Turbatae acies, probably by the death of Aconteus, who was evidently on the Latin side. Serv. says that these alternate advances and retreats are only the ordinary manœuvres of a cavalry engagement. Comp. 5. 580 foll. Sall. Iug. 59, “illi—non uti equestri proelio solet, sequi dein cedere, sed adversis equis concurrere, inplicare ac perturbare aciem.” Livy 22. 47 (of the battle of Cannae), “minime equestris more pugnae: frontibus enim adversis concurrendum erat.” Thus it would appear that Virg. has combined two modes of cavalry engagement, that which Sallust describes as the less regular with the more regular.
 They sling their shields behind to protect their backs in flight. Comp. Il. 11. 545, where Ajax, beginning to retreat, ό̀πιθεν σάκος βάλεν ἑπταβόειον. This was sometimes done as a sign of friendly intentions: Galba ap. Cic. Fam. 10. 30, “repente equum inmisi ad eam legionem quae veniebat e castris, scuto reiecto.” Livy 22. 48, “specie transfugarum cum ab suis, parmas post terga habentes, adequitassent.” For ‘vertunt,’ which is awkward after ‘versi,’ one of Ribbeck's cursives has ‘tendunt,’ giving ‘vertunt’ as a marginal variant.
 Agunt sc. “Latinos.” ‘Inducit’ on the enemy. Some MSS. (none of Ribbeck's) give ‘invasit,’ which was the reading of some early editions. ‘Princeps,’ see on 10. 254. ‘Asilas’ doubtless the Etruscan leader of 10. 175, as there is nothing to show that the one mentioned 9. 571 was a prominent person.
 This simile seems to be Virg.'s own: at least Il. 11. 305 foll., to which Heyne refers, bears no real resemblance to it. The object described, as Heyne remarks, is not the ebb and flow of the tide, but the alternate advance and retreat of the waves. For ‘procurrens’ Rom. and two of Ribbeck's cursives have ‘procumbens,’ while in Pal. the ‘rr’ of ‘procurrens’ is written over an erasure: but though ‘procumbens’ would not be inappropriate (Ribbeck appositely refers to G. 3. 240), the other seems better. ‘Procurrens alterno gurgite’ seems to mean advancing alternately, i. e. advancing and retreating by turns.
 Terras Med., ‘terram’ Ribbeck's other MSS. Wagn. restored ‘terram’ on internal grounds, which, as Forb. remarks, are very precarious: but the external reason seems sufficient. For ‘unda’ the old reading was ‘undam,’ which is found in Gud. and three other of Ribbeck's cursives, in one of them from a correction. Rom. has ‘suberigit,’ which can hardly be more than a clerical error, though it is apparently supported by Sil. 15. 155, “Corus Isthmon curvata sublime suberigit unda,” where however Wagn. plausibly suggests ‘superiicit.’ ‘Scopulos superiacit unda’ will then be a variety for “scopulis superiacit undam,” like “socios circumtulit unda” 6. 229, which Wagn. comp., though much less strong. Comp. the double construction of “figere,” “suffundere” &c.
 Retro with ‘fugit,’ ‘saxa’ with ‘resorbens’ (so Wagn. rightly against Heyne), ‘aestu’ probably with ‘revoluta.’ The force of the wave dislodges the stones and partially sucks them back. ‘Retro fugit’ above v. 405. The recurrence of ‘r’ and ‘s’ here and in vv. 624, 625 is doubtless intentional, and so perhaps that of ‘l’ in the next line.
 It is singular that the commentators generally should have taken no notice of this line, which certainly requires explanation. The meaning seems to be ‘twice, beaten back, they (the Tuscans) look behind them and cover their backs with their shields,’ ‘armis’ being constructed with ‘tegentes,’ which is nom. Comp. v. 619 above. “Eos qui eruptionem fecerant in oppidum reiiciebant,” Caes. B. C. 2. 2. Burm. on Ov. M. 2. 582 apparently understands ‘reiecti’ in the sense of ‘parmas reiectas habentes,’ but does not say whether it is constructed with ‘armis.’
 With the beginning of this line comp. 3. 37. Heins. read ‘totasque’ from a few inferior MSS.
 ‘Inplicuere:’ comp. Sall. Iug. 59, quoted on v. 618 above. ‘Vir virum legit’ was an old Roman phrase, which seems to have been originally applied to cases of conscription, where certain individuals were bidden to select other persons individually in order to make up an army: comp. Livy 9. 39, “lege sacrata coacto exercitu, cum vir virum legisset.” Id. 10. 38, “decem nominatis ab inperatore edictum ut vir virum legerent, donec sexdecim milium numerum confecissent.” Suetonius uses it twice of the filling up of vacancies in the Senate by a similar process, Aug. 35, 54, though in the latter passage the reading is not quite certain. Cic. Pro Mil. 21 uses it contemptuously to express the suitability of Clodius' companions to himself. Tac. H. 1. 18 makes Galba apply it to his adoption of Piso, “more divi Augusti et exemplo militari quo vir virum legeret.” Virg. evidently means it to be understood of man singling out man in hand-to-hand fighting, perhaps taking a hint from Il. 15. 328, which evidently was in his mind, ἔνθα δ᾽ ἀνὴρ ἑλεν ἄνδρα κεδασθείσης ὑσμίνης, as if ἕλεν ῀ εἵλετο. Comp. also Il. 4. 472, ἀλλήλοις ἐπόρουσαν, ἀνὴρ δ᾽ ἄνδρ᾽ ἐδνοπάλιζεν. Comp. Livy 22. 47, “in directum utrimque nitentes, stantibus ac confertis postremo turba equis, vir virum amplexus detrahebat equo.”
 For ‘tum vero’ in the apodosis comp. 5. 719 note. The omission of the verb after ‘et’ is peculiar, as it is coupled closely with a clause where there is a verb expressed, so that the construction is perhaps to be regarded as a harsh zeugma. For the sense comp. Il. 4. 450, ἔνθα δ᾽ ἅμ᾽ οἰμωγή τε καὶ εὐχωλὴ πέλεν ἀνδρῶν Ὀλλύντων τε καὶ ὀλλυμένων: ῥέε δ᾽ αἵματι γαῖα. For ‘in alto’ one of Ribbeck's cursives gives ‘largo,’ and some inferior copies have ‘multo.’ Serv. says “bene belli faciem demonstravit multa enumerando quae in alto sanguine velut natarent.”
 Gud. has a variant ‘semineces.’ The spelling ‘semanimes,’ which obtained before Wagn., is found in none of Ribbeck's MSS. See on 3. 244. ‘Pugna aspera surgit’ 9. 667. In one of Ribbeck's cursives these words are written over an erasure.
 Comp. 10. 892, where Mezentius' horse rears after being wounded between the temples. For ‘furit’ Rom. and originally Med. have ‘ferit,’ the correction in Med. being made by a late hand: but though “calce ferire” occurs Ov. F. 3. 755 of an ass kicking, it does not appear that ‘ferio’ is thus used absolutely. “Arduus equis furit” 7. 625.
 Ingentem animis, μεγάθυμον. Comp. Il. 21. 395, μέγας δέ σε θυμὸς ἀνῆκεν. ‘Ingentem corpore’ like “cornibus ingens” 7. 483. No commentator seems to notice ‘armis,’ which was doubtless supposed to be parallel to “ingentior armis” v. 124, or at any rate to denote that Herminius wore massive armour. But in what follows the stress is laid rather on the absence of defensive armour, and the mention of his shoulders, ‘nudi humeri,’ and the use of the word ‘armos’ v. 644, show that ‘armis’ here is from ‘armus.’ Thus this passage and 4. 11 (note) establish each other.
 ‘Deiicit,’ brings down from his horse, as it was a cavalry engagement. So vv. 665, 832 below, and doubtless 10. 753, where horses are mentioned in the context. It does not seem to be used by Virg. of simply bringing down from a standing position, like “sternere.” In 10. 319 the addition of “leto” makes the difference: comp. “demisere neci” 2. 85. In 9. 770 the head is cut from the standing body, and so falls from a height. Comp. v. 580 above. ‘Herminium:’ Serv. thinks Virg. had in his mind Herminius, the companion of Cocles at the bridge, who, we may remember, was engaged against the Etruscans. Melanchthon and others ap. Taubm. notice that the physique of Herminius here is that of a German or Gaulish warrior, and accordingly suppose Virg. to have been thinking of Arminius, whom they apparently assume to have been known to the poet while a hostage at Rome. Both suppositions are conceivable, and not incompatible. ‘Fulvus’ is the colour of a lion's mane, G. 4. 408.
 Tantus in arma patet can only mean ‘so vast the front he presents to the weapons of the enemy.’ Serv.'s note is doubtless to be read “tantum patebat in volnera, i. e. in hostilia tela tantus patebat.” Rom. has ‘tantum.’ There is still room for doubt about the connexion with the preceding line, which may either be, as Heyne seems to think, the surface he leaves exposed shows that he does not fear wounds, or, he trusts that the vastness of his frame will protect him, as if mere physical strength could blunt the edge of weapons. ‘Per armos:’ see on v. 641. The imitation in Stat. Theb. 7. 634 looks as if he took ‘armos’ here of Herminius' horse.
 Duplicat virum: Heyne comp. ἰδνώθη δὲ πεσών Il. 13. 618. “Duplicato poplite” 12. 927. ‘Transfixa’ of the spear, not, as commonly, of that which it pierces. Forc. cites “Ora ducis, quae transfixo deformia pilo Vidimus,” Lucan 9. 137. Comp. the double construction of “fixus,” “infixus” &c. ‘Dolore’ i. q. “prae dolore.” ‘Dolorem’ is found in one MS., and was read by some of the early editors, who either constructed ‘virum’ with ‘transfixa’ or read ‘viri’ instead. From this line to v. 692 Pal. is wanting.
 Partly repeated from G. 4. 218 (note). Comp. 9. 401.
[648-698] ‘The actions of Camilla. She kills many of the Trojans and their allies in various ways, by arrow, spear, or battle-axe.’