^{page 145 note 1} Achilles does in fact kill Troilus with a imitations of the passage (applied to other spear in some versions of the story (Stat. warriors) in Stat. Th. 10. 544 f.Google Scholar, Sil. 4. 254Google ScholarSilv. 2. 6. 32 f.Google Scholar, Eustath, . on Il. 24. 257); the f. have nothing decisive.Google Scholar

^{page 146 note 1} Varro, , De L.L. 5. 115Google Scholararma ab arcend quod his arcemus hostem; other Roman gram-marians differentiate between arma and tela (see Thes. L.L., s.v. anna). Of course the word very often means fighting equipment of all sorts, but we can fairly say that it is at least as likely to mean ‘armour’ as to mean ‘weapons’.

^{page 146 note 2} I should not press this point far if we were concerned with rapidly changing and eventful narrative; a poet who describes many related scenes in quick succession may not remember, or be concerned to remember, whether a warrior when last mentioned was wearing a helmet, or had two spears or one, or was accompanied by a charioteer or alone. But I do stress it very strongly in this case where Virgil is describing one single scene only, one picture whose visual impact is not blurred by actions leading up to it or away from it.

^{page 146 note 3} See Pearson's Intro, to the fragments of Sophocles' Troilus, and Roscher and Pauly Wissowa, s.v. Troilos. Mayer in Roscher regards Virgil's version as confused, but rightly denies that Virgil altered the legend so as to portray an armed combat. Lesky in P.-W. argues that Virgil followed an already existing legend involving armed combat, but the evidence for this is very faint indeed compared with the unarmed tradition.

^{page 146 note 4} See, apart from Roscher, and P.-W., , Ure, P. N., J.H.S. Ixxi (1951), 198 f. (on the Etruscan paintings on the Tomba dei Tori).Google Scholar

^{page 146 note 5} It is noticeable that the literary evidence always tells of Troilus ambushed while exercising his horses; it knows nothing of Polyxena and the spring, so frequent in art.

^{page 147 note 1} The evidence from Roman art suggests it (see P.-W., ad fin.), though it is much less ciear than the Greek evidence. It occurs too in Greek writers of the Roman period, Apollodorus (Epit. 3. 32Google Scholar) and Chrysostorn, Dio (II. 77–78).Google Scholar The latter specifically links the death of Troilus with the early period of the war, when the Greeks were not in possession of much of the Troad; for other-wise ‘Troilus would never have ventured outside the walls for exercise’.

^{page 147 note 2} Conceivably the Greek pictures of Tro-ilus riding one horse and leading another alongside may have given Virgil the idea of a chariot-team.

^{page 147 note 3} From the first century we may note Sen. Agam. 747 f. (an echo of Virgil) nimium cito / congresse Achilli Troile, Stat. Silv. 2. 6.Google Scholar 32 f. circum saevi fugientem moenia Phoebi / Troilon Haemoniae deprendit lancea dextrae, Silv. 5. 2 121 f.Google ScholarTroilus haud aliter gyro breviore minantes / eludebat equos (evidently an echo of Virgil's chariot).

^{page 147 note 4} So in Ausonius, (Epit. 18)Google Scholar, in a passage full of Virgilian reminiscence: Hectare prostrato, nec dis nee viribus aequis / congressu saevo Troilus Aeacidae, / raptatus bigis, fratris coniungor honori. Dictys (4. 9)Google Scholar also puts event after the death of Hector, ; Quintus of Smyrna (4. 430 f.)Google Scholar implies deliberate armed combat. Servius (on 1. 474) has a variant on the erotic element in the story as told by Lycophron.

^{page 147 note 5} Lesky in P.-W. thinks that the word congressus is significant, but it need mean more than ‘coming up against’, ‘meeting’.

^{page 149 note 1} The emotional structure and intensity of the passage has been well discussed by Plüss, Th., J.K.Ph. (1875), pp. 639–42.Google Scholar

^{page 149 note 2} It is possible that a mental association can here be made with the loss of the Palladium (Aen. 2. 166), another of the thr ‘dooms’ of Troy mentioned in the passage from Plautus cited above.Google Scholar

^{page 151 note 1} I am much indebted to Mrs. A. D. Ure and Mr. A. E. Wardman for their help in discussing with me many aspects of this article.