² ‘R.’ was at first used by Gercke, A., Studia Annaeana (Greifswald, 1900), pp. 21Google Scholar ff., to denote Geneva lat. 77 (12th cent.), but since then the Geneva MS. has been known as Z.

³ For fuller descriptions see: Lowe, G. and von Hanel, W., SB Wien, Phil.-hist. Cl. 111 (1885), 512–13;Google ScholarAntolin, G., Catalogo de los codices latinos de la Real Biblioteca. del Escorial, 3 (Madrid, 1913), 225–7.Google Scholar

⁴ The De dissentione etc., is in PL 171, cols. 989 ff., under the title Liber de querimonia et conflictu carnis et spiritus seu anitnae. In R this is followed by one or more poems of Hildebert, for the explicit (‘Angelus exultat homo gaudet tartara merent’) given by Hartel, Löwe-von and Antolìnas though it comes from the De dissentione is in fact the last line ofGoogle ScholarHildebert carm. min. 40 (ed. A. B. Scott); I am grateful to Mr. A. B. E. Hood for helping me to track this downGoogle Scholar

⁵ R has the shorter version; see Boutemy, A., Le Moyen Age 52 (1946), 243–56.Google Scholar

⁶ Text in Hartel, Lowe-von, op. cit., p. 513.Google Scholar

⁷ Text in Becker, G., Catalogi Bibliothecarum Antiqui (Bonn, 1885), p. 202;Google ScholarCatalogue général des manuscrits des bibliothèques publiques de France: Dèpartentents, Octavo series Vol. 2 (Paris, 1888), p. 398 (ed. Omont, H.).Google Scholar

⁸ Gercke, A., Seneca-Studien, Jahrb. f class. Phil., Suppl. 22.1 (Leipzig, 1895; repr. Hildesheim, 1971), p. 37; id., Studia Annaeana, p. 20; id., L. Annaei Senecae Naturalium Quaestionum Libri VIII. Edidit Alfred Gercke (Bibliotheca Teubneriana, Leipzig, 1907; repr. Stuttgart, 1970), p. xxxv.Google Scholar

⁹ On the true relationship of ST to P, see pp. 307–10 below.

¹⁰ See Muller, G., De L. Annaei Senecae quaestionibus naturalibus (Diss. Bonn,1886), pp. 12–13;Google ScholarGercke, , Seneca-Studien, pp. 37, 40–1, 49.Google Scholar

¹¹ I have also collated the whole of ADFGHPSTUVWZ, parts of E, and all or parts of certain other manuscripts. FW have been collated from the manuscripts themselves, the remainder from microfilm. I have been unable to obtain microfilm of B, so for its readings I am reliant on the reports of Fickert, Gercke, and Oltramare; and for λ I generally rely on Gercke. My sigla are the same as those of the editions of Gercke and Oltramare, with the following additions: D = Dublin, Trinity College 514 (13th cent.); W = Venice, Biblioteca Marciana, Lat. Z.268 (1548) (14th cent.).

¹² I follow Oltramare, P. in his edition (Sénequè, Questions naturelles (Budé edn., Paris, 1929; repr. 1961), pp.xii ff.), in calling this the Grandinem book order, after the first word of Book IVb. The order of our editions (I-IVa, IVb-VII) I call the traditional order.Google Scholar

¹³ The scribe of R omitted uerior at first, but then inserted it in the margin.

¹⁴ I have excluded from the above evidence those errors of PR which are also in STX, contaminated relatives of R (see below, pp. 307–10). If such errors are included, the evidence of R's relationship to P becomes clearer: e.g. PR and t (as defined on p.307 below) share the following transpositions in Book V: 4.2 exonerat cum sono; 10.2 et niues ponuntur (ponuntur et niues P²); 16.2 potest modo.

¹⁵ PR agree in error over 300 times, if errors also found in STX are included; and there are also numerous errors shared by R with STX which must have eluded the supposed contaminator.

¹⁶ The conclusions of the examination of the other sections of R must necessarily be anticipated here. The comparison with STX, relatives of R which are contaminated, is instructive (cf. pp. 303, 307–10 below).

¹⁷ A= Leiden Voss. lat. 0.55 (12th cent.); B = Bamberg IV.16 (12th cent.); V = Vatican Pal. lat. 1579 (12th cent.).

¹⁸ He says merely (op. cit., p. xxx), ‘J’ collationnè moi-meme … A, B, V, qui sont les mèilleurs temoins de δ.’

¹⁹ Cf. Gercke's, edition, pp. xxxi-xxxin, xxxviii.Google Scholar

²⁰ e.g. the following transpositions: I.pr.2 caliginem in qua uolutamur; 3.7 hebes et in firma est acies;11.21.4 Fulmen est quiddam plus quam fulguratio. Vertamus istud: fulguratio est paene fulmen; IVb.3.1 Grandinem hoc modo fieri quo apud nos fit glacies gelata nube tota si tibi affirmauero; 7.1 uim quandam potentem auertendae nubis ac repellendae esse in ipso sanguine. For more detail see my thesis, pp. 108–14.

²¹ Gercke, to judge from his apparatus ad loc., took a different view of the course of the corruption in first uarietatem was altered to repraesentationem in δ; then a scribe jumped from repraesentationem to repraesentat, producing δ's text. This view is implausible, for it gives no motive or explanation for the original alteration of uarietatem. The view taken above is confirmed by δ's text; see below.

²² Geist, H., De L. Annaei Senecae Naturalium quaestionum codicibus (Diss. Erlangen, publ. Bamberg, 1914), pp. 25–7.Google Scholar

²³ Castiglioni, L., ‘Contributo alla storia del testo delle Naturales Quaestiones di Seneca’, AFLC 18 (1951), 81–102; his suggestion is on p. 87, n. 2.Google Scholar

²⁴ Again I exclude errors also found in STX; cf. above, n. 14.

²⁵ Of course there may well have been various lost intermediaries. One might consider the hypothesis of a fluid exemplar, i.e. that δ was not a copy of a, but was a itself, with corrections and conjectures added (after an ancestor of R_a was copied, presumably). Some of the errors of ABV are neatly explained by this hypothesis, but the numerous transpositions found in ABV but not in Ra are only explained adequately on the assumption that.5 was a copy of a. For examples of such transpositions cf. those from Book I in the list on p. 299.

²⁶ The omissions are: 1.1.6 sic quanto … leuiora fulmina (after fulmina); 2.9 temperatum … intercisa est apparet (after apparet); 3.1–2 uarietatem. Altera … faciem arcus (after arcus); 3.6 specula sunt totidem (after totidem); 5.2 figura sobs neque (after neque); 5.8 ortu sobs … rubere uidemus (after uidemus); 5.11 ab eo a quo … quamuis non uideatur est (after uidetur); ita non est ergo et color (after ergo et color (ergo color et AB)); 6.3 conuenit colones … unum inter nos (before non conuenit); 11.55.3 sed cum hoc fieret tonitrua facto sunt (afterfacta sunt); 59.3 nos posse. Quemadmodum … contempta sunt; V11.23.1 nescio duo nostra aetate fecerunt (after fecerunt); 25.2 concentuni quendam … animae alius (after alius).

²⁷ I ignore the fact that in a few cases PR's reading is found in one or two other manuscripts (not, of course, in A, B, or V); though cf. n. 28 below.

²⁸ et per is also in TX and in E; in the latter perhaps by chance coincidence, or perhaps because of contamination in E from a text like STX's; cf. my thesis, pp. 52–8.

²⁹ Oltramare, , RPhil 45 (1921), 43–4, in another connection, uses to denote an ancestor of Z.Google Scholar

³⁰ We might have expected the three manubiae to be in simple ascending or descending order at 44.2 (as at 41.1–2), but, they are not. If this is thought unacceptable, one might conjecture that the three phrases, all beginning with quaedam, have somehow been transposed in the tradition. But the apparently anomalous order could have an ethical intention: Seneca (perhaps with halfan eye on Nero) may be deliberately playing down the use of severe punishment by sandwiching it between light punishment and mere warning; cf. his pointed exhortation (43.2) to take advice before punishing.

³¹ For a demonstration that the manuscripts other than derive from a common ancestor see my thesis, pp. 70–81. This will be fully dealt with in another article.

³² The differences are as follows (R = first text, Rf = footnote text): densae] depressae Rf; eiusmodi] quae eiusmodi Rf; quiet] qua R;pereussu] percursu R: percussa Rf; resplendent ] -et Rf; fueritl fiunt R; propior] prior R; remittet] -it Rf; efficiet] -it Rf.

³³ On ST, see Gercke, , Studia Annaeana, pp. 20–7, and his edition, pp. xxvii, xxxv.Google Scholar On X, see Claxton, M. T., ‘The place of Exeter Cathedral Manuscript 3549(B) in the textual tradition of Seneca's Naturales Quaestiones’ (University of Exeter M.A. Thesis, 1960, unpublished).Google Scholar, Claxton, op. cit., pp. 35 ff., tackles the problem of the precise relations of STX to each other and their common ancestor; but she is hampered by the incompleteness of Gercke's and Oltramare's reports of ST, on which she relies.Google Scholar

³⁴ One may ask why T has this lacuna and X does not. There are two main possibilities. Either T represents the state of the common ancestor of TX, and X has undergone subsequent contamination which restored the missing portion; or X represents the common ancestor, and T's lacuna is perhaps to be explained on the hypothesis of a fluid exemplar, T (or an intermediary) having been copied after some folios were lost. (This would have been very easy if some folios containing the missing portion had been stuck in to the common ancestor.) In the text above I assume that X represents the common ancestor, but even if this is not so, there is sufficient evidence in T itself to prove that the central portion of t comes from ŋ.

³⁵ f = Escorial N III 16 (13th cent.), twin of G; see Geist, , op. cit., pp. 9–10; my thesis, pp. 118–19.Google Scholar

³⁶ In general, where one has three manuscripts, ABC, of which AB and BC share significant errors, while AC do not, then there must, on the simplest possible account, be contamination in either A or C. Cf. Willis, J., Latin Textual Criticism, Illinois Studies in Language and Literature, 61 (Illinois, 1972), 17 ff.Google Scholar

³⁷ All editions have imitationem, but the sense requires mutationem.

³⁸ θ here denotes the consensus of DEFGH.

³⁹ In inmensum procedam. In inmensum is the normal expression (see ThLL 7.1.453.56–65), so one would expect it here; but for the sake of the clausula Seneca might have used in inmensa. Cf. how at V. 17.3 he uses in transuersa, which is otherwise poetic (see Lewis, and Short s. transuerto II A 2). Even if inmensa is correct, it is another case of t siding with θ.Google Scholar

⁴⁰ By ŋ I denote the common ancestor of DE, which was probably an ancestor of Gercke's ‘ŋ’; see my thesis, pp. 51 ff.

⁴¹ The correct text is ut et nauiget; see Axelson, B., Senecastudien, Kritische Bemerkungen zu Senecas Naturales Quaestiones, Lunds Universitets Arsskrift, N. F. 29. 1. 3 (Lund, 1933), 80 n. 10.Google Scholar

⁴² The question of contamination of t from ŋ is complicated by the fact that there is a possibility of contamination in the other direction as well (cf. n. 28 above), whichmakes it hard to tell in which direction the contamination has gone in a particular instance. But the presence of characteristic θ errors in t, and the origin of , makes it plain enough that there has been contamination from .

⁴³ The excerpts, in Grandinem book order, begin with: V.15.4 Nulli [UM ergo uulg.] mortuo … infodit; VI.29.1–2 Quid mirum est animos. non desipere Iconsipere uulg.I; I.pr.5 Quid est cur suscipiamus [suspic- uulg.] … ualitudinario.

⁴⁴ The excerpts, in Grandinem book order, begin with: V.18.6 Quae nos dementia … sepulturae; 18.9 Miseri quid … malum; 15.3.Quae tanta … possidendum quam quaerquam [quaerquam poss- uulg.].