¹ Gercke, A., Seneca-Studien, Jahrb. f. class. Phil., Suppl. 22.1 (Leipzig, 1895; repr. Hildesheim, 1971)Google Scholar; Studia Annaeana (Greifswald, 1900)Google Scholar; L. Annaei Senecae Naturalium Quaestionum Libri VIII. Edidit Alfred Gercke (Bibliotheca Teubneriana, Leipzig, 1907; repr. Stuttgart, 1970).Google Scholar

² CQ 8 (1914), 273–4.Google Scholar

³ On conjecture in Δ or 8 see CQ N.S. 28 (1978), 299–302. Several of the readings Gercke adduced are in fact not unique to A anyway, but we may let that pass.Google Scholar

⁴ Garrod argued against δ's text at II. 12.1, but his objections can be countered; he accepted δ's text in the other three places, but in each case δ is suspect. See the commentary on each passage in my thesis.

⁵ Diels, H. and Kranz, W., Die Fragmente der Vorsokratiker⁶ (Zurich-Berlin, 1952), i. 87, line 29. Compare the phrases spiritus infirmior and acrioris deniorisque spiritus later in the chapter. (Mr. Colin Hardie first pointed out to me that Δ's reading begs the question.)Google Scholar

⁶ Oltramare, P., ‘Le codex Genevensis des Questions Naturelles de Sénequè’, RPhil 45 (1921), 5–44Google Scholar, and his edition, Sénèque, Questions Naturelles (Budé edn., Paris, 1929Google Scholar; repr. 1961). Earlier judgements on Z were based on inadequate knowledge of its readings: Gercke, , Studia Annaeana, pp. 21–7Google Scholar (where Z is called R); edn., pp. xxvii-xxviii; Garrod, , op. cit., p. 275Google Scholar; Geist, H., De L. Annaei Senecae Naturalium quaestionum codicibus (Diss. Erlangen, publ. Bamberg, 1914), pp. 18–21.Google Scholar

⁷ Hence, perhaps, his failure to convince so exacting a critic as Axelson, who at one time said that he intended to examine Z afresh at some future date, and for the time being was leaving the question of Z's status open (Axelson, B., Senecastudien, Kritische Bemerkungen zu Senecas Naturales Quaestiones, Lunds Universitets Arsskrift, N.F., Avd. 1, Bd. 29, Nr. 3 (Lund, 1933), pp. 3–4Google Scholar). Later he abandoned his intention (id., Neue Senecastudien, Textkritische Beiträge zu Senecas Epistulae morales, Lunds Universitets Årsskrift, N.F., Avd. 1, Bd. 36 (Lund, 1939), p. 235 n. 24).Google Scholar

⁸ Castiglioni, L., ‘De Naturalium quaestionum codice Veneto Marc. XII 141 (cod. lat. 1548)’, Miscellanea Giovanni Galbiati i, Fontes Ambrosiani 25 (Milan, 1951), pp. 183 ff., noted that Z often shares readings with UW (for the sigla, see below), which he thought might indicate a common ancestor, or else contamination from a manuscript like Z in UW. But all the pertinent readings of ZUW are correct, or at least possibly correct, so he ought to have entertained the possibility that Z and UW derive them independently from the archetype.Google Scholar

⁹ Vottero, D., ‘Problemi di critica del testo nelle “Naturales Quaestiones”. Nota I. L'ordinamento dei libri’, AAT 107 (1980), 249–69; his comments on Z are on pp. 264–7.Google Scholar

¹⁰ Cf. CQ N.S. 28 (1978), 310; below, passim.Google Scholar

¹¹ I am grateful to him for giving me this opinion, by letter.

¹² When an asterisk is prefixed to the first mention of a manuscript, it indicates that i have not seen the manuscript or microfilm or photographs of it, and am reliant on information published by others.

¹³ von Jan, L., ‘Neueste Literatur des Seneca’, NJPhP 31 (1841), 261–72Google Scholar; L. Annaei Senecae opera … recensuit … Carolus Rudolphus Fickert (3 vols., Leipzig, 1842–1845).Google Scholar

¹⁴ Haase, followed by Gercke, needlessly transposed this section of the text to form III.12.2.

¹⁵ Seneca-Studien, pp. 55–71Google Scholar (where θ is called μ); Studio Annaeana, pp. 8–12; edn., pp. xxix-xxx, xxxiv.Google Scholar

¹⁶ Castiglioni, loc. cit., discerned a connection between UW, even though he was working with Gercke's and Geist's meagre reports of U. I shall generally ignore corrections by a later hand in W when the original reading is not in doubt.

¹⁷ The correct text, to be found in all editions before Gercke's, is: sed primum omnium, quia inbecilli oculi ad sustinendut comminus solem ignoraturi erant formam eius, bebetato ilium lumine ostendit. Gercke accepted the omission of ilium, but against this see Axelson, Senecastudien, p. 31Google Scholar.

¹⁸ See Oltramare, , RPhil 45 (1921), 21m. 1; edn., apparatus ad loc.Google Scholar

¹⁹ S.'s usual word for ‘flood’ is diluuium (III.27.1, 27.14, 29.1, 30.4), but eluuies or diluuies is presumably possible here, (eluuiei is conjectured by Gercke at III.26.8.) Once I considered illuuies for uies illam, but Mr. Reynolds points out that the word is rarely used to mean ‘flood’; TbLL cites only lust. and Drac. Remoter possibilities for uies illam are dies ilia (or ille) (cf. III.27.1 fatalis diluuii dies (dies diluuii Ψ), or even pernicies illa (cf. III.28.2 illa pernicies, 29.2tanta pernicies).Google Scholar

²⁰ RPhil 45 (1921), 20.Google Scholar

²¹ The asyndeton is an additional point in f's favour.

²² Alexander, W. H., ‘Seneca's Naturales Quaestiones: The text Emended and Explained’, University of California Publications in Classical Philology 13.8 (1948), 328, approves Z's reading here.Google Scholar

²³ Axelson, , Senecastudien, p. 40; E. H. Warmington cited in Corcoran's Loeb edn. ad loc.Google Scholar

²⁴ RPhil 45 (1921), 35; his edition does iot mention Z's readings here.Google Scholar

²⁵ On the gender of dies here see Fraenkel, E., Glotta 8 (1916), 59Google Scholar (= Kl. Beitr. i. 62).Google Scholar

³⁶ It is also theoretically possible that π is contaminated from a, or vice versa; but on the former supposition, it would be surprising that none of a's more outrageous errors found their way into π against either supposition, the agreements of aπ are errors, and almost exclusively the trivial sort of errors one does not expect to be transmitted by contamination.

²⁷ nee ira numinum aut caelum conuerti aut terram. Editors are fairly evenly divided between concuti and conuerti. Conuertere can mean almost ‘to convulse, destroy’; (see OLD s.v.2; ThLL 4.866Google Scholar.45 ff.; Hall on Claud. DRP 1.64–5), but since it is also used of the revolution of the heavens (OLD s.v.1; ThLL 4.854Google Scholar.25 ff.), it is inapposite here with caelum. There is perhaps a parallel at Cic, . Rep. 1Google Scholar.56 qui [sc. Iuppiter] nutu, ut ait Homerus [e.g. 11. 1.530], totum Olympum conuerteret; but unless Cicero has misunderstood one should maybe consider concuteret there. Axelson, , Senecastudien, p. 79, makes a further point against conuerti at NQ VI.3.1, namely that it gives a poor clausula. So concuti is preferable.Google Scholar

²⁸ Asa comparison, Fδ are not related, but share three transpositions: VI.3.3 pars magna; VII.6.1 umida aridaque; 11.2 quidem fusus.

²⁹ This florilegium is contained in: B.L. Add. 16608 (s. XIV); Douai 285 (s. XU); Douai 533 (s. XII/XIII); Troyes 215 (s. XH/XIII); Vatican Reg. lat. 1707 (s. XIII).

³⁰ See particularly Gercke, , Seneca-Studien, pp. 8–33Google Scholar; Studia Annaeana, pp. 35–44Google Scholar; Geist, , op. cit., pp. 21–31Google Scholar; also Oltramare, edn., pp. xxix-xxx; Castiglioni, L., AFZC 18 (1951), 87, n. 2.Google Scholar

³¹ On Abr.'s text of II.13.2, see Geist, , op. cit., pp. 35–6; his account of Abr.'s affiliations (pp. 23–4) is erroneous. In III.25.6 – IVa Abr. generally, but not always, agrees with π: of the errors of π in this section listed on p. 189 above, Abr. has the omission at III.27.1 and the errors at 111.27.11, 29.6, IVa.pr.ll, 2.7, but not the omissions at III.26.8, 27.4, 28.6. The occasional agreements with p (e.g. III.25.9 ob] ab), v (IVa.pr.20 samus] simus), U (e.g. IVa.pr.l tecum] iterum) and W (III.30.6 committent] commiscent) seem to show that Abr. is not a sincere, independent descendant of π, but contaminated. It is scarcely worth pursuing its precise affiliations and the source(s) of contamination; I have collated the whole section in question and nothing clear emerges.Google Scholar

³² On the asterisk see n. 12 above.

³³ Foerster, O., Handschriftliche Untersuchungen zu Senekas Epistulae Morales und Naturales Quaestiones, Würzburger Studien zur Altertumswissenschaft 10 (Stuttgart, 1936), pp. 46–56.Google Scholar

³⁴ Occasional separative errors show that neither B nor C is derived from the other (e.g. B omits II.12.5 haec; IVb.11.2 totum; C omits II. 1.1 caelum;quemadmodum). The status of Prag. is uncertain: Müller, G., De L. Annaei Senecae quaestioni-bus naturalibus (Diss. Bonn, 1886), pp. 5–9Google Scholar, tried to show that B and Prag. are twins; but Gercke, , Studia Annaeana, pp. 40–1Google Scholar, thought Prag. might be a copy of B, and none of Miiller's evidence rules this out. Rossbach, O., BPhW 27 (1907)Google Scholar, 1480, after re-examining Prag., declared it to be neither a copy of B nor a twin; one may trust his statement that it is not a copy, but one wonders what precisely he means by twin. Nothing is known of the relation between Prag. and C, but the similarity of the contents of B and Prag. suggests that they are closer to each other than to C (for they both contain the Asclepius and the Latin translation of Nemesius, besides the Natural Questions).

³⁵ In this manuscript the text of the missing portion of Books IH–IVa (taken from an η source, see next note) is tacked on to the end, after Book VII.

³⁶ η is a descendant of θ, and ancestor of DE and others; see § 3.2 below.

³⁷ This manuscript forms a close-knit group with Paris lat. 17911 and Vatican Pal. lat. 1541, all three having a similar text, and the same marginal notes. Cf. Hijmans, B. L., Mnem. Ser. IV. Vol. 21 (1968), 240–53CrossRefGoogle Scholar; Fohlen, J., Scriptorium 29 (1975), 62–6.CrossRefGoogle Scholar

³⁶ I am grateful to Dr. C. B. R. Pelling for inspecting this manuscript in the Escorial for me.

³⁹ The contents of Tours 693 are identical with those of Paris lat. 8717, except that the end of the Paris MS. is lost, so that one can merely assume it originally contained the last couple of items found in the Tours MS. Some of the contents are unusual, particularly the grouping of the first few items: Iuncta de S. Germano, Tabula alphabetica ex operibus Senecae; Albertinus Mussatus, Argumenta tragoediarum Senecae; Mart. 5.42; Seneca Apoc.;Anth. Lat. 660.

So it is a reasonable assumption that the Tours MS. is a twin or even a copy of the Paris MS., in which case the former may be safely assumed to have a text like the latter, which I have seen.

⁴⁰ The first three excerpts are: I.pr.1–pr.2 Quantum … unde lucet; 11.56.1–2 tonitrua nos … uti; III.25.12 Quis non grauissimas … creditur.

⁴¹ Formerly Vienna Rossianus IX.249; cf. Geist, , op. cit., pp. 29–31.Google Scholar

⁴² See references in n. 15, and Geist, , op. cit., pp. 9–16.Google Scholar

⁴³ 1 am grateful to the Provost and Fellows of Eton for transferring this manu- script to the Bodleian for me.

⁴⁴ e.g. the corrector of J (Oxford, St. John's College 36 (s. XIII)), using θ MS., expunges III. 30.6 pluribus locis and IVa. pr. 17 potentium.

⁴⁵ E was collated by I. F. Herel (whose collation appears in Fickert's edition as ‘e'), and by Fickert; recollated by Larisch (I–II.26), Diels (IVa.1–2), G. Müller (III–VII), and Gercke (11.27–59); and again by Oltramare. I have recollated parts of E.

⁴⁶ This manuscript is in two parts. The first, dated A.D. 1379, contains excerpts from the Natural Questions on fos. 32^v–36^v, and a complete text on fos. 39–127; the text of the excerpts is very similar to the full text (cf. Geist, , op. cit., pp. 11–14) and will be ignored here. The second part of the manuscript, dated A.D. 1378, contains a couple of excerpts from the Natural Questions on fo. 169^v: 1.17.4 Inuenta sunt … ipsos uidendi, and 1.6.5. (The former excerpt also occurs e.g. in Paris lat. 6395A (see p. 213), and in a paraphrase, falsely attributed to Augustine, in Schlägl 85 (s. XIV).)Google Scholar

⁴⁷ Diels, H., Seneca und Lucan, Abb. d. K. Akad. d. Wiss. zu Berlin, 1885Google Scholar, Phil.- hist. Abh. 3 (Berlin, 1886), p. 34; Müller, G., op. cit. (n. 34), pp. 5–13.Google Scholar

⁴⁸ The contents of Paris lat. 6387 are: Seneca, Natural Questions; Walter Map, Dissuasiones Valerii ad Rufinum ne ducat uxorem; anon., Expositio epistulae Valerii ad Rufinum; Mart. 10.47, 1.33; Jerome, Contra Iouinianunt; anon., Dicta Platonis; Dicta Aristotelis. These works, preceded anc followed by others, all occur in the first part of Cam., with one difference of order, namely that the Walter Map work comes after the poems of Martial, thus being separated from the Expositio. So if the Paris MS. is a copy of Cam., the scribe has rationalized the order by putting the Dissuasiones before the Expositio.

⁴⁹ The lines of Cam. omitted at first are IVb.13.10–11 esse putas … uenarum nee and II.3 5.1 ergo expiationes … sunt fata.

⁵⁰ My η is probably an ancestor of Gercke's η, which is the immediate common ancestor of the vulgares, my δ-η-η MSS. See Gercke, , Seneca-Studien, pp. 71–6Google Scholar; Studia Annaeana, pp. 8–9; edn., pp. xxix-xxx, xxxvi-xxxvii.Google Scholar

⁵¹ Gercke, , Seneca-Studien, pp. 59–60 (where F and H are called N and Q, and θ is called M).Google Scholar

⁵² The crucial folio, containing the end of Book IVa, is missing from m, but calculation shows that m probably had the same lacuna. Taking half a dozen surviving folios of m, I have counted that they contain between 175 and 202 lines of Teubner text.

The missing folio (between fos. 146 and 147) contained IVa.pr.22 -genti spectaculo to IVb.3.4 rotunda fit, 268 Teubner lines, far more than the other folios. But if m had the same lacuna as Hprv, it contained 186 Teubner lines, which is just right.

⁵³ The alternative, that DE are sincere representatives of η at the end of Book IVa, that the ancestor of Hmprv was also a copy of η), and that the errors of 17 were removed from H by contamination, is not worth considering, in view of the purity of H's text, and the amount of corruption and conjecture in η.

⁵⁴ There is a further possible link between the contents of Paris 6286 and P, for the former also contains, inter alia, Apuleius, De deo Socratis, and De Platonis dogmate, and the Asclepius. These works, with others, appear in Paris lat. 6634, which together with Vatican Reg. lat. 1107 and 1438, and P, at one time formed MS. HHH 22 in the library of St. Victor at Paris. (See Delisle, , Cabinet, ii. 231–2Google Scholar; Pellegrin, E., BEC 103 (1942), 79 and 95.) So Paris 6286 may derive from the time when Paris 6634 was still bound up with P. (I am grateful to Prof. R. H. Rouse for bringing this information about P's origins to my notice.)CrossRefGoogle Scholar

⁵⁵ Paris 6286 and P have the following: Book IVb, L. Annei Senece de naturalibus questionibus. Liber incipit: de grandine et niue; Book V, Liber secundus incipit de uentis; Book VI Liber tertius incipit de terremotu; Book VII, Liber quartus incipit de cometis; the remaining books have no title or incipit. Reims 872 has the same for Books IVb and VII; for V, Liber secundus de uentis; for VI, Liber tertius Lucilio de terre motu.

⁵⁶ The excerpts begin with: I.pr.6 Auaritia quicquid omnibus abstulit sibi ipsi negat; pr. 13 Deus est mens uniuersi. Deus est totum quicquid uides; pr.14 Nostri melior pars animus.

⁵⁷ Gercke, , Seneca-Studien, pp. 41–8;Google ScholarStudia Annaeana, pp. 13–9; edn., pp. xxviii– xxix, xxxiv–xxxv.Google Scholar

⁵⁸ The excerpts begin: Seneca in libra de naturalibus quaestionibus. De contemptu transitoriae uitae et omnium temporalium. Humano animo a bonis artibus robur uenit et contemplatione naturae (cf. VI.32.1). ‘Quid est enim cur… obuiam exire.’ (VI.32.2–12) Contra humanam curiositatem beneficiis naturae abutentem. ‘Derideantur … factum est.' (1.17) De diuersis bonorum morum instructionibus. ‘Inter nullos magis quam inter pbilosopbos esse debet aequa libertas'. (IVb.3.6).

⁵⁹ I am grateful to Dr. B. C. Barker-Benfield for drawing this manuscript to my attention.

⁶⁰ e.g. at the start of Book II there are the following λ readings: II.2.4 sua] sola; 6.3 est uox] inu. ord.; 11.1 sic diuisus] inu. ord.; 12.5 eodem] autem add.; 26.5 usti montis] inu. ord. F λ But at II.6.4, uim et uelocitatem neruis betrays contamination from 6. A similar picture emerges from other sample collations.

⁶¹ The excerpts, with Grandinem book order, start with: IVb.3.3–4 Quare autem … cauantur aqua; 3.6 Quare non … grandinem interesse; 5.1 Pauca enim … uincunt litigant.

⁶² The excerpts begin with: III.30.8 Cito nequitia … discuntur; IVb.4.3 Bruma lentas pluuias babet et tenues; 7.1 Tempes- tates munusculis redimuntur. Examples of λ errors: III.30.8 uirtus difficilis] inu. ord.; 1.16.3 fecisse se] sibi fecisse. Apart from the first excerpt, the Grandinem book order is followed, although the excerpts get no further than Book I.

⁶³ Seneca might have used the nom. sing, form ualles; but uallis occurs at Tro. 1124, Pha. 1133, and in E at Tro. 178 (tellus A).

⁶⁴ Gercke, edn., p. xl (misprinted as lx); cf. Seneca-Studien, , pp. 77–8.Google Scholar

⁶⁵ fuere is wrong. S. generally uses -re for -runt in perfects to improve the clausula, and here it spoils it (-i fuerunt uiri); see Axelson, , Senecastudien, p. 88.Google Scholar

⁶⁶ sua ought never to have been introduced into the text by Gercke.

⁶⁷ At the end of Book IVa the only manuscripts which count are ZPRUW. Gercke gives a very distorted view of the tradition at this point, for he treats all Z's true readings as conjectures.

⁶⁸ Borghese 188, fo. 77, has, amongst excerpts from other works, the quotations from Ovid and Lucretius at IVb.3.4, and II.31.1–2 Ceterum mira … omne consumitur. The only tenuous clue to origins is II.31.2 adnotandum (= πΛ).

⁶⁹ The excerpts, in traditional book order, begin: I.pr.5 O quam contenta res est homo non supra humana surrexerit; pr. 6 Intus (Virtus Ω) ista quam … cognitionem caelestium; pr.11 Punctum est … regnatis. There are a number of δ errors, but it is not clear whether the manuscript derives from δ, or is just contaminated from it.

⁷⁰ Haase, F., ed., L. Annaei Senecae opera (3 vols., Leipzig, 1852), II, p. iv.Google Scholar

⁷¹ In the margin of his copy of Seneca; see Fickert, , op. cit. (n. 13) II, p. xviii.Google Scholar

⁷² Seneca-Studien, pp. 48Google Scholar9; Studia Annaeana, p. 26; edn., p. xxxv.Google Scholar

¹³ See pp. 209 ff.; Gercke, , edn., p. xxxvii.Google Scholar

⁷⁴ See p. 204; Fickert, C. R., Prolegomena in nouam operum L. Annaei Senecae philosophi editionem, I (Leipzig, 1839), 38–9Google Scholar, 50; Gercke, , Seneca-Studien, p. 18.Google Scholar

⁷⁵ See p. 212; Gercke, , Seneca-Studien, p. 37.Google Scholar

⁷⁶ Gercke, , Seneca-Studien, p. 57.Google Scholar

⁷⁷ See Mynors, R. A. B., Catalogue of the manuscripts of Balliol College Oxford (Oxford, 1963), p. 108.Google Scholar

⁷⁸ They are: those cited by Fortunatus and Erasmus; Casaubon's MS., cited by Gruter; the Chiflerianus, Troncianus, and others of Dalecampius. Fortunatus does not identify individual manuscripts, so it is impossible to trace them. Erasmus' Britannicus shares some readings with D, but in the case of this and the others, too few readings are cited for a reliable investigation of their affiliations. Note that the identification of the Nicotianus of Opsopoeus with an extant Nicotianus, namely A (see Gercke, , Seneca-Studien, p. 11Google Scholar), may be a mistake, for one of the few of its readings which are recorded does not tally with A: 11.18 crassum] cussum A: quassum Nicot. If different from A, the lost Nicotianus was not much different.