A few MSS. give ‘laborem,’ a common variation.
 Iris is Juno's usual messenger, 5. 606., 9. 2, after Il. 18. 166 foll. She is here sent on an extraordinary mission, to do what is usually done by Proserpine. Whether there is any precedent for giving Iris a function in connexion with violent deaths, which of course formed a large class by themselves, or whether this is an isolated act on the part of Juno ‘omnipotens,’ does not appear. Perhaps we may compare the χθόνιον γέρας of Hermes, who is a sort of male Iris.
 ‘To separate the struggling soul from the limbs that cling to it.’ Cerda rightly remarks that the metaphor is thoughout from the palaestra. Comp. Ov. M. 6. 242, “et iam contulerant arto luctantia nexu Pectora pectoribus.” Not unlike is Lucr. 2.950, “Vitalis animae nodos e corpore solvit.”
 Henry rightly explains ‘fato’ of a natural, ‘merita morte’ of a violent death provoked by some action on the sufferer's part, comparing for the first “decessit . . . et quidem sponte . . . est enim luctuosissimum genus mortis quae non ex natura nec fatalis videtur” Pliny Ep. 1. 12, “qui partim fato partim ferro periere” Justin 9. 8, for the second “Ut caderem meruisse manu” above 2. 434, “Si nocentem innocentemque idem exitus maneat, acrioris viri esse merito perire” Tac. H. 1. 21. The opposite of these kinds of death is expressed in the next line, “misera ante diem, subitoque accensa furore:” she did not wait till fate summoned her: her death was not predestined but sudden. The distinction which Virg. intends is practical rather than philosophical, and the words employed must not be weighed too nicely. Serv. felt there was an inconsistency, asking how Dido's death ‘ante diem’ can be reconciled with Jupiter's declaration 10. 467, “stat sua cuique dies;” but his distinction between ‘fatum denuntiativum’ and ‘fatum conditionale’ scarcely removes it. In one of the passages quoted by Henry to substantiate the sense of ‘fato,’ Tac. A. 2. 71, the dying Germanicus says that if he were dying a natural death, “si fato concederem,” he should still have to complain that his end was premature. ‘Ante diem’ may be the Homeric ὑπὲρ μοῖραν Il. 20. 336; in the Iliad however things do not happen ὑπὲρ μοῖραν, as we are expressly told Il. 6. 487 foll., though supernatual interference is sometimes required to prevent such a catastrophe, as in Il. 2. 155., 17. 321., 20. 30, 336 &c. the ὑπὲρ μόρον of Od. 1. 33, 34 is not quite the same thing.
 Pal. and Gud. have ‘necdum,’ apparently mistaking the construction. The cutting off of the lock of hair (a custom referred to by Eur. Alc. 76, as well as by later writers) is explained by Turnebus Adv. 19. 17 from the analogy of sacrifices, where a few hairs are plucked from the forehead of the victim as part of the κατάργματα (“libamina prima” 6. 246), a dying man being regarded as a victim to the powers below. Perhaps we may illustrate also from the story of Nisus and Scylla. “Flaventis comas” above v. 590. ‘Vertice’ is emphatic, as the lock was taken from the crown of the head. Henry quotes Etym. M. s. v. ἀπεσκολυμμένος: σκόλλυς γὰρ ἡ θρὶξ ἡ ἐπὶ τοῦ ἄκρου ἣν ἐφύλαττον ἀκούρευτον, θεοῖς ἀνατιθέντες.
 Roscida and ‘adverso sole’ belong to the physical rainbow, which in 5. 601, 658., 9. 15 Virg., unlike Hom., makes the accompaniment of the mythological Iris. ‘Croceis’ apparently as the colour of light, v. 585. ‘Trahens’ apparently expresses length.Litus arenosum Libyae,’ I have examined ten of the Bodleian MSS., the same which I examined in reference to 5. 573 (see the Preface). Five of them read ‘ac Libyae,’ four ‘Libyae,’ one ‘ad Libyae.’ Those which read ‘Libyae’ are numbered respectively Auct. A. A. 1 (first half of 15th century), Auct. B. B. 1 (14th century), Auct. B. B. 2 (? apparently late), and Auct. F. 2. 5 (middle of 15th century). In A. A. 1 and B. B. 2 ‘ac’ is written above the line. In F. 2. 5 ‘ventoque’ appears for ‘ventosque,’ there being a blank space where ‘s’ has been erased. In B. B. 2 ‘volabat’ is written apparently by the same hand as the rest of the line, but at a later time, as if a blank space had been originally left and afterwards filled in. In A. A. 1 and B. B. 2 v. 257 precedes v. 256, but the order is corrected in the margin. The inverted order is also found in the text of one of the other MSS. which I examined, and in the margin of another. It appears then that the reading ‘Libyae,’ like ‘Trinacriis’ 5. 573 is at any rate prior to the invention of printing, so that it may have some better authority than critical conjecture.