfway to the right solution, need consideration.

1081. montivagum: cf. i. 404, ii. 597. According to Lucr.'s usual practice elsewhere it might be gen. plur. with ferarum. But ii. 597 montivago generi … ferarum is decisive for regarding it as neut. acc. with genus.

1082. genitam: picking up genus in 1081. But 'the begotten offspring of men' is an odd expression and it is possible that geminam OQ may be right, 'the twin offspring of men', i.e. male and female. Virgil has geminam … prolem in this sense in Aen. i. 274, which might be an imitation of Lucr.

pg 971

1083. squamigerum: gen. plur. cf. i. 162 squamigerum genus.

1085. cetera quae sunt: 'and the rest', sc. the other parts of a mundus, stars, clouds, and so on.

1086. numero … innumerali: see ii. 1054 n.

magis: 'rather' = potius.

1087–9. 'Since the deepset boundary mark of life as much awaits these and they are as much formed of a body which has birth as every race which exists in our world (hic) abounding in things after its own kind', and therefore, since they resemble the things of our experience in their creation and mortality, they must also be things of a class (see introductory note).

depactus terminus alte: metaphor from the boundary stone, driven deep in a field, a variation for alte terminus haerens i. 77, 596, v. 90, vi. 66.

1088. haec: sc. earth, sun, moon, etc.

nativo corpore constant: cf. ii. 452 fluvido quae corpore liquida constant.

1089. genus omne: 'every class'. The text of the rest of the line is uncertain. O has quod his generat in rebus abundans, Q his generatim; generatim must be right, but the phrase does not construe. The simplest emendation is Lachmann's est generatim, but est could not easily give his and reference is needed to things in our world. Bernays wrote hic generatim rebus abundat; hic is probably right, cf. ii. 1120 omnibus hicaetas debet consistere rebus, but abundans does not look like a corruption. Munro puts it right with hic generatimst rebus abundans. Diels and Martin have recently returned to his, following Munro in the rest of the line; this may be right, his meaning 'things such as we know'. Brieger's hic generatumst rebus abundans is improbable: genus generatumst seems an unnatural expression. Though the text is doubtful I think Munro is most likely to be right.

## (c) A digression against the theological view. 1090–1104.

1090. With this in mind we can see that nature, free from all divine tyranny, can accomplish all by itself. 1093. For, by the gods, who live a quiet life in peaceful serenity, who could guide the vast universe, keep all the many heavens turning and hold the reins of the deep, warm the earths with heavenly fires, 1101. be present everywhere to cause cloud and thunder and hurl his lightnings, with which he often destroys his own temples, and then retire to the desert to practise with the bolt which often misses the guilty and slays the innocent?

Lucr. pauses in his discussion at an important point, as he does in ii. 167–83, at great length in v. 110–234, and in a passage closely resembling the present in vi. 379–422, to protest against the theological view that the events of the world are controlled by divine pg 972agency; see Prol. IV, § 15. There are several passages where Epicurus similarly pauses and protests.1 The form of the argument here is traditional and almost every stage of it can be paralleled from Aristophanes Nubes (see commentary). Two points may specially be noticed, firstly that the passage which begins in a solemn and serious form degenerates in 1101 into irony and sarcasm, and secondly that, although he denounces the idea of the gods' interference he speaks of the gods themselves in terms of a fine reverence, which may be compared with ii. 646–51 and v. 1161–82.

1090. teneas: the idiomatic 2nd pers. subj.: 'you, the reader'. Cf. i. 327 n., and see Prol. V b, § 10 (a).

1091. dominis … superbis: 'tyrants', sc. the gods, if regarded as controlling the affairs of the world; cf. vi. 63 dominos acris adsciscunt. So ironically Cic. Tusc. i. 21. 48 liberatos enim se per eum (sc. Epicurum) dicunt gravissimis dominis, terrore sempiterno et diurno ac nocturno metu.

privata: 'free from': see i. 47 n.

1092. ipsa sua per se sponte: a very emphatic accumulation; cf. ii. 1058 n., and see Prol. V b, § 8.

dis … expers: 'freed from gods'; similarly constructed with abl. in vi. 1181 expertia somno, as also in Sallust, Cat. 33. 1 plerique patriae sed omnes fama atque fortunis expertes sumus; but usually in Lucr. and elsewhere with gen.

1093–4. For the tranquil life of the gods see ii. 645–51 (i. 44–9).

tranquilla … pace belongs in sense to the following relative clause, but is placed here for emphasis; cf. ii. 1133 n. This is better than to regard it as a descriptive abl. with pectora.

1094. aevum vitamque Avancius; aevo multamque OQ is a good example of the wrong division of words (cf. i. 32) coupled with the common confusion of i and l.

1095. immensi … profundi: each of these two words is probably to be taken as a substantive; for the former cf. i. 74 omne immensum peragravit and for the latter i. 1002 natura loci spatiumque profundi. Munro and Giussani prefer to take profundi as subs. and immensi as adj., the whole going with both clauses: quis regere immensi profundi summam, quis habere immensi profundi … habenas? But this seems unnecessarily complicated. In any case immensi summam is a notable oxymoron; it means here 'the universe', looked at as an infinite extent of space, as is made clear by profundi.

1096. indu manu = in manu, the archaic form being used for scansional reasons. It occurs elsewhere (see n. on i. 82) in compound verbs, e.g. indugredi, and substantives, e.g. induperator iv. 967, pg 973v. 1227, but here only as a preposition, v. 102 iacere indu manus is like it in sound, though it stands there for inicere.

moderanter: ἅπ. λεγ‎., but i. 323 moderatim, ii. 169 admoderate. For the formation of adverb from participle cf. iii. 779 praeproperanter, 1063 properanter,