2. For example: Martin, J. (Würzb. Jahrb. 1 (1946) 98–107Google Scholar)saw the connection between E. 1.45 and E. 6.4–5, and pointed out the parallel between the god's address to Tityrus and that of the Muses to Hesiod; Hanslik, Rudolf (WS 68 (1955) 5–19)Google Scholar developed these ideas and detected the casting of Octavian as Apollo; Pösch, Viktor (Die Hirtendichtung Virgils (1964) 9–15)Google Scholar expounded the essentially programmatic nature of E. 1.1–5. I had arrived independently at a formulation of these views before reading their work.

3. In general terms the contention that the nature of the poet's own activity is a major topic of the Eclogues is far more widely accepted than when I began to write. Indeed, many recent treatments go a good deal further than I would wish to in their speculations. Substantial contributions on this theme will be found in Schmidt, E. A., Poetische Reflexion: Vergils Bukolik (1972)Google Scholar, Berg, W., Early Virgil (1974)Google Scholar, Van Sickle, J., The design of Virgil's Bucolics (1978)Google Scholar, and Alpers, P., The singer of the Eclogues (1979)Google Scholar.

4. e.g. Williams, Gordon, Tradition and originality in Roman poetry (1968) 328–9Google Scholar.

5. I shall subsequently argue that there is no absence of a dedication in E. 1. It should also be noted that Bowersock, G. W. (HSCP 75 (1971) 73–80)Google Scholar has presented a powerful case for identifying the addressee of E. 6.13 as Octavian, not Pollio, with a consequent advancing of the terminus post quem for the publication of the collection to 35 B.C. This has led Clausen, Wendell (HSCP 76 (1972) 201–5)Google Scholar to argue that the historical circumstances which appear to lie behind the attitudes expressed in E. 1 best accord with a date for final composition not before 35 B.C. My own view of the nature of the poem had driven me to similar conclusions, although I would not share the degree of confidence which permitted Clausen to introduce the later dating without caveat into The Cambridge history of classical literature II (1982) 309Google Scholar. The principal obstacles to accepting such a view in the past have been (1) the testimony of the scholiastic tradition that the Eclogues were composed in the years 42–39 B.C., and (2) the a priori (and usually unstated) belief that since the expropriations of land in Northern Italy formed the background to the narrative of the poem it must have been written within months of their occurrence in 40 or 39 B.C. I have nothing to add to Bowersock's trenchant demolition of the scholiastic tradition (73). The second obstacle is equally fragile. No argument can be advanced to prove that Virgil could not have chosen the expropriations as a suitable background for a Roman pastoral poem three or four years after the events. Furthermore, even if the poem had been originally conceived and written in 40 or 39 it could still, as I observe below (n. 6), have been subject to rewriting to an incalculable extent.

6. It is in principle impossible to date the whole of a Latin poem, particularly one in a collection composed over a number of years and subsequently published as a single entity. It may be possible to fix a terminus ante quem through an inclusion or omission which is inconceivable in certain datable historical circumstances. Such opportunities are likely to be few and highly debatable. All that can normally be established, if suitable evidence is available, is the latest historical event or circumstances reference to which is detectable in the work. This provides only a terminus post quem for publication. Given Roman habits of composition, circulation and revision, the poem may contain material from several reworkings at different dates. All that we usually possess is the final version. It is the datum upon which criticism should be based. Neglect of this elementary principle is one of the prime sources of the vast and fruitless literature upon the relative dating of the Eclogues. The order in which Virgil composed the poems is of no literary significance whatsoever. What matters is the effect upon the reader which results from publishing them in their present form and present order. It should perhaps be added that these facts render any attempt to use these poems as historical or biographical sources hazardous in the extreme. Even so sober a reconstruction as that of Winterbottom, Michael (G & R 2nd ser. 23 (1976) 55–9)Google Scholar is in the end of limited value. For further discussion of the ‘achronicity’ of the Eclogues, see Van Sickle, J., CW 75 (1981) 72Google Scholar and the literature cited there.

7. Catul. 1.1, Hor. Carm. 1.1, Tib. 1.1, Prop. 1.1, Ov. Am. 1.1, etc.

8. Theoc. 3.2, 7.72. Van Sickle (n. 3, 119–23) draws attention to the importance of the relationship between Idyll 7 and Eclogue 1 (cf. Ramus 5 (1976) 23–4Google Scholar).

9. A point noted by Pöschl (n. 2) 10, who also draws attention to the sense-echo of Theoc. 7.88-9. This combination of elements of two Theocritean passages is immediately characteristic of Virgil's methods. Although the order of the poems in ancient editions of Theocritus is uncertain, it seems safe to assume that Idyll 1 always appeared at the head of the collection (cf. Gow, A. S. F., Theocritus I (1950) lxvi–ix)Google Scholar. The point of the echo is clearly to recall the opening words of his chosen model.

10. Cf. Smith, Peter L., TAPA 101 (1970) 506–7Google Scholar.

11. For what follows cf. Damon, Philip, Cal. Pub. Class. Phil. 15 (1961) 281–9Google Scholar and the earlier literature cited by him.

12. E. 1.4–5, 2.4–5, 6.10–11, 82–4, 8.22–4, 10.8.

13. It is true that Virgil would have found this form of prooemium also in Greek poetry (e.g. Il. Parv. Fr. 1) and that both here and at A. 7.37–45 he combines it with the appeal to a divine inspirer. This does not detract from the primacy given to the human composer at the beginning of each of his works.

14. Cf. Hermesian. Fr. 2. 75–8, Call., Epigr. 22, 3–4Google Scholar, [Mosch.] 3.20–2.

15. E. 1.1, 2.3, 3.12, 5.13, 9.9.

16. Theoc.. 9.20, 12.8.

17. For a similar learned transliteration in a programme poem, cf. Cairns, Francis, CQ n.s. 24 (1974) 97–8CrossRefGoogle Scholar, where it is shown that hirsutas feras (Prop. 1.1.12) is based on (Hom. Il. 2.743). Professor Kenney, E. J. (in a paper now published in ICS 8 (1983) 44–59Google Scholar and of which he kindly lent me a typescript) also points to lepidus = λεπτός at Catul. 1.1 (50 n. 23).

18. Williams (n. 4) 317–9. Kenney (n. 17) 49–50 directs our attention rather to the appearance of the φηγός in Callimachus' story of Acontius and Cydippe (Frr. 67–75 Pf.). He notes the collocation of fagus and the Callimachean tenuis at the opening of E. 1. His contention that Acontius and Cydippe was an important source for Virgil in Eclogues 2 and 8 and is alluded to through the mouth of Gallus at E. 10.52–4 lends significant support to my own view (p. 130) that the reference to it in the opening lines of E. 1 goes beyond the word fagus. See also Du Quesnay 145 n. 117.

19. Fr. 69: φηγοὶ Πανὸς ἄγαλμα

20. E. 1.2, 5, 2.31, 4.3, 6.2, 10.8, 63.

21. E. 1.10, 6.8; 3.84.

22. Cf. Brink on Hor. Ars 203.

23. Fr. 1.11 Pf., Epigr. 27.3 (λεπτός), Fr. 1.24 Pf. (, cf. Pfeiffer ad loc.).

24. Theoc. 3, 4.36, 38; E. 1.5, 30, 36, 2.14, 52, 3.81, 8.77-8, 101, 9.22. Note to /formosam … Amaryllida.

25. Alpers (n. 3) 72–3, observes how the studied symmetry of the opening exchange between Thyrsis and the goatherd in Idyll 1 mirrors the perfect understanding they have of each other. In Virgil the absence of such symmetry (even the repeated vocative, O Meliboee, is very different in tone from Tityre) varies this Theocritean pattern in a way which emphasises the contrast between his characters. See also Du Quesnay 52-3.

26. Cf. Dover, K. J., Theocritus (1971) lvi–iiGoogle Scholar.

27. Idylls 1, 4, 5, 6. 7 and 10. In Virgil Eclogues 1, 3, 5, 7, 8, 9 contain an encounter.

28. Note that Tityrus willingly and joyfully promises the frequent sacrifice of a lamb to the god (7–8) while Meliboeus has unwillingly and miserably had to abandon twin kids (14–15).

29. As observed in the scholiastic tradition: Schol. in Theoc. Vet. pp. 5.2-5, 11.18–20 Wendel.

30. For the possible literary significance of this word, see p. 131.

31. As Du Quesnay notes (136) the word libertas at Prop. 1.1.28 refers to freedom in poetic composition.

32. E.g. Hor. S. 2.6.1–15, Carm. 3.16.22–44 (cf. Carm. 2.16.37–40). All these passages are resonant with the ideas of Callimachean poetics.

33. Eclogues 1, 2, 6, 9(?), 10.

34. Cf. Berg (n. 3) 144–6 and the literature cited there. See also Skutsch, O., HSCP 73 (1969) 167Google Scholar.

35. If these views incur the wrath of critics of allegorisation of the Eclogues so be it. I regard the word ‘allegory’ as a red herring. Saying that Tityrus is Virgil's poetic ‘persona’ is merely a convenient shorthand for the statement that the ideas about poetic composition which the poem communicates emanate from Virgil.

36. Cf. e.g. Catul. 116 as expounded by Macleod, C. W. (CQ n.s. 23 (1973) 304–9)CrossRefGoogle Scholar, Hor. Carm. 3.30.

37. Cf. Schmidt (n. 3) 32–8.

38. Doubtless with an allusion to the equally programmatic cricket-cage being woven in the first poem of ‘Theocritus’ collection (Id. 1.52–3).

39. I detect an identical movement of thought in Prop. 2.13, which makes sustained use of the Callimachean programme (cf. Wilkinson, L. P., CR n.s. 16 (1966) 141–4Google Scholar): sat mea sat magna est si tres sint pompa libelli/ quos ego Persephonae maxima dona feram (25–6). Whatever the correct reading at the beginning of 25, an apparent contrast is clearly intended between the self-deprecating sat = a form of magnus and the self-assertive superlative maxima. The pompa may appear small, but it is actually great. We should not, therefore, translate maxima dona by ‘as all my offering’. These lines are a further example of the poem's central theme of the paradox of the ‘slender Muses’ (3) whose poetry will rival even the might of Homeric epic (37–8). There is, pace Camps ad loc., no irony in maxima, and of the passages which he cites as parallels only Prop. 2.9.40 is apposite since there the same thing is both great and small simultaneously, whereas at Prop. 4.1.34 and Ov. Fast. 6.263–4 the contrast is between past and present scale which is not nearly so paradoxical. Another example of the same movement of thought using Callimachean terms in Virgil is G. 4.6–7: in tenui labor; at tenuis non gloria, si quem/ numina laeva sinunt auditque vocatus Apollo. In tenui equivocates artfully between the smallness of the actual subjects, the bees (who have, by contrast, just been described magniloquently in anthropomorphic terms), and the stylistic sense of the word (which again contrasts with the ‘epic’ terms just used of the bees: magnanimos duces, populos, proelia). But the poet's inspiration will transmute this tenuis subject-matter and stylistic level into something great and glorious.

40. On otiosi see Segal, C. P., G&R n.s. 17 (1970) 25–31Google Scholar and on lusimus Fordyce ad loc.

41. GRBS 5 (1964) 196Google Scholar. The whole article is of great interest for my present theme. Clausen comes very close to identifying the Callimachean nature of Eclogue 1.

42. Martin (n. 2) 104, Cairns, Francis, CQ n.s. 21 (1971) 207Google Scholar, citing Thuc. 1.5.2, [Moschus] 3.82, Virg. E. 6.46, 62, Hor. S. 1.10.36, 2.5.41, Stat. Silv. 2.7.77.

43. Cf. Theoc. 4.4.

44. Cf. LS s.v. 1.2, and Coleman ad loc.

45. I do not mean that Virgil was unaware of the true significance of ᾶναλλα (although I believe there may have been some confusion in his view of [Theoc.] 9.3). Such an outrageous bilingual etymology is a highly sophisticated piece of wit at the expense of the doctrina which it itself manifests.

46. Other equally erroneous etymologies are to be found in the scholiastic tradition: cf. Schol. in Theoc. Vet. pp. 4.3–5, 10.5–26 Wendel.

47. The word errare may have suggested itself to Virgil through his awareness of the use of by Callimachus (Del. 4.176) and Theocritus (7.92) in the sense of ‘wander’ (without a flock). This interpretation is argued at length by Giangrande, Giuseppe (AQ 37 (1968) 508–11Google Scholar). He is probably right to accept Gow's interpretation of the choice of the word in Theocritus as being to express the specifically bucolic nature of the inspiration which Simichidas received. One should not, however, forget that a rustic locus and occupation is conventional in encounters between poets and prophets and the gods who inspire them (see p. 117).

48. Cf. Dover (n. 26) liv–v.

49. For a good survey of such initiation involving both poets and prophets and detailed interpretation of the Hesiodic vision, cf. West, M. L., Hesiod, Theogony (1966) 158–67Google Scholar. Also important is Cameron, Archibald, ‘The Form of the Thalysia’ (Miscellanea di Studi Alessandrini in memoria di Augusto Rostagni (1963) 291–307)Google Scholar.

50. SEG 15 (1958) n. 517. 11.20–52Google Scholar, cited hereafter as SEG + line number (cf. the literature cited in SEG and Kambylis, Athanasios, H 91 (1963) 129–50Google Scholar).

51. Vorsokr. I pp. 27–31Google Scholar.

52. Fr. 229 Pf. Branchus is not included in the lists in West (n. 49), but his experience clearly conforms to the type.

53. This was pointed out by Hanslik, (n. 2) 16–7 (cf. Ryberg, Inez Scott, TAPA 89 (1958) 117–8)Google Scholar. It is astonishing to find Hardie, Colin (The ancient historian and his materials, ed. Levick, Barbara (1975) 117)Google Scholar concocting an explanation based upon Tityrus' supposed position as an estate manager. Equally odd is the puzzlement of Putnam, M. J. C. (Virgil's pastoral art (1970) 44)Google Scholar at the use of the word puer when Tityrus is a senex. Puer was used of slaves without distinction of age. The phrase ut ante is more easily explained in terms of the surface narrative (e.g. Du Quesnay 135–6) but it is not beyond the bounds of imagination to suppose that Virgil is claiming ‘divine” approval for work already in progress. Ovid does precisely that, although in a jocular manner, at Fasti 4.15–6.

54. For a succinct account of the arguments, see Gordon Williams (n. 4) 311–2. (cf. Du Quesnay, 133–4). One may add the (colloquial?) use of deus to describe the great in Augustan Rome, e.g. Hor. S. 2.6.52. Du Quesnay (41–2) draws interesting parallels with Theocritus' praise of Ptolemy Philadelphus in Idylls 14 and 15. Note that in the passage which he regards as most striking (14.61–4) Ptolemy is described as φιλόμονσος

55. Cf. Moritz, L. A., CQ n.s. 18 (1968) 116–31CrossRefGoogle Scholar. Mr A. G. Lee points out to me that the references to Pollio and his espousal of nova carmina begin in the twenty-fifth of the forty-eight lines of the amoebaean contest in Eclogue 3. The arithmetical centre is not found so precisely, but the position and content of the lines suggest that here too we have a dedication expressed by mentioning the dedicatee in a central position.

56. On this formula and its application only to gods and rulers who are being treated as gods, cf. Levi, Peter, H 94 (1966) 76–8Google Scholar. His argument that E. 8.11 cannot therefore be a genuine reference by Virgil to Pollio is not conclusive (nor are any of his other arguments). A young poet writing in the Rome of about 40 B.C. which lacked a single god-like ruler might have used such an expression to Pollio then at the height of his success and might simply not have felt the need to remove it at a later date when the book was published and the position of Pollio was rather different. If, on the other hand, G. W. Bowersock (n. 5) is right in claiming that E. 8.6–13 are addressed to Octavian, the consequences are very interesting. The ‘first and last’ formula is then entirely appropriate. We have, moreover, a parallel to the relationship which I shall argue exists between the encounter with the deity in E. 1 and Apollo's instructions to Tityrus in E. 6. A standard motif is first presented in a covert fashion in E. 1 and then made explicit in the later poem, this later reference helping us to understand the earlier if we have missed or failed fully to appreciate it. The actual central position in E. 1 fulfills the ‘in the middle’ part of the formula, while E. 8 supplies ‘first and last’ in the normal way. One serious difficulty which I feel in accepting Bowersock's case and its consequences is that E. 8.6–13 do not seem to be worked into any of the systems of symmetrical correspondence in the book as a whole. In the light of my later observations on this subject, I find it hard to believe that such a major encomium of Octavian would have been left suspended in a place of so little significance.

57. For discussion and a survey of previous interpretations, cf. Wilkinson, L. P., The Georgics of Virgil (1969) 165–72, 323–4Google Scholar. Wilkinson is surely right to draw attention to the Pindaric imagery of the passage. But that does not preclude, as he seems to think, a simultaneous debt to Callimachus. Much earlier Greek literature was mediated to the Romans through the Alexandrians. The images of the way, chariot-driving, and envy were doubtless taken by Callimachus from the earlier Greek tradition (cf. e.g. apart from Pindar, Choeril. Fr. 1), but his particular formulations became canonical. One can almost assert a principle in these matters of post Callimachum ergo propter Callimachum (cf. Clausen (n. 41) quoted on p. 114). This is not to deny that Virgil may have been aware of these images as Pindaric but to assert that he cannot have failed to be aware of their use by Callimachus. I did not have sight of the excellent article by Thomas, Richard F. ‘Callimachus, the Victoria Berenices, and Roman poetry’ (CQ n.s. 33 (1983) 92–113)CrossRefGoogle Scholar until my work was completed. His discussion of the proem to Georgic 3 (a) offers full arguments for the position taken in this note, and (b) demonstrates a detailed exploitation by Virgil in a programmatic context of material from the Aitia, which enhances the plausibility of my claims for the use of similar techniques by the poet in an earlier work.

58. This is not the place to offer an interpretation of E. 6, which would require another work as long as the present. To substantiate the description given, I simply draw attention to a few salient points. It opens with Tityrus-Virgil's inspiration by Apollo in Callimachean style. It contains Gallus' inspiration by the Muses and Linus on behalf of Apollo in Hesiodic style with an encomiastic reference to Gallus' (Callimachean?) aetiological poem on the Grynean grove. A lively and very slightly risqué narrative introduces a song in which, whatever the rationale of the arrangement and selection of material, a great variety of poetic themes and stylistic treatments are artfully fitted together. It is a virtuoso piece and Apollo is to be found at its beginning and end as well as in the midst, if not the middle, of the song.

59. Silv. 1.4.19–37.

60. Hanslik, (n. 2) 18, identifies Octavian with Apollo making the additional point that in Hesiod the Muses are doing the work of Apollo in giving him instructions and so the source of the instructions in Virgil must also be Apolline. On Octavian's Apolline propaganda, cf. Lambrechts, P., La Nouvelle Clio 4/5 (1953) 65–82Google Scholar, Weinstock, Stefan, Divus Julius (1971) 12–15Google Scholar, esp. 15. Clausen (n. 5) 202–3, reminds us of Wissowa's observation on E. 1.44 that monthly celebration of the birthday of the ruler was a feature of Hellenistic ruler worship. See Du Quesnay (101–12) for a well-documented account of deification as a topos of the eucharistikon.

61. Suet., Aug. 70Google Scholar.

62. Hanslik (n. 2) 17, is right to say that there is no question of temporal sequence in the use of the word (i.e. the god is not the first of several persons to respond to Tityrus) and to emphasise the debt to Hesiod. Interpretations such as Hardie's (n. 53) 117, where the iuvenis is first of a number of masters (praesentis divas [41]), simultaneously ignore the conventions of rhetoric (see the next note) and of poetic symbolism. I cannot, however, share Hanslik's view that no real weight is to be attached to primus.

63. Du Quesnay 98, 134–5. For words denoting priority of action applied to a benefactor who is being thanked cf. e.g. Cic., Red. Sen. 8Google Scholar (princeps), 24 (princeps, primus), Red Pop. 16 (unus), Marc. 11–2 (unus), 23 (unus) – unus is, of course, a fortiori equivalent to primus. One may also note Stat. Silv. 4.2 where the priority words appear in order to emphasise the poet's gratitude but are transferred from the benefactor to his action or its results: primum (6) indicates that this is the first time Statius has been invited to dine with Domitian; haec aevi mihi prima dies, hic limina vitae (13) claims that the effect of the act has been to mark a new beginning of life for the poet. The aim in both places is the intensification of the expression of gratitude not the clarification of temporal questions. I return to this poem later. Ovid characteristically exploits both temporal and eucharistic senses at Pont. 3.3.29 in a programmatic context.

64. It is, as it were, a literary specialisation of the πρῶτος εύρετής concept which was at least as old as Aeschylus (cf. Kleingünther, Adolf, Phil. Suppl. 26 Heft 1 [1933])Google Scholar. The example of Epicurus at Lucr. 5.8 9 illustrates well the possible overlap of benefactor and innovator. Epicurus is being thanked in appropriate terms for the benefits which he has conferred on mankind, but these benefits have been made possible by his intellectual innovation. Many additional examples including the Callimachean variations can be found by consulting Wimmel, Walter, Kallimachos in Rom (Hermes Einzelschriften 16, 1960)Google Scholar index s.v. primus-Motiv.

65. 1.66 (Epicurus), 117 (Ennius), 926–30 (Lucretius), 3.2.9 (Epicurus), 5.9 (Epicurus).

66. Catul, 1.5: the praise of Nepos' work is implied praise of his own; unus (=primus) and ausus es (see below, p. 121) indicate the claim to priority. Hor. Carm. 3.30.13–4. Ep. 1.19.21–4. Prop. 3.1.3–4. Man. 1.4–6. Hor. Serm. 2.1.62–3.

67. 2.174–5 (Georgics), 3.8–15 (projected epic), 40–1 (Georgics), 4.565 (Eclogues). The first and last of these passages indicate the claim to priority (as Catullus did in part) by the use of the words audere/audax, the literary use of which I explain in n. 69.

68. There is doubtless an element of the eucharistikon present, but it is only one feature of the context. Ov. Met. 5.341–4 is a clearer case of gratitude to the innovative deity; cf. Lucr. 5.14–5, Virg. G. 1.147, Tib. 1.7.29–36, Ov. Am. 3.10.11–14, Fast. 4.401–2.

69. Enn. Ann. 217, Lucr. 1.67, Catul. 1.5, Virg. G. 1.40,2.175, Hor. Serm. 2.1.62–3, Prop. 2.10.5–6, cf. Hor. Ars 125–6, 286–7, Man. 3.1–3. In this specific sense of innovation the term has a Greek forerunner in the use of τολμᾶς at Call. Fr. 203, 19 Pf. where the reference is to his rashness in writing in many metres unlike other poets. The word is put in the mouth of an imagined opponent and so is presumably intended to be denigratory whereas in the Roman poets the sense is always one of boldness as a positive virtue. For less specific uses in literary contexts cf. Brink on Hor. Ars 10. The dangers of literal-minded historical or biographical readings of these poems are again illustrated by Hardie (n. 53), who takes audaxque iuventa to mean that Virgil displayed youthful boldness in his overt attacks on Octavian in Eclogues 1 and 9. (The word audax at Hor. Ep. 2.2.51–2 is equally devoid of political implications.) It is worth observing that such violence to the meaning of the text is provoked by a view that the poems are attacks upon Octavian, a view which can hardly survive a reading of them unprejudiced by the false beliefs about their date and Virgil's personal involvement in the expropriations derived from the scholiastic tradition.

70. Williams (n. 4) 328–9. The uniqueness of the circumstances holds given a reasonably tight definition of ‘contemporary’. On a rather broader scale of both time and genre the dedication is seen not to be compulsory (cf. White, Peter, JRS 64 (1974) 55Google Scholar; pp. 50–61 of that article give a thorough account of ancient book-dedications which illustrates the great variety of methods employed in addition to the direct address by name). Since, however, I detect a dedication in E. 1, I do not wish to pursue the point.

71. Similarly Cicero's Brutus and Academica are dedicated to Brutus and Varro respectively by their being introduced as principal characters whose appearance sparks off the conversation reported in the work (White (n. 70) 52) and not by any formal address. Even when there is no formal difficulty in addressing the dedicatee, he may be mentioned only in the third person or simply alluded to. The absence of a vocative or second person address does not make Paulus Maximus any the less the dedicatee of Horace's fourth book of Odes (Carm. 4.1.9–11). I would even suggest that Odes 3 is directed towards Maecenas by the concluding reference in the opening poem to Horace's Sabine farm (Carm. 3.1.47) which is made to symbolise his preference for peace with little (which in turn symbolises his poetic calling) as opposed to wealth with trouble. The encomiastic purpose is sufficiently served by the presence of the person at the beginning of the work. In exactly the same way there may be no effective difference between the use of Du-Stil and Er-Stil in religious utterances (cf. Norden, Eduard, Agnostos Theos ed. 4 (1956) 143–66Google Scholar). Nor does the postponement until line 42 of the word iuvenis which identifies the god tell against the view. Tullus has to wait nine lines to hear his name in Prop. 1.1 and Messalla more than fifty (two-thirds of the poem) in Tib. 1.1.

72. Cf. Fraenkel, E., Horace (1957) 242–6Google Scholar, Nisbet, R. G. M. and Hubbard, M., A commentary on Horace: Odes I (1970) 16–7Google Scholar.

73. Not to mention an explicit imitation with variation of the Hesiodic encounter at 6.64–73. This presumably corresponds to the situation in Callimachus where the visitation of Apollo (Fr. 1.21–28 Pf.) is separate from the dream of meeting the Muses (Schol. Flor. ad Fr. 2 Pf.). Such a correspondence may lend support to the correspondence for which I later argue between the double encounter with ‘Apollo’ in Virgil and the double ‘divine’ encounter in Theoc. 7.

74. Fraenkel (n. 72) 230.

75. Cf. Posch, Sebastian, Beobachtungen zur Theokritnachwirkung bei Vergil (1969) 17Google Scholar. The lists of reminiscences provided by Posch are extremely useful but not every reader will agree that he has assembled all or only the true reminiscences (e.g. Theocr. 1.1 / E. 1.1 is missing and I cannot detect any clear reminiscence in Theocr. 1.20/ E. 1.2).

76. Cf. n. 8.

77. Cf. Alpers (n. 3) 136.

78. In what follows I am indebted to the articles on Idyll 7 by Giangrande, Giuseppe (AQ 37 (1968) 491–533Google Scholar) and Williams, Frederick (CQ n.s. 21 (1971) 137–45CrossRefGoogle Scholar).

79. Bibliography in Williams (n. 78) 137, n. 3.

80. Cf. Giangrande (n. 78) 508–11.

81. Cf. Williams (n. 78) 144. One may add to his examples of the amusement of the unrecognised god the highly relevant case of the women who are encountered by Archilochus, (SEG, 29–32Google Scholar): [Archilochus] .

82. Lettres d'Humanité 3 (1944) 7l–147Google Scholar. For a survey of schemes which attempt to establish the existence of deliberately planned responsions and correspondences in the arrangement of the book see Rudd, Niall, Lines of Enquiry (1975) 119–44Google Scholar. While this offers, as intended, an interesting study of methodology its disappointingly negative conclusions result largely from a concentration upon schemes which are no longer seriously believed by anyone. Again, the Maury scheme is discredited by refuting the over-enthusiastic support for it of Otis, Brooks (Virgil (1964) 128–43)Google Scholar. It may well be that Otis goes too far in his attempt to provide a convincing explanation for the scheme in terms of the detailed treatment of subject-matter and an overall contrast between the first and second halves. This may be too subjective, but that does not diminish the credibility of more objective elements such as the general subject-matter of particular poems (e.g. the expropriations of land in 1 and 9), or their form (e.g. 3 and 7 amoebaean). Rudd does not consider the additional scheme which I am about to put forward.

83. The exactness of this equivalence can be illustrated from a non-symbolic context in Pompon, com. 57-9 = Macr. 6.4.13: vocem deducas oportet ut mulieris videantur / verba. – iube modo adferatur munus, ego vocem reddam / tenuem et tinnulam.

84. These connections are clearly stated by Klingner, Friedrich (H 62 (1927) 147–52Google Scholar, cf. now Virgil (1967) 27–32Google Scholar): cf. Liegle, Josef, H 78 (1943) 216Google Scholar and Anm. 6, Bömer, F., Würz. Jahrb. 4 (1949–1950) 61–2Google Scholar.

85. Cf. Theoc. 15.47, 17.16–19, Call. Del. 165–70 (where reference is made to the Ptolamaic title of Soter and to the way in which Ptolemy Philadelphus will reproduce the virtues of his father – just like Octavian and Julius Caesar). Cf. e.g. Virg. G. 1.24–5, Hor. Carm. 1.2.45–9 with Nisbet-Hubbard ad loc.

86. Cf. Alpers (n. 3) 190–1 for the emphasis in recent work on the poetic significance of Daphnis in Eclogue 5.

87. ZPE 25 (1977) 49 50Google Scholar. My own views on the structure of the Eclogues predate that article. Cf. now Thomas (n. 57).

88. Frr. 1–7.19, 112 Pf.

89. Aitia 3, Frr. 67 75 Pf. On the fame of this poem, cf. Pfeiffer ad Dieg. Z 1. Professor Cairns first (!) directed my attention to the possibilities explored here. Cf. also n. 18.

90. Cf. Cairns, Francis, CR n.s. 19 (1969) 131–4Google Scholar.

91. Φθόνος, Ap. 105–7; ασκανίη Fr. 1.17 Pf., Ep. 21.4.

92. E.g. Hor. S. 2.1.77, Carm. 4.3.16, Prop. 3.1.21, Ov. Am. 1.15.1, cf. Virg. G. 3.37.

93. Du Quesnay (69) offers additional, but not incompatible, explanations of miror in generic terms. On the critical vocabulary of Prop. 1.7 in general cf. Quadlbauer, Franz, H 98 (1970) 331–9Google Scholar. Professor Cairns drew my attention to the relevance of this poem.

94. The simile is, of course, again adapted by Virgil with great subtlety and applied to Dido (Aen. 1.498–502).

95. For the evidence, cf. Weinstock (n. 60) 93–9.

96. Suet. Nero 46.1, Plin. Nat. 36.41.

97. E.g. Liv. 1.16.7, 45.3, Ov. Am. 1.15.26.

98. Pace Dick, Bernard F., AJP 91 (1970) 285–7Google Scholar. There may be some sense in libertas of ‘freedom to compose’ (cf. n. 31) but I feel that the word relates mainly to the surface narrative and its function is to get Tityrus to Rome where he can meet the god. This would be legally credible on the basis of his seeking a magistrate with imperium who could manumit him under the procedure for manumissio vindicta (cf. Treggiari, Susan, Roman freedmen during the Late Republic (1969) 21–5Google Scholar). There is, of course, no mention of Tityrus' master who would have had to be present also and the god's response is quite unrelated to manumission. Libertas is a temporary signpost in the narrative whose importance is exhausted when it has served its immediate purpose of motivating the journey. An alternative and possibly more plausible account of the necessity of Tityrus' journey is provided by Du Quesnay (115 - 127). It is a remarkable piece of scholarly ingenuity and (in the best sense) imagination, but seems to me to go beyond what even a Roman reader could infer from the text. The conflicting explanations, all of them possessing some degree of implausibility, advanced over the years by competent scholars raise the question of whether the text implies a fully worked out legal situation.

99. For apples in love stories, cf. Acontius and Cydippe and Milanion and Atalanta; also e.g. Theoc. 3.10, Catul. 65.19, Virg. E. 3.64, 71, Prop. 1.3.24, Petron. Fr. 33.1.

100. Cf. Kambylis (n. 50) 134–5. The evidence is as follows: Hesiod – Th. 22 ποθ', assuming this refers to an occasion at some distance in the historical past (cf. West ad loc.); Archilochus – SEG 22; Epimenides – Vors. 1. pp. 27.23 - 28.1, assuming that being sent to look after the sheep by one's father is evidence of youth; Callimachus Fr. 1.21 Pf. , 37 παῖδας, cf. Schol. Flor. 18 with Pfeiffer ad. loc. Cf. also the stories about Pindar (Paus. 9.23.2) and Aeschylus (Paus. 1.21.2) and the adoption of the motif (doubtless with at least an element of autobiographical veracity) by Horace (Carm. 3.4.9 – 20, cf. 4.3.1 – 12), Ovid (Tr. 4.10.57–8), and Quintus of Smyrna (12.309–10; one wonders if he is deliberately going one better than Callimachus by receiving inspiration rather than , although Hom. Od. 11.319–20 may be the decisive factor).

101. Cf. Berg (n. 3) 191.

102. Du Quesnay's attempt (122) to deny that 33–5 are connected with 31–2 is incomprehensible to me. Had he included in his citation of Conington's note on 35 the preceding sentence, the logical incoherence of the position adopted would have been manifest. Tityrus is slightly illogical in blaming the town for the unfortunate results of his own infatuation with Galatea, but such behaviour is both psychologically plausible and may be justified by the programmatic implications indicated.

103. It is also fundamental to the humorous purpose detected by Giangrande (n. 78) in Theoc. 7 where much of the point depends upon the exposure of Simichidas the pseudocountryman by Lycidas and the poet.

104. Martin, (n. 2) 104, interprets non insueta pabula (50) as ‘customary art’ and mala vicini pecoris contagia (50) as ‘rivalry of the envious’. Berg, (n. 3) 150–3, carries this a good deal further with particular reference to Callimachus. As well as making additional points, I have to traverse this ground again so as not to leave my argument incomplete.

105. P. 113.

106. Hor. S. 1.4.11, 10.50, Call. Ap. 108–9.

107. For Callimachus, cf. p. 130, and n. 91; Prop. 3.1.21; cf. n. 92.

108. FGrH 1, p. 201Google Scholar, 3–22, cited as Conon 33.

109. Pfeiffer, II, 120, Barber, E. A. and Maas, Paul, CQ 44 (1950) 168Google Scholar, Trypanis, C. A., Callimachus: Aetia, etc. (1958) 170–1Google Scholar.

110. The staff and wreath are, of course, aetiological elements in the story, explaining the origins of the crown of laurel and wand which were used in the ritual of divination. Cf. Fontenrose, J. E., TAPA 64 (1933) 105Google Scholar. The whole article is of interest for the background to Callimachus' poem. Fontenrose suggests (103 n. 20) that Varro and Conon may have had different sources, one Callimachus and one Ephorus, as an explanation of any discrepancies between them (e.g. Apollo kisses Branchus in Conon, vice-versa in Varro).

111. One may add the observation that the encounter was clearly envisaged as taking place in Branchus' youth (cf. e.g. Conon, 33.4).

112. On the whole question of the plague and Branchus as healer cf. Pfeiffer ad loc. Pfeiffer's observation ‘at in Branchi fabula Milesia de hac re adhuc nihil notum est’ is entirely consistent with the Barber–Maas–Trypanis explanation. There is nothing in the story about animals being struck by a plague because what Apollo is giving is a pledge that they will not be struck. In this case his medical powers are put to preventive rather than curative use. Thus the questions of the animal plague and Branchus' subsequent healing powers (for which after a divine encounter cf. Epimenides, Vors. 1, p. 28.8–15Google Scholar), are quite separate and distinct. It is worth observing in anticipation of the parallel I am going to draw with Octavian/ Apollo in Virgil that Apollo's medical function was to the fore in his earliest appearances in Rome (cf. Gagé, Jean Apollon Romain (1955) 19–83Google Scholar).

113. . For the Arcadian connections of this cult title, which would render any allusion to the functions it denoted highly apposite in the Eclogues, cf. Cic. Nat. Deor. 3.57 with Pease ad loc.

114. For this concept in the Eclogues cf. Griffiths, Carlotta, PVS 9 (1969–1970) 5–10Google Scholar.

115. Call., Ep. 27Google Scholar.

116. Giangrande (n. 78) 491–2.

117. As observed earlier some of the points I am about to make were first mentioned by Martin (n. 2) 103. The argument has been developed further by Berg (n. 3) 150–3.

118. For much information on the traditional connection between bees and poets, cf. Berg (n. 3) 207 n. 60. He does not, however, mention the largest collection of such information, Robert-Tornow, W., De apium mellisque apud veteres significatione et symbolica et mythologica (1893)Google Scholar.

119. For Pindar cf. Paus. 9.23.2 (note the characteristic features of the divine encounter: Pindar is young, he is on a journey, the time is midday); other versions in Ael. VH 12.45, D.Chr. 64.23, Philostr. Im. 2.12. The popularity of such tales can be seen by their ascription in later times to Virgil (Vita Focae, 28-30) and Lucan (Vita Vaccae). Bees and honey are also, of course, closely connected with Apollo in both his prophetic and poetic functions, cf. Robert-Tornow (n. 118) 116. Theocritus relates such a story about a goatherd singer, Comatas, at 7.78–85. Honey is also the appropriate reward of the sweet-tongued singer at Theoc. 1.146–7.

120. Cf. Wellesley, Kenneth, CP 63 (1968) 140Google Scholar. Note further that Virgil seems fond of using rapidus in an etymologising way, e.g. rapidus vorat aequore vortex (A. 1.117); cf. Bartelink, G. J. M., Etymologisierung ij Vergilius (1965) 29Google Scholar.

121. For Alexandrian interest in rivers, cf. Callimachus' work (Frr. 457–9 Pf.), and Philostephanus of Cyrene's (Frr. 20–6 Müll.).

122. Possibly not priestesses, cf. Pfeiffer, Rudolf, History of classical scholarship from the beginning to the end of the Hellenistic age (1968) 125–6, 284Google Scholar. On this question and the passage as a whole, see the Commentary of Frederick Williams (1978).

123. S. 1.4.11, 10.50, cf. Sen., Con. 4Google Scholar, pref. 11. The figurative use of the muddy river in S. 1.1.54–60 is particularly interesting in the present context. That poem, like Eclogue 1, contains an implicit programme which is made explicit later in the book. The example of the river is set in an ethical discussion, but its use of the imagery of Callimachean poetics makes a literary point. The same thing can be more easily seen at the end of the poem, where the comparison of that rare occurrence, the contented man, to a conviva satur (with an implied etymology of satura, one of the names of the genre) leads into the Callimachean adherence of iam satis est and verbum non amplius addam (117–21).

124. Rh M. 99 (1956) 198–9Google Scholar.

125. Cf. Prop. 3.3.31 et Veneris dominae volucres, mea turba, columbae.

126. On the doctrina of these lines, see Du Quesnay 137–8.

127. I start from the position that, while the Eclogues are not in a full sense ‘Augustan’ poems, there is a continuity in thought and expression between them and Virgil's later works.

128. For the importance in that passage of ideas going back to Pindar and Callimachus, cf. n. 57. A preoccupation with the poet's task has been identified as an element in almost all of the surviving epinician odes of Pindar (Hamilton, Richard, Epinikion: general form in the Odes of Pindar (1974Google Scholar) index s.v. Poet's Task; cf. CR n.s. 27 (1977) 5). Its presence in the Eclogues other than 1 can be detected for certain at the following places: 2.31–9, 3.60–3, 84–7, 4.1–3, 53–9, 5.45–55, 81–4, 6.1–12, 64–73, 7.18–28, 8.1 13,9. passim, 10. passim. Conversely, Thomas (n. 57) emphasises the epinician nature of much of the imagery of the proem to Georgic 3.

129. The open, if obscure, encomium of Eclogue 4 is a special case and its extension of the conventions of the genre is signalled in the opening lines (1–3).

130. Wimmel (n. 64) 221–50.

131. Wimmel (n. 64) 216–8. It is interesting to note from that discussion that Propertius seems to have treated the end of Georgics 2 and the beginning of Georgics 3 as a single unit for the purposes of imitation. This lends support to my own adoption of the same viewpoint in analysing the similarities between this section of the Georgics and Eclogue 1.

132. Ecl. 4. 168–9. Raoul Verdière's attempts to explain this away are quite beside the poetical point (T. Calpurnii Siculi De Laude Pisonis, etc. (1954) 255Google Scholar).

133. For the interpretation by a Silver Latin poet of Augustan poems see Hardie, Alex, Statius and the Silvae (1983) 152–71, esp. 156–64Google Scholar on Silv. 3.2 and Hor. Carm. 1.3 (cf. Ov. Am. 2.11, Hor. Epod. 1).

134. There may be an elegant piece of variatio on his model in 41–2 (cf. E. 1.65). The locale of Corydon's exile would have been by a silt-bearing river in the far west as opposed to the far east. The conscious doctrina of the reference is guaranteed by the use of dicitur (cf. Fordyce ad Catul. 64, If.). It is not possible to say for certain whether Calpurnius recognises and exploits a literary significance in Virgil.

135. N. 63.

136. Elsewhere in the poem Statius plays the epic poet to the Emperor's hero (Hardie (n. 133) 6).

137. Professor A. J. Woodman drew my attention to this aspect of Tityrus' description of the deus. See Du Quesnay 133.

138. Latin satire: the structure of persuasion (1970) 133–4Google Scholar.