^{page 409 note 1} I must thank Professor F. H. Sandbach, Dr. G. E. R. Lloyd, Dr. J. F. Procope, and the editor of CQ and his adviser for their very helpful comments on earlier drafts of this paper. I must also thank Dr. H. M. Hine for permission to use his Oxford D.Phil. thesis ‘An edition with commentary of Seneca Natural Questions book two’ (1975; cited as ‘Hine’); 1 have generally adopted his text of book II, and owe much to his very thorough commentary.

For other books of Seneca's Naturales Quaestiones I have mainly used Oltramare's, P.Bude edition (Paris, 1961;Google Scholar cited as ‘Oltramare’), and for Aristotle's Meteoro-logica the Loeb edition of H. D. P. Lee (London, 1952).

^{page 410 note 1} The most detailed study I know is that of A. Brennecke, Animadversiones ad fontes Naturalium Quaestionum Senecae (Diss. Greifswald 1913; cited as ‘Brennecke’).

^{page 410 note 2} I must acknowledge much help from Hine, pp.387–402.

^{page 410 note 3} Cf. . ‘A latere eliditur’ (Oltramare's reading) I take to mean ‘is squeezed out by pressure from the side as clouds come together’, i.e. by the clouds exerting pressure from both sides on the exhalation at right angles to its resultant motion (… , Mete.369^a22–3). But Hine, p.391, points out that ‘a latere’ is not found in the reliable manuscripts, which have ‘alterum’; this makes no sense, and he reads ‘actarum’ (conjectured by Brennecke, p.3 7). I prefer ‘a latere’, as closer to Aristotle.

^{page 411 note 1} ‘Vi latus’ Gronovius, for MSS. ‘ut’ or ‘ubi lathis’. See Hine, pp.391 f.

^{page 411 note 2} See 369^a16–25, besides the lines quoted.

^{page 411 note 3} (middle, ‘holding them selves around’, unless E. W. Webster's is right-see The works of Aristotle translated, ed. Ross, W. D., vol. 3, Oxford, 1931).Google Scholar Hine, pp.393–5 (though admitting most scholars interpret as I do) thinks ‘the surrounding clouds’ means the clouds which originally surrounded the exhalation; but this is surely wrong: those clouds set the exhalation in motion (see n.3, p.410), so how can it be borne violently against them?

^{page 411 note 4} As Hine points out, this appears from ‘exprimi collisis nubibus’ (see below), whatever the correct reading and interpret ation of ‘coitu … eliditur’.

^{page 411 note 5} Seneca turns into spiritus, reasonably, since exhalation causes wind; and Aristotle's explanation does involve a mass of the stuff encountering flame; but Aristotle says nothing of moisture, nor of exhalation bursting. (I have been much helped by Hine's commentary here.)

^{page 412 note 1} i.e. ‘condensing to water’, cf. ^b14–16.

^{page 412 note 2} Kroll's conjecture for ‘esistit’.

^{page 412 note 3} So Haase, for MSS. ‘itaque’.

^{page 412 note 4} See373^b14–21. As Brennecke pp.28–9 points out, Seneca has already spoken of the effect of raindrops in causing rainbows in 1.3.5–6 (connecting this with 1.3.7 by ‘Aristoteles idem iudicat’). But not even 1.3.5–6 speaks of raindrops beginning to form.

^{page 413 note 1} This is reinforced by the fact that 2.12.3, reporting Anaxagoras and (appar ently) Empedocles, is evidently derived from Mete.369^b 12 ff., which follows my quotation on p.411. (See below, p.428 and n.l.)

^{page 413 note 2} There is not enough evidence to show how accurately Seneca is reporting Theophrastus.

^{page 413 note 3} A less accurate statement of Aristotle's exhalation theory than in 2.12.4 (see p.410).

^{page 413 note 4} Mete.2.8. (Seneca does not state clearly which exhalation is cause; but nor, in this chapter, does Aristotle. Theoretically, dry exhalation causes wind(360^b10ff. etc.), which causes earthquakes (366^a3 f. etc.); but in 2.8 moisture is also involved in producing the exhalation to which earth quakes are due (365^b24–7, 366^b9 f.).

^{page 413 note 5} Cf.Mete.341^b7 ff., quoted on p.410.

^{page 414 note 1} ‘Suetam’ Garrod; MSS. ‘suam’.

^{page 414 note 2} Pliny N.H. 2.96.

^{page 414 note 3} But perhaps we should read that the phenomenon , for the (surprising) of Mete.343 ^b 23. However, neither manuscripts nor ancient commentators support this, nor would it justify Seneca's ‘Aristoteles ait non trabem … fuisse’.

^{page 414 note 4} For another attempt to derive Seneca's statement (indirectly) from Mete, see Brennecke, pp.46 f.

^{page 414 note 5} See Brennecke, pp.29–30. Aristotle's explanation must have been too mathemat ical for Seneca.

^{page 415 note 1} Since Seneca seems to mean that Aristotle said ‘statim cometes … minatur’, but that in fact ‘annum totum suspectum facit’. However, he could mean that Aristotle said ‘non statim cometes … minatur, sed annum totum’ etc. This seems forced, but would be nearly correct; Brennecke, p.48, seems to understand Seneca thus.

^{page 415 note 2} 344^b34 ff., the ‘great comet’ (that of N.Q. 7.5.4) coincided with high wind; 368^b6 ff., the winds then caused an earth quake; cf. 343^bl f.

^{page 415 note 3} Seneca is of course thinking, as was Aristotle, of the view that wind causes earthquakes.

^{page 415 note 4} As to Theophrastus, the spurious De signis (34) says that comets . This adds nothing.

^{page 415 note 5} See Oltramare ad loc.; Brennecke, pp.49 and 50.

^{page 415 note 6} F131a Edelstein-Kidd, lines 33–6 (from Scholia to Aratus 1091):. Cf. F131b lines 20–2.

^{page 415 note 7} Oltramare, p.4; Brennecke, pp.18–9.

^{page 415 note 8} Some passages might conceivably be derived from some lost work; but it seems very improbable, because we hear of hardly any lost works likely to have discussed these topics. The only exception is N.Q. 7.28: comets as weather-signs may well have been discussed in the Aristotelian work on weather-signs (mentioned, for example, at Diogenes Laertius 5.26) or in Aristotle's own (which dealt with winds: Mete.363^a24. The extant Problems 26 is on wind but contains nothing relevant to Seneca). But it is unlikely that any Aristotelian work contradicted Mete.'s statement that comets are a sign of drought and may affect a year's weather, since chat seems to have been the regular view: see above, nn.4 and 6, p .415, and Aratus 109 3, .

^{page 416 note 1} Some scholars (e.g. Oltramare, pp.xvii and 4 f.; Brennecke, p.50) have argued that Seneca knew Aristotle's meteorology only at second hand.

^{page 416 note 2} Plutarch, De tranquillitate animi 464F, says that that work is based on his , and the younger Pliny tells how his uncle made voluminous notes from his reading, on which (evidently) he based his published works (Ep. 3.5.10, 15, 17). Seneca may have made similar notes.

^{page 416 note 3} Cf. p.426. Bennecke, pp.49 f., denies the possibility of bad memory, notes, or assistants, but for no good reason I can see.

^{page 416 note 4} I am concerned only with statements about Plato's and Aristotle's doctrines, ignoring (for example) biographical anec dotes and places where Plato is mentioned as an archetypal philosopher. I have mostly used the Bude editions of Seneca's works, from which (and sometimes also from the Teubner editions) I have taken many of the references to passages in Plato and Aristotle.

^{page 417 note 1} In discussing Ep.58 and Ep.65 I have made some use of E. Bickel ‘Senecas Briefe 58und65’ (Rheinkcbes Museum 103 (1960), 1–20),Google Scholar and also a little use of Scarpat, G., La lettera 65 di Seneca. 2nd edn. (1970).Google Scholar

^{page 417 note 2} Passages supporting the view that he did are Phaedo 102 d, distinguishing from , cf. 103 b () and Parmenides 130 b (distinguishing ). Interpreters disagree over whether ‘largeness in us’ (etc.) is really something distinct, in Plato's mind, both from the Form ‘largeness’ and from large objects (see G. Vlastos, ‘Reasons and causes in the Phaedo’, in his Platonic Studies (1973), pp.84 f. and notes). It at least seems clear that Plato did not develop this notion of immanent characters, nor does he in the dialogues call them ).

^{page 417 note 3} The Idea of the Good, and Ideas in general, are said in effect to excel everything and to exist ‘really’ in a way other things do not (Republic 509 a–b, ; Phaedrus 247 c-e ; cf. Seneca's ‘exsuperat omnia’, ‘per excellen-tiam esse’), and Ideas are called divine (e.g. Phaedo 80 b ); but they are not said to be deoi or Serfc. The Demiurge is called 6 fled? (Timaeus 30 a 1 etc.), but is not explicitly a supreme sort of being, since he is not said to be superior to the Forms (although many interpreters have supposed that he is so; see Grube, G. M. A., Plato's Thought (1935), pp.156 ff.).Google Scholar

^{page 418 note 1} On the second and fourth kinds see the two preceding notes. The first and third kinds are different ways of describing the Forms (which are , and immortal of sensible things: Republic 507b-510, 596 b; Timaeus 27 d-29); the fifth kind seems to be a way of referring to genera and species, which in Platonic terms would mean the Forms once again; the sixth kind, ‘quae quasi sunt’, suggests what is said of the Receptacle in Timaeus 49 a–51 a, and of the objects of Srfja which ‘partake in being and in not-being’ in the Republic (e.g. 478 e). (However, Bickel op.cit. gives rather different interpretations, especially of the fourth and fifth kinds; which shows how remote from Plato all this is.)

^{page 418 note 2} See Bickel, op.cit., especially p.4.

^{page 418 note 3} He identifies the Receptacle with (Timaeus 52 a–), which might be identified with void; but the latter identification is not made in the Timaeus.

^{page 418 note 4} To the Stoics, time and void are and therefore but not (SVF ii.329, 331; cf. Theiler, quoted by Bickel op.cit. p.4)-note that Seneca, Ep.58, does not accept this, allowing the classification of incorporalia among quod est, and evidently confining the class of which are not to figments such as centaurs (Ep.58.11–12, 15). (Bickel disagrees with Theiler, but his view (op.cit. pp.6 f.) seems to me much less plausible.)

^{page 419 note 1} This is not to say that they must be from the same source: as Bickel points out (op. cit., p.6), Ep. 58.16–20, unlike Ep.65, makes no use of the concept of materia, though we might expect it in a discussion of kinds of being.

^{page 419 note 2} Agent: Pbilebus 26 e identifies and Timaeus 29 a, the Demiurge is . Form: Pbaedo 100 c, of other things' beauty. Purpose: Timaeus 29 d–30 b. (On immanent forms see above, p.417, n.2.)

^{page 419 note 3} Cf. especially 50 a–b, comparing the and gold.

^{page 419 note 4} Aristotle Pfcys.209^bll ff., cf., e.g., Plutarch, De animae procreatione 1014 d-e, 1015 c; Albinus, Didascalicus, cap.8.

^{page 419 note 5} Cf. evidence cited in Scarpat, op.cit., pp.124 f.; cf. Bickel, op.cit., p.14.

^{page 419 note 6} Ep.65.7: ‘Deus intra se habet… plenus his figuris est, quas Plato ideas appellat.’

Cf. Philo Judaeus, De cherubim 49, God is , cf. De opificio mundi 17–20; Albinus, Didascalicus 9 etc. (On this theory see A. N. M. Rich, ‘The Platonic Ideas as the thoughts of God’, Mnemosyne ser.4, 7 (1954), 123–33; A. H. Armstrong ‘The background of the doctrine “That the intelligibles are not outside the intellect”’, Entretiens Hardt 5 (1960), 391–413;Google ScholarMerlan, P. and Chadwick, H. in Cambridge History of Later Greek and Early Medieval Philosophy (1967), pp.66 and 142.)Google Scholar

^{page 419 note 7} See Timaeus 28–30.

^{page 420 note 1} Its precise position among Middle Platonist theories is beyond the scope of this paper; see, for instance, Bickel, op.cit.; Merlan in Cambridge History of Later Greek … Philosophy, pp.54 f.

^{page 420 note 2} Ep.58.8: ‘Amicus noster … hodierno die dicebat’-this expression suggests verbal, not written communication.

^{page 420 note 3} Cat.2^a11 ff.; Top. 144^a28–36. Cf. An.Posf.91^b4 ff.; Metaph.B 995^b29–31, Z 1038^b17 ff.

^{page 420 note 4} Metaph.B 998^b22 ff., H 1045^b5 f., K 1059^b27–34.

^{page 421 note 1} See Porphyry's Isagoge, p.4.21–6 Busse. Cf. Bickel, op.cit., p.2.

^{page 421 note 2} The evidence is set out and discussed by J. Fillion-Lahille, ‘La colère chez Aristote’ (REA 72 (1970), 46–79). 1 must acknowledge the help of an unpublished paper, and much advice, by J. F. Procopé m this section.

^{page 421 note 3} E.N. 1106^b16–23, 1107^a8-ll.

^{page 421 note 4} Other Aristotelian parallels are listed by Fillion-LahUle, op.cit. 76–8; I have quoted some of the most striking.

^{page 422 note 1} Fillion-Lahille, loc.cit., compares certain Aristotelian passages with these; but the parallels are far from close.

^{page 422 note 2} e.g. Fillion-Lahille, op.cit., pp.46 ff.

^{page 422 note 3} Besides Seneca, see Philodemus, De ira, col.XXXI, 31 ff.; Cicero, Tusc.4.19. 43 ff., off.1.25. 89; Lactantius, Inst. 6.19.

^{page 422 note 4} Tusc. 4. 19. 43.Google Scholar

^{page 422 note 5} Bonitz, H., Index Aristotelicus (1870), s.v. . See Top.l56^a31 ff., . Cf. Rhet.1378^a31–3 (similar, but more elaborate).Google Scholar

^{page 423 note 1} See Sandbach, F. H., The Stoics (1975), p.157.Google Scholar

^{page 424 note 1} For instance, it may make men (etc.; 953 33 ff.), or . (954^a31–3).

^{page 424 note 12} Quoted in F. S. Forster's Oxford Translation of Problems (1927) ad. loc.

^{page 424 note 13} This, Préchac's text (Budé, 1962), is partly conjectural, but the sense is not in doubt.

^{page 425 note 11} So at N.Q. 1.1.7–9, 1.3.7 ff., De ira 1.6.5–one might add N.Q. 7.5.4–5, where, though Aristotle is mentioned twice, the second mention is a question, not a statement of fact.

^{page 425 note 2} As at N.Q. 1.8.6, Ep.58.9 ff., Ep.44.4. Similarly, at Ep.58.22 the citation ends sooner than we might suppose after the last mention of Plato.

^{page 425 note 3} N.Q. 6.13.1, De ira 1.9.2, 3.3.1 and l,Ad Marciam 23.2.

^{page 425 note 4} Ep.65.4–6.

^{page 425 note 5} e.g. Ep. 58.22 (p.418)j ad Marciam 23.2 (p.423); Tranq. 17.10 (p.424); perhaps Ben. 4. 33.1, N.Q. 5. 18.16 (p.423).

^{page 426 note 1} He may have modified his report of Aristotle in N.Q. 7.5 A for his own purposes: see p.414.

^{page 426 note 2} N.Q. 6.13.1 is a possible example (p.413).

^{page 426 note 3} e.g. De ira 1.9.2, presented as direct quotation from Aristotle, where Philodemus and Cicero speak of ‘Peripatetics’, reporting their view in oratio obliqua (pp.421 f.).

^{page 426 note 4} See p.420; Bickel, op.cit. p.7.

^{page 426 note 5} See Ziegler's RE article ‘Plutarchos’ (1951), cols. 914–28; Helmbold, W. C. and O'Neil, E. N., Plutarch's Quotations (1959),Google Scholar preface; Russell, D. A., Plutarch (1972), pp.42 ff.Google Scholar

^{page 427 note 1} See Guthrie, W. K. C., History of Greek Philosophy I (1962), 41–3;Google ScholarKirk, G. S. and Raven, J. E., The Presocratic Philosophers (1957), pp.3–7.Google Scholar

^{page 427 note 2} See e.g. Kirk and Raven, loc. cit. cit.

^{page 427 note 3} For discussions see Hine pp.384–7, 433, and, on N.Q. 2.12.3, Brennecke pp.32–4.

^{page 427 note 4} .Cf. 369^b 19–20: .

^{page 428 note 1} Since like Mete, it also quotes what is evidently Empedocles' theory (though without naming him), and Seneca goes on to give Aristotle's own theory in 2.12.4.

^{page 428 note 2} MSS. have ‘Anaxandrus’; Seneca has just given Anaximenes' and Anaximander's theories in 2.17–18, so cannot mean either of them; and Anaxagoras is the only pre-Socratic whose lightning theory is said by any source to have involved (save for Archelaus, who adopted Anaxagoras' theory: Aetius 3.3.5 DK 60 A 16).

^{page 428 note 3} But some MSS. of Hippolytus read for .

^{page 428 note 4} Aristotle, Mete.339^b22, 369^b14; Cael. 270^b24, 302^b4.

^{page 429 note 1} Aristotle Cael.294^b 13 ff. (DK 13 A 20) etc.

^{page 429 note 2} See frags. 12, 15, and Theophrastus, Sens. 59 (DK 59 A 70).

^{page 430 note 1} The earthquake theories of Anaxagoras and (as seems likely: v. DK 12 A 28 and n.) Anaximenes are also reported by Ammianus 17. 7.11–2: I ignore this here, as no esti mate of Ammianus' accuracy is needed for my inquiry.

^{page 430 note 2} N.Q. 4b.9, on how different bodies absorb heat, and 5.2 (DK 68 A 93a), on wind.

^{page 430 note 3} Mentioned only at Mete. 349^a16 ff., 353^b6 ff., 355^a21 ff., 360^a17 ff.

^{page 430 note 4} Anaximander (DK 12 A 24 etc.), Anaximenes (DK 13 A 7(7) etc.), Xenophanes (DK 21 A 46, B30), Heraclitus (DK 22 A 1 (10)), Anaxagoras (DK 59 A 42(11) etc.), Diogenes of Apollonia (DK 64 A 17), Democritus (see n.2 above), and Metrodorus of Chios (DK70A 18).

^{page 430 note 5} Vict. 2.38 and Nat. Puer. 2.24–5.

^{page 430 note 6} Nu.404–7.

^{page 430 note 7} e.g. Aetius 3.3

^{page 431 note 1} This is confirmed by other passages where Aristotle and Aetius report the same theory: compare Mete.342^b25 ff. with Aetius 3.2.2 (DK 59 A 81), Mete. 345^a13 ff. with Aetius 3. 1.2 (DK 41.10, 58 B 37c). (Mete. 348^b12–4 neither confirms nor contradicts Aetius 3. 4.2 (DK 59 A 85). I ignore passages where Aristotle reports a theory without saying who held it.)

^{page 431 note 2} Cf. Bicknell, P. J., ‘Seneca and Aetius on Anaximander's and Anaximenes’ accounts of thunder and lightning', Latomus 27 (1968), 181–4.Google Scholar

^{page 432 note 1} The clue must have been in the lost beginning of book 4b.

^{page 432 note 2} For a detailed discussion see Hine, pp.419–32; for a different view, Bicknell, op.cit.

^{page 432 note 3} Oltramare, with most editors, reads ‘Ita, ut Anaximenes’, thus making Anaximenes' theory an illustration of the preceding sentence, which compares thunder to the noise of red-hot iron quenched in water; but Hine omits Ita, pointing out (pp.420 f.) that Anaximenes' theory makes a bad illustration and that Ita is not found in the best manuscripts.

^{page 433 note 1} Hine pp.419 f. says that ‘omnia ad spiritum retulit’ means that spiritus was the first principle of Anaximander's cosmology. But I cannot believe that Seneca would introduce (and dismiss) so abruptly, and with so little explanation, a point so irrelevant to the context: Hine compares 3.13.1, on Thales' first principle, but that is much more explicit (see below). If ‘omnia’ cannot mean ‘all the phenomena I am talking about’, then (as Hine alternatively suggests) haec or ista must be added to the text.

^{page 433 note 2} Imaginary dialogue: e.g. N.Q. 2.1.3–4; 2.13.3; 2.14.1; 2.25–6; 2.28.1. Details and rare variant phenomena: e.g. 1.1.15 (thunder from a clear sky); 2.26.8–9 (lightning on clear nights); 2.27 (varieties of thunder); 2.30.1 (thunder and lightning associated with volcanic eruptions); 2.40 (varieties of thunderbolt).

^{page 433 note 3} Diogenes of Apollonia, according to N.Q. 2. 20.1, explained thunder unaccompanied by lightning, and Mete. 369^b1 ff. explains the of thunder. I do not think any of the three questions in ‘Quare inaequalia … in sonum valuit’ are answered elsewhere in any account of a pre-Socratic or in Aristotle.

^{page 434 note 1} Cf. , unparal leled in Aristotle Metaph. A 983^b6 ff. (DK 11 A 12).

^{page 434 note 2} ‘Of small value’ say Kirk and Raven (Presocratic Philosophers, pp.5 f.) of Hippolytus' report of Thales.

^{page 434 note 3} West, M. L., Early Greek Philosophy and the Orient (1971), p.210, also refers to Thales Aetius 3. 15.9 . But we can hardly regard this as independent evidence for Thales when Aetius does not name him.Google Scholar

^{page 435 note 1} e.g. Kirk and Raven, op.cit., p.92; Lloyd, G. E. R., Polarity and Analogy (1966), p.308.Google Scholar Others have rejected it, e.g. Diels, H., Doxograpbi Graeci (1879), p.255.Google Scholar

^{page 435 note 2} Cf. Kirk and Raven, op.cit., p.85; Guthrie, History of Greek Philosophy 1.45.

^{page 435 note 3} The same applies to the statement in N.Q. 3. 14.1 that the earth's floating on water accounts for the supply of water to maintain the rivers.

^{page 435 note 4} This is Oltramare's text, which includes emendations; but 1 do not think there is doubt about the main point, that the Etesians cause the flood by holding up the Nile in its course.

^{page 435 note 5} A Latin summary of a Greek work attributed in antiquity to Aristotle (see Aristotelis fragmenta, ed. V. Rose (1886), frg.248).

^{page 435 note 6} For a brief account of these sources see Diels, Doxograpbi Graeci, pp. 226–9.

^{page 435 note 7} See Postl, B., Die Bedeutung des Nil in der romischen Literatur (1970), p.65.Google Scholar

^{page 435 note 8} See, e.g., Diels, op.cit., p.227.

^{page 436 note 1} The case is different with Thales' earthquake theory, for which Seneca, Aetius, and Hippolytus may well depend on a late Hellenistic source.

^{page 436 note 2} Posidonius, ed. by L. Edelstein and I. G. Kidd, vol. 1 (1972).

^{page 436 note 3} I should state, however, that I understand from Professor Kidd that he regards 2.55 as ‘of possible relevance’ for Posidonius, but not a direct report of him.