stand, though the possibility is infinitesimal; but in that case we must suppose that he wrote the lines in the order in which we have them, but that on revision he would have left out vv. 624-627, if not v. 628.

624.] "Pedes apparat ire " 10. 453. "It campis" 4. 404. Pars arduus' &c. is a mixture of two common constructions, "pars ardua and " 'pars ardui" (nom. masc.), resulting, no doubt, from the use of 'pedes immediately before. For arduus' comp. v. 285, "Sublimes in equis redeunt.' 'Pulverulentus furit' should, as Forb. says, be taken closely together, as if it were "furendo pulverem excitat:" this will remove the difficulty of the two epithets, ' arduus altis equis' alone being opposed to 'pedes.'

625.] Requirunt' i. q. "quaerunt:" comp. v. 460 &c. Macrob. Sat. 6. 1 quotes from Ennius (A. 6. fr. 10), "Balantum pecudes quatit: omnes arma requirunt."

626.] Tergunt' was introduced by Heyne, from a misunderstanding of the critical notices of Pierius and Heins. It is really supported only by a correction in Gud., by some MSS. of less authority, including Canon. and Balliol, and by some notices in the grammarians, who speak of it as equally admissible with 'tergent,' though Serv. says 'tergo tergis' is obsolete. 'Levis' and 'lucida' are of course proleptic.

627.] Serv. quotes Suetonius de Vitiis Corporalibus to prove that 'arvina' is the hard fat between the skin and the flesh: others however, he says, gave the name to lard. The only other writers who use it, Prudentius and Sidonius Apollinaris, make it a synonyme for fat. Subigunt,' rub down, and so, whet. So it is used of kneading bread: see Freund. For 'securis' see on v. 510.

628.] This line describes the march of the assembling troops. Signa ferre,' to advance, 8. 498 (comp. "inferre signa,"

625

630

"conferre signa "), with the notion also of course of the pride taken in displaying or advancing the standards. Med. has iubet,' the same variety as in 4. 498, from which Heins. plausibly suggests 'lubet,' a variety actually found along with 'iuvat' 9. 514. But the copyist may have thought of 8. 498, "signaque ferre iubent." Generally we may comp. Hor. 1 Od. 1. 23, "Multos castra iuvant et lituo tubae Permixtus sonitus," which may have been in Virg.'s mind.

629.] 'Quinque adeo,' literally, as many as five great cities: but adeo' should in fact merely be rendered by an emphasis on 'quinque: see on v. 427., 3. 203. 'Positis incudibus:' they set up anvils, for which they had previously had no occasion. Forb. comp. Ov. F. 4. 473" Antraque Cyclopum positis exusta caminis."

630.] Tela novant' like "transtra novant" 5. 752. Comp. Hor. 1 Od. 35. 38, “O utinam nova Incude diffingas retusum in Massagetas Arabasque ferrum." Atina is apparently regarded by Virg. as a Latin city, though it seems to have been originally Volscian, while historically it was Samnite (Dict. G. s. v.). Tibur is called 'superbum' doubtless with reference to its position, though Serv. fancies there is an allusion to an answer given by the Roman senate on one occasion to a Tiburtine embassy, "superbi estis."

Crus

631.] Ardea' above v. 411. tumeri would seem to be the inhabitants of Crustumerium, the people being mentioned instead of the town on metrical grounds: they are however generally called "Crustumini," and the place is sometimes called "Crustumium," which would have suited the metre. For the questions about its origin see Dict. G. It was said to have been conquered by Romulus along with Antemnae and Caenina, all of which took up arms to avenge the rape of their women at the Consualia (Livy 1. 9 foll.). There are similar questions about the origin of Antemnae (Dict. G.). Sil. 8. 365 calls it

'prisco Crustumio prior." It was SO called from its position "ante amnem,' below the confluence of the Anio and Tiber.

632.] The alliteration, which was doubtless intended, may remind us of Lucr. 2. 618, "Tympana tenta tonant palmis et cymbala circum Concava." Tegmina capitum' vv. 689, 742 below.

6

633.] The frame-work of the shield was made of twisted osiers, which were covered with hides and finally bound round with metal (Dict. A. Clipeus'). Hence iréa is used for a shield Eur. Supp. 697, Tro. 1201. We may say either that "umbo" is put for the whole shield, or that 'cratis umbonum' is the wicker-work to which the boss was fitted. Lersch §. 31 comp. Caes. B. G. 2. 33 "partim scutis ex cortice factis aut viminibus intextis, quae subito, ut temporis exiguitas postulabat, pellibus induxerant," arguing that it was the "scutum," not the "clipeus," that was so constructed. Others comp. the Persian yéppov, Hdt. 7. 61, of which crates' may be a translation. "Cráteras aenos 9. 165.

دو

634.] It matters little whether 'lento argento' be taken with 'ocreas' as a material abl. or with 'ducunt.' The former would be more symmetrical: but the latter is more in accordance with Virg.'s love of variety, as if "ducunt thoracas aere" had preceded, and is supported by Pliny 7. 37, "[Alexander] edixit ne quis ipsum alius quam Lysippus ex aere duceret ;" Appul. Flor. 1, "qui solus effigiem regis Polycletus aere duceret," quoted by Forb. 'Ducere' in these cases is to extend by beating, and answers nearly to our sense of ἐλαύνειν, as "ducere murum (1.423) does to another. The spondaic metre expresses the slowness of the process. Forb. asserts that ocreas is to be pronounced as a dissyllable, being apparently misled by one or two of the early writers on metre, who quote this line as one which would be purely spondaic if

635

'ocreas' were so pronounced or changed e. g. into "lamnas" (see Marius Plotius De Metris pp. 247, 251 of Gaisford's Scriptt. Latt. Rei Metr., Diomedes p. 495). Serv.'s remark "est autem spondaicus et reciprocus versus " ("reciprocus" rhyming, alluding to "lento-argento ") only means that there is a spondee in the 5th place. "Leves ocreas" 8. 624, where the metal is gold and electrum: comp. 11. 488.

635.] Huc cessit:' the esteem in which agriculture was held is swallowed up in the enthusiasm for war. They cease to make agricultural implements, and perhaps, though this is not expressly stated, turn them into warlike weapons, as in G. 1. 508, where the imagery is generally parallel (comp. especially v. 506, "non ullus aratro Dignus honos"). With 'huc cessit,' which implies that one feeling has passed into the other, comp. 8. 395, "fiducia cessit Quo tibi, diva, mei ?" Virg. seems generally to have had in his mind the description in Lucr. 5 of the gradual discovery of the use of metals, "Nunc iacet aes, aurum in summum successit honorem" (v. 1275: comp. the following lines, where 'honore' occurs twice), "Inde minutatim processit ferreus ensis Versaque in opprobrium species est falcis aenae (vv. 1293, 4, quoted as parallel by Macrob. Sat. 6. 1).

636.]Recoquunt ensis' like "diffingas ferrum" Hor. 1 Od. 35. 39, quoted on v. 630 above.

637.] Iamque' second in a clause 3. 588., 5. 225. "Classica" G. 2. 539, here used in its strict sense of the sound of the horn which called the Romans together to battle or other assemblies. 'Tessera,' Dict. A. s. v., originally a cube used as a token or tally, hence, as here, a watchword, which passes, 'it,' from man to man. The preparations are over, and the actual march begins.

638.] Tectis' from his house, from the chamber where it is stored.

639.] Cogit:' brings together, implying

that a pair are to be yoked. The more ordinary expression would be "sub iuga cogit," as in Moretum 113. Stat. Theb. 7. 136 (quoted by Forb.) has “alienaque cogunt Ad iuga cornipedes," an imitation which shows that 'ad iuga' here does not go with 'frementis,' like "fremit ad caulas" 9. 60. "Auro trilicem Loricam :" see on 3. 467.

"

640.] "Fidum ensem 6. 524. 641-646.] Sing, Muses, of the Italian chiefs and their followers.'

641.] This invocation is of course from that in Il. 2. 484 foll., preceding the catalogue of the ships. It is generally briefer than its prototype; but the subject to be narrated is indicated at greater length. The notion is that an account involving much statistical detail requires the special aid of the goddesses of memory and song. 'Pandite Helicona' like "panditur domus Olympi " 10. 1, as if the gates of Helicon held in song. Wagn. well comp. Bacchylides, fr.14 Bergk, ovde yàp pậσTоv åρþýтwv ἐπέων πύλας ἐξευρεῖν. The notion is probably a complex one, of free utterance and of expounding things unknown. The former is the image in Pind. O. 6. 45, which Wagn. thinks inapplicable, xpǹ τοίνυν πύλας ὕμνων ἀναπιτνάμεν αὐταῖς: comp. the opening of Whytehead's Cambridge Installation Ode, "Fling the gates of music wide, Hold back no more the rush of song." "Movete' like "moveo V. 45, whether in any way connected with the preceding image is not clear. There is a plausible variant 'monete in fragm. Vat. and Gud., recognized also by Serv.: comp. v. 41. Wagn. rightly rejects it, but is perhaps too scrupulous when he questions its Latinity. Heins. conj. "cantuque monete," which is actually found in Canon.

642.] Exciti,' the long penult. as in 3. 676., 10. 38: comp. v. 623 above. 'Bello,' prob. dat., as in v. 482 &c. Comp. generally vv. 37 foll. above.

643.] Iam tum,' even then, before the

,

640

645

great historical period of Rome. 'Tantum' was a reading before Pierius. 644.] Comp. G. 2. 167 foll. of the nations and families of Italy. 'Alma' i. q. "parens." "Quibus arserit armis' expresses generally what is expressed more in detail in the two previous lines, 'arma' being the Virgilian accompaniment to 'viri,' as in the first line of the Aeneid and elsewhere. 'Arserit' probably includes both martial enthusiasm and the flashing of armour.

645.] ὑμεῖς γὰρ θεαί ἐστε, πάρεστέ τε, LOTE TE Távтa Il. 1. c. Virg. has chosen two verbs which suggest the connexion of memory with the Muses: see on v. 45, E. 7.19.

646.] ἡμεῖς δὲ κλέος οἷον ἀκούομεν, οὐδέ Ti toμev II. 1. c.

647-654.] Mezentius and his son Lausus lead an army from Agyllae in Etruria.'

647.] Init bellum,' begins the war. Prima pares ineunt gravibus certamina remis 5. 114. 6 Tyrrhenis ab oris :' "oriundo Tuscus, non qui nunc de Tuscia venit, quem antea pulsum a civibus constat" Serv. This is possible enough (comp. G. 3. 2, "Pastor ab Amphryso"): but even if the words are constructed with 'init,' they need not be pressed to mean that Mezentius came direct from Etruria. The name of Mezentius comes from the early legends, though his part in the story was differently represented by different persons, some making him kill Aeneas in a battle subsequent to the Trojan settlement in Latium. See Heyne, Excursus 3 on Book 8, Lewis 1, p. 338. So, according to another account, Lausus was killed later by Ascanius (Dict. M. s. v.). Serv. remarks on the whole catalogue that Virg. mentions some whom he does not name afterwards in his narrative, while some who are named afterwards are not mentioned here, which he calls 66 poetae affectatio, nam amblysiam" (apparently some word con