^{page 185 note 1} Cf. Norden, , Aeneis VI, p. 262;Google ScholarBüchner, , R.-E. viii A, col. 1353.Google Scholar

^{page 185 note 2} Wiechmann, , De Aeneidos libri II compositione, progr. Potsdam, 1876, p. 17.Google Scholar

^{page 185 note 3} qui a Tucca et Vario Bergk; Tucca et Varius C (but Bergk emended obliti, unnecessarily, to sublati). It is amusing to ind Gruppe (Minos, p. 175) defending Virgil's executors from a charge of amnesia, ind Hartmann (Mnemosyne N.s. xxxiii [1905], 144) happily accepting the possibility as: learing up the whole problem. Birt (Kritik und Hermeneutik, p. 160) apparently swallowed he MS. reading (‘Varius und Tucca hatten lie vergessen’).Google Scholar

^{page 185 note 4} Ed. Harv., p. 2;Google Scholar Thilo-Hagen's presenta tion of the text is confusing (see Fraenkel, , J.R.S. xxxviii [1948], 131,Google Scholar where 567–88 are referred to as ‘the interpolation’).

^{page 185 note 5} So Servius auctus: the shorter version runs ‘ut enim in primo diximus, aliquos hinc versus constat esse sublatos, nee immerito’ (the rest is common to both versions).

^{page 185 note 6} Cf. Kviĉala, , Neue Beiträge ztir Erklärung der Aeneis (Prag, 1881), pp. 33 f.;Google ScholarPascal, , Graecia Capta (Firenze, 1905), p. 121;Google Scholar Sabbadini states roundly (in his text of 1937) ‘hos versus Vergilius ipse delevit’.

^{page 186 note 1} Prolegomena, p. 92 (‘propter tam futtiles causas’).Google Scholar

^{page 186 note 2} e.g. by Gossrau (1846); Mancuso, , Classici e Neolatini, 1911, pp. 31 ff.;Google ScholarNoack, , Rheinisches Museum N.F. xlviii (1893), 420 ff.Google Scholar

^{page 186 note 3} Hartmann, I.c., p. 441.

^{page 186 note 4} Classical Philology i (1906), 222.Google Scholar

^{page 186 note 5} Fleckeisens Jahrb. für classische Philologie, xxxi (1885), 399; unfortunately Baehrens's own ideas of how to rewrite Virgil scarcely recommend him as a critic.Google Scholar

^{page 186 note 6} I.c., p. 431.

^{page 186 note 7} I.c., p. 27.

^{page 186 note 8} Op. cit., p. 180.Google Scholar

^{page 186 note 9} Aeneis VI, p. 262.Google Scholar

^{page 186 note 10} Virgils epische Technik (3rd ed. 1928), p. 45.Google Scholar

^{page 186 note 11} Plautinische Forschungen (2nd ed., 1912), p. 42,Google Scholar and see his even more acrimonious remarks in DerMonolog im Drama (1908), p. 5,Google Scholar n. 1. Leo classes the passage with die false ending of the Andria, or with ‘Lucili quam sis mendosus’, etc. (Hor. Sat. 1. 1 o, lines which still have defenders: see Burck's supplement to Kiessling-Heinze, 6th ed., p. 411).Google Scholar

^{page 186 note 12} Peerlkamp had a simple solution: he ejected this passage, and 589–623, and vi. 493–547. He made an illuminating comment: ‘ipsi recentiores Europaei, Vida, Fracastorius, Lotichius, multique Neerlandi, ea saepe protulerunt quae Virgilius libera pro suis agnosceret.’ Libens, indeed!

^{page 187 note 1} Palmer, L. R., Mnemosyne, 3rd series, vi (1938), 368 ff. But he does not explain why so cardinal a passage should have had so casual a history.Google Scholar

^{page 187 note 2} Aeneis VI, Anhang xi. 1.3 (p. 454).Google Scholar

^{page 187 note 3} T.A.P.A. lvi (1925), 172 ff., an interesting and important paper.Google Scholar

^{page 187 note 4} C.R. xli (1927), 123.Google Scholar

^{page 187 note 5} e.g. 9. 677, 678, 12. 757, 769, 771 (with fixam et, 773). Neither Shipley nor Johnson takes account of Virgilian ‘paragraph’- structure (e.g. in the group 3. 188, 200, 222), but this scarcely weakens their case.

^{page 187 note 6} Cf. Braum, O., De monosyllabis ante caesuras hexametri latini collocatis, diss. Marburg, 1906, p. 22.Google Scholar

^{page 187 note 7} e.g. 5. 737, 743, 744, 747, 7. 66, 69, 70; G. 1. 149, 154, 159; G. 2. 347, 348, 355.

^{page 187 note 8} The frequency of et—nine times in these lines, and immediately preceding the caesura in six—might seem suspicious. But in 6. 212–31 (a highly finished passage, which even Peerlkamp did not eject) et occurs eight times (and there is another in 211), in four cases immediately before the caesura.

^{page 187 note 9} Cf. Shipley, I.c., p. 182.

^{page 187 note 10} It reappears in Auson. 31. 306, Prudent. Symm. 2. 681.

^{page 187 note 11} Shipley notes how many ‘common’ compound verbs appear once only in Virgil: ‘one might as well question 4. 297 because of praesensit’.

^{page 187 note 12} It seems to me highly probable that there is an actual reminiscence here of Lucr. 3. 1018 ff. ‘at mens sibi conscia factis / praemetuens adhibet stimulos torretque flagellis,/ nee videt interea qui terminus esse malorum / possit nee quae sit poenarum denique finis.’

^{page 188 note 1} For a remarkable passage involving cinis see 5. 785 ff. ‘non media de gente Phrygum exedisse nefandis / urbem odiis satis est nee poenam traxe per omnem / reliquias Troiae: cineres atque ossa peremptae / in-sequitur’: how the wolves might have fallen on these lines if there had been any hint of mystery about their genuineness—exedisse, poenam traxe per omnem, the ossa of an urbs!

^{page 188 note 2} All these passages show how appropriate satiare is in the context of 587.

^{page 188 note 3} Op. cit., p. 45, n. 1.Google Scholar

^{page 188 note 4} Cf. Madvig, , Adversaria Critica, ii. 107,Google Scholar against the reading focum servat in Ovid, F. 6. 317: ‘non agitur de assidua ad focum commoratione.’ I suspect that Aeneas‘ instruction in 2. 711 ‘longe servet vestigia coniunx’ really means that Creusa is to keep fairly close behind.

^{page 189 note 1} The curious suggestion was made by Heidtmann that erranti … ferenti refers to Helen, not to Aeneas. This scholar, whose text of this book was published at Wesel in 1882, ejected 230 lines as spurious, boasting that he had beaten Peerlkamp's record.

^{page 189 note 2} It has been suggested that arts is dative, with invisa (cf. Gerloff, , Vindiciae Vergilianae, diss. Iena, 1911, pp. 35, 38,Google Scholar and Palmer, , I.c., p. 375, n. 2). But this is no more justifi able than it would be to take ratibus superbis as dative with invisam in 4. 540. Arts must be ablative.Google Scholar

^{page 189 note 3} Note the juxtaposition in 7. 570 f. ‘Erinys, / invisum numen’: Helen is no numen, but she is an Erinys.

^{page 189 note 4} Unless possibly in 7. 577, where Servius interprets igni as ‘ipso fervore deditionis’. Cf. Val. Flacc. 1. 748 ‘saevos irarum concipit ignes’.

^{page 189 note 5} Note Ovid, Met. 6. 708 ‘arserunt agitati fortius ignes’ (of love).

^{page 189 note 6} Perhaps also Prop. 1. 11. 5 (but see Bailey, Shackleton, Propertiana, p. 32).Google Scholar

^{page 190 note 1} Birt, , op. cit., p. 161, n. 1.Google Scholar

^{page 190 note 2} This parallel was noted by Forbiger; but exstinctum following nefas there is a conjecture only, though a probable one (by Gronovius), based on this passage, for MS. est tunc.

^{page 190 note 3} Plautinische Forschungen, p. 43, n.Google Scholar

^{page 190 note 4} Op. cit., p. 161, n. 1.Google Scholar

^{page 190 note 5} Fairclough, , I.c., p. 223;Google Scholar see Norden, , ad loc.Google Scholar

^{page 190 note 6} Noack, , I.c., p. 425, n.Google Scholar

^{page 190 note 7} Belling, , Studien über die Compositionskunst Vergils in der Aeneide (Leipzig, 1899), p. 178.Google Scholar

^{page 190 note 8} Propertiana, p. 39;Google Scholar he regards vices as equivalent to ultionem, noting Servius on 2. 433.

^{page 190 note 9} Cf. Palmer, , I.c., p. 377;Google ScholarPascal, , op. cit., p. 120,Google Scholar n. 2; Hatch, , Classical Philology liv (1959), 255 f.CrossRefGoogle Scholar Gossrau comments ‘sceleratae poenae sunt eae, quibus novum scelus committitur …; ita Orestes in Clytaemnestra sceleratas poenas sumpsit’.

^{page 191 note 1} Op. cit., p. 181.Google Scholar

^{page 191 note 2} See Helm, , Philologische Wochenschrift, 1934, col. 1419Google Scholar where this line of Virgil is adduced; Frere concurs; Vollmer, more probably, takes seriem as object of merentis. Immerens is passive in Fulgentius, , aet. mund. p. 153.Google Scholar 4. Wackernagel, who holds these lines to be interpolated, seems to take merentis in 585 as a passive (Vorlesungen über Syntax, i. 286).Google Scholar

^{page 191 note 3} -que with coniugium is the link between ibit and videbit; cf. 6.683 ‘fataque fortunasque virum moresque manusque’, where the grouping is similar (fata, fortunas virum, moresque manusque).

^{page 191 note 4} Löfstedt, , Syntactica, i. 69;Google Scholar it seems especially common near Trier (Fahnestock, and Peaks, , T.A.P.A. xliv [1913], 80).Google Scholar

^{page 191 note 5} Ovid, Met. 4. 61 has been adduced, but patres need not mean parentis either there or in the passages noted by Langen on Val. Flacc. 1. 150.

^{page 192 note 1} Natos too has been criticized, since in Homer there is an only child, Hermione; but other legends mention a son, Nicostratus (Apollodorus, Bibl. 3. 11. 1, where see Frazer's note for this and other versions of Helen's family). But this is pedantry: nati is conventional and a natural convention.

^{page 192 note 2} See Fraenkel, , J.R.S. xxxviii (1948), 137.Google Scholar

^{page 192 note 3} Mackail: habet h. v. laudem, / exstinxisse nefas; Itamen et…; Janell: habet h. v. laudem: / extinxisse nefas tamenet… (similarly Ladewig Deuticke, but with a full stop after laudem).

^{page 192 note 4} Birt, , op. cit., p. 161, n. 1, reads: habet h. v. laudem, extinxisse nefas, tamen; et. …Google Scholar

^{page 192 note 5} In his text of 1886, inaccessible to me.

^{page 192 note 6} Op. cit., p. 182;Google Scholar B. ingeniously observes ‘die Wiederholung des Wortes … ist in dem ganzen Abschnitt so wenig vermieden, dass es nichts ausmacht, ob es noch einmal mehr stent’.

^{page 192 note 7} P. explains that Aeneas intended Helenam inflammas inicere.

^{page 192 note 8} Plautinische Forschungen, p. 42, n. 3; L. adds ‘das ist aber wohl zu gut für diesen Poeten’.Google Scholar

^{page 192 note 9} Ancient Lives of Vergil, p. 24.Google Scholar

^{page 193 note 1} Pascal, , op. cit., p. 120, n. 2, following a suggestion by Heyne.Google Scholar

^{page 193 note 2} I.c., pp. 399 f.

^{page 193 note 3} Firmicus Maternus, de err. prof. relig. 16. 2 ‘quid enim meretur aliud parricida, nisi ut perpetua (con)tinuatione flammarum ante sententiam dei cotidie flammis ultricibus concremetur?’; Rufinus, hist. 3. 7. 8 ‘si possent ultrices poenarum flammas lacrimarura ubertate restinguere’; Valerianus Cemenelensis, hom. 8. 5 ‘ultrices criminum Hammae’. But it must be granted that Firmicus Maternus has two clear Virgilian echoes in the de errore (3. 1 ‘pro iniuria ipretae fecerat formae’, 6. 7 ‘Liber … cum semiviro comitatu fugiens’).

^{page 193 note 4} Op. cit., p. 161, n. 1.Google Scholar

^{page 193 note 5} Bailey, Shackleton, Propertiana, p. 126, explains differently.Google Scholar

^{page 193 note 6} s.v. expleo, col. 1717. 42.

^{page 194 note 1} Op. cit., p. 47.Google Scholar

^{page 194 note 2} Birt, (op. cit., p. 161, n. 1) objects to the absence of inquam or the like; but its presence would have been frostig in the extreme. Yet Birt himself (p. 131) adduces this passage to illustrate Hor. Epod. 4. 11 ff., where liberrima indignatio breaks into speech (with no inquit) in a context not unlike in tone.Google Scholar

^{page 194 note 3} Again in 407, and nowhere else in Virgil; it recurs in Sil. 2. 210, Val. Flacc. 8. 445.

^{page 194 note 4} Cf. Leo, , Ausgewählte kleine Schriften, ii. 52.Google Scholar

^{page 194 note 5} See Norden, , Aeneis VI, Anhang III. A. 2.Google Scholar

^{page 194 note 6} Cf. Kviĉala, , op. cit., pp. 33 ff.; with exstinxisse … sumpsisse … explesse compare 4. 603 ff. (fuisset… tulissem … implessem … exstinxem … dedissem).Google Scholar

^{page 195 note 1} Cf. Heinze, , op. cit., p. 50.Google Scholar

^{page 195 note 2} Cf. Walter, O., Die Entstehung der Halb-verse in der Aeneis, diss. Giessen, 1933, p. 34.Google Scholar

^{page 195 note 3} Hermes li (1916), 145 ff.Google Scholar

^{page 195 note 4} Peerlkamp regarded 589–623 as spurious. If the Venus-scene had been inter polated, which of course is nonsense, it must have happened very early, for non tibi Tyndaridis fades (601) occurs in a papyrus dated by E. G. Turner to the first century A.D.: see Studi in onore di Aristide Calderini e Roberto Paribeni (1957), ii. 157 ff.,Google Scholar and Milne, J. G., J.H.S. xxviii (1908), 125. Even Heidtmann retained the scene, in a mangled form.Google Scholar

^{page 196 note 1} deo V, above dea; in M dea has been corrected to deo; in P dea is a supralineal correction of de; Schol. Veron. on v. 467 quotes the line with deo. Servius (here and on 1. 382, 4. 228, 7. 498), Macrobius (3. 8. 1), Aelius Donatus (on Ter. Ad. 894, Eun. 875) support deo. Ti. Claudius Donatus read abducente dea (wicked Venus persuading her son to desert his post). The balance is strong in favour of the masculine.

^{page 196 note 2} See Thes. L.L. s.v. deus, col. 890. 16 ff., for the masculine used of a male or female numen (but Catull. 61. 64 is wrongly classified). Both Servius and Macrobius quote Calvus ‘pollentemque deum Venerem’ (these scholia are closely related, cf. Servius on 7. 498); this may have been a neoteric mannerism, since the corresponding Greek usage of is frequent in Callimachus (see Pfeiffer's index).

^{page 196 note 3} See Heinze, , op. cit., p. 48.Google Scholar

^{page 196 note 4} Heinze denies this, adducing instead Eur. Or. 1388. There is no reason why both passages should not lie behind the phrase, as well as Ennius, Sc. 71 V. ‘Lacedaemonia mulier, Furiarum una’, which Fraenkel (on Aesch. I.c.) thinks may go back to Euripides.

^{page 196 note 5} Eur. Andr. 627 ff.,Google ScholarAr. Lys. 155 f.;Google Scholar the story goes back to the Ilias Parva. Cf. Gerloff, , op. cit., pp. 25 ff.;Google Scholar Mancuso, I.c., p. 25; Noack, I.c., p. 428.

^{page 197 note 1} Cf. Bill, , Classical Philology xxvii (1932), 168 f.CrossRefGoogle Scholar

^{page 197 note 2} T.A.P.A. lvi (1925), 184.Google Scholar

^{page 197 note 3} See Aulus GelUus, 13. 21. 4 (Probus), 1. 21.2 (Hyginus); cf. Bill, I.c.

^{page 197 note 4} Cf. Donatus, vita 33 ‘recitavit et pluribus, sed neque frequenter et ea fere de quibus ambigebat, quo raagis iudicium hominum experiretur’.

^{page 197 note 5} Jahn, retaining habet haec in 584, adds ‘ich habe … die folgenden drei Verse als Dittogramm ausgesondert, bin aber meiner Sache keineswegs sicher’.

^{page 197 note 1} Cf. Funaioli, , Studi di letteratura antica II. i. 245.Google Scholar