¹ ‘Euripides, Supplices 71–86 and the Chorus of “Attendants”’, CQ 40 (1990), 340–8Google Scholar. I am again indebted to Dr Diggle for his helpful criticisms of my first draft, and further suggestions.

² Diggle, J., Euripidis fabulae, ii (Oxford, 1981)Google Scholar; Collard, C., Euripides Supplices (Groningen, 1975)Google Scholar, cf. also his Teubner edition (1984).

³ In 45f. Collard obelizes only the first three words. In 51 I prefer his colon after χειρν; however the syntax is interpreted (n. 22 below), the structure of this stanza-pair favours internal punctuation lighter than Diggle's full-stop.

⁴ Modern editors have rightly rejected the mixed iono-choriambic interpretation attested by L: 2ion | 2ion _^ | 2ch || 3ion _^ | 3ion _^ | 2ion _^ | 2ch (presumably favoured as avoiding word-overlaps). But their lineations have been oddly varied: Dale (Lyric Metres ², 123) has dimeters apart from a concluding tetrameter (though she calls the first three dimeters a ‘hexameter’). Diggle and Collard both lineate so as to show a tripartite pattern of six, six and four metra. The former subdivides both the 6ion runs as tetrameter plus dimeter, despite the more natural 2|4 division, both having diaeresis thus in both stanzas; Collard follows Murray and others in dividing the first as three dimeters, the second (see further in n. 7 below) as two trimeters. The ionics here replace the anapaestic dimeters with which choruses often begin in other plays; cf. also the iambic dimeters in 71–86 (art. cit. 342–5).

⁵ On catalexis in relation to period-end, see Parker, L. P. E., CQ 26 (1976), 14–28.CrossRefGoogle Scholar

⁶ For my definition of ‘colon’ (not synonymous with ‘verse’), see my commentary on Orestes, p. xx. In a ‘dicolon’ (often but not necessarily printed as a ‘distich’) the diaeresis between the constituent lengths can either shift or disappear.

⁷ It may be difficult to distinguish between syncopated and catalectic ionic metra of the form . But there can be no doubt here about catalexis in the sixth and sixteenth metra; and the pattern favours similar interpretation of the twelfth (…-сι μλη ∼ …-τε δμοιс; the ‘ overrun’ in the antistrophe at … οὒτε δμοιс | προθμαν is like similar overruns in the second antistrophe, see below). By contrast, the seventh metron (†ἄνομοι† ∼ καταδρύμ-) is certainly syncopated within the period, and we should not wish to suggest catalexis in both the ninth and twelfth by dividing at φθιμνων | νεκύων ∼ τ γρ; ἂ | φθιμνονс (giving 3ion _^ |3ion _^).

⁸ A similarity unlikely to be fortuitous in a metrical sequence otherwise apparently modelled on the Parodos of the Persae; cf. CQ 40 (1990), 340Google Scholar, and further on 69–70 below.

⁹ ‘Ιсμηνν (not 'Ιсμ-), see Mastronarde, , ed. Phoenissae (Teubn. 1988), xxiifGoogle Scholar. For the text of 58–9 (ὂссον Tr²) and 60–2 (where Collard prefers λιссμεθ’) see Diggle, , Studies on the Text of Euripides (1981), 2–3Google Scholar. As to punctuation, I have removed the usual but superfluous comma preceding in 60; for or ὦ following an imperative, cf. CQ ibid., 345 with nn. 34–5.

¹⁰ Both Diggle and Collard, after Wilamowitz, Dale and others, treat the stanza as 7ion _^ ∥ 7ion _^ ∥ 10ion _^. This is only superficially attractive, since (a) the 7ion _^ lengths, apart from their overall equality, are in other respects amorphous (divided by Diggle as 4 + 3|4 + 3, and by Collard, again following Murray, as 3+ 2 + 2|2 + 3 + 2); (b) the division after 7ion _^ is presumably intended to give symmetrical periods with catalectic close, but there can be no period-end at μετ νυν | δс (an indivisible unit) in the strophe. [Diggle further contrives to get the metra at verse-end in all three places, by dividing the 10ion _^ sequences as 4 + 3 + 3, rather than 3 + 3 + 4; thus further obscuring the difference between catalexis and syncopation (cf. n. 7 above).]

¹¹ L's divisions match the phrase-lengths in the strophe, and give another eight-verse stanza. But there is still an overlap at δνс |τνχαν in the antistrophe; and the dimeter-arrangement with indentations arguably shows more clearly the structure of both stanzas and their relationship with 42–7/48—53. Since the 3ion lengths occur in pairs, it is likely that Euripides was himself conscious of the ambivalence (3×2 = 2×3). Cf. the 3an phrase-lengths which frequently occur in anapaestic systems, where tine dimeter is the fundamental, but not sacrosanct, unit; West, M. L., BICS 24 (1977), 89–94.Google Scholar

¹² For such initial syncopation, Dr Diggle reminds me of Ba. 64 'Αсαс π γс (Hermann γααс) and 69.

¹³ i.e. rhetorical ‘pause’, as defined and analysed in relation to period-end by Stinton, in CQ 27 (1977), 27–66CrossRefGoogle Scholar; or we may prefer simply to speak of ‘enjambment’.

¹⁴ See n. 8 above.

¹⁵ Against Murray, who unconvincingly defended the text as ‘clamores confusos’. We agree also in rejecting all proposals that give a text in which the dead corpses are leaving their own limbs unburied.

¹⁶ The only other Euripidean instances claimed by Diggle of prodelision after -αι (both conjectural, though probable) are of ()γώ following -ομαι Hel. 953, I.A. 1396 and (he now adds) I.A. 900 (Markland). [The position is different if the ι is adscript, as at 69 ταλαναι 'ν χερ for which see Diggle, Studies, 33]

¹⁷ The seriously disturbed text of Hel. 1152–4 (a passage which I hope to discuss elsewhere) may well involve some corruption in the relevant words (πνουс … καταπνμενοι); and in any case νμονс and πνονс are different kinds of object.

¹⁸ Professor West has drawn my attention to the corruption of ἄνομον to ἄνεμον at Hes. Th. 307, where allegorical interpretation was a factor.

¹⁹ The imperative ντω occurs at S. Ichn. 70. The participle occurs in Ar. Vesp. 369 (ἄνοντοс), and possibly in S. O.C. 523 (ἄνων Hermann, for ἄκων; a conjecture meriting more attention than it has received); cf. also A. Ch. 798 (νμενον Emperius, -νων M). The optative ἄνοιс is plausibly restored in A. fr. 161.2 (Dobree; Stob. ναοῖс, λβοιс Σ Hom. I 158 et Eust.). For Euripides we have at present only Andr. 1132 λλ' οὐδν ἦνεν (ἦνον Borthwick), but that suffices to show that ἄνω, equivalent to νύω, was not a verb alien to his vocabulary.

²⁰ Especially after verbs like κελεύω; often also after сτε, as in 204 etc. (PCPS 36 (1990), 199 n. 37).Google Scholar

²¹ Kühner-Gerth ii.402; for omission in particular of the accusative antecedent of a clause with nominative relative pronoun, cf. Alc. 338 сτνγ μν ἥ μ' ἕτικτεν. Relative pronoun plus finite verb is commonly equivalent to article plus participle, and vice versa. So here οἳ καταλεπονсι can (in theory at least) be understood as τοὺс καταλεπονταс.

²² The antecedent of ἅ may be either the first person variously implicit in the context as a hole (including the extended object of сιδοсα), or specifically the ‘ I ’ implicit in the elliptical τ γρ; (sc. πθω).

²³ The suggestion offered is ex. gr. (some may prefer to look for a suitable noun denoting the Thebans). For сχτλιοс ‘hard-hearted, merciless’ (LSJ s.v. 1.2) followed by a relative clause, cf. Od. 5.118ff., etc.; it is a word that naturally begins a colon, as at Alc. 470 сχετλω πολιν ἕχοντε χαταν (followed by an ionic sequence).

²⁴ Cf. Kannicht on Hel. 176–8.