Pal. and two other good MSS. have ‘dictabat,’ as in 9. 323 some have “vastabo” for “vasta dabo,” varieties which support Lambinus' “nuda dabant” for “nudabant” in Lucr. 5.970. The imperfects are intended to show that while he was speaking he moved neither hand nor eye. Virg. doubtless took his description from Od. 3. 281, where Menelaus' pilot dies by a visitation of Apollo in the performance of his duty, πηδάλιον μετὰ χερσὶ θεούσης νηὸς ἔχοντα. Comp. also Od. 5. 270 foll., αὐτὰρ ὁ πηδαλίῳ ἰθύνετο τεχνηέντως, Ἥμενος: οὐδέ οἱ ὕπνος ἐπὶ βλεφάροισιν ἔπιπτεν. For ‘clavum’ Med. a m. p. gives ‘clavo,’ a natural variation, which might also be accounted for by the form ‘clavom,’ found in Pal. a m. p. and adopted by Ribbeck.
 For ‘nusquam’ one MS. (Hamb. 1 a m. sec.) gives ‘numquam,’ which Wagn. was inclined to adopt: but Forb. rightly refers to Hand. Turs. 4, p. 349, where however the most apposite parallel, Plaut. Bacch. 5. 2. 84, rests on a false or doubtful reading. ‘Nusquam discedere’ is a phrase found more than once in Cic. where we might have expected ‘numquam’ (Ep. Att. 5. 11): and so Virg. has already used ‘nusquam abero’ 2. 620. There is however generally some little force in the substitution, which here there can hardly be said to be.
 A branch is used by the god as the best instrument for sprinkling, as by Medea Apoll. R. 4. 156 foll. in putting the dragon to sleep. Heyne reminds us of the lustral bough, 6. 230. For the image of dew used in connexion with sleep see on 1. 692.
 ‘Soporare,’ to affect with sleep, is commonly applied to making persons, drowsy, more rarely, as here and 6. 420, to imparting soporific properties. The transition is sufficiently natural, especially in poetry, and may be illustrated by Shakspeare's ‘insane root that takes the reason prisoner.’ No illustration has been quoted of this supposed soporific effect of the waters of Styx. Perhaps the poet, having mentioned Lethe, added Styx, to show that this was not an ordinary sleep, but a baleful and fatal one. So Serv. “morte plenum.”
 Cunctanti of resistance 6. 211, G. 2. 236. Heyne rightly remarks that sleep may be said with equal propriety to bind or to relax the eyes. Comp. 9. 189 “somno vinoque soluti,” 10. 418 “leto canentia lumina solvit.” Here there is a special propriety in the image, as opposed to the unremitting tension which Palinurus had kept up. “Natantia lumina” G. 4. 496.
 Vix followed by ‘et’ 2. 692 note. Burm. erroneously took ‘cum’ in v. 858 as ‘quum,’ which would involve the awkwardness of referring ‘superincumbens’ to ‘quies,’ not to speak of other objections. ‘Primos’ has really the force of ‘primum,’ as in 1. 723., 3. 69: but it is also meant to be taken of those limbs, or that part of them, which were first affected by sleep. We should say ‘sleep had scarcely begun to relax his limbs,’ looking at the process as separable into parts, though the effect of each part would extend equally to the whole body: Virg. chooses to suppose one part of the body affected before another
 We need not, with one or two of the later editors, press Virg., as if the breaking away of the rudder and a part of the stern were unlikely in itself and inconsistent with v. 868, where Aeneas manages to perform the part of pilot. The account is at least consistent with 6. 349 foll.
 For ‘saepe’ Med. and one or two other MSS. read ‘voce,’ doubtless, as Wagn. remarks, from a recollection of such passages as 6. 506., 10. 873. ‘Saepe’ is confirmed, as he observes, by 4. 384, “nomine Dido Saepe vocaturum.”
 Some MSS. (including Pal. and Gud.) give ‘in auras,’ which would be the stronger expression of the two, ‘into the sky’ rather than ‘sky-ward:’ see Wagn. Q. V. 10. 1. ‘Ad’ is supported by G. 1. 408, “qua se fert Nisus ad auras.” ‘Sustulit’ is connected closely with ‘ales,’ almost as if it had been “sustulit alis,” as in v. 657 above.
 Iamque adeo: 2. 567 note. ‘Scopulos:’ Hom. (Od. 12. 39 foll., 166 foll.) says nothing about rocks: he speaks of the island of the Sirens, but in detail we hear merely of a meadow, with a pile of human bones. Virg. has apparently introduced ‘scopulos’ from a wish to rationalize the story, as if the real danger was from shipwreck. Accordingly he drops all mention of the song, employs the epithet ‘difficiles’ (comp. Cic. Div. Verr. 11, “scopuloso difficilique in loco,” where however another reading is ‘scrupuloso’), and describes the waves as even then plashing among the rocks. ‘Quondam’ is another instance (see on 3. 700, 704) of Virg. voluntarily or involuntarily separating the time he is writing of from the old heroic age.
 Tum referring to ‘iamque,’ not contrasted with ‘quondam.’ ‘Rauca’ qualifies ‘sonabant,’ as Wagn. remarks. The recurrence of the hissing sound is doubtless intentional. “Sale saxa peresa” Lucr. 1.326. Perhaps Virg. imitates Apoll. R. 2. 553, “ἤδη δέ σφισι δοῦπος ἀρασσομένων πετράων Νωλεμὲς οὔατ᾽ ἔβαλλε, βόων δ᾽ ἁλιμυρέες ἀκταί”.
 Concussus v. 700 above.
 This and the following line are the words of Aeneas, as we learn from the beginning of the next book. Heyne thought them spurious: but the only charge he brings against them, except that of frigidity, is that they are inconsistent with the fact, Palinurus having met his fate precisely because he refused to trust the sea and take his natural rest—a charge at once answered by Aeneas' ignorance of the circumstances of the case. ‘Pelago sereno’ is a singular expression (in Stat. Silv. 3. 2. 10 the reading is doubtful): but Virg. doubtless felt that ‘caelo’ paved the way for the extension of the epithet.
 Nudus et, an erroneous reading, took possession of the early editions before Pierius. ‘Nudus’ apparently combines the two notions of uncovered by the water (comp. E. 1. 61, “Et freta destituent nudos in litore pisces”) and unburied. Comp. Soph. Ant. 409, πᾶσαν κόνιν σήραντες ἣ κατεῖχε τὸν Νέκυν, μυδῶν τε σῶμα γυμνώσαντες εὖ. ‘Ignota’ as opposed to a grave in his own country. To be buried in a foreign land would have been a sorrow (comp. Soph. El. 1141, Catull. 66 (68). 99 &c.): to lie unburied in a foreign land was sorrow upon sorrow. ‘Arena’ is significant, as the corpse would be thrown up on the shore, and lie there. Serv. and Probus (quoted by Pomp. Sabinus) preserve a tradition, which Ribbeck follows, that Virg. added to this book vv. 1, 2 of Book 6, but that Tucca and Varius, or some one else (for the versions of the story vary) transferred them to their present place. The story is easily explained if we suppose with Conrads (see Introduction to this Book) that the present book is an afterthought, Virg. having originally intended to bring Aeneas direct from Carthage to Italy. In that case vv. 1, 2 of Book 6 really belong to the later draught, and in that sense are separable from the lines which they at present precede. As the poem now stands, no reader of taste will, I think, wish to disturb the arrangement of the ordinary editions, supplying as it does an affecting close to the book, which would be spoiled by carrying our thoughts on to Aeneas' safe arrival. The apparent abruptness of the opening of the next book, ‘Sic fatur,’ is doubtless due to an imitation of the opening of Il. 7, Od. 13. In concluding Book 3 Virg. chose an opposite course: but his object there was precisely the contrary: he did not wish his readers to dwell on Aeneas' last words about the death of Anchises, and so purposely carried them farther, that they might end with a sense of repose.