Respicit seems to imply that they had proceeded some way towards Elysium, as at the point of divergence Tartarus would be before them. ‘Sub rupe’ is apparently from the Homeric description of the junction of Styx and Cocytus, Od. 10. 515, πέτρη τε, ξύνεσίς τε δύω ποταμῶν ἐριδούπων.
 See on 2. 234, a passage which, like this, enables us to discriminate between ‘murus’ and ‘moenia.’ It signifies little whether we suppose that here we are intended to conceive of one large building or of several. In any case we are meant to imagine a tower or Bastille. The wall that surrounds it is from Hesiod, Theog. 726, τὸν (Τάρταρον) περὶ χάλκεον ἕρκος ἐλήλαται.
 Phlegethon acts as a moat, apparently outside the walls. ‘Torrentibus’ is probably meant to suggest the notion of a torrent as well as that of scorching flame. So “pice torrentis ripas” 9. 105., 10. 114. ‘Flammis’ probably with ‘ambit’ rather than with ‘rapidus.’
 The full name of the river is Pyriphlegethon, Od. 10. 513, Plato, Phaedo p. 113 B. It is called ‘Tartareus’ like Acheron v. 295, but with more propriety, as it is specially the river of the place of torture. For ‘torquetque saxa’ after ‘quae ambit’ see on G. 2. 208. The ‘sonantia saxa’ may come from a misunderstanding of Plato l. c. Πυριφλεγέθοντα, οὗ καὶ οἱ ῥύακες ἀποσπάσματα ἀναφυσῶσιν, ὅπη ἂν τύχωσι τῆς γῆς, where ἀποσπάσματα are not fragments of rocks, but parts of Pyriphlegethon, which are said to be disgorged by lavastreams in different parts of the earth. We may comp. the description of an earthly torrent Lucr. 1.288, “volvitque sub undis grandia saxa,” and G. 3. 254, “Flumina, correptosque unda torquentia montis.” Rom. has “tonantia.”
 Columnae i. q. ‘postes,’ being apparently chosen as better adapted to the gigantesque style of description. Heyne comp. Il. 8. 15, ἔνθα σιδήρειαί τε πύλαι καὶ χάλκεος οὐδός, said of Tartarus. ‘Adamas’ is the common poetical word for the hardest substance, e. g. Aesch. Prom. 6, “ἀδαμαντίνων δεσμῶν ἐν ἀρ᾽ῥήκτοις πέδαις”, which will also illustrate v. 553.
 The meaning is that neither men nor gods can make the gates of Tartarus open when once closed. ‘Bello’ is the reading of Rom., fragm. Vat., and Pal., ‘ferro’ of Med. With Ladewig I prefer the former, which is more picturesque, and avoids the awkwardness of ‘ferrea’ following. Ladewig thinks there is an intentional ascent from the violence of men to the battle of the gods. But there is much to be said for ‘ferro,’ which is supported by 9. 137, “ferro sceleratam exscindere gentem,” and by 2. 463. ‘Bello’ has been erroneously introduced for ‘ferro’ by Med. in 12. 124, as Wagn. remarks.
 The stronghold has a tower, like Priam's palace 2. 460. ‘Stat,’ as Forb. remarks, combines the notions of height and fixity. ‘Ad auras,’ as if ‘surgit’ or ‘se tollit’ had preceded. Here and in v. 561 ‘aurae’ of course stands for the atmosphere of the lower world. Serv. says sensibly “auras inferis congruas intelligamus,” telling us at the same time that Pollio supposed that Aeneas and the Sibyl brought some of the upper air with them.
 Tisiphone G. 3. 552. Her bloody robe is from Il. 18. 538, where it is said of ὀλοὴ Κήρ in the middle of a fray εἷμα δ᾽ ἔχ᾽ ἀμφ᾽ ὤμοισι δαφοινεὸν αἵματι φωτῶν, “accommodatius sane in pugna,” as Heyne remarks. She is represented in battle 10. 761, evidently from the same passage in Hom. ‘Palla’ note on 1. 648.
 Tisiphone is meant to act as porter or sentinel, v. 575. ‘Servat’ G. 4. 459. ‘Exsomnis’ i. q. ‘insomnis.’ It is used by Hor. 3. Od. 25. 9 in the sense of ‘starting from sleep,’ but it would be too much to assume with Forb. that such is its natural meaning. All we can say is that while words compounded with ‘in’ may be called negative, like those with ἀ in Greek, those compounded with ‘ex,’ like those with ἀπό, may be called privative: but in poetical language at any rate the two are virtually equivalent. Rom. has ‘insomnis.’ “Noctes atque dies” v. 127 above.
 ‘Tractae catenae’ probably nom. pl., though it might be gen. sing. Wagn. remarks that ‘stridor’ practically supplies the place of a verb; we are probably however meant to borrow one grammatically from the former sentence, though of course it would be possible to understand a verb substantive. ‘Ferri’ is explained by ‘catenae.’
 Strepitu—haesit is restored by Wagn. from Med., Pal. a m. s. for ‘strepitum—hausit,’ Gud., Pal. a m. p., fragm. Vat. a m. s. Rom. has ‘strepitum—haesit.’ There is but little to choose between the two: ‘haesit’ however seems slightly preferable, as ‘hausit’ would apparently make his listening to the shrieks subsequent to his stopping. Wagn. comp. “paulum aspectu conterritus haesit, Continuitque gradum” 3. 597, “subitoque aspectu territus haesit” 11. 699—passages however which might possibly have suggested the variety to a transcriber here. Henry prefers ‘hausit’ as more picturesque. I have sometimes fancied that an opposite corruption may have taken place in 11. 864, “Audiit una Arruns, haesitque in corpore ferrum,” where ‘hausit,’ though found in no MS., would be somewhat more vivid.
 Urguentur, the ‘scelerum facies,’ which includes ‘scelerati.’ ‘Quis’ is recalled by Wagn. from Med., Gud., Pal. am. s. &c. for ‘qui,’ which Heins. had introduced from Rom., Pal. a m. p. &c. See note on E. 1. 19. Ribbeck reads ‘clangor’ from Pal. and Gud., but ‘plangor’ is obviously preferable. ‘Ad auras’ v. 554 note. Here it suggests a verb, and in effect supplies its place. Ribbeck reads ‘ad auris’ from Pal.
[562-627] ‘The Sibyl explains that this is Tartarus, which is never entered by the good, she herself having only seen it when introduced to her office by Hecate. Criminals are given over to Rhadamanthus, who compels them to confess, and delivers them to the Furies: then the gate is opened, and they are thrown into a tremendous abyss. All the great criminals, demigods and heroes, are suffering there, and all the guilty of later times. The forms of crime are innumerable, and so are the punishments.’
Insistere with acc. G. 3. 164.
The ‘limen’ is called ‘sceleratum’ as
‘Tartara’ are called ‘impia’ v. 543; but
there is also a reference, as Germ. points
out, to the threshold as the special seat of
the Furies. He comp. Ov. M. 4. 453 foll.,
where Juno goes down to Orcus to seek
“Carceris ante fores clausas adamante
Deque suis atros pectebant crinibus angues.
Quam simul agnorunt inter caliginis umbras,
Surrexere deae: sedes Scelerata vocatur.
 v. 118.
 Deum may either be used generally, the punishments being supposed to have the sanction of the whole body of gods (comp. “fata deum” v. 376 &c.), or specially, indicating that the punishments were frequently inflicted at the instance of one or other of the gods, e. g. on Tityos, v. 595, who offered violence to Latona. A few MSS. have ‘loci poenas,’ and Heins. ingeniously conjectured ‘reum,’ i. e. ‘reorum.’ ‘Per omnia duxit’ is to be understood literally, as Forb. remarks, referring to vv. 582, 585.
 Rhadamanthus, the brother of Minos (Il. 14. 321), in Hom. is placed in the Elysian fields, apparently as a kind of president (Od. 4. 564). In Plato, as we have seen on v. 430, he is the judge of the Asiatic dead. Heyne remarks that his office here answers rather to that of the ‘Triumviri Capitales’ at Rome, or to that of the Eleven at Athens, than to that of a judge, as the spirits are presumed guilty before being committed to him, and he tortures them into confession and inflicts or superintends their punishment. It is difficult to say whether or no Wagn. is right in placing a comma after ‘habet:’ but on the whole the apposition, ‘durissima regna,’ seems more in the manner of Virg. (comp. 3. 106, 272., 11. 252), and it has perhaps some additional force as a separate clause. ‘This is the empire of Rhadamanthus, and a stern and savage one it is.’ ‘Haec’ then, as Wagn. says, will be used generally, like τάδε in Greek. In any case ‘durissima regna’ will refer at least as much to the character of Rhadamanthus' rule as to that of the regions subjected to it. So far as the reference is local, it extends to the whole of Tartarus, the ‘moenia’ of v. 549.
 Castigatque auditque, a ὕστερον πρότερον, perhaps intended to express the summary character of Rhadamanthus' justice, punishment following at once on examination. ‘Dolos’ seems to be put generally for crime, which is conceived of as skulking from justice and pleading not guilty. A more special instance of the same thing follows in the next clause. Forb. notes the general use of ‘fraus’ for injury.
 ‘Has put off to the late hour of death,’ not a strictly accurate expression, as Virg. means not a death-bed confession, but a suppression of guilt till it is revealed in the other world. ‘Piaculum’ of a crime is as old as Ennius and Plautus, see Forc. Its use here may be meant to suggest that the confession has been delayed till earthly expiation is too late, at the same time that it suits ‘distulit,’ as what is really put off is not the crime but the confession and atonement. Expiation must now be made in the lower world.
 Tisiphone, as we saw on v. 556, is the ‘dweller on the threshold;’ the meaning here accordingly seems to be that Rhadamanthus consigus the guilty to her, and she opens the door through which they pass to their doom. ‘Accincta’ seems merely to mean armed, as we cannot suppose that Tisiphone carried a scourge at her girdle. See on v. 184. So “omnis facibus pubes accingitur atris” 9. 74. Serv. explains it grotesquely of the long lash coiling round her as she wields it. There is a similar passage in 2. 612 foll., “Iuno . . . sociumque furens a navibus agmen Ferro accincta vocat,” but it does not seem to help us to explain that before us. We might explain ‘accincta flagello,’ ‘girt up for wielding the lash,’ like “se praedae accingunt” 1. 210, “accingunt omnes operi” 2. 235, but the parallel 9. 74 is against this.
 Quatit is not constructed with ‘flagello,’ but there can be little doubt that ‘accincta flagello’ is meant to indicate the kind of ‘shaking’ meant. Heyne seems right in supposing that Virg. was thinking of shaking the scourge, which is the common expression; he intended however also to give the image of the victim driven as it were from side to side by the force of the blows (“pulsat versatque” 5. 460), and writhing and shrinking under them. Comp. 12. 337, “Talis equos alacer media inter proelia Turnus Fumantis sudore quatit,” where perhaps the notion of scourging is meant to be combined with that of rapid motion (“cursu quatiunt” G. 3. 132), especially as ‘insultans’ follows in the next line, showing that Virg. there, as elsewhere, is imitating himself. Cerda comp. Val. F. 7. 149, “Ipsum angues, ipsum horrisoni quatit ira flagelli,” an obvious imitation of Virg. For ‘torvos’ some of Pierius' MSS. gave ‘tortos,’ a very plausible reading, supported by an intermediate correction in Pal. Tisiphone apparently has a scourge in one hand and serpents in the other, as in the imitation in Val. Fl. Heyne refers to a similar picture of Tisiphone, Stat. Theb. 1. 112, “Tum geminas quatit ira minas, haec igne rogali Fulgurat, haec vivo manus aera verberat hydro.” If we suppose the serpents themselves to be the scourge, we may say that in her right hand she grasps the culprits. This would agree with “verberat” in the line just quoted. and with the reading ‘tortos,’ with which Ribbeck comp. “torto verbere” 7. 378, G. 3. 106.
 It would seem as if the other Furies were called to carry away the culprit; but it may be to assist in the torture. The former view however is supported by the author of the Axiochus, who says of the guilty, ἄγονται πρὸς Ἐρινύων ἐπ᾽ Ἔρεβος καὶ χάος διὰ Ταρτάρον. ‘Agmina,’ see note on 4. 469, 470.
The description is continued:
when the culprit is handed over to the
Furies, then, and not till then, is the
adamantine door of the prison opened.
Serv. says, “Mittuntur, inquit, post verbera
ad aeternum supplicium. Et est
secutus ordinem iuris antiquum. Nam
post habitam quaestionem in Tullianum
ad ultimum supplicium mittebantur.”
Another interpretation, also mentioned
by Serv. and accepted by the earlier commentators,
and now by Ribbeck, supposed
these words to be the poet's, as if, just as
the Sibyl was speaking, the gates flew
open, and afforded a glimpse of the scene
within; but this would be inconsistent
with what follows, where the Sibyl calls
attention to the sentry at the gate, whom
Aeneas can see, and then proceeds to
speak of the horrors within, which he
cannot see. “Foribus cardo stridebat
aenis” 1. 449. Milton's well-known imitation
(P. L. Book 2. 879 foll.) will bear
“On a sudden open fly
With impetuous recoil and jarring sound
The infernal doors, and on their hinges grate
Harsh thunder, that the lowest bottom shook
 Custodia, i.q. “custos,” as in 9. 166, for “custodes.” So we say ‘watch’ and ‘sentry’ for ‘watchmen’ and ‘sentinels.’ The “custos” is Tisiphone, not, as Süpfl thinks, a Hydra who is compared to another and fiercer one within.
 Vestibulo and ‘limina’ are important, being contrasted with ‘intus.’ Henry remarks that the three degrees of horror are Tisiphone on the threshold, the Hydra within, and the terrific depth of Tartarus, ‘Tartarus ipse.’ ‘Limina servet’ v. 402., 2. 567. ‘Facies’ of a monster 8. 194.