GIFT OF * UNIVERSITY OF CALIFORNIA PUBLICATIONS IN CLASSICAL PHILOLOGY Vol. 2, No. 9, pp. 173-226 June 1, 1912 NOTES ON THE TEXT OF THE CORPUS TIBULLIANUM BY MONROE E. DEUTSCH UNIVERSITY OF CALIFORNIA PRESS BERKELEY UNIVERSITY OF CALIFORNIA PUBLICATIONS. Note. — The University of California Publications are offered in exchange for the publi- cations of learned societies and institutions, universities, and libraries. Complete lists of all the publications of the University will be sent upon request. For sample copies, lists of publications or other information, address the MANAGER OF THE UNIVERSITY PRESS, BERKELEY, CALIFORNIA, U. S. A. All matter sent in exchange should be addressed to THE EXCHANGE DEPARTMENT, UNIVERSITY LIBRARY, BERKELEY, CALIFORNIA, U. S. A. CLASSICAL PHILOLOGY.— Edward B. Clapp, William A. Merrill, Herbert 0. Nutting, Editors. Price per volume $2.50. Cited as Univ. Calif. Publ. Class. Phil. Vol. 1. 1. Hiatus in Greek Melic Poetry, by Edward Bull Clapp. Pp. 1-34. June, 1904 ~ - ?0.50 2. Studies in the Si-Clause. I. Concessive Si-Clauses in Plautus. H. Sub- junctive Protasis and Indicative Apodosis in Plautus. By Herbert O. Nutting. Pp. 35-94. January, 1905 _ _ 60 3. The Whence and Whither of the Modern Science of Language, by Ben}. Ide Wheeler. Pp. 95-109. May, 1905 .25 4. On the Relation of Horace to Lucretius, by William A. Merrill. Pp. 111-129. October, 1905 _ .26 5. The Priests of Asklepios, a New Method of Dating Athenian Archons, by William Scott Ferguson. Pp. 131-173. April 14, 1906 (reprinted September, 1907) _ - , .50 6. Horace's Alcaic Strophe, by Leon Josiah Richardson. Pp. 175-201. March, 1907 _ 26 7. Some Phases of the Relation of Thought to Verse in Plautus, by Henry Washington Prescott. Pp. 205-262. June, 1907 _ .69 Index, pp. 263-270. Vol. 2. 1. Some Textual Criticisms on the Eighth Book of the De Vita Caesarum of Suetonius, by William Hardy Alexander. Pp. 1-33. November, 1903 30 2. Cicero's Knowledge of Lucretius 's Poem, by William A. Merrill. Pp. 35-42. September, 1909 10 3. The Conspiracy at Rome in 66-65 B.C., by H. C. Nutting. January, 1910 10 4. On the Contracted Genitive in I in Latin, by William A. Merrill. Pp. 57-79. February, 1910 _ 25 5. Epaphos and the Egyptian Apis, by Ivan M. Linforth. Pp. 81-92. August, 1910 10 6. Studies in the Text of Lucretius, by William A. Merrill. Pp. 93-150. June, 1911 50 7. The Separation of the Attributive Adjective from its Substantive in Plautus, by Winthrop L. Keep. Pp. 151-164. June, 1911 15 8. The 'OapiuTis of Theorrjws, T>y Edward B. Clapp. Pp. 165-171. Oc- tober, 1911 15 9. Notes on the Text of the Corpus Tibnllianum, by Monroe E. Deutsch. Pp. 173-226, J.a>e, 1912 50 GRAECO-ROMAN ARCHAEOLOGY. (Quarto.) Vol. 1. The Tebtunis Papyri Part 1. Edited by Bernard P. Grenfelh Arthur S. Hunt, and J. Gilbart Smyly. xix -+- 674 pages, with 9 collotype plates. 1902. £2 5s, $16. Vol. 2. The Tebtunis Papyri, Part 2. Edited by Bernard P. Grenf ell and Arthur S. Hunt, with the assistance of Edgar J. Goodspeed. xvi+485 pages and 2 collotype plates, with map. 1907. Vol. 3. The Tebtunis Papyri, Part 3. Edited by Bernard P. Grenfelh Arthur S. Hunt, and J. Gilbart Einyly. (In preparation.) For sale by the Oxford University Press (Henry Frowde), Amen Corner, London, E.O. (£2 5s), and 91-93 Fifth avenue, New York ($16). Copies for exchange may be obtained from the University Press, Berkeley. UNIVERSITY OF CALIFORNIA PUBLICATIONS CLASSICAL PHILOLOGY Vol. 2, No. 9, pp. 173-226 June 1, 1912 NOTES ON THE TEXT OF THE CORPUS TIBULLIANUM BY MONEOE E. DEUTSCH PREFACE The following paper was submitted to the faculty of the University of California in August, 1910, as a dissertation for the degree of Doctor of Philosophy. Since that time it has been revised somewhat, but it is essentially unchanged from its original form. I desire to acknowledge here my great indebtedness to Professor William A. Merrill, to whom I owe my instruction in the field of textual criticism. Monroe E. Deutsch. Berkeley, California, December, 1911. ASS I, 3A< 2, jrES - 88. Dig •CUSS] I, 5, 76. I, 6, 7. I, 9, 25. I, 10 , 37. II, 2, 17-2S ; II, 3, 61. II, 4, 43. II, 5, 47. III, 6, 3. III, 12 (IV, 6), 19-20 III, 14 (IV, 8), 5-6 244548 174 University of California Publications in Classical Philology. [Vol. 2 I, 2, 88 Proposed reading: at tu, qui laetus rides mala nostra, caveto mox tibi : non mi uni saeviet usque deus. Here A V and 6 unite in reading non unus; P, et iratus; and £, non vanus. None of these seems possible here ; so the majority of the editors follow the early Italian scholars in reading non uni. 1 The change from uni to unus is thus explained by Belling, Prolegomena, p. 39, note: "Der libr. arch, schrieb unus, da in seiner Vorlage t das i dem Abkurzungszeichen fur us ahnlich sah." Heyne easily accounts for it thus: "Scilicet primum s adhaeserat ex sequenti voce: turn unis mutatum fuisset in unus." As to the interpretation of uni, Golbery declares : ' ' Sensus est : Amor non uni mihi, sed et tibi saeviet, quamvis me nunc laetus irrideas. " The sense is perfectly good, but does Tibullus ever use unus in this way? The following are the instances of the word in the Corpus Tibullianum : I, 6, 23 illam sequar unus III, 6, 32 una serena dies. ad aras. Ill, 19 (IV, 13), 5 uni mihi. I, 2, 9 uni mihi. Ill, 10 (IV, 4), 19 in uno I, 2, 58 de me uno. corpore. Ill, 7 (IV, 1), 142 una per ostia (Heinsius). In other words, unus is never used substantively in the Corpus Tibullianum, and of the six well-authenticated instances of the word, four show it modifying some form of ego. Moreover, in the elegies of Tibullus himself (if we assume III, 19 to be by Tibullus), four of these six uses of unus occur, in all of which unus modifies some form of ego. This makes the substantive use appear improbable here. Moreover, the substantive use of adjectives in the dative i However, among the other suggestions are : uni is — Goerenz ; unis — F. W. Eichter (in vers. German. 1831); vacuus — Huschke; in me — Hiller (but Hiller reads non unus in Corpus poetarum Latinorum, 1893) ; ullus — Rabus (Observationes in Tibulli carmina, p. 6. Aug. Vind. 1837); in nos — Leo (Philol. Unters. 2[1881], p. 39); "caveto: mox tibi nam durus saeviet usque deus" — Fr. A. Rigler; "caveto: mox tibi iam lusus saeviet usque deus" — Kemper (Quaest. Tibull., Monasterii, 1857, p. 46); mitis — Baehrens. 1912] Beutsch. — Notes on the Text of the Corpus Tibullianum. 175 singular is rare in the Corpus Tibullianum. The following list includes all instances wherein an adjective in the dative singular stands alone, whether it is used substantively or the noun which it modifies is to be supplied from the preceding clause : 2 absenti I, 8, 53. amanti I, 3, 65; I, 5, 57; II, 4, 15. anhelanti I, 8, 37. avarae II, 4, 35. fatenti I, 6, 29. fesso I, 10, 42. formosae Ill, 10 (IV, 4), 4. misero I, 6, 2; I, 8, 23; II, 4, 4; III, 19 (IV, 13), 20. nulli I, 6, 77; III, 5, 7. roganti I, 4, 55. securae II, 4, 50 (dative or genitive). sopitae II, 6, 38. suae I, 4, 75; II, 5, 103. tardo Ill, 7 (IV, 1), 92. tenero I, 8, 51. Of these twenty-two instances, but nineteen are in Tibullus' own elegies (even if one assumes III, 19 to be his). It will be further noted that of these nineteen, eight are participles; among the remaining eleven, comprising as they do all instances wherein the adjective in this form stands alone, we find misero used four times, and suae twice. If these were set aside from the list, but five isolated examples would remain in the elegies of Tibullus. But even if we take the whole list, exclusive of participles, the number is exceedingly small, too small to justify the introduc- tion of the substantive use of an adjective which is never used substantively in the whole Corpus Tibullianum. Aside from this objection to the use of uni alone, the usual reading would demand for uni the meaning of "one and the same man," a meaning not found in Tibullus. On the other hand, as the instances of Tibullus' use of unus show, we should expect it to modify some form of ego. More- over, the meaning of the passage demands something that is equivalent to "uni mihi" (quoting from Golbery's note), which 2 Based upon the Index Verborum in Hiller's edition. 176 University of California Publications in Classical Philology. [Vol. 2 very group of words is found twice in Tibullus. With this thought I have suggested the insertion of mi before uni; we have then exactly the phrasing that fits the sense of the passage. It is true that the form mi is not found in the Corpus Tibul- lianum. In Propertius, 3 however, the shorter form is found in I, 12, 19; II, 18, 30; II, 22, 1; II, 22, 18; II, 30, 25; IV, 1, 62; IV, 8, 53; IV, 11, 47. In Catullus 4 the use is exceedingly fre- quent. Lucretius employs mi in I, 924 and III, 105 ; it appears in the Aeneid VI, 104 and 123. That the form never appears in the Corpus Tibullianum as we now find it, is after all not strange, when one notes these figures : Propertius.s Corpus Tibullianum.s me (ablative) 41 11 That is, the ratio of the uses of me (ablative) in the Corpus Tibullianum to the uses in Propertius is 11 : 41. Propertius using mi but eight times, at this same ratio the Corpus Tibullianum would contain mi only twice and a fraction. In other words, mi should by no means be expected frequently in the Corpus. The elision of the monosyllable which the proposed reading would introduce, has the following parallels in our text : I, 2, 56 si in. Ill, 6, 25, qua est. I, 2, 58 me uno. Ill, 7 (IV, 1), 75 si inter- I, 4, 56 se implicuisse. rupto. I, 6, 59 te adducit. Ill, 7 (IV, 1), 179 se accin- I, 6, 61 me adfixa. gere. I, 7, 9 me est (sine marte III, 7 (IV, 1), 182 me ad- ibi — Baehrens). versa. II, 1, 69 se exercuit. Ill, 19 (IV, 13), 19 me au- II, 3, 5 cum aspicerem (dum dacius. — Heyne). II, 3, 61 qui abducis (text uncertain). Of these fourteen instances ten are in Tibullus, 6 one in Lygdamus, and three in the Panegyric; it will be further noted that, of the ten in Tibullus, seven involve personal or reflexive pronouns. 3 J. S. Phillimore, Index Verborum Propertianus. * Neue- Wagener, Formenlehre, II, 349 foil, s Index Verborum in Hiller, Albii Tibulli Elegiae. 6 If we assume III, 19 to be by Tibullus. 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 177 The elision of the monosyllable has therefore sufficient warrant in the usage of Tibullus. It will be noted that the proposed reading, as to elision and word-order, exactly parallels me uno, Tib. I, 2, 58. As mi is not found in the Corpus Tibullianum, for examples of elided mi we shall have to look elsewhere. They are frequent enough, e.g. Lucr. I, 924 ; Hor. Serm. I, 1, 101 ; I, 2, 57 ; I, 3, 23 ; I, 4, 108 ; I, 9, 50 ; II, 6, 27 ; Hor. Epist. I, 18, 112. That mi could be readily omitted in MSS (1) because elided and (2) because written in compendium, is readily apparent. For example, on Horace Serm. I, 9, 50 (nil mi officii, inquam) Orelli 's apparatus criticus contains the statement : mi a-yRo- mihi F omisit g( Codex Gothanus, B61). Again, Ennius in C.ic. de Oratore III, 58, 218 reads : 7 Turn pavor sapientiam omnem mi exanimato expectorat. The MSS read thus: "mihi L, mi P, om. M." It is of interest that in both of these instances mi suffers elision. If it be further objected that we should then have but a solitary instance of mi in the whole Corpus Tibullianum, it might be noted that there is but one instance in Persius, I, 2. 8 The use of the dative case (uni or mi uni) with saevire has always been defended by Ov. Her. IV, 148 qui mihi nunc saevit, sic tibi parcat Amor. But Leo 9 objects to its use, brushing aside this instance by declaring that mihi saevit is influenced by tibi parcat. In reply it -might be noted that mihi saevit precedes tibi parcat, and hence would not be so likely to be influenced by the construction with parcere as if the reverse order were found ; moreover, in the text before us we have caveto tibi, a use of the dative which is not uncommon; according to Leo's reason- ing, then, it would not be surprising if this use of the dative influenced the other construction so as to read mi uni saeviet. i Mueller-Friedrich. s ' ' Vocabula Satirarum Persii ' * in Auli Persii Flacci Satirarum Liber, edidit Otto Iahn. 9 F. Leo, Phil. Unters., II, 34-9. 178 University of California Publications in Classical Philology. [Vol. 2 For in this passage the dative with cavere precedes and would therefore far more easily influence the construction with saevire. Leo's second objection to the usual reading is that, even if the dative were used, he would expect nobis. This expectation is due, I presume, to mala nostra of the preceding verse. Compare how- ever the following passages : I, 2, 11-12 Et mala si qua tibi dixit dementia nostra, ignoscas: capiti sint precor ilia meo. I, 4, 77-8 me, qui spernentur, amantes consultent: cunctis ianua nostra patet. So also I, 5 41 me and I, 5, 44 nostra puella; and I, 9, 42-3 ipse tuli! and munere nostro. These are but instances, selected at random, of what is common in Tibullus. Finally, we should quite expect mi set off against tibi, for this is not uncommon in Tibullus : 10 I, 2, 97 At mihi parce, Venus : semper tibi dedita, etc. I, 1, 59 te spectem, suprema mihi cum venerit hora. I, 6, 3 quid tibi saevitiae mecum est? Cf . also Hor. Serm. I, 4, 116 causas reddet tibi ; mi satis est, etc. I should accordingly read uni with the editors, but insert mi before it. I, 5, 76 Proposed reading: utere quaeso, dum licet : in liquida nam tibi linter aqua est. AV read as above, though without est; G also omits est, and substitutes for nam the verb nat, which is likewise the reading of Vm2 in the margin. This last reading is accepted by practically all editors of Tibullus. Guyet proposed stat tibi, which was suggested by Prop. II, 9, 30 ant mea si staret navis in Oceano. Rossberg 11 would read nunc for nam. 12 Nam is approved by Maurenbrecher, 13 who quotes the readings of the MSS thus : io So too in this distich tu and rides are set off against mala nostra. ^Jahrb. f. Philol., 119 (1879), p. 77. 12 Cf. the variant readings in II, 4, 12: nunc P, nam A, iam f. is Philologus, 55 (1896), 439. 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 179 "nat G, nam AV, was ohne Anstoss ist." Belling 14 in his text reads nam, without adding est, however, which seems to me essential to the verse. Cartault 13 reads non, and Heyne suggested that if nam be retained, we should read "it liquida nam tibi linter aqua." The change from nam to nat is easily understood. The posi- tion of nam as the third word in its clause would seem impossible ; the t beginning tibi would influence the change of nam to nat, especially since this verb would appear fitting with linter. The clause being thus supplied with a verb, est would naturally be dropped. This loss would be made easier by the fact that the verse is a rather long one for a pentameter. 16 It is true that nam stands first in its clause in practically every instance of its use in the Corpus Tibullianum. But we read as follows in III, 4, 43-4 : salve, cura deum: casto nam rite poetae Phoebusque et Bacchus Pieridesque favent. And in II, 4, 12 the reading of A is perfectly acceptable : omnia nam tristi tempora f elle madent. Some editors, however, read nunc 17 and others iam. 18 In other words, in three separate passages in the Corpus Tibullianum, A reads nam in a position subsequent to the first in its clause ; of these one is accepted by the editors, a second by a number, and a third rejected by practically all. If we can accept the authority of A in III, 4, 43 for a deferred nam, the postponement of nam should not in itself be an argument for the change here, when resting on the same MS authority. Moreover, deferred nam is sufficiently common in the other poets of the period. We find the following instances : 19 i* H. Belling, Albius Tibullus, Untersuchung und Text, Berlin, 1897. Cf. his argument for nam in his Prolegomena, p. 63. 15 Tibulle et les auteurs du corpus Tibullianum. Paris, 1909. !« For those MSS (AV) that read nam we should only have to explain the loss of est. The length of the verse has already been mentioned, and the ease with which est is dropped may be seen by noting the instances of such loss mentioned on page 183. l* e.g. Baehrens, Haupt-Vahlen, Hiller. is e.g. Broukhusius, Heyne, Voss, L. Mueller. is Emilius Schuenke, De traiectione coniunctionum et pronominis relativi apud poetas Latinos, Kiliae, 1906. 180 University of California Publications in Classical Philology. [Vol. 2 Catullus XXIII, 7 nee mirum: bene nam valetis omnes XXXVII, 11 puella nam mi, quae meo sinu fugit LXIV, 301 Pelea nam tecum pariter soror aspernata est Virgil: Georgics IV, 16 omnia nam late vastant ipsasque volantis Aeneid I, 444 sic nam^o fore bello egregiam et facilem victu 518 quid veniant: cunctis nam lecti navibus ibant 731 Iuppiter, hospitibus nam te dare iura loquuntur III, 379 prohibent nam cetera Pareae scire Helenum. IV, 421 solam nam perfidus ille te colere, arcanos etiam tibi credere sensus VI, 667 medium nam plurima turba hunc habet IX, 803, aeriam caelo nam Iuppiter Irim demisit X, 585 dicta parat contra, iaeulum nam torquet in hostis. XII, 206 dextra sceptrum nam forte gerebat Cir. 221 sonitum nam fecerat illi 458 omnia nam potius quam te fecisse putabo. Cat. (Priapea) 3, 5 huius nam domini colunt me Cat. 4, 10 Clio naw2i certe Candida non loquitur. Ov. Her. XI, 61 fratris nam 22 nupta futura es Hor. Epod. 14, 6 deus, deus nam me vetat 17, 45 et tu, potes nam, solve me dementia Serm. II, 3, 20 olim nam quaerere amabam 41 primum nam inquiram quid sit furere 302 insanire putas? Ego nam videor mihi sanus. II, 6, 78 si quis nam laudat Arelli sollicitas ignarus opes Epist. II, 1, 186 aut ursum aut pugilis; his nam plebecula gaudet. Carm. I, 18, 3 siccis omnia nam dura deus proposuit IV, 14, 9 milite nam tuo Drusus Genaunos, implacidum genus Propert. IV, 8, 23 Serica nam-z taceo vulsi carpenta nepotis In this list it will be noted that nam is found in the third place in its clause in Virgil Aen. IX, 803 and XII, 206, and in Horace Carm. I, 18, 3. This position of nam is therefore by no means an impossible one. If we restore nam then, the clause stands without a verb. Would an ellipsis of est here be in accordance with the style of Tibullus? The following list 24 contains passages wherein there is an ellipsis of esse, noted in the Corpus Tibullianum : I, 1, 75 hie ego dux milesque bonus. I, 3, 5 non hie mihi mater. 20 sic nam F M 2 R a 2 b 2 : signam M 1 P 2 y 1 : signum y 2 deteriores pauci. 2i Elionam M : Clio tarn Casaubonus. 22 Text doubtful; above reading, Palmer's. 23 Serica nam taceo — Beroaldus ex emend.; serica nam tacto — V; si riga- nam tacto ceteri. 2* Based on the text of Hiller (1885). 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 181 I, 3, 7 non soror. I, 3, 9 Delia non usquam. I, 3, 43 non fixus in agris .... lapis. I, 4, 23 gratia magna Iovi. I, 6, 33 quid tenera tibi coniuge opus? I, 7, 9 foil. Tarbella Pyrene testis et Oceani litora Santonici, testis Arar Rhodanusque celer magnusque Garunna, Carnutis et flavi caerula lympha Liger. I, 7, 44 sed chorus et cantus et levis aptus amor, sed varii flores et frons redimita corymbis, fusa sed ad teneros lutea palla pedes et Tyriae vestes et dulcis tibia cantu et levis oecultis conscia cista sacris. I, 10, 3 turn caedes hominum generi, turn proelia nata. I, 10, 9 non arces, non vallus erat. I, 10, 26 25 hostiaque e plena rustica porcus hara. I, 10, 63 quater ille beatus quo tenera irato flere puella potest. II, 1, 63 hinc et femineus labor est, hinc pensa colusque. II, 1, 67-8 ipse interque greges interque armenta Cupido natus et indomitas dicitur inter equas. II, 1, 79 a miseri, quos hie graviter deus urget! II, 1, 79-80 at ille felix, cui placidus leniter adflat Amor. II, 3, 19 o quotiens ausae, caneret dum valle sub alta, rumpere mugitu carmina docta boves! II, 3, 27 Delos ubi nunc, Phoebe, tua est, ubi Delphica Pytho? II, 3, 32 fabula sit mavult quam sine amore deus. II, 4, 11 nunc et amara dies et noctis amarior umbra est. II, 4, 45 At bona quae nee avara fuit. II, 5, 15 te duce Romanos numquam frustrata Sibylla. II, 5, 107 ars bona! II, 6, 13 iuravi quotiens rediturum ad limina numquam! III, 19 (IV, 13), 1126 tu mihi curarum requies, tu nocte vel atra lumen, et in solis tu mihi turba locis. Aside from the above in the elegies of Tibullus, the following were found in the Corpus : III, 1, 20 ; III, 1, 26 ; III, 1, 27 ; III, 2, 5; III, 2, 5-6; III, 4, 30; III, 4, 51-2; III, 4, 83; III, 4, 94; III, 6, 19; III, 6, 43; III, 7 (IV, 1), 9-10; III, 7 (IV, 1), 25; III, 7 (IV, 1), 25-6; III, 7 (IV, 1), 32; III, 7 (IV, 1), 37; III, 7 (IV, 1), 40; III, 7 (IV, 1), 81; III, 7 (IV, 1), 87; III, 7 (IV, 1), 107 foil.; Ill, 7 (IV, 1) 180; III, 7 (IV, 1), 198-9; III, 9 (IV, 3), 7; III, 10 (IV, 4), 23; III, 15 (IV, 9), 1; III, 16 (IV, 10), 3-4. 25 Text uncertain. 26 Assuming that it may be by Tibullus. 182 University of California Publications in Classical Philology. [Vol. 2 In the list of ellipses in Tibullus, a large number occur where a form of esse either immediately precedes or follows, e.g. I, 7, 44 ; I, 10, 9 ; II, 1, 63 ; II, 3, 27 ; II, 3, 32 ; II, 4, 11 ; II, 4, 45, and here may be classed (as following I, 3, 5) I, 3, 7 and I, 3, 9. A second group comprises instances of the omission of esse in compound verb-forms, e.g., I, 10, 3; II, 1, 67-8; II, 3, 19; II, 5, 15; and II, 6, 13. A third group comprises those cases wherein a predi- cate noun or adjective occurs, e.g. I, 1, 75; I, 3, 43; I, 7, 9 (bis) ; I, 10, 26; I, 10, 63; II, 1, 79; II, 1, 80; II, 5, 107; and III, 19 (IV, 13), 11. A fourth group comprises those instances where the construction demands a form of the verb esse (e.g. dative of the possessor) ; such we find in I, 3, 5; I, 4, 23; and I, 6, 33. These four groups comprise all instances of the ellipsis of esse in Tibullus. Restated they are : 1. "When esse immediately precedes or follows. 2. Compound verb-forms. 3. "When a predicate adjective or noun occurs. 4. "Where the construction demands a form of esse. As none of these conditions holds true in the verse under con- sideration, it seems to be impossible to admit of an ellipsis of esse here. The following table 27 shows the position of est in the penta- meter in the elegies of Tibullus : Elision No elision Before diaeresis 928 132 Begins second foot 429 233 After diaeresis 634 Ends verse 430 Ends first foot 131 Begins verse 435 18 13 27 Based on Killer's text. 28 1, 1, 22; I, 1, 34; I, 4, 32; I, 5, 68; I, 8, 76; II, 1, 46; II, 3, 24: II, 5, 36; III 19 (IV, 13), 2. 2» I, 2, 16; I, 2, 42; I, 9, 20; III, 19 (IV, 13), 4. so I, 4, 4; I, 10, 4; II, 6, 10; II, 6 44. si I, 5, 46. 32 II, 3, 74. 33 II, 3, 2; II, 4, 52. 34 I, 3, 36; I, 6, 66; II, 3, 16; II, 3, 36; II, 4, 24; III, 19 (IV, 13), 16. 35 I, 6, 44; I, 8, 64; I, 9, 24; II, 1, 30. 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 183 The most common position of est in the pentameter of Tibul- lus where it causes elision, is just before the diaeresis; this, how- ever, would put nam in the fourth place in its clause, a position in which it is apparently not found in classical Latin. Next in frequency are the position at the end of the verse, and that at the beginning of the second foot. The latter being impossible, if we retain the MS reading, I have accordingly placed est at the end of the verse. This position moreover furnishes an easy explanation of its loss. Omissions of est are not unparalleled in the MSS of the Corpus Tibullianum: I, 1, 34, magno Fr. Par. 1. magno est A Par. 2. Ill, 2, 7 pudor est G. est om. A. Ill, 3, 20 invidia est Par. invida quae A. Moreover the ease with which it may be omitted at the end of the verse is shown in Propertius II, 14, 1 ; II, 18, 25 ; and II, 34, 55. The verse with this wording suggests, even more than in its usual form, Horace Epist. I, 18, 87 dum tua navis in alto est, frequently cited as parallel. Moreover in Terence Hauton Timorumenos 343-7 we find a resemblance in language : quod boni .... datur, fruare dum licet : nam nescias .... eius sit potestas posthac an numquam tibi. I, 6, 7 Proposed reading: ilia quidem per multa negat, sed credere durum est. This reading is that of all the MSS save that they unite in reading tarn for per. All the early editors followed the MS read- ing, but we see signs of dissatisfaction with it in Scaliger's sug- gestion of iam multa, Burmann's insimulata, and Santen's delicta. Heyne, however, was the first to object seriously to the reading, and he hazarded both sua furta and iurata. The former reading need not detain us, but the latter is decidedly worthy of consideration. It was suggested to Heyne by a careful read- 184 University of California Publications in Classical Philology. [Vol. 2 ing of Ovid's Tristia II, 447 foil., which verses repeat a consider- able portion of this elegy in words which are very like those of Tibullus. The first two verses of Ovid 's paraphrase run : credere iuranti durum putat esse Tibullus, sic etiam de se quod neget ilia viro. Heyne saw that no word corresponding to Ovid's iuranti ap- peared in this verse in Tibullus, and inserted the only form possible from the standpoint of sense and meter iurata, a form which is found, for example, in Propertius I, 8, 27. The word has met with the approval of L. Mueller, Baehrens, H. Bubendey, Belling, H. Magnus, 36 B. Maurenbrecher 37 and Postgate. 38 Hiller proposed and read mihi cuncta, 39 comparing III, 7 (IV, 1), 129, but cf. Belling, Prolegomena zu Tibull, p. 75. Cartault (1909) reads quam multa, though earlier 40 he favored iurata. The only reading 41 thus far proposed that is worth con- sideration is Heyne 's iurata, but to it I have two objections. 1. Palaeographical. How can the change of iurata to tarn multa be explained ? They have only their last two letters in common. We must note, moreover, that all the MSS unite in the reading tarn multa. 2. The view that some word corresponding exactly to iuranti must appear in Tibullus is not sound, for Ovid handles this passage in the main quite freely. To be sure, some verses are strikingly alike; compare e.g. verses 25-26 of this elegy of Tibullus : saepe, velut gemmas eius signumque probarem, per causam memini me tetigisse manum. with Ovid's version of the same (451-2) : saepe, velut gemmam dominae signumve probaret, per causam meminit se tetigisse manum. But even here, where the resemblance is so great, there are variations that cannot be explained as due to a change of person 36 Berliner philol. Wochenschrift, 5 (1885), 589. 37 Philol., 55 (1896), 450. 38 Cf. Classical Review, 9 (1895), 77. so But tarn multa in the Corpus poetarum Latinorum (1893). *° A propos du corpus Tibullianum. 4i F. Wilhelm (J. P. P., 151 [1895], 114) and G. Friedrich (Wissen- schaftliche Beilage sum Jahresbericht des evang. Gymnasiums in Schweid- nitz, Ostern, 1898) defend tarn multa. 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 185 or the exigencies of meter. "Why, for example, did Ovid write -ve instead of -que? "Why did he not retain the plural gemmasf And if we pass from this to the passages where the correspondence is not so great, we can easily see the freedom with which Ovid has handled his original. Thus Tibullus reads in verses 19-20 : neu te decipiat nutu, digitoque liquorem ne trahat et mensae ducat in orbe notas. but Ovid (in 453-4) as follows: utque refert, digitis saepe est nutuque locutus, et tacitam mensae duxit in orbe notam. It should be noted that neu te decipiat has no expression corre- sponding to it in Ovid's lines, that nutu is joined with the following clause, and that digito .... liquorem .... trahat appears as digitis . ... est ... . locutus; moreover words are added, some to be sure to keep the original speaker in mind, as utque refert, but others with no such purpose, as saepe and tacitam. The freedom of Ovid's treatment being thus apparent, we have no right to demand that some form of iurare appear in Tibullus ; if we had, should we not similarly try to insert negat, or some form thereof, in the corresponding verse of Ovid? All we can say is that the one word, iuranti, in Ovid corresponds roughly to the clause ilia quidem .... multa negat in Tibullus ; in other words, though iurare itself may not appear in Tibullus, the idea of an oath should naturally suggest itself. The reading per multa, which is now proposed, was suggested by Tibullus I, 2, 38 : perque deos omnes se meminisse neget. In other words, Tibullus uses the expression per aliquem negare, from which of course it is but a step to per aliquid negare. In his comment on the above verse of Tibullus, Heyne says : " h. 1. iuret per deos omnes se non meminisse," that is to say, negare per is a synonym of iurare per .... non. i2 The Corpus Tibullianum is exceedingly fond of the use of per in asseverations : 42 Cf. Martial XI, 94, 7: "Ecce negas iurasque mihi per templa Tonan- tis. ' ' 186 University of California Publications in Classical Philology. [Vol. 2 I, 5, 7-8 parce tamen, per te furtivi foedera lecti, per Venerem quaeso compositumque caput. I, 4, 25-6 perque suas impune sinit Dictyima sagittas adfirmes, crines perque Minerva suos. II, 6, 29 parce, per immatura tuae preeor ossa sororis. So also III, 1, 15-16, III, 6, 47-8, and III, 11 (IV, 5), 7-8. The use therefore of per in asseverations in general and with negare in particular is paralleled in Tibullus. What then would per multa negare mean? Consider such a plea as that in Virgil Aen. IV, 314-8 : per ego has lacrimas dextramque tuam te (quando aliud mihi iam miserae nihil ipsa reliqui), per conubia nostra, per inceptos hymenaeos, si bene quid de te merui, fuit aut tibi quicquam dulce meum, miserere domus labentis. Or turn to that in Silius Italicus V, 82 foil. : Iliacas per te flammas Tarpeiaque saxa, per patrios, consul, muros, suspensaque nostrae eventu pugnae natorum pignora, cedas oramus superis. If we desired to describe these appeals briefly, should we not say that in each case the speaker ' ' per multa oravit ' ' ? We have moreover the oath in the Aeneid XII, 197 foil. : haec eadem, Aenea, terram, mare, sidera, iuro Latonaeque genus duplex Ianumque bifrontem, vimque deum infernam et duri sacraria Ditis. There is also that in Juvenal XIII, 78 foil. : per Solis radios Tarpeiaque fulmina iurat et Martis frameam et Cirrhaei spicula vatis, per calamos venatricis pharetramque puellae perque tuum, pater Aegaei Neptune, tridentem; addit et Herculeos arcus hastamque Minervae quidquid habent telorum armamentaria caeli. These oaths could readily be summarized by saying "per multa iuravit." It can hardly be objected that there are absolutely no examples in Latin of asseverations where per is followed by the neuter of an adjective or pronoun used substantively, for we find in Servius on Aen. X, 45: "Sallustius in primo postremo ipsos colonos per miserias et incerta humani generis orare." Per incerta is parallel to per multa. Somewhat analagous are also 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 187 Stat. Theb. XI, 368-9 per si quid in ilia dulce domo and Quint. Declam. TV, 22 per ego, si fas est, quicquid feci, .... We find also in Auct. ad Herennium IV, 52 : 'parce, ' inquit, 'et per quae tibi dulcissima sunt in vita, miserere nostri. '« The change palaeographically is not a great one. Multa is retained ; per, through the influence of the following word, might readily have converted r into m. Both words are of exactly the same length ; and the interchange of p and t appears also e.g. in Tibullus I, 10, 36 where for puppis A (pupis and pauppis, Par.) turpis, the reading of £, is accepted by scholars. I, 9, 25 Proposed reading: ipse deus tacito permisit vina ministro, ederet ut multo libera verba mero. MSS leve AV (lene Vm2), Iqne G. lena, laeva, leva, seva, saeva £. Readings : lingua Eigler, followed by L. Mueller,*-* Haupt-Vahlen, Belling (Tibullus), and F. Wilhelm.45 verba and in verse 26 lingua for verba — Francken. 4 ** nonne suggested by Lachmann. saepe proposed by Muretus: followed by Hiller." 4 ? . - lene Aldus, Sessa, Muretus, Grasser, Maittaire, Lachmann, Gruppe, Baehrens. vela Scaliger, followed by Broukhusius and Volpi. vela magister Guyet. laeva Voss. frena Burmann II and Huschke. lora Santen. lena Statius, Passerat, Heyne, Huschke, Golbery, Dissen, and Kemper. laeve Nemethy. tormentum admovit lene ministro** Cartault (1909). 43 So Baiter and Kayser: Orelli reads "per ea quae, etc." 44 L. Mueller ascribes this reading to exc. Par. « N. Jahrb. f. Phil. u. Paed., 151 (1895), 769. ^Mnemos. n.s., 6 (1878), 187. 4 ? But in the Corpus poetarum Latinorum (1893), leve. «F. Jacoby, Berliner philol. Wochenschrift, 29 (1909), 1467, calls it die plumpe Heriibernahme aus Horat. c. Ill, 21, 13." 188 University of California Publications in Classical Philology. [Vol. 2 Postgate impales lene as corrupt, suggesting that lenae was perhaps written as an explanation of ministro; but he proposes as a possible reading vina. Prior to seeing Postgate 's note, this .reading had appeared to me the correct one, and the fact that it has suggested itself to so eminent a scholar has of course intensified my belief in the soundness of vina. 49 If we assume that lene, whether as adjective or adverb, is impossible here, we have in vim a word that fits the thought admirably. The distich translated would read: "The god him- self granted wine to the silent servant, to cause him to utter free- spoken words through the influence of abundant liquor." The expression ipse deus .... permisit vina is paralleled by Ov. Fast. I, 403 vina dabat Liber. That both vinum and merum are found in the same distich is not in any wise abnormal in Tibullus : I, 2, 1 Adde merum vinoque novos compesce dolores. I, 5, 37-8 saepe ego temptavi curas depellere vino: at dolor in lacrimas verterat omne merum. So also Prop. II, 33, 31-32 : tuque, o Eurytion, vino, centaure, peristi, nee non Ismario tu, Polypheme, mero. Both of the other instances of permittere in the Corpus Tibullianum 50 show the verb with an object, though of course the expressions" are not parallel. "While there are no instances in Tibullus of permittere with an object and also a clause of pur- pose, yet he shows several instances of dare with both : I, 8, 29-30 det munera canus amator, ut foveat molli frigida membra sinu. I, 6, 13 tunc sucos herbasque dedi, quis livor abiret. For a concrete noun as object of permittere, may be compared Lucan VII, 123-4 : Sic fatur et arma permittit populis. Palaeographically the change is not difficult, even if we feel by no means certain of Postgate 's theory that lenae was a gloss on ministro. Vina and leve (reading of AV) are words of exactly 40 For Postgate 's discussion, see Classical Beview, 19 (1905), 213-4, and 23, 186-7. so HI, 7 (IV, 1), 92, and III, 16 (IV, 10), 1-2. 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 189 the same length. Confusion of initial I and v is also found in the MSS of Lucr.' Ill, 95 (locaturn — Marullus; vocatum — OQ) and Lucr. V, 12 (locavit — L; vocavit — OQ). For the ease with which e and i are interchanged, compare : I, 2, 65 posset — G; possit — A. I, 2, 79 magni — A; magnae — others. I, 4, 9 fuge te — others; fugite — Fr. T, 4, 40 vincit — f; vincet — AVG. The v and n are easily interchanged ; note the various MS read- ings of this very word, as well as I, 8, 2 lenia V, levia A, and I, 8, 57 lenis G, levis A. The final a for e is found in the reading of several of the inferior MSS. In I, 1, 5 the same error is found, vita, P. Fr., apearing as vite in A. Moreover the other readings proposed are not satisfactory. Lena and lingua introduce a very peculiar word-order : we should then have the subject of the subordinate clause inserted in the main clause. There are, it is true, examples throughout the Corpus Tibullianum of the deferring of the conjunction 51 of the subordinate clause. 52 But the only instances I have noted wherein one or more words belonging to the subordinate clause are placed before a word belonging to the main clause, are the following : 53 (1) Where the main clause consists of but one word (a verb) and no conjunction 54 is employed : I, 2, 12 capiti sint precor ilia meo. 67 ille licet Cilicum victas agat ante catervas. 3, 83 at tu casta precor maneas. 4, 53-4 rapias turn cara licebit oscula. 6, 56 sit precor ilia levis. 9, 40 sit precor exemplo sit levis ilia tuo. 9, 49 ilia velim rapida Vulcanus carmina flamma etc. Ill, 6, 26 quid valeat laesi sentiat ira dei. si See Emiliiis Schuenke, Be traiectione coniunctionum et pronominis relativi apud poetas Latinos, Kiliae, 1906. It does not, however, deal with books III and IV. 52 Goldbery defends the word-order (reading lena) by Hor. Serm. I, 1, 88, which is not at all conclusive: (1) because it is only an instance of the trajection of the relative, and (2) were it an even more complicated order, such order in Horace would not justify its introduction into the text of Tibullus. 53 Of course sentences are not included wherein the whole subordinate clause precedes the main clause. 54 We have, however, an interrogative pronoun in the verse from Lygdamus. 190 University of California Publications in Classical Philology. [Vol. 2 This use is quite characteristic of the Corpus Tibullianum and might be a development of the purely parenthetical use, as in II, 3, 74 mos precor ille redi. III, 12 (IV, 6), 8 sed iuveni quaeso mutua vincla para. (2) This isolated case : , I, 4, 25-6 perque suas impune si nit Dictynna sagittas adfirmes, crines perque Minerva suos. Leaving this passage for the present, we find besides the follow- ing instances of peculiar word-order : II, 3, 14 quidquid erat medicae vicerat artis amor. III, 16 (IV, 10), 5-6 solliciti sunt pro nobis, quibus ilia doloris ne cedam ignoto maxima causa toro. In these last two instances, however, instead of a thrusting for- ward of a portion of the subordinate clause into the main clause, we have the reverse condition, i.e. where a word of the modify- ing clause (which precedes the clause on which it depends) is delayed and placed in the main clause; but in both instances of the delayed word, a modifying adjective (medicae, II, 3, 14, and ignoto, III, 16 [IV, 10], 6) causes the mind to remain in suspense, waiting for the noun each is to modify (artis, II, 3, 14, and toro, 111,16 [IV, 10], 6). I, 4, 25-6, to recur to that passage, resembles class 1, cited above, in that the subordinate clause has no conjunction; more- over, the only portion of the subordinate clause that has thrust itself forward is per suas : the mind, as in the two instances just discussed, waits for the noun which suas is to modify, which appears in sagittas, the first word of the subordinate clause. Nowhere, therefore, in the Corpus Tibullianum do we find a single word taken from out of the subordinate clause (and in fact the subject of that clause), and embedded in the main clause — a condition which would be demanded by reading lena or lingua. Nonne never appears in the Corpus Tibullianum; -ne is always appended to the first word in the sentence and verse, save in III, 11 (IV, 5), 20, where it appears twice in an indirect ques- tion, but in its usual position in the clause. The chief difficulty in the reading saepe (and it seems to me insuperable) is the palaeographical one. But it may also be 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 191 noted that its position is unusual. The word is found in the Corpus Tibullianum 55 nineteen times, 56 of which eighteen are in books I and II, and one in III, 6, 4 ; of these uses, eighteen precede the verb and only one follows it, this being I, 6, 21 where we have the expression quam saepe. (Note that the deferred quam cito in I, 4, 28 also follows its verb.) Again, of these nineteen uses, fourteen appear as the first word in the verse and word- group, two as the second word in the verse and sentence (I, 9, 61 and II, 5, 35) and only three later in the sentence and verse, of which two involve the expression quam saepe (I, 6, 21 and I, 8, 53) and the third is at the earliest possible position in its clause' (II, 3, 59). Consequently the use of saepe in Tibullus is decidedly against its employment here in this position in the sentence and verse, even aside from the palaeographical diffi- culties and Wilhelm's well-taken criticism 57 that it is superfluous when employed with permisit, which is here used, like iussit, as a gnomic aorist. Francken's insertion of verba here and the substitution of lingua for verba in the following verse appear changes altogether too violent to be probable. The fact therefore that the readings thus far proposed are so unsatisfactory, should make us welcome Postgate's suggestion of vina all the more. I, 10, 37 Proposed reading: illic {parcel) ustisque genis ustoque capillo errat ad obscuros pallida turba lacus. For the words suggested the reading of AVG is percussisque ; P, perscissisque ; and £, perculsisque. The bulk of the editions read as A or P, percussisque appearing, for example, in Muretus, Huschke, Maittaire, Lachmann, Dissen, Baehrens, and Postgate. Perscissisque is the reading of Guyet, Scaliger (editions of 1600 and 1607), Volz, Francken, 58 L. Mueller, Hiller, Ramsay, and ss III, 14 (IV, 8), 6 is too unsettled to cite in this list. 56 See the Index Verborum in Hiller 's edition. 57 N. Jahr. f. Phil. u. Pacd., 151 (1895), 769. ssphilol, 28 (1869), 573. 192 University of California Publications in Classical Philology. [Vol. 2 Schulze. In the editions before 1600, Scaliger read percisisque. Heinsius' conjecture of exesisque was adopted by Broukhusius, Heyne, and Becker. Voss, Bach, and Bauer read perculsisque. Livineius and Dousa P. suggested pertusisque, which Postgate introduced into his Selections. Lachmann's conjecture, rescissis- que, 59 was approved and accepted by Gruppe, Haupt-Vahlen, 60 and Jacoby. Goerenz read peresisque, which is unmetrical. Moreover Eabus proposed percoctisque ; Belling, praescissisque; B. Fabricius, exustisque, which was also read by G. Nemethy, 61 who changed in addition ustoque to tostoque. Nemethy had previously (1905) proposed perfossisque. Many scholars accept the readings named with reluctance. Thus Ramsay would prefer exesisque, save that "it involves a great change in the text." Wratislaw and Sutton declare per- cussisque corrupt, but add that "no satisfactory emendation has been proposed." Dissen also calls it "corrupta lectio" and feels that such a word as exesisque gives the sense required : ' ' consilio loci accommodatissima, pr. excavatis, absumta carne per ignem rogi." To take percussis as "struck with fear" is bold; moreover the linking of such a thought with usto capillo is exceedingly un- natural; to give it the meaning "driven in, i.e. hollow, sunken" seems entirely unwarranted. As to perscissis, it would be appro- priate if used of the mourners, not of the dead themselves. In the same way percussis with its literal meaning would fit the mourners but not the dead. Voss' interpretation of perculsis, "mit Wangen voll Todesangst, d.i. blass und verzerrt, " is ludicrous. Exactly the thought that such scholars as Dissen suggest, "excavatis, absumta carne per ignem rogi," would be obtained by reading ustis. "Parce!" is the expression falling from the poet 's lips as the grewsome picture presents itself to his mind. "There (ah, spare me!) with cheeks burned away and hair burned away, the wan throng wanders by the dark pools. ' ' so Cf. Eleinere Schriften, II (1876), 147. «o In the edition of 1904 percussisque appears. «i Cf. Ehein. Museum, 64 (1909), 471. 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 193 No instance of ustae genae has come to my notice, but a rather interesting parallel is found in Propertius III, 12, 26 : exustaeque tuae mox, Polypheme, genaefi* One may cite Propertius IV, 1, 44 for some points of similarity : et verita est umeros urere flamma pios. Moreover, the burning of the funeral-pyre is also referred to in Propertius III, 15, 46 : , te solam et lignis funeris ustus amem. Propertius IV, 7, 7-9 (of Cynthia's ghost) is sometimes cited with reference to this passage : eosdem habuit secum quibus est elata capillis, eosdem oculos lateri vestis adusta fuit, et solitum digito beryllon adederat ignis. It must be noted, however, that the very first verse of this quota- tion contradicts our usto capillo. In Tibullus ' elegy the horrors of death are portrayed, and the picture that he is seeking to present is of atra mors. Therefore he represents the pallida turba of ghosts wandering by the obscuros lacus, mere skeletons with hair and flesh burned from them. It is not the dead individual whose loved image is being recalled as it was in the prime of life or as last seen on earth, but the throng of the dead, each as the flames of the funeral- pyre have left him. And as this horrible picture comes before his mind, Tibullus shrinks back and pleads ' ' parce ! ' ' Just so he pleads for escape from death in I, 3, 51, parce, pater. The picture here being more grewsome, the "parce!" falls from his lips even before he depicts the scene. The words parce! and parcite! are frequently found in Tibul- lus : in books I and II 63 we find thirteen instances as compared with five in all of Propertius. 6 * As an appeal for protection we find at milii parce, Venus (I, 2, 97). In I, 3, 51 we have just seen that it is used with reference to death; there mihi is also omitted as here. The dative is likewise omitted in I, 4, 83 ; I, 5, 7 ; I, 9, 5 ; and II, 6, 29. Of these I, 5, 7 most nearly resembles the 62 Here genae = " die Augenhohlung oder das Auge" (Eothstein). 63 Cf. Index Verborum in Hiller's edition. 64 Cf. Phillimore, Index Verborum Propertianus. 194 University of California Publications in Classical Philology. [Vol. 2 proposed reading, because not only is the dative omitted, but no vocative is expressed : paree tamen, per te furtivi foedera lecti per venerem quaeso compositumque caput. In II, 6, 29 also, both vocative and dative are omitted with parce, but as the vocative, dura puella, appeared in the preceding verse, its omission is less striking. One does not have to search far in Tibullus for examples of such repetitions as ustis — usto. Of this very verb, urere, there are three examples in Tibullus : II, 4, 5-6 et seu quid merui seu quid peccavimus, urit. uror, io, remove, saeva puella, faces. II, 2, 3 urantur pia tura focis, urantur odores. I, 9, 15 uretur facies, urentur sole capilli. Of these three examples, the first and last are interesting as not employing urere in precisely the same form in the two uses ; this is also the case in the proposed reading. No example of elided parce appears in the Corpus Tibul- lianum, but we find parce oculis in Propertius IV, 9, 53 and similar elisions in Juvenal VIII, 117 and Seneca Her. Get. 1447. Sentences of command limited to the imperative appear in II, 1, 87 (ludite), II, 3, 79 (ducite), II, 5, 121 (adnue). Parenthetical expressions we find for example in I, 3, 26 (an elided memini) and I, 8, 69 (moneo). Moreover the use of such an expression to impart vividness appears in II, 2, 10 : en age (quid cessas? adnuit ille) roga. The change from (parce!) ustisque to percussisque is very easily explained palaeographically. The moment the scribe failed to see that he had two words, and regarded them as one, 65 it was easy for the first syllable of parce to be altered to per. The interchange moreover of a and e is frequent in the MSS. To change t of ustisque to another s was natural, owing to the presence of this letter twice in the word, and the fact that the word percussis is a common one. Tibullus then represents a ghastly throng as wandering by 65 For examples of similar errors in incorrect division into words, see page 216. 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 195 the pools of the lower world ; in other words, his picture of the dead is closely akin to that in Ovid Ibis 143-4 : turn quoque factorum veniam memor umbra tuorum, insequar et vultus ossea forma tuos. II, 2, 17-22 Proposed reading : vota cadunt : 'utinam strepitantibus advolet alis flavaque coniugio vincula portet Amor, vincula, quae maneant semper, dum tarda senectus inducat rugas inficiatque comas.' hie veniat (Natalis, ades) prolemque ministret, ludat et ante tuos turba novella pedes. These verses read in A exactly as above (disregarding vinculaque manent) save that ades appears as avis. In but few of the editions is the MS reading retained, and, in addition to difficulties of reading, those of interpretation also arise in con- sidering these verses. Utinam is read by most editors, but viden ut, the conjecture of Guyet and Heinsius (suggested by II, 1, 25) appears in the editions of Broukhusius, Volpi, Heyne, Voss, Huschke, Golbery, L. Mueller, and Postgate. 66 Baehrens proposed ut iam and altered to advolat and portat; these readings Woltjer and Jurenka followed. The inferior reading, trepidantibus, which Broukhusius in- troduced into his text, is now universally rejected and need not detain us. On the score of a reading cadant £, Haupt proposed vota cadant utinam! str., etc. Postgate also independently made this suggestion, 67 but later 68 rejected it and proposed vota cadant tibi, nam. Verse 21 is the most altered of all the verses. The MS read- ing hie (hec G) veniat natalis avis prolemque ministret, is fol- 66 Selections. In his complete edition, Postgate reads utinam, comment- ing however "quod vix sanum. " Dissen, while not placing viden ut in the text, approves of it as "praeclara coniectura. " 67 Journal of Philology, 25 (1897), 51. C8 Journal of Philology, 26, 184 foil. 196 University of California Publications in Classical Philology. [Vol. 2 lowed by Goldbery and Haupt-Vahlen, 69 and, with a change to prolesque, by Rothstein. 70 The MS reading with the capitaliza- tion of Nat alts, is followed by Lachmann 7] and Gruppe. Drenck- hahn would punctuate hie veniat natalis, avis prolemque minis- tret. The earlier scholars, e.g. Muretus and Scaliger, read hue veniat natalis avis, prolemque ministret. Heinsius proposed hue veniat natalis, avis prolemque ministret, and hac veniat natalis avi prolemque ministret, and also the latter with venias and ministres. The last of these (with a capitalization of Natalis) was accepted by Broukhusius, Heyne, 72 Voss, the Del- phin Classics, and Becker. This conjecture, hac .... avi, but with the verbs retained in the third person, is approved by Huschke, Dissen, Hiller, L. Mueller, Jacoby, 73 and Jurenka. Bauer reads hue venias, Natalis avis, prolemque ministres. Baehrens (in Tib. Bl., p. 89) suggests hie veniat Natalis avis (avof) prolesque ministret Ludat ut ante, etc.; in his edition, however, we find haec veniat genialis avis; prolesque ministret, in which Woltjer follows him. Graef thinks the dfstich spurious or that the genuine words of the hexameter (perhaps interea, Natalis, ave prolemque ministra) have been very seriously cor- rupted. Eveniat was Housman's suggestion for the beginning of the verse; Postgate's haec {hec G) valeat (which he incor- porated in his Selections) 7 * and haec veniat, Natalis, avis, etc., which appears in his complete edition. Belling read sic 75 veniat natalis avis, prolesque ministret. In the final verse, et found in practically all of the editions is converted into ut (£) by Voss, Huschke, Baehrens, and Woltjer ; and Dissen also approves of it. «9 H. Magnus, Berliner philol. Wochenschrift, 13 (1893), 1550 declares the MS reading "gewiss nicht unecht. " to Cf. Bursian Jahr., 51, 336 foil. 7i Cf. Eleinere Schriften, II (1876), 147. 72 Heyne was, however, dissatisfied with the reading and proposed: "Sic venias, Natalis, avis" or "Hie veniat natalis avis, prolemque minis- tret" (the MS reading). 73 Jacoby: "1st avis nicht vielleicht nur durch Dittographie von -alis entstanden und hat ein anderes Wort verdriingt?" 74 Cf. Journal of Philology, 26 (1898-9), 184 foil. In the Classical Be- view, 9 (1895), 74-8, he favored prolesque ministret . . . . ut. 70 E. Ehwald, Deutsche Litteraturzeitung, 16 Jahrg. (1895), 937-40, had proposed: "sic veniat natalis avis." See also Philol., 54 (1895), 459. 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 197 "Utinam," according to Postgate, "is clearly wrong, as the appearance of Love is the sign that the prayer is granted ' ' ; and similarly Broukhusius declares: "post vota rata, quid hie faciat utinam, optandi particula, vix bene perspici potest." It is such objections that have caused suspicion to be cast on, or changes to be made in utinam. Let us rather, however, take the text as it stands, and con- sider the meaning of the individual words from their context. Vota would of course be the prayers of Cornutus, those which he was urged to make in verse 10 and the contents of which the poet foretells (note the tense of optabis) in verses 11 to 16 in- clusive. Passing over cadunt for the time being, we come to utinam. Here we evidently have the vota themselves. 76 That is, the long-delayed prayers of Cornutus begin with the word utinam and continue to the end of verse 20. As we are waiting for the prayers, vota at once brings them sharply before the mind and utinam introduces them. What now of cadunt f It evidently cannot mean "frustranea et irrita sunt," 7 - 7 for this would make their subsequent statement entirely out of place, and would moreover be contradictory to verses 9-10 : adnuat et, Cornute, tibi quodcumque rogabis. en age (quid cessas? adnuit ille) roga. But, with very few exceptions, cadunt is interpreted rata sunt, eveniunt, and yet, as has been stated, for the prayers to follow immediately on a statement that they "are realized," is hardly natural. The difficulty, it appears to me, is to be met by taking a different interpretation of cadunt, rather than by altering utinam. Does cadere ever mean precisely ratum esse? Becker (Elegeia Romana) says: "Scire tamen velim quo alio loco cadere i.e. accidere positum sit pro evenire, ratum esse (in Erfiillung gehen)." Is it not rather true that cadere when applied to such words as verbum, vox, votum, merely means "fall 76 Cf . Hor. Serm. II, 6, 59-62 : perditur haec inter misero lux non sine votis: o rus, quando ego te adspiciam? quandoque licebit nunc veterum libris, nunc somno et inertibus horis, ducere sollicitae iucunda oblivia vitae? Cf. also Hor. Serm. II, 6, 1. 77 Delphin Classics. 198 University of California Publications in Classical Philology. [Vol. 2 from the lips," and are not the meanings "be lost," "fail," or the like due to some word used with cadere that has tinged its meaning? Compare Horace Epist. I, 18, 12 sic Herat voces et verba cadentia tollit. E. C. "Wickham translates "picks up every word he lets drop." The verse from Horace gives us "verba cadunt" untinged by a modifying adjective, adverb, or phrase. "The words fall; he picks them up." Juvenal VI, 440 reads verborum tanta cadit vis, which is translated by John Delaware Lewis "such a power of words falls from her." If verba cadunt has this meaning, clearly vota cadunt can also. In Ovid Her. 3, 98 at mea pro nullo pondere verba cadunt, verba cadunt simply means "the words fall"; the idea that they are vain comes from "pro nullo pondere" ("as of no weight" — Palmer). Propertius contains three passages that are similar: I, 10, 24 neu tibi pro vano verba benigna cadant. I, 16, 34 at mea nocturno verba cadunt Zephyro. I, 17, 4 omniaque ingrato litore vota cadunt. In each of these the futility of the words or prayers rests not in cadere, but in pro vano, nocturno Zephyro, and ingrato litore; they are vain because they do not reach their goal 78 and fall either on the wind or the shore. In the only passage in Tibullus resembling that under dis- cussion, I, 6, 85, we find : haec aliis maledicta cadant. The word aliis mentioning the goal that the maledicta are to reach, serves to differentiate this passage from that under con- sideration. Aside from that however, "May these imprecations fall upon others ! " is perfectly clear, and while we may translate : "May these curses come true in the case of others!" the first rendering is satisfactory. 79 The meaning of cadere here defended is that first maintained (as far as known to me) by Becker (Elegeia Bomana), so who 's Eothstein on Prop. I, 10, 24. "9 Cf. Becker, Elegeia Bomana. so A view similar to this is put forth by Belling, Wochenschrift f. Philol., 15 (1898), 459, who however makes no mention of Becker's statement. Thus Belling: "vota cadunt — Es muss heissen 'werden ausgesprochen. He does not discuss the following verses. > > t 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 199 suggests that cadunt here means " pronuntiantur, funduntur. " In so far I agree with Becker; he appears, however, to take cadunt as an historical present, and thus deals with the following verses: "lam, inquit, pronunciata sunt vota. Utinam iis re- spondens advolet quam celerrime Amor, portetque felicissima coniugii vincula." If, however, one examines carefully the sentence introduced by utinam, he will find that he has here the prayer in expanded form that was foretold in verse 11 : auguror, uxoris fidos optabis amores. The idea in optare is of course found in utinam and the follow- ing subjunctives; amores is expressed in advolet .... Amor; uxoris we find altered to coniugio, and fidos is amplified into vincula, quae maneant semper, etc. In other words, we have in full that prayer which was anticipated and briefly summarized. It is not, then: "The prayers have been uttered; etc.," but rather : ' ' The prayers come : ' May Love fly, etc. ' ' ' This prayer is concluded at the end of verse 20 and then the poet continues : hie veniat (Natalis, ades) prolemque ministret. Hie can only refer to Amor whose coming has just been prayed for; that is, the poet's "hie veniat" is his own reiteration of Cornutus' "advolet .... Amor." The same pronoun, hie, is the subject of ministret. The term ministrare, as applied to the god Amor, Tibullus used in I, 10, 57 at lascivus Amor rixae mala verba ministrat, and it also appears in III, 12 (IV, 6), 12 fallen- dique vias mille ministret Amor. That moreover Amor should be called on to bring offspring, is in harmony with II, 1, 83-4, where he is to be called to the herd to make it prolific : vos celebrem cantate deum (Amor, v. 80) pecorique vocate, voce palam pecori, clam sibi quisque vocet. The change of the MS reading avis to ades is in itself but a slight one, and is also in some measure defended by the fact that the reverse mistake {ades written instead of avis) is found in the MSS of Tibullus, II, 1, 34. 81 Ades here of course is equal to fave : and it is used after a prayer or wish, as in III, 3, 31-33 : 81 Cf. also Jacoby's note, on page 196 of this paper. 200 University of California Publications in Classical Philology. [Vol. 2 haec alii cupiant; liceat mihi paupere cultu securo cara coniuge posse frui. adsis et timidis faveas, Saturnia, votis. The direct address to the deity after the prayer, and the refer- ence to vota are parallel to this passage. The use of adesse closely associated with some form of votum is found e.g. in Ovid Fast. Ill, 256 ; Trist. Ill, 1, 78 ; Her. VI, 151-2 ; Seneca Here. Fur. 645-6 ; Phaedra 423 ; so too with precibus, Ovid Am. II, 13, 21. That in a poem dealing with a birthday, the poet should at the end recur to the Natalis, addressing him directly, is closely paralleled by III, 11 (IV, 5), where verses 19 and 20 return to a direct address to Natalis : at tu, natalis, quoniam deus omnia sentis, adnue. In this passage moreover adnue is parallel to ades; and, besides, these verses conclude a prayer for mutual love as in our elegy : (verses 7-9) mutuus adsit amor, per te dulcissima furta perque tuos oculos per geniumque rogo. mane geni, cape tura libens votisque faveto. Verses 13-16 contain a reference to the vinculo that in our pas- sage Love is to bring. Ill, 12 (IV, 6) is addressed to Natalis luno; in it verses 13 and 14 bid dea casta, adnue, and verses 7 and 8 are closely parallel to portions of the elegy under discussion : at tu, sancta, fave, neu quis divellat amantes, sed iuveni quaeso mutua vincla para. In I, 7 (one of Tibullus' own elegies) the last distich makes a direct address to Natalis : at tu, natalis multos celebrande por annos, candidior semper candidiorque veni. The coupling of the appeal to Natalis and the prayer for progeny also bears some resemblance to I, 7, where we find in I, 7, 55 at tibi succrescat proles, etc., and in 63^1, just quoted, an appeal to the natal god. The appeal is perfectly natural in the light of the following statement in Roscher's Ausfiihrliches Lexicon der Griechischen und Edmiscken Mythologie, I, 1615: "Dem Genius lag dem Gesagten gemass vornehmlich die Tutel 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 201 des Ehebettes ob, welches nach ihm standig lectus genialis hiess (so an vielen Stellen: bei Catull 64, 47 aueh pulvinar geniale; torus bei Liv. 30, 12, 21 ist unsicher). Nuptiis sternitur in honorem genii erklart Festi epit. 94. . . . Beim Laberius (v. 54 R.) heisst derselbe generis nostri parens d.i. 'das die Familie von einer Generation zur andern erhaltende Prinzip.' (Preller)." Cf. Arnobius 2, 67 cum in matrimonium convenitis, toga sternitis lectulos et maritorum genios advocatis . . .'.! And so too Preller (Rom. Mythol. 1, 78) speaks of the lectus genialis "wo der Genius der Familie segnend und befruchtend waltet, dass es dem Hause nie an Kindern fehle." And finally with the interpretation thus far given of the pas- sage, tuos (v. 22) would clearly refer to Natalis just preceding. II, 3, 61 Proposed reading: at tibi, dura, (seges, Nemesis, quia ducit ab urbe) persolvat nulla semina terra fide. In this distich the reading of A is : at tibi dura seges nemesis qui abduc 82 ab urbe, persolvat nulla semina terra fide. The reading of V is also nemesis, but in Vm2 and G we read nemesim. For qui, £ have quae. Instead of abduc, V reads abducit; £, abducis. Some inferior MSS read quia ducis, where A has qui abduc. All editors, as far as I know, read Nemesim or Nemesin 83 and change abducit or ducit to the second person. 84 (Rigler proposed quae abduxit). A typical reading is that of Hiller: at tibi dura seges, Nemesim qui abducis ab urbe, persolvat nulla semina certa fide. 82 So Hiller and Wilhelm: abducit say Baehrens and Postgate. See Wochenschrift fur Mass. Phil., 23 (1906), 1148. 83 In Berliner phil. Wochenschrift, 29 (1909), 1467, Cartault (1909)— whose work I have not seen — is said to read ' ' At tibi dura seges Nemesis qui abducit ab urbe. ' ' 84 Save F. Wilhelm, N. Jahrb. f. Phil. u. Paed., 151 (1895), 770, who defends abducit, its subject being qui which in turn refers to tibi, by Tib. I, 2, 33. This does not appear analogous. But in Rhein. Museum 59 (1904), 283, he discusses the passage, reading abducis. Belling (Prolego- mena, 62) approves of quia ducit, but in his "Tibullus" he reverts to quae abducis. 202 University of California Publications in Classical Philology. [Vol. 2 Tibi, thus taken, refers to the rich lover just mentioned. The reading ccrta (ascribed to Pucci) is a substitute for terra O. 85 The following variations in the readings are found : at changed to sit Rossbach, followed by Baehrens. seges changed to Ceres N. Heinsius, followed by Broukhusius, Joh. Schrader, Heyne, Voss, Golbery, Dissen, L. Mueller, and Belling. qui changed to quae (f) Aldus, Scaliger, Grasser, Broukhusius, Mattaire, Joh. Schrader, Heyne, Voss, Golbery, Dissen, Rigler, Belling, Postgate. quia ducis f quae ducis f Aldus, Grasser, Voss. Nemesim changed to Venerem Joh. Schrader. terra changed to certa Lachmann, Gruppe, Hiller, Haupt-Vahlen, L. Mueller, R. Ullrich, F. Wilhelm. For terra Statius and Canter proposed tecta. As has been pointed out, practically none of the editors keep Nemesis and abducit, and most of them make at least one other alteration in the MS reading. The reading proposed follows A in detail, save that quia (£) ducit appears for qui abducit; in other words, an effort is made to restore the text by dropping but a single letter of the MS reading and in one instance altering the division into words. The translation would run : ' ' But to thee, cruel one, because 'tis the crop, Nemesis, that leads thee from the city, may the earth prove false and fail to give back the seeds." The word quia appears in Tib. I, 4, 13, and the frequency of its use in Propertius 86 warrants it appropriateness in elegiac- poetry of this period. The trajection of conjunctions in Tibullus is exceedingly common, as this list 87 witnesses : 88 cum 1, 1, 47; I, 1, 59; I, 2, 65; I, 3, 9; I, 4, 33; I, 5, 12; I, 7, 21; I, 7, 61; I, 10, 8; II, 1, 47; II, 3, 29; II, 5, 14; II, 5, 109. dum 1, 2, 4; I, 2, 73; I, 3, 25; I, 3, 56; II, 3, 19. dummodo 1, 1, 58. modo 1, 2, 31. ss But cf. Postgate, Journal of Philology, 26 (1898-9), 186. so See Phillimore, Index Verborum Propertianus. 87 Derived from E. Schuenke, Be traiectione coniunctionum, etc., Kiliae, 1906. 88 This list is confined to books I and II. 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 203 ne 1, 1, 67; I, 3, 21; I, 5, 5; I, 6, 20; I, 8, 29; I, 9, 17. ni 1, 4, 63. qua II, 2, 16; II, 5, 96. quam (saepe) 1, 6, 21. qui (all forms of the relative) 1, 1, 50; I, 3, 7; I, 7, 3; I, 7, 13; I, 9, 24; I, 9, 72; I, 10, 1; I, 10, 6; II, 4, 10; II, 4, 45; II, 5, 16; II, 5, 40; II, 5, 68; II, 6, 5; II, 6, 8. quicumque 1, 2, 39; I, 4, 39; I, 10, 59; II, 2, 13. quod 1, 3, 79. quotiens II, 6, 13. si 1, 1, 44; I, 2, 71; I, 4, 15; I, 4, 49; I, 6, 23; I, 6, 33; I, 6, 74; I, 8, 61; I, 9, 19; II, 4, 33; II, 4, 53. ubi 1, 4, 31. ut 1, 1, 18; I, 2, 2; I, 4, 52; I, 6, 16; I, 9, 26; I, 10, 48; II, 1, 2; II, 1, 15; II, 1, 50; II, 3, 45; II, 5, 72; II, 6, 42. "While there are no instances of a postponed quia in Tibullus, it may be noted that the causal conjunction quod is found thus deferred. But the trajection of quia is sufficiently common in other authors to justify this position here. It is found in : Plautus As. 386; Cure. 225; Men. 513; Mil. 54; Trin. 1165. Lucretius 1, 169; I, 176; I, 221; II, 607; III, 278; III, 364; III, 746; III, 1070; IV, 92; IV, 241; IV, 355; IV, 694; IV, 1242; V, 357; VI, 349; VI, 353; VI, 841; VI, 909; VI, 1059. Terence Eec. 681. Cicero Aratea 12. Virgil Aen. VIII, 650. Horace Serm. I, 3, 92-3 and I, 9, 51; Epist. II, 3, 295 and II, 3, 376; Carm. IV, 9, 28. Aetna 284. In its opening as well as in the trajection, the verse resembles Propertius III, 18, 31-2 : at tibi, nauta, pias hominum qui traicis umbras, hue animae portent corpus inane tuae. In the trajection and the appearance of another vocative in the subordinate clause, Virgil Aen. XII, 179-180 is parallel: tuque inclute Mavors, cuncta tuo qui bella, pater, sub numine torques. Tibullus himself in II, 5, 113-4 has a distich that closely re- sembles the one under discussion : at tu, nam divum servat tutela poetas, praemoneo, vati parce, puella, sacro. 204 University of California Publications in Classical Philology. [Vol. 2 The beginning at tu resembles the at tibi under discussion ; in each instance we have the causal clause closely following ; and in each instance we have a delayed vocative. Moreover, in the same way in which the quia-cl&use here pre- cedes the major portion of the main clause, it appears in the only other instance of the word 's use in Tibullus, I, 4, 13 : hie, quia fortis adest audacia, cepit. Cf. also I, 3, 57-8 and Propertius IV, 2, 11 and IV, 10, 47. The object of ducit is so clearly apparent that it would seem needless to express it in this verse. So we find abducere without an object in III, 9 (IV, 3), 5: sed procul abducit venandi Delia cura. And still nearer at hand, we have in verse 79 of this very elegy : ducite: ad imperium dominae sulcabimus agros. If abducere should be thought preferable to ducere here, it would be well to notice that neither with abducit, III, 9 (IV, 3), 5, nor abducta, III, 14 (IV, 8), 7, is the locus a quo mentioned. Tibul- lus writes, however, (I, 2, 43) hanc ego de caelo ducentem sidera vidi, using the simple verb; and in the following instances Propertius uses ducere together with the locus a quo : II, 13, 6 aut possim Ismaria ducere valle feras. II, 1, 56 ex hac ducentur funera nostra domo. Durus, as applied to Nemesis, is perfectly appropriate, "nam" (as Pichon 89 says) "duri dicuntur qui amorem oblatum respuunt neque precibus commoventur." Thus in Tibullus I, 8, 50 we find in veteres esto dura, puella, 00 senes. And in II, 6, 28 Nemesis is directly called so : ei mihi, ne vincas, dura puella, deam. She is also called saeva puella in II, 4, 6. Other passages where mistresses are termed dura are: Propertius I, 1, 9-10 I, 7, 6; I, 17, 16; II, 1, 78; II, 22, 11; II, 22, 43; II, 24, 47 IV, 2, 23; Ovid Am. I, 9, 19; Her. XX, 5; Ars Am. II, 527 Fast. IV, 111 and VI, 120. While dura is not used substantively in Tibullus, yet note that it is in the last passage cited (Ovid Fast. VI, 120) : 89 Rene Pichon, De sermone amatorio apud Latinos elegiarum scriptores, Paris, 1902. From it the following instances are drawn. »o Note the sense-pause between dura and puella. So dura and seges in this passage. 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 205 viderat banc Ianus, visaeque cupidine captus ad duram verbis mollibus usus erat. Moreover the authors of the Corpus (including Tibullus himself) use other adjectives substantively in the vocative case, often indeed with forms of tu : I, 9, 65 nee tu, stultissime, sentis. II, 1, 81 sancte, veni dapibus festis. III, 8 (IV, 2), 3 at tu, violente, caveto. Ill, 10 (IV, 4), 9 sancte, veni. Ill, 12 (IV, 6), 7 at tu, sancta, fave. It may also be noted that in this very elegy (II, 3), we find in the immediate vicinity the following feminine forms of adjectives used substantively: formosas (65) and mea (77). As to the form Nemesis, Postgate 91 admits that the vocative in -is of such Greek words is found in prose and comedy, "but," he assumes, "it was not tolerated in the refined compositions of the Hellenizing poesy." The form Nemesi is, however, of course "metrically unavailable," and we do find the vocative Nemesis in Auson. Idyll. 8, 41 (III, 5, 41 in the edition of R. Peiper) mitibus audi auribus hoc, Nemesis. Moreover, the statement in Charisius I, 17 implies no such limitation: "Mysis o Mysis Terentius, ut o crinis funis cinis. Graeci demunt s litteram, nostri parem nominativo vocativum servant." And, most im- portant of all, in this passage Nemesis is the reading of AV. In the text as reconstituted seges assumes a very prominent position and justly so, for it is the wealth of the country, its crops and vintage, that lead Nemesis away from the city. And it is just because the wealth of the country has enticed her away that the poet prays that the earth may fail to send up a harvest. The method in which the text became corrupted is apparent : dura, standing next to seges, was taken to modify it; because tibi was supposed to refer to the lover just mentioned, it was easy for quia to be made into the relative qui, the antecedent of which was tibi. The a remaining, together with ducit, under the influence of ab urbe, easily became abducit. si Classical Review, 23, 186-7. Tbe vocative form is discussed in Neue- Wagener Formenlehre, 1, 443. 206 University of California Publications in Classical Philology. [Vol. 2 II, 4, 43 Proposed reading: Sed veniet tibi mors, nee erit qui lugeat ullus, nee qui det maestas munus in exsequias. Where the above reading proposes sed, all the MSS (as far as known to me) unite in the reading seu, nor is it known that seu has ever been questioned. Veniet has not escaped thus, however ; the reading of A (veniet) appears as veniat in VG, and this latter reading is accepted by Scaliger, Broukhusius, Voss, Wun- derlich, Bach, and Baehrens. Moreover Scaliger, Dousa, and Broukhusius alter nee erit to neque sit. 02 Belling, too, though not incorporating it in his text, approves of nee sit. 93 By these changes we should secure a succession of present subjunctives to harmonize with eripiant (40), spectent (41), and addat (42). Veniet (and far more, erit) has maintained its position (1) because the reading of A; (2) as Dissen points out, "neque enim dubium venturam mortem"; (3) because of its agreement with erit. 9i Reading then veniet, we should have the verse : seu veniet tibi mors, nee erit qui lugeat ullus. First of all, the verse seems suspicious because of the thought : "Or if death comes." 95 Bui death will come; how can it be stated conditionally? Tibullus could of course have easily referred to an early death conditionally, or a long-deferred death, but hardly to death in general. Thus in the former two cases, we find in the Panegyric, III, 7 (IV, 1), 205-6 : seu matura dies celerem properat mihi mortem, longa manet seu vita. »2 The earlier editions (e.g. Aldus and Muretus) have neque erit. »3 Philol., 47, 382. 94 Of the examples cited by Voss wherein a present subjunctive in the protasis is found coupled with a future indicative in the apodosis, Virg. Eel., IV, 58, "Pan .... mecum si ... . certet, Pan .... dicet se . . . . victum," should evidently be read dicat. In Prop. II, 26, 29, clearly the fact stated in the protasis is unlikely to come to pass (therefore present subjunctive); but should it come true, the conclusion would inevitably be realized (therefore future indicative). But in our passage the first is absolutely certain, is not in any sense unlikely. »6 Cranstoun translates: "When death shall come," but no authority is cited for this meaning. 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 207 But as to the inevitableness of death, listen to Tibullus I, 1, 70 : iam veniet tenebris mors adoperta caput. and I, 10, 34 : imminet [mors] et tacito clam venit ilia pede. The time may be doubtful, but never the fact : I, 1, 59-60 te spectem, suprema mihi cum venerit hora, te teneam moriens deflciente manu. So too III, 2, 9 foil, and III, 3, 9-10. 96 The only conditional expressions noted in the Corpus Tibullianum in connection with death are : I, 3, 53 quod si fatales iam nunc explevimus annos. Ill, 5, 31-2 vivite felices, memores et vivite nostri, sive erimus seu nos fata fuisse velint. But in neither of these is it doubted that death will surely come, the only element of doubt being that of its coming at a specific time. The second quotation, for example, reads: "Live happily and live remembering me whether (at that time) I shall be living or the fates should wish my life to be no more." In other words, "if I shall be living at a definite time" (expressed or implied) is quite a different thing from saying "if death shall come." The second objection to the verse lies in the word seu. Sive and seu are found in pairs 97 in the following passages in the Corpus Tibullianum : I, l, 11-12. Ill, 7 (IV, l), 45-6. I, 2, 17-18 III, 7 (IV, 1), 66. I, 2, 33. Ill, 7 (IV, 1),. 79-80. I, 10, 21-22. Ill, 7 (IV, 1), 95. II, 4, 5. Ill, 7 (IV, 1), 96. II, 6, 3. Ill, 7 (IV, 1), 101-3. III, 1, 6. Ill, 7 (IV, 1), 159-60. Ill, 1, 26. HI, 8 (IV, 2), 9-10. Ill, 3, 14. Ill, 8 (IV, 2), 11-12. Ill, 4, 11-12. Ill, 9 (IV, 3), 1-2. Ill, 5, 32. We find a group of three in III, 7 (IV, 1), 25-6 and a group of ss Cf. Prop. II, 13, 17, "quandocumque igitur nostros mors claudet ocellos. " 97 i.e. sive (or seu) used twice, or sive with seu. 208 University of California Publications in Classical Philology. [Vol. 2 five in III, 7 (IV, 1), 205-9. The only uses of a single seu 98 or sive are the following : I, 6, 21 exibit quam saepe, time, seu visere dicet sacra Bonae maribus non adeunda Deae. Ill, 7 (IV, 1), 94 seu libeat. The first (the only one in Tibullus) is to be translated "even if"; 99 the second, "or if." The first meaning is not possible here; it would take the fact of death, already conditioned so abnormally, and place it even farther in the range of improba- bility. The second is, of course, no criterion for the usage of Tibullus, as it occurs in the Panegyric ; moreover it would retain the conditional force which is so unnatural here. Draeger 100 cites the following examples of a single sive or seu = " or if" in the poets: Horace Od. I, 6, 19; I, 15, 24; III, 27, 61; Serm. II, 1, 59; perhaps Virgil Aen. XI, 327; Tibullus II, 4, 43 ; 101 IV, I, 93 ; Propertius III, 21, 8 and 33 ; IV, 6, 81 and Ovid Her. X, 97. The use is therefore a rather rare one, and in the meaning here suggested has no warrant in Tibullus' diction. The objections to seu are therefore two-fold : first and more important, as to sense, and second, as to the use of seu in Tibullus. For it the reading sed is proposed. This involves the change of but a single letter, and the alteration of the d can easily be accounted for by the initial letter of the following word. We should then have (after the wishes of verses 39-42) : "But death will come to thee, and there will be no one to mourn. ' ' In other words, the adversative conjunction sed with the future indica- tive, stating a fact as unmistakable after a previous wish, acquires the tone of "but however that may be," "but however my wish may turn out ' ' ; that is, it discusses the destined future state even if the wish be not fulfilled. For this use, compare Horace Serm. II, 1, 42-6 : »8 Excluding the passage under discussion, of course. 99 See Dissen's note. This use is classed by Draeger under the follow- ing: "Dies sive, welches fur vel si steht, ist zuweilen, aber nur in der Sprache der Juristen und der Dichter, einem voraufgehenden Satze oder Ausdrucke coordinirt, der zwar nicht die Form, aber den Sinn eines Bedingungssatzes hat." ioo Historische Syntax der Lateinischen Sprache, II, 149. ioi The passage here under discussion. 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 209 o pater et rex Iuppiter, ut pereat positum rubigine telum, nee quisquam noceat cupido mini pacis! at ille qui me commorit (melius non tangere, clamo), flebit et insignis tota cantabitur urbe. Somewhat analogous is Propertius I, 8, 9 foil. : O utinam hibernae duplicentur tempora brumae, et sit iners tardis navita Vergiliis, nee tibi Tyrrhena solvatur funis harena, neve inimica meas elevet aura preces! atque ego non videam tales subsidere ventos, cum tibi provectas auferet unda rates, ut me defixum vacua patiatur in ora crudelem infesta saepe vocare manu! sed quocumque modo de me, periura, mereris, sit Galatea tuae non aliena viae. Cf . also Lygdamus III, 5, 27-8 : atque utinam vano nequiquam terrear aestu! languent ter quinos sed mea membra dies. Ill, 4, 62-3 also bears on this use of sed: a pereat, didicit fallere si qua virum. sed flecti poterit: mens est mutabilis illis. We find an excellent parallel, though the adversative conjunc- tion is not expressed, in I, 9, 11 foil. : at deus ilia in cinerem et liquidas munera vertat aquas, iam mihi persolvet poenas, pulvisque decorem detrahet et ventis horrida facta coma; uretur facies, urentur sole capilli, deteret invaiidos et via longa pedes. The position of sed at the beginning of the sentence and verse is sufficiently justified by I, 10, 15 ; I, 10, 53 ; II, 1, 31 ; II, 4, 24, etc. As to the uniting of three coordinate members by nee between the first and second, and also between the second and third, one may readily compare the following: I, 1, 37-8 adsitis, divi, nee [neu A] vos e paupere mensa dona nee e puris spernite fictilibus. I, 1, 71-2 iam subrepet iners aetas, nee amare decebit, dicere nee cano blanditias capite. II, 3, 11-13 pavit et Admeti tauros formosus Apollo. nee cithara intonsae profueruntve comae, nee potuit'euras sanare salubribus herbis. 210 University of California Publications in Classical Philology. [Vol. 2 And again, for sed .... nee .... nee, a pretty close parallel may be found in this very elegy, II, 4, 33-4 : sed pretium si grande feras, custodia victa est, nee prohibent claves, et canis ipse tacet. In other words, by reading sed we remove the thought of death as conditional and a rather suspicious use of seu in Tibullus. We introduce a word, the loss of which may easily be accounted for palaeographically, and which fits the sense of the passage and is in harmony with Tibullus' usage. II, 5, 47 Proposed reading: ecce mihi lucent rutilis incendia castris. This reading is simply a restoration of the spelling of A and V. Rutulis, as spelled in £, 102 is found in every edition of Tibullus known to me. The first suggestion that another read- ing might be preferable is found in Statius: "In scriptis non- nullis, Rutilis. Forte rutila, fulgentia dixit, ut Horat. in I od. 103 'seu te fulgentia signis Castra tenent.' " The only other scholar, as far as known to me, who suspected Rutulis, was Carlo Pascal, 104 who proposed ecce mihi lucent rutilis incendia classis. 105 It is, to be sure, but a slight change from rutilis to Rutulis, and the reference to the early conflicts of Aeneas in Italy might easily suggest a mention of his enemies, the Rutulians. Read- ing RuUdis then, three possible interpretations of the passage are suggested. First, that the Rutula .... castra was the city of Ardea, 106 the burning of which is mentioned, for example, in Ovid Met. XIV, 572 foil. : 107 102 What appears in G, I do not know, but from Baehrens' critical note I infer that Rutulis is its reading. 103 Hor. Od. I, 7, 19-20. loiBivista di Filologia, 17 (1889), 452-4. 105 This reading takes rutilis as an adjective of the third declension, but, as Cartault points out, its use here would be "avec une faute de quantity. ' ' ice Suggested as a possibility by Heyne. io7 It is of course true that in some respects the two passages resemble each other. 1912] Beutsch. — Notes on the Text of the Corpus Tibullianum. 211 tandemque Venus victricia nati arma videt, Turnusque cadit, cadit Ardea, Turno sospite dicta potens. Quam postquam barbarus ignis abstulit, et tepida latuerunt tecta favilla, congerie e media turn primum cognita praepes subvolat, et cineres plausis everberat alis. In answer to this interpretation, it may be said that no passage is cited where Ardea is termed castra or Rutula castra; and even if there were, there is in our elegy not the slightest mention of Ardea, so as to bring it to mind when the Rutula castra is referred to. In the second place, Cyllenius, Wunderlich, Voss, and Huschke take incendia = faces, namely, those that are now ready in the camp of the Rutulians to attempt to burn the fleet of the Trojans. This is referred to in Virgil Aen. IX, 69 foil, and Ovid Met. XIV, 530 foil. It is true, to be sure, that in each of these accounts incendia practically = faces; but it is far more accurate to say that in the two passages cited the word means "flames" and the context shows that faces are thought of. Virgil Aen. IX, 71-2 sociosque incendia poscit ovantis atque manum pinu flagranti fervidus implet. Ovid Met. XIV, 539-40 ' Irrita sacrilega iactas incendia dextra, Turne,' ait.ios In our passage there is not the slightest hint that faces are alluded to; but, even beyond this, would it be likely that in her prophecy the seeress would refer, and that too in such impas- sioned language, merely to the preparations for the burning and not to the actual flames amid the ships of the Trojans ? Cf . Ovid Met. XIV, 532-4 : iamque picem et ceras alimentaque cetera flammae Mulciber urebat, perque altum ad carbasa malum ibat, et incurvae fumabant transtra carinae. The third interpretation, stated by Heyne 109 and accepted by modern scholars, is thus phrased by Postgate : ' ' The burning of Turnus' camp is not mentioned in Virgil." One might infer 108 The word faces appears at the very beginning (531) of the account of the attempted burning. loo « « Ergo necesse est, fuisse inter veteres de Troianorum rebus in Italia fabulas etiam aliquam narrationem de castris Turni ab Aenea captis et incensis, etsi a Virgilio praeteritam" (Heyne). 212 University of California Publications in Classical Philology. [Vol. 2 from this that it is mentioned in some other author, but, if so, the various editors have entered into a conspiracy of silence. It is surely true that there may have been a firing of the Rutulian camp by the Trojans, but what right have we to alter the read- ing of the MSS in order to get an allusion to an event, the only mention of which appears in this very word? Recurring to the whole passage, we find that the Sibyl refers only to the more important events in the future history of the Trojans, the events familiar to everyone who has even a slight knowledge of the classics. Thus in verse 40 we have the refer- ence to the exiled Aeneas bearing the Trojan sacra in his ships ; in verses 43 and 44 the death of Aeneas, the purification in the Numicius and his becoming "deus indiges" 110 are all mentioned. And in the following verses we have allusions to Turnus' death, Laurentum, Lavinium, the founding of Alba Longa, Rhea Silvia and the love of Mars. Finally the wide domain of the future Roman state is foretold. In this summary of important events in the history of the Trojan settlement in Italy, ecce mihi lucent .... incendia would suggest but one thing, i.e. the attempt of the Rutulians to burn the Trojan fleet and camp. Cf. Virgil Aen. IX, 69 : classem, quae lateri castrorum adiuncta latebat. This attempt, frustrated only by the gods, is described at length in the Aeneid IX, 69 foil, and Ovid Met. XIV, 530 foil. As has been mentioned previously, the ships were actually on fire, so that the term incendia is perfectly appropriate. Let us now turn to the word rutilus. It is used frequently in Latin as an epithet of fire : Virgil Aen. VIII, 430 rutili tris (radios) ignis. Virgil Georg. I, 454 rutilo .... igni. Ovid Fast. Ill, 285-6 Ecce deum genitor rutilas per nubila flammas spargit. Ovid Met. IV, 402 foil. Tecta repente quati, pinguesque ardere videntur lampades et rutilis collucere ignibus aedes. So too Ovid Met. XI, 435-6 and XII, 294-5 ; Her. Ill, 64 ; and Stat. Theb. IV, 5-6. From this it would be but a step to the use of rutilus in connection with objects reddened by fire. We no Cf. Ov. Met. XIV, 597-608. 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 213 find in Sil. Ital. XVI, 232 sed prorupturis rutilabant aequora flammis (of the dawn) ; Stat. Theb. X, 84CM: reads: ardua mox torvo metitur culmina visu, innumerosque gradus, gemina latus arbore clusus, aerium sibi portat iter, longeque timendus multifidam quercum flagranti lumine vibrat; arma rubent una clipeoque incenditur ignis. And, while it is not strictly parallel, mention may be made of Ovid Met. XII, 276-7, ferrum igne rubens (red-hot). The adjective "red" is not a permanent epithet of the camp; it is simply rutilus because of the fire. Thus in Virgil Georg. Ill, 358-9 we read nee cum praecipitem Oceani rubro [Sol] lavit aequore currum. Forbiger's note is: "rubro aequore pro rubescente a sole, igneo. " Again Horace Od. I, 2, 2-4 run as follows : et rubente dextera sacras iaculatus arces [Pater] terruit Urbem. E. C. Wickham's comment is "red from the flames of the bolt which he is launching," and the note in the edition of Orelli- Baiter-Hirschf elder reads "a repercusso fulminum rubore. " But a far bolder use is that in Valerius Flaccus V, 450-1 : et iam rutilis correpta venenis implicat igne domos. The lexicon of Facciolatus and Forcellinus comments : ' ' rutilis ; h.e. incendia portantibus. " On the basis of such passages, par- ticularly Horace Od. I, 2, 2-4 and Virgil Georg. Ill, 358-9, it seems possible to take rutilis .... castris as the camp reddened by flames. The meaning of rutilis is surely made clear by lucent and incendia. If it be urged that a reference to the attempted burning of the fleet is out of place in a prophecy wherein the future triumphs of the Trojans and their descendants are mentioned, it may be noted that several other unfortunate circumstances are alluded to, though in each case a compensation for it is mentioned. In verse 42 the lares are termed errantes, though in the same verse vocat .... Jwspita terra. The death of Aeneas which is implied is compensated for by his deification. The ships 214 University of California Publications in Classical Philology. [Vol. 2 are termed fessae, but Victory at last comes to them. So might one not feel here that for the attempted burning alluded to in verse 47 the death of Turnus is retribution ? ecce mihi lucent rutilis incendia castris: iam tibi praedico, barbare Turne, necem. Ill, 6, 3 Proposed reading: aufer et ipse meum pariter medica arte dolorem. This verse, on the reading of which all the best MSS agree, has been the happy hunting-ground of the emendator. The reading of AVG Plant, is pariter medicanclo. 111 This reading is found in editions before Broukhusius, as well as in Maittaire and Lachmann. Statius made the change of medicando to medicande, which was accepted among others by Guyet, Brouk- husius, Voss, Golbery, Dissen and F. Wilhelm. 112 Waarden- burg's patera medicante has been chosen by Santen, Gruppe, Haupt- Vahlen, L. Mueller, Hiller, and Postgate. 113 Among other readings suggested are : Huschke — medicate; Heyne — pater et medicare ; 114 Baehrens — pater o, medicare; Birt (ad hist, hexam. Lat. symb. Bonnae 1876, p. 47) — patera medicare. Belling in his Prolegomena suggests "hue ades atque meum pariter medi- care dolorem," and in his Albius Tibullus: Untersuchung und Text, pariter temptate. Postgate 's suggestion is pariles medicate. 113 The reading of the MSS (pariter medicando) seems syntacti- cally, and, more especially, metrically impossible (i.e. at this period of Latin). 115 Medicande, involving the slightest change, would mean "you who also need to be cured," and a reference in "pariter comp. " Hiller says. 112 N. Jahrb. f. Phil. u. Paed., 147 (1893), 769 foil. us But note Postgate 's objection to patera in the Classical Beview, 9 (1895), 77, and the Journal of Philology, 25 (1897), 59. ii4 Heyne says: "Wacker Medebach. in Amoen. p. 72 nuper emendare vidimus: 'Adfer et ipse merum, pater, et medicare dolorem.' " Heyne also suggested (though admitting its remoteness from the MS reading): "Aufer et indomitum succo medicante dolorem." us See Belling, Prol., 76-7. 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 215 to Bacchus' love for Ariadne is suggested. 116 But love's pain is to be cured by forgetfulness ; does Bacchus, the successful lover, need that sort of remedy? As Belling 117 puts it, "jener Aus- druck, der den Gott als gegenwartig liebeskrank, als ungliick- lich Liebenden vorstellt, (ist) unertraglich schief." The love affair with Ariadne (see verses 39 foil, of this elegy) is past ; he is therefore not now medicandus. Moreover it is not an adventure in which Bacchus' situation as represented by Lygdamus is in any wise comparable with his own. The conjectures which introduce the word patera are un- happy; the patera 115 is the libation bowl, not the cup the lover will use to drain draughts of healing for his love. I have therefore suggested for medicando, medica arte. The expression appears in the Corpus Tibullianum, II, 3, 14 : quidquid erat medicae vicerat artis amor. Medicus appears also in III, 10 (IV, 4), 3-4: crede mihi, propera: nee te iam, Phoebe, pigebit formosae medicas applicuisse manus. The appropriateness of the expression is apparent, pariter look- ing forward to the next verse, as others have noted. "Do thou even thyself remove in like wise my pain by thy healing skill; often has love perished, conquered by thy gift." For similarities in expression one may compare : Propertius III, 17, 3-4 118 tu potes insanae Veneris compescere fastus, curarumque tuo fit medicina mero. Tibullus I, 2, 1 adde merum vinoque novos compesce dolores. Ovid Bern. Am. 75-7 119 Te precor incipiens, adsit tua laurea nobis, carminis et medicae, Phoebe, repertor opis; tu pariter vati, pariter succurre medenti. Ovid Bern. Am. 131-2 temporis ars medicina ferest: data tempore prosunt, et data non apto tempore vina nocent. Ovid Bern. Am. 135 nostrae medicabilis arti. lie For a defence of medicande, see F. Wilhelm's articles, N. Jahrb. f. Phil. u. Paed., 1893, p. 769, and 1895, p. 775. ii7 Prol., p. 76. us Cf. also Prop. I, 5, 27-8; II, 1, 57; III, 17, 9-10. us Cf. Rene Pichon, Be sermone amatorio, etc., Paris, 1902 (sub medicina, mederi, medicus). 216 University of California Publications in Classical Philology. [Vol. 2 The change palaeographically from medica arte to medi- cando 120 is not difficult to explain. The ending -do is clearly due to the following word, dolorem. (The corruption of -te to -do is also found in Ovid Her. IX, 126: tegente P; tegendo G.) I have noted the following instances in the Corpus Tibullianum where the error in the IMS reading may to some extent have been due to the beginning of the following word: 121 following word I, 3, 87 at Par; ac A circa. I, 7, 13 an f ; at A te. Ill, 7 (IV, 1), 78 erroris F; errorum A miseri. Ill, 16, (IV, 10) 6 ne f ; nee A cedani. II, 4, 2 paterna G; paterve A vale. The confusion between arte and ante appears in Tib. II, 1, 24 where for ante, £ read arte. In Prop. II, 3, 42, for in arte D (the accepted reading), NFV read in ante. The error in the MS reading (if the conjecture be accepted) is in large measure due to an improper division of words, two words being written as one. The following instances of this very mistake appear in the MSS of Tibullus : 122 I, 1, 44 si licet f ; scilicet A Par. 1, 5, 7 per te G; parce A. I, 6, 40 et fluit f ; effluit A. I, 8, 39 iuvant quae V; iuvatque A. II, 4, 17 et qualis f ; equalis A. II, 6, 16 st licet f ; scilicet A. III, 1, 15 per vos G; parvos A. Ill, 4, 87 canis anguinea G Cuiacianus Plant. , ; consanguinea A. cams anguma Postgate {termine A (supra scr. ab al. manu ge). tergemine, G Cuiacianus. Ill, 7 (IV, 1), 189 ante actos (accepted ._. , T " . :„ J accitos V Cuiacianus. I accitus A. Ill, 17 (IV, 11), 1 pia cura f; placitura A. At other times the improper division of words has corrupted the reading by breaking , up one word into two. The MSS of Tibullus show the following instances of this sort of corruption : 120 A possible order of corruption is: medica arte — medicarte — medicante — medicando. 121 There is no attempt at completeness in this list. 122 There is no attempt at completeness in this list. 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 217 I, 4, 29 deperdit Par.; te perdit A. I, 7, 9 Tarbella Scaliger; tua bella A. I, 9, 19 divitiis f; viciis AV. Ill, 7, (IV, 1), 39 castrisve Par. Plant.; cartis ne A. Ill, 7 (IV, 1), 73 more f Plant.; fef ore A. ttt n /tit -i \ -i ^o • ^ • o i • ( * e « iunctum A Plant. Ill, 7 (IV, 1), 103 seiunctim Salmasius;J . ^ | seu vmctum Par. Finally error is caused by improperly dividing words, even where the number of words employed is not altered, e.g. II, 6, 21, in which credit aratis appears in Fr. as credita ratis. It is apparent, then, how numerous the instances are wherein the text of the Corpus Tibullianum has suffered through an improper word-division ; the other changes in medica arte are easily explainable palaeographically. The expression is more- over in harmony with the elegiac diction and appropriate to the sense of the passage. Ill, 12 (IV, 6), 19-20 Proposed reading: si iuveni grata est, veniet cum proximus annus, hie idem votis mutuus adsit amor. MSS : sis F; si AVG A (of Lachmann) ; sit G 2 ; sic f. iuveni AFVG. veniet AVG Cuiacianus; et veniet f; ac veniet f; adveniet G 2 Plant. iam vetus O Plant, (where mutuus stands in verse 20). esset O ; adsit f Plant. ; ut sit G 2 . The reading of verse 19 is seriously disputed ; in verse 20 all editions seen read adsit save Scaliger 's, which retain esset. 123 Vetus is also generally read; 124 Prien 125 however proposed ratus, which Baehrens incorporated in his text 126 and H. Graef 127 approved. C. M. Francken 128 is also dissatisfied with vetus. 12 3 Statius — ut sit: Baehrens — "fort, extet": Cartault (1909) — exstet and vobis for votis. 124 As far as I know, no one alters iam. !25 Die Symmetric und Eesponsion der rbmischen Elegie, p. 8. 12 e gee also Baehrens' Tibullische Blatter, p. 91. 12 7 Annotationes ad Tibullum (Particula altera), Memel., 1885. ^Mnernos. n.s., 13 (1885), 185. 218 University of California Publications in Classical Philology. [Vol. 2 It is, however, in the first verse that we find the main diffi- culties, and so the less unanimity of agreement on the part of scholars. The various important readings are as follows: sit iuveni grata, adveniet — Aldus, Sessa, Muretus, Grasser. sic iuveni gratum, veniet — Statius, Kraffert (with colon after gratum) .129 sic iuveni grata, veniet — Scaliger.130 sic iuveni gratis, veniet — Heinsius,i3i Broukhusius, Volpi, Heyne. 132 sit iuveni grata, ac, veniet — Passerat, J. Dousa, Maittaire, Delphin Classics, Voss, Wunderlich, Bach, Golbery, Hertzberg, L. Doe- derlein. Heyne also approves of this reading, after changing the order of the preceding distichs. sternuit ilia: ratum est — Herel (see Heyne 's edition), sic iuveni placeat: veniet — Huschke. si, iuveni grata veniet — Lachmann, Dissen. si. iuveni gratae veniet — Lachmann proposed, sis iuveni cara — Dissen proposed. sis, Iuno, grata C ac | veniet — Gruppe, Prien, Graef. So also, sub- ) et ( stituting ut (Eberz) before veniet — Hil- ler,i33 Postgate,i3* Jurenka. diva, veni grata, ut verteret cum — Baehrens.135 sit iuveni grata, et veniet — Vahlen (ed. V), Jacoby. sit iuveni gratum: veniet — Mueller, si iuveni gratum, veniet — Eigler. sis iuveni grata; adveniet — Rossbach. sit iuveni grata: veniet — K. P. Schulze.iso sit iuveni grata, ut, veniet — A. Otto. 13 ? sic iuveni gratae, veniet — Belling.iss sic iuveni grato veniet — Ehwald.1 39 sis iuveni grata ac — Cartault. 129 Beitrage zur Kritik und Erlcldrung lateinischer Autoren, III Teil, Aurich, 1883. 130 Scaliger: "grata hoc est gratare. " 131 Heinsius: "gratis iuveni votis adsit Amor iterum anno proxime venturo. ' ' 132 But he was dissatisfied with this reading. 133 In the Corpus poetarum Latinorum (1893), he reads: "sis iuveni grata, ut. ' ' is* In his Selections; in his edition he marks the verse corrupt, suggest- ing "fort, grate." 135 Cf. Tibullische Blatter, p. 91. - ' 136 Beitrage zur Erklarung der romischen Elegiker, Berlin, 1893, p. 22 ; and Wochenschrift fur Phil., 2 (1885), 598. He also thought that perhaps votis should be iunctis. 137 Zeitschr. f. Gymnasialwesen, 39 (1885), 225. 138 C. M. Francken (Mnemos. n.s., 13, 185) believes that the thought should be to this effect: sic iuveni gratae, veniet cum proximus annus, his palam votis arbiter adsit Amor. wPhilol, 54 (1895), 458-9. 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 219 This list, an incomplete one, should perhaps be sufficient to deter one from a fresh attempt, but it shows at any rate that the reading of the distich is not yet settled. In the reading proposed the word est is inserted after grata; the text at present is unmetrical, and it is obvious that est could readily be dropped, particularly if si became altered to sis or sit, thus demanding the elimination of the following verb. 140 This clause is the protasis of the condition, and the citm-clause is a part of the apodosis, i.e., "if she is dear to the youth, then when the next year comes, etc." Iuveni is of course Cerinthus, as in verse 8 of this elegy; and the feminine subject of est implied in grata is the docta puella of verse 2, ilia of verses 5, 10 and 16. There is, to be sure, no certain instance of gratus referring to a person in the Corpus Tibullianum; though it may be that in II, 5, 68 the reading of A (grata) is correct, and Herophile Phoebo grata would be parallel to this passage. We do find gratus with a dependent dative in III, 7 (IV, 1), 8-9, but modify- ing a non-personal noun : etiam Phoebo gratissima dona Cres tulit. But outside of the Corpus Tibullianum the instances are numerous enough : Ovid Fast. Ill, 467-8 inter eaptivas facie praestante puellas grata nimis Baccho filia regis erat. Am. II, 19, 30 factast, quam fuerat, gratior ilia Iovi. Fast. Ill, 495 vitio tibi gratior ipso est. Met. VIII, 771 Nympha sub hoc ego sum Cereri gratissima ligno. Met. XIII, 528 proles gratissima matri. Horace C. I, 10, 19-20 superis deorum gratus (Mercurius) et imis. C. Ill, 9, 1 donee gratus eram tibi. Propertius I, 2, 31 his tu semper eris nostrae gratissima vitae. 141 In other words, we have that meaning that Pichon refers to when he says, "Ad personas hoc verbum (gratus) refertur quae carae sunt amantibus. ' ' 140 Voss says: "In anderen (Handschriften) war et veniet verderbt in est veniet und eveniet, ' ' and again : " In den meinigen verschrieb einer ' Sic iuveni grata est'; einer 'si iuveni grata est.' " Broukhusius also says: "grata est unus St." According to Heyne, Guelf. 3 read: "Si iuveni grata est, veniet cum etc." Belling (Prol., p. 70) says: "Soviel ist sicher, dass an der Commissur der Worte grata und veniet ein Fehler der tiber- lieferung steckt. " 1*1 Note also Prop. I, 12, 7, "olim gratus eram." 220 University of California Publications in Classical Philology. [Vol. 2 The fact that the CMm-clause is taken as modifying the fol- lowing wish and not as dependent upon the protasis which it immediately succeeds, may seem unusual, and yet there are several instances of such a use in the Corpus Tibullianum. Thus in I, 4, 49-50 : nee, velit insidiis altas si claudere valles, dum placeas, umeri retia ferre negeDt. the second clause does not depend upon the first, which, it may incidentally be noted, is a si-clause as here. Again, in I, 6, 41-2 : quisquis et occurret, ne possit crimen habere, stet procul ante, alia stet procul ante via. the we-clause modifies, not the preceding, but the succeeding clause. And I, 8, 65-6 : dum mini venturam fingo, quodcumque movetur, illius credo tunc sonuisse pedes. shows the same usage. And the elegy preceding the one under discussion, namely III, 11 (IV, 5) shows something rather similar in verses 5-6 : iuvat hoc, Cerinthe, quod uror, si tibi de nobis mutuus ignis adest. for the second subordinate clause does not depend upon the clause immediately preceding. Let us now turn to the meaning of the clause as a whole : " If she is dear to the youth, then, etc." It is, however, clearly felt that the love is reciprocated, or else verse 7 : at tu, sancta, fave, neu quis divellat amantes. would be inappropriate ; and even more convincing is verse 11 : nee possit cupidos vigilans deprendere custos. Therefore the expression means: "If she is dear to the youth (and she surely is), then, etc." The previous elegy, which has so many correspondences with this one, contains a thought closely parallel to this in verses 9-10 : mane Geni, cape tura libens votisque faveto, si modo, cum de me cogitat, ille calet. 1 * 2 Here si modo .... ille calet is parallel to si iuveni grata est; note further that the conclusion to this protasis is in part votis 1*2 But the expressions of confidence in Cerinthus' love are far stronger in III, 12 (IV, 6). 1912] Deutsch. — Notes on the Text of the Corpus Tibullianurn. 221 faveto, as in verse 20 of our passage it is votis .... adsit amor. In other words, it is only if he loves her (and the writer of this elegy feels that he does) that the speaker wishes the prayers granted. Next stands the clause veniet cum proximus annus. As these prayers are offered upon her birthday, clearly this must mean "when her next birthday comes." So also in verse 20 the vota are those to be offered up on that next birthday, as Heyne points out: "Vota versu ultimo intellige, quae puella proximo anno eadem die conceptura est." These birthday prayers are fre- quently alluded to in the Corpus : thus II, 2, 9-10 (note also vota, v. 17) ; III, 11 (IV, 5), and in particular v. 9 votisque faveto; III, 12 (IV, 6) itself; and I, 7, the birthday poem to Messalla. It is interesting to compare Ovid Tristia III, 13, and particularly verse 18 : concipiamque bonas ore favente preces? as well as Ovid Tristia V, 5, in which after the introductory verses the prayer begins in verse 13 : optime Natalis, quamvis procul absumus, opto candidus hue venias dissimilisque meo, etc. The expression votis adsit amor means, I believe, "may love be present at her (next birthday-) prayers," i.e., "may the love still exist when the prayers are offered up on her next birthday. ' ' While there are not a great many instances of adesse with its subject non-personal as well as the dative used with it, still a number of them is found. 143 In most of these, however, the meaning of adesse does not appear very different from that of esse. But in the following instances the true force of adesse is clearly shown : Virgil Aen. X, 547-8 vimque adfore verbo crediderat. 144 Ovid Her. XIV, 10 quaeque aderant sacris, (pater) tendat in ora faees. lam vetus appears inappropriate to me, as it did to Prien, Baehrens, and Francken. If the love is "already old," not a hint to that effect is given in the elegies; if, on the other hand, iam vetus is taken to mean, as it usually is, "by that time old," what * 4 s See Thesaurus Linguae Latinae, assum. 1 44 Conington, "he thought that his prowess would second his word." 222 University of California Publications in Classical Philology. [Vol. 2 point is there in the addition of these words? What does the prayer gain by having this idea added to it ? The change palaeographically to mutuus is much less difficult than it appears at first sight. If in copying, the m were separated from the rest of the word, we should have remaining utuus. If we place beside each other utuus and uetus, the close resemblance becomes apparent ; both are not only words of five letters, but in fact contain exactly the same letters, with a single exception, arranged in a slightly different order. The m remain- ing by itself would then be expanded into the "Flickwort" iam. The word mutuus is a favorite one in this group of elegies; thus in verse 8 of this elegy, we have mutua vincla, and in III, 11 (IV, 5) we have in verse 6 mutuus ignis adest and in the next verse this very group of words, mutuus adsit amor. This very correspondence may be urged as an objection to the proposed reading, but it might be claimed that such, an objec- tion would in some measure apply to the substitution of adsit for the MS esset, which gives us these two words in precisely the same order as in III, 11 (IV, 5). And yet this change is made by practically every editor of the Corpus Tibullianum. More- over, in III, 11 (IV, 5) we have mutuus ignis adest followed immediately by mutuus adsit amor But, aside from this, is not the prayer that the mutual love may only last, a frequently recurring one in these elegies? Cf. e.g. Ill, 10 (IV, 4), 16: tu modo semper ama: salva puella tibi est. This fear that Cerinthus' love may not last we find again in III, 17 (IV, 11). The distich then as a whole would mean: "If she is dear to the youth (and she really is), then at the coming of the next year may this same mutual affection exist when the birthday- prayers are offered up." 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 223 III, 14 (IV, 8), 5-6 Proposed reading: iam, nimium Messalla mei studiose, quiescas : non tempestivae saepe, propinque, viae. The reading above is that of the MSS, save that for non £, AVG read neu. Sepc propinque is read in compendium in A. The reading given is moreover that of the large majority of the editors, who, however, place a comma after quiescas, and remove that after saepe. Those who read thus are Muretus, Scaliger, Grasser, Broukhusius, Maittaire, Heyne, Huschke, Golbery, Lachmann, Dissen, Gruppe, L. Mueller, Haupt-Vahlen, and Belling; 145 of these, however, Heyne, Huschke, and Dissen feel that something is wrong in the pentameter. Heyne proposes non tempestivam sic properare viam. Baehrens and Postgate 140 both mark saepe propinque as corrupt, though in his Selections, Postgate accepts non .... saeve propinque (the conjecture of R. Unger) as a "stop-gap." Baehrens 147 suggested perge monere (retaining neu), which Hiller 148 accepted and inserted into his text. Jurenka would read : neu tempestivae saepe moneto viae. Rigler, at first accepting non tempestivae, saeve propinque, viae, later proposed non tempestivae, quae procul urbe, viae. Ayrmann, Vit. Tib., § 82 suggested non tempestivae simque pro- pinqua viae. Francken read for saepe propinque, parce, pro- pinque. % ** Cartault (1909) proposed seu tempestivast, sive propinqua via. 150 In these various readings viae is taken either as a genitive us But with no punctuation save a colon after viae. i* 6 He proposes in the Journal of Philology, 25, 64, semper amice for saepe propinque. **' Baehrens thought the following possible: "quiescas Ceu tempestivae, saeve propinque, viae." i*8 In the Corpus poetarum Latinorum (1893) it is marked corrupt. lioMnemos. n.s., 13 (1885), 186. iso Statius suggested fugae for viae. Heyne made a second suggestion: "non tempestive saepe propinque tuae. " Voss read as the majority of the editors, substituting however nee for non, and tempestive for tem- pestivae. After this paper was in print, I noted the proposal of J. J. Hart- man, Mnemos. n.s. 39 (1911), 399, Non tempestiva est, saeve propinque via. 224 University of California Publications in Classical Philology. [Vol. 2 depending upon quiescas, 151 or as depending upon propinque. Thus Scaliger: "Qui saepe te aceingis itineri intempestivo." The reading suggested by me follows the MS reading, save in substituting non for neu. A colon is, however, placed after quiescas, and by putting a comma both before and after propinque, saepe is completely separated from it in construction. The verse non tempestivae saepe, propinque, viae would then be translated: "Journeys are often unseasonable, kinsman." The only other instances of quiescere in the Corpus Tibul- lianum are II, 4, 49, and II, 6, 30, and in neither of these do we have a noun used directly with it or an infinitive, but the use in each case parallels that proposed : II, 4, 49 et 'bene' discedens dicet 'placideque quiescas.' II, 6, 30 sic bene sub tenera parva quiescat humo. «■ The ellipsis of sunt which has been assumed is quite in accord with usages in the Corpus Tibullianum. In discussing the read- ing of I, 5, 76 (page 182), the instances of ellipses of esse in the Corpus Tibullianum are classified under four heads, of which the third "comprises those cases wherein a predicate noun or adjective occurs," which is the condition we have in this verse. The following instances fell under this head : I, 1, 75 hie ego dux milesque bonus. I, 3, 43 non fixus in agris, qui regeret certis finibus arva, lapis. I, 7 9 foil. Tarbella Pyrene testis et Oceani litora Santonici, testis Arar Khodanusque celer magnusque Garunna, Carnutis et flavi caerula lympha Liger. I, 10, 26 (text dubious) hostiaque e plena rustica porcus hara. I, 10, 63 quater ille beatus quo tenera irato flere puella potest. II, 1, 79 a miseri, quos hie graviter deus urget! II, 1, 80 at ille felix, cui placidus leniter adflat Amor. II, 5, 107 ars bona! III, 2, 5 non ego firmus in hoc. Ill, 4, 51-2 tantum cara tibi quantum nee filia matri, quantum nee cupido bella puella viro. Ill, 4, 94 isque pater, quo non alter amabilior. i5i Postgate, Selections: "viae gen. after quiescas. 'cease from un- reasonable travel,' i.e., give up the journey, an imitation of the Greek construction." 1912] Deutsch. — Notes on the Text of the Corpus Tibullianum. 225 III, 6, 19 nee torvus Liber in illis. Ill, 6, 43 felix, quicumque dolore alterius disces posse cavere tuo. Ill, 7 (IV, 1), 9-10 et cunctis Baccho iucundior hosjtes I cams. Ill, 7 (IV, 1), 40 nee tamen hie aut hie tibi laus maiorve minorve. Ill, 7 (IV, 1), 107 foil. testis mihi victae fortis lapydiae miles, testis quoque fallax Pannonius, gelidas passim disiectus in Alpes, testis Arupinis et pauper natus in arvis. Ill, 19 (IV, 13), 11 tu mihi curarum requies, tu nocte vel atra lumen, et in solis tu mihi turba locis. In the group of elegies by Sulpicia we find the following in- stances of an omission of esse : III, 15 (IV, 9), 1: in a compound verb-form. Ill, 16 (IV, 10), 3-4: where another form of esse is found in the same sentence. In other words, ellipses of esse of this type are frequent in the Corpus Tibullianum, and while none appears in the poems by Sulpicia, yet in this small group of elegies there are two instances of the omission of esse. It may be objected that the thought in this verse is not a very profound one, but generalizations and those too not very abstruse are not uncommon throughout the Corpus. Thus, for an example of such a generalization following a command (cf. this distich), we have in III, 10 (IV, 4), 15: pone metum, Cerinthe: deus non laedit amantes. Compare also : III, 2, 6 frangit fortia corda dolor. Ill, 3, 20 falso plurima vulgus amat. Ill, 3, 21 non opibus mentes hominum curaeque levantur. Ill, 4, 63 mens est mutabilis illis. Ill, 4, 76 vincuntur molli pectora dura prece. Ill, 6, 34 difficile est tristi fingere mente iocum. Ill, 14 (IV, 8), 3 dulcius urbe quid est? I, 4, 28 non segnis stat remeatque dies. I, 4, 77 gloria cuique sua est. While it must be admitted that saepe holds this position in the verse nowhere else in the Corpus Tibullianum, yet it may be noted that of fifty-two instances 152 of saepe in Propertius, 153 152 Including II, 25, 12. 153 See Phillimore, Index Verborum Propertianus. 226 University of California Publications in Classical Philology. [Vol. 2 twenty-one appear in exactly the position saepe has here, i.e., immediately following the diaeresis in the pentameter, and thirty- one in all other possible positions. Moreover I have come upon two cases where saepe and pro- pi nquus stand next to each other, and in neither case does saepe modify propinquus : Propertius IT, 6, 7 quin otiam t'alsos fingis tibi saepe propinquos. Martial IX, 54, 11-12 (which is more interesting because the words appear in the same position in the pentameter as in the verse under consideration): mittimus ergo tibi parvae minuscula chortis. qualia si recipis, saepe propinquus eris. As to the exact relationship that Messalla bore to Sulpicia, that cannot of course be determined with certainty, but Haupt's 154 suggestion that Valeria, the sister of M. Valerius Messalla Corvinus (the Messalla of this elegy), who married Servius Sulpicius, 155 was the mother of Sulpicia, seems likely. Transmitted January 8, 1912 is* Haupt, Opuscula, iii, 502-3, number lxii, which is the same as Hermes, 5 (1871), 32-4. 155 Cf. Ill, 16 (IV, 10), 4. UNIVERSITY OF CALIFORNIA PUBLICATIONS— (Continued) EGYPTL4.N ARCHAEOLOGY. (Quarto.) Vol. 1. The Hearst Medical Papyru3. Hieratic Text in 17 facsimile plates in collotype, with Introduction and Vocabulary, by George A. Reisner. 48 pages. 1905 25 marks Vol. 2. The Early Dynastic Cemeteries of Naga-ed-DSr, Part I, by G. A. Reisner. 172 pages, 80 plates, 211 text-figures. 1908 -75 marks Vol. 3. The Early Dynastic Cemeteries at Naga-ed-D6r, Part IL by A. 0. Mace, xi + 88 pages, with 60 plates and 123 text-figureB. 1909 50 marks For sale by J. C. Hinrichs Verlag, Leipzig, Germany. Copies for exchange may be obtained from the University Press, Berkeley. SEMITIC PHILOLOGY.— William Popper, Editor. Vol. 1. 1907-. (In progress.) 1. The Supposed Hebraisms in the Grammar of Biblical Aramaic, by Herbert Harry Powell. Pp. 1-55. February, 1907 „ 75 Vol. 2. 1909— (In progress.) 1. Ibn Taghri Birdi: An-Nujum az-ZahirS fi Muluk Misr wal-Kahira (No. 1 of Vol. 2, part 2). Edited by William Popper. Pp. 1-128. Sep- tember, 1909 - 1.60 2. Idem (No. 2 of Vol. 2, part 2). Pp. 129-297. October, 1910 1.50 The publication of this text will be continued. European orders for the parts of this volume as published may be sent to Late E. J. Brill, Ltd., Leiden. MODERN PHILOLOGY.— -Charles M. Gayley, Lucien Foulet, and Hugo BL Schilling, Edi- tors. Price per volume $2.50. Vol. 1. 1. Der Junge Goethe und das Publikum, by W. R. R. Pinger. Pp. 1-67. May, 1909 50 "2. Studies in the Marvellous, by Benjamin P. Kurtz. Pp. 69-244. March 17, 1910 $2.00 3. Introduction to the Philosophy of Art, by Arthur Weiss. Pp. 245-302. January 12, 1910 _ - 50 4. The Old English Christian Epic, by George A. Smithson. Pp. 303-400. September 30, 1910 - 1.00 Vol. 2. 1. Wilhelm Busch als Dichter, Kunstler, Psychologe, und Philosoph, von Fritz Winther. Pp. 1-79. September 26, 1910 75 2. The Critics of Edmund Spenser, by Herbert E. Cory. Pp. 81-182. June 30, 1911 1-00 3. Some Forms of the Riddle Question and the Exercise of the Wits in Popular Fiction and Formal Literature, by Rudolph Schevill. Pp. 183-237. November 2, 1911 : 50 MEMOIRS OF THE UNIVERSITY OF CALIFORNIA (Quarto). Vol. 1. 1. Triassic Ichthyosauria, with special reference to the American Forms, by John C. Merriam. Pp. 1-196; plates 1-18; 154 text-figures. Sep- tember, 1908 < - $3.00 2. Silva of California, by W. L. Jepson. Pp. 480; plates 85. December, 1910. $9; buckram, $10; carriage extra. Other series in Classical Philology, Education, Engineering, Entomology, Egyptian Archaeology, Graeco-Roman Archaeology, History, Modern Philology, Philosophy, Semitic Philology. UNIVERSITY OF CALIFORNIA CHRONICLE.— An official record of University life, issued quarterly, edited by a committee of the Faculty. Price § 1 per year. Current volume No. XTv". Address all orders or requests for information concerning the above publications to The University Press, Berkeley, California. European agent for the series in American Archaeology and Ethnology, Classical Phil- ology, Education, Modern Philology, Philosophy, and Semitic Philology, Otto Harrassowitz, Leipzig. For the Memoirs, and the series in Botany, Geology, Pathology, Physiology, Zoology and also American Archaeology and Ethnology, R. Friedlander & Sohn, Berlin. THIS BOOK IS DUE ON THE LAST DATE STAMPED BELOW AN INITIAL FINE OF 25 CENTS ™m B B n ASSESSED FOR F A. LURE TO» TH.S BOOK ON THE DATE DUE. THE PENALTY ^AY L AND R ro SE T ° 5 ° CENTS ° N ™ E ™«™ oCerdue. ° N ™ E SEVENTH day SEP 28 1*32 LD 21-50m-8,32 / 244548