¹ I am indebted throughout to Professor Hugh Lloyd-Jones, and on individual passages to ProfessorCollard, C., DrDiggle, J. and MrReeve, M. D., for valuable advice and criticism. Part I of this article appeared in JHS xcvi (1976) 121–45Google Scholar.

² Cf. Dale, , LMGD 115Google Scholar.

³ See Griffith, M., The authenticity of ‘Prometheus Bound’ (1977)Google Scholar; cf. Parker, L. E., ‘Catalexis’, CQ 26 (1976) 20CrossRefGoogle Scholar.

⁴ So Wilamowitz, , GV 508–9Google Scholar; Dale, , BICS Suppl. 21.1 (1971) 18–19Google Scholar.

⁵ Cf. Barrett, on E. Hipp. 840Google Scholar. It occurs in mss. at E. Andr. 834 = 838, but is easily emended there (see ad loc. below, p. 143); in A. P.V. 576/95 there is certainly some corruption.

⁶ In E. Hec. 912 = 921, corresponds with but divide with normal licence. In any case the responsion is impossible, as is while is unexampled as a form of dochmiac. (Conomis, N. C., Hermes 92 [1964] 35Google Scholar, cites E. Or. 1247 = 1267, but this is best taken as iambic tripody; see di Benedetto ad loc.)

⁷ Dain's, supplement in O.T. 1217Google Scholar is clearly wrong, but a supplement giving a long to suit his analysis could readily be found.

⁸ Lobeck's (with ) is on the wrong lines: it is dishonour, not vengeance, that Ajax dreads.

⁹ So Dale, l.c. (n. 4).

¹⁰ Wilamowitz also omits after and divides which he takes with But ‘I am involved with imbecile plunderings together with these,’ sc. slaughtered beasts, gives an odd sense to We then need cf. 427 and the Homeric This might even be right, but the further changes are the last straws to break the back of the camel's credulity, if it is not broken already.

¹¹ Fraenkel, was careful to point out (‘Kolon u. Satz, II’, NGG [1933] 319–20Google Scholar = Kl. Beitr. i 93–4) that the rules governing the position of in classical prose do not necessarily apply to other genres. In practice they generally seem to hold for verse, and the few exceptions should probably be emended (e.g. H.F. 235 ib. 665–6 transpose [Hermann]), or explained as a means of emphasis (see on Phil. 702, p. 135 below).

¹² Dale, reads πατέρων and scans as anapaests (LMGD 138)Google Scholar.

¹³ Page has pointed out that even is in fact extremely rare, though it seems a straightforward variation of (see n. 20, on S. Phil. 683–6 below).

¹⁴ In Phil. 828 εὐαὲς ἔμīν ἔλθοις, text and colometry are quite uncertain.

¹⁵ See Dale, , ‘Lyrical Clausulae in Sophocles’, in Greek Poetry and Life (essays presented to Gilbert Murray) (1935), 195Google Scholar = Collected Papers (1969) 13.

¹⁶ Probably also S. Phil. 832–3 = 848–9, 835–6 = 851–2. 838 854 can hardly be other than (on the correption, see BICS 22 [1975] 106 n. 22); is certainly dochmiac, and so I suspect is 830 = 846 (possibly followed by mol. =8). Cf. Dale, , LMGD ²117Google Scholar f. See also Kaibel, , Electra 147Google Scholar; Parker, L. E., CQ 18 (1968) 258CrossRefGoogle Scholar f.

^16a Cf E. Cycl. 501 (Kells) is too strong for the context. In any case ‘importune’ is not the same as and would not be so used in the first person (contrast A. P.V. 1002–3).

¹⁷ For the slight zeugma (with ἐισηεύδειν sc. δεī, not ἔχεί λόγον) cf. 649, O.T 241, 818, O.C. 1402–4.

¹⁸ On the rarity of the lengthening before mute and liquid see Barrett, , Hippolytus pp. 310Google Scholar, 435.

¹⁹ Other examples are given by Dale, , who holds that all such irregular clausulae are explained by the metrical context (‘Lyrical clausulae in Sophocles’, in Greek Poetry and Life (Essays presented to Gilbert Murray) [1935] 200Google Scholar f. = Collected Papers [1969] 19 f.). Cf. ‘More rare verse-forms’, BICS 22 (1975) 101–3.

²⁰ According to Dale, (o.c. 199 = 18)Google Scholar, this responsion ‘can be dismissed at once; Sophocles could never have set a regular to match a “limping” iambic at the close of a stanza’. Her assertion is ill-founded. The licence is not frequent enough in Sophocles for any inference to be drawn from its absence at stanza-end. It is not frequent in Euripides, either, and Hipp. 741/751Google Scholar with the exact parallel, noted above (p. 129), offered by mss. at Med. 159/183 is perfectly good evidence for Sophocles' practice, despite her denial (ib. n. 2).

²¹ At Phil. 827 = 845, however, the metre is unclear.

²² See Fraenkel, on Ag. 7Google Scholar, 681 ff.; Dornseiff, F.Pindars Stil (1921) 107Google Scholar ff.

²³ Dindorf, followed by Diggle, J., CR 16 (1966) 262Google Scholar.

²⁴ Nor with Diggle's ἔλασεν, cf. Austin, C. and Reeve, M. D., Maia 22 (1970) 2–3Google Scholar.

²⁵ Or -αν … ἀν- if Musgrave's ἄντυγα is right, which it may well be. ‘Rim’ is nearer to wheel than ‘frontlet’, and the lexicographers' explanations, e.g. Hesych. ἀμπυκες…ἢ τροχοί οὕτως Σ. ἐν φιλοκτήτῃ, διά τὸ κυκλοτερές, could derive from this passage. However, ἅμπυξ does not really mean ‘wheel’ either, so it is better to leave the text. I am not convinced by Robinson, D. B.'s explanation of ἅμπυξ in C.Q. 19 (1969) 42–3CrossRefGoogle Scholar, that since it means (1) (gold) diadem, (2) horse's frontlet, (3) bridle, it suggests that Ixion's wheel is round, fiery and a curb on his passions.

²⁶ For the position of the predicative between and cf. O.C. 716 d where Jebb rightly takes with (Cf. also CRBS 17 [1976] 327 ff.)

²⁷ Marginalia Scaenica (1955) 110 ff.

²⁸ Short before initial rho is very rare in tragedy (see Dawe, R. D., Studies on the text of Sophocles [1973] i 299Google Scholar ff., and add E. Suppl. 380 πάντα ῥύῃ) but it is occasionally found before ῥέζειν; cf. also Solon fr. 23.20 D. = 34.8 W., where [ῥέζ]ειν seems the best supplement (so Diehl).

²⁹ PCPS n.s. 6 (1960) 52, cf. Sappho and Akaeus, 81.

³⁰ Long, A. A., Language and thought in Sophocles (1968) 134Google Scholar n. 73, cl. Miller, H. W., ‘Medical Terms in Greek Tragedy’, TAPA 75 (1944) 165Google Scholar.

³¹ πόθος cannot here refer to more general desires such as hunger; that is reserved for the next stanza.

³² ‘Uber ein Gesetz der indogermanischen Sprache’, IF i (1892) 333–446= Kl. Sehr, i 1–104; cf. Dover, K. J., Greek Word Order, p. 14Google Scholar.

³³ Thuc. i 45. 3 and Hdt. i 85. 2 cited by Classen ad. loc. are slightly different, but they may be formed on the same model; cf. Hdt. viii 90. 1 also cited by Classen, where the order is regular because is second word. Wackernaget observes that the regular hyperbaton of enclitics is more frequent in the Ionic dialect, though not confined to it.

³⁴ Wackernagel, , o.c. 393–402Google Scholar; cf. Fraenkel, E., ‘Kolon u. Satz II’, NGG Phil-Hist. (1933) 319Google Scholar = Kl. Beitr. i 93–4. See also on Aj. 408 (p. 128 n. 11), and below.

³⁵ So Jebb, rightly, since is not equivalent to

³⁶ There is of course lack of responsion at the end of the line. This is neatly cured by Campbell's does not occur elsewhere in tragedy, but then neither does except in O.C. 64. But is also a variant here in G and Q. (Easterling, P. E., CQ 19 [1969] 76CrossRefGoogle Scholar). This can hardly be a metrical conjecture (though cf. Trach, i 118 A, and see PCPS n.s. 13 (1967) 51); it may simply be a gloss on but it could be an old reading.

^36a Cf. also Ar. Ach. 640, Eg. 405 (verb first); Ach. 215, Eg. 707, 855; S. O.C. 1174.

³⁷ Eustathius', οὒτε τι ῥεξας might just be due to a recollection of Od. iv 690Google Scholar οὒτε τινα ῥέξας ἐξαίσιον οὒτε τι εἰπών | ἐν δήμῳ. But as Jackson, says Eustathius certainly did have access to an independent tradition, cf. ἀμφίσταμαι at El. 192Google Scholar.

³⁸ Phil. 194 Philostr. jun. (Imag. 17, 859 K): Philoctetes knew where the altar of Chryse was because he had been there before with Heracles; Euripides, Philoctetes ap. Dio Chrys. Or. 59, 9: Philoctetes showed the Greeks where the altar was, cf. S. Phil. hyp. 4–5. The coincidence of the later sources with Euripides is enough to show that this version is not a late invention. Moreover, a series of vase-paintings (the earliest c. 430) show Heracles with Philoctetes (named) and/or Iolaus or Lichas at an altar, most probably the altar of Chryse (see Hooker, E. M., JHS 70 [1950] 35–42CrossRefGoogle Scholar). Mrs Hooker plausibly suggests that these were inspired by Euripides' play (431), in which the incident may well have been mentioned, rather than a large painting of c; 440 as Schefold thought. This previous association of Philoctetes and Heracles may belong to an early form of the legend, as Wilamowitz, and Robert, C. held (Herakles II 80Google Scholar; Gr. Heldensage 599 n. 3); in any case it is earlier than Sophocles' Philoctetes. The mere fact of Philoctetes receiving the bow from Heracles would of course suffice to make him ‘Heracles' man’, and thus anticipate Heracles' entry in the play; but the phrase has more point if it implies the earlier association of the two. (Cf. Sen. Here. Oet. 1717 Alcidae comes, with ib. 1603–6 umerisque tela /gestat et notas populis pharetras / Herculis heres.)

³⁹ Cf. LMGD 73, 100.

⁴⁰ will then mark another allusion in tragedy to the pre-Zenodotean version of Il. 1–5 shown by Pfeiffer to have been current in the fifth century, cl. A. Suppl. 800, S. Ant. 29–30, E. Ion 504, Hec 1078 (History of Classical Scholarship (1968) 112–3). I owe this point to Professor Lloyd-Jones.

⁴¹ BICS 22 (1975) 88–95.

⁴² A more sceptical view of ‘hexamakra’ is now taken by Diggle, in PCPS n.s. 20 (1974) 22–4Google Scholar.

⁴³ Possibly (enopl.) but this is also rare in Euripides.

⁴⁴ For examples of repeated μά see Pfeiffer, on Callim. fr. 194Google Scholar. 105–6, to which add Men. Dysc. 666–7 (these references, and that in n. 46 below, I owe to Dr Diggle).

⁴⁵ I am indebted to Mr R. A. S. Seaford for this point.

⁴⁶ So Franke, F., Commentationum de Cyclope Euripidis criticarum et grammaticarum spec, i (1829) 32Google Scholar (Θἰερά, vel potius θαἰερά), with reasons and the parallels for form and metre. The conjecture seemed worth reviving in this note, since it is clearly better than others since.

⁴⁷ As Diggle, suggests, Maia 24 (1972) 345Google Scholar.

⁴⁸ See Zuntz, , Inquiry 38Google Scholar, al. That the alteration is early is shown by the agreement of P. (Mr R. A. S. Seaford tells me that the correction is l ² or l ³, but that this may be simply a clarification of an obscure compendium in L.)

⁴⁹ The first objection (which I owe to Dr Diggle), is decisive, and the second cannot easily be met: will not do, since is not found either, though said to be Attic by Choeroboscus, (in Theod. 1.248 H)Google Scholar; with in apposition, is unlikely to be right since in a sexual context is intransitive only at P.N. x 81.

⁵⁰ I owe this point to Mr Seaford.

⁵¹ o.c. (n. 47). He compares (after Rossì, L. E.) Alcaeus fr. 346.1Google Scholar (L.-P.) The passage is certainly relevant, but does not show that is object rather than subject in Cycl. 514. Nightfall as a time for love is an obvious topos.

⁵² πάλαι is due to Dindorf, καί to Hartung, the combination to Seaford.

⁵³ This is only possible when the word to supply can be inferred either from what the speaker has already said (see Diggle, 's own remarks on aposiopesis in PCPS n.s. 15 [1969] 57Google Scholar) or from the context; e.g. Men. Epitr. 442–3 is explained by 435–6

⁵⁴ The point is made by an apparent exception at S. O.C. 209–11:

The bare negative is enough to tell the chorus what kind of utterance is to follow. Slightly different is S. El. 854–7:

Here the simple question needs no such indication.

⁵⁵ This might be just a joke, like singing ‘for I'm a jolly good fellow’. But though it is doubtless the victor's friends who would strike up the καλλίνικος (cf. Σ P.O. ix 1), there is no reason to think the victor would feel inhibited from joining in.

⁵⁶ In P.P. x 21f. is obviously quite different.

⁵⁷ Austin, C. and Reeve, M. D., Maia, 22 (1970) 11–12Google Scholar.

⁵⁸ Austin and Reeve, with some justice, question whether the conditional can bear this nonrestrictive meaning. My doubts are not wholly allayed by Diggle, 's paraphrase ‘if the flautist strikes up a tune, I like to dance’, as I do not know an exact parallel. Bacch. 135Google Scholar is presumably a case in point, if it refers to the god. Possibly si quidem, as e.g. in P.O. ix 25–7 where the clause is strictly causal. But perhaps should be read.

⁵⁹ As e.g. at P.O. i 1 ff., B. iii 85 ff. cf. Dornseiff, F., Pindars Stil 1921, 96Google Scholar ff.). Diggle's examples and his comments on them give the impression that the type (or is the only form of priamel, whereas its variety is manifold, as Bundy, Elroy L., in particular, has shown (Studia Pindarica i, Univ. of California Publ., 1962)Google Scholar. Dr Diggle tells me that he does not intend to give any such impression.

⁶⁰ See PCPS n.s. 22 (1976) 74.

⁶¹ Cf. O. T. 896 though this too can also be interpreted in persona (see Bain, D. M., ‘Audience address in Greek tragedy’, CQ 25 [1975] 16CrossRefGoogle Scholar n.).

⁶² This is not to say that the priamel here hinges on a contrast between the speaker and others, as so often, e.g. P.O. i 111–3

⁶³ See Dittenberger's note.

⁶⁴ This suggestion was endorsed by Lloyd-Jones in his review of Barrett, 's edition (JHS 86 [1966] 164 f.)Google Scholar.

⁶⁵ It makes no difference whether we write (whence M. Gr. ) with Fraenkel and others, or (sc. ) with Page; cf. Wackernagel, , Syntax ii 166Google Scholar, cited by Page ad loc.

⁶⁶ Conomis, N. (o.c. [n. 6] 45)Google Scholar maintains that brevis in longo without pause is extremely rare in dochmiacs. I noted on S. Trach. 1008 ff. (JHS 96 [1976] 144 n. 87) that this appears to be true of Sophocles, but not of Aeschylus or Euripides (see CQ.27 [1977] 46 f.).

⁶⁷ See Vian, Francis, La Guerre des Géants (1952) 251Google Scholar; also pp. 63–7, 200 f. The first part of this note is largely derived from Vian's book.

⁶⁸ Gigantomachies on vases begin, and are most common, in about the decade before the middle of the sixth century. This fits well enough with the traditional date for the foundation (or ‘revival’) of the Panathenaea by Pisistratus (566); see Beazley, , The development of Attic black-figure (1951)Google Scholar ch. viii; Davison, J. A., JHS 78 (1958) 27Google Scholar.

⁶⁹ For this interpretation see Vian, , o.c. 200Google Scholar.

⁷⁰ Aristophanes, as Vian observes (184 f.), draws indifferently on both legends for his parody in the Birds. For references to Hellenistic and Roman writers see Vian, p. 173.

⁷¹ Denniston, indeed, after a list of examples in drama (GP 49)Google Scholar, concludes: ‘There are, then, not a few cases in the dramatists where interrogative is placed late’. But in such a case as Eum. 745 the vocative forms a separate colon, beginning a new one. If we exclude such cases, never comes later than fourth word in tragedy (here, given an incision at it is sixth, otherwise twelfth). There is one case in comedy: Ar. Ec. 462 Denniston compares ib. 462 The postponements in Plato, which Denniston says are much freer than in other prose writers, are often to be explained in the same way by a separate initial colon, e.g. Phlb. 27B Fraenkel's demonstration that in prose always obeys Wackernagel's rule (see above, p. 128, n. 11).

⁷² In this note and the following notes on the Supplices I had the early benefit of Professor Collard's commentary, which he kindly allowed me to see in typescript.

⁷³ Equally plausible is Kirchhoff's, οὔτ' <οὖν> ἐν ϕθιμένοις as DrDiggle, points out to me, comparing Andr. 329 ἐν ϕθιμένοις' href=https://scholar.google.com/scholar?q=Kirchhoff's,+οὔτ'+<οὖν>+ἐν+ϕθιμένοις+as+DrDiggle,+points+out+to+me,+comparing+Andr.+329>Google Scholar, 731, I.A. 1437, and L's οὔτ' for οὔτ' in Hec. 1244. He also suggests οὐ ζώντων ἀριφμουμένα cl. Ba. 1317, an interesting possibility.

⁷⁴ if correct, would mean: ‘(goodbye to) any legitimate marriage in Argos appearing for my children‘. Evadne is then mourning her own children's blighted prospect of marriage, as tragic heroines do elsewhere (Alc. 165 ff., 318 fr.; H.F. 476 ff.; Hel. 282 f., 933. I am assuming some such supplement as does not mean ‘goodbye to’, see Diggle, , PCPS n.s. 20 [1974] 8Google Scholar n.). however is very odd, and I believe corrupt. Diggle would restore the passage so that Evadne is still referring to her own children. But Evadne's children have no place in this incident: the whole monody concerns her love for her husband and the happiness in marriage they have lost. My tentative ex. gratia restoration would be: meaning any marriage of which the children are i.e. with both parents surviving, unlike Capaneus and herself.

⁷⁵ Presumably the ‘attractive smell’ is like that of Barine in Hor. C. 2.8. 23 f. tua ne retardet aura mantos, though this is hardly appropriate in our passage. (I owe this example to Dr Diggle.)

⁷⁶ Professor Collard's note as published is slightly different, though not in substance.

⁷⁷ C. G. Haupt: M: Kirchhoff.

⁷⁸ It is impossible to translate adequately, since no equivalent English word has the requisite verbal force. ‘Plume’ would suggest what smoke looks like rather than what it does. A more abstract use seems to be required at Callim, . Lav. Pall. 124Google Scholar of omens, though this may be technical. McKay, K. J., The Poet at Play (1962) 48Google Scholar n. 2 compares penna = ‘omen’ at e.g. Prop. 3.10, 11, with Butler's note.

⁷⁹ For this type of lyric trimeter without caesura cf. Or. 966, 989.

⁸⁰ As Diggle, J. points out in his note on the passage (PCPS, n.s. 15 [1969] 57–9)Google Scholar, the articular form of the relative pronoun is very rare when notrequired by metre (see also Reeve, M. D., GRBS 11 [1970] 285Google Scholar f.)

⁸¹ is analogous and Od. xvi 472 is a sound example of the sense required, but the word is common enough for the argument ex silentio to be valid. The dominance of the special usage can be seen from A. Cho. 3 and from Aeschylus' defence of the line against Euripides' charge of tautology (Ar. Ran. 1154–65).

⁸² Cf. Page, D. L., Actors' Interpolations in Greek Tragedy (1934) 76Google Scholar.

⁸³ Cf. on Cycl. 672–5 above, and see Diggle, , o.c. (n. 80) 57Google Scholar.

⁸⁴ As Diggle observes, though he seems not to recognise that Murray construes in the same way as himself (cf. Austin, C. and Reeve, M. D. in Maia 22 [1970] 15Google Scholar).

⁸⁵ is used sometimes of the sprinkling, cf. Od. iii 445 sometimes of the stage after the sprinkling, cf. Ar. Av. 559 where the genitive denotes the offering, as in I.T. 56, 1154, cf. D.xxi 114 For the technical sense of cf. S. Trach. 764, and see Wilamowitz, , Sappho und Simonides (1913) 152 n. 3Google Scholar.

⁸⁶ also occurs in a fourth-century inscription (IG vii 235, 25) with the genitive used of the offerings like But the use is not attested in literature, and the more general sense of is needed here (just as in Ag. 1450 the general sense of is needed to refer back to the chorus' prayer, while the technical sense gives the line its point).

⁸⁷ So Diggle, who puts brackets round the clause The hyperbaton is in itself perfectly possible; but with Murray's pointing the parenthesis breaks up the run of the sentence, while with Diggle's the relative clauses impede it.

⁸⁸ So Diggle, (o.c. 58)Google Scholar, though he would now prefer, with Elmsley, to take the lines as an explanation of νόμοισί(ν).

⁸⁹ As Diggle, , after Wecklein, , points out (o.c. 57)Google Scholar, though he retains the line in his final version.

⁹⁰ As Hermann remarked, cl. Hdt. iv 103. For the relation between Artemis Ταυρική and Τανροπόλος, see Farnell, L. R., The Cults of the Greek States (1896) ii 240Google Scholar ff., 251–5.

⁹¹ read by the Budé editors, avoids this snag, but alien close to dactyls is normally iambic, so is an unlikely sequel.

⁹² For some philosophical formulations see Kirk, G. S., Heraclitus, the Cosmic Fragments (1954) 310Google Scholar.

⁹³ Hermann, who makes this point, read which he preferred to Seidler's but see GP 517.

⁹⁴ As Dr Diggle points out to me. He prefers Burges' ὂ'σα, with the explanation given above.

⁹⁵ In Virgil the point is rather more sophisticated; see Hubbard, M., PCPS n.s. 21 (1975) 53Google Scholar ff., esp. 61.

⁹⁶ Cf. Lucas, D. W., Aristotle's Poetics (1968) 100Google Scholar f. It was from the first two occupants of the Delphic oracle, combined in one (Gaia-Themis), that Prometheus learned to foretell the future and to validate his prophecies by his knowledge of the past (cf. A. Eum. 1–4, P.V. 209–11, 824–5).

⁹⁷ Better perhaps ἐς ἐμέ, an improvement suggested by Dr Diggle. The corruption would be just as easy.