37. QVAM POTVIT ... MAXIMA. For the idiom compare Cic Fam XIII vi 5 'quam maximas ... gratias agat' and ND II 129 'gallinae ['hens'] ... cubilia sibi nidosque construunt eosque quam possunt mollissime substernunt'.

37. GRATVS ABVNDE EST. Apparently the only instance in classical poetry of abunde modifying an adjective. The prose authors cited by the lexica are Sallust, Livy, Valerius Maximus, Curtius, the elder Pliny, and Quintilian. Abunde elsewhere in Ovid only at Met XV 759 'humano generi, superi, fauistis abunde!' and Tr I vii 31 'laudatus abunde'.

38. FINEM PIETAS CONTIGIT ILLA SVVM. 'That act of piety has reached its objective', that is, has made the giver gratus.

39-42. For the sentiment compare EP III iv 81-82 'haec [sc laudanda uoluntas] facit ut ueniat pauper quoque gratus ad aras, / et placeat caeso non minus agna boue'.

41-42. GRAMINE PASTA FALISCO / VICTIMA TARPEIOS INFICIT ICTA FOCOS. Compare iv 29-32 'templaque Tarpeiae primum tibi sedis adiri ... colla boues niueos certae praebere securi, / quos aluit campis herba Falisca suis'.

42. INFICIT. 'Stain'. Inficere in the context of a sacrifice also at Met XV 134-35 '[uictima ...] percussa ... sanguine cultros / inficit' and Hor Carm III xiii 6.

44. PRINCIPIBVS ... VIRIS. A fixed colloquial idiom: OLD princeps¹ 5 cites Plautus Amphitruo 204 'delegit uiros primorum principes' and Hor Ep I xvii 35 'principibus placuisse uiris non ultima laus est'. There was a parallel expression principes feminae: see Pliny NH VIII 119 and Tac Ann XIII 42 (Suillius compares himself to Seneca) 'an grauius aestimandum sponte litigatoris praemium honestae operae adsequi quam corrumpere cubicula principum feminarum?'.

45. CARMINA VESTRARVM PERAGVNT PRAECONIA LAVDVM. Praeconia in a similar context at Tr II 65 'inuenies uestri praeconia nominis illic [in the Metamorphoses]'; used with peragere at Tr V i 9 'ut cecidi, subiti perago praeconia casus'.

45. LAVDVM. 'Deeds meriting praise'; compare 87 'tuas ... laudes ... recentes'. The meaning is found even in prose: see Caesar BC II 39 4 'haec tamen ab ipsis inflatius commemorabantur, ut de suis homines laudibus libenter praedicant' and the other passages cited at OLD laus¹ 3b.

46. ACTORVM. AVCTORVM (BCHL) is possible enough; but actorum accords better with the preceding laudum.

46. CADVCA. 'Impermanent'. The sense is frequent in Cicero: see Rep VI 17 'nihil est nisi mortale et caducum praeter animos' and Phil IV 13. Elsewhere in Ovid the usual sense of the word is 'ineffectual': see Fast I 181-82 'nec lingua caducas / concipit ulla preces, dictaque pondus habent' and Ibis 88 'et sit pars uoti nulla caduca mei'. Similar uses at Her XV 208 & XVI 169.

47. CARMINE FIT VIVAX VIRTVS, EXPERSQVE SEPULCRI / NOTITIAM SERAE POSTERITATIS HABET. For the immortality given by verse, compare from Ovid Tr V xiv 5 (to his wife) 'dumque legar, mecum pariter tua fama legetur' and EP III ii 35-36 (to those friends who assisted him) 'uos etiam seri laudabunt saepe nepotes, / claraque erit scriptis gloria uestra meis'. The topic is closely related to that of the poet's own immortality, for which, in Ovid, see xvi 2-3 'non solet ingeniis summa nocere dies, / famaque post cineres maior uenit' and Met XV 871-79.

For other poets' treatment of the immortality given by verse, see Prop III ii 17-26, Hor Carm IV ix, Pindar Nem VII 11-16, Gow on Theocritus XVI 30, and Murgatroyd on Tib I iv 63-66.

47. VIVAX VIRTVS. Compare Hor AP 68-69 'mortalia facta peribunt, / nedum sermonum stet honos et gratia uiuax'.

47. EXPERSQVE SEPVLCRI. The diction of this line is very elevated: Professor R. J. Tarrant compares Met IX 252-53 (Jupiter speaking of Hercules) 'aeternum est a me quod traxit, et expers / atque immune necis' and Cons Liu 59-60 'Caesaris adde domum, quae certe funeris expers / debuit humanis altior esse malis'. The following line's notitiam ... habet is in comparison an anticlimax.

49. TABIDA CONSVMIT FERRVM LAPIDEMQVE VETVSTAS. Iron and flint were proverbial for hardness: compare x 3-4 'ecquos tu silices, ecquod, carissime, ferrum / duritiae confers, Albinouane, meae?', Her X 109-10, AA I 473-76, Met XIV 712-13, Fast V 131-32, Tr IV vi 13-14, and EP II vii 39-40; other passages are cited by Smith at Tib I iv 18 'longa dies molli saxa peredit aqua'. At I 313-16, Lucretius, discussing the invisible wearing away of substances, says 'stilicidi casus lapidem cauat, uncus aratri / ferreus occulte decrescit uomer in aruis, / strataque iam uolgi pedibus detrita uiarum / saxea conspicimus'.

51. SCRIPTA FERVNT ANNOS. The phrase completes the sentence begun in the previous distich, as is shown by the parallel passages Am I x 61-62 'scindentur uestes, gemmae frangentur et aurum; / carmina quam tribuent, fama perennis erit' and Am I xv 31-32 'ergo cum silices, cum dens patientis aratri / depereant aeuo, carmina morte carent'.

51. FERVNT. 'Withstand'; the same sense at Tr V ix 8 'scripta uetustatem si modo nostra ferunt', Cic Am 67 'ea uina quae uetustatem ferunt', Silius IV 399-400 'si modo ferre diem ... carmina nostra ualent', and Quintilian II 4 9 'sic et annos ferent et uetustate proficient'.

51-53. AGAMEMNONA ... THEBAS. The two great cycles of Greek heroic mythology. The same conjunction at Am III xii 15-16 'cum Thebae, cum Troia foret, cum Caesaris acta, / ingenium mouit sola Corinna meum' and Tr II 317-20 'cur non Argolicis potius quae concidit armis / uexata est iterum carmine Troia meo? / cur tacui Thebas et uulnera mutua fratrum / et septem portas sub duce quamque suo'; compare as well Prop II i 21 '[canerem ...] nec ueteres Thebas nec Pergama, nomen Homeri'. Lucretius, arguing that the world was created at a definite moment, wrote 'cur supera ['before'] bellum Thebanum et funera Troiae / non alias alii quoque res cecinere poetae?' (V 326-27).

52. QVISQVIS CONTRA VEL SIMVL ARMA TVLIT. The leaders of the Greeks and Trojans.

The line's structure parallels 54 'quicquid post haec, quicquid et ante fuit'. Both are conspicuous by their lack of adornment.

55. DI QVOQVE CARMINIBVS, SI FAS EST DICERE, FIVNT. This is possibly a reference to Herodotus II 53, where Herodotus says that Homer and Hesiod established the Greek pantheon; for Ovid's borrowings from Herodotus, see at iii 37 opulentia Croesi (p 189). The same idea previously in Xenophanes (fr. 11 Diels).

The line looks ahead to 63-64 'et modo, Caesar, auum, quem uirtus addidit astris, / sacrarunt aliqua carmina parte tuum'.

55. SI FAS EST DICERE. Ovid here apologizes for the shocking statement he is making. Up to this point poetry has helped give lasting fame to what was already a fact, but here poetry is actually making something happen (or appear to happen). At Am III xii 21-40 Ovid similarly describes how poets created the myths.

57-64. Ovid follows the same sequence in the Metamorphoses, describing the separation of Chaos at I 5-31, the attack of the Giants at I 151-55, Bacchus' conquest of India at IV 20-21 & 605-6, and Hercules' capture of Oechalia at IX 136; he foretells Augustus' apotheosis at XV 868-70. Professor R. J. Tarrant points out that these lines may well be referring specifically to the earlier poem.

57-58. SIC CHAOS EX ILLA NATVRAE MOLE PRIORIS / DIGESTVM PARTES SCIMVS HABERE SVAS. 'Thus we know Chaos now has its divisions after having been arranged in order from the famous mass that was its previous nature'. Ovid describes the separation of the elements at Met I 25-31 and Fast I 103-10; see also Ecl VI 31-36.

I take illa ('famous') to refer to the familiarity through the poets and philosophers of the notion of the separation of Chaos into the four elements. Alternatively, Professor A. Dalzell points out to me that illa could have a pejorative sense.

58. DIGESTVM. 'Separated'. At Met I 7 Ovid calls Chaos 'rudis indigestaque moles'.

59. ADFECTANTES CAELESTIA REGNA GIGANTAS. At Am III xii 27 Ovid, speaking of false legends created by the poets, says 'fecimus Enceladon iaculantem mille lacertis'.

In his youth, Ovid had attempted but later abandoned a poem on the battle of the Giants against Jupiter 'designed to glorify Augustus under the guise of Jupiter' (Owen Tristia II p. 77): the language he uses at Tr II 333-40 seems too explicit to be a mere instance of the love-poet's defence of his subject-matter: 'at si me iubeas domitos Iouis igne Gigantas [Heinsius: Gigantes codd] / dicere, conantem debilitabit onus. / diuitis ingenii est immania Caesaris acta / condere, materia ne superetur opus. / et tamen ausus eram; sed detrectare uidebar, / quodque nefas, damno uiribus esse tuis.[20] / ad leue rursus opus, iuuenalia carmina, ueni, / et falso moui pectus amore meum'. He refers to the same poem again at Am II i 11-18 'ausus eram, memini, caelestia dicere bella / centimanumque Gyen—et satis oris erat— / cum male se Tellus ulta est, ingestaque Olympo / ardua deuexum Pelion Ossa tulit. / in manibus nimbos et cum Ioue fulmen habebam, / quod bene pro caelo mitteret ille suo— / clausit amica fores! ego cum Ioue fulmen omisi; / excidit ingenio Iuppiter ipse meo'.

The actual descriptions of the Giants' rebellion in Ovid's surviving poems are brief (Met I 151-62 & 182-86, Fast V 35-42), but references to the rebellion are frequent (Met X 150-51, Fast I 307-8, Fast IV 593-94, Fast V 555, Tr II 71, Tr IV vii 17, EP II ii 9-12). The accounts at Met V 319-31 of the flight of some of the gods to Egypt and at Fast II 459-74 of Venus' flight to the Euphrates are no doubt derived from Ovid's earlier researches.

59. ADFECTANTES. 'Unlawfully seeking to obtain'; compare Met I 151-52 'neue foret terris securior arduus aether, / adfectasse ferunt regnum caeleste Gigantas' and Fast III 439 'ausos caelum adfectare Gigantas'. This sense is found in prose: compare Livy I 50 4 'cui enim non apparere adfectare eum imperium in Latinos?'. At Livy I 46 2 the word is used without the conative sense: 'neque ea res Tarquinio spem adfectandi regni minuit'.

59. GIGANTAS Heinsius. The manuscripts have GIGANTES, which Lenz, Wheeler, and André print. In classical Latin poetry, Greek nouns of the third declension with plural nominatives in -ες and plural accusatives in -ας retained these endings. Housman 836-39 gives many instances where metre demonstrates an accusative in -ας. In Ovid when such an ending occurs, some manuscripts commonly offer the normalized -es; at Tr II 333, as here, all manuscripts offer Gigantes, again corrected by Heinsius.

Such apparent violations of the rule as Fast I 717 'horreat Aeneadās et primus et ultimus orbis', Fast III 105-6 'quis tunc aut Hyadās aut Pliadas Atlanteas / senserat' and Virgil G I 137-38 'nauita tum stellis numeros et nomina fecit, / Pleiadās, Hyadās, claramque Lycaonis Arcton' are of course no real exceptions, the lengthening of short closed vowels at the ictus being permitted (Platnauer 59-62).

60. AD STYGA NIMBIFERI VINDICIS IGNE DATOS. 'Hurled to the underworld by the lightning-bolt of cloud-gathering Jupiter'. This was Jupiter's first use of the weapon: see Fast III 439-40 'fulmina post ausos caelum adfectare Gigantas / sumpta Ioui: primo tempore inermis erat'.

60. NIMBIFERI VINDICIS IGNE is my correction of the manuscripts' NIMBIFERO and NVBIFERO. The unmodified uindicis and modified igne of the manuscript readings might be defended by EP II ix 77 'quicquid id est [whatever Ovid has committed], habuit moderatam uindicis iram', but uindicis is there defined by the following 'qui nisi natalem nil mihi dempsit humum', and moderatam is a more suitable epithet for iram than is nimbifero for igne in the present passage., At Tr II 143-44 'uidi ego pampineis oneratam uitibus ulmum, / quae fuerat saeuo fulmine tacta Iouis', the manuscripts divide between saeuo and saeui, which has a good claim to be considered the true reading; in any case, Iouis is less in need of a defining adjective than uindicis in the present passage. Finally, the genitive here is strongly supported by Ibis 475-76 'ut Macedo rapidis icta est cum coniuge flammis, / sic precor aetherii uindicis igne cadas'.

The corruption may have been induced by a wish to introduce interlocking word order: for a similar instance see at ii 9 Baccho uina Falerna (p 164). But in fact substantive and epithet are constantly found linked at the caesura of the pentameter: the strong break in the metre at that point no doubt made the construction more readily acceptable there than in other positions.

I have printed nimbiferi in preference to nubiferi because Jupiter is linked with nimbi at two other passages. The first of these is Am II i 15-16 'in manibus nimbos et cum Ioue fulmen habebam, / quod bene pro caelo mitteret ille suo', and the second Met III 299-301, where Ovid describes Jupiter's preparations to descend on Semele: 'aethera conscendit uultuque sequentia traxit / nubila, quis nimbos immixtaque fulgura uentis / addidit et tonitrus et ineuitabile fulmen'.

61-62. SIC VICTOR LAVDEM SVPERATIS LIBER AB INDIS ... TRAXIT. Bacchus' conquest of India is also mentioned by Ovid at Fast III 465-66 'interea Liber depexos crinibus Indos / uicit et Eoo diues ab orbe redit', Fast III 719-20, and Tr V iii 23-24.

61-62. VICTOR should be taken both with Liber and Alcides.

61-62. LIBER ... ALCIDES. The same pairing (both times in the context of Augustan panegyric) at Aen VI 801-5 'nec uero Alcides tantum telluris obiuit, / fixerit aeripedem ceruam licet, aut Erymanthi / pacarit nemora et Lernam tremefecerit arcu; / nec qui pampineis uictor iuga flectit habenis / Liber, agens celso Nysae de uertice tigris' and Hor Carm III iii 9-15. Ovid may have made similar mention of Bacchus and Hercules in his panegyric of Augustus.

61-62. SIC ... LAVDEM ... ALCIDES CAPTA TRAXIT AB OECHALIA. Hercules attacked and captured Oechalia in order to carry off Iole, the king's daughter. This was his last exploit, for it led to Deianira's sending him the poisoned robe which caused his death. The capture of Oechalia is also mentioned at Her IX passim (the poem perhaps not by Ovid) and Met IX 136-40.

62. OECHALIA. For the quadrisyllable ending to the pentameter, see at ii 10 Alcinoo (p 164).

63. AVVM. Augustus. In AD 4 Augustus adopted Tiberius (son of Livia's first husband, Ti. Claudius Nero), and Tiberius adopted Germanicus, son of his brother Drusus.

63. QVEM VIRTVS ADDIDIT ASTRIS. Compare Aen VIII 301 (of Hercules) 'salue, uera Iouis proles, decus addite diuis'.

Augustus died on 19 August AD 14; on 17 September the Senate decreed caelestes religiones for him (Tac Ann I 10 8; Fasti Amiternini, Antiates, & Oppiani, at Ehrenberg-Jones 52). Augustus' apotheosis is also mentioned at ix 127-32 and xiii 23-26.

64. ALIQVA ... PARTE. The same phrase in the same metrical position at Fast I 133-34 (Janus speaking) 'uis mea narrata est. causam nunc disce figurae: / iam tamen hanc aliqua tu quoque parte uides'.

64. CARMINA. Ovid is referring to his own poems (in Latin and Getic) on Augustus' apotheosis, also mentioned at vi 17-18 'de caelite ... recenti ... carmen', ix 131-32 'carmina ... de te ... caelite ... nouo', and xiii 25-26.

65-66. SI QVID ADHVC IGITVR VIVI, GERMANICE, NOSTRO / RESTAT IN INGENIO, SERVIET OMNE TIBI. Compare Prop IV i 59-60 'sed tamen exiguo quodcumque e pectore riui / fluxerit, hoc patriae seruiet omne meae', which Ovid is clearly imitating. Hertzberg ad loc conjectured RIVI for our passage, which may well be right; but uiui seems to agree better with restat.

67. VATIS ... VATES. For an extreme instance of Ovid's favourite figure of polyptoton (Quintilian IX 3 36-37), see the account at Met IX 43-45 of Achelous' wrestling-match with Hercules: 'inque gradu stetimus, certi non cedere, eratque / cum pede pes iunctus, totoque ego pectore pronus / et digitos digitis et frontem fronte premebam'. Other instances of polyptoton with uates at Fast I 25 (to Germanicus) 'si licet et fas est, uates rege uatis habenas' and EP II ix 65 (to Cotys, king of Thrace, apparently a writer of poetry) 'ad uatem uates orantia bracchia tendo',

67. VATES. Approximately nine hundred lines survive of a version of Aratus generally attributed to Germanicus, who might have been composing the poem at the time Ovid was writing: Augustus' apotheosis is mentioned at 558-60. It is possible however that Tiberius was the poem's author: he is known to have written a Conquestio de morte L. Caesaris and to have composed Greek verse (Suet Tib 70). For a full discussion see the introduction to Gain's edition of the Aratus.

69-70. QVOD NISI TE NOMEN TANTVM AD MAIORA VOCASSET, / GLORIA PIERIDVM SVMMA FVTVRVS ERAS. Compare Met V 269-70 (the Muses to Minerva) 'o nisi te uirtus opera ad maiora tulisset, / in partem uentura chori Tritonia nostri'.

There is a striking parallel to this passage in Quintilian's address to Domitian in his catalogue of poets: 'hos nominamus quia Germanicum Augustum ab institutis studiis deflexit cura terrarum, parumque dis uisum est esse eum maximum poetarum' (X i 91-92).

70. GLORIA PIERIDVM SVMMA. Gloria similarly used at EP II xi 28 'maxima Fundani gloria, Rufe, soli', Aen VI 767 'proximus ille Procas, Troianae gloria gentis', and Val Max IV iii 3 'Drusum ... Germanicum, eximiam Claudiae familiae gloriam'. The term was used in particular of fine cattle: see AA I 290 'candidus, armenti gloria, taurus', Pan Mess (Corp Tib III vii) 208 'tardi pecoris ... gloria taurus' and Aetna 597 'gloria uiua Myronis' (on Myron's Cow see at i 34 ut similis uerae uacca Myronis opus [p 158]).

71. SI DARE R. J. Tarrant. The manuscripts' SED DARE is a possible reading; but Professor Tarrant's slight change removes the awkwardness of nec tamen following immediately upon sed.

71. MAVIS IF^2ul MAIVS BF¹. Either of the two variants could be read from CMHLT. The preferable reading is mauis, since it links more closely to potes in the pentameter, and would be especially liable to corruption after maiora two lines previous. I have found no good parallel for singular maius 'a more important thing': for the plural OLD maior 5 cites from verse Fast IV 3 'certe maiora canebas' and its model, Ecl IV 1 'paulo maiora canamus'.

72. NEC TAMEN EX TOTO DESERERE ILLA POTES. Graecinus was another of Ovid's addressees who, while a soldier, kept up his other pursuits: 'artibus ingenuis [=lībĕrālibus], quarum tibi maxima cura est, / pectora mollescunt asperitasque fugit. / nec quisquam meliore fide complectitur illas, / qua sinit officium militiaeque labor' (EP I vi 7-10).

72. EX TOTO. 'Altogether'. Compare EP I vi 27-28 'spes igitur menti poenae, Graecine, leuandae / non est ex toto nulla relicta meae'. The idiom was probably subliterary: the only instances from the time of Ovid cited by OLD totum 2 are Celsus III 3 71b 'neque ex toto in remissionem desistit' and Columella V 6 17 'antequam ex toto arbor praeualescat'.

73. NVMERIS ... VERBA COERCES. 'You arrange words in metrical patterns'. Similar wording at Cic Or 64 'mollis est enim oratio philosophorum ... nec uincta numeris ['not in rhythmic prose'], sed soluta liberius'.

Professor E. Fantham points out to me that Ovid may also be playing on numerus 'military contingent' (OLD numerus 9): 'you draft words in squads'.

75-76. NEC AD CITHARAM NEC AD ARCVM SEGNIS APOLLO, / SED VENIT AD SACRAS NERVVS VTERQVE MANVS. Apollo is similarly described at Met X 107-8 (of Cyparissus) 'nunc arbor, puer ante deo dilectus ab illo / qui citharam neruis et neruis temperat arcum'.

76. VENIT = conuenit. In Latin verse a simple verb can carry the sense of any of its compounds, even when this sense is quite different from the usual meaning of the simple verb. Compare Catullus LXIV 21 'tum Thetidi pater ipse iugandum Pelea sensit', "where it is plain that iugandum is for coniugandum, and this leads the reader to the conclusion that sensit is for consensit, where the omission decidedly affects the sense" (Bell 330).

The line should not be taken as an instance of the expression uenire ad manum (OLD uenio 7c), since the idiom's sense 'be convenient' does not fit the context here: for the sense compare Livy XXXVIII 21 6 'quod [sc saxum] cuique temere trepidanti ad manum uenisset' and Quintilian II xi 6 'abrupta quaedam, ut forte ad manum uenere, iaculantur'. Venire in manus offers a somewhat more satisfactory meaning, almost equivalent to 'have, hold' (compare Cic Q Fr II xv [xiv] i 'quicumque calamus in manus meas uenerit' and Persius III 11 'inque manus chartae nodosaque uenit harundo'), but seems to be a separate idiom.

79. QVAE QVONIAM NEC NOS. 'Since she continues to give poetic inspiration to myself as well as to you'. Quae quoniam seems very prosaic, but Ovid uses the phrase again at Tr I ix 53-54 'quae [sc coniectura] quoniam uera est ... gratulor ingenium non latuisse tuum'.

79-80. VNDA ... VNGVLA GORGONEI QUAM CAVA FECIT EQVI. Hippocrene, the spring of the Muses, said to have been created by the hoof-beat of Pegasus. Similarly described at Met V 264 'factas pedis ictibus undas', Fast V 7-8 'fontes Aganippidos Hippocrenes, / grata Medusaei signa ... equi' and Persius prol 1 'fonte ... caballino'.

80. VNGVLA ... CAVA. Professor J. N. Grant points out to me the possible borrowing from Ennius Ann 439 Vahlen³ 'it eques et plausu caua concutit ungula terram'.

80. GORGONEI ... EQVI. The same phrase in the same metrical position at Fast III 450 'suspice [sc caelum]: Gorgonei colla uidebis equi'. For the birth of Pegasus from the blood of the Gorgon Medusa, see Met IV 784-86,

81. COMMVNIA SACRA TVERI. Sacra similarly used of poetry at Tr IV i 87, Tr IV x 19 'at mihi iam puero caelestia sacra placebant', EP II x 17 'sunt tamen inter se communia sacra poetis', and EP III iv 67 'sunt mihi uobiscum communia sacra, poetae'. For tueri 'observe, maintain' compare Cic Tusc I 2 'mores et instituta uitae resque domesticas ac familiaris nos profecto et melius tuemur et lautius'.

82. ISDEM STVDIIS IMPOSVUISSE MANVM. Similar phrasing at Tr IV i 27-28 'non equidem uellem ... Pieridum sacris imposuisse manum'.

82. IMPOSVISSE has the sense of the present infinitive, as is shown by tueri in the previous line; compare as well ii 27-28 'uix sumptae Musa tabellae / imponit pigras, paene coacta, manus'. For the idiom, see Platnauer 109-12. It is particularly frequent in the latter half of the pentameter, immediately before the disyllable: compare, from many instances, AA III 431-32 'ire solutis / crinibus et fletus non tenuisse decet' and Tr IV viii 5-12 'nunc erat ut posito deberem fine laborum / uiuere, me nullo sollicitante metu, / quaeque meae semper placuerunt otia menti / carpere et in studiis molliter esse meis, / et paruam celebrare domum ueteresque Penates ... inque sinu dominae carisque sodalibus inque / securus patria consenuisse mea'. The idiom, although more common in elegiac verse, is also found in epic: compare Aen X 14 'tum certare odiis, tum res rapuisse licebit'.

83. LITORA PELLITIS NIMIVM SVBIECTA CORALLIS. Compare ii 37 'hic mea cui recitem nisi flauis scripta Corallis'. Strabo mentions the Coralli as inhabiting the region near Haemus (VII 5 12); they are rather obscurely described at Val Fl VI 89-94 'densique leuant uexilla Coralli, / barbaricae quis signa rotae, ferrataque dorso / forma suum ['of pigs'], truncaeque Iouis simulacra columnae; / proelia nec rauco curant incendere cornu, / indigenas sed rite duces et prisca suorum / facta canunt ueterumque, uiris hortamina, laudes'.

Nothing else is known of the tribe.

83. PELLITIS. Elsewhere in Ovid only at x 2 'pellitos ... Getas'.

83. NIMIVM SVBIECTA. Compare vi 45 'nimium nobis conterminus Hister'.

85. VLLO M ILLO BCFHILT. Illo is not a possible reading, since of course most parts of the empire would have been less isolated than Tomis. Ovid does not specify a preferred place of exile at either Tr IV iv 49 'nunc precor hinc alio iubeat discedere' or EP III i 29-30 'non igitur mirum ... altera si nobis usque rogatur humus', nor in any of the passages listed in the next two notes.

86. QVI MINVS ... DISTET. For this constant prayer of the exiled Ovid, see Tr II 575-78 (the concluding lines) 'non ut in Ausoniam redeam, nisi forsitan olim, / cum longo poenae tempore uictus eris; / tutius exilium pauloque quietius oro, / ut par delicto sit mea poena suo', Ibis 28, EP III i 4 & 85, EP III iii 64, EP III vii 30, EP III ix 38, and EP III ix 1-4 'Quod sit in his eadem sententia, Brute, libellis, / carmina nescio quem carpere nostra refers, / nil nisi me terra fruar ut propiore rogare, / et quam sim denso cinctus ab hoste loqui'.

86. DISTET FHILM^2c. Lenz and André print DISTAT (BCT); however, the defining subjunctive seems to be required, and is supported by EP II viii 36 'daque procul Scythico qui sit ab hoste locum'.

87. LAVDES. See at 45 laudum (p 268).

88. MAGNAQVE QVAM MINIMA FACTA REFERRE MORA. At EP III iv 53-60 Ovid speaks of how a poem of his on a recent triumph has been late in being written, and will be late in reaching Rome: 'cetera certatim de magno scripta triumpho / iam pridem populi suspicor ore legi. / illa bibit sitiens lector, mea pocula plenus; / illa recens pota est, nostra tepebit aqua. / non ego cessaui, nec fecit inertia serum: / ultima me uasti distinet [scripsi: sustinet codd] ora freti. / dum uenit huc rumor properataque carmina fiunt / factaque eunt ad uos, annus abisse potest'.

90. SOCERO PAENE ... TVO. See at 11 eadem mihi filia paene est (p 262).

IX. To Graecinus

The poem begins with Ovid's wish that his letter might arrive on the day Graecinus becomes consul (1-4). He imagines himself present when Graecinus enters his magistracy; since he will not be there, he will at least in his mind imagine Graecinus carrying out his consular functions (5-56). He then speaks of Graecinus' brother Flaccus, who will succeed him as consul ordinarius for 17: the two brothers will take pleasure in each other's office (57-65). He describes the brothers' devotion to Tiberius, and asks for their assistance in obtaining his removal from Tomis (65-74). The mention of his exile serves as a bridge to the topic of his life in Tomis. Flaccus can attest to the hardships Ovid endures, since he was recently stationed in the area (75-86). Once Graecinus has learned of these hardships from Flaccus, he should ask what Ovid's reputation in Tomis is. He will learn that Ovid is well liked, and has even received public honours (87-104). His loyalty to the imperial family is well known: Flaccus may have heard of this, Tiberius will eventually learn of it, but Augustus has certainly observed it from heaven; Ovid's poems are perhaps inducing Augustus to yield to his prayers (105-34).

The poem is the longest in the book, and combines several almost unrelated sections dealing with a number of subjects. The first section of the poem, the celebration of Graecinus' nomination to the consulship, is very heavily indebted to IV iv, Ovid's first poem on Sextus Pompeius' election to the consulship. The section detailing Flaccus' presence near Tomis owes something to IV vii, the letter to Vestalis. The description of Ovid's reputation in Tomis is new, and shows a softening of his attitude towards his fellow-townsmen, but the description of his piety to the imperial family owes much to III ii, a letter of thanks to Cotta for the gift of images of the members of the family. The poem's discursiveness and large number of derived elements suggest a hasty composition.

1. GRAECINE. Graecinus became a frater Arualis in 21 (CIL VI 2023); the C. Pomponius Graecinus of CIL XI 5809 (Iguvium) seems not to have survived to enter the Senate (Syme HO 74-75). Graecinus is not mentioned in literary sources apart from Ovid, but his brother Flaccus was rather more famous: see at 75 (p 308).

3. DI FACIANT looks like a colloquial expression. Other instances at iv 47-48 'di faciant aliquo subeat tibi tempore nostrum / nomen', Tr V xiii 17, and Prop II ix 24.

3. AVRORAM here is virtually equivalent to diem; it is not found elsewhere in the poetry of exile, but compare Fast I 461 & II 267-68 'tertia post idus nudos aurora Lupercos / aspicit'.

3. OCCVRRAT. 'Arrive', as commonly: compare Cic Phil I 9, Livy XXXVII 50 7 'ad comitiorum tempus occurrere non posse', and Pliny Ep VI xxxiv 3 'uellem Africanae [sc pantherae] quas coemeras plurimas ad praefinitum diem occurrissent'.

4. BIS SENOS = dŭŏdĕcim, metrically difficult because of its initial three consecutive short vowels. Roman poets avoid using the usual names for numbers above nouem, with the obvious exceptions of centum and mille; sometimes, as here, metrical exigencies left them with no alternative. For bis seni (sex) Tarrant at Sen Ag 812 bis seno ... labore cites Ennius Ann 323 Vahlen², Ecl I 43, Aen I 393, Prop II xx 7, Met VIII 243, Fast I 28, Sen Tro 386 & Oed 251, and from Greek Callimachus Aetia I fr. 23 19 Pfeiffer.

6. TVRBAE. Compare iv 27 'cernere iam uideor rumpi paene atria turba'.

7. IN DOMINI SVBEAT PARTES. Partes = 'function'; see at ii 27 uix uenit ad partes ... Musa (p 170). For subeat 'undertake' compare Quintilian X i 71 'declamatoribus ... necesse est secundum condicionem controuersiarum plures subire personas' and the passages cited at OLD subeo 7b.

8. FESTO Burman IVSSO BCMFHIL IVSTO T, sicut coni Merkel. Iusso has been explained since Merula as meaning that Ovid hopes the letter will arrive on the day it is told to; but the word seems rather strange, and lacks the point it has in the passages cited by Ehwald (KB 64), AA II 223-24 'iussus adesse foro, iussa maturius hora / fac semper uenias, nec nisi serus abi' and Prop IV vi 63-64 (of Cleopatra) 'illa petit Nilum cumba male nixa fugaci, / hoc unum, iusso non moritura die' (she would commit suicide at a time of her own choosing), or at Aen X 444 (cited by Owen in 1894) 'socii cesserunt aequore iusso', where iusso stands by hypallage for iussi. The meaning of iusto is inappropriate for the present passage, as will be seen from Suet Tib 4 2 'retentis ultra iustum tempus ['the time allowed'] insignibus'. Burman's conjecture festo was not placed in the text even by its author, but it seems a reasonable solution to the difficulty. For it Burman cited 56 'hic quoque te festum consule tempus agam'; see as well Fast I 79-80 'uestibus intactis Tarpeias itur in arces, / et populus festo concolor ipse suo est'. The corruption of so straightforward an epithet may seem unlikely, but compare Prop IV xi 65-66 'uidimus et fratrem sellam geminasse curulem; / consule quo, festo [Koppiers: facto codd] tempore, rapta soror'.

9. ATQVI unus e duobus Hafniensibus Heinsii. The ATQVE of BCMFHILT is possibly right. For the adversative sense here required, OLD atque 9 cites Plautus Aul 287-88 'atque ego istuc, Anthrax, aliouorsum dixeram, / non istuc quod tu insimulas', Mer 742, and Ter Heaut 189 (apparently a misprint for 187 'atque etiam nunc tempus est') from comedy, but from the classical period only Cic Att VI i 2 'ac putaram paulo secus' and Fam XIV iv 5 'atque ego, qui te confirmo, ipse me non possum', and instances of ac tamen at Fam VII xxiii 1, Caesar BC III 87 4, and Tac Ann III 72. In view of the doubtful status of adversative atque at the time of Ovid and the ease of corruption of atqui to atque I have followed Heinsius in reading atqui. Heinsius similarly restored atqui from his codex Richelianus for the other manuscripts' atque at Tr II 121-24 'corruit haec ... sub uno ... crimine lapsa domus. / atqui ea sic lapsa est ut surgere, si modo laesi / ematuruerit Caesaris ira, queat'; and atque is found for the correct atqui in some manuscripts at Hor Sat I ix 52-53 '"magnum narras, uix credibile!" "atqui / sic habet"' and EP I ii 33-34 'atqui / si noles sanus, curres hydropicus', and in most manuscripts at Ep I vii 1-5 'Quinque dies tibi pollicitus me rure futurum / Sextilem totum mendax desideror. atqui, / si me uiuere uis sanum recteque ualentem, / quam mihi das aegro, dabis aegrotare timenti, / Maecenas, ueniam'.

10. SINCERO. 'Unbroken'.

12. SALVTANDI MVNERE ... TVI. Professor R. J. Tarrant points out to me the notably prosaic use of the defining gerundive.

13. GRATATVS has the force of a present participle, as is shown by cum dulcibus ... uerbis; André mistranslates 'après t'avoir félicité, je t'embrasserai avec des mots tendres'. The perfect participle of deponent verbs takes past or present meaning indifferently, according to context.

16. VT CAPERET FASTVS VIX DOMVS VLLA MEOS seems strange, as does Némethy's explanation 'poeta elatus superbia tectum uertice tangere sibi uidetur'. Perhaps the distich means something like 'on that day I would be filled with a pride which no ancestry, no matter how illustrious, could justify'.

16. FASTVS. 'Haughtiness'—Wheeler. The same sense at AA II 241-42 'exue fastus, / curam mansuri quisquis amoris habes' and Aen III 326-27 (Andromache speaking) 'stirpis Achilleae fastus iuuenemque superbum ... tulimus'. Ovid generally uses fastus of the arrogance of women to their suitors (Am II xvii 9, Met XIV 762, Fast I 419); the word is not found elsewhere in the poetry of exile.

17. DVMQVE LATVS SANCTI CINGIT TIBI TVRBA SENATVS. Compare iv 41 'inde domum repetes toto comitante senatu'; Ovid is here obviously referring to the earlier procession from the new consul's house.

20. LATERIS ... LOCVM is a strange phrase, but is made easier by latus ... cingit in 17. Compare also such passages as Met II 448-49 'nec ... iuncta deae lateri nec toto est agmine prima' and Aen X 160-61 'Pallas ... sinistro / adfixus lateri'. It is possible that latus here means 'companion', as at Martial VI lxviii 4 'Eutychos ille, tuum, Castrice, dulce latus'.

20. HABVISSE is equivalent to habere, as is shown by esse in the preceding line. For the idiom, see at viii 82 imposuisse (p 282) and xi 2 habuisse (p 361).

21. TVRBA QVAMVIS ELIDERER. Elidere similarly used of a crowd's jostling at Sen Clem I 6 1; an extended description at Juvenal III 243-48.

23. PROSPICEREM. Owen in his second edition, Wheeler, and Lenz follow Ehwald (KB 64) in printing B's ASPICEREM. Ehwald argued that prospicerem, 'survey from a distance', was inappropriate in view of the preceding turba quamuis eliderer. But the verb should be taken not with the pentameter that precedes, but with the one that follows, 'densaque quam longum turba teneret iter': prospicerem seems very appropriate. Riese conjectured RESPICEREM 'look back at', but emendation seems unnecessary.

Compounds of specere (the simple verb is used by Plautus and Ennius) are peculiarly liable to confusion: prospicere is similarly corrupted to aspicere in some manuscripts at Met III 603-4 'ipse quid aura mihi tumulo promittat ab alto / prospicio' and Met XI 715-16 'notata locis reminiscitur acta fretumque / prospicit', and other instances of variation of prefix will be found at Met II 405, VI 343, XI 150, XIV 179, XV 577, 660 & 842, Fast I 139 & 461, V 393 & 561, and Her XIX 21.

25-26. Heinsius and Bentley questioned the authenticity of these lines, but the distich does not seem lame enough to warrant excision, and tegeret (see below) is paralleled elsewhere.

25. QVOQVE MAGIS NORIS. 'Listen: this will make you understand better'. Ovid is very fond of quoque magis and the corresponding quoque minus, particularly at line-beginnings. He generally uses the formula to denote the emotion which information he then gives should induce. Compare Met I 757-58 '"quo"que "magis doleas, genetrix" ait, "ille ego liber, / ille ferox tacui"', Met III 448-50 (Narcissus to his reflection) 'quoque magis doleam, nec nos mare separat ingens ... exigua prohibemur aqua', Met XIV 695-97 'quoque magis timeas ... referam tota notissima Cypro / facta', Tr I vii 37-38, and EP I viii 9-10 'quoque magis nostros uenia dignere libellos, / haec in procinctu carmina facta leges'; similar instances of quoque minus at Met II 44, VIII 579, 620 & 866, and EP III ii 52. The present passage shows the same idiom, but with the difference that a subordinate clause (quam me uulgaria tangant) depends on the verb (noris) introduced by the quoque magis clause.

The same formula is used with a different sense, the quoque being an ablative of degree of difference, at Am III ii 28 and Met IV 64 'quoque magis tegitur, tectus magis aestuat ignis'.

EP II v 15-16 'quoque magis moueare malis, doctissime, nostris, / credibile est fieri condicione loci' reads oddly; something has probably been lost from the text after the hexameter.

25. VVLGARIA. 'Commonplace, ordinary'. Compare Hor Sat II ii 38 and Cic De or II 347 'neque enim paruae [sc res] neque usitatae neque uulgares admiratione aut omnino laude dignae uideri solent'.

25. TANGANT. 'Impress'; compare Her V 81 'non ego miror opes, nec me tua regia tangit', Her VI 113, Her VII 11, Met IV 639, Met X 614-15 'nec forma tangor (poteram tamen hac quoque tangi), / sed quod adhuc puer est: non me mouet ipse, sed aetas', and Fast V 489, as well as Her XVI 83. For tangere with a neuter plural subject see Aen I 462 'mentem mortalia tangunt'.

26. TEGERET. There are twenty trisyllabic pentameter endings in Tibullus, thirty in Propertius, but only five in Ovid, all in the Ex Ponto: I i 66 faciet, I vi 26 scelus est, I viii 40 liceat, III vi 46 uideor, and this passage (Platnauer 15-16). Quadrisyllabic endings are similarly frequent in the poetry of exile: see at ii 10 Alcinoo (p 164).

27. SIGNA ... IN SELLA ... FORMATA CVRVLI. For signum 'bas-relief' see at v 18 conspicuum signis ... ebur (the phrase also of the curule chair).

28. NVMIDAE SCVLPTILE DENTIS OPVS. Professor R. J. Tarrant points out to me the clear imitation of Prop II xxxi 12 'ualuae, Llbyci nobile dentis opus'.

28. NVMIDAE ... DENTIS edd NVMIDI ... DENTIS codd. The masculine first declension substantive Numida is occasionally used as an adjective: compare AA II 183 'Numidasque leones' (some manuscripts read Numidosque) and Juvenal IV 99-100 'ursos ... Numidas'. André prints Numidi, citing a nominative Numidus at CIL VIII 17328, the variant at AA II 183, and Apicius VI 8 4 'pullum Numidum' (where there is a variant Numidicum, which André printed in his 1974 edition of Apicius). But given the support for the first-declension form offered by the Juvenal passage and the better manuscripts of the Ars Amatoria, the danger in adducing a doubtful passage of Apicius and a single inscription to determine poetic usage, and the ease of corruption to the second declension, it seems better to assume that Ovid here used the first declension form.

Numidae ... dentis is high poetic diction: compare Met XI 167-68 'instructam ... fidem gemmis et dentibus Indis', Catullus LXIV 47-48 'puluinar ... Indo ... dente politum', Prop II xxxi 12 (quoted above), and Statius Sil III iii 94-95 'Indi / dentis honos'.

28. SCVLPTILE. The word does not seem to occur again in Latin until Prudentius Steph X 266.

29. TARPEIAS ... IN ARCES. See at iv 29 Tarpeiae ... sedis (p 208).

30. DVM expresses purpose; if it were temporal, the verb would be cadit instead of caderet: compare 17-18 'dumque latus sancti cingit tibi turba senatus, / consulis ante pedes ire iuberer eques'.

31. SECRETO represents Ovid's response to the bidding fauete linguis. The word is frequent in comedy, but is very rare in verse, being virtually confined to satire (Hor Sat I ix 67, Juvenal I 95).

31-32. MAGNVS ... DEVS = Iuppiter Optimus Maximus. Compare AA II 540 'eris magni uictor in arce Iouis'.

33. TVRAQVE MENTE MAGIS PLENA QVAM LANCE DEDISSEM. The same notion of sincerity of feeling being more important than size of gifts at viii 35-40.

34. TER QVATER ... LAETVS. 'Infinitely happy'; compare Prop III xii 15 'ter quater in casta felix, o Postume, Galla!', Aen I 94 'o terque quaterque beati', AA II 447-48, and Tr III xii 25-26 'o quater et quotiens non est numerare beatum / non interdicta cui licet urbe frui!'. The phrase is common in Ovid, but he generally uses it to mean 'several times': compare Am III i 31-32 'mouit ... terque quaterque caput', Met II 49, Met IV 734 'ter quater exegit repetita per ilia ferrum', Met VI 133, Met IX 217, Met XII 288, Fast I 576, and Fast I 657 'ter quater euolui signantes tempora fastos'.

35. HIC. 'Hier auf dem Kapitol'—Ehwald (KB 65). The idiom is somewhat strange, but seems well enough supported by Met XIV 372-73 '"per o, tua lumina" dixit / "quae mea ceperunt, perque hanc, pulcherrime, formam"' and Her XVI 137, passages cited by R, J. Tarrant at Sen Ag 971 'dummodo hac ['your'] moriar manu'. Compare as well Prop I xi 17-18 'non quia perspecta non es mihi cognita fama, / sed quod in hac omnis parte ['at Baiae'] timetur [codd: ueretur Lachmann] amor' and Fedeli ad loc.

36. MITIA ... SI ... FATA DARENT. 'If the Fates had been kind, and given'.

36. VRBIS editio Aldina 1502 VERBIS codd. Ius urbis = ius urbis habitandae; compare Met XIII 471-72 'genetrici corpus inemptum / reddite, neue auro redimat ius triste sepulcri [=sepeliendi]'.

37-38. MENTE ... OCVLIS. Similarly contrasted at Met XV 62-64 'isque, licet caeli regione remotos, / mente deos adiit et, quae natura negarat ['Medic. rectius' (Heinsius): negabat codd] / uisibus humanis, oculis ea pectoris hausit'.

38. NON ITA CAELITIBVS VISVM EST. 'The gods decided otherwise'. Compare xi 7 'non ita dis placuit', Met VII 699, Tr IV viii 15-16 (Ovid had hoped for a peaceful and happy old age) 'non ita dis uisum est, qui me terraque marique / actum Sarmaticis exposuere locis'. These passages are probably all echoes of Aen II 426 'dis aliter uisum'.

40. IVVET B^pcCMFHILT FORET B^ac 'unde uerum eliciendum'—Riese. But the correction is by the original hand (Owen suggested that the error was induced by foret at the end of the preceding distich), and iuuet is unobjectionable: Ovid is explaining his admission in the previous line that the gods were perhaps just in his case—claiming he was innocent, that is, that the gods had been unjust, would be of no assistance to him.

41. MENTE TAMEN, QVAE SOLA DOMO NON EXVLAT, VSVS. See at iv 45 qua possum, mente (p 211).

41. QVAE SOLA DOMO NON EXVLAT. Similar wording at Tr III iv 45-46 'Nasonisque tui quod adhuc non exulat unum / nomen ama'.

41. DOMO NON EXVLAT. Domo is my conjecture for the transmitted LOCO, which is strange and difficult to construe. FOCO is also possible; but the singular would be unusual. For domo compare Ter Eun 610 'domo exulo nunc'.

42. PRAETEXTAM FASCES ASPICIAMQVE. The -que logically belongs with fasces, joining it with praetextam: such dislocations are common in the pentameter because of its strict metrical requirements.

According to the manuscripts the preceding line ends with VTAR; I have printed Heinsius' VSVS, since there would otherwise be an asyndeton between utar and aspiciam. There are similar errors at 57 and xi 15 (cedet for cedens; peruenit for perueniens): here we may have a deliberate alteration by a scribe who did not understand the force of the delayed enclitic and sought a verb to couple aspiciam with.

44. DECRETIS Korn SECRETIS codd SECRETO Wheeler. Korn's conjecture makes the pentameter an amplification of the hexameter, a common pattern in Ovid; its corruption to secretis would be easy. Ehwald (KB 39-40) retained secretis, citing Tac Ann III 37 'secreta ['solitary designs'—Grant] patris mitigari' and Pliny Pan 53 6 (we should rejoice in our present good fortune under Trajan, and weep at the tribulations endured under previous emperors) 'hoc secreta nostra ['our private thoughts'], hoc sermones, hoc ipsae gratiarum actiones agant'. But in a list of the consul's public functions such a deviation of subject seems inappropriate. Wheeler's secreto is a little forced: 'my mind ... shall fancy itself present unseen at your actions'. Ehwald objected that Korn did not explain what his conjecture meant; but decernere was used of the consuls' judicial decisions (Cic Att XVI xvi a 4(6) 'consulum decretum').

45. LONGI ... LVSTRI. The epithet seems to have no special force: compare iv 23 'longum ... annum'.

45. REDITVS HASTAE SVPPONERE. See at v 19 reditus ... componet (p 219).

46. CERNET PM^2c, Gothanus membr. II 121 (saec xiii) CREDET BCFHILT. Cernet seems preferable to credet as continuing the image of uidebit in 43.

46. EXACTA CVNCTA LOCARE FIDE. Graecinus will be careful and incorruptible in assigning taxation contracts. For fide compare v 20 'et minui magnae non sinet urbis opes'; for exacta compare Suet Tib 18 'cum animaduerteret Varianam cladem temeritate et neglegentia ducis accidisse ... curam ... solita [scripsi; confer Liu XXVII 47 1 'multitudo ... maior solita' solito codd] exactiorem praestitit'.

48. PVBLICA QVAERENTEM QVID PETAT VTILITAS. The consul acted as chairman of the Senate, proposing the order of the day, and asking the senators in order of seniority for their sententiae on the appropriate action for the question under discussion.

48. PVBLICA ... VTILITAS. 'The people's interest'. For utilitas compare Met XIII 191 'utilitas populi', Cic Part Or 89 'persaepe euenit ut utilitas cum honestate certet', Cic Sul 25 'populi utilitati magis consulere quam uoluntati', and Livy VI 40 5 & VIII 34 2 'posthabita filii caritas publicae utilitati'.

49. PRO CAESARIBVS = pro Caesarum factis. Compare Res Gestae 4 'ob res a me aut per legatos meos auspicis [=auspiciis] meis terra marique prospere gestas quinquagiens et quinquiens decreuit senatus supplicandum esse dis immortalibus. dies autem per quos ex senatus consulto supplicatum est fuere DCCCLXXXX'.

49. CAESARIBVS. Tiberius, Germanicus, and Drusus. Similarly used at EP II vi 18 (to Graecinus) 'omnia Caesaribus [Augustus and Tiberius] sic tua facta probes'.

49. DECERNERE GRATES. 'Propose (in the Senate) the decreeing of thanks'. The sense of decernere is common in prose: see Cic Prou Cons 1, Att VII i 7, and the other passages at OLD decerno 6.

49. GRATES appears occasionally in prose (Tarrant at Sen Ag 380 reddunt grates cites Livy XXIII 11 12, Curtius IX 6 17, and Vell Pat II 25 4), but in hexameter and elegiac verse is the necessary representative for grātĭās.

51. CVM IAM FVERIS POTIORA PRECATVS. For potior 'more important' compare Caesar BC I 8 (a reported remark of Pompey) 'semper se rei publicae commoda priuatis necessitudinibus habuisse potiora', Livy VIII 29 2, and the many passages at OLD potior² 4. The usage belongs to prose: Ovid elsewhere and Virgil always use potior to mean either 'more powerful' or 'preferable'.

53-54. SVRGAT ... DETQVE. The apodosis of an implied condition: 'If you prayed for me, the fire would rise'.

53. SVRGAT AD HANC VOCEM PLENA PIVS IGNIS AB ARA. The same favourable omen at Met X 278-79 (Pygmalion has finished his prayer to Venus) 'amici numinis omen, / flamma ter accensa est apicemque per aera duxit'.

53. PLENA ... AB ARA. Another indication of Graecinus' devotion to the Caesars.

53. PIVS. 'Holy'; compare pia tura at Am III iii 33, Met XI 577, and Tr II 59, pia sacra at Tr V v 2, and pio ... igne at Tr V v 12.

54. LVCIDVS. Proleptic: 'The flame-tips would become bright and furnish a good omen for your prayer'.

55. NE CVNCTA QVERAMVR. 'So that not everything I say will be a complaint'.

57. LAETITAE EST LT. Most manuscripts have LAETITIA EST. Similarly at Met VIII 430 'illi laetitiae est cum munere muneris auctor' most codices read laetitia est. Heinsius thought LAETITIAE possibly correct here, as might be the case also in the Metamorphoses: laetitiae could easily have been misread as laetitia ē [=est], with laetitiae est as a later correction.

58. FRATER. L. Pomponius Flaccus (PIR¹ P 538), consul ordinarius for 17. As the greater honour would indicate (Graecinus was consul suffectus), Flaccus was more prominent than his brother and, unlike Graecinus, is several times mentioned in literary sources outside Ovid. At II 129 Velleius Paterculus speaks of Flaccus' ability and modesty, and Suetonius (Tib 42 1) names him as a drinking-companion of the emperor, made propraetor of Syria by Tiberius. Tacitus says that Flaccus proposed the supplicationum dies following the discovery in 16 of Libo's plot against Tiberius (Ann II 32 3); at Ann II 41 2 he names Flaccus as consul at the time of Germanicus' great triumph in 17, and at VI 27 3 mentions Flaccus' death in 34 while propraetor of Syria. For Flaccus' special mission to Thrace shortly after the time this poem was written, see at 75 (p 308).

EP I x is addressed to Flaccus, but gives little information except that Flaccus had, like Graecinus, given help to Ovid (37-40). Ovid's relations with Flaccus were clearly not as intimate as those with his brother.

59-60. The distich may be an interpolation, or at least deeply corrupted in its present form. Professor E. Fantham points out to me that the construction of die with both summo ... Decembri and Iani is awkward, and that dies Iani does not seem to be used elsewhere in Latin literature. The tense of suspicit is strange as well: a future would normally be expected here.

61. QVAEQVE EST IN VOBIS PIETAS. 'Your family-feeling is so great that ...' The same idiom at Met V 373 'quae iam patientia nostra est', EP I vii 59, EP II ii 21-22 'quaeque tua est pietas in totum nomen Iuli, / te laedi cum quis laeditur inde [=ex illis] putas', and Hor Sat I ix 54-55 'quae tua uirtus, / expugnabis'. The sense is frequent in prose (OLD qui¹ A 12).

The expression is used as a simple relative with the implication of size only from context at Tr III v 29 'quaeque tibi linguae est facundia, confer in illud' and Tr III vi 7-8 'quique est in caris animi [codd: animo fort legendum; uide ad 91] tibi candor amicis— / cognitus est illi quem colis ipse uiro'.

61-62. ALTERNA ... GAVDIA. Flaccus will first rejoice to see Graecinus become consul; then Graecinus will have the pleasure of seeing Flaccus consul.

64. BINVS seems sufficiently confirmed, as Ehwald points out (KB 51-52) by bis ... bis in the preceding line; BIMVS, conjectured by Heinsius and found in certain late manuscripts, seems ingenious but unnecessary. Ehwald compares Ecl III 30 'bis uenit ad mulctram, binos alit ubere fetus'.

64-65. HONOR ... INGENS. At vii 17 Ovid calls the rank of primipilaris 'titulus ... ingens'.

65-66. MARTIA ... ROMA. The same phrase at Tr III vii 52 and EP I viii 24; compare as well Aen I 276-77 'Romulus ... Mauortia condet / moenia'. Mars, father of Romulus and Remus, was peculiarly the god of Rome: compare Fast I 39-40 & III 85-86 'Mars Latio uenerandus erat, quia praesidet armis: / arma ferae genti remque decusque dabant'.

The reference to Mars is very apt in view of the primarily military nature of the republican consul's office.

67. MVLTIPLICAT TAMEN HVNC GRAVITAS AVCTORIS HONOREM. Flaccus had been nominated for the consulship by Tiberius.

For language and sentiment compare Met VIII 430 'illi laetitiae est cum munere muneris auctor'.

67. GRAVITAS is linked with Hercules at Met IX 270, with Jupiter at Met I 207 (considered suspect by Merkel) and II 847, with all the Olympian gods at Met VI 73, and with Augustus at Tr II 512. Underneath the ostensible connection to Jupiter at Met II 846-47 'non bene conueniunt nec in una sede morantur / maiestas et amor' Professor R. J. Tarrant sees an allusion to Augustus.

69-70. IVDICIIS IGITVR LICEAT FLACCOQVE TIBIQVE / TALIBVS AVGVSTI TEMPVS IN OMNE FRVI. Compare EP II vi 17-18 (to Graecinus) 'quodque soles animo semper, quod uoce precari, / omnia Caesaribus sic tua facta probes'.

70. AVGVSTI = Tiberii; his name in inscriptions is TI·CAESAR·AVG (Sandys 235).

71. CVM FILT QVOD BC VT MH QVVM Weise. The archetype was illegible at this point, and the manuscripts offer various supplements. Of these cum seems the most appropriate. Ehwald favoured quod (KB 48), but all except one of the passages he cited are instances of quod superest or quod reliquum est. The one relevant passage he cited was Fast II 17-18 (to Augustus) 'ergo ades et placido paulum mea munera uultu / respice, pacando si quid ab hoste uacat'. Many manuscripts however offer uacas (for which compare Prop II xxxii 7 'quodcumque uacabis'), and the corruption to the third person seems an easy one. Vacare in general does not seem to occur with an expressed impersonal subject.

71. CVRA PROPIORE. The same phrase at Met XIII 578-79 'cura deam propior luctusque domesticus angit / Memnonis amissi'.

73. SI QVAE DABIT AVRA SINVM. 'If some wind should give the opportunity of filling my sails'. Quae is my correction for QVA (CMFHIL), which would make the sentence mean 'If the wind should in some way ...'. The difficulty here is with the apparently already existing aura: what breeze is Ovid referring to? QVEM (BT) presents the same difficulty ('If the breeze should offer any opportunity ...') and in any case looks like a scribal correction. I take qua to be an unmetrical form corrupted from the rare form quae of the indefinite adjective. For the form, compare Ter Heaut 44 'si quae [Bembinus (saec iv-v): qua recc] [sc fabula] laboriosast, ad me curritur', Hor Sat I iv 93-95 'mentio si quae [uar qua] ... te coram fuerit, defendas, ut tuus est mos', Hor Sat II vi 10 'o si urnam argenti fors quae mihi monstret', and CIL I 583 37 'SEIQVAE CAVSA ERIT'. Quae in the present passage offers the same notion of a fresh breeze rising as is found at viii 27-28 'quamlibet exigua si nos ea [sc ara] iuuerit aura, / obruta de mediis cumba resurget aquis' and Tr IV v 19-20 'remis ad opem luctare ferendam / dum ueniat placido mollior aura deo'.

Quae should possibly be written at Met VI 231-33 'praescius imbris ... rector / carbasa deducit ne qua leuis effluat aura', but Professor R. J. Tarrant points out that qua can be defended by taking leuis to mean 'nimble', a sense supported here by effluat. A strong case could be made for reading quae at Hor Carm III xiv 19-20 'Spartacum si qua potuit uagantem / fallere testa'.

73. SINVM. Sinus in the sense of 'sail' is common enough (Am II xi 38, AA III 500, Fast V 609, and Aen III 455 & V 16; the origin of the metonymy seen at Prop III ix 30 'uelorum plenos ... sinus'); but the brachylogy here 'opportunity of filling my sails' is remarkable.

73. LAXATE editio princeps Romana IACTATE codd. Korn, Lenz, and André print the manuscript reading, and Korn offers three parallel passages in its defence, none of which stands up to examination. The first is EP III ii 5-6 'cumque labent alii iactataque uela relinquant, / tu lacerae remanes ancora sola rati', where iactata means 'storm-whipped'; compare Statius Theb VII 139-41 'uento / incipiente ... laxi iactantur ubique rudentes'. At Cic Tusc V 40 (a Spartan to a wealthy sea-merchant) 'non sane optabilis quidem ista ... rudentibus apta fortuna', 'Well, your fortune depends on your cables, and I don't think it something to be sought for', iactare does not appear. The third passage, Virgil G II 354-55 'seminibus positis superest diducere terram / saepius ad capita ['roots'] et duros iactare bidentis', hardly seems relevant.

For laxate rudentes 'let out the sails' Heinsius cited Aen III 266-67 'tum litore funem / deripere excussosque iubet laxare rudentis' 'Next he commanded us to fling hawsers from moorings and uncoil and ease the sheets' (Jackson Knight), Aen VIII 707-8 'uentis ... uela dare et laxos iamiamque immittere funis', Cic Diu I 127, Lucan V 426-27 'pariter soluere rates, totosque rudentes / laxauere sinus', and Lucan IX 1004.

74. E STYGIIS ... AQVIS. Similar phrasing at Met X 697 'Stygia ... unda, Met XI 500 'Stygia ... unda', Aen VI 374 'Stygias ... aquas', Aen XII 91 'Stygia ... unda', and Cons Liu 410 'Stygia ... aqua'.

Ovid often uses the phrasing of his exile: see Tr I ii 65-66 'mittere me Stygias si iam uoluisset in undas / Caesar, in hoc uestro non eguisset ope', Tr IV v 22, EP I viii 27 'careo uobis, Stygias detrusus in oras', and EP II iii 44 'a Stygia quantum mors [codd: sors Heinsius] mea distat aqua?'. For Ovid's exile as the equivalent of death, see at vi 49 qui me doluistis ademptum (p 243).

75. PRAEFVIT HIS ... LOCIS MODO FLACCVS. At Ann II 64-67 Tacitus reports how, following the death of Augustus, Rhescuporis attacked and imprisoned his brother Cotys (addressee of EP II ix), alleging a plot against himself; on their father's death, the kingdom of Thrace had been divided between them, Cotys receiving the better regions. Tiberius insisted that Rhescuporis release his brother and come to Rome to explain the situation; Rhescuporis then killed his brother, claiming it was a suicide. 'nec tamen Caesar placitas semel artes mutauit, sed defuncto Pandusa, quem sibi infensum Rhescuporis arguerat [scripsi: arguebat M], Pomponium Flaccum, ueterem stipendiis et arta cum rege amicitia eoque accommodatiorem ad fallendum ob id maxime Moesiae praefecit'; the previous service mentioned by Tacitus is no doubt the command Ovid is here referring to.

Flaccus succeeded in trapping Rhescuporis and bringing him to Rome; he was found guilty and sent in exile to Alexandria, where he died. Velleius Paterculus placed the episode first in his list of memorable events of Tiberius' reign (II 129); it is briefly mentioned at Suet Tib 37 4.

75. FLACCVS. 'Ab hoc Flacco uolunt quidam Valachiam ['Wallachia'] fuisse dictam olim Flacciam, quod nomen sensim corruptela sermonis transiit in Valachiam. Vide Georgii a ['von'] Reychersdorff Chorographiam Transyluaniae. pag. 33 [first published in 1595; see British Museum Gen Cat 200 383] qui addit hinc [sic] adhuc Romanum ibi sermonem durare, licet admodum corruptum. sed hae fabulae'—Burman. Clearly the existence of Rumanian was not widely known in Western Europe at the time Burman wrote.

77. MYSAS GENTES = Moesos. Strabo (VII 3 10; cited by André) claims a common origin for the Μοισοί of Europe and the Μυσοί of Asia. For the Greek form, compare Ovid's use of Getes for Geta and Sauromates for Sarmata.

78. ARCV FISOS ... GETAS. For the bow as the typical Getic weapon, see iii 52 'arcu ... Gete", EP III v 45 'Getico ... arcu' and Ibis 635 'Geticasque sagittas'.

78. ENSE. The gladius, typical weapon of the Roman legionary. For the precise equivalence of the two terms, see Quintilian X i 11. In Ovid's poetry, the proportion of instances of ensis to instances of gladius is about 90:30; in the poetry of exile, it is 21:3. For a discussion of ensis/gladius, with statistics, see Axelson 51; the only poets to admit gladius more freely than Ovid are Lucan and Juvenal.

79. TROESMIN Heinsius TROESMEN C TROESENEN B¹ TROEZEN uel similia codd plerique. Troesmis, the modern Galaţi, is located on the north bank of the Danube, about 160 kilometres inland from Aegissos (Tulcea). Heinsius did not have the assistance of CIL V 6183-88 & 6195, but seems nonetheless to have conjectured that Troesmin was a possible reading ('sed legendum, Τρωισμὶς uel Τρωσμίς'). Korn was the first to place Troesmin in the text.

79. CELERI VIRTVTE. 'With a bold surprise attack'.

80. INFECITQVE FERO SANGVINE DANVVIVM. Compare the similar description of Vestalis' recapture of Aegissos: 'non negat hoc Hister, cuius tua dextera quondam / puniceam Getico sanguine fecit aquam' (vii 19-20).

80. DANVVIVM. According to Owen at Tr II 192 this, and not DANVBIVM (the reading of the manuscripts), is the spelling certified by the inscriptions. Manuscripts divide between the two spellings at Hor Carm IV xv 21 and Tac Germ I 1.