² The most explicit choral βοηδρομ⋯α is Held. 73ff.Google Scholar (with 121 βοηδρομ⋯сαс); cf. O. Taplin, , The Stagecraft of Aeschylus (Oxford, 1977), pp. 218–20Google Scholar, who shows that the ‘shout’ motivating an entry is often a lament. Note that here, unlike Med. 131Google Scholar ἒκλυον… βο⋯ν, the root βοα- appears only in the simile at the very end of the antistrophe (190 ⋯ναβοι), if with Badham (and as argued below) we excise ⋯νεβ⋯αсεν in 184. The dramatic technique is much more sophisticated.

³ Compare, e.g., El. and Or., in both of which the Parodos has exchanges with a soloist (Electra) in its first pair of stanzas, but is initiated as usual by the chorus; I.T., where Iphigeneia goes to fetch the chorus before taking the lead in strophic lamentation; and Andr. and Ion, where the monody preceding the Entry is in a different metre. Hypsipyle (later than Hel.) is the nearest parallel, if the chorus enters there in the middle of the first strophe (Bond, , pp. 61ff.Google Scholar; Taplin, , p. 64).Google Scholar

⁴ The probability of innovation is enhanced by metrical novelty (n. 22 below). ‘Unexpected entry’, cf. comm. on Or. 71–125, 380–4Google Scholar, and Taplin, , pp. 11–12.Google Scholar In effect, this chorus of human Φ⋯λαι arrives just in time to provide the proper ‘antiphonal’ element in the dirge (cf. I.T. 179ff.Google Scholar, etc.; comm. on Or. 960–1012Google Scholar) in place of the wished-for otherworldly chorus. For the timing and stagecraft of their entry, see further in nn. 7 and 11 below.

⁵ Heidelberg, 1969Google Scholar; text: i.139–41; commentary: ii.59–94.

⁶ Dale's plausible suggestion. Sirens, a fortiori Sirens with instruments, may have been a new theme in tragedy. For the new sepulchral motif of ‘Grabsirenen’, variously portraying these otherworldly creatures as musicians and/or mourners (foreshadowing Baroque putti and angels), see Weicker, G., Der Seelenvogel in der alien Litteratur and Kunst (Leipzig, 1902), pp. 171ff.Google Scholar, and further below.

⁷ To avoid too long a hiatus before the Chorus start singing, they must at least come into view during the strophe (see further in n. 11 below). Helen, however, only becomes aware of their entry during the antistrophe (her attention having been focused on the Tomb of Proteus?); and they do not see her (they continue to speculate about what they have heard) until she addresses them at 191.

⁸ For the stylish paregmenon, see Breitenbach, W., Untersuchungen zur Sprache der euripideischen Lyrik (Stuttgart, 1934; repr. 1967), pp. 221ff.Google Scholar; for the correction οἶκ(τ)ον, see Zuntz, , p. 42Google Scholar, and further below.

⁹ Kannicht compares the epic and Pindaric (⋯να)β⋯λλεсθαι, but see also LSJ s.v. καταβ⋯λλω 11.7 ‘lay down as a foundation’. The sense ‘begin (a song)’, perhaps neologistic here with κατα-(next in Callimachus fr. 392 Pf.), may owe something to analogy with κατ⋯ρχεсθαι (cf. comm. on Or. 960 κατ⋯ρχομαι сτεναγμ⋯ν κτλ).

¹⁰ Such interrogative hesitation is especially appropriate to formal threnody: cf. Alexiou, M., The Ritual Lament in Greek Tradition (Cambridge, 1974), pp. 161 ff.Google Scholar

¹¹ On the ‘new metre’ see further below (with n. 22). For Eur.'s penchant for ‘highlighting’ artificialities (of various kinds), cf. comm. on Or. 1214–15Google Scholar, and Winnington-Ingram, R. P., ‘Euripides: poiêtês sophos' Arethusa’ 2 (1969), 132.Google Scholar The whimsically artificial effect here will have been the more pronounced if the Chorus begin to enter silently quite early during the strophe, fitting their processional steps to the rhythm of Helen's lament (as accompanied by the αὐλητ⋯с?), but not yet seeing the singer (n. 7 above). Cf. my discussion of the unusual ‘hushed’, ‘tiptoeing’ Parodos in Or. (comm. p. 104). On the instrumental accompaniment, an interesting issue in this context, see further in n. 44.

¹² For ⋯μιλ- in contexts of ‘projection’, cf. 1471–2 αμιλληс⋯μενοс †τροχι τ⋯ρμονι δ⋯сκου†, and Hyps. fr. 764 ⋯ξαμ⋯λληсαι κ⋯ραс for the construction with acc., cf. also 546, 1386–7, Andr. 336–7Google Scholar, Hec. 271Google Scholar, Or. 38Google Scholar (comm. p. 89).

¹³ αἰαî Hermann, Murray. Hermann's alternative αἰαî αἰαî is plausible as an interpretation of ἒ ἒ (cf. Ph. 1284Google Scholar) and as a pair of αἰ⋯γματα (cf. 186), but unappealing as an expedient for padding out a third hexameter. Colometric ‘transposition’ of ἒ ἒ is wrongly ascribed by editors to Triclinius. It is clear from a photograph of L that Tr. was merely concerned to indicate (with the siglum στ [sic] and an enlarged initial letter) the beginning of the strophe at пτεροΦ⋯ροι. He therefore erased what stood before пτερ- in the right-hand column and rewrote ἒ ἒ: (sic) after, but separated by a substantial space from, п⋯νθεσιν: (sic) in the left-hand column. There is no reason to doubt that what Tr. erased before пτεροΦ⋯ροι was ἒ ἒ: (sic). The size of the erasure is exactly the same as that occupied by ἒ ἒ:(sic) in L's presentation of Hel. 661 and 662Google Scholar (in each case, as here, before a п‘Duo’ p. 62).Google Scholar

¹⁴ Not to be supported by Supp. 77/85, as I hope to show in a forthcoming article, to be published in the next issue of CQ, on a number of controversial issues in Supp. 71–86. There are several other points of contact, meriting cross-reference: see below, nn. 30, 48, 55, 57–8.

¹⁵ For the transmissional significance of L's colometric indications, either verse-end in the left- or right-hand column or the ‘divider’: (sic), or both, see Zuntz, , p. 212Google Scholar, and further in n. 17 below.

¹⁶ We can properly modify L in respect of the ‘γρ’ readings in 171a and 185 (see Zuntz, , p. 43Google Scholar, and further below): αἲ αἲνοιс entered the text as a misreading of AIΛINOIC, and θρνον as a gloss that extruded ἒλεγον. Other Tricliniana will be considered in due course. I adhere for convenience to the familiar Barnesian line-numeration, with the addition of ‘171a’ and ‘185a’ (Kannicht does not explain why he prefers to call 184 ‘183a’ 185 ‘184’ and ⋯ τι пοτ' ἒλακεν ‘185’). Earlier printed editions from Portus onwards had thirteen-line stanzas numbered 166–78 and 179–91, following the sub-Triclinian lineation of P and the Aldine.

¹⁷ ‘The division of cola in lyric passages is as much a part of the tradition as its wording, and although these divisions – being due to the Alexandrian editor and not to the poet – are not binding on the modern critic, they are always deserving of attention and may occasionally assume crucial importance’ (Zuntz, , p. 212Google Scholar). That needs some qualification as to ‘the Alexandrian editor’: Zuntz's study of Hel. 625–97Google Scholar in the light of P. Oxy. 2336 established that the amount of colometric agreement virtually guarantees that 77 and L have a common ancestor, presumably the Alexandrian text as lineated by Aristophanes of Byzantium c. 200 b.c.; but he should have focused more attention on the visible colometric divergences in 634–5 and 650–1 (‘Duo’ pp. 51, 59Google Scholar). These, and other features in L's text of 625–97 (‘Duo’ n. 56), suggest that we must visualize a ‘sub-Aristophanic’ ancestor of L with a fair number of deviant, yet still rational, verse-divisions. Caution is the more necessary in a text convicted of corruption. My ‘ancestral lineator’ is not necessarily Aristophanes in respect of every single verse-division; but in principle the erroneous iambic interpretation of our passage probably does go back to 200 b.c. (at least).

¹⁸ But L's reading is often misreported as the vox nihili ⋯μ⋯θ', without mention of the verse-division, пαρ ⋯μ⋯ θ' is a possible phrase in itself (it evidently satisfied Triclinius), and could be a quite ancient error; for the normality of elided τε at the beginning of a verse, cf. Zuntz, , p. 232.Google Scholar But the chances are that the lineator correctly read εμε-/θεν, overlapping thus in accordance with a misinterpretation which treated -θεν υпο μελα- θρα νυχια пαι- as a symmetrical iambic dimeter ; cf. n. 20 below.

¹⁹ e.g. 194 να⋯ταс 'Аχαιν τιс ~ 213, αἰὼν δυсα⋯ων τιс ~ 213 αἰὼν δυсα⋯ων τιс, cf. 193 ~ 213, 200 ~ 219, 242, 352, 355. The notation ‘ρα’ is convenient, though is really only a syncopated form of tr; cf. West, , p. 103Google Scholar, Dale, , LM², p. 93.Google Scholar

²⁰ As suggested below, the lineator's text of 175–7 may have run пεμψειε ΦερсεΦαссα Φονι (a)/αχαριτ(a) ιν εпι δακρυсι пαρ εμε-/θεν κτλ. (n. 18 above). That is consistent with the present hypothesis, while accounting for χαριταс and the hypermetric verse п- Ф- Φονια.

²¹ It is unlikely that he edited пαιναс (anticipating Triclinius), with -/αναс in responsion with… κλαγγα(ι)с or -/γαιсιν the sing. пαινα is so obviously superior (pace Dale). Other anomalies suggest that he interpreted the paradosis separately for each stanza, despairing of proper responsion. For the compound problem at…μ⋯χαλα γ⋯αλα/κλαγ⋯с…, see n. 73 and pp. 95ff.

²² West, (GM 102)Google Scholar rightly draws attention to the novelty of the genre here: ‘But in Helen we suddenly get this’ (followed by citation of 167–78). Apart from its special prominence in Hel., the examples cited (cf. also Wilamowitz, , pp. 269ff.Google Scholar, and Dale, , LM², pp. 92–3Google Scholar) are in Euripidean plays later both chronologically and alphabetically.

²³ For the ‘non-integral’ (appositive) use of the internal accusative, cf. Barrett, on Hipp. 752–7Google Scholar, Diggle, in Dionysiaca… studies… presented to Sir Denys Page… (Cambridge, 1978), pp. 171–2.Google Scholar ‘Congruent tears’ cannot define the action μ⋯λοιτε.

²⁴ Wilamowitz appears to have begun a new clause at αἰλ⋯νοιс т', with only a comma after μ⋯λεα. But he did not elucidate that, so as to explain why Helen should wish Persephone to send ‘congruent δ⋯κρυα…μ⋯λεα’, or how μουсεīα… could be added to that object-phrase.

²⁵ Wilamowitz had treated пαρθ⋯νοι…Λ⋯βυν‖λωτ⋯ν… λ⋯βηι as two periods of six and nineteen metra (the latter already overweight), with an unnatural period-end in mid phrase.

²⁶ West defines пνγοс as ‘a very long period in uniform rhythm’. The longest пν⋯γη occur in comedy, e.g. Ar. Ach. 266–79Google Scholar (Dale, , LM², p. 198)Google Scholar, which runs to thirty-three iambic metra (following a three-dimeter period with catalexis). Note that it not only has the proper rhythmical uniformity – it also falls into sections, demarcated by syntax-breaks at verse-end, of only ten, fourteen and nine metra.

²⁷ Dale desperately (one supposes) offered the plea that ‘this is an operatic aria whose words must not be expected to bear too close a scrutiny of their meaning’. Formally, this is the opening of the Parodos; and in any case Eur. was not Timotheos. I hope to show that every single word pulls its weight.

²⁸ It is immaterial whether the long ancipitia feature in an iambic metron or a trochaic . I might have added 171a (Φ⋯ρ-); but West contrives to scan с⋯ριγγαс ἣ Φ⋯ρμιγγαс as ia pa, removing any vestige of isometric balance in the isometric phrases (cf. n. 32 below).

²⁹ There is uncertainty at 191 ~ 210, but we do not have to follow Triclinius and Murray in writing αἰα 〈αἰα/ὦ〉 δα⋯μονοс пολυсτ⋯νου to correspond with ἰὼ ἰώ /θ⋯ραμα βαρβ⋯ρου пλ⋯ταс. Kannicht accepts Wilamowitz's reduction of ἰὼ ἰώ to ὦ in 191, giving a syncopated trimeter (sp lk). ἰὼ [ἰώ] θ⋯—~ αἰα 〈ὦ〉 is a possibly better alternative (ba lk, cf. Or. 965/76 with ἰώ ὦ); ἰώ is appropriate to this address (cf. n. 7 above). There are some long ancipitia in the conjectures considered by Kannicht at 358–9 τι τε †с⋯ραγγ' ⋯οιδα⋯ сεβ⋯ζον† Пριαμ⋯δαι пοτ' ⋯μΦ⋯ βουсτ⋯θμονс; but that is enough to condemn them. Accepting с⋯ριγγ- (Musgrave) and Hermann's сεβ⋯ζον 〈τι〉 (τι omitted before п), I see nothing wrong with τι τε с⋯ριγγαс (pl., see n. 44 below) -/δαι (local dat.) сεβ⋯ζοντι Пρια-/μ⋯δαι κτλ (much the same overlapping cretic pattern as Tro. 1091f./1110f.); for Paris' devotion to ‘piping’ on Ida, cf. I.A. 573–8Google Scholar; for the ironical use of сεβ⋯ζειν, cf. Med. 155Google Scholar (= colere, not venerari, cf. Fraenkel, on A. Ag. 1612).Google Scholar

³⁰ Supp. 71–86 (n. 14 above) and A. Pers. 115–25Google Scholar are earlier exx. of ia-tr with only short ancipitia; Or. 960–81Google Scholar is another such stanza-pair. Bacchyl. 19 (Wilamowitz, , 393)Google Scholar shows a similar principle operating in a distinctive enoplian genre with almost all the link-ancipitia short. There is further work to be done on theoretically anceps positions which are consistently treated as short, whether in particular places in the verse (esp. before diaeresis) or more extensively (esp. in conjunction with resolution and/or syncopation); cf. comm. Or. pp. 113, 288.Google Scholar

³¹ Dale nodded uncharacteristically when she wrote: ‘No one could think λωτ⋯ν ἢ с⋯ριγγαс αἰλ⋯νοιс in responsion to θ⋯λпουс' ⋯μΦ⋯ δ⋯νακοс ἓρνεсιν very satisfactory’. Somehow losing sight of κακοс, she failed to see that ἢ с⋯- [ἢ Ф⋯ρ-]… κακοîс corresponds perfectly with θ⋯λпουс'… ἒρνεсιν.

³² We should be happy to be spared the repetition ἢ -ιγγαс ἢ -ιγγαс. Isometric jingles are, indeed, a familiar feature of E.'s lyric style; but is not a metrical unit. Contrast the properly balanced units in 348 с⋯ γ⋯ρ ⋯κ⋯λεсα, /с⋯ δ⋯ κατ⋯μοсα, Or. 170Google Scholar οὐκ ⋯Φ' ⋯μν/οὐκ ⋯п' оἴκων, Hec. 1099Google Scholar пο τρ⋯пωμαι;/пο пορευθ; and many similar examples cited by Diggle, on Phaethon 99.Google Scholar

³³ Trimeters and/or monometers first appear in 191/210 (n. 29 above) and 195/214 (). Further variations appear in 229ff. and 330ff. Cf. the looser texture of the second stanza-pair in Or. 140–207 (comm. pp. 106, 112).Google Scholar

³⁴ cf. A. Pers. 115f./120f., Ag. 974f./988f. For the cadential (catalectic) effect of the lekythion in trochaic contexts, cf. Parker, L. P. E., ‘Catalexis’, CQ 26 (1976), 21ff.CrossRefGoogle Scholar

³⁵ 7–14 is a single period, all the verses ending with except for the final lk. It does not follow that we need to indent the whole of 8–14. Indentation is best used only when the colon (or dicolon etc.) is not coterminous with the verse; a practice which, applied to this patterned sequence of eight dimeters, shows better the symmetrical treatment of verse- and colon-end in str. and ant. ‘Cola’ are by definition delimited by word-end (comm. Or. p. xxGoogle Scholar). The word-end can be after a prepositive or before a postpositive (‘weak diaeresis’); but elision counts as word-overlap, as at 181 θ⋯λпου-/с' ⋯μΦ⋯.

³⁶ Dale (LM ², p. 93Google Scholar) called it ‘a sort of inverted ithyphallic’. It is more to the point to think of the ithyphallic in connection with the unusual verse (tr pa) in 10 and 13, which at once hints at and cheats expectation of a clausula (‘not yet’).

³⁷ For…… within a period in late Eur., cf. comm. on Or. 985–7 and 1377–9Google Scholar. But cf. also Or. 167/188Google Scholar (comm. p. 112), where I prefer an emendation that gives 2ba rather than 2ia.

³⁸ For 4ba runs, cf. comm. on Or. 1294–5Google Scholar. This one could, of course, be analysed as palimbacchiac with word-overlaps; but Eur. often has iambic cola in the middle of mainly trochaic sequences and vice versa, and there is no good reason for not lineating them as such; cf. comm. on Or. 982–4, 1369–74, etc.Google Scholar

³⁹ Kannicht’s… ἒρνε' /ἒνθεν… is otherwise unacceptable: there is no other elision of this type at verse-end in 167–251 or 330–74, and the elision is particularly unwelcome where a pause is appropriate. For vindication of the transmitted ἒρνεсιν, see further below.

⁴⁰ Paregmenon, cf. n. 9 above; for the favourite δακρυ- δακρυ-, cf. 195, 365, Or. 335, 1308Google Scholar, Tro. 605Google Scholar; Breitenbach, , pp. 222–4.Google Scholar

⁴¹ Those who have given п⋯μψειε a subject other than Φερс⋯Φαссα have also emended μουсεα, e.g. Hermann, who accepted Fix's μοс' ⋯εсα and wrote ΦερсεΦ⋯ссαι (cf. n. 59 below); other proposals only merit oblivion. The collocation of acc. and nom. n. pls. may be unusual, but so too is the new metrical genre with its need for short pendent syllables. See further below, where μουсεα (nom.) is more fully interpreted, in conjunction with the proposal п⋯μψαιτε.

⁴² The metrical pattern might seem to favour two self-contained phrases; but I slightly prefer Wilamowitz's punctuation (no commas before the one after Σειρνεс), which more poetically treats νε⋯νιδεс as adjectival.

⁴³ For Dale's suggestion, cf. n. 6. As in Or. 311 ffGoogle Scholar. there is ad hoc ‘myth-determination’ in the attributes. Homer's Sirens had been feminine and two in number (Od. 12.39, 52, etc.Google Scholar), and in Alcm. 30 Page the Muse is ⋯ λ⋯γηα Σηρ⋯ν; but for male Sirens, cf. Weicker, , figs. 19 (Orpheus-like) and 72Google Scholar, also Simonides 607 Page. In art sirens occur often in pairs or groups of three (Weicker, , 162Google Scholar). The number here is naturally indefinite: cf. Or. 317ff.Google Scholar, where the Eumenides are similarly indefinite in number as a ‘thiasos’, though in Or. 408 and 1650Google Scholar they are a triad (see comm.). Sirens are properly ‘chthonian’, but they were not canonically ‘daughters of Earth’ (‘honorific’: cf. Δι⋯с in Hipp. 534Google Scholar, Or. 5Google Scholar); in S. fr. 861 Radt they are Φ⋯ρκου κ⋯ραι. For ‘avian’ threnody, cf. 1107ff., El. 151ff.Google Scholar, Tro. 146ff.Google Scholar, Ph. 1515ff., etc.Google Scholar

⁴⁴ Plural ‘pipes’ can, of course, be played by a single instrumentalist (cf. Ion 498Google Scholar), whether as ‘twin pipes’ or ‘Pan-pipes’. I assume, but cannot prove, that Helen's ἂλυροс ἒλεγοс was accompanied in performance by the αὐλητ⋯с (the instrumental accompaniment perhaps beginning at 167; cf. n. 11, also comm. on Or. 145–6Google Scholar). If that is right, it may well be deliberate that the phrasing here is consistent both with a wish for an appropriate instrumental accompaniment to Helen's lament (highlighting the artificiality?) and with a wish for ‘congruent’ wind-accompaniment for the dirge as performed by the Sirens in the other world.

⁴⁵ cf. Ph. 190–2, 808–10, 1579–80Google Scholar, Ba. 421, I.A. 1036–9Google Scholar; Breitenbach, , p. 246Google Scholar. The ‘framing’ effect here is like Ba. 526–7Google Scholar ἲθι, Διθ⋯ραμβ', ⋯μ⋯ν ἂρсενα τ⋯νδε βθι νηδ⋯ν and 547–8 τ⋯ν ⋯μ⋯ν δ' ⋯ντ⋯с ἒχει δώματοс ἢδη θιαсώταν.

⁴⁶ Others may have supplied δ⋯κρυсι or θρ⋯νοιс or п⋯νθεсι here, see above. The omission of γ⋯οιс in the Aldine, and thereafter in all the earlier printed editions, is interesting, but presumably accidental.

⁴⁷ Kannicht gives alternative explanations of the dat. with μ⋯λοιτ': ‘entweider (1) das Ziel (vgl. Cho. 935, Ant. 233–4, Io 846) oder (2) den Beweggrund (vgl. Phoen. 1043) des μολεν'. (1) leans on false parallels (personal datives with μολεν virtually ‘come to pass’); (2) confuses the invocation to the Sirens with the threnody to be sung by the Sirens with Helen.

⁴⁸ μ⋯λοιτε is in line, prima facie, with 1111f. ἒλθ' 〈ὦ〉…θρ⋯νων ⋯μο⋯ ξυνεργ⋯с Supp. 73–4Google Scholar ἲτ' ὦ…(n. 58 below), Hyps. I.iv.9 ἒλθοι (above); also H.F. 787fGoogle Scholar. βτε…сυναοιδο⋯ ν⋯μΦαι But a conventional deity-invocation with ‘come’ is less appropriate here, since (a) the Sirens are to sing ‘in the halls of Night’; (b) they can be benign creatures in that milieu, but one does not wish for close contact with them in this world. For a possibly analogous ancient corruption, cf. the responsion-flaw at ἢλυθε in Or. 813.Google Scholar

⁴⁹ Passages illustrating that ‘if you wish to play a mournful tune, you use the flute and not the lyre’ are collected by Diggle, in PCPhS 20 (1974), 11–12Google Scholar; cf. also Haldane, J. A., JHS 85 (1965), 40–1CrossRefGoogle Scholar. A lyre could be used for a lament, as by Orpheus, in Hyps. 1.iii.9Google Scholar and by Apollo, in H.F. 348Google Scholar; but in both passages there is oxymoron in the atypical use of the instrument. In Hyps. 1.iv.5ff. ‘sweet-toned lyre-lament’ is by implication inappropriate to express the singer's grief.

⁵⁰ cf. n. 6 and n. 55 below. Pairs of sirens with pipe(s) and lyre (Weicker, , 172Google Scholar) symbolize complementary musical genres (cf. Alc. 446–7Google Scholar). Weicker's earliest sepulchral example (Berlin no. 755) dates from the late fifth century.

⁵¹ Weicker, , figs. 13, 83ff., etc.Google Scholar

⁵² Beazley, , ARV ²557, 120Google Scholar, illustrated in his Der Pan-Maler (Berlin, 1931), pl. 6Google Scholar. Weicker illustrates a single piping siren of the later, less avian type in fig. 89.

⁵³ See Kannicht; but he surprisingly omits AP 6.348.1–2 αἲλινον ὠκυμ⋯ρωι με λεχωΐδι τοτο γεγρ⋯Φθαι/τс Διοδωρε⋯ου γρ⋯μμα λ⋯γει сοΦ⋯ηс, which may suggest how the adjectival use was developed. It is still possible there to take the initial αἴλινον as conventionally exclamatory, but the phrasing αἴλινον…τοτο…γρ⋯μμα is at least on the way to meaning, or being understood as, ‘this doleful inscription’.

⁵⁴ cf. Kühner-Blass, i. 251Google Scholar. The non-Attic -εεινόс is associable with the non-Attic neuter ἔλεοс (ibid, ii.296). Since ἔλεοс is masc. in Attic, the form ἔλεινόс is not necessarily a contraction of ⋯λε-εс-νόс; and, even if it is so in origin, it does not follow that it was so regarded by Eur. (especially in a metrically innovatory context). One would, of course, welcome a parallel; but ⋯λεινόс occurs elsewhere in Attic poetry only at Hel. 992Google Scholar ⋯λεινόс ἥν ἄν μλλον ἤν δραсτήριοс, ?1164 (conj. Nauck), S. Tra. 528Google Scholar, O.T. 672Google Scholar, Phil. 870Google Scholar (-с) and 1130.

⁵⁵ πάεα in 684 (‘Duo’ p. 67Google Scholar) is the appropriately chosen word for Leda's ‘suicidal grief’. For сπαραγμόс, cf. 374, 1089, Hec. 653–6Google Scholar, Or. 961–2Google Scholar (with comm.); Collard, on Supp. 48–51aGoogle Scholar, Denniston, on El. 146–9Google Scholar, and further in my forthcoming article on Supp–86 (n. 14 above). The ‘blood’-motif, for the moment implicit, becomes explicit in 175 below. For the ‘rhythm’ of tears and сπαραγμόс, cf. Hipp. 1464Google Scholar δακρύων πίτυλοс (comm. Or. p. 360Google Scholar) and Tro. 1235–6Google Scholar ἄραсс ἄραссε κρτα, πιτύλουс διδсα χειρόс. κομμόс is more obviously rhythmical. For ‘Klagende Grabsirenen’ thus (variously) engaged, see Weicker, , 173ff.Google Scholar; many analogous mourners can be seen on the monuments in Westminster Abbey.

⁵⁶ A μονсεον is properly a place sacred to the Muses and/or devoted to musical performance (cf. Ὠιδεον). For the plural, cf. 1107–8 where the nightingale is exquisitely ‘sitting within perched halls of song (μονсεα κα⋯ θάκονс ⋯νίζονсαν) beneath leafy coverts’, and Ar. Ran. 93Google Scholar χελιδόνων μ-, parodying the Euripidean κιсс⋯νων μονсεον (fr. 88; ⋯ηδ- Meineke, χελιδ- codd.), in a reference to twittering poets. For the ‘personifying’ use of the place-word Kannicht rightly compares εὐνατήριον (Or. 590Google Scholar, Antiope 48.101 Kambitsis) and βονλεντήρια (Andr. 446Google Scholar, A. S.c.T. 575), but fails to show that such metonymy can be used when the persons concerned are on the move. ‘Monasteries’ can denote monks, but not itinerant monks. The μονсεον of the Sirens is located in Hades (cf. Pl. Crat. 403dGoogle Scholar); and they cannot be referred to in the same breath as ‘halls of song’ and as ‘sent’ from Hades to Helen.

⁵⁷ cf. A. Ag. 283 etc.Google Scholar (of the fire-beacon), and comm. on Or. 617Google Scholar. For the ‘choric’ point, cf. Supp. 73–5 (next n.).Google Scholar

⁵⁸ ‘Go (i.e. dance) as fellow-mourners the dance which…’ (to be discussed further, see n. 14). The similar predicative phrase ξυνωιδο⋯ κακοс…ξυναλγηδόνεс is metonymic there also (abstract ‘grievings’ for personal ‘mourners’).

⁵⁹ So Hermann and others (n. 41 above). But to write Φερсεφ⋯ссαι introduces an unacceptable long anceps at either πεμ or -сαι.

⁶⁰ Surprisingly, no one seems to have proposed here the easy, but probably wrong, alteration of λάβη(ι) to λάβοι.

⁶¹ H.F. 1061, 1133Google Scholar (CQ 38 [1988], 90–2Google Scholar), Or, 319, 621, etc.Google Scholar; Breitenbach, , pp. 236–8.Google Scholar

⁶² cf. φόνια μνсαρά El. 1178Google Scholar, ἄνομοс ἄχαριс Andr. 491Google Scholar, ἄπολιс ἄφιλοс I.T. 220Google Scholar, etc. For the advanced (so emphatic) position in the clause, before ἴνα and widely separated from the noun, cf. Or. 162ffGoogle Scholar. ⋯τ' |⋯δίκαсε.

⁶³ The separation here by eleven words is near the upper limit (cf. n. 45). For the Corruption by false assimilation, cf. Alc. 424 (corr. Diggle).Google Scholar

⁶⁴ Note that, if the lineator read πεμψειε φερсεφαссα φονι(α) | αχαριτ(α) ιν επι δακρνсι παρ εμε | θεν…, his verses in 175–7 were rational iambic dimeters (n. 20), but not if he read αχαριταθ ιν with the same division after εμε.

⁶⁵ cf. comm. on Or. 174–8 and 1225–6Google Scholar (δμα Ννκτόс).

⁶⁶ Diggle rightly rejects Markland's metrically questionable νέκνсι μελομέναν the sense ‘unhappy music for the dead’ is unexceptionable there. Neither is there any need here for παινα νέκνсι μελόμενον, after Lobeck (παιναс ν- μελομένουс) and others.

⁶⁷ I take the gen. there to be objective (‘paean honouring the dead Agamemnon’). For a somewhat similar equivalence of gen. and dat., cf. comm. on Or. 123.Google Scholar

⁶⁸ ἔτνχον, cf. Or. 1426Google Scholar (like ⋯τύγχανον Or. 866Google Scholar, Ba. 215Google Scholar). ⋯μφί ‘near, by’, as in 359, I.T. 6Google Scholar, Or. 1310Google Scholar, etc. As in Hipp. 125ffGoogle Scholar. the ὔδωρ is presumably fresh water, though the epithet gives it a marine colour (cf. 1502, I.T. 7Google Scholar). We have been told in the Prologue both that the palace of Proetus is near the Nile (1–3), and that it is on the island of Pharos (5, del. Dingelstad). We are not expected to refer to a map.

⁶⁹ A characteristic ‘chiaroscuro’, cf. comm. on Or. 821–2Google Scholar. Note also the more ‘tripping’ rhythm, with three resolutions in the second verse.

⁷⁰ ⋯ν⋯, cf. comm. on Or. 329–31Google Scholar (‘environmental’). ἔλιξ, usually substantival, is applied in poetry to ‘curly’ κόμαι, tender and/or luxuriant in growth, of various kinds. The treatment of ἔλιξ in LSJ needs revision. The rendering ‘on the tangled grass’ (followed by Dale) is plainly wrong here; and ἔλικα does not = βον in Ba. 1170 (see Dodds). ἔλιξ (A) and ἔλιξ (B) are scarcely distinct words. For the variously adjectival and substantival use, cf. Eng. ‘spiral’.

⁷¹ Discussed in CQ 38 (1988), 94f.Google Scholar, where I missed the ἔρειсμα point (‘supported on’) in defending ⋯μφί against Elmsley's excision. For the word-order there, cf. Andr. 511Google Scholar μαсτοс ματέροс ⋯μφ⋯ сс.

⁷² For the ⋯π⋯ κοινο preposition, cf. comm. on Or. 1449–51Google Scholar, Bruhn, , Anhang §171 viGoogle Scholar (also ibid. §237 for some phrases like ‘the sun and his golden rays’ = ‘the sun's golden rays’), and Kiefner, G., Die Versparung (Wiesbaden, 1964), pp. 27ffGoogle Scholar. χρνсέαιсι|〈ξὺ〉ν αὐγαсι (instrumental) may be a possible alternative for those who still prefer the Triclinian ⋯λίον.

⁷³ The metrically identical verses θάλπουс' ⋯μφ⋯ δόνακοс ἔρνεсιν and κλαγγ⋯с Παν⋯с ⋯ναβοι γάμουс probably reflect the lineator's colometric interpretation; this false internal responsion is unlikely to be fortuitous. But it is conceivable that the lineator intended |πουс(a) here and either κλαγ-|γα(ι)с (similarly) or κλαγ-|γαιсι (~ παι-|ανα) in 189f. We cannot be sure that all the lineator's verse-divisions, especially if in the middle of a word, have been correctly transmitted (cf. n. 17).

⁷⁴ Kannicht impossibly punctuates ἔνθεν οἰκτρ⋯ν ⋯νεβόαсεν as a separate sentence. Helen did not shout either from the waterside or as a consequence of the ‘chance’ described (ἔτυχον); we cannot understand the unexpressed subject as ‘she (Helen)’; and the following asyndeton is intolerable. Attempts to save ⋯νεβόαсεν by adding words have failed miserably. Murray's ‘〈and of my mistress〉, whence she shouted…, I heard… what she shouted…’ is grotesquely clumsy, though Dale looked for the truth on similar lines.

⁷⁵ The legitimacy of in responsion with is unpersuasively argued by West (GM 103f.Google Scholar). It needs to be emphasized that the attested lineation is evidence that ⋯ τι ποτἔλακε were not treated as corresponding metra in antiquity.

⁷⁶ In support of ⋯νεβ⋯αсεν…οα…⋯ναβοι Kannicht compares Il. 3.2–8 (cited in n. 87 below). Such parallels can equally be invoked to show how $⋯νεβ⋯αсεν could have entered the tradition falsely, perhaps by way of the margin.

⁷⁷ See nn. 11 and 44 above; for λακεν of pipe-accompanied song, cf. Alc. 346fGoogle Scholar. пρ⋯с Λ⋯βυν λακεν αὐλν.

⁷⁸ See Seaford, R. A., ‘The Tragic Wedding’, JHS 107 (1987), 106–30.CrossRefGoogle Scholar

⁷⁹ Murray's interpolation of ‘my mistress’ (n. 74) was misconceived for a different reason, as Dale observed. After the ladies of the Chorus have entered, Helen recognizes and addresses them in 191–2 as ‘Greek κ⋯ραι, victims of barbarian piracy’, and laments her fate; they, on their side, are already familiar with Helen's famous story, and they address her respectfully as п⋯τνια (224). It does not follow that they are Helen's servants (for the use of п⋯τνια Dale might have compared Or. 1249Google Scholar). Rather, ‘outsider’-status has already been suggested by entry from outside the palace with motifs reminiscent of the choral entries in Med. and Hipp, (for other female choruses of sympathetic ‘outsiders’, cf. El., Ph., Or., I.A.); and it will be confirmed at 313, where they ask пс δ' εὐμενε⋯δ' ⋯ν δ⋯μοιс ἒχεс;

⁸⁰ μυχ⋯θεν also A. Ag. 96Google Scholar; cf. οἲκοθεν, θ⋯ραθεν (first at Andr. 953Google Scholar), κλιс⋯ηθεν Il. 1.391, etc. (Kühner-Blass, ii. 308.4Google Scholar). For λακεîν with an adverbial expression of place, cf. also Hec. 1110Google Scholar, IT. 976Google Scholar, Or. 329Google Scholar. Note that АΥΛΑΘΕΝ could have dropped out quite easily after ΕΛΑΚΕΝ, either before or after the text was lineated. Could it also, perhaps, have played some part in generating ANEBOACEN?

⁸¹ For the multivalent ‘hall/court’ constituted by the stage-building and the acting-area immediately in front of it, cf. comm. Or. p. xl.Google Scholar Entering choruses allude to the αὐλ⋯ in El. 168 (ποτ⋯ с⋯ν ⋯γρ⋯τειραν αὐλ⋯ν) and I.T. 128Google Scholar (πρ⋯с с⋯ν αὐλ⋯ν); cf. also Hec. 171Google Scholar, Ph. 1536Google Scholar, Or. 1257Google Scholar, Ba. 630.Google Scholar

⁸² cf. A. Eum. 256. We may prefer the tribrach word for exact responsion, but we have no right to demand it in the light of the adjacent -εсα (~ δ⋯κρυсι).

⁸³ In its musical sense (LSJ s.v.II)Google Scholar a ν⋯μοс is never simply a song, but rather a type of song, or ‘song-mode’ (in accordance with the root idea of ‘custom’), nearly always with a defining epithet (⋯ρθιοс, Βοιώτιοс, βακχεοс, etc.), and with an expressed or implied definite article (cf. A. S.c. T. 952Google Scholar ⋯πηλ⋯λαξαν Ἀρα⋯ τ⋯ν ⋯ξὺν ν⋯μον).

⁸⁴ cf. Or. 1468Google Scholar φυγ⋯〈δ〉ι δ⋯ ποδ⋯(s.v.l.), 1499 δραπ⋯ταν…π⋯δα, Hel. 1301Google Scholar δρομ⋯δι κώλωι.

⁸⁵ cf. Phaethon 171 Diggle ἵει δ᾽ ⋯φ᾽ ⋯πτ⋯ Πλει⋯δων ἔχων δρ⋯μον, Rh. 798Google Scholar ἵεсαν φυγι π⋯δα. With δρ⋯μον γοερ⋯ν we might then compare Thuc. 4.128 ⋯ν φοβερι ⋯ναχωρ⋯сει, but it would be more natural to write γοερ⋯с (fem., cf. Ba. 992Google Scholar Δ⋯κα φανερ⋯с, etc.; Kühner-Blass i.535–6). ‘Running’, however, is not the relevant point here; and no support should be looked for in the possibility that Triclinius found a transmitted ποδ᾽' in the margin nearby (see below).

⁸⁶ Not ⋯ρεсι ὅμαδον, since an overlap at ὅ-|μαδον would conflict with the pattern of articulated dimeters; but word-corruption and wrong word-order can go hand in hand. ὅμαδον could have been skipped before ⋯ρεсι and/or transposed within the verse (bringing ὅμ- nearer to γοερ⋯ν); φυγαδαγαμον is then a garble of φυγα…μαδον, perhaps helped by illegible suprascription.

⁸⁷ ὅμαδον…ἱεсα γοερ⋯ν thus frames the phrase, with immediate repetition in the simile of the first of the paired nouns in the leading clause: cf. Il. 3.2–3 Τριεс μ⋯ν κλαγγι τ᾽ ⋯νοπι τ᾽ ἴсαν ⋯ρνιθεс ὥс, | ἠΰτε περ κλαγγ⋯ γερ⋯νων π⋯λει οὐραν⋯θι πρ⋯ (cited by Kannicht as supporting ⋯νεβ⋯αсεν…οα…⋯ναβοι).

⁸⁸ cf. Hes. Th. 119 Τ⋯ρταρ⋯ τ᾽ ἠερ⋯εντα μυχι χθον⋯с εὐρυοδείηс (and Supp. 545, 926Google Scholar, Tro. 952Google Scholar, H.F. 37Google Scholar, Ion 1239Google Scholar, etc.). μυχαλ⋯с is acceptably in line with words like ⋯παλ⋯с, ⋯μαλ⋯с τροχαλ⋯с, for the accentuation of such adjs. (several of them rare), see Chandler, , Greek AccentuationGoogle Scholar, §389. μ⋯χια τ⋯ρταρα might have been understood similarly, but the Hesiodic sense of μ⋯χιοс is ⋯ν μυχι οἰκ⋯αс or ναο (cf. West on Op. 523 and Th. 991). For virtually synonymous cognate adjs., cf. πιναρ⋯с (El. 184)Google Scholar and πινώδηс (Or. 225).Google Scholar

⁸⁹ Presumably cognate with γ⋯ηс (also usually pl.), cf. Ba. 13Google Scholar and A. Supp. 550Google Scholar; I.T. 1235Google Scholar, P. N. 10.56, Ar. Thesm. 110Google Scholar, etc. ‘Concave places’, able to hold soil and water, afford the only cultivable ground in much of Greece. The archetypal ‘Pythian’ passage is Hes. Th. 499 Πυθο ⋯ν ⋯γαθ⋯ηι, γυ⋯λοιс ὑπ⋯ Παρνηссοο which follows the usual formulaic pattern if understood as ὑπ⋯ γ- γ- (so West; but there is no need to write ὕπο, cf. n. 71 above and comm. on Or. 94)Google Scholar; likewise h.Ap. 396Google Scholar γυ⋯λων ὑπ⋯ Π -(cf. h.Hom. 26.5 Νὐсηс) and P. Py.. 8.63 Πυθνοс ⋯ν γυ⋯λοιс. The characteristically Euripidean use of γ⋯αλα as ‘vox propria’ in reference to the Pythian precinct, as in Andr. 1093Google Scholar θεο γ⋯μοντα γ⋯αλα, Ion 76, 220, 233, 245Google Scholar, Ph. 237Google Scholar μεс⋯μφαλα γ⋯αλα Φο⋯βου, may reflect a misinterpretation of Hes. Th. 499Google Scholar as ‘holy Pytho, (the) γ⋯αλα under Parnassos’. I.A. 1052Google Scholar χρυс⋯οιсιν … ⋯ν κρατ⋯ρων γυ⋯λοιс seems to combine the Homeric use of γ- to denote concavities in metalwork with the kind of gen. first seen in S. Phil. 1081 π⋯τραс κο⋯λαс γ⋯αλον. The sing. γ⋯αλον is unusual there, and the true meaning of Phil.'s envoi to the cave may well be something like ‘O my territorial dell (γ⋯αλον = γ⋯ηсх) of hollow rock’, the sense ἄντρον2 or μυχ⋯с being given by the phrase as a whole.

⁹⁰ Except perhaps by Eur. himself in Or. 331Google Scholar, where I now think that ἵνα μεс⋯μφαλοι λ⋯γονται μυχο⋯ combindes the ideas μεс⋯μφαλοс ⋯ст⋯α (Ion 462)Google Scholar and μεс⋯μφαλα γ⋯αλα (Ph. 223)Google Scholar in a new formulation, perhaps influenced, like so much in Or., by S. Phil. (see last n.).

⁹¹ Dittography might account for μ⋯αλα (L^ac) before γ⋯αλα (though without explaining where the μ came from); but the transmitted status of μ⋯χαλα. is likely to have been confirmed by Triclinius (see below).

⁹² Widely accepted (cf. Jebb, on S. Phil. 1081Google Scholar); but μ⋯χατοс is a late form (first in Ap. Rhod. and Callim.), and we certainly do not need a superlative here. Canter, of course, lacking contrary evidence, judged that μ⋯χαλα. was a vox nihili.

⁹³ Many neuter nouns are adjectives used substantially, with a tendency then for them to retract their accent (Chandler, , §§340, 346Google Scholar); indeed, it is plausible to derive γ⋯αλα from a hypothetical γυαλ⋯с ‘concave’ (γυαλ⋯с in fact occurs, but apparently with a different sense, in a fragment of Callimachus).

⁹⁴ If I am right in postulating a variant by misquotation (like ‘fields’ for ‘woods’ in Lycidas line 193), the error may be very ancient indeed. We can already observe a convergence of γυ- and μυχ in late Eur. and Soph. (nn. 89–90 above). As to the hypermetric verse, it remains possible, but not probable, that the lineator intended γοερον…κλαγ-|as a trimeter (cf. n. 73).

⁹⁵ Note that the dative also allows us to understand the force of οῒα (s.c. αἰ⋯γματα) as continuing to the end of the simile. But κλαγγ⋯с (int. acc.) is good enough to be an ancient error; or the lineator may have read κλαγγαс (for inferences from the lineation, see n. 73). Professor West interestingly suggests κλαγγсι; but see Barrett, on Hipp. 101.Google Scholar

⁹⁶ ‘Shout for’, cf. Ph. 1155Google Scholar βοαι⋯ πρ κα⋯δικ⋯λλαс, A. Ch. 402Google Scholar (see Garvie, ), S. Tra. 761Google Scholar; ‘shout against’, cf. the passive ⋯ναβοι in Or. 103Google Scholar (with comm.). Both, in different ways, express the idea ‘shout aloud concerning’.

⁹⁷ cf. Zuntz, , pp. 49f. (and passim).Google Scholar

⁹⁸ For the ambivalence of ‘γρ’, = γρ⋯φεται (κα⋯) or γραπτ⋯ον, see Zuntz, , pp. 42, 130ff.Google Scholar

⁹⁹ Ibid. pp. 185ff., 192.

¹⁰⁰ cf. Zuntz, , pp. 194f.Google Scholar

¹⁰¹ The line encircling κλαγγ⋯с is visibly fainter in the photograph than similar deletions elsewhere.

¹⁰² (i) φ⋯νια 〈φ⋯νια〉. Tr. was capable of doubling words metri gratia; but for the typical instance cited by Zuntz, on p. 194Google Scholar (παε 〈παε〉 at Rh. 685Google Scholar), he is likely to have found evidence in some manuscript (sic O). The false doubling here could owe something to the colometric transmission: if the uncial ancestor had divided the dimeters at ΦONI/A, there could have been some duplication in the transcription (ε) which Tr. misinterpreted, (ii) γ⋯μον for γ⋯μων improved only the metre; but collation may have shown -ων to be a recent error, (iii) παινα 〈с〉 was doubtless conjectural (cf. n. 21); but someone before Tr. could have conjectured it to go with αχαριταс (sic).

¹⁰³ cf. n. 85. It could, of course, simply be someone's guess at the noun required with φυγ⋯δα ίεсα

¹⁰⁴ For the scansion thus of δ⋯νακοс as , cf. the Byzantine misinterpretation of Or. 146 λεπτο δ⋯νακοс,ὠ φ⋯λα…(part of a dochmiac dimeter) as iambic, presumably ; Smith, O. L., Scholia metrica anonyma in Euripidis Hecubam, Orestem, Phoenissas (Copenhagen 1977), p. 11.Google Scholar