72. TITVLOS. 'Claims to glory'; compare Met VII 448-49 (to Theseus) 'si titulos annosque tuos numerare uelimus, / facta prement annos' and Met XII 334 'uictori titulum ... Dictys Helopsque dederunt'.

73. VETAT ILLE PROFECTO. 'I am quite certain that he does not allow ...'

74. TRANQVILLI ... TEMPORIS implies sed non temporis aduersi.

75. CONDITVR A TE. Ovid does not elsewhere use a person as the object of condere, although at Tr II 335-36 he uses a person's achievements as object: 'diuitis ingenii est immania Caesaris acta / condere'.

76. TANTVS QVANTO L TANTO QVANTVS B^acCFHIT^pc TANTVS QVANTVS M^2c TANTO QVANTO B^pcT^ac QVANTO TANTVS fort legendum. The transmitted reading, tanto quantus, can be construed: Professor E. Fantham translates 'a man so great as should have been sung with this mighty style'. This however subordinates Theseus to Albinovanus, while the purpose of the line is to emphasize Theseus' greatness. Tanto quanto is generally printed: it is acceptable enough (compare EP II ix 11-12 'regia, crede mihi, res est succurrere lapsis, / conuenit et tanto, quantus es ipse, uiro'), but is very weakly attested, and does not explain the transmitted reading. I have printed L's tantus quanto; quanto tantus might also be read.

76. QVANTO ... ORE. For os 'grandness of utterance' Professor R. J. Tarrant compares Am II i 11-12 'ausus eram, memini, caelestia dicere bella ... et satis oris erat'.

78. INQVE FIDE THESEVS QVILIBET ESSE POTEST. For the use of mythological figures as character types, compare RA 589 'semper habe Pyladen aliquem qui curet Oresten' and Martial VI xi 9-10 'ut praestem Pyladen, aliquis mihi praestet Oresten. / hoc non fit uerbis, Marce: ut ameris, ama'.

79-82. Professor R. J. Tarrant points out to me how the example of Theseus balances the comparison with Ulysses at the start of the poem. Earlier Ovid argued against a difference of scale between his own case and the mythic figure's: here he insists on it.

79. HOSTES ... DOMANDI. For lists of these enemies, see Her II 69-70 'cum fuerit Sciron lectus toruusque Procrustes / et Sinis' and the Athenians' hymn of praise to Theseus at Met VII 433-50.

79. CLAVAQVE. For Theseus' club see Her IV 115-16 (Phaedra to Hippolytus) 'ossa mei fratris claua perfracta trinodi / sparsit humi' and Her X 77 'me quoque, qua fratrem, mactasses, improbe, claua'. Ovid mentions the club of Hercules about a dozen times.

80. VIX ILLI. For uix 'with difficulty' OLD uix 1 cites Fast I 508 'uix est Euandri uixque retenta manu'.

Most editors print VIX VLLI (BCT), which is possible enough. Vix illi seems rather more forceful, however, as making the point that even Theseus was able to make the dangerous journey only with difficulty, and that before him the road was impassable. Compare Met VII 443-44 'tutus ad Alcathoen, Lelegeia moenia, limes / composito Scirone patet'.

81. OPEROSA. The word in the sense 'troublesome' seems confined to prose except for this passage and Her II 63-64 'fallere credentem non est operosa puellam / gloria; simplicitas digna fauore fuit'.

83. PERSTAS IPF^2ul. Compare Tr IV i 19-20 'me quoque Musa leuat Ponti loca iussa petentem: / sola comes nostrae perstitit illa fugae' and Tr V xiv 19-20 'quae ne quis possit temeraria dicere, persta [uar praesta] / et pariter serua meque piamque fidem'. PERSTAS, the reading of most manuscripts, would have no acceptable meaning in the present passage; it has no object, and the intransitive meaning, 'stand out', is clearly inappropriate. The error may have been induced by Tr IV v 23-24 'teque, quod est rarum, praesta constanter ad omne / indeclinatae munus amicitiae'; more probably, it is an aftereffect of praestandus in 81.

83. INDECLINATVS governs amico. The only other instance of the word in classical Latin seems to be Tr IV v 24, quoted at the end of the last note.

84. LINGVA QVERENTE. Ovid elsewhere uses persons as the subject of queri, except for similar uses of metonymy at xiv 26 'littera de uobis est mea questa nihil' and Tr V xi 1-2 'Quod te nescioquis per iurgia dixerit esse / exulis uxorem, littera questa tua est'.

XI. To Gallio

1. GALLIO. Junius Gallio[24], adoptive father of the younger Seneca's elder brother, is often cited by the elder Seneca, who considered him one of the four supreme orators of his time (Contr X praef. 13). At Suas III 6-8, Seneca discusses Gallio's fondness for the Virgilian phrase plena deo (which, oddly, is not found in our text of the poet), and quotes Gallio as saying that his friend Ovid was also very fond of the phrase. Quintilian and Tacitus did not share Seneca's high opinion of Gallio: Quintilian criticizes the lack of restraint in his style (IX ii 92), while at Dial 26 1 Tacitus has Messalla say how he prefers 'G. Gracchi impetum aut L. Crassi maturitatem quam calamistros ['curling irons' = 'excessive ornament'] Maecenatis aut tinnitus Gallionis'.

In AD 32 Gallio proposed in the Senate that ex-members of the Praetorian guard be permitted to use the theatre seats reserved for members of the equestrian order; this resulted in a bitter and sarcastic letter from Tiberius to the Senate attacking Gallio's presumption; he was first exiled, then brought back to custody in Rome after it was decided that Lesbos, chosen by him, was too pleasant a place of exile (Tac Ann VI 3; Dio LXVIII 18 4).

1. EXCVSABILE. The word is extremely rare, and is not found in verse outside the Ex Ponto: compare I vii 41-42 'quod nisi delicti pars excusabilis esset, / parua relegari poena futura fuit' and III ix 33-34 'nil tamen e scriptis magis excusabile nostris / quam sensus cunctis paene quod unus inest'.

2. HABVISSE could have the usual past sense of the perfect infinitive, but more probably is equivalent to habere: compare ix 20 'gauderem lateris non habuisse locum' and see at viii 82 imposuisse (p 282).

3-4. CAELESTI CVSPIDE FACTA ... VVLNERA. 'Wounds inflicted by no human weapon'. The cuspis is attributed to Mars at Am I i 11, to Neptune at Met XII 580, and to Athena at Fast VI 655. At Sen Ag 368-71 'tuque, o magni nata Tonantis / inclita Pallas, / quae Dardanias saepe petisti / cuspide terras', R. J. Tarrant cites HF 563 (Dis), HF 904 & Phaed 755 (Bacchus), HO 156 (Hercules), and Juvenal II 130 (Mars). Professor Tarrant points out to me that the cuspis does not seem to be attributed to Jupiter, no doubt because the fulmen was too firmly established as his weapon. Ovid is therefore not making his customary specific equation of Augustus with Jupiter.

4. FOVISTI. Fouere was a technical term in medicine for bathing something in a liquid (Cato Agr 157 4, Celsus IV 2 4, Columella VI 12 4). The word occurs in this sense in poetry: see Met II 338-39 'nomen ... in marmore lectum / perfudit lacrimis et aperto pectore fouit', Met VIII 654 (perhaps spurious; the passage is one where textual doublets occur), Met X 186-87 (Hyacinthus has just been struck by Apollo's discus) 'deus conlapsos ... excipit artus, / et modo te refouet, modo tristia uulnera siccat', Met XV 532 'et lacerum foui Phlegethontide corpus in unda', and Aen XII 420 'fouit ea uulnus lympha longaeuus Iapyx'.

5. RAPTI. The word could be taken to mean 'dead'; compare xvi 1 'Nasonis ... rapti', where the context shows this is the meaning, and EP I ix 1-2 (to Cotta Maximus) 'Quae mihi de rapto tua uenit epistula Celso, / protinus est lacrimis umida facta meis'. For the similarly ambiguous use of ademptus, see at vi 49 qui me doluistis ademptum (p 243).

6. QVOD QVERERERE. For the phrase, compare Am I iv 23-24 (Ovid is listing the signals his girl should use at the dinner-table) 'si quid erit de me tacita quod mente queraris, / pendeat extrema mollis ab aure manus', Tr V i 37 (of Fortune) 'quod querar, illa mihi pleno de fonte ministrat', Her XIX 79, and Her XX 34 & 94.

7-8. PVDICA / CONIVGE. Being pudica, she deserved to survive—Professor E. Fantham points out to me here Ovid's use of what could be called the quid profuit topic.

The reference to Gallio's wife seems rather cool in tone. For some very warm descriptions of recently deceased wives, see Lattimore 275-80.

8. NON HABVERE NEFAS. This sense of habere, very common in prose, does not seem to occur elsewhere in Ovid; but Professor R. J. Tarrant cites Aen V 49-50 'dies ... adest quem semper acerbum, / semper honoratum ... habebo'.

9. LVCTVS = causae luctus. Other instances of this sense of luctus, which seems to be confined to poetical passages of great emotional content, at Met I 654-55 (Inachus to Io) 'tu non inuenta reperta / luctus eras leuior', Met IX 155, and Aen VI 868 (Aeneas has just seen Marcellus) 'o nate, ingentem luctum ne quaere tuorum'.

10. LECTAQVE CVM LACRIMIS SVNT TVA DAMNA MEIS. Compare EP I ix 1-2 (quoted above at 5 rapti) and Fam IV v 1 (Ser. Sulpicius Rufus to Cicero) 'Postea quam mihi renuntiatum est de obitu Tulliae, filiae tuae, sane quam pro eo ac debui grauiter molesteque tuli communemque eam calamitatem existimaui'.

10. TVA DAMNA. Compare Fast II 835-36 (Lucretia has just killed herself) 'ecce super corpus communia damna gementes / obliti decoris uirque paterque iacent' and Tr IV iii 35 'tu uero tua damna dole, mitissima coniunx'.

11. SED NEQVE SOLARI PRVDENTEM STVLTIOR AVSIM. Compare Fam IV v 6 'plura me ad te de hac re scribere pudet, ne uidear prudentiae tuae diffidere'. For the opposite reasoning, see Sen Cons Marc 1 1 'Nisi te, Marcia, scirem tam longe ab infirmitate muliebris animi quam a ceteris uitiis recessisse et mores tuos uelut aliquod antiquum exemplar aspici, non auderem obuiam ire dolori tuo'.

12. VERBAQVE DOCTORVM NOTA. Compare EP I iii 27-30 (to Rufinus, who has written him a letter of consolation on his exile) 'cum bene firmarunt animum praecepta iacentem, / sumptaque sunt nobis pectoris arma tui, / rursus amor patriae ratione ualentior omni, / quod tua fecerunt scripta retexit opus', and Sen Cons Marc 2 1 'scio a praeceptis incipere omnes qui monere aliquem uolunt, in exemplis desinere'.

13-14. FINITVMQVE TVVM ... DOLOREM / IPSA IAM PRIDEM SVSPICOR ESSE MORA. Compare EP I iii 25-26 'cura quoque interdum nulla medicabilis arte est— / aut, ut sit, longa est extenuanda mora', Fam IV v 6 'nullus dolor est quem non longinquitas temporis minuat ac molliat', and Cons Marc 8 1 'dolorem dies longa consumit'. For a variation of the theme, see Cons Marc 1 6 'illud ipsum naturale remedium temporis, quod maximas quoque aerumnas componit, in te una uim suam perdidit'.

The topic of time as the healer of pain is common in ancient literature from New Comedy on: see Tarrant on Sen Ag 130 'quod ratio non quiit, saepe sanauit mora', Otto dies 6, and Kassel 53.

13. SI NON RATIONE. Ratio similarly used to counter strong emotion (without success) at EP I iii 27-30 (quoted at 12), Met VII 10-11 (Medea falls in love with Jason) 'ratione furorem / uincere non poterat', and Met XIV 701-2 (similar phrasing for Iphis' falling in love with Anaxarete).

14. IPSA ... MORA. 'By the mere passage of time'.

15-16. DVM TVA PERVENIENS, DVM LITTERA NOSTRA RECVRRENS / TOT MARIA AC TERRAS PERMEAT, ANNVS ABIT. Similar phrasing at EP III iv 59-60 'dum uenit huc rumor properataque carmina fiunt / factaque eunt ad uos, annus abisse potest'.

15. PERVENIENS is my correction for the manuscripts' peruenit. The perfect tense of peruenit conflicts with the following permeat and abit. It might be argued that the perfect is acceptable, since Ovid is speaking of a past event; but he would not have used the perfect of an action which took place over a considerable period of time. For perueniens ... permeat referring to a past event, compare Ovid's use of the present uenit in the very similar passage EP III iv 59-60 (quoted at the end of the last note).

The postponement of permeat to the following line made the corruption of dum ... perueniens to dum ... peruenit simple enough.

17. TEMPORIS OFFICIVM EST SOLACIA DICERE CERTI. Here Ovid says that words of comfort should not be offered too late; at RA 127-30 he says they should not be offered too early: 'quis matrem, nisi mentis inops, in funere nati / flere uetet? non hoc illa monenda loco est. / cum dederit lacrimas animumque impleuerit aegrum, / ille dolor uerbis emoderandus erit'.

For the same concern with time as in the present passage and medical imagery similar to that in 19-20, see Cons Marc 1 8 and Cons Hel 1 2 'dolori tuo, dum recens saeuiret, sciebam occurrendum non esse, ne illum ipsa solacia irritarent et accenderent; nam in morbis quoque nihil est perniciosius quam immatura medicina. expectabam itaque, dum ipse uires suas frangeret et ad sustinenda remedia mora mitigatus tangi se ac tractari pateretur'. See as well the passages cited at Kassel 52-53: from modern literature he quotes Sterne Tristram Shandy III 29 'Before an affliction is digested consolation ever comes too soon;—and after it is digested—it comes too late: so that you see ... there is but a mark between those two, as fine almost as a hair, for a comforter to take aim at'.

18. DVM DOLOR IN CVRSV EST. Compare RA 119 'dum furor in cursu est, currenti cede furori' and Met XIII 508-10 (Hecuba speaking) 'in cursuque meus dolor est: modo maxima rerum ... nunc trahor exul, inops, tumulis auulsa meorum'.

18. AEGER. The substantive aeger is quite common in both verse and prose, but always with the meaning 'physically ill'; even when used, as here, with a transferred meaning, the sense of metaphor is still present. Compare RA 313-14 'curabar propriis aeger Podalirius herbis, / et, fateor, medicus turpiter aeger eram', EP I iii 17 'non est in medico semper releuetur ut aeger', and EP III iv 7-8 'firma ualent per se, nullumque Machaona quaerunt; / ad medicam dubius confugit aeger opem'.

The adjective, however, is used by the poets from Ennius on (Sc 254 & 392 Vahlen³), particularly in the phrases mens aegra and animus aeger, to indicate a state of mental anguish. Compare, from Ovid, Tr III viii 33-34 'nec melius ualeo quam corpore mente, sed aegra est / utraque pars aeque', Tr IV iii 21, IV vi 43 & V ii 7, EP I iii 89-90 'uereor ne ... frustra ... iuuer admota perditus aeger ope', I v 18 & I vi 15 'tecum tunc aberant aegrae solacia mentis', and Ibis 115; from other poets, compare Cons ad Liuiam 395, Hor Ep I viii 8, and Aen I 208 & IV 35. The same use of the adjective is found occasionally in the historians (Sallust Iug 71 2, Livy II 3 5, etc).

19. LONGA DIES = tempus. Compare Met I 346, Met XIV 147-48 (the Sibyl to Aeneas) 'tempus erit cum de tanto me corpore paruam / longa dies faciet', and Tr I v 11-14 'spiritus et uacuas prius hic tenuandus in auras / ibit ... quam subeant animo meritorum obliuia nostro, / et longa pietas excidat ista die'.

19. VVLNERA MENTIS. Ovid is fond of this metaphorical sense of uulnus; see Met V 425-27 'Cyane ... inconsolabile uulnus / mente gerit tacita', Tr IV iv 41-42 'neue retractando nondum coeuntia rumpam / uulnera: uix illis proderit ipsa quies', EP I iii 87-88 'nec tamen infitior, si possent nostra coire / uulnera, praeceptis posse coire tuis', and EP I v 23 'parcendum est animo miserabile uulnus habenti'. To judge from Seneca, the metaphor was usual in treatises of consolation: 'antiqua mala in memoriam reduxi et, ut scires [Schultess: uis scire codd] hanc quoque plagam esse sanandam, ostendi tibi aeque magni uulneris cicatricem' (Cons Marc 1 5), 'itaque utcumque conabar manu super plagam meam imposita ad obliganda uulnera uestra reptare' (Cons Hel 1 1).

20. FOVET Heinsius MOVET codd. For the meaning of fouet see at 4 fouisti (p 361). Mouet here is to some extent supported by Ovid's use of such verbs as tangere and tractare in contexts like that of the present passage; compare EP I vi 21-22 'nec breue nec tutum peccati quae sit origo / scribere; tractari uulnera nostra timent', EP II vii 13, and EP III vii 25-26 'curando fieri quaedam maiora uidemus / uulnera, quae melius non tetigisse fuit'. But tractare and tangere are neutral in force, while mouet here would mean 'disturb', as at Hor Carm III xx 1-2 'Non uides quanto moueas periclo, / Pyrrhe, Gaetulae catulos leaenae?' and Lucan VIII 529-30 'bustum cineresque mouere / Thessalicos audes bellumque in regna uocare?'. As Professor R. J. Tarrant comments, if mouet were read in the present passage, intempestiue would lose the appropriateness it has when fouet is read: there is no proper time to "disturb" a wound.

20. NOVAT. Similar phrasing at Tr II 209 'nam non sum tanti renouem ut tua uulnera, Caesar' and RA 729-30 'admonitu refricatur amor, uulnusque nouatum / scinditur'.

21. ADDE QVOD. Professor E. Fantham points out to me how extraordinary the occurrence of this phrase in the last distich of the poem is. Of the twenty-five instances of the idiom in Ovid's poems[25], none except the present passage occur in the final distich of a poem or book. The other examples all occur in the middle of an argument, or lead into another distich containing a final injunction or proof of an argument. As Professor J. N. Grant suggests to me, this poem therefore furnishes another example of Ovid's favourite device of unexpectedly altering a poem's tone in the final distich, for a discussion of which see at xiv 61-62 (p 427).

21. MIHI BF¹ TIBI MHILTF² om C. As Burman saw, mihi must be the correct reading, the perfect subjunctive acting as a past optative: 'certe ego mihi praeferrem: utinam mihi, mentionem facienti noui tui coniugii, uerum illud omen uenerit, neque fallar, sed tu iam uxorem duxeris, ut ego uoueo'. Tibi is hardly possible, since an omen to Gallio indicating that he had remarried would be superfluous.

XII. To Tuticanus

Tuticanus[26] (known only from the Ex Ponto) seems from the testimony of the poem (19-30) to have been a close friend of Ovid; he is mentioned again at xiv 1-2 and xvi 27. It is reasonable to suppose that, like Sextus Pompeius, he had previously been unwilling to allow Ovid to mention him in his verse.

The poem opens with a discussion of the difficulty of fitting Tuticanus' name into elegiac verse: Ovid could split the name between verses, or alter the quantity of one or another of the name's syllables, but neither procedure would be acceptable to Ovid or to his readers (1-18). He has known Tuticanus since early youth; they assisted each other in their verse (19-30). He is quite certain that Tuticanus will not desert him (31-38). He should use his influence with Tiberius to assist Ovid; but Ovid is so confused after his hardships that he cannot suggest precisely what Tuticanus should do; he leaves this to Tuticanus' judgment (39-50).

The appeal for assistance is a constant theme of the poetry of exile; and the recalling of their assisting each other with their poetry is paralleled by EP II iv, in which Ovid recalls how he used to submit his verse to Atticus for criticism, and by Tr III vii, Ovid's letter to his stepdaughter Perilla, whom he assisted when she first began writing verse. The opening discussion of the metrical difficulty of Tuticanus' name finds parallels elsewhere in Latin and Greek literature (see at 1-2), but is remarkable for its fullness. The explanation for this fullness may well be Tuticanus' being a fellow poet: he would be amused by the use of his own name for the witty discussion of the handling of metrical difficulties with which he himself would be familiar enough.

1-2. QVOMINVS IN NOSTRIS PONARIS, AMICE, LIBELLIS, / NOMINIS EFFICITVR CONDICIONE TVI. A constant problem for the Latin poets was the impossibility of using words with cretic patterns (a long syllable, followed by a short syllable, followed by another long syllable) in hexameter or elegiac verse. The fact played an important part in determining Latin poetic vocabulary; for instance, such an ordinary word as femina, cretic in its oblique cases, is usually represented through metonymy by such words as nurus and mater. Proper names presented a special problem, which could however occasionally be solved through the use of special forms or circumlocutions; hence such lines as 'cumque Borysthenio liquidissimus amne [=Bory̅sthĕnē] Dirapses' (x 53) and 'Scipiadas [=Scīpĭōnes], belli fulmen, Carthaginis horror' (Lucretius III 1034). Sometimes, as in the present passage, such avenues were not available, and the poet was simply unable to use the name he wanted. From Greek authors Marx, commenting on Lucilius 228-29, cites Critias fr. 5 'οὐ γάρ πως ἦν τοὔνομα ἐφαρμόζειν ἐλεγείῳ' Archestratus fr. 29 (Brandt) 'ἰχθύος αὐξηθέντος ὃν ἐν μέτρῳ οὐ θέμις εἰπεῖν' and Ep Gr 616 (Kaibel) 'οὐ γὰρ ἐν ἑξαμέτροισιν ἥρμοσεν τοὔνομ' ἐμόν' In Latin, the best-known reference to this difficulty is Hor Sat I v 86-87 'quattuor hinc rapimur uiginti et milia raedis, / mansuri oppidulo, quod uersu dicere non est'. On the passage Porphyrion comments 'Aequum Tuticum significat [this is disputed by modern commentators, since the town's known location does not fit with Horace's indication; no certain candidate has been proposed], cuius nomen hexametro uersu compleri [codd: contineri fort legendum] non potest. hoc autem sub exemplo Lucili posuit. nam ille in sexto Saturarum [228-29 Marx] sic ait: "seruorum est festus dies hic, / quem plane hexametro uersu non dicere possis"'. In his comment on the passage from Horace, Lejay cites Martial IX xi 10-17 (Martial wanted to mention Flavius Ĕărĭnus, whose name starts with three consecutive short vowels) 'nomen nobile, molle, delicatum / uersu dicere non rudi uolebam: / sed tu, syllaba contumax, rebellas. / dicunt Eiarinon tamen poetae, / sed Graeci, quibus est nihil negatum, / et quos Ἆρες Ἄρες decet sonare: / nobis non licet esse tam disertis / qui Musas colimus seueriores', Rutilius Namatianus 419-22 (of Vŏlŭsĭanus [short 'o', 'u', and 'i'] Rufius) 'optarem uerum complecti carmine nomen, / sed quosdam refugit regula dura pedes. / cognomen uersu ueheris [Préchac: ueneris uel uenens codd], carissime Rufi; / illo te dudum pagina nostra canit', and Apollinaris Sidonius Carm XXIII 485-86 'horum nomina cum referre uersu / affectus cupiat, metrum recusat'.

Professor C. P. Jones cites the discussion at Pliny Ep VIII iv 3-4. Pliny, writing to Caninius, who is composing a poem in Greek on the Dacian war, discusses the difficulty of using barbara et fera nomina in the poem: 'sed ... si datur Homero et mollia uocabula et Graeca ad leuitatem uersus contrahere extendere inflectere, cur tibi similis audentia, praesertim non delicata sed necessaria, non detur?'.

For a further discussion of the topic, see L. Radermacher, "Das Epigramm des Didius", SAWW 170,9 [1912] 1-31.

1. QVOMINVS is rare in Augustan verse; but compare AA II 720 'non obstet tangas quominus illa [sc loca] pudor'.

3. AVT BC AST MFHILT. The false reading was probably induced by a failure to understand the meaning of aut 'otherwise', for which compare iii 21 'aut age, dic aliquam quae te mutauerit iram', Met VII 699, Met X 50-52 'hanc [sc Eurydicen] simul et legem Rhodopeius accipit heros, / ne flectat retro sua lumina donec Auernas / exierit ualles; aut inrita dona futura', and Tr I viii 43-45 'quaeque tibi ... dedit nutrix ubera, tigris erat. / aut mala nostra minus quam nunc aliena putares'.

2. CONDICIONE. 'Nature'. Compare Lucretius II 300-1 'et quae consuerint gigni gignentur eadem / condicione et erunt et crescent uique ualebunt'.

4. SI MODO. 'If, that is ...' Compare 43-44 'quid mandem quaeris? peream nisi dicere uix est, / si modo qui periit ille perire potest'.

5. LEX PEDIS. 'The rules of metre'. Lex used similarly at Hor Carm IV ii 10-12 'per audaces noua dithyrambos / uerba deuoluit numerisque fertur / lege solutis', Cic Or 58 'uersibus est certa quaedam et definita lex', and Columella XI 1 1.

5. FORTVNAQVE. The sense of the word is difficult. It seems, as Professor R. J. Tarrant notes, to combine the idea of 'condition, state' (compare for example Aen II 350 'quae sit rebus fortuna uidetis') with that of 'unfortunate circumstances', giving the general sense 'the fact that you have the bad luck to possess a metrically impossible name'. Three lines before, Ovid used nominis ... condicione tui; and in the present line he seems to have been influenced by the common phrase condicio et fortuna, 'allotted circumstances of life', for which compare Cic Off I 41 'est autem infima condicio et fortuna seruorum', Mil 92 'in infimi generis hominum condicione atque fortuna'. At II Verr I 81 Cicero similarly adapts the expression to suit his context: 'Lampsacenis ... populi Romani condicione sociis, fortuna seruis, uoluntate supplicibus'.

7. NOMEN SCINDERE. That is, split the name so that the hexameter (uersus prior) would end in Tūtĭ- and the following pentameter (uersus minor) begin with -cānŭs. Such word-divisions are not permissible in Augustan verse; from earlier poetry Professor C. P. Jones cites Ennius Ann 609 Vahlen³ 'saxo cere comminuit brum'.

8. HOC = nomine tuo.

9-14. Ovid lists the three possible ways of scanning the name so as to remove the cretic: Tūtĭcănus, Tŭtĭcānus, and Tūtīcānus.

9. MORATVR = longa est. The TLL cites Velius Longus VII 55 5 Keil 'hanc ... naturam esse quarundam litterarum, ut morentur et enuntiatione sonum detineant'.

11. ET BCHI^acLT NON M NEC FI^pc. Nec, printed by some editors, cannot by itself be correct, for there is no negative with the corresponding producatur in the following distich. A negative is implicitly supplied for potes ... uenire and producatur by 15-16 'his ego si uitiis ...', but Professor R. J. Tarrant is possibly right to suggest that nec should be read both here and (replacing aut) at the beginning of 13.

W. A. Camps (CQ n.s. IV [1954] 206-7) has pointed out that it is somewhat odd that 'The first two possibilities are introduced, in lines 7 and 9, in terms that disclaim them at once' and that 'the third and fourth possibilities are added without disclaimer ... in terms that would be quite appropriate to serious suggestions'. He suggests reading at, so that 11-12 represent an imaginary rejoinder to Ovid's rejection of the possibilities already suggested; Ovid's rejoinder is given at 15 'his ego si uitiis ...'. But at potes is difficult: Ovid could have written 'at, puto, potes', speaking in his own person to raise an objection he would then counter, or he could have represented Tuticanus as saying 'at ... possum'; but it is hard to see how he could have written 'at potes'.

13. PRODVCATVR MHI VT DVCATVR LTB²F^2ul VT DICATVR B¹CF¹. Producere is the correct technical term for 'lengthen'; compare Quintilian VII ix 13 'productio quoque in scripto et correptio in dubio relicta causa est ambiguitatis' & IX iii 69 'uoces ['words'] ... productione tantum uel correptione mutatae'. Vt ducatur is unlikely to be right. Ducatur could certainly stand for producatur (although this would destroy the balance with the following correptius), but the verb is clearly indicated as a potential subjunctive by the preceding potes ... uenire; and ut (which would in any case be taken as correlative with ut in line 12) cannot stand with this construction. Vt dicatur, Ehwald's preferred reading ('dicatur et sit secunda [syllaba] productâ morâ longa'—KB 68), is even less likely to be right, since dicere in this context could only mean 'pronounce', as at Cic Or 159 '"inclitus" dicimus breui prima littera, "insanus" producta'.

13. EXIT. Exire similarly used of words being uttered at Her VIII 115-16 (Hermione speaking) 'saepe Neoptolemi pro nomine nomen Orestae / exit, et errorem uocis ut omen amo'. OLD exeo 2d gives other instances from Cicero (Brutus 265), Seneca (Ben V 19 4), and Quintilian (XI iii 33), but from verse outside Ovid only Martial XII xi 3, where the word has a somewhat different meaning: 'cuius Pimpleo lyra clarior exit ab antro?'.

14. PORRECTA is equivalent to longa, and belongs to secunda (sc syllaba) by hypallage. Compare Quintilian I vi 32 'aut correptis aut porrectis ... litteris syllabisue' & I vii 14 'usque ad Accium et ultra porrectas syllabas geminis, ut dixi, uocalibus scripserunt [that is, they wrote uiita for uita and so on; such spellings occur sometimes in inscriptions]', and Rutilius Lupus I 3.

15. VITIIS. Vitium similarly used for faults of diction at AA III 295-96 'in uitio decor est: quaerunt male reddere uerba; / discunt posse minus quam potuere loqui', Cic de Or I 116, and Quintilian I v 17, a discussion of the shortening and lengthening of vowels; this he includes among the 'quae accidunt in dicendo uitia'. Ovid is probably combining this sense with that of 'poetic weakness', for which compare Tr I vii 39-40 'quicquid in his igitur uitii rude carmen habebit, / emendaturus, si licuisset, eram' and the use of uitiosus at xiii 17 and Tr IV i 1 and IV x 61.

16. MERITO PECTVS HABERE NEGER. 'People would quite rightly say that I was ignorant'. Compare Met XIII 290-91 & 295 (Ulysses is speaking of Ajax's claim to the arms of Achilles) 'artis opus tantae rudis et sine pectore miles / indueret? neque enim clipei caelamina nouit ... postulat ut capiat quae non intellegit arma!'.

17-18. MVNERIS ... QVOD MEVS ADIECTO FAENORE REDDET AMOR. Adiecto faenore = 'with interest added on'; Ovid will make up for his past negligence by sending Tuticanus more than one poem ('tibi carmina mittam'). It is clear from the opening distich of poem xiv that Ovid sent the poem to Tuticanus very soon after the composition of xii: 'Haec tibi mittuntur quem sum modo carmine questus / non aptum numeris nomen habere meis'.

A similar use of faenus at EP III i 79-81 'nec ... debetur meritis gratia nulla meis. / redditur illa quidem grandi cum faenore nobis'.

The variant AGER (TM²I²) for amor was clearly induced by such passages as Tib II vi 21-22 'spes sulcis credit aratis / semina quae magno faenore reddat ager', RA 173-74 'obrue uersata Cerealia semina terra, / quae tibi cum multo faenore reddat ager', and EP I v 25-26 'at, puto ... sata cum multo faenore reddit ager': these passages refer to the original meaning of faenus ('faenum appellatur naturalis terrae fetus; ob quam causam et nummorum fetus faenus est uocatum'—Festus 94 Muller, 83 Lindsay).

18. REDDET GCMIT REDDIT BFHL. Numerous instances of similar corruptions in Lucan and Juvenal given by Willis (166-67), who remarks 'The general trend seems to be from other tenses to the present, and from other persons and numbers to the third person singular'.

19. QVACVMQVE NOTA. 'With whatever method of indicating your name is possible'. For the collocation of nota and nomen, see Aen III 443-44 'insanam uatem aspicies, quae rupe sub ima / fata canit foliisque notas et nomina mandat'.

Luck joins the phrase with the following tibi carmina mittam, but the construction seems somewhat cumbersome; it is probably better to retain the comma after nota and take the phrase with teque canam.

20-22. PVERO ... PVER ... FRATRI FRATER. For Ovid's use of polyptoton, see at viii 67 uatis ... uates (p 278).

23. DVXQVE COMESQVE. The same phrase at Tr III vii 18 (to his stepdaughter Perilla) 'utque pater natae duxque comesque fui' and Tr IV x 119-20 (to his Muse) 'tu dux et comes es, tu nos abducis ab Histro, / in medioque mihi das Helicone locum'.

24. FRENA NOVELLA. For the image, see at ii 23 frena remisi (p 169). Nouellus is a rare word in poetry. In prose, the word is often used of young plants or farm animals; and here frena nouella may well be a metonymy for frena nouellorum equorum. Alternatively, the word could be equivalent to noua 'new, unfamiliar', as at Fast III 455 'iamque indignanti noua frena receperat ore'. In either case, Ovid is clearly referring to the beginning of his poetic career.

25. SAEPE EGO CORREXI SVB TE CENSORE LIBELLOS. Compare Tr III vii 23-24 (to Perilla) 'dum licuit, tua saepe mihi, tibi nostra legebam; / saepe tui iudex, saepe magister eram'. Censore was probably still felt as a metaphor; the only precedent given at OLD censor 2b is Hor Ep II ii 109-10 'at qui legitimum cupiet fecisse poema / cum tabulis animum censoris sumet honesti', which is virtually a simile.

26. SAEPE TIBI ADMONITV FACTA LITVRA MEO EST. Similar phrasing in a similar context at EP II iv 17-18 (to Atticus) 'utque meus lima rasus liber esset amici, / non semel admonitu facta litura tuo est'.

27. DIGNAM MAEONIIS PHAEACIDA ... CHARTIS. 'A Phaeacid worthy of the Homeric original you were translating'. It is clear from xvi 27 that Tuticanus produced a translation rather than a new work in imitation of Homer: 'et qui Maeoniam Phaeacida uertit'.

27. MAEONIIS = 'Homeric', Homer being considered a native of Maeonia (Lydia). The same use at RA 373 'Maeonio ... pede', EP III iii 31-32 'Maeonio ... carmine', and Prop II xxviii 29 'Maeonias ... heroidas'; the word in this sense perhaps brought into standard poetic vocabulary by Horace (Carm I vi 2 'Maeonii carminis', Carm IV ix 5-6 'Maeonius ... Homerus').

27. CHARTIS = carminibus. Compare AA II 746 'uos eritis chartae proxima cura meae'. The metonymy is not found in Virgil or Propertius, but compare Lucretius IV 970 'patriis ... chartis' = 'Latinis uersibus' (I 137) and Hor Carm IV ix 30-31 'non ego te meis / chartis inornatum silebo' (where Kiessling-Heinze point out that chartis refers to the poem in its published state being transmitted to others, rather than to the poem at its moment of composition).

28. CVM TE PIERIAE PERDOCVERE DEAE. For the poet's being divinely taught, compare Prop II x 10 & IV i 133, Her XV 27-28 'at mihi Pegasides blandissima carmina dictant; / iam canitur toto nomen in orbe meum', and the disclaimers at Prop II i 3 and AA I 25-28 'non ego, Phoebe, datas a te mihi mentiar artes, / nec nos aeriae uoce monemur auis, / nec mihi sunt uisae Clio Cliusque sorores / seruanti pecudes uallibus, Ascra, tuis'. The topic is an important one in ancient literature, the most influential passages being the opening of Hesiod's Theogony (referred to in the passage just cited) and the beginning of Callimachus' Aetia.

29. TENOR. 'Course'; the same use at Her VII 111-12 (Dido speaking) 'durat in extremum uitaeque nouissima nostrae / prosequitur fati qui fuit ante tenor'.

29. VIRIDI ... IVVENTA. Ovid is perhaps imitating Aen V 295 'Euryalus forma insignis uiridique iuuenta'. Similar phrasing at AA III 557 'uiridemque iuuentam', Tr IV x 17 'frater ad eloquium uiridi tendebat ab aeuo', and Tr III i 7-8 'id quoque quod uiridi quondam male lusit in aeuo / heu nimium sero damnat et odit opus'; at the last passage Luck aptly cites Met XV 201-3 'nam tener et lactens puerique simillimus aeuo / uere nouo [sc annus] est; tunc herba nitens et roboris expers turget'.

30. ALBENTES ... COMAS. For the synecdoche compare Callimachus Ep LXIV (=Anth Pal V xxiii) 5-6 'ἡ πολιὴ δὲ / αὐτίκ' ἀναμνήσει ταῦτά σε πάντα κόμη'.

Ovid would have been about sixty years of age at the time of this poem, old by Roman standards; but his father lived to ninety, and was survived by his wife (Tr IV x 77-80).

30. INLABEFACTA occurs in classical Latin only here and at viii 9-10 'ius aliquod faciunt adfinia uincula nobis / (quae semper maneant inlabefacta precor)'.

31-32. QVAE NISI TE MOVEANT, DVRO TIBI PECTORA FERRO / ESSE VEL INVICTO CLAVSA ADAMANTE PVTEM. Compare Her II 137 'duritia ferrum ut superes adamantaque teque', Her X 109-10, and Met IX 614-15 (Byblis on her brother) 'nec rigidas silices solidumue in pectore ferrum / aut adamanta gerit'.

Professor R. J. Tarrant notes the unexpected shift in the thought of the poem: earlier it was Ovid who was guilty of delaying in sending Tuticanus any sign of his friendship. Ovid might be postponing the real point of the letter for reasons of tact: Tuticanus has acted as though his long association with Ovid meant nothing to him, but Ovid does not want to complain of this openly, and so stresses his own failure to send Tuticanus a letter.

33-36. The set of adynata is remarkable for the way Ovid makes each of them relate to his own hardships; even Boreas and Notus have a specific connection, since Ovid complains so often of the climate of Tomis.

35. TEPIDVS BOREAS ... SIT. A comparable inversion of nature described at Ibis 34 'et tepidus gelido flabit ab axe Notus' (before Ovid will forgive his enemy).

35. PRAEFRIGIDVS appears here for the first time in Latin; it occurs later in Celsus and the elder Pliny. Praegelidus, however, is found at Livy XXI 54 7.

36. ET POSSIT FATVM MOLLIVS ESSE MEVM. The personal reference in the last element of the series of adynata is a clear break with the conventions of the topic. The last (and therefore greatest) curse in the Ibis has a similar personal reference: 'denique Sarmaticas inter Geticasque sagittas / his precor ut uiuas et moriare locis'.

37. LAPSO FHILT LASSO BCM. Lapso ... sodali seems to me the preferable reading, since it contrasts Ovid's former life in Rome with his disgrace and exile; but lasso is well attested and can be construed easily enough. Unfortunately, parallels from the poems of exile are of little use, since in most of them the one word could easily be read for the other: 'tu quoque magnorum laudes admitte uirorum, / ut facis, et lapso [uar lasso] quam potes adfer opem' (EP II iii 47-48), 'fac modo permaneas lasso [uar lapso], Graecine, fidelis, / duret et in longas impetus iste moras' (EP II vi 35-36), 'regia, crede mihi, res est succurrere lapsis [uar lassis], / conuenit et tanto, quantus es ipse, uiro' (EP II ix 11-12), 'digne uir hac serie, lapso [uar lasso] succurrere amico / conueniens istis moribus esse puta' (EP III ii 109). Professor R. J. Tarrant cites similar variants in the text of Seneca at HF 646 & 803 and Thy 616 & 658.

A clear decision can be made, however, for the phrase res lassae; it is certified as the correct term by the parallel phrase res fessae, for which see Aen III 145 'quam fessis finem rebus ferat' and Aen XI 335 'consulite in medium et rebus succurrite fessis', cited by Luck at Tr I v 35. For res lassae in Ovid, compare Tr I v 35 'quo magis, o pauci, rebus succurrite lassis', Tr V ii 41 'unde petam lassis solacia rebus?', EP II ii 47 'nunc tua pro lassis nitatur gratia rebus', and EP II iii 93 'respicis antiquum lassis in rebus amicum'; in each of these passages lapsis is found as a variant for lassis. Similarly, the sixth-century codex Romanus reads lapsis at Virgil G IV 449 'uenimus hinc lassis quaesitum oracula rebus'.

38. HIC CVMVLVS NOSTRIS ABSIT ABESTQVE MALIS. Festus defines cumulus as a heap added to an already full measure (s.u. auctarium, 14 Muller, 14 Lindsay). The transferred sense is common in Cicero (Prou Cons 26, S Rosc 8, Att XVI iii 3), and is found elsewhere in Ovid at EP II v 35-36 'hoc tibi facturo uel si non ipse rogarem / accedat cumulus gratia nostra leuis' and Met XI 205-6 'stabat opus: pretium rex infitiatur et addit, / perfidiae cumulum, falsis periuria uerbis'.

38. ABSIT ABESTQVE. The more natural abest absitque cannot be placed in a pentameter.

39. PER SVPEROS, QVORVM CERTISSIMVS ILLE EST. Similar line-endings at Ibis 23-24 'di melius! quorum longe mihi maximus ille est, / qui nostras inopes noluit esse uias' and EP I ii 97-98 'di faciant igitur, quorum iustissimus ipse est, / alma nihil maius Caesare terra ferat'.

40. QVO ... PRINCIPE. Professor R. J. Tarrant points out that Augustus must here be meant, since it appears from 20 that Ovid and Tuticanus were contemporaries: Tuticanus must by the time of the poem's writing have been in later middle age, rather late to be prospering only under Tiberius. T. P. Wiseman (268) has suggested that Ovid's Tuticanus might be the son of a Tuticanus Callus known to have been senator before 48 BC.

41-42. EFFICE ... NE SPERATA MEAM DESERAT AVRA RATEM. 'See to it that the breeze I hope for does not fail to come to my ship'. Deserere generally refers to something failing one that was originally operative: compare Cic Att VII vii 7 'nisi me lucerna desereret' ('if the lamp were not going out'—Shackleton Bailey), Plautus Mer 123 'genua hunc cursorem deserunt' and the other passages cited at OLD desero 2b. But sperata indicates that the breeze cannot yet be present; other instances of the same metaphor at viii 27-28 'quamlibet exigua si nos ea iuuerit aura, / obruta de mediis cumba resurget aquis', ix 73 'et si quae dabit aura sinum, laxate rudentes', and Tr IV v 19-20 'utque facis, remis ad opem luctare ferendam, / dum ueniat placido mollior aura deo',

43. QVID MANDEM QVAERIS. Similar wording at EP III i 33-34 (to his wife) 'quid facias quaeris? quaeras hoc scilicet ipsa [Riese: ipsum codd]: / inuenies, uere si reperire uoles'.

Ovid's pretense of not knowing what to tell Tuticanus to do was an ingenious solution to his friends' complaint that he was constantly repeating the same instructions to them (EP III vii 1-6). Professor R. J. Tarrant points out the balance with the poem's start, where Ovid pretends not to know how to address Tuticanus.

43. PEREAM NISI DICERE VIX EST. Similar doubt expressed at Tr IV iii 31-32 'quid tamen ipse precer dubito, nec dicere possum / affectum quem te mentis habere uelim'. Peream nisi, which Ovid plays on in the next line, is colloquial and foreign to poetic diction: instances at OLD pereo 3b.

44. SI MODO QVI PERIIT ILLE PERIRE POTEST. Similar phrasing at Tr I iv 27-28 'uos animam saeuae fessam subducite morti, / si modo qui periit non periisse potest'.

45. NEC QVID NOLIMVE VELIMVE. Compare Met XI 492-93 'nec se ... fatetur / scire ratis [codd: satis fort scribendum] rector ... quid iubeatue uetetue' and Tr I ii 31-32 'rector in incerto est nec quid fugiatue petatue / inuenit'.

46. NEC SATIS VTILITAS EST MIHI NOTA MEA. 'And I am at a loss to know what is to my advantage'. Satis strengthens the sentence: compare Ter Hec 877 'ego istuc sati' scio', 'I know that very well'. For utilitas, see at ix 48 publica ... utilitas (p 300).

48. SENSVS here means 'judgement' or 'good sense', as at Prop II xii 3 'is primum uidit sine sensu uiuere amantes' and Val Max I vi ext 1 'si quod uestigium in uecordi pectore sensus fuisset'. Elsewhere in Ovid sensus carries the meaning 'awareness, consciousness'.

48. CVM RE codd CVM SPE Heinsius. Cum re, 'along with my fortune', seems somewhat out of place; but Burman pointed out that consilium et res seems to have been a Latin phrase, citing Sallust Iug 74 'neque illi res neque consilium aut quisquam hominum satis placebat' and Ter Eun 240-41 'itan parasti te ut spes nulla relicua in te siet tibi? / simul consilium cum re amisti?'.

50. QVAQVE VIA VENIAS AD MEA VOTA, VIDE. This is a provisional restoration of the line. The manuscript reading which most closely approaches this text is that of L and F³, QVAQVE VIAM FACIAS AD MEA VOTA, VIDE; the other manuscripts have the same text, except that QVOQVE is found in some for quaque, while for uide there are the variants MODO, VADO, and VALE.

My restoration is based on 6 'quaque meos adeas est uia nulla modos' and Fast I 431-32 (Priapus approaches the sleeping nymph Lotis) 'a pedibus tracto uelamine uota / ad sua felici coeperat ire uia'.

Before Professor E. Fantham brought this passage to my attention, I had thought that M's quoque uiam facias ad mea uota modo was correct. Modo is weak and does not fit well with the preceding qua ... parte, but at least is acceptable Latin; for quo ... modo compare Med 1-2 'Discite quae faciem commendet cura, puellae, / et quo sit uobis forma tuenda modo' and Ibis 55-56 'nunc quo Battiades inimicum deuouet Ibin, / hoc ego deuoueo teque tuosque modo'.

The image in quoque ... uado ['ford'] is rather strange, and for this sense of the word Ovid seems to have used the plural (Met III 19; Met IX 108). At Fast IV 300 'sedit limoso pressa carina uado', uado means 'river-bottom'.

Ovid does not end any one of his dozens of verse epistles with uale, so the reading of FTI^2ul must be discounted.

If my restoration is correct or nearly correct, the original corruptions would have been of uia to uiam and of uenias to facias; the latter corruption might have been a deliberate interpolation to procure a governing verb for uiam, or might have been a misreading of or conjectural restoration for a damaged archetype. The variant quoque for quaque and the different variants for uide would have been secondary corruptions, unless they also were the result of a damaged archetype.

50. VIDE. For uide at the end of the pentameter, compare EP II ii 55-56 'num tamen excuses erroris origine factum, / an nihil expediat tale mouere, uide'. It must however be said that uide is somewhat strange following the subjunctive quaeras.

XIII. To Carus

Certain elements of the poem, such as the flattering references to Carus' poetry and the request for his help, are commonplaces of the poetry of exile; the list of the members of the imperial family is similarly paralleled in Ovid's other poems (see at 25-32 [p 400]). Ovid nowhere else explicitly describes any of his Getic poems.

1. MEMORANDE BMFHILT NVMERANDE C. For memorande compare Tr I v 1 'O mihi post nullos umquam memorande sodales'. Numerande is in itself acceptable enough: see ix 35 'hic ego praesentes inter numerarer amicos'.

2. QVI QVOD ES, ID BCFI QVI QVOD ID ES MH QVIQVE QVOD ES LT. For the use of id, Ehwald (KB 47) cited Fast II 23-24 'quaeque capit lictor domibus purgamina uersis ['swept out'] / torrida cum mica farra, uocantur idem [sc februa]', Hor Sat II iii 139-41 (of Orestes) 'non Pyladen ferro uiolare aususue sororem / Electram, tantum male dicit utrique uocando / hanc Furiam, hunc aliud', Sen Ben I 3 10 'id quemque uocari iubent', and Tac Germ 6 'definitur et numerus: centeni ex singulis pagis sunt, idque ipsum inter suos uocantur' ['they are called "The Hundred"']'.

Quique quod es is, however, an attractive reading: compare Tr I v 1-2 'O mihi post nullos umquam memorande sodales, / et cui praecipue sors mea uisa sua est'. Quique quod is obviously prone to haplography; on the other hand, it could be a rewriting of qui quod id es, which is itself presumably a simple corruption through interchange of qui quod es id. I therefore print qui quod es id, although with some hesitation.

2. VERE. 'Justly'. For the same adverb used once again of names "properly" applied, see Tr V x 13-14 'quem tenet Euxini mendax cognomine litus, / et Scythici uere terra sinistra freti'.

2. CARE. Luck among others believes that Carus is also addressed at Tr III v 17-18 'sum quoque, care, tuis defensus uiribus absens / (scis "carum" ueri nominis esse loco)'; but it seems excessively ingenious to make Ovid say 'I call you carus instead of your real name, Carus'. Still, as Professor R. J. Tarrant points out to me, the passage is odd, in that Ovid elsewhere uses care only in conjunction with another vocative (compare viii 89 'care Suilli' and Tr III iv 1-2 'care quidem ... sed tempore duro / cognite'); care may have been used as a metrical equivalent to the suppressed name, in the way the "cover names" in elegy correspond to the shape of the alleged actual names of the women. Unlike care, carissime is often found by itself (Tr I v 3, III iii 27, III vi 1, IV vii 19 & V vii 5; EP II iv 21 & IV x 3).

2. AVE occurs in Ovid only here and at RA 639-40 'nec ueniat seruus, nec flens ancillula fictum / suppliciter dominae nomine dicat "aue!"', and is not common in writing. It was, however, frequent in everyday speech, as is clear from Sen Ben VI 34 3 'uulgare et publicum uerbum et promiscuum ignotis "aue"'.

3. SALVTERIS MFT SALVTARIS BCHIL. Ovid usually employs the subjunctive in indirect questions; this is demonstrated by metre at such passages as Fast VI 385-86 'increpat illos / Iuppiter et sacro quid uelit ore docet', Tr II 294 'sustulerit quare quaeret Ericthonium', Tr II 297-98 'Isidis aede sedens cur hanc Saturnia quaeret / egerit Ionio Bosphorioque mari', Tr V xiv 1-2 'Quanta tibi dederim nostris monumenta libellis ... uides', EP I i 55-56 'talia caelestes fieri praeconia gaudent, / ut sua quid ualeant numina teste probent' and EP II vii 3 'subsequitur quid agas audire uoluntas'.

I have found two passages where metre demonstrates that Ovid used the indicative in an indirect question, Met X 637 'quid facit [codd plerique: quod facit recc quidque agat Heinsius quid factum Merkel quid uelit Nick quid facti Rappold dissidet Korn quid sciat Slater] ignorans amat et non sentit amorem' and EP I viii 25-26 'sed memor unde abii queror, o iucunde sodalis, / accedant nostris saeua quod arma malis'. But in the first passage faciat would have an ambiguous meaning, since it could represent either quid facio or quid faciam, and in the second ăbĭĕrim with its short 'a', 'i', and 'e', would be metrically intractable.

It is difficult to say whether the scribes were more prone to influence by the subjunctive normal in classical Latin prose, or by the indicative of the Romance languages and of ecclesiastical Latin. I print the subjunctive in view of Ovid's usual practice, and in particular because of EP I ii 5 'forsitan haec a quo mittatur epistula quaeras' and EP III v 1 'Quam legis unde tibi mittatur epistula quaeris?'. But Professor R. J. Tarrant notes that the need for a dependent subjunctive would be more strongly felt with quaerere in these two passages than with the index of the present passage.

Not all poets were as strict as Ovid in using the subjunctive in indirect questions. Propertius at III v 26-46 has the following verbs in a series of indirect questions: temperet, uenit, deficit, redit, superant, captet, sit uentura, bibit, tremuere, luxerit (from lugere), coit, exeat, eat, sint (uar sunt), furit, custodit, descendit, potest.

3. COLOR HIC. 'The style of this opening'. Ovid is presumably referring to its playful tone. Compare Tr I i 61 (to his poem) 'ut titulo careas, ipso noscere colore', at which Luck cites Martial XII ii 17-18 'quid titulum poscis? uersus duo tresue legantur, / clamabunt omnes te, liber, esse meum'.

Color is not found in precisely this sense until Horace. For a discussion of its development, see Brink at Hor AP 86 operumque colores.

4. STRVCTVRA. This passage is the first instance cited by OLD structura 1b of structura in this transferred sense, which becomes common in Silver prose, particularly Quintilian (I x 23, VIII vi 67, IX iv 45). Lewis and Short point out that Cicero uses the word in similar contexts only as a simile: compare Brut 33 'ante hunc [sc Isocratem] enim uerborum quasi structura et quaedam ad numerum conclusio nulla erat', Or 149 'quasi structura quaedam', and Opt Gen 5 'et uerborum est structura quaedam'.

There are two instances in Ovid of struere with a similar meaning, both from the Ex Ponto. One is from line 20 of this poem ('structa ... uerba'), while the other is at II v 19 'structos inter fera proelia uersus'.

5. MIRIFICA is a colloquialism. Common in the letters of Cicero, the word (according to TLL VIII 1060 52) is not found in Livy, Vitruvius, Celsus, Curtius, or Tacitus. The only poets apart from Terence and Ovid cited as using the word are Accius, Ausonius, and the author of the Ciris (although the passage where the word occurs, 12-13, is corrupt); see also Catullus LIII 2, LXXI 4, and LXXXIV 3. For a discussion of mirificus, see Axelson 61, and of the similarly colloquial mirifice Hofmann 78.

5. PVBLICA = 'usual, ordinary'. Compare Am III vii 11-12 'et mihi blanditias dixit dominumque uocauit, / et quae praeterea publica uerba iuuant', AA III 479-80 'munda, sed e medio consuetaque uerba, puellae, / scribite: sermonis publica forma placet', and Sen Ben VI 34 3 (quoted at 2 aue).

6. QVALIS ENIM CVMQVE EST. A common phrase in the poets when they speak of their own verse: compare Catullus I 8-9 'quare habe tibi quidquid hoc libelli / qualecumque', Hor Sat I x 88-89 'quibus [sc amicis] haec, sunt qualiacumque, / arridere uelim, doliturus, si placent spe / deterius nostra' (at which Bentley cited the present passage), Martial V lx 5 'qualiscumque legaris ut per orbem', and Statius Sil II praef 'haec qualiacumque sunt, Melior carissime, si tibi non displicuerint, a te publicum accipiant; sin minus, ad me reuertantur' (both passages cited by Munro, Criticisms 5).

7. VT TITVLVM CHARTAE DE FRONTE REVELLAS. The same hypothetical case at Tr I i 61-62 'ut titulo careas, ipso noscere colore; / dissimulare uelis, te liquet esse meum' and EP II ix 49-52 (to King Cotys) 'nec regum quisquam magis est instructus ab illis [sc the liberal arts] ... carmina testantur, quae si tua nomina demas / Threicium iuuenem composuisse negem'.

7. CHARTAE. See at xii 27 chartis (p 380).

7. REVELLAS 'tear away' is surprisingly strong in its overtones. It is found only here in the poems of exile, six times in the other elegies, and fifteen times in the Metamorphoses.

8. QVOD SIT OPVS VIDEOR DICERE POSSE TVVM. 'I think I could say which work was yours'. Heinsius' QVID SIT OPVS VIDEAR is a strange error: the interrogative adjective is acceptable enough, while the notion of the subjunctive must of course be contained in posse, not in the verb that governs it.

11. PRODENT AVCTOREM VIRES. 'His strength will reveal the poet's identity'. The same sense of prodere at Met II 433 'impedit amplexu nec se sine crimine prodit', Met XIV 740-41 'adapertaque ianua factum / prodidit', and Am I viii 109 'uox erat in cursu, cum me mea prodidit umbra'. Vires again used of poetic skill at Tr I vi 29 'ei mihi non magnas quod habent mea carmina uires', Tr IV ix 16 'Pierides uires et sua tela dabunt', EP III iii 34, and EP III iv 79.

13. DEPRENSA. Deprendere 'recognize, detect' is also found at Met II 93-94 'utinamque oculos in pectore posses / inserere et patrias intus deprendere curas' and Met VII 536-37 'strage canum primo uolucrumque ouiumque boumque / inque feris subiti deprensa potentia morbi', as well as at Livy XLII 17 7 (uenenum) and Celsus III 18 3 '[phrenetici ...] summam ... speciem sanitatis in captandis malorum operum occasionibus praebent, sed exitu deprenduntur'. This seems to be a semi-medical sense; Professor R. J. Tarrant suggests that colore may bear the secondary meaning 'complexion' in this passage.

15. TAM MALA THERSITEN PROHIBEBAT FORMA LATERE. For Thersites' ugliness, see Il II 216-19 'αἴσχιστος δὲ ἀνὴρ ὑπὸ Ἴλιον ῆλθε· / φολκὸς ἔην, χωλὸς δ' ἕτερον πόδα· τὼ δέ οἱ ὤμω / κυρτώ, ἐπὶ στῆθος συνοχωκότε· αὐτὰρ ὕπερθε / φοξὸς ἔην κεφαλήν, ψεδνὴ δ' ἐπενήνοθε λάχνη'.

For the modern reader, Thersites' ugliness is hardly his leading characteristic; but at EP III ix 9-10 Ovid again refers to his appearance: 'auctor opus laudat: sic forsitan Agrius [his father] olim / Thersiten facie dixerit esse bona'. Other mentions of Thersites' ugliness at Lucian Dial Mort XXV (Thersites argues that he is now as handsome as Nireus) and Epictetus Diss II 23 32 (Thersites is contrasted with Achilles), to which Professor C. P. Jones adds from Greek epigram Greek Inscr. Brit. Mus. IV ii 1114; other citations from late Greek authors at PW V A,2 2457 18-38 & 2464 23-66 and Roscher V 670 23 ff.

16. NIREVS. For the beauty of Nireus, see Il II 671-74 'Νιρεὺς αὖ Σύμηθεν ἄγε τρεῖς νῆας ἐί̈σας, / Νίρεὺς Ἀγλαί̈ης υἱὸς Χαρόποιό τ' ἄνακτος, / Νίρεύς, ὃς κάλλιστος ἀνὴρ ὑπὸ Ἴλιον ῆλθε / τῶν ἄλλων Δαναῶν μετ' ἀμύμονα Πηλεί̈ωνα'. This is the only mention of Nireus in the poem; but Demetrius (Peri Hermeneias 62; cited by Cope at Aristotle Rhet 1414a) remarks that because of Homer's use of epanaphora (the repetition of Nireus' name) and dialysis (asyndeton) 'σχεδὸν ἅπαξ τοῦ Νιρέως ὀνομασθέντος ἐν τῷ δράματι μεμνήμεθα οὐδὲν ἧττον ἢ τοῦ Άχιλλέως καὶ τοῦ Ὀδυσσέως'. Ovid mentions Nireus again at AA II 109-12 'sis licet antiquo Nireus adamatus Homero ... ingenii dotes corporis adde bonis'; see also Hor Epod XV 22 'forma ... uincas Nirea', Hor Carm III xx 15 (where Nireus is paired with Ganymede) and Prop III xviii 27 'Nirea non facies, non uis exemit Achillem'; from Greek epigram Professor C. P. Jones cites Peek Griech. Versinschr. 1728 (Merkelbach ZPE 25 [1977] 281).

16. CONSPICIENDVS. The word is metrically suited to the second half of the pentameter, before the disyllable: compare Tib I ii 70 & II iii 52, Fast V 118 & V 170, and Tr II 114.

17. MIRARI SI is a colloquialism: most of the passages from verse cited at TLL VIII 1067 14 are from Plautus and the hexameter poems of Horace; from Propertius compare II iii 33 'haec ego nunc mirer si flagret nostra iuuentus?' and from Ovid Her X 105 'non equidem miror si stat uictoria tecum' and Tr I ix 21 'saeua neque admiror metuunt si fulmina'.

19. A PVDET, ET GETICO SCRIPSI SERMONE LIBELLVM. The rest of the distich after a pudet explains the exclamation ('I have even written ...'), and so the punctuation should mark the break. The idiom is different from the et pudet et construction seen at xv 29 'et pudet et metuo ['I am both embarrassed and afraid'] semperque eademque precari' and Tr V vii 57-58 'et pudet et fateor ['I confess with embarrassment'], iam desuetudine longa / uix subeunt ipsi uerba Latina mihi'.

The only other instance of independent a pudet in Ovid is AA III 803-4 'quid iuuet et uoces et anhelitus arguat oris; / a pudet, arcanas pars habet ista notas', which, however, Professor R. J. Tarrant suspects is part of an interpolation.

19. GETICO ... SERMONE. Ovid repeatedly claims to have learned Getic and Sarmatian: compare Tr III xiv 47-48 'Threicio Scythicoque fere circumsonor ore, / et uideor Geticis scribere posse modis', Tr V vii 55-56 'ille ego Romanus uates—ignoscite, Musae!— / Sarmatico cogor plurima more loqui', Tr V xii 58 'nam didici Getice Sarmaticeque loqui', and EP III ii 40 (identical to Tr V xii 58).

It is of course not possible to prove that Ovid did or did not learn Getic and write poetry in that language. But in the absence of other evidence, it seems better to suppose that he did learn the language since (a) he claims to have do so, (b) Latin and Greek would hardly have been widely spoken in the region, and (c) a man with Ovid's linguistic facility would have had little difficulty in learning the languages of the region.

20. STRVCTAQVE ... VERBA. Compare Cic de Or III 171 'struere uerba', and see at 4 structura (p 393).

20. NOSTRIS ... MODIS. Ovid did not use native rhythms, but instead used Latin metres.

21. ET PLACVI. Luck compares EP I v 63-64 'forsitan audacter faciam, sed glorior Histrum / ingenio nullum maius habere meo', but it is clear enough from the context that Ovid was there speaking of his Latin poetry.

21. GRATARE. Gratari is extremely rare in Latin, being found only in the poets and historians; grātŭlāri was of course not available (except for grātŭlŏr) for use in dactylic verse. Other instances of the word in Ovid at ix 13 'gratatusque darem cum dulcibus oscula uerbis', Her VI 119 'nunc etiam peperi; gratare ambobus, Iason!', Her XI 65, Met I 578, VI 434, IX 244 & 312, and Fast III 418.