² The text is D. L. Page's O.C.T. (Oxford, 1972), which at 984–6 is actually the text of the Cod. Florentinus Laur. 31.8. Triclinian variants will be mentioned when the occasion warrants.

³ First in his bilingual edition (Greek-German) of Agamemnon (Berlin, 1885)Google Scholar and again in his editio maior of Aeschylus (Berlin, 1914).Google Scholar

⁴ Mazon in his edition of Aeschylus in the Collection des Universités de France (Paris, 1925); Groeneboom (Groningen, 1944) and Fraenkel (Oxford, 1950) in their respective editions of this play.

⁵ This edition (Oxford, 1957), begun by J. D. Denniston and completed by D. L. Page, is referred to throughout as Denniston-Page.

⁶ ‘Aeschylus, Agamemnon 984–6’, JHS 86 (1966), 166–7.Google Scholar

⁷ H. Lloyd-Jones takes the clause the same way in his prose version of the play (Englewood Cliffs, NJ, 1970), but without endorsing the text adopted by Wilamowitz. F. Paley in his edition of Aeschylus (3rd edn; London, 1870) also interprets the clause in this way.

⁸ Il. 2.216, 249, 492, 673 and 23.297. The phrases λα⋯ν ⋯γαγ⋯νθ᾽ ὑπ⋯ τεχοс (Il. 4.407), ὑπ⋯ τεχοс ἰ⋯νταс (12.264), and ὑπ⋯ πτ⋯λιν αἰπ⋯ τε τεχοс ἵξεсθαι (11.181–2) do suggest assault, but they also occur in rather different contexts and make explicit reference to the τεχοс.

⁹ For ὑπ᾽ Ἴλιον cf. E. Or. 648. For ὑπ᾽ Ἰλ⋯ωι cf. A. A. 860, 882, 1439, Ch. 345; E. Andr. 1182, El. 881, Hec. 764, Hel. 288, Or. 58, 102.

¹⁰ Lloyd-Jones, op. cit., gives ‘Time has passed its youth since…’, and Denniston-Page hesitates between ‘Time passed its prime when…’ and ‘Time has passed its prime since…’ (but also wonders whether χρ⋯νοс is to be construed with παρ⋯βηсεν at all).

¹¹ S. O.C. 618 and E. Heracl. 900, Supp. 787, Antiop. fr. 222 N². See de Romilly, J., Time in Greek Tragedy (Ithaca, NY, 1968), p. 37.Google Scholar Whether or not they are by Aeschylus, neither Pr. 981 nor Eu. 286 ought to be taken as referring to abstract time-in-general. In Pr. 980–1: Ἑρ. τ⋯δε Ζεὺс τοποс οὐκ ⋯π⋯сταται. | Πρ. ⋯λλ᾽ ⋯κδιδ⋯сκει π⋯νθ᾽ ⋯ γηρ⋯сκων χρ⋯νοс, the sense, as the preceding verse, which is often omitted in citations, shows, is ⋯ γηρ⋯сκων χρ⋯νοс (sc. ⋯μν), i.e. ‘the time that ages for us teaches everything = as we grow older we learn everything [and Zeus is no exception]’ (see Romilly, , p. 45Google Scholar). Eu. 286 is an interpolation and so out of its context, but nothing about it indicates a reference to time-in-general and similar passages suggest supplementing it: χρ⋯νοс καθαιρε π⋯ντα γηρ⋯сκων ⋯μο (sc. ⋯μν), i.e. ‘The life span that ages at the same time as we do destroys all’. Cf. especially S. O.C. 7–8, and more generally the material cited in n. 12 below.

¹² For this way of personifying the time of living things see Ed. Fraenkel on A. 105f.; de Romilly, J., op. cit., pp. 41–9Google Scholar; and Fraenkel, H., ‘Die Zeitauffassung in der frühgriechischen Literatur’, repr. in Wege und Formen frühgriechischen Denkens (Munich, 1968), pp. 21–2.Google Scholar

¹³ H. Fraenkel seems to have noticed this difficulty; he renders ‘wir sind alt und sorgenvoll geworden durch diese Ausfahrt’, i.e. he takes the χρ⋯νοс to be the χρ⋯νοс of the members of the Chorus. But this is most unlikely. After all the Chorus were already old when the expedition left, too old to join it.

¹⁴ G. Thomson in his edition of the Oresteia (2nd edn rev.; Amsterdam and Prague, 1966)Google Scholar is one editor who does question this assumption, but his own version of the passage requires a rewriting too extensive ever to be convincing.

¹⁵ Not only is it impossible to get this sense from χρ⋯νοс … παρ⋯βηсεν, but any supposition that χρ⋯νοс ⋯πε⋯ alone could convey it here was refuted by Ed. Fraenkel in his note on the passage.

¹⁶ In his posthumous edition of Aeschylus (2nd edn; Berlin, 1859). This view has been very influential; a recent instance is Hogan, J., A Commentary on the Complete Greek Tragedies: Aeschylus (Chicago, 1984).Google Scholar

¹⁷ J. Davies had earlier conjectured ἄταс (as reported by N. Wecklein in the Appendix to his edition of Aeschylus [Berlin, 1885–93]). Hermann restored ἄγα for ἄτα earlier in the play (131) in the list of emendations appended to W. von Humboldt's translation (Leipzig, 1816). So far as I know, no one has proposed ἄγαι here, although H. L. Ahrens did suggest ⋯γ⋯ in ‘Studien zum Agamemnon des Aeschylus’, Philologus 1. Supplbd. (1860), 593–4.Google Scholar

¹⁸ In his posthumous edition of Aeschylus; the same emendation was incorporated by E. A. J. Ahrens in his text of the plays (Paris, 1842). Triclinius has ⋯π⋯, to which there are metrical objections (see Denniston-Page) nor is it otherwise attractive. Willink suggests that ΕΠΕΙ might have first come on the scene as a gloss on ΕΥΤΕ.

¹⁹ The change from -αιс to -αс is not especially difficult in itself, but the change from ⋯μβολαс ψαμμ⋯αιс to ⋯μβολαс ψαμμ⋯αс is rather less easy because it involves dissimilation of endings; likewise the change from -εμβολοιс ψαμμ⋯οιс cannot be regarded as an easy slip. And Triclinius' reading εμβολοιс ψαμμ⋯αс, being almost certainly an emendation of his own, is of no help in explaining the preceding -αс.

²⁰ C. J. Blomfield, F. H. Bothe, A. Emperius, W. Kayser, H. Weil, and S. Preuss have at one time or another favoured ψαμμ⋯αιс or ψαμμ⋯αιсιν; and D. S. Margoliouth proposed ψαμμ⋯οιс (all as reported by Wecklein). No one appears actually to have proposed ψαμμ⋯ων.

²¹ The supplement is not printed in the text of their edition probably because of the uncertainty of the corresponding line in the strophe, which they leave obelized. Denniston-Page, perhaps relying on Wecklein's Appendix, attribute 〈τ⋯〉δ᾽ to C. J. Blomfield, but he himself in his edition of the play (Leipzig, 1823) attributes it to J. Dorat (as well as to Th. Stanley and R. Porson), so that it must come, if not from Dorat, at least from some other scholar of his time (cf. Fraenkel, , Prolegomena, p. 35Google Scholar). The loss of ΤΑ may have been eased by the sequence ΙΤΑΔ, in which Ι resembles Τ, and Α resembles Δ; but there are enough cases in which a letter or two disappears without apparent help from its surroundings, in this play, e.g. (according to Page's text) at 195: 〈τε〉, 434: 〈τιс〉, 546: 〈μ᾽〉, 1252: 〈μακ〉.

²² See Greek Particles ², p. 540.Google Scholar

²³ op. cit., p. 538. He cites a number of passages in which this occurs. For δ⋯ as an allowed collocation, in which, however, each particle seems to retain its own force, again see Denniston, , pp. 548–9, 552.Google Scholar

²⁴ For the construction cf. 584: ⋯ε⋯ γ⋯ρ ἥβαι τοс γ⋯ρουсιν εὐμαθεν, ‘To be a good learner always retains its youth for old men’.

²⁶ See the references in n. 12 above.There are, ofcourse, other possibilities than χρ⋯νοс … ἄγαι. There is χρ⋯νοс … ἄγαс, which is slightly less likely palaeographically (and of course Triclinius' ⋯κ⋯ταс cannot be counted in its favour); and there is χρ⋯νωι … ἄγα, which is also inferior palaeographically. I have also considered variations with ἄται (or ἄτα or ἄταс). They are easier palaeographically, but all involve the objection as to sense that while ἄγα might be thought to have lost its force when the expedition got away from Aulis, there is no obvious reason for saying the same of ἄτα. They would also be at odds with what we know of the metrical practice of Aeschylus, who in the extant plays, although he makes frequent use of the lecythion, rarely uses the form with long middle anceps , and never in a strophe, only in an antistrophe in response to (see Supp. 385 = 374, 803 = 795; Ch. 616 = 606, 74 = (?)69).

²⁶ For ἄγα here see n. 17 above.

²⁷ For the definition of enallage and the related concept hypallage see Bers, V., Enallage and Greek Style (Leyden, 1974), pp. 1–5, 70–1.Google Scholar

²⁸ op. cit., pp. 49–58.

²⁹ Ahrens, H. L., op. cit, p. 592, showed that ⋯μβολαс can mean ‘stowings’, and this seems the only possible interpretation in this passage. The stowings, or rather the ropes, are ‘sandy’ in particular because of the sand that has accumulated about the moorings piled up by the on-shore winds during the long delay before departure.Google Scholar

³⁰ Bers adopts the expression Wortfelder from Kastner, W., Die griechischen Adjektive zweier Endungen auf -ΟΣ (Heidelberg, 1967).Google Scholar One might try to connect this expression with the Wortfeld: Water, but admittedly this is open to question. However, Kastner, , p. 72Google Scholar, does connect the expression ναΐοιсιν ⋯μβολαс (A. Pers. 297, 336) with this Wortfeld. One might also try to find in the adjective ψ⋯μμιοс a link with Kastner's Wortfeld: Geographische Bezeichnungen und Verwandtes.

³¹ I owe particular thanks to C. W. Willink for pointing this out to me.

³² op. cit., p. 114: ‘Adjektive, die an sich schon poetisch waren, wie δ⋯λιοс, φ⋯νιοс etc., werden durch die Motionslosigkeit stilistisch noch gehoben.’

³³ Enallage is not found in a few of the Wortfelder in which adjectives of two endings are found, but the discrepancy is minor and of no significance for the argument here.

³⁴ In view of the close connection in thought between (2) and (3) Willink suggests, I think correctly, that a colon rather than a full stop would be desirable in the Greek text after 987.