I wish to thank Alessandro Barchiesi, Mina Cerruti, Gian Biagio Conte, Andrea Cucchiarelli, Paolo Fedeli, Luigi Galasso, Francesca Lechi, Gianpiero Rosati, Alessandro Schiesaro, Fabio Stok and Gareth Williams for their useful criticism and suggestions. The majority of them do not completely agree with all of my arguments (the most striking exception is Signora Cerruti, who incidentally is my mother); I of course take full responsibility for any remaining weaknesses.

1. Stephen Hinds has very sharply considered the implications of this allusion: ‘Booking the Return Trip: Ovid and Tristia 1’, PCPS 31 (1985), 14–16Google Scholar; see also Williams, G.D., ‘Representations of the Book-Roll in Latin Poetry: Ovid, Tr. 1.1.3–14 and Related Texts’, Mnemosyne 45 (1992), 187fCrossRefGoogle Scholar.

2. I have briefly discussed some examples (from the Heroides and the Ars Amatoria) of quaero as a signal of the interpretational activity of Ovid’s model reader in ‘Strategies of Tension (Ovid, Heroides 4)’, PCPS 41 (1995), 8fGoogle Scholar.

3. André, J. (ed., trans.), Ovide: Tristes (Paris 1968), 3Google Scholar.

4. ‘[L]a lezione (quaerenti plura legendum) tràdita dai manoscritti…non è del tutto congruente col contesto: peralto nessuno fra i vari tentativi esperiti dai filologi può dirsi risolutivo.’ Lechi, F. (ed., trans.), Ovidio: Tristezze (Milan 1993), 64 n.8Google Scholar.

5. Luck, G. (ed., trans.), P. Ovidius Naso: Tristia, Band I (Heidelberg 1967), 29fGoogle Scholar.

6. Hall, J.B., in his new edition of the Tristia (Stuttgart and Leipzig 1995Google Scholar), also neutralises the tension through emendation: atque ita te cautus quaerenti plura legendum / ne, quae non opus est, forte loquare, dato (‘and thus be careful to present yourself for reading by the person who wants to know more, lest you chance to say what you ought not to’; see his apparatus, p. 2).

7. Wheeler, A.L. (ed., trans.), Ovid: Tristia, Ex Ponto (Cambridge MA and London 1924Google Scholar; 2nd ed. revised by G.P. Goold, 1988), 5. Wheeler read legendus: ‘for he who asks more must read you.’ The use of the gerund with a complementary object in the accusative, in place of the more usual gerundive, is infrequent but attested in Latin of all ages: cf. e.g. Lucr. DRN 1.111 aeternas…poenas in morte timendumst, and above all Virg. Aen. 11.230 pacem Troiano ab rege petendum; see Hofmann, J.B. and Szantyr, A., Lateinische Syntax und Stilistik (Munich 1965Google Scholar), #202 pp.373f. with Risch, E., Gerundivum und Gerundium: Gebrauch im klassischen und älteren Latein Entstehung und Vorgeschichte (Berlin and New York 1984), 186–88CrossRefGoogle Scholar. Goold’s translation makes perfect grammatical sense of the Latin; he probably thinks that the rhythm of the line tells in favour of the second construction. I believe that both the constructions are present. A true syntactical ambiguity. It is self-reflexive: the same phrase which invites the reader to ‘seek/read more’ can be read in two different ways.

8. Hinds (n.1 above), 19.

9. Cf. also A.A. 3.41f. quid uos [sc. the heroines who committed suicide out of love] perdiderit dicam? nescistis amare:/defuit ars uobis… ‘Shall I say what it was that ruined you? You did not know how to love; you lacked the art [/Ars]’.

10. H.B. Evans, Publico Carmina: Ovid’s Books from Exile (Lincoln and London), 58.

11. Williams, G.D., Banished Voices: Readings in Ovid’s Exile Poetry (Cambridge 1994), 105Google Scholar; we shall turn later to the other explanation considered by Williams, that of Marg, W., ‘Zur Behand-lung des Augustus in den Tristien’, Atti del Conv. Int. Ovidiano, Sulmona 1958 (Rome 1959), ii.349fGoogle Scholar.

12. Williams (nil above), 105.

13. Luck (n.5 above) and Hall (n.6 above) wrongly prefer the variant officium…meum.

14. The only addressees whom I find plausibly described are the sons of Messalla Corvinus (Tr. 4.4, perhaps to Messalinus; 4.5, and perhaps 5.9, to Cotta Maximus: see Syme, R., History in Ovid [Oxford 1978], 76Google Scholar), and possibly also Atticus: see Galasso, L. (ed.) P. Ovidii Nasonis Epistulae ex Ponto Liber II (Florence 1995), 29fGoogle Scholar.

15. Another means by which Ovid conveys this message is the storm with which the Tristia open immediately after the introductory poem. What could be more allegorical than a storm? What could be less credible for Ovid’s reader than the poet’s claim to have written Tristia 1 in the midst of terrible elemental fury, as if he were a Eumolpus ante litteram ? The reader is forced to ‘read more’ when he encounters Ovid in the midst of a storm (Tr. 1.2, 1.4), after having read at Tr. 1.1.81–88 that the ira of Augustus was like a storm. Perhaps Ovid tells us not of what occurred during his journey into exile, but of what occurred at Rome before the voyage? Surely Ovid gets off to a programmatic start by beginning his new production on an allegorical footing. I would have liked to discuss this topic in a paragraph entitled ‘Lector in incerto est’. For now, I refer the reader to Cucchiarelli, A., ‘La nave e l’esilio (allegorie dell’ ultimo Ovidio)’, MD 38 (1998Google Scholar).

16. On the figure of the naive reader, a reader ready to be exploited by the text, swallowing the bait and falling into the interpretational traps laid for him, a reader who exists as both a textual strategy and as an extra-textual being (i.e. in a bibliography), see my ‘Ovidio e la preconoscenza della critica. Qualche generalizzazione a partire da Heroides 14’, Philologus 141 (1997Google Scholar) and ‘Ovid, Apollo, and the Foreknowledge of Criticism (Ars 2.493–512)’, CJ 93 (1997Google Scholar).

17. Cf. Galasso (n.14 above), 25f.

18. Marg (n.l 1 above), ii.349f.

19. For the use of demere in this context cf. Ex Pont. 1.8.73f. terra uelim propior nullique obnoxia bello/detur: erit nostris pars bona dempta malis (‘I would wish that I was granted a land nearer home and one not exposed to war; a large part of my sufferings will then be removed’). Cf. further e.g. Tr. 2.185f. and 575–78, 3.8.21f., 4.4.51f., Ibis 27f., Ex Pont. 2.7.63f., 3.1.4 and 85, 3.7.30, 3.9.38, 4.8.85f.. But it is unnecessary to accumulate passages; see Ex Pont. 3.9.1–4, where Ovid claims that someone complains because his verse contains ‘nothing but petitioning to be transferred to a land nearer home and talk of the large number of enemies surrounding him’.

20. Sarcastic: Augustus presents himself as uindex libertatis (‘champion of liberty’; cf. RGDA 1.1 rem publicam…in libertatem uindicaui), but here his activity as a uindex is explained in terms of his taking Ovid’s freedom away from him.

21. Nagle, B.R., The Poetics of Exile: Program and Polemic in the Tristia and Epistulae ex Ponto of Ovid (Brussels 1980), 171Google Scholar.

22. Williams (n.l 1 above), 50–99.

23. Williams (n.l 1 above), 52.

24. Williams (n.l 1 above), 70.

25. The term used by A. Barchiesi in connection with the Fasti; see IL poeta e il principe: Ovidio e il discorso augusteo (Roma-Bari 1994), 69–129Google Scholar.

26. On this issue see Citroni, M., Poesia e lettori in Roma antica (Rome-Bari 1995), 440–74Google Scholar.

27. A figure known only from this Ovidian elegy; cf. Syme (n.14 above), 81 and 88, and see Helzle, M., Publii Ovidii Nasonis Epistularum ex Ponto Liber IV: A Commentary on Poems 1 to 7 and 16 (Hildesheim 1989), 184Google Scholar, on 4.16.7.

28. On the ambiguity of laudes see Williams (n.ll above), 95 n.99; cf. also Akrigg, G.M. (ed.), The Last Poems of Ovid: A New Edition, with Commentary, of the Fourth Book of the Epistulae ex Ponto (Diss. Toronto 1985), 400Google Scholar.

29. Williams (n.l 1 above), 95.

30. For the irony of nouitas cf. Claassen, J.-M., Poeta, Exsul, Vales: A Stylistic and Literary Analysis of Ovid’s Tristia and Epistulae ex Ponto (Diss. Stellenbosch 1986), 207Google Scholar: ‘The illogic of acceptance of Augustus as Jupiter from Tl. 1.81–84 onwards and sporadic wishes for (or threats of) his apotheosis, as the accompaniment of prayer (e.g. T5.2.52, T5.5.61–64, T5.8.29, T5.1.25–27) has its final incongruous reversal with the reference (P4.13.24) to the nouitas of his godship.’ It must be said, however, that nouitas can refer to the newness of Augustus’ godship only on a secondary level of reading; literally, the nouitas (‘novelty’, ‘strangeness’) is Ovid’s undertaking: to sing imperial praises in Getic.

31. Hartog, F., ‘Salmoxis: le Pythagore des Gètes ou l’autre de Pythagore?’, ASNP 8 (1978), 23Google Scholar. The important point is that for the Getae it was not a matter of simple belief in the soul’s immortality but of belief in a real and personal form of deification post mortem to which not just rulers but also ordinary people had access; on the theme see especially Coman, I.G., ‘L’immortality chez les Thraco-Géto-Daces’, RHR 198 (1981), 243–78CrossRefGoogle Scholar, who concludes his analysis thus (277): ‘L’immortalité thrace en général, et géo-dace en paiticulier, a donc constitué une caractéristique de première importance chez un peuple trés ancien, non sans grandeur, et surtout connu dans l’Antiquité parce qu’il paraissait avoir su s’assurer une immortalité divine.’ Cf. also Eliade, M., De Zalmoxis à Gengis-Khan (Paris 1970), 31–80Google Scholar; Bianchi, U., ‘Dualistic Aspects of Thracian Religion’, HR 10 (1971), 228–33Google Scholar. The main testimony for the deified Zalmoxis and on the immortality of the Getae is supplied by Herod. 4.93f., 5.4, Strabo 7.29, Diod. 1.94.2, Plat. Charm. 156d (on the Thracian doctors who were able to confer immortality), Apul. Apol. 26, Mela 2.18, Solin. 10.1–3, Mart. Cap. 6.656, etc.

32. Herod. 4.93; on the expression cf. I.M. Linforth, ‘οί άθανατίζοντε<, (Herodotus iv.93–96)’, CP 13 (1918), 23–33; Eliade (n.31 above), 40 and n.22. Herodotus then explains: ‘they believe themselves (or become) immortal in this way: they think that they are not affected by death, and that in departing from life one goes to join Zalmoxis, a divine being’ (4.94.1). On the Herodotean passage cf. also Barié, P. ‘“Interpretatio” als religionspsychologisches Phänomen’, AU 28 (1985), 82–86Google Scholar.

33. Cf. Ex Pont. 1.5.66,3.5.28.

34. Apul. Met. 5.8; but cf. already Sen. Dial. 6.4.1.

35. Cf. Barchiesi (n.25 above), 30: ‘Scuotendo le faretre i Geti segnalano la loro ammirazione per un tipo di poesia che si rivela perfettamente adeguata alla loro cultura… La riconciliazione tra Ovidio e Augusto ha partorito una poesia barbara per uditori barbari vibranti di armi.’ In considering the irony implicit in Ex Pont. 4.13 from a perspective which is slightly different but perfectly compatible with my own (see pp.28–30), Barchiesi suggests that the designation poeta Getes alludes to the proverbial ‘Geta’, the slave of comedy (cf. e.g. Prop. 4.5.44), ‘il servo scaltro che sotto sotto raggira i suoi padroni’. Cf. A.A. 3.332 (pater uafri luditur arte Getae, ‘the father is taken in by the tricks of the crafty Geta’); ‘come poeta getico, Ovidio si sta piegando con reverenza alia divinità di un personaggio definito pater Augustus (v.25), e questo potrebbe essere l’ultimo trucco di “Geta” alle spalle del suo signore’ (p.30).

36. Cf. Akrigg (n.28 above), 406.

37. Allowance must be made for what was surely the natural attitude of the Getae towards Augustus; in this respect A. Grisart (in Atti [n.l 1 above], i.80) makes an important point in stating that ‘[L]a simpatia dei Geti per Ovidio si spiega anche pensando che essi vedevano in lui una vittima del sovrano del populo che li opprimeva.’

38. Horn. Il. 2.673–75: ‘Nireus, the most handsome man who came to Troy from among all of the Danaans, after the excellent son of Peleus. But he was a coward, and a small host of men followed him’. Cf. e.g. Quint. Inst. 3.7.19 …quibusdam bona uitiis corrupta odium attulerunt, ut Nirea in bellem, Plisthenen inpudicum a poetis accepimus (‘the corruption of their good qualities by other defects has brought dislike upon certain people, just as we are told by the poets that Nireus was a coward, Plisthenes immoral’).

39. Cf. Galasso (n.14 above), 183.

40. Cf. also Barchiesi (n.35 above), 29: ‘[L]e sue poesie sono uitiosa, tecnicamente imperfette (…;ma l’aggettivo si può rendere anche come “perverse”, “depravate”)…’

41. On Brutus see Helzle (n.27 above), 136–38. On Ex Pont. 3.9 see Block, E., ‘Poetics in Exile: An Analysis of Epistolae ex Ponto 3.9’, CA 1 (1982), 18–27Google Scholar; Williams (n.l 1 above), 86–89.

42. On the clear absurdity of the geographical, climatic and anthropological claims which Ovid makes for Tomis see Williams (n. 11 above), chapter 1, on ‘The “unreality” of Ovid’s exile poetry’.

43. Nugent, G., ‘Tristia 2: Ovid and Augustus’, in Raaflaub, K.A. and Toher, M. (eds.). Between Republic and Empire (Berkeley and Los Angeles 1990), 239–57Google Scholar; Barchiesi, A., ‘Insegnare ad Augusto: Orazio, Epistole 2.1 e Ovidio, Tristia II’, MD 31 (1994), 149–84Google Scholar; and Williams (n.l 1 above), 154–209.

44. At Tr. 2.281 and 458 pecco is used in the erotic sense of ‘being unfaithful’; at 4.4.9 nil ego peccaui, of disclosing the identity of his addressee (perhaps Messalinus); at 5.13.11 the addressee ‘sins’ in rarely writing to Ovid; at Ex Pont. 4.1.5 Ovid ‘sins’ in naming Sextus Pompeius in his verse. On the use of pecco/peccatum in reference to Ovid’s fall see also Galasso (n.14 above), 207, on Ex Pont. 2.3.33.

45. From this perspective the final touch which the Metamorphoses lacks and of which Ovid writes in Tr. 1.7 could also refer to a lack of self-critical intervention on the poet’s part in the Augustan sections of the work. The final touch could have eliminated the irreverent ‘anti-Augustan’ element, but Ovid fails to apply it; cf. 1.7.39f. quicquid in his igitur uitii rude carmen habebit / emendaturus, si licuisset, erat (‘all the faults which the rough poem [rudis also suggests “uncouth” in the sense of “untamed”, “unruly”] will still contain, the author would have corrected, had he been permitted to do so’). On the political incorrectness of the Metamorphoses cf. recently Nugent (n.43 above), esp. 256f.

46. Already Polybius cites the praising of Thersites and criticism of Penelope as examples of subjects treated in the schools; see Walbank, F.W., A Historical Commentary on Polybius (Oxford 1967), 404fGoogle Scholar. on 12.26b.5; Pernot, L., La Rhétorique de l’éloge dans le monde greco-romain (Paris 1993), i.44Google Scholar and n.

47. See my note on Her. 9.69f. (Florence 1995), 115f. (praise of Busiris).

48. At Am. 2.6.41 Thersites serves as an example of the worse outliving their betters, but no particular reference is made to his ugliness; Thersites attended Protesilaus’ funeral, while Hector did not survive his brothers (for the injustice of Thersites surviving his betters cf. Soph. Phil. 438–52). At R.A. 482 Thersites is evoked in an amusing reconstruction of Agamemnon’s words in Iliad 1 (‘if I am your king, and if Briseis does not spend her nights with me, then let Thersites become your ruler’); Agamemnon’s words themselves presuppose those of Thersites when in Iliad 2, the latter accuses the king of ‘wanting some young girl, to sleep with and to keep as his private property’ (232f.; cf. Pinotti, P. [ed.], P. Ovidio Nasone: Remedia Amoris [Bologna 1988], 230f.Google Scholar).

49. On the figure of Thersites, in addition to the standard reference works (V. Gebhard, RE 5 A2 [1934], 2455–71; Schmidt, J., Roschers Lexicon 5 [1916–24], 665–75Google Scholar), there is a vast bibliography; see Seibel, A., ‘Widerstreit und Ergänzung: Thersites und Odysseus als rivalisierende Demagogen in der Ilias’, Hermes 123 (1995), 388Google Scholar n.12. On Thersites in Latin literature see Spina, L., ‘Tersite a Roma’, Vichiana 13 (1984), 350–63Google Scholar with La Penna, A., Tersite censurato e altri studi di letterature fra antico e moderno (Pisa 1991), 113–29Google Scholar.

50. See Chantraine, P., ‘A propos de Thersite’, AC 32 (1963), 18–27Google Scholar; cf. Kirk, G.S. (ed.), The Iliad: A Commentary. Volume I: Books 1–4 (Cambridge 1985), 138CrossRefGoogle Scholar: ‘Thersites is a “speaking” name formed from θέρσος,, the Aeolic form of Ionic θάρσος implying either boldness or rashness—in his case, obviously the latter.’ If the exilic works are a Thersites, and are Ovid’s offspring, the name of their father, Agrius, may also be significant: ‘what better “significant name” for the father of such an agitator than Aypioc,?’ (H.D.Rankin, Thersites the malcontent, a discussion’, SO 47 [1972], 47n.40Google Scholar).

51. One aspect of the Homeric description may have special relevance for Ovid. Thersites is XcoX6?…exepov 7t68a (‘lame-in one foot’); cf. Tr. 3.1.1 If. clauda quod alterno subsidunt carmina uersu,/uel pedis hoc ratio, uel uia longa facit (‘as for the fact that my limping couplets fall away in each second line, it is due either to the form of the metre or to the length of my journey’). Cf. Ex Pont. 4.5.3.

52. Cf. for the ‘disorderly words’ of Thersites as a significant detail Gell. NA. 1.15.11: neque non merito Homerus unum ex omnibus Thersitam ἀμετρoεπἆ et ἀκριτóμυθov appellat uerbaque illius multa et ἄκoσμα strepentium sine modo graculorum similia esse dicit. quid enim est aliud ἐκoλώα ? (‘And Homer deservedly calls Thersites, and him alone of all, unbridled of tongue and a confused babbler, and he says that his many disorderly words resemble the unrestrained squawking of jackdaws. For what else does ἐκoλώα mean?’)

53. Thersites’ ugliness is explicitly mentioned in surprisingly few passages; cf. Quint. Inst. 3.7.19, where Thersites is set alongside Irus among figures scorned because of their corporis ac fortunae…mala. The focus is on his ugliness at Anth. Lat. 749.13 (i.2, p.271 Riese); cf. Apul. Flor. 3.4 Helm (text uncertain), Boeth. Herm. I p.96.1 Meiser, Cassiod. Hist, trip. 6.44 (PL 69.1060 A). Thersites is thus usually assumed to be ‘the worst of men’, but his words (as opposed to his appearance) tend to characterise him as such. See Spina (n.49 above) for a full list of citations of Thersites in Latin literature. The Ilias Latina (136f.) takes up both aspects of his character: Thersites, quo non deformior alter/uenerat ad Troiam nee lingua proteruior ulli (‘Thersites, the most ugly man of all, had come to Troy, and no one had a more impudent tongue’).

54. Cf. Arctinus, Aethiopis; see Bernabé, A. (ed.), Poetarum Epicorum Graecorum Testimonia et Fragmenta (Leipzig 1987), 67fGoogle Scholar. For discussion of the different traditions regarding Thersites see Rankin (n.50 above), 39ff. That Ovid wants to suggest that he has read an Iliad furnished with Alexandrian commentary seems to me clear from the later reference (Ex Pont. 3.9.23f.) to Aristarchus’ philological work on Homer (Aristarchus’ sole appearance in Ovid).

55. With a punch, e.g. schol. ad Soph. Phil. 445 (364.1 Iff. Papageorgios); Quint. Smyrn. 1.741ff.; schol. ad Lycophr. 999 (312.8ff. Scheer); Eustath. Il. 208.2ff. = i.317 van der Valk; Tzetz. Posthom. 205 and on Lycophr. 999 (312.1 If. and 23f. Scheer); with a spear only in Lycophron (1001). On the death of Thersites see Morelli, G., ‘La morte di Tersite nella “Tabula lliaca” del Campidoglio’, in Pretagostini, R. (ed.), Tradizione e innovazione nella cultura greca da Omero al età ellenistica: Scritti in onore di Bruno Gentili (Rome 1993) 1143–53Google Scholar.

56. Cf. Ahl, F., ‘The Art of Safe Criticism in Greece and Rome’, AJP 105 (1984), 174–208Google Scholar; Williams (n.ll above), 158–61; Orlando, F., Illuminismo e retorica Freudiana (Turin 1982), 164ffGoogle Scholar.

57. Williams, G.D., The Curse of Exile: A Study of Ovid’s Ibis, PCPS Suppl. Vol. 19 (Cambridge 1996Google Scholar), passim, esp. Chapter 1 ‘In search of Ibis’. Before this, much excellent material could be found in Watson, L., Arae: The Curse Poetry of Antiquity (Leeds 1991Google Scholar).

58. Cf. e.g. the passages collected by Nisbet, R.G.M. and Hubbard, M., A Commentary on Horace Odes Book I (Oxford 1970Google Scholar) on 1.34.12 and 1.35.2.

59. The text of line 30 is very uncertain; see Hall’s apparatus (n.6 above). Hall prints rogata (Owen) petam (MSS.); dabit is T. Faber’s conjecture.

60. Cf. e.g. Virg. G. 1.500ff., Hor. C. 1.2.45f. (with Nisbet and Hubbard [n. 58 above], 45 ad loc.), Prop. 3.11.50; see further Veil. 2.131, Sen. Dial. 11.12.5, Luc. 1.45f., Calp. Sic. 4.145f, Sil. 3.625ff., Stat. Silu. 1.1.105ff., 4.2.22, 57ff., Plin. Ep. 10.1.1, Mart. 13.4. In Ovid cf. M. 15 448f., 838f., 868–70; Tr. 2.57, 5.5.61f., 5.11.25f., Ex Pont. 2.8.41f.

61. Cf. Baehrens, E., ‘Ire = sterben’, Glotta 5 (1914), 98Google Scholar, citing Lucr. DRN 3.526 hominem paulatim cernimus ire and Stat. Silu. 2.1.218f. ibimus…ibimus as the sole classical examples of this use of ire; but others are to be found (e.g. Prop. 4.7.23; cf. TLL s.v. 636.33ff, 643.40, 82; OLD s.v. 4b). It is probably a colloquialism; see Donat. on Virg. Aen. 3.10; Van Dam, H.J., P. Papinius Statius: Silvae Book II (Leiden 1984), 179Google Scholar.

62. Or orb; see on the distinction Galasso (n. 14 above), 215f.

63. Cf. Helzle (n.27 above), 103, who, however, speculates on the ‘real’ identity of the addressee (presumably a poet) in a manner unacceptable to me.

64. See Watson (n.57 above), 42ff.

65. Tr. 2.133f., tristibus inuectus uerbis (ita principe dignum)/ultus es offensas, ut decet, ipse tuas (‘inveighing against me with stern words—as befits a ruler—you yourself, as is appropriate, avenged your injury’); Ex Pont. 2.7.56, addita sunt poenis aspera uerba meis (‘harsh words were added to my punishment’).

66. The word ibis occurs elsewhere in the exile poetry only at Tr. 1.1.1 and 4.2.49.

67. The space at my disposal forces me to be brief, but each elegy addressed to an enemy can, I suggest, be viewed as a source of insults and invective against Augustus; I aim to set out the material on this topic rather more fully in a later work.

68. The first time that I heard it suggested that ‘Ibis is Augustus’ was in informal conversation with Alessandro Schiesaro. I thank him for arousing my suspicions. Schiesaro will discuss the Ibis from this perspective in a work soon to be published. The medieval tradition to which Brunetto Latini refers in the passage of Tresor cited in the epigraph above was pointed out to me by Francesca Lechi.

69. The concept of the ‘figure of Freudian substitution’ is elaborated and developed by Orlando (n.56 above) passim (a book which is very important for understanding Ovid’s exile poetry). If we would want to take up (in terms of substitution) Ovid’s attitude towards Augustus in the rest of the exile poetry, the formula would be: I/do not attack/him; i.e. I revere him, but in a way so excessive and grotesque that who wants to understand will understand.

70. That the Roman reader was ready to grasp the double sense of Ibis/ibis is clear, but also finds strong confirmation from the fact that Horace had begun the Epodes with the word ibis which, considered as a form of title (as often happened to the first words of a Roman poetic work; e.g. arma uirumque = the Aeneid), probably constituted an ‘allusion to Callimachus’ invective poem Ibis’ (S.J. Heyworth, ‘Horace’s Ibis: On the Titles, Unity, and Contents of the Epodes’, PLLS 7 [1993], 86). A ‘trasparente giuoco di parole fondato sul titolo Ibis’ was detected by Rostagni, A., Ibis: Storia di un Poemetto Greco (Florence 1920), 29Google Scholar on Tr. 4.9.19 (ibunt) and 21 (ibit).