3.1.21–4

What Propertius says of poetry is refuted by his scribes. In v. 23 ΠΛ would have him say ‘age makes all things greater after one's death’; N prefers ‘it [sc. honos] fashions greater things after the death of old fame’. It requires no great penetration to see which of these readings is the better, but even the best is not wholly free of fault. There are two problems here essentially. In the first instance, N, commonly esteemed as the most ‘honest’ of the manuscripts, offers two singularly terrible variants for words which in ΠΛ are perfectly sound and in fact indispensable.Footnote ² How is this to be squared with its typically impeccable reputation? To dismiss fame … uetuste with A. E. Housman as a mere ‘careless blunder’ will not do:Footnote ³ N does not blunder this badly without reasonable cause, and editors are neglecting their charge if they do not make at least an attempt to explain it.Footnote ⁴ Secondly – a point made by Heyworth – post obitum is an otiose repetition after the same in v. 22, and fails to give good sense in its second appearance.Footnote ⁵ That the value of a person's works is appreciated after death is a thought already expressed in v. 22 with respect to Propertius and in v. 24 with respect to everyone else. Verse 23 seems to contain a statement of wider application, namely that age increases the worth of all things, period: a general truth in contrast to the particular circumstance described in vv. 21–2, and more than a mere echo of maius ab exsequiis nomen in ora uenit in v. 24.Footnote ⁶ That post obitum is corrupt, and N's fame and uetuste some desperate attempt to fix it, is the inevitable conclusion, and a successful answer to this crux must start from that as a premise. But, says Heyworth, ‘[i]t is difficult to see what else could fill the line; “time makes everything greater” is the required sense, and an insertion that was not tautologous would impair the verse.’Footnote ⁷ Ezra Pound added ‘regardless of quality’ to his paraphrase, but this is quite unlikely to be what Propertius actually said (as if he conceived his own poetry to be of doubtful quality).Footnote ⁸ I for my part would be satisfied with an adverb meaning merely ‘in the end, eventually’, and for post obitum therefore suggest postremo. Errors of perseveration (so the experts call them) are frequently triggered by a likeness in the beginnings of words; and in this case postremo could well have prompted the scribe of Ω to recall post obitum from the previous verse and so write it out in the next. That postremo can bear the sense of ‘in the end, eventually’ is illustrated through several parallels gathered by the authors of the OLD s.v. 3, as well as a contextually relevant fragment of Enn. Ann. 16.5 (Skutsch) (= Gell. NA 9.14.5):

‘In the end, the long age of day shall put an end to’ probably the monuments discussed in the foregoing fragment, 16.4 (Skutsch). Whether postremo in Ennius’ text truly means just ‘in the end, eventually’ cannot be known of a surety, but it is difficult to see what else it could signify, and all translators render it in this sense.Footnote ¹⁰ However, despite being a very apposite parallel for the meaning of postremo posited here, Enn. Ann. 16.5 appears to contain the only other instance of postremo in the sense of ‘eventually’ to be found in hexameters of the archaic and classical periods, save perhaps Sil. Pun. 13.28.Footnote ¹¹ Lucretius has it seventeen times, but always in the sense of ‘last of all, finally (in a logical sequence)’ (OLD s.v. 1b).Footnote ¹² In other words, postremo is very close to being a poetical hapax legomenon. In any other poet that fact might render its conjecture an illegitimate proceeding, but the Sprachkunst of Propertius is a very different piece of work from the polished idiom of a Tibullus or an Ovid, and archaisms and/or colloquialisms of this kind are scarcely rare. To cite just one further ‘hapax’, at 3.24.17 the verb resipiscimus has somehow survived into the thirteenth century without getting corrupted in either branch of the tradition, though it is not elsewhere to be found in classical Latin verse. Further poetical hapaxes include argutat at 1.6.7, uenumdata at 3.19.21 and sirpiculis at 4.2.40 (elsewhere only at Columella, Rust. 10.1.1.305), to name just three.Footnote ¹³ Between the evidence of Enn. Ann. 16.5 and Propertius’ penchant for occasional colloquialism/archaism, and in view of the generally gnomic quality of vv. 25–6 (which certainly lends itself to an old-timey parlance), postremo here is established on perfectly good linguistic ground, and seems (at least to me) to be as close to a solution to this crux as we are likely to get. For it also carries with it a plausible explanation of the garbled text of N. Assume that the corruption of postremo into post obitum was obvious to the scribe that made it and prompted him to write -remo above or in the general vicinity of obitum; it is possible then to explain fame … uetuste as a mishandled attempt at effecting this correction. The diligent scribe of N saw remo in his exemplar, and bethinking himself of 2.3.41 si quis uult fama tabulas anteire uetustas, corrected the verse according to his honest if defective interpretation of the ductus.Footnote ¹⁴ The rest of the manuscripts, ΠΛ, simply ignored what appeared to be the meaningless ablative or dative form of the Latin word for ‘oar’. Or, in other words,

became in Ω

which in N was made

and in ΠΛ remained

Such – upon my theory – was the genesis of these variants. For another example of ‘corrections in Propertius gone wrong’, compare the corruption of 3.5.35 plaustra Bootes by N to flamma boon and by F to flamma palustra, the latter of which is evidently a mislaid correction of flamma back to plaustra.Footnote ¹⁵ In this instance F's mistake is telling: if its testimony were all we had to go on, how many editors would dare to print the emendation plaustra Bootes?

3.4.11–18

I present these lines in the order in which they are given by the manuscripts, but concede that some verses may be out of place and others perhaps are missing.

Verse 18 presents two problems of interpretation, the first of which arises from subter, the second from captos. With subter it is the word order that causes issues: the adverb at first sight appears to go with captos, but because the ‘spoils’ of captive chieftains were normally borne above them, not below, editors take it with arma. On this Butler and Barber have the following to say: ‘subter governs arma; but the distortion of the natural order is not merely harsh, but ambiguous, since subter should prima facie govern captos … duces’.Footnote ¹⁶ ‘Ambiguity’ is not really the right word for this phenomenon. Prepositions normally go with the noun to which they are closest, and to understand them otherwise is to fight against the current of the syntax.Footnote ¹⁷ sedere moreover is a questionable verb to apply to captives in a triumph, for normally they are represented as being led in chains before the triumphator's chariot: Res gest. 4.3 in triumphis meis ducti sunt ante currum meum reges aut regum liberi nouem, Liv. 3.29.4 ducti ante currum hostium duces, 34.52.9 hostiae ductae et ante currum multi nobiles captiui obsidesque, 39.7.2 duces hostium duo et quinquaginta ducti ante currum. It is conceivable that Propertius envisaged the chieftains as sitting upon a vehicle in the procession with the trophies situated high above them;Footnote ¹⁸ but if that is what the poet had in mind, one wonders why he would choose to signify it by subter, which at first sight gives the very opposite of his meaning, instead of supra, which gives it exactly.Footnote ¹⁹ If supra is indeed what Propertius wrote, then arma is the subject of sedere (OLD s.v. 6a ‘(of things) to be placed, rest, sit, lie’), and the sense is simply: ‘… and the arms sitting above the chieftains’. This combination of sedere and super/supra is found elsewhere in Propertius, as at 2.30.8 et grauis ipsa super libera colla sedet, and the two are even placed in the same metrical seat as here by Martial 12.70.2 et supra togulam lusca sederet anus. The reverse corruption, of subter to supra, happens to occur at App. Verg. Cul. 155; and Edward Courtney makes a convincing case in favour of reading supra for subter at Cul. 75, which error he explains as arising from the likeness of sup̄ (supra) and sup̄t (supter).Footnote ²⁰ The same corruption appears to have happened at Pers. 3.39–40 et magis auratis pendens laquearibus ensis | purpureas subter ceruices territus, where the emendation of Sandford (Reference Sandford1890), supra for subter, is well supported by Hor. Carm. 3.1.17 destrictus ensis cui super inpia | ceruice pendet, Quint. Decl. min. 305.14 non licet tibi gladium supra ceruices diu tenere, and Sil. Pun. 7.702–3 ferrumque super ceruice tremiscens | palluit infelix subducto sanguine Maurus. I turn now to consider the second difficulty of this line, the repetition of captos after capta in v. 16.

captos in v. 18 is a vacuous epithet: the reader knows from the context that these chieftains are ‘captive’, and to state the fact over again serves no discernible purpose. In v. 16 however the adjective is exquisitely apt: by oppida capta the poet intends to signify the simulacra of conquered cities carried in the procession; ‘captured’, therefore, in both a literal and figurative sense.Footnote ²¹ In default of a compelling excuse for the repetition, a repetition so obvious that Propertius could hardly have overlooked it, one must assume that captos is an error for another epithet or participle, probably uinctos. The chieftains are nowhere else in this poem represented as ‘enchained’, but it is an image so common to descriptions of triumphs that its absence here is keenly felt. Only a handful of witnesses need be called to make the point; I start with Cic. Pis. 60:

Tib. 1.7.5–6:

Ov. Ars am. 1.215–16:

Ov. Tr. 4.2.19–22:

And especially Ov. Pont. 3.4.101–6:

To the last of these parallels it may not be amiss to add, in view of Cic. Pis. 60 uincti ante currum duces, Tib. 1.7.6 euinctos … duces, Ov. Ars. am. 1.215 duces onerati colla catenis, Pont. 2.1.3–4 totque tulisse duces captiuos addita collis | uincula, that uiros in v. 104, colourless and unspecific, is, I conceive, probably a mere banalisation of duces, occasioned by the similar sound and appearance of uinctos preceding. Precisely the same mistake occurs at Verg. Aen. 8.120 and Ov. Tr. 2.365, where duces is in both places reduced to uiros by certain aberrant manuscripts.Footnote ²³

Returning to the matter at hand, the same parallels aforecited tend to support the reading of uinctos in Prop. 3.4.18. uinctus, uictus and captus are often confused, as at Prop. 2.10.18 capta] uicta, Ov. Met. 14.299 uictos] captos, Tr. 3.10.61 capta] uincta.Footnote ²⁴ Further examples include Ov. Pont. 4.13.45 sic capto Latiis Germanicus hoste catenis, where M. Akrigg has convincingly emended capto (meaningless with catenis for a modifier) to uincto, and Tib. 1.1.55 me retinent uinctum formosae uincla puellae, where uinctum is a corruption of Postgate's captum owing to uincla at the end of the line (cf. Hor. Carm. 4.1.38 iam captum teneo).Footnote ²⁵ But to go on adducing further examples here would be supererogatory, for captos in v. 18 is plainly due to the reminiscence of capta in v. 16.Footnote ²⁶

3.8.11–18

Rather than present these lines with Butrica's grege si (thus Heyworth), I have retained gregibus, albeit with an obelus on either side of it.

In the absence of a seu or a si, one naturally interprets v. 13 as the apodosis of a conditional completed by the alternative protases of vv. 14–16: ‘She surrounds herself with troops of protectors, whether she dashes down the thoroughfare like an inspired Maenad, or is driven mad by frightening dreams, or is moved to melancholy by the portrait of a girl.’ This interpretation is clearly erroneous, but it is one of only two that the paradosis will admit of. The other is to take v. 13 as a sentence by itself (and thus to punctuate with a full stop after euntem), and to construe vv. 17–18 as the apodosis of a conditional whose protases consist in vv. 14–16. The reader is constrained to choose between these two unhappy alternatives because, as E. Wistrand has pointed out, in Latin one cannot coordinate two full secondary clauses by a single seu between them.Footnote ²⁷ And the reason for this hardly needs stating: without a si or seu to mark it, the first protasis of a conditional that contained more than two further protases would be indistinguishable from its apodosis. Paolo Fedeli appears to miss this point when he adduces for the omission of the initial seu Prop. 2.6.33–4:

Citing this in defence of v. 13 is to confuse the issue. Nobody questions whether a single seu or siue may satisfy for a simple disjunctive coordinator of noun, adjective or verb phrases.Footnote ²⁹ Lines such as Hor. Carm. 1.3.15–16 quo non arbiter Hadriae | maior, tollere seu ponere uult freta and Sat. 2.8.16–17 Albanum, Maecenas, siue Falernum | te magis adpositis delectat, habemus utrumque are proof enough of that. The question is whether a single seu can link alternative conditional clauses (ideally without causing any ambiguity of sense), such as vv. 13–16 must be. In the absence of a direct parallel, this must be assumed impossible.

Most modern editors accept this analysis of the problem, and nearly all of them print a conjecture for gregibus of NΠΤ, gregi of SΓ and gregis of Y (curious variants in the Λ branch, mind). G. P. Goold (Reference Goold1990) and S. Viarre (Reference Viarre2005) adopt Heinsius and Postgate's correction, grege seu; Heyworth prefers Butrica's grege si on palaeographical grounds.Footnote ³⁰ grege seu or grege si is clearly what the sense requires, but the rhythm of both seems rather odd. There are ten other places in Propertius where at the start of a line I find a molossus or a choriamb followed by a pyrrhic followed by a monosyllable:

Four of these (1.4.13, 3.16.9, 3.22.25, 4.4.1) have et in the crucial position; the rest Propertius could not have arranged otherwise without either employing different vocabulary (2.5.5, 2.29.1, 2.29.17, 2.13.41, 3.11.51) or giving undue emphasis to an unemphatic word (4.10.41 Rhaeno genus hic). Here, in v. 13, there are not the same constraints. In fact, I really cannot conceive why our poet should wish to place custodum so far ahead of seu. Is it so important that ‘protectors’ be the first word of the line? Is seu ever trajected so far from its usual position? For what reason, when it was perfectly within his power, would Propertius not write simply seu grege custodum, and thereby attain not only the usual word order, but also the quadruple anaphora of seu? Would he have missed a trick like this? The author of CIL vi.21521, vv. 31–4, certainly did not:

A parallel for the trajection of seu over two words (or even over one word) is not forthcoming from the poetry of Propertius. Instead, the reader is confronted with scores of passages of the following shape (1.20.7–12):

Parallels could be multiplied.Footnote ³² Only J. S. Phillimore, displeased with the tautological function of circa, appears to have hit upon the right order with his conjecture siue ea custodum gregibus se stipat euntem. Closer to the text would be Heinsius’ emendation in reverse order: seu grege custodum circa se stipat euntem.Footnote ³³ Editors of Propertius are more comfortable with the transposition of couplets than of words within a couplet, but the phenomenon is common enough in Latin literature generally to present no difficulty here in point of palaeographical likelihood.

3.10.13–14

Propertius instructs his girlfriend to wake up for her birthday.

There is an ambiguity in the hexameter. ac primum must be taken as a unit, meaning ‘but first’. somnum however invites an epithet, and readers (this reader certainly) may be tempted to take primum as such a one. Commentators may try to keep them from error by pointing to the status of ac primum as a set phrase (cf. e.g. Verg. G. 3.166, Ov. Met. 11.20). ‘But so is primum somnum’, readers will say, quoting Prop. 1.3.3 qualis et accubuit primo Cepheia somno, Verg. Aen. 1.469–70 Rhesi … tentoria … primo quae prodita somno, Phaedr. 3.10.31 sopita primo quae nil somno senserat and Stat. Theb. 10.115 primosque hortantia somnos. Servius on Verg. 1.470 interprets this phrase primum somnum to mean the ‘beginning of night’, i.e. prima pars noctis; J. Henry with greater probability takes it to mean in every instance ‘the first part of sleep’, that is, ‘sleep having only just begun’. He illustrates this sense by quoting Sil. Pun. 9.90 ecce sub aduentum noctis primumque soporem, Verg. Aen. 2.268 prima quies and Valerius Flaccus 8.81 primi percussus nube soporis.Footnote ³⁴ Now Henry's interpretation of primus somnus accords perfectly with Prop. 1.3.3 qualis et accubuit primo Cepheia somno (primarily because of accubuit), but it makes havoc of 3.10.13 ac primum pura somnum tibi discute lympha. For how can Propertius’ girlfriend shake off ‘the beginning of sleep’ if the morning has only just begun (vv. 1–2 mane … sole rubente)? All these considerations taken together suggest that somnos should be read for somnum in v. 13. The grammatical ambiguity is then dissolved, the meaning is left unchanged, and a parallelism is established between somnos here and somnos in v. 25 dulciaque ingratos adimant conuiua somnos. For the plural in connection with dis-/excutere, compare Ov. Ib. 156 excutiam somnos uisus adesse tuos, [Sen.] Oct. 123 excutit somnos and Stat. Theb. 2.129 somnosque excussit.Footnote ³⁵ Errors of assimilation, whereby a noun or adjective assumes the case of an adjacent word, are some of the commonest in the Propertian tradition: 1.11.15 amoto] amota NATΓYC due to puella following; 1.18.10 puella] puellae Γ due to nequitiae… tuae; 1.20 Gigantei] Gigantea Ω due to ora; 2.27 ruinas] ruinam FLT due to flammam preceding.

3.11.39–46

Propertius speaks to Cleopatra. For this passage I give vv. 40 and 46 as they appear in Ω, without the emendations printed by Heyworth and other editors.

This long sentence is compact of several problems, and as there is a good chance of getting lost in the maze of them, it may help to give a summary at the outset: (1) the sentence beginning with v. 39 and ending with v. 46 is without a finite verb; (2) the scansion of dare in v. 46 as an iamb before statuas is unparalleled in Propertius and scarcely paralleled in elegy;Footnote ³⁶ (3) ausa in v. 41 governs five syndetically coordinated infinitives between vv. 41–5, but dare in v. 46 it governs in asyndeton (a discomposing irregularity); (4) the meaning of una Philippeo sanguine adusta nota in v. 40 is obscure and the syntax ambiguous. Problems (2) and (3) are usually answered by the inserting of <et> (ς) between dare and statuas in v. 46. But et produces its own problems and for various reasons is unlikely to be right. In the first place, as Heyworth remarks, the trajection of et beyond second position finds no adequate parallel in Propertius.Footnote ³⁷ Secondly, the use of et to connect one clause with another, while in that same clause an entirely different et is used to connect two nominal phrases, creates a certain amount of visual confusion for the reader, especially as the former of these two instances of et would have to sit directly in front the noun statuas, implying a correlative usage where no correlative is meant. In other words, the statement iura dare et statuas inter et arma Mari at first sight reads as if it meant ‘dispensing justice amidst both the statues and the arms of Marius’: a perplexing and unhappy inelegance to introduce by conjecture.Footnote ³⁸ Finally, the addition of et fails to resolve our initial problem: the sentence spanning vv. 39–46 remains without a finite verb, and in order to fashion one for it editors are compelled to supply either <est> (ς) or <es> (Heyworth) after nostro in v. 41. The loss of such a small word is not unimaginable, but neither is it easily accounted for. If a solution were found that could answer all three problems by a single change – the missing verb, the prosody of dare, the asyndeton of v. 46 – that would be much more attractive than the printing of piecemeal supplements such as <et> and <est> or <es>.

Such a solution was proposed in 1880, and editors have been sleeping on it ever since. Arthur Palmer, improving on the humanist emendation of daret for dare (46), conjectured dares, punctuating vv. 39–46 as a rhetorically charged question and assuming a ‘polemical’ use of the imperfect subjunctive in an interrogative sentence.Footnote ³⁹ The sense would be ‘harlot-queen of Canopus, did you really think (39 scilicet), daring to do X Y Z, that you would have dispensed laws amidst the statues and the arms of Marius?’ This use of the present and imperfect subjunctive in questions of a polemical colour is sufficiently well attested to make Palmer's conjecture a compelling answer to our three problems.Footnote ⁴⁰ dares for dare is the slightest of changes (s lost by haplography before statuas) and gives greater force and purpose to some of the other elements of vv. 39–46, such as scilicet in v. 39, termed ‘feeble’ in its present state by Heyworth.Footnote ⁴¹ The result is a long, suspenseful period with a barbed conclusion: Cleopatra, says Dio 50.5.4, really did conceive a desire to dispense justice on the Capitol (ποιεῖσθαι τὸ ἐν τῷ Καπιτωλίῳ δικάσαι), and to be denied this opportunity must surely have rankled in her heart. Finally, if Palmer's dares is accepted, meretrix regina Canopi etc. in vv. 39–40 is then allowed to serve as the vocative which fugisti in v. 51 presupposes but does not currently have.Footnote ⁴²

I turn now to consider the text of v. 40, the difficulty of which has long been acknowledged, but never satisfactorily answered. Butler and Barber list three possible interpretations of the appositional phrase una Philippeo sanguine adusta nota:

(1) is difficult for a further reason beside the absence of a dative in Romae: namely that Propertius has just called the assassination of Pompey a notam itself four lines ago (36 tollet nulla dies, hanc tibi, Roma, notam), and so to call Cleopatra ‘the one mark [of disgrace] branded upon Rome’ verges on the contradictory.Footnote ⁴⁴ Shackleton Bailey and editors since have therefore preferred (3), primarily on the ground that, though Cleopatra may not have been ‘the only disreputable Lagid’, yet una elsewhere can mean something like praecipua, and so the meaning ‘the one disreputable Lagid above all others’ remains possible.Footnote ⁴⁵ Still, una … nota remains an unhappy choice of words after notam in v. 36; and the fact that adusta ordinarily requires a dative to govern, not an ablative, and the fact that inusta nota, not adusta nota, is the established idiom, do not help to cement one's confidence in the text. Moreover, is it not a curiously mixed metaphor that has the ‘blood’ of Philip ‘branded’ by Cleopatra's misdeeds? And why should Propertius be particularly concerned with the honour and the good breeding of the line of Philip anyway?Footnote ⁴⁶

Preference for any interpretation of the text, be it (1), (2) or (3), appears to have arisen rather out of resignation at the difficulty, than from any genuine confidence in the meaning of these words.Footnote ⁴⁷ Emendation is therefore attempted by Phillimore, who adopting Lachmann's transposition of vv. 36 and 40 rewrites the line as illa uelim proprio sanguine adusta natet; Heyworth with little more deference to the ductus prints una Philippei sanguinis usta nota, while accepting that ‘the corruptions are not easy to explain’.Footnote ⁴⁸ When one observes that nota has been repeated from the end of tollet nulla dies hanc tibi, Roma, notam (36), rather like a Perseverationsfehler, it will be seen that these conjectures are targeted at the wrong part of the line.

The first step towards a restoration is to recognise that Philippeo does not here mean ‘of Philip II of Macedon’, and that interpreters have gone very far astray in thinking that it does. Philippeo sanguine refers not to the Macedonian blood that runs in the Lagid dynasty; rather it refers, as at Manilius 1.909, to the blood spilt at the battle of Philippi.Footnote ⁴⁹ The carnage of this engagement is often mentioned in poetical descriptions of the Roman civil wars, frequently before Cleopatra and Actium are themselves introduced. Compare for instance Manilius 1.907–19:

Compare also Ov. Met. 15.822–8:

If ‘by the bloodshed of Philippi’ is the sense of Philippeo sanguine, then adusta cannot stand: one is not ‘burnt’ or ‘branded’ by the spilling of soldiers’ blood. Whatever participle is to replace it must be one that can accept sanguine as an ablative of the instrument; and the only word that can serve this end and look anything like adusta (i.e. aduſta) is, I believe, adulta, in the sense of ‘nourished’ or, by a metaphorical extension, ‘strengthened’ (cf. OLD s.v adolesco 5).Footnote ⁵⁰ The collocation of this with sanguine is supported by an allusive couplet of Ausonius, 12.19.1–2 Green (from the Epitaphia heroum):

‘Get thee hence, unknowing stranger, and shun that myrtle tree: the battlefield was nourished with my blood.’Footnote ⁵¹ The probability that Ausonius is here consciously or unconsciously repeating a phrase of Propertius’ is increased when one observes that certain other of his verses do the same: compare e.g. Auson. Parentalia 21.6 maesti carminis obsequio and Prop. 1.8.40 blandi carminis obsequio; Auson. Prof. Burd. 1.28 fundunt excisi per caua buxa gradus and Prop. 4.8.42 iactabat truncas ad caua buxa manus.Footnote ⁵² In any case, whether the echo be intentional or not, it affords good evidence for the Latinity of Philippeo sanguine adulta, ‘nourished’ or ‘strengthened by the bloodshed at Philippi’.Footnote ⁵³ (adulta moreover recalls adultera, an epithet appropriate of Cleopatra as Propertius imagined her, given meretrix in v. 39.) Now if, as seems necessary, adulta be accepted for adusta, then nota must be altered as well: neither a brand nor a stain can be ‘nourished’ or ‘strengthened’ by the blood of Philippi. Strike it out and you have a line which runs as follows:

What part of speech could end a pentameter like this? The possibility of a finite verb is excluded by two considerations: (1) as Palmer's dares is required to resolve the metrical and syntactical problems of v. 46, printing a verb at the end of v. 40 would split vv. 39–46 into two sentences and so deprive dares of an explicit subject (it would have to depend merely on ausa in v. 41); and (2) the only preterites of iambic prosody in Latin are bibit, dedit, fidit, fluit, fuit, ruit, scidit and stetit (none of which but fuit makes very much sense with adulta), but as the second person is required so as to make incestri meretrix regina Canopi the vocative which fugisti in v. 51 looks back to, no perfect or imperfect verb (excepting dabas) is metrically possible. So, our missing disyllable is probably not a finite verb. An adverb likewise is unpromising: una … adulta would then have to depend entirely on meretrix regina, and no adverb that I can think of is so good as to compensate for the imbalance. A present participle or adjective, unless it were a very apt one, would be superfluous next to Philippeo sanguine adulta, so it seems Propertius did not write one of these either. The final word of v. 40 then can only be a noun: singular and feminine in agreement with una, standing in apposition to incesti meretrix regina Canopi, and expressive of a relation to the battle of Philippi (and/or Pharsalus) that is either peculiar to Cleopatra (una = unica, OLD s.v. 7 ‘the one and only’) or more suited to her than to any other (una = praecipua, OLD s.v. 8 ‘one in particular’). The constraints of metre, grammar and sense limit the options to a very small number indeed.Footnote ⁵⁴ Out of all the possibilities, lues, ‘scourge’, leaps out as one promising solution.Footnote ⁵⁵ This word is frequently used in a pejorative sense of odious persons, including such threatening individuals as Hannibal (Sil. Pun. 16.622–3 non dira illa lues notis iam moenibus urbis | adsiliet). Cicero likewise combines pestis, a synonym of lues, with adulta (in the sense ‘so far advanced’) to describe the person of Catiline: Cat. 1.31 delebitur non modo haec tam adulta rei publicae pestis, uerum etiam stirps ac semen malorum omnium. However, though as a stand-alone description of Cleopatra lues works quite well, it may be doubted whether ‘scourge’ goes quite so well with una and adulta as some of the other possibilities. nurus suggests itself as a lookalike of nota, but while Cleopatra could well be described as ‘the one married woman who profited by the bloodshed at Philippi’ (the other wives having lost their husbands in the fighting, Rome her soldiers in a futile civil war), adulta does not seem quite the right word to convey this sense, nor does ‘married woman’ square neatly with meretrix in the previous verse. The only disyllabic noun that can adequately describe Cleopatra by herself, as well as complete the metaphor suggested by ‘growing fat on the blood(shed) of Philippi’, is, I believe, canis.Footnote ⁵⁶ ‘Dog’ has a very long history of use as a term of reproach in Latin literature as well as Greek. In scenic verse as well as prose it is hurled without qualification at odious persons: Ter. Eun. 803 ain uero, canis?; Petron. 74.10 ultimo etiam adiecit: ‘canis’. The implication of it is sometimes one of shamefulness, as in Catull. 42.16–17 quod si non aliud potest, ruborem | ferreo canis exprimamus ore ‘But if nothing else can be done, let us at least squeeze out a blush from the bitch's face.’ In Prop. 1.11.40 that association would be appropriate after incesti meretrix … Canopi, but also annexed to this word is the idea of boldness, derived from the target's ‘latrandi, mordandi, sim. libid[o]’, as O. Prost puts it (TLL iii.258.21–2): Plaut. Bacch. 1146 est illic clam mordax canis, Men. 837 me ab laeua rabiosa femina adseruat, 933 uxorem suam esse aiebat rabiosam canem. In Cleopatra's case a connotation of ‘audacity’ would lead well into the accusation that follows in v. 41, ausa Ioui nostro latrantem opponere Anubim ‘daring to oppose barking Anubis against our Jupiter’.Footnote ⁵⁷ Nor would she be the only famous woman to be so denominated. Of Hecuba, for instance, Plautus (Men. 717–18) has Sosicles say, omnia mala ingerebat quemquem aspexerat: | itaque adeo iure coepta appellari est Canes.Footnote ⁵⁸ For Cleopatra the insult may have a special relevance, owing to the prominence in the Roman mind of ‘dog-gods’ in Egyptian culture: Prudentius, Contra Symmachum 2.354–5 quasdam uicta dedit capitis Cleopatra canini | effigies; Luc. 8.831–2 nos in templa tuam Romana accepimus Isim | semideosque canes et sistra iubentia luctus; Juv. 15.7–8 illic aeluros, hic piscem fluminis, illic | oppida tota canem uenerantur [sc. Anubim].

Now by itself canis would be a satisfactory insult, but combined with una Philippeo sanguine adulta it takes on a further significance. By ‘nourished on the blood of Philippi’ Propertius presumably envisaged Cleopatra as one dog out of many (OLD s.v. una 8) waxing fat on carrion: a metaphor for the rising strength and opportunism of the East, growing in proportion as Rome is withered by civil war. This is a commonly expressed anxiety in Roman literature; the locus classicus is Hor. Epod. 7.3–10:

The idea of ‘presenting’ the nation's blood as a spectacle to foreign nations, whereby to feed the souls of the slain, is a topos often employed by Lucan: 1.9 gentibus inuisis Latium praebere cruorem, 39 Poeni saturentur sanguine manes, 6.310–11 Poenorumque umbras placasset sanguine fuso | Scipio. While an authority for describing Rome's eastern foes as ‘jackals’ may be wanting,Footnote ⁵⁹ the image of dogs and vultures seeking carrion is a common one: Cic. Mil. 33 cruentum cadauer … nocturnis canibus dilaniandum relinquisti; Liv. 41.21.6 cadauera intacta a canibus ac uolturibus tabes absumebat; Luc. 9.141–2 nam corpus Phariaene canes, auidaeque uolucres | distulerint; Juv. 4.111–12 et qui uulturibus seruabat uiscera Dacis | Fuscus. A very long list of prosaic and poetical descriptions of hounds which ‘laniant homines, devorant corpora’ is supplied by Probst, TLL iii.254.18–40; and a particularly relevant parallel for Propertius’ sanguine adulta canis discovers itself in Ov. Met. 3.140 canes satiatae sanguine erili ‘hounds sated with the blood of their master, Actaeon’. That adulta may serve for nutritae or satiatae is shown by Ausonius’ imitation, telorum seges est sanguine adulta meo, which too connects sanguine with blood spilled in battle. Here, in Prop. 3.11.40, the metaphorical sense of adulta must also be present (OLD s.v. adolesco 5 ‘to attain power, become strong’; TLL i.802.1–35), a meaning illustrated in connection with warfare by Minucius Felix 25.4 iam finitimos agro pellere, ciuitates proximas euertere cum templis et altaribus, captos cogere, damnis alienis et suis sceleribus adolescere cum Romulo regibus ceteris et posteris ducibus disciplina communis est, the essential part of which is translated by G. H. Rendall as ‘to wax fat on losses inflicted, and crimes committed’.Footnote ⁶⁰ What Propertius means in 3.11.40, in other words, is that the losses which Rome sustained at Philippi served to strengthen, relatively speaking, the hand of Cleopatra.

If nota is emended thus, it will be noticed that canis and Canopi are placed in ‘vertical juxtaposition’, an arrangement which in Virgil and the Alexandrian poets is often used to suggest a shared etymology.Footnote ⁶¹ It may be that Propertius intended the same idea: to etymologise the Egyptian city through one of Egypt's familiar associations.Footnote ⁶² The fact that Canopus was generally thought to be derived from Greek Κάνωβος, the pilot of the ship that carried Menelaus to Troy, by such authors as Statius and Silius Italicus does not pose an insurmountable obstacle to this theory.Footnote ⁶³ Propertius was free to draw connections between such words as he pleased, and in this particular matter he may not be alone. For a similar juxtaposition obtains in a passage of Statius’ Silvae, where Canopus and Cerberus (= Anubis) are placed in close connection (3.2.111–12): curue Therapnaei lasciuiat ora Canopi, | cur seruet Pharias Lethaeus ianitor aras ‘why the shores of Spartan Canopus wanton, why the watchdog of Lethe guards Pharian altars’. While not a sufficient reason in itself to emend nota, the fact that a vertically juxtaposed attempt at etymologising Canopus through canis would obtain if it were indeed emended is an interesting consequence.

To sum up, many thoughts and meanings are constellated in this line: (1) the long history (from Homer to Catullus) of canis/κύων as a reproach of notorious women; (2) the cultural association in the Roman mind between Egyptian monarchs and their ‘dog-gods’; (3) the conception of Rome's enemies as ‘jackals’ who profited by her self-inflicted wars; and (4) perhaps an attempt at ‘etymologising’ Canopi through Latin canis. Whether these various strands of thought make for a more interesting and more Propertian verse than the present reference to Philip of Macedon's ‘branded bloodline’ must be left for others to decide. But one thing may be added in conclusion. The Ptoemphani, a people who dwelled on a certain side of the upper Nile, were said to ‘worship a dog for a king’ (Plin. HN 6.192 canem pro rege habent).Footnote ⁶⁴ It would have been an easy bigotry for Propertius to have turned this curious fact into a contumely against Cleopatra, una Philippeo sanguine adulta canis ‘the bitch that grew fattest on the blood of Philippi’.