STOP Early Journal Content on JSTOR, Free to Anyone in the World This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in the world by JSTOR. Known as the Early Journal Content, this set of works include research articles, news, letters, and other writings published in more than 200 of the oldest leading academic journals. The works date from the mid-seventeenth to the early twentieth centuries. We encourage people to read and share the Early Journal Content openly and to tell others that this resource exists. People may post this content online or redistribute in any way for non-commercial purposes. Read more about Early Journal Content at http://about.jstor.org/participate-jstor/individuals/early- journal-content . JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people discover, use, and build upon a wide range of content through a powerful research and teaching platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please contact support@jstor.org. TEXTUAL PROBLEMS IN ARISTOTLE'S METEOROLOGY By F. H. Fobbs In an article published in the Classical Review for December, 1913, I unsuccessfully attempted, on the basis of an examination of the first book only, to answer the question whether the MSS of Aristotle's Meteorology fall into two groups or into three — a question important because of Bekker's and Ideler's close dependence upon Parisinus 1853 (E). 1 Since writing that article, I have been able to revise from the MSS themselves three collations which I had first made from rotographs — Vaticanus 1027 (H), Vaticanus 258 (N), and Parisinus suppl. grec 314 (Hayduck's W) — and to collate, also from the MSS themselves, books ii-iv of the four MSS already mentioned, 1 Close as Bekker's dependence is, one is scarcely warranted in saying with Thurot (Revue Archeologique, XX [1869, 2], p. 420) that he followed E wherever the evidence did not prevent. See : 339al9 tyyvT&ru] Bek. tyyirara E 314 Id. 339611 twv re] Bek. T«wEId. 339626 ti] Bek. re 314 om. E Id. 341al5 t6] Bek. om. E Id. (Cf. 347627, where the MSS are reversed: Sepphv] Ti BeppbvE Bek. Id.!) 342a33 ffXiox xai ffeXiixTjx] Bek. ^Xtos xoi r)s 8' o&rrjs [358a21], &airep elwonev, rijs re &tji«6w5oi;s avadv/uaaeus Kal rrjs ^rjpas, orav avvunriTai eis vify-q Kal vhwp, avay- MS has the form Trpos which Bekker wrongly attributes to F); Parisinus Coislin. 166: ircpl /xiv oSv [338a20] .... Ttpdrov [348634] (the catalogue wrongly implies that this first section runs through book iii) ; book iv (twice). Ambrosianus R. 119. sup.: irepl piv otv [338a20] .... Sepivats [343ol5], rb Si [368a33] . . . . t6 ISSap [384621]; Vaticani 115; 313; Vaticanus 319: book iv; Vati- canus Ottobon. 76; Escorialensis ¥-1-11: 8ti piv [350622] . . . . iv iroWQi [351625] (this eleventh-century fragment I have collated from rotographs) ; Bernensis (Bibl. Bongars.) 579, 16. Textual Problems in Aristotle's "Meteorology" 191 Katov lnirtpChapfiaveaivbp£vos tS>p ivoTrrpav, Kai p-qbtpiav exoov aiadrjriip haipeaiv, irpds Se 777 yfi juaXXcw Sid rb vrjvepMrepov elvat,' irvevp-aros yap ovtos otiic elvat, araaiv 4>avepbv. itapa, Se tovto pihaiva 17 ixonevq 7repic6epeia, Sid Tr/v heivrjs \evKbTiyra boKovaa elvat jueXcwTepa. ir\eopaias Se yiyvoprat, at aXcj> irepi ri\v aekqvqv Sid rb rbv r\kiov depporepov 6vra BcLttop SiaXveip rds owrdcreis tov atpos. Thurot (op. cit., XX, 418) remarks that Alexander, instead of suspecting a transposition of 21-25, contents himself with saying that Aristotle Sid piaov .... e'ur&v .... aaacpearepav ewoi-qcxe rfiv Xe&.p (146, 3-5). An examination of Alexander's words shows that the lines of Aristotle criticized are not 21-25 but 23-25; and I suspect that Thurot's "21" may be a misprint for "23"; if so, it is undeniable that 23-25 (irpds Se t% yrj . . . . avtpbv) come more 192 F. H. Fobbs naturally after p,e\avripa (27), where Vicomercatus and Ideler put them; but it is just as undeniable that this transposition is not indispensable to the sense. If Thurot really contemplated trans- posing 21-25, his suggestion is not likely to find many supporters. Whichever he intended, the instance is inconclusive. The general rule (383618-19) that objects which contain more water than they contain earth (e.g., milk) are thickened by fire, those which contain more earth than they contain water (e.g., clay) are coagulated is not easy to apply to oil: if in oil there is more water than earth, we should expect oil to be coagulated by cold; if more earth than water, we should expect it to be coagulated by fire; as a matter of fact, it is coagulated by neither, but is thickened by both (but cf. 38767 and 388a31-32). iraxwerai piv ovv [383632] vir' a/Mt>OTipuv, ov i-ripaiverat. d' inr' ovberipov' ovre yap b tjAios ovrt rb \fsvxos Iftpaivet' oil pbvov hum yKiaxP° v > &AX& ical Stmt, aipos loriv. ov ZqpaivtTat. Serb vSap ovS' bf/erai inrb irvpos, ori ovk arpl^et. Sia 7X1- ffxP 0Tr l' ra - I Q pl ace OI vSup, F H N (supported by the Vetus Versio) read tEhaiov; vdup, the lectio difficilior, found in all the other MSS (supported by the Nova Versio), stood without doubt in the arche- type; and was explained by Alexander as to kv t ^Xaico vScop (so also Vicomercatus and Barthelemy S.-Hilaire). Yet Ideler and Thurot (op. cit., XX, 418-19; XXI, 401) find in vSup an insuperable diffi- culty, and read $hawv; though Thurot adds quite rightly that the presence of eXcucw in our MSS is due to correction. The difficulty of taking vScop as to iv r<£ iKaJUfi v8ap seems to me lessened by the con- text and contents of the sentence in which it occurs; and even if we had not oi>d' ^erai 1 and Std yKio-xpoTtjTa, I believe that vSosp could be defended; as Capelle has said (Hermes, XL VII [1912], 517), Aristotle does not always express himself correctly; his words are often open to misunderstanding. Cumbustible things are either inflammable or non-inflammable: &m Se \oyi.o-Ta 6X«toi. All the other MSS, including 314, have hf/erai. Textual Problems in Aristotle's "Meteorology" 193 vypbv Kal Si' o\ov avvexh eanv, & avvexh S£ Kal '{Karrov rj Sore \oya iroiijaai' 6 8e TafJavurds tj? piv ofrrws tj; 5' eKeivws ex«. These last four lines, which Alexander does not paraphrase, Thurot (op. tit., XXI, 404) is tempted to delete; he does not understand "how the inflammability or non-inflammability of a body can depend upon the distribution of the aqueous element, and how copper has not enough of the aqueous element to make it inflammable," and in support of his doubt he appeals to Aristotle's definition of inflammables. If he had only gone a little farther back, to the definition of combustibles (387 avco t6tt<>), not to be found in our text; yet these words and also the next lemma — irepl pev ovv tov avco tottov — must, at least in his MS, have stood between avruv and ■Kf.pl (341a9). And at 346a9 between hpnrifav and the following Kal there must, in Olympiodorus' MS, have stood the words dvdyKti to'ivvv tcov avrSsv peyio~Twv kvk\o)p judXio-ra rf/v pekXovaav tovto iroiiiopav (77, 3-4) and XPV 7<*P tovto, tva ToXkri Kivqcns fj Sid t6 piyedos yi.yvop.ivt] Kal irkeiova tt)v e^a\f/i.v irovf)o-Q (77, 7-8) ; but in this case, unlike the first, the lacuna is not equally great in all the MSS (see Stiive's apparatus): whereas E 314 JFi have merely eKpiTi^eiv, Fcorr. m. i Matrit. Ambros. H N have iKpin^eiv aipa t« /cat diaicplveiv Sid t6 tov kvk\ov peyedos (re om. F Matrit. Ambros. CO rr. H N; SiaKpi- veiv td Ambros.i Siaicplvet. did Ambros. cor r. m. 1 SiaKpiveiv Sid Ambros.rec.; ptyedos F t peyedos deva [sic] F cor r. m. i)- That the lemma xPV 7<*P tovto, etc., was fully believed by Olympiodorus to be a part of the text is shown by his clear labeling of the four eri- xeipij/xara; but the fact that there is no trace of these lemmata in Alexander and Philoponus makes it unsafe to lay much weight upon their presence in Olympiodorus. 1 II The relationship of the nine chief descendants of this lost arche- type has been in part determined : 2 H and N have been found very closely related — more closely than appears from the apparatus of Bekker, whose collation of H was apparently less careful than that of N, and who consequently out of some 260 errors peculiar to H N has attributed four to H alone 3 and 44 to N alone; 4 Laurentianus 1 On Olympiodorus' MS or MSS see Susemihl, Bhein. Mus., XL (1885), 564, note 3. s Classical Review, XXVII, 252. s 340621; 372all; 378al9; 389620. * Title iteredpuv; 339al5; 340al0; 343619; 344al6; 344a34; 34561; 346a36 346636 ("35" in Bekker's apparatus, through a misprint) ; 349611; 349614; 350616 351a36; 354a6; 355a34; 35565; 357al6; 358al0; 358al5; 359al5; 35968; 361620 362all; 363al5; 363633 om; 364a32; 365627; 366a25; 368a3; 368a7; 371al6 372al8; 376a26; 376a31; 37661; 376622; 378a2; 379613; 380al5; 381a7; 382a3 383623; 387al7; 390a22. The error as regards the title is shared by Brandis, who has misled Ideler and Capelle. Capelle writes (Hermes, XL VIII [1913], 330, note 3) : "In den Bekkerschen Hss. lautet die Uberschrift des 1. Buches M€T€wpo\oycKav & (Der Angabe Bekkers, dass in N der Titel MeTeiipuv laute, widerspricht die Idelers Textual Problems in Aristotle's "Meteorology" 195 81.1 (S) I now disregard, believing it to be a descendant of the Ambrosianus — at least in book i, for which alone I have collated S; S's occasional freedom from errors of the Ambrosianus, as at 33%22 f 35, 345al3, is probably an incidental result of contamination from E and Parisinus suppl. 314 (Classical Review, XXVII, 252, n. 1). E and 314 1 have been seen to stand together; and so also J and F. I praef. p. xiv, der nach Brandis mitteilt dass in N die Uberschrift heisse ix rusv fUTeapoKoyiKur f}i/}\lov [Brandis has |8i/SX.]d.) Die anderen Bucher haben keine Uberschrift; eine subscriptio hat iiberhaupt keins." The facts are:_ (1) book i is headed in E aptaroTekovs tpCKoffotpov \ jierewpoKoicwv [sic] tujv eis 8 \ to a; in F tA _ov dpioTOTeXous nereupoXoytKuv, a; in H apurro fieredpuiv a (Bekker implies that H has p.ereupo\oyLK&v) ; in N dpurrorfKovs fiereiipuy-.-a; (2) in N the title of book ii is: A ipurroT^Xovs neredpuv; f); of book iii: dpicrTore' lureibpuv, y (this is in red ink, and A appears to have been added afterward, though by the first hand) ; of book iv; dpurrore' tierei&puv, S (the d- is in red, the rest in black) ; (3) N has a subscription: tAos tuv S flifSXlwv tuv | pxTewpwv I rod ipuTToriXovs. 1 Parisinus supplement grec 314, s. XIV (so Omont; ss. XII-XIII according to an early cataloguer's insert) is bombycinus; binding calf, with heavy wooden boards. The first three gatherings are lost; of the 35 gatherings which remain, all are quater- nions except ty' (ff. 94-99) and Xt} (ff. 292-95). The signatures, standing at the bottom of the first rectos, have in some cases been cut off by the binder's knife. The folia were correctly numbered by an old hand (22-100, 1-100, 1-92; these figures have been made consecutive by a modern hand) ; and some of them have since been wholly or in part torn out — 41, 130-32, 156-59, 166, 293-95. The second part of the first feuille-de- garde and the first part of the second are likewise lost. The pages are 247-48 mm. high, 168-73 mm. wide; each contains one column of text 185-90 mm. high and 112- 20 mm. wide. In ff. 22-96 I have noted no trace of horizontal lines for the text; from f. 100 on, the horizontal lines were made by a sharp point from marginal pin- pricks (mostly trimmed off) . The number of lines to a column varies from 28 to 36, averaging about 30, with 60 or 70 letters to the line. The first hand ends at 96r; a second hand has done lOOr to 129r and 172r to the end; from 133r to 165b the script, though much like that of the first hand, is probably by a third. There are marginal scholia in red from 22r to 40» and from lOOr to 133r, in black from 126b to 142». The titles are in red, and there is occasionally red ornamentation at the beginning or at the end of a work; red capitals are frequent, not only in the margin but also in the body of the text. The letters of the numerous diagrams and sometimes also parts of the diagrams themselves are done in red. Contents: 22r-40o &piotot4\ovs irepl yevtoetas Kal rirapTov; 89r-96r dpurroTAous, vepl xpw/idrwp; 100)^129r dpicrrorAov* trepl \f/vxys; 133r-142» dpiffroriXovs irepl alcd-fyrewv Kal alc6rjT&v; 143r— 146r apiffTortXovs, irepl ixv^/x7]s xal tov fivtifMvefeiv; 146r— 154tf apiaTorpKovs, irepl tiirvov Kal iyprjybpaetas Kal tjjs ko#' iirvov /Mwriirijs; on 154t> begins apurTorfkovs irepl yuxKpo- /3i6tijtos Kal ftjaxujSiArijTos ; on 165t> ends the De iuventute et senectute, vita et morte; 173r— 292a aXe^dvdpou d(ppodi(ri4ws tGjv &pkttot£\ovs p^rewpoXoyiKtov inr6p.vi)/m (a false start was made on 172r) ending with tjs (Hayduck 226, 22). This last section was collated by Kalbfleisch for Hayduck's edition of Alexander. 196 F. H. Fobes A similar group of two MSS is formed by the Matritensis 1 and the Ambrosianus 2 in the first two books and by the Matritensis and Parisinus 1880 3 in the last two. 1 Matritensis 41 (olim 185) is described by Graux, "Notices sommaires des manu- scrits grecs d'ljspagne et de Portugal, edited by Martin (Nouvelles archives des missions scientifiques et littiraires, II, Paris, 1892, 109) ; also by Graux-Martin in "Frag- ments inedits de Lydus xepi Suxrripeiwv" (Revue de Philologie, XX [1896], 23-35), and in "Figures tirees d'un manuscrit des Me'ttoTofogiques d'Aristote" (ibid., XXIV [1900], 5-18, with plates). It is a quarto MS (the dimensions I do not know) and is assigned by Graux to s. XII. 'Ambrosianus E.93.sup. (olim 325), s. XII, is an incomplete parchment MS bound in brown leather and rebacked. The gatherings — more of them are quaternions than anything else — are in some cases irregular. The greater part of the first two have been lost; f. 2 should stand after f. 8; ff. 3-8 area trinion; and the gathering beginning 9r is signed S; at the end a number of folia have been lost between 332 and 333, and one folio after 334. The gatherings are constructed in the usual way — flesh side to flesh side and hair side to hair side, with the flesh side outside. The signatures, placed in the lower left corner of the first rectos, have in most cases been lost in trimming. Lost also are most of the old folio numbers, which stood in the lower right corner of the rectos. Since the mutilation of the MS a modern hand has numbered the folia in pencil, with no error except at f. 43, which it at first skipped and later marked "42a"; the folio marked "43" should then have been 44, and so on to the last folio preserved ("334"), which should have been 335. The pages are 262 mm. high, 180 mm. wide; each contains one column of text 195-200 mm. high, 122-25 mm. wide. The hori- zontal lines, made with a sharp point, number 22 to the column. The script, standing below the lines, is full and round, slanting a little to the right. The Aristotle text is distinguished from the Alexander by capitals at the beginnings of sections and by guillemets in the margin. Diagrams are done in red and black. There are corrections by the first hand, in the same brown ink used for the text, and by a later hand, in a darker ink which is readily distinguishable. There are also glosses, some in red and some in black; those in red occur at the beginning of book i only, and are, I believe, by the first hand. Contents: (on the second recto of the paper feuille-de-garde) "Alexandri Aphro- disiensis | in Aristot Metheora Comment. | Codex optimus seculo XII scriptus | E.93" ; on a parchment sheet which follows the feuille-de-garde is written, among other things, "fuit ex libris Octaviani ferrarij"; then two paper sheets, mostly blank; at the top of lr a late hand has supplied a title, 'Egrry-ficrem &Ke£&v8pov d0po6Wws (sic) «'s t4 /j*Te apuTTOT (\ovs ; f. 1 begins to tov 65otos pjyeffos (Hayduck 8, 10; one folio must have been lost since the MS was examined by Tschiedel for Hay- duck; see Hayduck, p. vi) and ends irXijpT) irvpbs (339624); f. 2 begins s (132, 17); following f. 334 come two paper sheets and the second half of the feuille-de-garde, all blank. 3 Parisinus 1880 (olim Medic-Reg. 2617) — Thurot's "tai" and Hayduck's B — s. XII, is a parchment MS bound in red leather with gilt tooling. The gatherings are quaternions and trinions, with an irregular gathering of three folia at the end; Textual Problems in Aristotle's "Meteorology" 197 The close relationship of the Matritensis and the Ambrosianus may be illustrated by the following instances, in each of which all the other MSS agree in the right reading: 341a33 yap om. Matrit. Ambros. 345ol3 a-apo] 7r«pi Matrit. i Ambros. 1. 357a24 Si] yap Matrit. Ambros. 360a35 koi om. Matrit.i Ambros.i. 363o7 ow«x«s] awcxw Matrit. Ambros. 363a25 Si] Si xal Matrit. Ambros. 368612 tov om. Matrit. Ambros. 36962 t« om. Matrit. Ambros. No less close is the connection of the Matritensis with Parisinus 1880; here also I quote only instances in which all the other MSS agree in the right reading: 372o3 tV om. Matrit. 1880. 372611 vjrapxovo-i(v)] wapx« Matrit. 1880i. 372630 yap] 1880var. Si Matrit. I88O1. 373al2 p.iv om. Matrit.i 1880. 376a22 yap om. Matrit. I88O1. 379626 <5>s om. Matrit. I88O1. 381625 ftiv om. Matrit. 1880. 382ol6 jrpos] on Matrit. I88O1. 38262 tovtov] rovrmv Matrit. I88O1. 384a28 816 koi] 810 Matrit. 1880. 388624 jravra] irdvra fiiv Matrit. 1880. 390ol6 Si om. Matrit. 1880. The nine MSS fall then into four groups of two each: H and N, E and 314, J and F, the Matritensis and the Ambrosianus in the first the quaternions are constructed in the usual way; in the trinions the first recto is flesh side, the second hair side, the third varies. The folia have been correctly num- bered by a modern hand ("1" to "253"). The pages are 260 mm. high, 180 mm. wide; each contains one column of text 189-203 mm. high, 119-21 mm. wide. Dia- grams are in red ink. The ink of the text varies between brown and brownish black; there are scholia by the first hand, and some by a later hand using darker ink. Here, however, it is less easy than in the Ambrosianus to distinguish the corrections of the two hands. In respect to signatures, horizontal lines, script, method of distinguishing the Aristotle text from the Alexander, the MS closely resembles the Ambrosianus; and is probably the work of the same scribe, though Hayduck does not suggest this. Contents: feuille-de-garde and two blank leaves; at the top of f. lr various notes by a late hand and the figures "F. 40," "No. 47 secundae," "1063," "2617"; title: d\ef <£PSpov &poSiaUias els \y tCov iiereupoKoyiKdv (in ornate capitals); ff. lr— 114r the complete text of Meteorology Hi with Alexander's comment; ff. 115r-253» the complete text of Meteorology iv with Alexander's comment. 198 F. H. Fobes two books, the Matritensis and 1880 in the last two. How many- families do these four groups represent? This question, which would naturally have come up some time ago, I plan to defer still longer, pleading as my excuse the remarkable circumstance that the stemma of the first two books and that of the last two are different: Whereas in the first two books the parent of the Matritensis and the Ambrosianus belongs with H N, in the last two the parent of the Matritensis and 1880 belongs with J F. 1 If, in some of the instances which follow, the reading given as correct makes no better sense than the reading given as incorrect, my excuse is that, since E 314 rarely agree in demonstrable error with any other MS or group of MSS, the appearance of the reading of E 314 in another group is prima facie evidence of its having stood in the archetype. I quote only instances in which the other four MSS agree in the right reading: 342626 mipa. rZv] Trap' Matrit. Ambros. H N. 349o21 avip.ov om. Matrit. Ambros. H N. 351o36 Uriporepav] £rjp5>v Matrit. Ambros. H N. (Bekker's apparatus wrongly implies that H has £r)porepo)v.) 355ol8 wav] vav 6 iJAxos Matrit. Ambros. H N. (Bekker errs in saying that H N have ttSs 6 rjXios.) 362all fiopau. piv] (p..fi. F) (3opoxi Matrit. Ambros. HN. (Bekker wrongly implies that H has fiopeai p*v.) 367o23 apavpov] apavporepov Matrit. Ambros. H N. 370627 Kal avrb dvayKrj h>] dvay/07 kvkXov JF Matrit. I88O1 koI avrb avayKt) kvkXov edd. Aristotle is speaking of the whirlwind (Stvos), whose origin he illustrates by the little eddies frequently seen on the street or in a gateway: when a current is continuously advancing against an obstruction, each particle in the current, upon reaching a certain point, is deflected in a certain way; if the obstruction is of a certain shape, the particle may be repeatedly deflected until it finds itself advancing again in its original direction; if each succeeding particle has followed the same course, the series of deflections has resulted in a single motion, a h; tovto 8' ia-rl kvkXck- 0x1 yap pJa ] oiSeVo) J F Matrit. 1880. 373a7 y) to (ante A A) om. E CO rr. J Fi Matrit. 1880. 37369 ovk om. J l Fi Matrit. 1880i. (6 in Bekker's apparatus is proba- bly a misprint for ovk.) 376a28 5^] ^ JF Matrit. 1880 CO rr. (Bekker errs in attributing the reading jpr to H also.) ior I88O1 it 314. 383627 w8v J F Matrit. 1880. 385ol4-16 eAarov avqhxTov (dveAaTov) are thrown back 76 letters, behind dpavo-rbv, by J F Matrit. 1880. (H does not share the error dve\oTov attributed to it by Bekker.) 385a30 TrdvTa] roSra J F Matrit. 1880. The total number of instances in which the Matritensis and the Ambrosianus agree in probable error with H N I reckon at about 35; the total number in which the Matritensis and 1880 agree in probable error with J F, at about 90. Since, however, the stemma is in neither case absolutely clear-cut — the Matritensis and the Ambrosianus agreeing with J F in some dozen cases, and the Matritensis and 1880 agreeing with H N in some half-dozen — I give in full the evidence against my view: 339635 ij/uv e'vTeS&v] e\ ^. J F Matrit. Ambros. 342630 rives Kal KaXovp.€VO>v HvOayoptiiov] Tives KaXovfitvuiv II. J F Matrit. Ambros. KaXov/jLtvwv II. rives H N. 35765 tov] del. (ut videtur) Hcorr. om. J F Matrit. Ambros. 363ffll6 juet£v xai E jxdt,u> Kal irXiltav Kal Fi jaet£a> Kal irXtita Kal J Matrit. Ambros. 364a26 curoAei7rei Te] dwroAeiWei Fi cHroAeiVeTai J Matrit. Ambros. 366620 yi(y)veTai yap) yap yi(y)veTcu J Fi Matrit. Ambros. 367634 7rpde«Ti ttoAAcikis ^fiepas] vds] iravroSairriv &viqv J F Matlit. Ambros. 368625 6 om. J F Matrit. Ambros. 368633 yt(y)v£Tcu eraer/tos] yC(y)vovrai aao-fiol J F re c. Matrit. Ambros. yivovrat and oikus before consonants, the cases in which E 314 read oi5ra> and the other MSS read ovtus are only about half as numerous as those in which E 314 read ofrrws and the other MSS read ovtu. In the similar matter of iras and airas there are two instances on each side. Where the alternative readings concern only a change in the order of words — disregarding cases where the desire to avoid hiatus might have been a motive to correction — it is almost impossible to say, in half the cases, that the one order is more simple and the other more complex; in the other half the suspicion of simplification falls less often on E 314 than on the other MSS. E 314 much more often omit the article where the other MSS have it, than vice versa; and in the majority of these cases the omission makes lectio difficilior. 1 These facts, however, I take to indicate, not that the other MSS form a single family, but that E 314 have preserved the tradition faithfully in many instances in which the other families have inde- pendently suffered correction; for it is unlikely that the Meteorology was ever completed, much less polished for publication. 2 And in the remaining cases of minor alternatives — a miscellaneous lot — E 314 can claim no more lectiones difficiliores than can the other MSS. I give a few of those which show the suspicion of correction which rests on E 314: 340633 avmrepai] av«> E 314 edd. (The tokos here referred to is the same as the avw totos of 340630; but this is no warrant for changing avotrepto to aW.) 350628 eis] «V F CO rr. E 314 edd. («s, the lectio difficilior, is certainly permissible, either as a substitute for em [Eucken, Ueber den Sprach- gebrauch des Aristoteles, pp. 32-33] or as meaning "into a small body of water." «r' dA/yov, as interpreted by Ideler and S.-Hilaire, becomes almost equivalent to Kara fUKpov.) 1 At 341ol5, where E has ffeppiv, the lectio difficilior, and 314 and the other MSS and Bekker have rb depp&v, Ideler thinks the article is better omitted; at 347627, a similar instance, where E 314 and Bekker have ri> 6eppi>v and the other MSS have depixbv, Ideler retains the article. * Thurot, op. cit., XX, 416. Textual Problems in Aristotle's "Meteorology" 203 35265 aev(v)dovs itvai Tims] toiis fiev devdous ttvcu. Ei edd. Bonitz (Aristo- telische Studien, II, 420) tovs fiev dewdovs ttvai twos Fcorr. 314. (The very bold construction tis /iev . . . . 6 8e apparently stood in the parent of E 314. In its favor, though not conclusive, is the fact that the reason following is, strictly speaking, not a reason why some rivers never fail and some do fail, but a reason why some never fail. tovs 8e nrj may therefore have been an afterthought.) 358630 dXX' $ iXXk E 314 edd. (dU' y is lectio difficilior, both in itself and because of the preceding dAA' rj, which means "except." It falls in the third of the classes enumerated by Professor Cook Wilson, "On the Use of dXV % in Aristotle," Classical Quarterly, III [1909], 121 ft.) 35962 roikmv] totW oh E 314 edd. (See Kuhner-Gerth, § 546, 5, £; Diels, op. tit., p. 15.) 375622 ' fj] s irepi dfovo. rip/ i' fj Matrit. 1880corr. oxTirtp aiova ttjv €' a>v 17 E ' rjv rj 314. vxnrtp 5£ova rijv i' fj ^c, "the line rjk"]; but, by the same token, the latter is lectio difficilior.) 384o30 As] Sxnrtp E 314 Fcorr. edd. (Possibly the parent of E 314 some- times avoided this use of ms; cf. 383o5 is] olov E 314 Bek.) These minor instances, then, do not by any means prove that there are not three families of MSS. A number of more serious cases, which can be adduced against my supposition, I plan to examine in greater detail. 344ol2 aipas cet. (In order to defend the supposition that there are three families, we must suppose that opd. In either of the last two cases there is hendiadys.) 349a20 that] Ei (ut videtur) 314 F va r. oVra cet. (ovra, involving a violent anacoluthon of a kind for which, so far as I can find, there is no parallel, must have stood in the archetype; correction to «ku would be easy. Thurot [op. tit., XX, 419] seems to have been under a 204 F. H. Fobes misapprehension as to the reading of E, though Bekker's report is correct. Cf. Physics 208525; Diels, p. 22.) 354625 wavrwiv tcr^arov d6' &s ol irXtUTTOi] iravriov io~xarov dO' a>s ol irXa<7T0( tovto Matrit. tovtiov iayarov d$' o>s ol jrXeioToi tovto Ambros. rovrtov toyaTov d6' o>s 7r\eto-TOi tovto H N tovtwv layaTov dre s ttoXAoi Olympiodorus. iiravrosv means " of all the four elements of the xdo>ios," not "of all the spheres or layers." It was this ambiguity which prob- ably led to the gloss or correction tovtwv, whether in the archetype or, independently, in latter MSS. For ol irX&rroi, as well as for Aristotle, fire was the last of the cosmic elements; but the former thought of fire as pervading also the heavens, where the latter puts aldyp. See Gilbert, Die meteorologischen Theorien des griechischen Altertums, Leipzig, 1907, p. 664.) 360a26 o-wurrairo av 6] edd. o~uvuTTaxTO av Ei awurraJLTO av ova) Ecorr. owiotoit' av 6 avu> 314 awiaTaTai 6 avm cet. ( must have stood in the archetype, avm springing from &v 6 either by an error of sight or by an error of hearing. 314 per- haps represents a correction made by the parent of E 314; E alone ventured to drop the objectionable ova>.) 376622-28 t&v pitv .... ivifuivcv om. E Alexander 6 tjXuk .... ivfaivtv om. 314. Olympiodorus had at least a part of these words. (Cf. Ideler and S.-Hilaire, ad he.; Thurot, op. cit, XX, 419. If, in spite of Alexander, these words belong here, the instance is irrelevant to our purpose; if they are an insertion, or if they belong somewhere else, the difficulty can be explained away by supposing that in the archetype they stood in the margin or were in some way marked as questionable.) 377ol0 iKBjrnt.fipias pr/re (w>"')] W to-OTax<»S ptj8* (/078') E 314 Alex, text Philop. lemmata. 364630 am raCra 8wa/x«] t. &. 8. F &. S. t. E 314 Alex, lemma Olympiod. lemma edd. 375634 to K/Zjj] Olympiod. lemma om. E t 314 Alex. text. 376all ^H (post ■nj's)] jHic Erec. 314 Alex, text Olympiod. text. (^ 8 . . . . Kfi. are omitted by Ei; but the preceding JUk offers an easy explanation of the omission if we suppose the parent of E to have read JSk.) 377al0 oo-<{>] E 314 Alex, text oo-av Matrit. oo-ov cet. (See above, p. 204.) 377al0 KarmTtpov] KarajTepa) E 314 Alex, text KaTwepos Bek. 1 See Diels, p. 23 ; Hicks, De anima, pp. lxxix-lxxxii. 206 F. H. Fobes 377a29 irept iraprjXiW] Alex, text Olympiod. lemma rrap' i)Xiiov N irept Traprj\iov Ei 314 Alex, lemma edd. irtpl rwv 7rapj;AiW Ecorr. (A clear instance of the unreliability of the lemmata.) 389a25 8a] xp« E i XPV E <=orr. 314 Alex. text. These instances are to some extent offset by occasional agreement of Alexander with some other group; for, as in the De anima (see Hicks, p. lxxxii), Alexander does not agree with E alone: 354625 iravTiav] tovtwv J F Ambros. H N Alex, text to Olympiod. lemma. (See above, p. 204.) 356al2 iraxriv] iraA.iv HN Alex. text. (The reading of H is possibly due to correction; Bekker, who collated H before it was repaired, implies that it reads irao-iv.) 372a8 Kepawvptvov] Kepawvp.a>mv Ecorr. Fcorr. H N Alex. text. 375a2 oivuajv] iKvdcrdai Hi N Ecorr. SvvcurOat. SiiKvturOai, Hvar. 1880corr. Alex, text SvvaxrOai SuKviixrOax rr/v oij/iv Fcorr. 382a20 8' eAAaVov] 314 Matrit. Se Aottov E t St A«W Ecorr. J F 1880 Si Aorrdv Kal H N Alex. text. In book i, for which alone I have collated the Aldine Aristotle, I note only one instance which could be used to support Susemihl's words (Rhein. Mus., XL, 564, note 3), "Die Quelle der Aldina steht jedoch der des Alexandras von Aphrodisias noch ungleich naher als E"; and that one is indecisive: 345a23 rnxud)K\u>v] Olympiod. lemma rd t)iju.k{>k\ioi> Matrit. ed. Aid. Alex. text. I do, however, note one agreement in real error between the Aldine Aristotle and the Aldine Alexander: both omit the necessary Kal at 342a35. Hence I suspect that Susemihl, writing before the publication of Hayduck's Alexander, may have been misled by contaminated lemmata in the Aldine Alexander. If Susemihl is right, 314 gains in value (see above, p. 201, n. 3). To judge once more by the first book alone, Alexander is no more closely related to the Aldine Aristotle than are Philoponus 1 and Olympiodorus. 2 1 340624 \eyofiivov] \cyo/i4vov Kal Ka\ov/Urov 314 ed. Aid. Philop. text Ixopivov Olympiod. lemma. 2 34165 (arliy) Kal roiriav Kal iroWuv &\\wr Philop. cod. V lemma roirinv fori Kal tQv JXXw Philop. cod. M lemma i o-tkyvrp Ambros.i ravra Karot atX-qvrts N Olympiod. lemma. 342a36 Si ko\ 1*% J V lemma 8k ml E 314 M lemma &(8') i™ F Matrit. Ambros. H N. 346612 £Km(/x)irpa/i£V)j5] JFi (utvidetur) Matrit. Ambros. H N V lemma e/MrijMrpa/teVijs E 314 F CO rr. M lemma (Bekker errs in attributing this reading to N). (Stephanus' Thesaurus gives no instance of ctori/Mrpavai from the classical period; but the form is attested for 367ol0.) Olympiodorus agrees with E 314 less consistently than does Alexander: in many cases he agrees with E alone; in some cases with F alone, with Matrit. Ambros. H N (in book i), with J F Matrit. 1880 (in books iii and iv) ; in one or two cases with 314, with H, with N, with HN, with Matrit. 1880 (in book iii). If then the archetype of the Meteorology MSS antedates Alex- ander, we must think of both Alexander and Olympiodorus as work- ing with more than one MS or with a MS containing variants. If it does not, we must infer that some MSS and MS families have been independently corrected from Alexander and from Olympiodorus; or that they have been independently corrected on older MSS now lost; or that the archetype had numerous variants. 1 If the arche- type antedates Philoponus, Philoponus' MS belonged to the E 314 family; if it is later than Philoponus, we must infer that the E 314 family has been corrected on Philoponus or on some MS preserving the tradition of Philoponus' MS, or that the archetype had variants. In any case, the evidence afforded by the commentators, so far as I am acquainted with it, does nothing more than to cast suspicion upon the three-family theory. If in the Meteorology, as, according to Diels, in the Physics, the archetype is later than the commentators, and if, as in the Physics, the condition of the MSS is due to a com- bination of the various processes suggested, the suspicion cast becomes negligible. The chief advantage of knowing the date of the archetype would then be that it would afford a basis for judging 1 Diels, p. 19. 208 F. H. Fobes the value of the commentators' texts — f or it is of course conceivable that an early commentator may have used a text distinctly poorer than a late MS. 1 In trying to fix the date of the archetype I have looked first for passages which the archetype must have had and the commentators lack and for passages which the archetype must have lacked and the commentators have. But I find only the four Olympiodorus lemmata mentioned above in I and the three passages adduced by Thurot (op. tit., XX, 419; XXI, 404). Of these last three, 376622-28 and 387627-31 have been discussed above and are inconclusive; there remains 384614-15, where &are irpbs irff^v &/jufx>Tipa)P deXrai' Sid a\vTov (37 letters) is not found in any of the MSS nor in either of the Greek-Latin translations. Olympiodorus probably had these words before him (cf. 313, 1 ff.), and Alexander paraphrases: Sore ■jrp&s rfy tttj^lv aptfroripuv avT$ 8ei ' 8id oi)K ?s tXaiov ks oi irtXyoi. and as 6 Kepapos omitted by E H N might each have filled a nine-letter line. Textual Problems in Aristotle's "Meteorology" 211 How much light the MSS of the De generatione et corruptione may throw upon their archetype, I do not know; Bekker's four MSS all bear traces of a nine-letter line. From Table I, based wholly on Bekker's apparatus, I exclude omissions of seven letters or less and omissions in which homoeoteleuton or the like might obviously have been a factor. Out of 18 omissions represented in this table it will be noted that all but six come within one of being multiples of nine. Similarly out of 28 omissions in which homoeoteleuton may have played a part all but 15 come within one of being multiples of nine. This evidence speaks much more clearly than does the evidence of table i Letters Omitted By E By F By H By L 8 1 1 1 3 1 1 9 3 4 10 11 1 12 1 14 i l l 20 25 27 i 35 l 1 53 l 62 1 73 l 185 1 the Meteorology; if it is permissible to argue from the archetype of one work to the archetype of another, we may see here a further indication of a nine-letter line in the archetype of the Meteorology. The position of the Escorial fragment 1 in the stemma cannot be exactly determined. It is not an ancestor or a descendant of any MS which I have collated. It shares one trivial error with N, one slight error with J F, one serious error (though possibly originating in a variant) with Matrit. Ambros. As to the relative value of the MSS, my collation of the lemmata and citations of the commentators has persuaded me that E, in spite of its countless individual errors and the countless corrections by various hands not easily to be distinguished, is more valuable than J, which I formerly preferred. E and J are of course the two most rotographs. see my former article. I have since collated this fragment from 212 F. H. Fobes valuable MSS. J and F are more important in books i and ii, where they stand alone, than in books iii and iv, where they are supported by the Matritensis and 1880: H and N are more important in books iii and iv, where they stand alone, than in books i and ii, where they are supported by the Matritensis and the Ambrosianus. F, which Bekker valued highly, is contaminated; 1 in books i and ii it affords 1 F is contaminated not only from E 314 but also from N: 341624 toO om. F, Ni. 344o9 4a.lves] Oiairep E x om. Ecorr. 314. 360ol4 ical] Kal to Ej om. Ecorr. 314. 375614 ird/MTTav] Traaav Ecorr. 314. 379o5 Sk irdvroiv tSiv SXhwv Toirav] y&p twv SKKiav airivrav Ei yi.p Toirmiv awdrruv Ecorr. 314 Si tuv &X\wv dirivruv Bek. 387620 •$• fapao-x&rOat N . wapixeadai S. Ei edd. S.s Par. 2032 Vat. Urb. 37. 353618. (The reading attributed to F is not found in any of Bekker's MSS nor in any of the ten others which I have collated.) 358o30. (The reading attributed to H is not found in any of Bekker's MSS nor in any of the four others which I have collated.) 358614 S