Hostname: page-component-6766d58669-88psn Total loading time: 0 Render date: 2026-05-17T03:18:07.669Z Has data issue: false hasContentIssue false
  • English
  • Français

EMENDATIONS IN THE DIRAE AND THE LYDIA

Published online by Cambridge University Press:  14 April 2023

Boris Kayachev
Boris Kayachev*
Affiliation:
Faculty of Classics, University of Oxford
*
boriskayachev@gmail.com
Rights & Permissions [Opens in a new window]

Abstract

This article argues that the text of the Dirae and the Lydia is even more corrupt than current editions give reason to believe, and attempts to emend about a dozen passages.

Keywords

Latin poetrybucolic poetryAppendix VergilianaDiraeLydiatextual criticism

Information

Type
Research Article
Information
The Classical Quarterly , Volume 72 , Issue 2 , December 2022 , pp. 703 - 718
Check for updates
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
Copyright
Copyright © The Author(s), 2023. Published by Cambridge University Press on behalf of The Classical Association

The Dirae and the Lydia are among the most corrupt poems in the Appendix Vergiliana, and at the same time among the least favoured by commentators.Footnote 1 The only line-by-line commentary published in the twentieth century is to be found in van der Graaf's 1945 monograph; yet while the book offers a number of useful essays on the metre and style of the two poems, the commentary itself lacks detail and is extremely conservative in its treatment of textual issues.Footnote 2 Subsequent publications usually address individual cruces, but do not scrutinize the text in a more systematic manner.Footnote 3 As a result, while a consensus exists that the poems are full of severely—sometimes irreparably—corrupt passages, the remaining superficially intact portions of the text tend to be accepted without questioning. To give just one example, all modern editions print Lydia 112 in its transmitted form o fortunati nimium multumque beati, despite the fact that multum after nimium is culpably anticlimactic and even though the correction nimiumque has been available since the Renaissance.Footnote 4 It may appear unlikely that a scribe could miscopy a word that he had just copied correctly, but we should bear in mind that nimium and multum, having the exact same number of strokes, can in certain hands look almost indistinguishable.Footnote 5 While the lack of critical attention to the textual problems of the Dirae and the Lydia can account to a large extent for such oversights, the situation is aggravated by the fact that the poems do not have a clear narrative frame that could provide either scribes or editors with some sort of control over the text. This can most clearly be seen in the case of the refrains, which both in sense and in syntax are detached from their immediate context and, as some of the following discussions attempt to show, are thus especially prone to be corrupted by scribes (and far less so to be corrected by editors). Without aiming at a comprehensive treatment, in what follows I mainly focus on identifying and solving problems that have been ignored in recent scholarship, while also hoping that the discussions will cumulatively have some weight in demonstrating that the text of the Dirae and the Lydia is even less certain than the available editions might let us believe.

DIRAE 14

In the first stanza of the curses proper (9–13) the speaker execrates his fields, pastures, orchards, vineyards, woods and streams; the second stanza (15–18) is a close repetition, with only slight variations, and it is introduced by a refrain that makes this explicit (14):Footnote 6

rursus et hoc iterum repetamus, Battare, carmen.

A long-acknowledged problem is the apparent superfluity of either rursus or iterum at 14. In principle, ‘anew and again’ is not as such an impossible turn of phrase, and the majority of scholars seem to understand the line in this way. The intervening et hoc, however, makes such a construal rather strained: thus positioned (rursus et hoc iterum), et would normally go either with hoc (‘and this also’) or possibly with the clause as a whole (‘and let us repeat the song, anew again’), but neither option seems to make good sense here. Naeke ascribed to rursus an imperative force (‘once more! and let us repeat the song again’), but this is unwarranted.Footnote 7 While the transmitted text may not be entirely unacceptable, it seems more likely that rursus et is an error for tristius, probably initiated by the omission of rubricated T; rıſtıuſ would then be read as rurſuſ, with et added subsequently to restore the metre.Footnote 8 Thus emended (tristius hoc iterum repetamus), the refrain will become more similar to 54 (= 71) tristius hoc, memini, reuocasti, Battare, carmen Footnote 9 and 75 tristius hoc rursum dicit mea fistula carmen. We may note that in all these cases tristius occurs in a refrain introducing the repetition of the previous stanza, and arguably this second stanza can be seen to be phrased in more forceful terms; accordingly, tristius should be taken predicatively: ‘let us repeat this curse more grimly’.

DIRAE 24

The speaker aims lines 20–4 specifically at his grasslands:

haec Veneris uario florentia serta decore,
purpureo campos quae pingunt uerna colore
(hinc aurae dulces, hinc suauis spiritus agri),
mutent pestiferos aestus et taetra uenena;
dulcia non oculis, non auribus ulla ferantur.

There are some uncertainties throughout the passage,Footnote 10 but the general sense seems clear: the plants growing on the meadow should no more produce fragrant flowers but should rather diffuse poisonous substances. Line 24 is intended to sum this up: as van der Graaf translates, ‘may nothing sweet ever reach eyes and ears’.Footnote 11 There are two potential difficulties here. First, until the present line the passage does not refer to sound, only to colour and smell: hence Heinsius's naribus for auribus, which editors mention even if they rarely accept it.Footnote 12 Second, as was likewise felt by Heinsius, the verb too may be inappropriate. As transmitted, the line can be construed in two ways: the datives oculis and auribus may depend on either the verb ferantur (literally ‘may move against the eyes and ears’) or the adjective dulcia (‘pleasant to the eyes and ears’). Neither construal, however, is convincing. The former requires for ferri a meaning that, as far as I can see, is unparalleled (‘to present itself to’). Taking the datives with the adjective is altogether preferable, but, again, interpreting the verb presents difficulties. One option is to take ferri to mean something like ‘to be borne through the air’: this might work for sounds (if reading auribus) or smells (if reading naribus), but hardly for what one sees, unless we imagine Lucretian simulacra. The alternative is to understand ferri as a synonym of gigni, and, as I suggest below, this is indeed the sense we need. However, while there is no lack of contexts of the type of Verg. Ecl. 4.38 feret omnia tellus, I am unable to parallel the reverse passive construction with ferri meaning ‘to be produced’ (with or without the ablative of agent); the idiom seems to be confined to the active voice.Footnote 13 What can ferantur have replaced? Heinsius suggested two possible corrections: serantur, which is easy palaeographically but less than ideal in sense (to begin with, meadows do not need to be sown), and creentur, which gives a reasonable sense but fails to account for ferantur.Footnote 14 I propose genantur: ‘may there grow nothing pleasant’.Footnote 15 Incidentally, this will also favour Heinsius's naribus (plants can be pleasant to smell, but far less ordinarily so to listen to).Footnote 16 The corruption will have begun with a scribe, unfamiliar with the rare form, changing genantur to gerantur (as at Lucr. 4.143 and 159), which would then mistakenly be corrected to ferantur.Footnote 17

DIRAE 25 (= 47)

Above we discussed a refrain which announces that the following stanza is a repetition of the preceding one (14); now we come to consider a different refrain which, by contrast, signals a change of topic (it occurs twice, at 25 and 47):

sic precor, et nostris superent haec carmina uotis.

The line does not yield immediate sense. Naeke considers two options, but both are absurd: ‘let this song outlast my prayer’ (‘praestent efficaciam haec mea carmina tum quoque, quum mecum mea vota mortua erunt’) and ‘let this song exceed my prayer’ (‘Optat poeta ut plus praestet hoc carmen suum, quam ipse dicendo vovere possit’).Footnote 18 Besides, while the former may be possible grammatically (OLD s.v. supero 7), the latter would require a different syntax (superare meaning ‘to surpass’ takes a direct object). Heinsius conjectured nostris superentur carmina uotis and nostri superent haec carmina uoti, but I fail to see how either is a plausible improvement.Footnote 19 Mähly thus seems to be right that the problem lies in the verb, even if his own proposal, spirent, is hardly less objectionable than the paradosis.Footnote 20 Giardina's subeant is, I believe, a move in the right direction, as it finally makes sense of the juxtaposition of carmina and uotis: sic precor and nostris … uotis refer back to what precedes (note that both 24 dulcia non oculis, non naribus ulla genantur and 46 qua nostri fines olim, cinis omnia fiat are wishes), whereas haec carmina points forward to what follows (‘così io prego, e in aggiunta ai miei voti vengano questi versi, questa mia poesia’).Footnote 21 In both cases the refrain in fact signals a change of theme: from meadows to woods at 25, and from a fire to a flood at 47. I suggest that Giardina's proposal can be improved upon by writing suberunt, which is slightly closer to the paradosis (for either corruption, compare the variant superant for suberant at Ov. Met. 10.593).Footnote 22 More importantly, the future indicative would be more suitable to introduce a new stanza (compare perhaps 30 resonabit) than the jussive subjunctive (for hortative subjunctive, compare 1 and 14 repetamus, 2 canamus, 97 reuocemus, but jussive subjunctive in reference to the speaker's performance is unparalleled in the Dirae).Footnote 23

DIRAE 79

The speaker invites heavy rainfall to flood the estate (79–81):

cum delapsa meos agros peruenerit unda,
piscetur nostris in finibus aduena arator,
aduena, ciuili qui semper crimine creuit.
79 cum delapsa Reitzenstein: unde elapsa codd.

Line 79 delapsa … unda probably implies rainwater running down the mountains (compare 76) rather than falling down from the sky, and this could account for the point of peruenerit: ‘when the stream originating in the mountains has reached the valley’. The resulting picture, however, faces two objections. First, it presents the natural disaster destroying the speaker's estate as an evolving process, but this is not how the poet seems to imagine it elsewhere.Footnote 24 Second, if we do conceive of the flood in such dynamic terms, the transition between 79 and 80 appears somewhat abrupt: when the stream reaches the fields, it has yet to submerge them, before anyone could fish there. Both these points are given additional weight by the use of the prefixed form, which lays emphasis on the process accomplished rather than the resulting state achieved. While I admit that these objections may not make peruenerit quite impossible, it seems worth voicing an alternative: perfuderit, which finds a parallel in 51 spissa campos perfundat harena. The error will have been due to perfuderit being written right over aduena in the next line, resulting in the scribe's inadvertently replacing -fud- in perfuderit with -uen- from aduena.Footnote 25 The one potential objection could be that the same verb, with a different prefix, has already been used in this stanza at 76 (teneant diffuso gurgite campos), but the poet does not in fact avoid such repetition (in addition to 51 and 76, note 43 diffusis ignibus, 49 auras diffunditis agris, 63 tuas infundimus aures, 68 diffundite campis).

DIRAE 66

It has long been recognized that 66 is not at home in its transmitted context, and it has plausibly been suggested that the line originally belonged to the passage concluding the curses proper (82–5), though the exact place has been disputed (see below):

o male deuoti, praetorum crimina, agelli,
tuque inimica tui semper Discordia ciuis,
exsul ego indemnatus egens mea rura reliqui,
miles ut accipiat funesti praemia belli?
82 praetorum cod. Esc. T II 9: pratorum SL: parcarum M: raptorum Scaliger

Yet before we turn to reconsider the original position of 66, I propose to discuss briefly a well-known problem that affects the middle of 82. One branch of the manuscript tradition offers impossible pratorum (SL), which a manuscript descending from the same branch corrects to praetorum. This has been a popular solution, adopted among others by Kenney, but a crucial obstacle is the fact that praetors had nothing to do with the settling of the veterans.Footnote 26 Courtney has defended Parcarum of the other branch of the tradition (M), by adducing Stat. Achil. 2.45 indecores, fatorum crimina, cultus and Dracontius, Orest. 453 iners, Parcarum crimina, pastor.Footnote 27 In both these cases, however, there is a point in blaming either Fate (Achilles was hidden on Scyros to avoid his destined death at Troy) or the Parcae (Aegisthus was born of incest), whereas in our passage such a reference could only convey a general disappointment at how things have turned out. Ellis cites prauorum from the 1482 Zarotti edition, but it is too vague a term; Heinsius proposed praedonum, which is more specific but arguably too derogatory.Footnote 28 Scaliger's raptorum is, I suggest, a somewhat easier correction and more to the point.Footnote 29 There is a category of curse tablets known as ‘prayers for justice’, whose characteristic feature is that, rather than wantonly cursing their victims, they ask for divine retribution in quasi-legal terms.Footnote 30 A common subgroup are tablets that ‘devote’ stolen objects to deities, who are asked, as their new owners, not to let the thieves benefit from their illegally acquired property.Footnote 31 This is the logic of 82: the speaker has ‘devoted’ (deuoti) his farmlands (to the underworld, as we shall see), because they have been unjustly taken away from him (raptorum crimina). This analysis of 82 can help us reconsider the long-standing problem of the original position of 66. It appears certain that 66 nil est quod perdam ulterius: merita omnia ditis (Kenney's text) does not belong in its transmitted place (on the one hand, it intrudes into a coherent passage; on the other, it has the air of a closural statement), but simply deleting it, as Kenney does, may not be the optimal solution.Footnote 32 First, a few words are in order on the text of the line itself. Since he obelizes it, Kenney prints it as it appears in the manuscripts, but merita is difficult to construe; an easy and compelling correction is Putsche's merito.Footnote 33 It finds support in an interesting bilingual curse-tablet from Spain (first century b.c. – first century a.d.), which reads in part: τοῖς κατὰ ᾍδην δίδωμι παραδίδωμι Νεικίαν καὶ Τειμὴν καὶ τοὺς ἄ[λ]λους οἷς δικαίως κατηρασάμην (recto) – deuotos defixos inferis Timen et Niciam et ceteros quos merito deuoui supr[a (verso).Footnote 34 The point of merito (δικαίως) is that the inscriber has suffered injustice from Nicias and Time and now hands them over (παραδίδωμι) to the gods of the underworld as judges: the curse is not a wanton attempt to hurt but a request for just revenge. In our case it is not the perpetrators who are ‘devoted’ to the underworld, but what they unjustly took: the speaker cedes his property rights on the estate to Dis, who should now claim it as his own. This reading is also supported by perdam, a verb that is often used in ‘prayers for justice’ in reference to stolen possessions.Footnote 35 While lines 82–5, transitional between the curses proper and the final farewells, are as such a likely home for 66, the fact that both 82 and 66 play on the conventions of a ‘prayer for justice’ strengthens the case for a connection. Ribbeck placed 66 in front of 82, but, on the one hand, the vocatives of 82–3 are a good way to signal the beginning of a new section while, on the other, on its own 66 is not an effective summation of the preceding curses.Footnote 36 Goodyear moved it after 83, which is an improvement.Footnote 37 Yet after 85 is, I suggest, a more plausible place. On the one hand, making 66 answer the question of 84–5 brings out the logic of a ‘prayer for justice’ in a clearer way: ‘Have I parted with my farm, so that the soldier could have it as his own? Yes, everything has been taken from me, but it now all belongs to Dis.’ On the other, the similarity of the beginnings of 85 (miles) and 66 (nil est) could account for the latter's omission.

DIRAE 88

The speaker casts a last glance over his estate (86–8):

hinc ego de tumulo mea rura nouissima uisam,
hinc ibo in siluas: obstabunt iam mihi colles,
obstabunt montes, campos audire licebit.

As Goodyear points out, the second half of 88 can hardly be right: ‘The phrase as it stands is fatally ambiguous, able to mean either “the plains will be able to hear” or “it will be possible to hear the plains”, and no less fatally obscure.’Footnote 38 Earlier corrections—humanist campos nec adire, Putsche's campis nec abire, Sillig's campos haud ire, Diggle's campos laudare—fail to convince.Footnote 39 More recently, Watt has suggested camposque uidere negabunt, which appears to produce an apposite sense: ‘the mountains will not let me see the plains’.Footnote 40 As he points out, the confusion between forms of audire and uidere is widespread, and the change of negabunt to licebit can also be paralleled. The one weakness of Watt's proposal is that, with an infinitive, negare means not ‘to forbid to’ but ‘to refuse to’ (OLD s.v. nego 4). To obtain the required sense, we should adopt Watt's camposque uidere but write uetabunt instead of negabunt.Footnote 41 The corruption of uetabunt to licebit may be due to the scribe's eye inadvertently jumping down to licebit at the end of 103.

DIRAE 97

The final stanza of the Dirae is introduced by the following refrain (97):

extremum carmen reuocemus, Battare, auena.

At first sight it appears unobjectionable, but in fact it has two potential problems. One is that bare auena at the end of the line is weak; it has the feel of an afterthought or even a mere metrical filler (contrast 14 or 54 [= 71], which lack such an instrumental ablative). Heinsius proposed changing extremum to extrema: this improves the syntax, but the harsh metonymy extrema auena for ‘my last playing on the panpipe made from reeds’ does not appeal.Footnote 42 There is, in fact, a further irregularity, even if on its own it may not be objectionable: elsewhere in the poem, the form carmen usually occurs in the sixth foot (14, 30, 54, 71, 75; line 19 is the one exception, which is suspect also for other reasons) and is usually preceded by Battare in the fifth (except at 75). Herrmann therefore had a point when he proposed swapping carmen and auena; since, however, auena between extremum and reuocemus violates the metre, we have to ask what else can have stood there.Footnote 43 As 14 and 54 (= 71) suggest, that would be a good place for an adverb; I have considered borrowing iterum from these refrains, but it has two weaknesses: on the one hand, it would make the corruption to carmen somewhat less likely; on the other, extra emphasis on the idea of repetition may be less welcome at the close of the poem. I suggest that the adverb we need is tandem: it is reasonably similar to carmen in both shape and sound (the corruption will, of course, have been influenced by carmen at the end of the line), and it appears appropriate in a refrain introducing the final stanza. The second problem is indicated by the fact that all other refrains that refer to the following stanza with carmen (carmina) define it with hoc (haec): 14 tristius hoc iterum repetamus, Battare, carmen, 25 (= 47) sic precor, et nostris suberunt haec carmina uotis, 30 hoc Footnote 44 mihi saepe meum resonabit, Battare, carmen, 54 (= 71) tristius hoc iterum reuocasti, Battare, carmen, 75 tristius hoc rursum dicit mea fistula carmen. We may also note that the refrains containing the idea of repetition (14, 54, 71, 75) imply specifically the repetition of the preceding stanza. What does the final refrain mean? Without the deictic hoc referring to what follows, extremum carmen can most naturally be interpreted as ‘the conclusion of the song’; while this is not in itself an implausible meaning, the point of reuocemus is lost: ‘let us repeat the end of the song’ makes no sense, because the end of the song has not yet been performed. I suggest that we should restore hoc here as well: extremum hoc tandem reuocemus, Battare, carmen, ‘Battarus, let us at last repeat this stanza as the final one’;Footnote 45 the implication will then be that 98–101 reiterate, as in fact they do, 4–8.

LYDIA 107–8, 109–10

The speaker of the Lydia envies the fields and meadows in which he used to spend time with his girlfriend, but which now have her company to themselves (107–10):

uos nunc illa uidet, uobis mea Lydia ludit,
uos nunc alloquitur, uos nunc arridet ocellis,
et mea submissa meditatur carmina uoce,
cantat et interea, mihi quae cantabat in aurem.

An overlooked difficulty concerns the sense and construal of 107 uobis mea Lydia ludit: what case is uobis, and what does ludit mean? Van der Graaf translates: ‘my Lydia plays with you’; and explains in the commentary: ‘Lydia's play is specified in 114, 117, etc. (cf. Catullus’ second poem).’Footnote 46 This translation, however, confuses two different senses of the English construction, ‘to play with a friend’ and ‘to play with a toy’: in a context such as this we expect the former, but the Latin (ablative without cum) can only convey the latter. In his revision of Fairclough's Loeb, Goold translates: ‘in you my Lydia plays’; but a locative interpretation of uobis is likewise inapposite, since the fields and meadows addressed here are virtually personified (besides, it is doubtful that the ablative of a pronoun can have a locative sense).Footnote 47 Before proposing an alternative construal, I wish to call attention to another potential cause for concern: there seems to be no logic in the order of the activities listed in 107–8 (107a uidet, 107b ludit, 108a alloquitur, 108b arridet). In and of itself, this may not be a sufficient reason for intervention; but if we swap the second hemistichs of 107 and 108, by way of an experiment, the advantages of the new arrangement are obvious. First, the triple anaphora of uos nunc is now uninterrupted. Second, the progression of ideas from 107a uidet (‘she sees you’) to 108b arridet ocellis (‘she smilingly looks at you’) to 108a alloquitur (‘she speaks to you’) obtains a clear logic (I stop at 108a for the moment). And third, the new position of 107b finally enables us to construe the phrase in a way that makes sense. Placed between alloquitur and meditatur (and cantat), ludit can likewise be taken to refer to verbal/vocal performance, with uobis being a dative: ‘my Lydia plays/performs for you’; what Lydia performs is explained in 109–10. Turning back to an earlier point, we can now take the climactic progression of ideas one step further: she admires you, she smiles at you, she speaks to you—and she sings for you.Footnote 48 The corruption is probably due to homoearchon: having copied the first hemistich of the first line, the scribe would jump down to the second hemistich of the second line; he would then add the omitted hemistichs in the margins, but the subsequent copyist would fail to restore them in their original place.

As we have established, 109–10 are intended to explicate 107b uobis mea Lydia ludit, but they involve some difficulties too. First, 109 appears to refer to Lydia's repeating the songs which the speaker used to sing to her, but there are reasons to doubt this interpretation. On the one hand, though less crucially, these carmina would most naturally be love songs addressed to Lydia, and it will be odd to imagine her repeating them to herself. On the other, meditatur should normally mean ‘to compose, rehearse’, and the same idea seems to be implied by submissa … uoce (Lydia does not sing these songs aloud, because they are not ready yet); this suggests that carmina should be Lydia's own.Footnote 49 While in principle it could make sense for 109 and 110 to refer to Lydia's singing her lover's and her own songs in turn, the contrast may in fact be between polishing new songs (109) and performing old ones (110).Footnote 50 We may thus provisionally mark mea as suspect (it could be a reflex of mea in 107b). A second difficulty concerns the articulation of the two lines as a whole; we could expect them to be saying that Lydia meditatur and cantat in turn, but, as it stands, the passage cannot be construed in this way. A glaring problem is 110 interea ‘in the meantime’, which makes no plausible sense in the context and for which Wagner's interdum is a compelling correction.Footnote 51 While interdum ‘sometimes’ is clearly a fitting word for our context, on its own it may not be sufficient to restore a plausible text. On the one hand, the articulation of the passage would be clearer if the first clause were likewise introduced by interdum or a synonym, so as to establish a bipartite interdum … interdum construction, even though this may not be strictly necessary. Writing nunc for et at 109 might be a possibility,Footnote 52 but it would clash with the repeated nunc in 107–8, where it has a temporal force (compare 124 et uobis nunc est mea quae fuit ante uoluptas); writing modo for mea—which, as argued above, is also suspect for another reason—seems a better option.Footnote 53 On the other hand, as part of a bipartite construction meaning ‘now … now’, interdum introducing the second clause cannot be preceded by et; besides, its normal position would be at the very head of its clause.Footnote 54 This suggests that we should write interdum cantat; the error is probably due to the omission of cantat, which was subsequently restored in the wrong place (with et added to fix the metre).

LYDIA 115–16

The speaker envies the farmlands in which Lydia will walk (114–17) –

aut roseis uiridem digitis decerpserit uuam
(dulci namque tumet nondum uitecula Baccho)
aut inter uarios, Veneris stipendia, flores
membra reclinarit teneramque illiserit herbam.
116 ueneris recc.: uenerem MSL | stipendia SL: spumantia M

A long-recognized problem concerns the middle of 116: while it seems virtually certain that between uarios and flores we should have an appositional phrase comparable to Columella 10.96 uarios, terrestria sidera, flores and that, accordingly, Veneris is probably right, stipendia has duly been suspected. The term is markedly ‘unpoetisch’: other than in Plautus and until Prudentius, it only occurs three times in poetry: Ennius (Ann. 215 Poeni stipendia pendunt) and Catullus (64.173 dira ferens stipendia tauro) use it as a technical term for either ‘soldiers’ wages’ or ‘war reparations’ (respectively); Horace (Epod. 17.36 quod me manet stipendium?) gives it a less precise sense (‘fine’), but its down-to-earth flavour is clearly what he is after.Footnote 55 In our context the term is problematic at least on two counts: first, it does not fit stylistically; second, the exact literal sense underlying the metaphor of Veneris stipendia is not clear (‘wages paid to Venus’? ‘tax levied in Venus’ interest’?).Footnote 56 If that were not enough, stipendia is, in fact, only the reading of one branch of the manuscript tradition (SL). Building on M's spumantia, some scholars have attempted to restore a participial phrase in agreement with 117 membra (such as Venerem spirantia Eichstädt, simulantia Sillig); but this results in a rather awkward syntax.Footnote 57 As noted above, it seems clear that we need a noun, and Shackleton Bailey proposed spiracula (‘Venus breathes through the flowers, causing their fragrance’): this is absurd, but points in the right direction.Footnote 58 I suggest that we should write spectacula, ‘Venus’ show’; the underlying image is that of spring as a festival held by, or in honour of, Venus.Footnote 59

Once we have restored spectacula, the jingle with uitecula (directly above it), though not quite impossible, becomes a cause for concern.Footnote 60 There are two further reasons to suspect uitecula, even if neither constitutes a decisive objection. First, the diminutive uiticula is not used in classical verse, and is quite rare in prose too.Footnote 61 Second, the metonymy is rather harsh (‘the vine swells with juice’).Footnote 62 I tentatively suggest that uitecula is an error for uindemia: written exactly over (specta)cula, (uīde)mia could easily produce (uite)cula;Footnote 63 for tumet uindemia, compare Auson. Mos. 195 uindemia turget. This conjecture may not be necessary in the sense that it fixes an unquestionably corrupt text, but it is fairly easy and simultaneously solves all the issues (admittedly minor): it removes the unattractive jingle (if spectacula is right); eliminates a conspicuous lexical rarity; and replaces the unparalleled metonymy with a standard one (‘ripe grapes do not yet swell with sweet wine’).

LYDIA 121

The speaker then describes the enjoyment which Lydia's company brings his estate in his absence (119–22):

gaudebunt siluae, gaudebunt mollia prata
et gelidi fontes, auiumque silentia fient,
tardabunt riui †labentes currite lymphae,
dum mea iucundas exponat cura querelas.

Kenney obelized a portion of 121 not because no correction could be found, but because it was uncertain which correction should be adopted.Footnote 64 Some scholars have in fact accepted the paradosis, taking labentes currite lymphae as a parenthesis and suggesting that labi means here something like ‘to flow gently, quietly’.Footnote 65 Yet this is far-fetched: it is quite doubtful that labentes can have this pregnant sense, and the Orphic topos requires that the rivers stop their flow rather than merely reducing its violence. It seems thus more or less certain that the imperative currite is corrupt, and one solution has been to change it to sistite (humanist). This brings us to a more general objection: a parenthetic apostrophe is out of place here. To begin with, the speaker is separated from the situation he is describing both spatially and temporally: it will happen in the future, and in a place which he has left. It is true that the entire first part of the Lydia is addressed to the abandoned farmlands, but the speaker never implies that they can actually hear him; an imperative does involve such an expectation. Perhaps more importantly, the imperative currite will jar with the future indicative tardabunt: an assertion should not depend on the fulfilment of a command. This is not merely a linguistic issue: the point of the passage is that the rivers, enchanted by Lydia's singing, will stop flowing of their own accord, not in response to the speaker's request. The likeliest correction for currite seems therefore to be currere of the 1534 Aldine edition, to be construed with tardabunt. With this change, the line can in principle make sense: ‘the flowing waters of the river will stop to run’. Yet both the general sense and the context (note 119 gaudebunt siluae) suggest that it would be preferable to have riui as the subject, and Scaliger proposed turning labentes … lymphae into an ablative: tardabunt riui labenti currere lympha, ‘the rivers will stop to run with their flowing water’.Footnote 66 Again, this is not impossible, but one cannot avoid the feeling that the expression is vague and wordy. I suggest that we need the accusative: tardabunt riui labentes currere lymphas, ‘the rivers will make their flowing waters stop their course’.Footnote 67 This has the advantage of giving both riui and lymphae a clear role to play, which moreover can be paralleled in the Dirae: note especially 67 flectite currentes lymphas, uaga flumina, retro, but also 48 undae, quae uestris pulsatis litora lymphis, where similarly ‘streams’ (flumina) and ‘waves’ (undae) appear as active agents which have control over their ‘waters’ (lymphas/lymphis).Footnote 68

LYDIA 175

After the references to Jupiter's premarital affair with Juno (166–8) and to Venus’ affair with virginal Adonis (169–74), Aurora is added to the list of divine seducers (175–6):

non Aurora nouos etiam plorauit amores
atque rubens oculos roseo celauit amictu?

Although it is true that she eventually lost her lovers Orion, Cephalus and Clitus, in one way or another, the idea that ‘Aurora was unhappy in love’ (suggested by plorauit) seems out of place here.Footnote 69 Hiding one's eyes is a gesture of shame (note also rubens, ‘blushing’), and while in principle the implication could be that Aurora was ashamed of crying (compare Stat. Achil. 1.233–4 udaque celat | lumina), this would give too much prominence to an action that could only indicate her affairs in the most indirect way, if at all. The most natural reason for Aurora's feeling of shame would be her indecorous behaviour, an idea which plorauit can hardly imply. A passage from Ovid provides an illuminating parallel (Ars am. 3.83–4):

Latmius Endymion non est tibi, Luna, rubori,
nec Cephalus roseae praeda pudenda deae.

Ovid playfully denies that Luna's and Aurora's ‘blushing’—the rising moon can be reddish, and so is dawn—was caused by their affairs with mortal lovers; the idea rejected by Ovid must be implied in the Lydia.Footnote 70 What can plorauit be concealing? One possibility is to assume that the line should refer to Aurora's enjoying her relationships with men other than her husband, and gustauit or libauit could convey such a sense (compare 168 gaudia libauit dulcem furatus amorem). This may not be entirely impossible, but the context suggests that the passage should be making a different point: not that Aurora was unfaithful to Tithonus, but that she seduced unexperienced young men.Footnote 71 I have thought of uiolauit, yet while uiolare can in principle imply rape, with amores it would more naturally mean ‘to betray love’ than ‘to seduce a lover’. I suggest that raptauit would be a more suitable term, for which compare Ov. Her. 15.87–8 hunc ne pro Cephalo raperes, Aurora, timebam, | et faceres, sed te prima rapina tenet, Eur. Hipp. 454–5 ἀνήρπασέν ποτε | ἡ καλλιφεγγὴς Κέφαλον ἐς θεοὺς Ἕως.Footnote 72 While Cephalus may well be implied here in the first place, the vagueness of nouos … amores, as well as the frequentative raptauit, could hint that this was not an isolated occasion and other young men had likewise been abducted by Aurora.Footnote 73 One difficulty remains: the point of etiam should be ‘did not Aurora too?’ (that is, as well as the other deities mentioned earlier), but positioned between nouos and raptauit it can more naturally be construed with either of them. A conceivable option would be to write nonne nouos Aurora etiam, but the elision at the penthemimeres does not appeal. I tentatively propose replacing etiam with totiens instead, which will make Aurora a suitable capping for Jupiter and Venus; for the idea, we may compare ps.-Apollodorus’ report that, after Eos slept with Ares, Aphrodite in revenge made her repeatedly fall in love with men one after the other (1.27 ἐποίει γὰρ αὐτὴν Ἀφροδίτη συνεχῶς ἐρᾶν, ὅτι Ἄρει συνευνάσθη).Footnote 74

Footnotes

This article was written as part of a project that has received funding from the European Union's Horizon 2020 Research and Innovation Programme under the Marie Skłodowska-Curie grant agreement No. 840190. I am enormously grateful to Stephen Heyworth, who read and commented on an earlier version of the article, and to Tristan Franklinos, who discussed with me some of the ideas presented here. I should also like to thank CQ's editor Bruce Gibson and the anonymous reviewers for criticisms and suggestions.

References

1 The Dirae and the Lydia are transmitted as one text, but there are compelling reasons for taking 1–103 and 104–83 as two separate poems, as was first realized by Jacobs, F., ‘Ueber die Diras des Valerius Cato’, Bibliothek der alten Litteratur und Kunst 9 (1792), 56–61Google Scholar (he identified the latter as the Lydia of Valerius Cato: the idea has been widely rejected since then, but the title is convenient). Jacobs adduced three fundamental arguments against the unity of 1–183: (1) the Dirae proper (1–103) exhibits a clear ring composition, with the last stanza (98–103)—identified as such (extremum) in the refrain (97)—echoing the first stanza (4–8); (2) the situations in the two poems are essentially different (in the Dirae the speaker is exiled from his farm, in the Lydia he is dying of lovesickness); (3) the style (‘Ton’ in Jacobs's terms) likewise differs notably between the two poems. Subsequent scholars developed and expanded these arguments, which even led some to believe that the two poems are not by the same author; note especially Rothstein, M., ‘De Diris et Lydia carminibus’, Hermes 23 (1888), 508–24Google Scholar and cf. e.g. Fraenkel, E., ‘The Dirae’, JRS 56 (1966), 142–55Google Scholar, at 151: ‘Nowadays it is also generally agreed […] that the two poems, the Dirae and the Lydia, cannot have been written by one and the same man. Metre, style, and mentality […] are all entirely different’, though Goodyear, F.R.D., ‘The Dirae’, PCPhS 17 (1971), 30–43Google Scholar, at 39 is right to be more cautious. C. van der Graaf, The Dirae (Leiden, 1945), 127–34 tried to defend the unity of 1–183, as well as Virgilian authorship; however, the analytical position has been upheld in subsequent scholarship: see especially Schutter, K.H.E., ‘De Lydia et Diris carminibus’, Mnemosyne 6 (1953), 110–15CrossRefGoogle Scholar; B. Luiselli, ‘Pseudovirgilio: le Dirae e la Lydia’, in Luiselli, B., Studi sulla poesia bucolica (Cagliari, 1967), 117–144Google Scholar; Van den Abeele, E., ‘Remarques sur les “Dirae” et la “Lydia” de l’“Appendix Vergiliana”: les arguments de Van der Graaf sont-ils irréfutables?’, RhM 112 (1969), 145–54Google Scholar; Pantoja, M. Rodríguez, ‘Dirae–Lydia: los argumentos de la métrica’, in Santini, C., Zurli, L., Cardinali, L. (edd.), Concentus ex dissonis: scritti in onore di Aldo Setaioli (Naples, 2006), 577–92Google Scholar. The few dissenting opinions that have been voiced in recent years succeed at best in demonstrating that there exist meaningful points of contact between the Dirae and the Lydia, possibly implying that they come from a single collection, but not that they are one poem: see A. Salvatore, ‘Da un «dramma» politico a un dramma esistenziale: le Dirae dell’Appendix Vergiliana e il problema dell'unità’, in F. Del Franco (ed.), Storia, poesia e pensiero nel mondo antico: studi in onore di Marcello Gigante (Naples, 1994), 594–604; S. Lorenz, ‘Inuideo uobis, agri: mea gaudia habetis: bukolische Verwünschungen und elegische Eifersucht in den Dirae’, in N. Holzberg (ed.), Die Appendix Vergiliana: Pseudepigraphen im literarischen Kontext (Tübingen, 2005), 1–27; M. Stachon, Tractavi monumentum aere perennius: Untersuchungen zu vergilischen und ovidischen Pseudepigraphen (Trier, 2014), 178–200; N. Holzberg, ‘Einführung’, in F. Zogg (ed.), Appendix Vergiliana: lateinisch–deutsch (Berlin, 2020), 11–42, at 16–18.

2 Van der Graaf (n. 1). Fraenkel (n. 1) and Goodyear (n. 1) are important treatments, both of which offer a critical text with ample notes, often surpassing van der Graaf's comments in detail and depth; both, however, are selective, and besides only cover the Dirae proper. A number of annotated bilingual editions have also appeared, including G.P. Goold's revision of Fairclough, H.R., Virgil Aeneid VII–XII, Appendix Vergiliana (Cambridge, MA, 2000)Google Scholar, Iodice, M.G., Appendix Vergiliana (Milan, 2002)Google Scholar and K. Rupprecht, ‘Dirae’, in F. Zogg (ed.), Appendix Vergiliana: lateinisch–deutsch (Berlin, 2020), 44–59; but these are of little relevance when dealing with textual issues. The commentary by A.F. Naeke, Carmina Valerii Catonis (Bonn, 1847), dated though it is, remains fundamental for the textual criticism of the Dirae and the Lydia.

3 Giardina, G., ‘Nuovi emendamenti al testo delle Dirae e della Lydia pseudovirgiliane’, QUCC 92 (2009), 167–73Google Scholar is one exception.

4 The closest parallel is supplied by Mart. 8.3.17 scribant ista graues nimium nimiumque seueri; cf. also e.g. [Tib.] 3.6.21 conuenit iratus nimium nimiumque seueris, Ov. Her. 1.41 ausus es, o nimium nimiumque oblite tuorum. In general, the reduplication of nimium (nimis) is quite frequent, whereas the only parallel for pairing nimium with multum I am aware of comes from Prosper of Aquitaine (fifth century a.d.), Epigr. 82.3 a quo longinquus multum est nimiumque remotus, where the order multum – nimiumque is less objectionable, though likewise one cannot avoid the suspicion that multum is a misreading of nimium.

5 The corruption can be paralleled in Maximianus 2.1 en dilecta mihi nimium formosa Lycoris, where some manuscripts read multum for nimium. An alternative possibility is suggested by Trappes-Lomax, J., ‘Sub aqua, sub aqua: submerged repetition in Ovid and Rutilius Namatianus’, PCPhS 51 (2005), 86–9Google Scholar, at 86 n. 1, who assumes that nimium was omitted by haplography and multum was then supplied to fill the metre (nimium nimiumque → nimiumque → nimium <multum>que).

6 Here and below, I start with the text by Kenney, E.J., ‘Dirae (Lydia)’, in Clausen, W.V., Goodyear, F.R.D., Kenney, E.J., Richmond, J.A., Appendix Vergiliana (Oxford, 1966), 1–14Google Scholar, which as a rule is sufficiently close to the paradosis, without being slavishly so; when necessary, I supply a concise critical apparatus of my own. The edition by A. Salvatore, ‘Dirae’, in A. Salvatore, Vivo, A. De, Nicastri, L., Polara, G., Appendix Vergiliana (Rome, 1997), 1–23Google Scholar, though offering a somewhat fuller account of the manuscripts, does not improve on Kenney's text.

7 Naeke (n. 2), 40–1.

8 Note that here T (Paris. Lat. 8069) further corrupts Rursus to ursus and similarly reads ristius for Tristius at 75 (cf. also estica for Pertica at 45, ulcius for Dulcius at 71, nuideo for Inuideo at 123). In general, manuscripts of the L family more or less systematically begin refrains with rubricated capitals, even though this need not imply that the archetype did the same.

9 At 71 the manuscripts read dulcius (instead of tristius), which Kenney (n. 6), 8 retains, though noting in the apparatus criticus: ‘non satis explicatum’; in support of tristius, cf. Goodyear (n. 1), 43. The corruption is palaeographically transparent: T → D, rı → u, ſ → l, t → c (line-initial capitals are particularly variable in shape in minuscule manuscripts, and in this case we may be dealing with the confusion between their insular forms, Ꞇ and Ꝺ); the fact that dulcis is a frequent term in the poem (22, 24, 49, 89 [bis], 98) has no doubt played a role too. Dilherr, J.M., Satyrici minores de corruptis seculi moribus (Jena, 1636)Google Scholar, B 5 seems to have been the first to suggest tristius in print, though it already appears as a marginal variant in Boccaccio's autograph copy, cod. Laur. 31, 33 (see the apparatus criticus in R. Giomini, Appendix Vergiliana [Florence, 1953], 72).

10 See the comments in van der Graaf (n. 1), 19–20.

11 Van der Graaf (n. 1), 2.

12 Heinsius in P. Burman, ‘Valerii Catonis Dirae in Battarum’, in P. Burman, Anthologia veterum latinorum epigrammatum et poematum, vol. 2 (Amsterdam, 1773), 647–78, at 654; the corruption could be triggered by the loss of initial n after non. Note e.g. van der Graaf (n. 1), 20: ‘Heinsius’ naribus is unnecessary, because the singing of the birds and probably also the susurrus of the wind is involved.’

13 See TLL 6.1.553.13–554.39. The empirical fact that this (very frequent) sense of ferre is unattested in the passive should suffice to demonstrate that ferantur cannot mean ‘to be produced’ in our context. My tentative explanation is that this is an inherited feature of the idiom, derived from the fact that ferre in the sense ‘to produce’ (OLD s.v. fero 25: ‘To bear, yield [fruit, produce, etc.]’) is a resultative for ferre in the sense ‘to be pregnant’ (OLD s.v. fero 10), just as, in general, ferre ‘to bring’ is the resultative for ferre ‘to carry’ (cf. A. Pârvulescu, ‘IE. *bhrā́tēr “brother”’, IF 101 [1996], 99–106, at 104): for semantic reasons, ferre ‘to carry in the womb’ cannot normally be used in the passive (its object is almost an internal accusative: note that the verb can be used absolutely, Plin. HN 10.179 cameli XII mensibus ferunt, or with uentrem, Varro, Rust. 2.1.19 equa enim uentrem fert duodecim menses), and, at least historically, ferre ‘to be fruitful with’ apparently involves the same construction (note that it can likewise be used absolutely: Cato, Agr. 6.2 ferundo arbor peribit). In Greek, φέρειν in the sense ‘to produce’ (LSJ s.v. φέρω V) seems likewise to be used in the active voice only (and can likewise be used absolutely), which supports the idea that this is an inherited feature (in general, the sense ‘to carry in the womb’ is attested for *bher- across different Indo-European languages: see e.g. E.P. Hamp, ‘The Indo-European roots *bher- in the light of Celtic and Albanian’, Zeitschrift für Celtische Philologie 39 [1982], 205–18, at 207–9; note that in English too the idiom ‘a tree bears fruits’ cannot be turned into the passive).

14 It is true that the substitution of creentur with ferantur, especially at the end of a line, is not impossible to imagine; but ferantur does not point to creentur in any way.

15 The non-reduplicated variant genere (for gignere) is almost exclusively confined to two authors, Lucretius (3.797; 4.143, 159) and Varro (Rust. 1.31.5, 40.1; 2.2.19, 6.3), and both may constitute relevant precedents. On the one hand, the Dirae poet engages with Lucretius, and appears to borrow his phrasing, elsewhere (note e.g. 6 rerum discordia gliscet, which evokes Lucr. 6.366 discordia sit rerum, while also echoing his frequent use of gliscere, a verb largely avoided by Augustan poets). On the other, Varro's usage suggests that the form may have survived as a more or less technical term in farmers’ idiom (a parallel of a sort is provided by 9 sterilescant: this appears to be a technical term, which otherwise is only attested in Pliny's Natural History).

16 Cf. his comment in Burman (n. 12), 654: ‘flores oculos ac nares oblectant; nihil illis rei cum aure.’

17 It should be noted though that ferantur is only the reading of one branch of the manuscript tradition (M), whereas the other reads forantur (FL).

18 Naeke (n. 2), 54–5. For the former, cf. van der Graaf (n. 1), 21: ‘The meaning is: may my songs of imprecation surpass my curses as to their duration, may even after my death my curses live on’ (which strangely turns the text upside down); yet in the Dirae ‘songs’ (carmina) and ‘curses’ (uota) are one and the same thing, so one can hardly imagine what it should mean for the speaker's carmina to outlast his uota.

19 Heinsius in Burman (n. 12), 654.

20 J. Mähly, Review of Ribbeck, O., Appendix Vergiliana (Leipzig, 1868)Google Scholar, Heidelberger Jahrbücher der Literatur 63 (1870), 769–96 and 801–39, at 777–8, paraphrasing: ‘Mögen diese meine Gesänge meinen Verwünschungen die gehörige Kraft einhauchen, mögen sie meinem Gefühl adaequat sein.’ In particular, it is difficult to see why the reader should take uotis as a dative rather than an ablative: ‘may these songs breathe with my prayers’ is a fairly plausible (if nonsensical) turn of phrase.

21 Giardina (n. 3), 168.

22 OLD s.v. subsum 1d: ‘to be subjoined (in a document); to attach (to a heading)’. Note that the Dirae speaker thinks of his composition in terms of a written text: 34 nostris potius deuota libellis.

23 Pinkster, H., The Oxford Latin Syntax. Volume 1: The Simple Clause (Oxford, 2015), 502CrossRefGoogle Scholar states that the jussive subjunctive can be ‘found in transition formulas – for example, in Varro and Pliny the Elder’, citing Naas, V., Le projet encyclopédique de Pline l'Ancien (Rome, 2002), 209–24Google Scholar; but Naas's dossier rather suggests that in Pliny the first-person hortative subjunctive (8.1 transeamus, 10.212 transeamus, 11.284 transeamus, 23.166 transeamus, 26.164 reddamus) or future (7.215 praeuertemur, 8.229 dicemus, 14.150 dicemus, 15.138 dicemus, 20.1 ordiemur and cogemus, 28.267 praeuertemur, 29.143 dicemus, 30.149 praeuertemur) or the third-person future (4.122 dicentur, 9.1 dicentur, 9.186 dicentur, 24.188 dicetur, 25.1 dicetur, 33.1 dicentur) are the rule, whereas the third-person jussive subjunctive is at best an exception (34.1 dicantur).

24 Lines 42–4 which describe the gradual destruction of the different parts of the estate by a fire is a different case.

25 Cf. my discussion of Lydia 115 and 116 below.

26 Cf. Fraenkel (n. 1), 149.

27 See Courtney, E., ‘A new text of the Appendix’, CR 17 (1967), 42–6Google Scholar, at 43; Courtney, E., ‘A miscellany on Latin poetry’, BICS 29 (1982), 49–54Google Scholar, at 51.

28 R. Ellis, Appendix Vergiliana (Oxford, 1907), 5–ivv. (cf. R. Reitzenstein, Drei Vermutungen zur Geschichte der römischen Litteratur [Marburg, 1893], 40); Heinsius in Burman (n. 12), 664.

29 J. Scaliger, Publii Virgilii Maronis appendix (Lyon, 1572), 440.

30 On ‘prayers for justice’ as a type of curse tablets, see especially H.S. Versnel, ‘Beyond cursing: the appeal for justice in judicial prayers’, in C. Faraone, D. Obbink (edd.), Magika Hiera: Ancient Greek Magic and Religion (Oxford, 1991), 60–106; Versnel, H.S., ‘Prayers for justice, East and West: new finds and publications since 1990’, in Gordon, R.L., Simón, F.M. (edd.), Magical Practice in the Latin West (Leiden, 2010), 275–354Google Scholar.

31 See e.g. Tomlin, R., ‘Cursing a thief in Iberia and Britain’, in Gordon, R.L., Simón, F.M. (edd.), Magical Practice in the Latin West (Leiden, 2010), 245–73Google Scholar.

32 Kenney (n. 6), 8 brackets the line, commenting: ‘uersus siue genuinus siue suppositicius siue corruptus his quid sibi uelit nescio: del. Kenney’.

33 See Putsche, K., Valerii Catonis poemata (Jena, 1828), 90Google Scholar; for the elision, cf. Verg. Aen. 6.188 quando omnia.

34 First published in Curbera, J.B., Delage, M. Sierra, Velázquez, I., ‘A bilingual curse tablet from Barchín del Hoyo (Cuenca, Spain)’, ZPE 125 (1999), 279–83Google Scholar.

35 See J.N. Adams, ‘British Latin: the text, interpretation and language of the Bath curse tablets’, Britannia 23 (1992), 1–26, at 3.

36 O. Ribbeck, ‘Dirarum carmen’, Index scholarum in Academia Regia Christiana Albertina 1867.2, 1–14, at 8.

37 Goodyear (n. 1), 41–2.

38 Goodyear (n. 1), 36.

39 Putsche (n. 33), 95; Sillig, J., ‘Review of K. Putsche, Valerii Catonis poemata (Jena, 1828)’, Jahrbücher für Philologie und Paedagogik 4 (1829), 17–35Google Scholar, at 33; J. Diggle in Goodyear (n. 1), 36 n. 2.

40 Watt, W.S., ‘Notes on the Appendix Vergiliana’, Eikasmos 12 (2001), 279–92Google Scholar, at 280.

41 For pairing uetare with uidere, cf. Ov. Rem. am. 438 et uidit quae mos ipse uidere uetat.

42 Heinsius in Burman (n. 12), 666. Note that auena is usually defined by epithets referring to its physical properties as a stalk of reed: e.g. Verg. Ecl. 1.2 tenui, [Verg. Aen.] proem 1 gracili, Tib. 2.1.53 arenti, [Tib.] 3.4.71 perlucenti, Calp. Ecl. 1.93 tereti, Laus Pisonis 234 sterili, Val. Fl. 4.386 leni; occasionally the reference is to the sound: Calp. Ecl. 3.60 acerbae, Nemes. 3.11 stridentis.

43 Herrmann, L., ‘Trois poèmes de P. Valerius Cato’, Latomus 8 (1949), 111–44Google Scholar, at 129. The likeliest explanation for the corruption seems to be that, first, carmen was repeated in its transmitted place from the end of the line and, next, the original carmen was conjecturally replaced with auena (note that a part of auena is erroneously introduced in the text of the Dirae on at least one other occasion, at 21 pingunt uerna [Heinsius: auena codd.] colore). It may be worth pointing out that cod. Monac. 21562 (copy of m) reads, before correction, extremum carmen reuocemus, Battare, carmen: this is, of course, an easier error (especially in view of the frequency of Battare, carmen earlier in the poem), but it illustrates the same process.

44 This is a conjecture by Ribbeck (n. 36), 5 for the transmitted nec, which is untenable; however, the line may be more gravely damaged.

45 Cf. the opening of Virgil's final eclogue (10.1): extremum hunc, Arethusa, mihi concede laborem, ‘grant me, Arethusa, this work as the last one’.

46 Van der Graaf (n. 1), 6 and 31.

47 Fairclough (n. 2), 397; in the original Loeb, H.R. Fairclough, Virgil, vol. 2 (London, 1918), 473 translated: ‘with you my Lydia plays’.

48 The one potential objection could be the repetition of mea at close distance (uobis mea Lydia ludit | et mea submissa meditatur carmina uoce), but, as I suggest below, mea at 109 is probably corrupt. In any event, it will not have much weight in view of 122–4 dum mea iucundas exponat cura querelas. | inuideo uobis, agri: mea gaudia habetis, | et uobis nunc est mea quae fuit ante uoluptas and 157–9 sacratamque meae uittam temptare puellae | immatura mea cogor nece soluere fata? | istius atque utinam facti mea culpa magistra.

49 The line could in principle be taken to mean ‘she rehearses songs I wrote for her to sing’, but the idea seems strained.

50 The exact nuance of meditari in reference to musical composition or performance is elusive (OLD s.v. meditor 7 ‘To work over (a song, etc.) in performance’ is not very helpful), but it never seems to be used of formal singing in front of an audience and rather implies solitary practising (Verg. Ecl. 6.82–3 omnia, quae Phoebo quondam meditante beatus | audiit Eurotas is telling: the Eurotas overhears Apollo). By contrast, cantare is the standard term for bucolic singing; see M. Lipka, Language in Vergil's Eclogues (Berlin, 2001), 21–2.

51 Wagner, W., ‘On Ribbeck's Virgil’, TPhS 12 (1867), 198–239Google Scholar, at 235. Earlier scholars took interea to mean interdum ‘sometimes’, but this is unwarranted; explaining interea as ‘while I am away’ does not produce a plausible sense.

52 I owe the suggestion to Stephen Heyworth.

53 For the corruption, cf. e.g. Maximianus 1.214 quaeque breuis fuerat iam modo [v.l. mea] longa mihi est, 4.22 cantabam dulces quos solet illa modos [v.l. meos]; cf. further Stat. Theb. 2.450 si modo [v.l. mihi] notus amor, Maximianus 3.41 sit modo [v.l. mihi] certa fides. For carmina undefined by an epithet, cf. Ov. Met. 14.341 quae [sc. Canens] dum feminea modulatur carmina uoce.

54 I cite all occurrences of a bipartite modo … interdum construction in poetry I am aware of: Hor. Sat. 1.9.9 ire modo ocius, interdum consistere, 2.7.7–8 pars multa natat, modo recta capessens, | interdum prauis obnoxia, Prop. 2.15.5–6 nam modo nudatis mecum est luctata papillis, | interdum tunica duxit operta moram, Ov. Met. 2.188–90 animo metitur utrumque | et modo … | prospicit occasus, interdum respicit ortus, 3.77–8 ipse modo immensum spiris facientibus orbem | cingitur, interdum longa trabe rectior adstat, 13.540–2 duroque simillima saxo | torpet et aduersa figit modo lumina terra, | interdum toruos sustollit ad aethera uultus, Moretum 29–31 modo rustica carmina cantat | agrestique suum solatur uoce laborem, | interdum clamat Scybalen. Less systematically, I have also checked examples of the same construction produced by interdum combined with other discourse markers, and am likewise unable to cite any cases of the second clause introduced by et.

55 Cf. G. Trimble, ‘A commentary on Catullus 64, lines 1–201’ (Diss., Oxford, 2010), 286.

56 Apul. Met. 9.20 prima stipendia Veneri militabant nudi milites suggests a more natural sense that Veneris stipendia could have (‘service in Venus’ army’), but it is impossible in our context; here we could expect something like ‘offerings to Venus’ (cf. Lucr. 1.17–18), but stipendia cannot convey this sense.

57 H.K.A. Eichstädt, Valerii Catonis Dirae (Jena, 1826), 18; J. Sillig, Review of Eichstädt (this note), Jahrbücher für Philologie und Paedagogik 1 (1826), 333–43, at 339.

58 Bailey, D.R. Shackleton, ‘Notes on minor Latin poets’, Phoenix 32 (1978), 305–25CrossRefGoogle Scholar, at 309.

59 For the idea, cf. Lucr. 1.17–18 tibi suaues daedala tellus | summittit flores and, with greater detail, 5.737–40: the earth is adorned with flowers to celebrate Venus’ advent; spectaculum is readily used in a variety of figurative ways, cf. Verg. G. 4.3 admiranda tibi leuium spectacula rerum (of bees), Hor. Sat. 2.2.26 picta pandat spectacula cauda (sc. the peacock), Man. 1.737 dum noua miratur propius spectacula mundi (sc. Phaethon), Aetna 156 nulla daret miranda sui spectacula tellus (of volcanoes), Paulinus of Nola 10.176 spernentes uarias, rerum spectacula, formas (of the pleasing appearances of material things as opposed to spiritual ones).

60 Naturally, this could be considered an argument against proposing spectacula at 116 in the first place; as I go on to suggest, however, uitecula is not unproblematic on its own.

61 Only five occurrences in classical prose (as a deprecatory term for grapevine as such: Cic. Nat. D. 3.86, Plin. Ep. 1.24.4; ‘young vine’: Columella 4.12.1; ‘vine-twig’: Plin. HN 24.98, 27.44). By contrast, uiticula has yielded a rich harvest of descendants across Romance languages (see Französisches etymologisches Wörterbuch 14.552–7 s.v. uiticula); this makes it likely that the scribe could be familiar with the term from everyday usage and could therefore introduce it by error.

62 The usual idiom is ‘the grape cluster swells with juice’, which involves no metonymy (at least, not at that side of the image, though of course ‘wine’ for ‘juice’ is a metonymy): Ov. Am. 1.15.11 mustis uua tumebit, Tr. 5.3.36 incluso plena sit uua mero, German. fr. 4.17 Libra tumescentes musto bene percoquit uuas, Mart. 13.68.2 turget adhuc uiridi cum rudis uua mero. Ov. Tr. 4.6.9 extensis tumeat facit uua racemis (‘the cluster swells with grapes’ for ‘grapes swell on the cluster’) does involve a metonymy but a different and less bold one. Sil. Pun. 7.164 grauidae cui nectare uites (‘the vine is heavy with juice’ for ‘the vine is heavy with grapes’) and Sen. Oed. 157–8 non plena suo uitis Iaccho | bracchia curuat (‘the vine is replete with wine’ for ‘the vine is replete with grapes’) come closest to our example in that they have uitis as the subject, but, unlike turgere or tumere which can only have the physical meaning ‘to swell’, both grauidus and plenus can have the figurative sense ‘abundant with’ as opposed to the physical ‘filled with’. In other words, the rule is observed that, in this kind of context, turgere and tumere always have uua as the subject.

63 The same kind of corruption has taken place twice elsewhere in our passage: first, ueneris, written over teneram, became uenerem; second, teneram, written under uenerem, became tenerem in one branch of the manuscript tradition (SL).

64 Kenney (n. 6), 11: ‘inter tot ueri similia uerum latet’.

65 See e.g. van der Graaf (n. 1), 33.

66 Scaliger (n. 29), 442.

67 For construing tardare with an accusative and infinitive, cf. OLD s.v. tardo 1b; note especially TrRF adesp. 1.74.6 illum ut maeror tardaret sequi.

68 J.P. Postgate, Select Elegies of Propertius (London, 18852), lxvii–lxix discusses the more general phenomenon (which he calls ‘disjunctiveness’) when the same object or idea is referred to by two synonymous terms in a seemingly pleonastic way, citing in particular Prop. 1.11.11–12 aut teneat clausam tenui Teuthrantis in unda | alternae facilis cedere lympha manu, but this may rather parallel Scaliger's correction of our line. The Dirae poet has a clearer and more specific image in mind, in that flumina and undae are conceived of as entities (deities?) endowed with agency (note that they are addressed in the vocative), whereas lymphae merely refers to their physical substance.

69 Pace van der Graaf (n. 1), 42.

70 The implication is probably that, when Luna and Aurora appear in the sky, they have just left their lovers, so it is natural to interpret their redness as a sign of embarrassment. Note that the catalogue of divine lovers, which Aurora closes, was opened with Luna, who was said to spend every day with Endymion (142–5); the latter passage is severely corrupt, but Sonntag, M., ‘Pseudo-Vergil Lydia 39–41’, WKPh 13 (1896), 478Google Scholar must be right about its intended sense: ‘Während der Sonnengott die Nacht benutzen muß, um die Geliebte aufzusuchen, bleibt der noctivaga Phoebe nur der Tag zu solchem Verkehr übrig.’ This strengthens the case for taking 176 to imply a previous night spent with a lover (while epic decorum presupposes that Aurora spends nights with her husband Tithonus).

71 To begin with, this is what is at issue for the speaker: he is punished for seducing virginal Lydia (156–8). The mythological examples—Jupiter raping Juno before they were married (166–8), and Venus making young Adonis her lover (169–71)—point in the same direction.

72 The corruption is probably due to the scribe's memory of 139 plorauit amores; but internal transposition (ra-pta- → plo-ra-) may also have played a role.

73 Note e.g. Od. 15.251 Κλεῖτον χρυσόθρονος ἥρπασεν Ἠώς, Hom. Hymn 5.218 Τιθωνὸν χρυσόθρονος ἥρπασεν Ἠώς, [Apollod.] 1.27 Ὠρίωνος δ’ Ἠὼς ἐρασθεῖσα ἥρπασε καὶ ἐκόμισεν εἰς Δῆλον.

74 For totiens in reference to a person's behaviour, cf. e.g. Verg. Aen. 4.536 quos ego sim totiens iam dedignata maritos, Prop. 2.33.5 quae dea tam cupidos totiens diuisit amantes, Ov. Met. 1.606 deprensi totiens iam nosset furta mariti.