STOP Early Journal Content on JSTOR, Free to Anyone in the World This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in the world byJSTOR. Known as the Early Journal Content, this set of works include research articles, news, letters, and other writings published in more than 200 of the oldest leading academic journals. The works date from the mid-seventeenth to the early twentieth centuries. We encourage people to read and share the Early Journal Content openly and to tell others that this resource exists. People may post this content online or redistribute in any way for non-commercial purposes. Read more about Early Journal Content at http://about.istor.org/participate-istor/individuals/early- journal-content . JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people discover, use, and build upon a wide range of content through a powerful research and teaching platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please contact support@jstor.org. THE TEXT OF THE CULEX By J. S. Phillimorb The controversy commenced by Skutsch in his Aus Vergils Friihzeit does not flag: not a few of the greatest living German scholars have made contributions to it, which, whatever eventual verdict shall decide the question of authorship, have at least done much for the elucidation of a very interesting body of poetry. This paper is not directly concerned with the main question, but I must in a few sentences summarily premise one or two points to define its relations to the main question. 1. I have no doubt that Virgil wrote the Culex. Against the arguments briefly outlined by Skutsch in an appendix to Aus Vergils Friihzeit and most conveniently and completely displayed by VoUmer in his pamphlet Die kleineren Gedichte Vergils (Miinchen, 1907), there is not and there never has been any serious objection except the plea that the Cvlex is unworthy of Virgil. That Virgil could not and did not write any juvenilia unworthy of the author of the Bucolics, the Georgics, and the Aeneid seemed good logic to the phi- losophy of the last century. But nowadays such an assertion as this (which may be read in Teuffel-Schwabe Engl. tr. § 230.2), "nothing is left us but to assume that Lucan, Martial and Statins were mistaken in identifying the extant Culex with the one written by Virgil," only survives as a curious monument of the fatuity of idealist criticism. But though Virgil's authorship is not disproved by the apparent unworthiness of the poem, the ease for the defense may be strength- ened if by amending systematically within the rules of common critical probability, the text can be brought nearer to the measure of technical accomplishment which we expect of Virgil even in youth. Many people will probably confess that at a first reading and for many readings after, the Culex appeared to be a jungle of solecisms and inconsequences; absurdly constructed and often uncouth in meter. Yet there were glimmerings of beauty in it sufficient to move [Classical Philology V, October, 1910] 418 The Text of the Cvlbx 419 Spenser to do it into English. And I am convinced that the pallor and flatness, the strange futility as of a bungling artist who cannot get a thing said although he keep talking round it — all this means not that the Culex is a bad picture, but that the picture is sorely in need of cleaning. 2. I premise then that the Culex is a gravely, exceptionally corrupt text. Though the Bembinus is a MS of the ninth century, this early date is no guarantee for the textual integrity of any work unprotected by either of the two great preservatives — ecclesiastical interest (e.g. the Bible and the Fathers), and the usage of the schools during the period of the fifth to the ninth centuries (e.g. Aeneid or Terence). To prove the point that we have to do with a deeply corrupted tradition, it is enough to instance almost at random: e.g. 330 BP gives iam oicon as; S pamoicon es; the correct iam Ciconas is only in V and Cors. In 332. V and Vossianus alone preserve any traces of the true reading Zanclaea; Bembinus and the rest give metuenda. In 324 for Hectoreo victor lustravit corpora Troiam all our MSS give Hectora lustravit victor de corpora Troiam which shows not honest corruption but bad patching, and so on: instances are innumerable. In a good text such as Virgil or Terence it is folly to multiply conjectures; in a moderately good text such as Propertius, the onus probandi is strong on each conjecture; but in a text as bad as I believe the Culex to be, and starting from the axiom that Virgil is the author, I venture to think we may a priori plausibly redress the poem to conform metrically and in style of language to a period between the Lucretian-CatuUine and the matured Virgilian. That is to say, such broken-backed verses as: perfidiam lamentandi mala : perfide multis and iret inevectus caelum super: omne propinquo and mente prius docta fastidiat et probet illi .... stand self-condemned, inviting emendation. No further preliminary apology is needed for submitting in the briefest possible form a long 420 J. S. Phillimore series of suggestions. If in even one or two of them I can convince the reader that emendation gives Latinity and coherence to what was solecistic nonsense, he will, I hope, admit the probability that the Culex only needs more emendation to be restored to coherence and Latinity throughout. 1-41: The structure of this exordium suggests that to vss. 11-23 Latonae magnique lovis decus, aurea proles containing the double invocation of Apollo and Pales, there were afterward added vss. 1-10 and 24-41, to serve some occasional pur- pose of the author's. I suppose he had been praised to Octavian as a poet and was pressed to produce something which he might offer to the venerandus sanctus -puer:^ he had the Culex in his desk and furnished it with a suitable new headpiece. The style smacks of the rhetorical school: e.g. the conclusion Sed nos ad coepta feramur 41 is not unlike Auctor ad Herennium (Lib. iii. praef.): nunc tu fac attentum te praebeas: nos proficisci ad instituta pergemus 3-6: These lines are an old crux. Leo quotes from Manil. ii. 522 a valuable illustration of consonet ordo. The following passage from Macrob. Sat. v. xiv. 11 is worth quoting too: item divinus tile vates res vel paulo vel multum ante transactas, opportune ad narrationis suae seriem revocat, ut et historicum stilum. vitet, nan per ordinem digerendo quae gesta sunt, nee tamen praeteri- torum nobis notitiam subtrahat. The Culex is mock-heroics; the "detailing of the argument" {ordo notitiae) is to "harmonize" with the manner of history, " bur- lesque-wise" {per ludum). And now follows a corrupt bit: The mock-solemn quisquis erit culpare iocos Musamque {read lusumque) paratus, pondere vel culicis levior famaque feretur; makes one expect an equally mock-solemn "warning off " of unworthy hearers, the profani: ducumque voces (Med.) "\ ductum voces (V) V licet invidus adsit. ducum voces (codd. ceteri) ) ' Leo cites examples to support sanctus. Venerandus is applied to Alexander the Great by Q. Curt. iii. vi. 16. The Text of the Culex 421 I suggest that on the analogy of Caesar dum canitur, quaeso, Juppiter ipse vaces Prop. iv. vi. 14. and (perhaps) mendax fama, vaces (Housman) ibid. iv. ii. 19 (vaces DV noces N voces PL) we should here read: notitiae. Doctrina, vaces licet: invidus absit. A truce to serious Art, and A vaunt Envy! For the use of licet in polite invitation cf. Prop. i. viii. 29; ii. xxii. 23. This would make Baehrens' conjecture in vs. 9 docta graviore sono tibi Musa loquetur more pointed. The detestable rhythm of the line as it stands might be corrected docta: dabunt secura suos mihi tempora fructus. This at least falls pat enough with the eventual facts of Virgil's life. Vss. 20-22 et tu Sancta Pales ad quam ventura recurrit agrestum bona secura sit cura tenentis fetura sit cura tenentes sors cura secura tenensque sors securaque cura tenentem Such, with minor variations of detail which may be studied in the apparatus of Professor Ellis or of Ribbeck, are the data for this problem. Read: et tu Sancta Pales, ad quam tutela recurrit agrestum, bona sis : tecum sit cura tenentis aerios nemorum cultus siluasque virentes. Tutela is, I think, an old conjecture, though I cannot find an author for it. Bona sis will recall to any reader sis bonus o! felixque tuis. For tecum tenentis I adduce Prop. ii. xxx. 25: libeat tibi, Cynthia, niecum rorida muscosis antra tenere iugis. quod si forte tibi fiierit fatorum cura meorum A.L. epigr. 965. 422 J. S. Phillimobb Virgil invites Pales to be kind (sit euro) to one who shares her haunts. The next line confirms it: te cultrice vagus silvas feror inter et antra. 24-41. vs. 24: et tu cui meritis oritur fiducia cartis BV chartis T castis Cant, canis Q Read tantis. Chartae have nothing to do with the case; it is great deserts which raise a high confidence in Octavius. Vss. 35, 36: moUia sed tenui pede currere carmina versu versum Most editors have adopted Heinsius' decurrere, but even his great authority does not reconcile me to the probability of an original decurrere being altered into pede currere by dittography. Look at the next verse: vlribus apta suis Phoebo duce ludere gaudet gaudent Of what subject can this most naturally be predicated? Surely of the Muses sporting with Musagetes Apollo. Read: moUia sed tenui pede currere: carmina Musae viribus apta suis Phoebo duce ludere gaudent. The Muses like playful themes appropriate to their strength, so let slender feet have soft ground, i.e. dainty subjects, to run upon. Similarly Propertius says: mollia sunt parvis prata terenda rotis (iii. iii. 18). Vss. 37-40: hoc tibi sancte puer memorabilis et tibi certet gloria perpetuum lucens mansura per aevum et tibi sede pia maneat locus, et tibi sospes debita felicis memoretur vita per annos. Considering: non frustra meditantur, habent memorabile quod sit in CatuU. Ixii. 13, 1 am sorry Professor Ellis either altered memorabilis et, beyond the almost inevitable change to msmorabile sit, or did not The Text of the Cvlex 423 conjecture meditabimur outright. Accepting Buecheler's sit tibi certe, I would read: hoc tibi, sancte puer, memorabile; sit tibi certe gloria perpetuum Culicis mansura per aevum ' and in vs. 40: debita felicis numeretur vita per annos. Vss. 42-57 : There is much room for suspicion in many of these verses, but corruption can hardly be doubted in vss. 48, 49 : iam silvis dumisque vagae iam vallibus abdunt corpora iamque omni celeres e parte vagantes Perhaps iam silvis dumisque fuga est, though the ablative is rather harsh. Vs. 50: How can a goat's bite, which poisons and destroys vegetation, be called tenerf And the recurrence of morsu in vs. 54 increases suspicion. Eclogue vii. 6, dum teneras defendo a frigore myrtos suggests tondebant tenerae viridantia germina myrtus. Vss. 58-97: Praises of the country life. This passage is con- fessedly a reminiscence of Lucretius ii. 23-36, and in close relation — parental, say Skutsch and VoUmer; filial, say the opposing school — to Georg. ii. 458-74. How close exactly is the correspondence in the detailed articulations of the two passages? The answer to this question is a main factor in the problem of emending the Culex text in some of these surely corrupt lines. The scheme of Georg. ii. 458-74 is as follows: O fortunatos . ... si ... . worm< agricolas! quibus .... fundit .... victum .... tellus si non .... domus .... salutantum vomit undam, nee .... inhiant .... postes .... inlusasque .... vestes nee fucatur lana .... nee corrumpitur usus olivi; at secura quies ete at otia, ete. non absunt: illic saltus ae lustra, et . . . . iuventus, sacra .... sanetique patres: per illos lustitia .... vestigia fecit. 1 lucens recurs in vs. 41 and vs. 74. 424 J. S. Phillimoee The scheme of Culex 58-78 is as follows (prescinding from 79-97, the materials of which have been licked down and condensed so that Cul. 79-85 = Georg. 459-60 and Cul. 80-97 = Georg. 467-73): O bona pastoris siquis si non .... fuerint si nitor non .... angit .... picturaeque decus .... nee .... manet .... nee referunt .... nee pretio est at prosternit corpus .... atque ilium illi sunt That is to say, the two agree in exclamatory invocation: Cul. has 2 si non +3 nee + at + ilium and illi; Georg. has 1 si non + 3 nee + 2 at +illic and illos. So close a correspondence as this certainly favors an indicative verb in vs. 62: si non Assyrio j fuerint > ^.^ j^^^^ ^^ ( iiunt ) where Ellis conjectures fervent, Schrader fulgent. I suggest that /enwni is the right word; as an anagram it accounts for fuerint, as expressing an effect produced on the spectator it matches with animum angit avarum in vs. 64. Cicero (Brut. 67.226) uses ferire absolutely, for "to strike " some object readily understood from the context; here it would merely mean the striking color of the Bi^a^a "which only an Attains' purse can command." But is it not really a yet closer correspondence than appears in our texts? O fortunatos .... agxicolas O bona pastoris .... (the noun is missing). Where is it to be sought? Rhythm, the clue of the Georgic parallel, and the presence of a weak and obscure point in the phrase as it stands — all these indications point to omnia luxuriae spretis incognita curis omnia = oia readily allows the restoration of what the sense also claims: otia. O bona pastoris .... otia. The Text of the Culex 425 Next to compare the reserves which qualify the respective ejacu- lations : sua si bona norint says the Georgic; and the Culex (58-59) O bona pastoris (si quis non pauperis usum mente prius docta fastidiat et probet illis) otia Several things excite suspicion here: (1) for siquis non fastidiat we require siquis non fastidit; (2) the vicious rhythm of et probet; (3) the vagueness of mente prius docta "a previously learned mind"; (4) the inelegance of pauperis standing as noun so near to pastoris; (5) the clumsy pair of verbs fastidiat et probet; (6) the pointlessness of illis. Five of these objections are removed by reading si quol non pauperis usus fastidia praebet ovilis. A remedy for mente prius docta is not so simple; but supposing menteprius to be an anagram for temperius, the small consequential change of docta into docto gives us a plausible Latin phrase to express "precociously learned," the contrary of o^ifiaOi^^ as it were a irpaniiaOrji;. O the jolly idleness of a shepherd — unless precocious learning puts a man out of conceit with the business of the humble sheep fold: O bona pastoris (si quoi non pauperis usus temperius docto fastidia praebet ovilis) Vs. 60: otia, luxiuiae spretis incognita curis quae lacerant avidas inimico pectore mentes ! mentes in vs. 61 now becomes unexceptionable, whether Leo's defense of inimico pectore be judged adequate, or Ellis' beautiful conjecture nimia cuppedine be preferred. Vs. 95 : I fear the pretty word tempe is only a mistake for semper, like the sempe cited by Ribbeck in vs. 265. If so, the puzzling fontis represents a proper name of some nymph. It might be, e.g., . . . . o gratissima semper Lotis Hamadryadum. semper gratissima is a Propertian endearment (i. ii. 31). 426 J. S. Phillimoee Vs. 107: iam medias operum partis evectus erat sol cum densas pastor pecudes cogebat in umbras ut procul aspexit luco residere virenti, Delia diva, tuo, quo quondam, etc. Satis turbata haec says Sillig. Professor Ellis holds that the protasis which opens at vs. 109 only finds an apodosis at vs. 158. A long sentence ! At vs. 109 perhaps : haut procul a speculis, luco residere verentes The flock are shy of entering the divinity-haunted grove; so for their siesta the shepherd drives them only a little way down from the hilltop where they have been feeding. They rest and sleep among the thickets of the open hillside {dumis of vs. 155); he enters the mystical lucus. Vss. 114-20: et Satyri Dryadesque chorus egere puellae naiadum coetu tantum non ] , , [ Hebrum ( horpneus ) restantem tenuit ripis silvasque canendo, quantum te pernigre morantem diva chorea Emendators of these lines have taken a false direction, I think, by not recognizing that Virgil throughout the Culex declines Orpheus as a Greek noun: gen. Orpheos, vs. 269; dat. — as I hope to show — Orphei, vs. 279; voc. Orpheu, vs. 292. The slightest of changes will restore sense to these lines: Naiadum coetu laetum non Orpheos Hebrum ars tantum tenuit ripis silvasque canendo quantum te, etc. Not Orpheus' cunning so captivated Hebrus who rejoices in his com- pany of Naiads, as these captivate Diana. Vss. 131-33: posterius cui Demophoon aeterna reliquit I periida \ perfidiam lamentandi mala < perfidia > multis ( perfide ; perflde Demophoon et nunc deflende puellis. Read: posterius cui Demophoon aeterna reliquit perfidia lamenta "Audi, male perfide! multis " perfide Demophoon, et nunc deflende puellis!" The Text of the Culex 427 Audi etc. are the words of the lament. Some allusion escapes us in this verse. One of the contemporary poets used Demophoon as his nickname; Propertius addressed ii. xxii to him; and this line of the Culex is surely echoed in Prop. iv. vii. 13 : perfide, nee cuiquam melior sperande puellae. For vs. 119 Nodell's suggestion pernox deserves to be recalled, though pemocte would be better. Vs. 140: ilicis et nigrae species et laeta cupressus etfleta Ellis; nee laeta vulg.; lethaea Gifanius. Perhaps acuta; cf. Ov. Met. iii. 155. Vallis erat picels et acuta densa cupressu. Vss. 154, 155: at circa fessae passim cubuere capellae excelsisque ) super | j • i -i. Aa . ^ i t dumis quos leniter adnans excesisque 1 supra ) aura susurrantis poscit confundere venti. The goats are lulled by the whispering wind in the bushes. Read: excessitque sopor dumis, etc. Slumber issued from the bushes. Vss. 157-201: The Serpent's attack. Vs. 168: toUebant aurae venientis ad omnia visus. If so desperate a place may excuse a bold conjecture I would suggest: torrebant furiale minantis adonia visus "his glances of frantic menace scorched the (herb) adonium," southernwood, or whatever is the modem name of it. Vs. 174: metabat sese circum loca cum videt fingens adversum recubare ducem gregis. Even Virgil's notorious fondness for the word ingens can hardly justify it here. One MS reads anguis; Ellis conjectures hiscens. The context appears to require a word expressing rage: for instance amens, "beside himself." Amens looks like an "artlessly violent" conjecture, but there are two other passages in the Culex in which I hope to show that where the MSS offers us a word beginning with 428 J. S. Phillimobe ic- or ig- we must read a word beginning with a-. From these I infer that at some period our MS tradition went through a stage of some pre-Caroline script in which a- was so written as to be readily- mistaken for ic- or ig-. See notes on vss. 302 and 326. Vss. 176, 177: probably intendere and infringere are historic infinitives. Vss. 182-86: cui cunpta paranti parvulus hunc prior umoris conterret alumnus et mortem vitare monet per acumina. namque qua diducta genas pandebant lumina gemmis gemmas hac senioris erat mature pupula telo iota lev!, cum prosiluit, etc. Read cui cuncta parantur d.Aen. ii. 132: mihi sacra parari. The miserable rhythm of namque in the 6th foot and the clumsiness of phrase in vs. 185 are alike mended by reading: nam qua diduc exorabile mortis. ( divae ) Almost the only service that Baehrens performed for the text of the Culex by conjecture was to extract ius out of the diu of S. But he did not complete the emendation. This passage is one of those which Skutsch minutely discusses; he shows that magis, so far from being inept, is particularly pointed. For the argument runs thus: Orpheus' power had carried all before it, even Persephone, even Death, willy nilly consented; Eurydice on her part knew better than to run any risks, but it was Orpheus, Orpheus himself who rather than any of them deserved to be called crudelis, crudelis. That is haec eadem potuit Ditis te vincere coniunx; Eurydicenque ultro ducendam reddere non fas, non vetat invitae ius exorabile mortis. Vss. 294-95: I suggest graviter (si Tartara nossent) peccatum meminisse grave est. It is grievous to remember what was (had but Tartara understood it) a grievous fault. Vss. 301-3 : These lines, as the editors print them, make nonsense. 1) Assidet referens ignis is absurd. 2) Sociatae gloria sortis does not indicate the two men with whom the passage deals, the Aeacids, Achilles and Ajax. The Text op the Cvlex 433 Let me next observe that we have in all this digression an apiareia of Ajax, with only so much mention of Achilles as is barely necessary; and finally, that we have in these next eighty lines the densest growth of corruption in the whole Culex. assidet ( hac ) iuvenis ( sociat de i j I hoc \ } sociate > gloria [ fortis. ( sociat quem ) doria ) S alone keeps doria. Now if you take the letters aciuiieis, by adding three vertical strokes above the line at the points I have indicated by dots, you get achilleis — a first ray of meaning. Next suppose sortis to repre- sent a noun agreeing with achilleis — a presumption justified by the ordinary structure of a hexameter: .... Achilleis sociat .... doria sertis, and the line completes itself Cassida Achilleis sociat se doria sertis altera. A second helmet, a Dorian helmet, shares in Achilles' garlands (of victory). Compare ille secundus apex bellorum etproxima cassis. — Stat. Silv. v. ii.47. And now for the unmeaning referens — Ajax does not carry back the fires or anything else — let us take a cue from the mistake of the MSS in Manilius ii. 863 and try" inexcessumque furens a navibus ignis Argolicis Phrygios torva feritate What is the verb? And what is the subject of that verb? The MSS reading repulsos (or refvlsos) favors the notion that we ought to replace a second person singular, repulsas. If so, there is yet a vocative wanting. One word in the sentence so far reconstructed remains otiose: ignis. Ajax savagely repels the Phrygians, not necessarily their fires rather than themselves. Now if the reader will recall what I premised in speaking of vs. 174, and in there advocating amens for what the MSS gi\e as ingens; and further compare vs. 326 where the MSS reading in the last foot is icta, he will see that the one word required to make sense in vs. 302 and in vs. 326 is Aias or Aiax. (Perhaps the Greek form of the name is likelier.) 434 J. S. Phillimoke At first blush it is an "artless violence" to read Aias for igis and Aias for icta; but, if comparing vs. 174, aens for Igens, in all three instances a postulated confusion of a-* with ic- or ig- restores a significant and necessary word to the sentence, I submit that there is some cumulative probability in restoring these lines as follows: cassida Achilleis sociat se Doria sertis altera; inexcessumque furens a navibus, Aias, Argolicis Phrygios torva feritate repulsas. O quis non referat talis divortia belli, quae Troiae videre viri videreque Grai Teucria cum magno manaret sanguine tellus, et Simois Xanthique liquor, Sigeaque per te (praeter codd.) litora, cum, Troes, saevi vos Hectoris Ira truderet in classis, etc., etc. (ducts vulg.; duos codd., exc. V, Cors.) Vs. 314 classibus ambustis flamma f lacrimante f daretur. Editors have conjectured flagrante, superante, crepitante, lacerante, etc. Paleographically easier and better in sense would be bacchante. Vss. 322-26: hoc erat Aeacides vultu letatus honores Of these words, hoc erat .... honores bear a look of ultimate integrity while they argue corruption on the surface. Hoc erat cries out for a relative clause to complete the idiom. Read hoc erat, Aeacides, quo tu laetatus ovares Here was something for you Aeacid (Ajax) to exult over; and the other Aeacid exults because, etc. Vss. 325-26: rursus acerba fremunt, Paris hunc quod letat et huius arma dolis Ithaci virtus quod concidit icta. fremunt demands a subject, which is but thinly concealed in et huius, namely Achivi. ' The confusion may readily happen in the English or Irish roundhand of the seventh to the eighth centuries, where a is written cc. I may remark, by the way, that a similar confusion helps to account for the error in Statins Silv. v. iv. 14. at nunct heusaliquist longa sub noete puellae braechia nexa tenens Read nescioquis. It was the corruption of ci into a which gave rise to aliqwis. heus is of course absurd; heu, signis does not account for the MS reading; and heu, si aliquis gives an impossible elision of the monosyllable. The Text of the Culex 435 For ida in the next line I have abeady suggested that Aias is to be substituted. The couplet regains point and relevancy if we read rursus acerba fremunt, Paris hunc quod letat, Achivi ; arma dolens, ira victus quod concidit Aias Possibly Ithaci might stand. Vs. 330: the missing word may be monstrum atrox Laestrygona monstrum and perhaps it was mostru which helped to foist the ludicrous Molos- sis (molosis codd.) into the next verse, where marinis alone seems applicable. Vss. 337-38: reddidit heu Graius poenas tibi, Troia, ruenti, Hellespontiacis tibi reddidit obrutus undis is perhaps the secret of the disconcerting obiturus reddidit in which the first part seems to make plain nonsense. Vss. 340-41: neque propriae fortunae munere dives iret / inevectus ) , ,.,}■. . y caelum super omne propinquo tendit ) in evectus i frangitur invidiae telo decus. To break the sentence with a fuUstop at super gives a harsh and improbable construction; and the analogy of vss. 209 and 229 illo ah eventu and parilis ad eventus favors a reading in eventus with a contemporary propinquos. Perhaps ne quis ceu (or uti) propriae fortunae munere dives iret in eventus caelum super omne propinquos, frangitur invidiae telo decus. Lest any man rich in the favor of fortune, as though fortune were his own to count upon, go to meet future chances in heaven-outsoaring pride, glory is shattered by the weapon of (divine) envy. Vs. 348: undique mutatur caeli nitor, omnia ventis omnia turbinibus sunt anxia. Surely corrupt. Suppose the author wrote undique mutatis caligant aequora ventis, omnia turbinibus obnoxia. 436 J. S. Phillimoke TibuUus has insanis cautes obnoxia ventis (ii. iv. 10). For the lengthening of -us cf. opus 395, genus 400. Vs. 357: omnis in aequoreo fluctuat naufragia fluctu omnis in aequoreo fluitat iam naufraga fluctu (V. Voss.) omnis in aequoreo .... fluit atia naufrage luctu (Cors.) fluctuat omnis in aequoreo naufragia luctu (ceteri codd.) fluctu appears to be sound in the sixth foot; if so, fluere, fluitare, fluctuare would appear to be ineligible for the middle of the line. luctari (cf. Aen. v. 220; Catalepton xiii. 26; Senec. ad Marc. x. 6) fits the case : omnis in aequoreo luctatur naufraga fluctu. Vs. 363: Curtius et medius quem quondam sedibus urbis perhaps Curtius et quem diis quondam .... entibus urbis supplying some participle, e.g. poscentibus. Vs. 368: If Horatius Codes be the hero to whom this Une alludes one may read fluminibus devota dedit qui corpora flavis. Vss. 369-71: iure igitur talis sedes pietatis honores istarum piadasque duces quorum devota triumphis moenia \ ^^^^ '^ [ Libycae Carthaginis horrent. ( romanis ) There is strong presumption of a lacuna between 369 and 370. Surely the first five letters of 370, as it stands in most MSS, signify that the lost line ended with instar. In vs. 371 I see no reason to suspect romanis; the necessary change is more easily made in vs. 370. iure igitur talis sedes pietatis honores .... instar, Scipiadasque duces duo quos devota triumphis moenia Eomanis Libycae Carthaginis horrent. Cf. Lucr. iii. 1034. scipiadas, belli fulmen, Carthaginis horror. The Text of the Cvlex 437 Vss. 379-80: , \ immemor audis naec ) et tamen utuadis \ dimitte somnia / , . erramus utuada ) dimittere somnia ) It had occurred to me that the words faia minuta lay concealed in the debris with which vs. 380 begins; but the presence of somnia, which I believe to be a genuine word, forbids the introduction of another noun into the phrase. Perhaps the Gnat warns the Shepherd that this dream is a true and creditable dream, inrap ovk ovap (as indeed appears presently, for the Shepherd loses no time, once risen, in fulfilling its wishes), in fact a morning dream, matutina somnia. I take it that by the loss of the initial, matutina went into atutina and this again into a tri tua or a tri utua. Accepting matutina we have three possibilities haec immemor audes matutina vagis dimittere somnia ventis? or haec immemor audes matutina videns dimittere somnia ventis? or haec immemor audis ? matutina cave dimittere somnia ventis. Vss. 376-79: ergo quam causam mortis iam discere vite Terberibus saevae cogunt ab indice Poenae cum mihi tu sis causa mali nee conscius adsis sed tolerabilius cures. (cures Ellis; curis codd.). Read ergo quam causam mortis iam discere ni te verberibus saevae coguntur ab indice Poenae, cum mihi tu sis causa mali? etc. The index from whom the cruel Poenae are bound to learn the true cause of his death is the Gnat himself: What cause can I tell them but you, the shepherd, since you are the cause of my woe, and you do not assist me? i.e. he will reluctantly bring vengeance on the shepherd, conscius adesse is a Propertian phrase (I. x. 12). 438 J. S. Phillimore Vs. 338: sibi tribuere is found in Ov. Met. vi. 89. nomina summorum sibi qui tribuere deorum The meaning is "with as much energy as his aged powers could claim." Propertius varies it: quaecumque aderant in corpore vires (iii. xv. 23). Vs. 391: capit for incipit seems unexampled. A slight change gives us non formare locum carptim piget. Or possibly latenti .... capiti piget. "The undiscoverable person" of the dead Gnat, (a good example of locus — a tomb in Anth. Lat. (Riese) 1067 : immatura meo perlege fata loco). Vs. 392: "To serve the purpose of an iron (blade or tool) he had recourse to . . . . " what? Capulum say the MSS. Surely the bacvlum which the shepherd had leaned on innixus baculo (vs. 98). For the phrase cf. impositum est non comparentis in usum partis ebur (Ov. Met. vi. 410). Vs. 393 : gramineam vlridi foderet de caespite terram. iam memor inceptum peragens sibi cura laborem congestum cumulavit opus. Festus says Acerra: ara quae ante mortuum poni solebat. Memor cura is suspicious in view of curae memor in 398. One might read gramineam viridi ut foderet de caespite acerram. iam memor inceptum peragens Libitina laborem, etc. Vs. 400: et violae omne genus. Violare genus saves the rude elision. Vs. 408: non illinc Narcissus abest. After hie five times repeated in the catalogue of flowers, illinc cries out to be emended to non ille hinc Narcissus abest cui, etc. The Text of the Culex 439 I do not like to conclude without expressing my debt of gratitude to Professor Ellis for his edition, which while it brilliantly solves some problems in the Cvlex offers a most convenient survey of the con- ditions for attacking others. And I have to thank Mr. H. W. Garrod of Merton College, Oxford, for kindly reading a proof and suggesting some improvements and additional evidences. Thb University Glasgow