¹ There is a thorough account of Erinyes etc. byxWiist, E. in RE Suppl. viii. 82–166Google Scholar, though he tends to assume that the various names are interchangeable; so does Gruppe, O., Griechische Mythologie und Religionsgeschichte II (Munich, 1906), 763–8Google Scholar. This fault, at least, is avoided byHarrison, J. E., Prolegomena to the Study of Greek Religion ³ (Cambridge, 1922), 213–56Google Scholar, and Farnell, L. R., The Cults of the Greek States v (Oxford, 1909), 437–13Google Scholar. Dietrich, B. C., Death, Fate and the Gods (London, 1965), 91–156Google Scholar, is of some use as a source of references.

² I am using the word ‘cult’ loosely. In some places, including Colonus, we know only that there was a sacred site at which offerings could be paid, and not that they were paid regularly.

³ Nor by the schol. on Od. 11. 271 = Androtion, FGrHist 324 F 62, but this source may not be of any value; see Jacoby, F., FGrHist IIIB (Suppl.), i. 169 f., ii. 154 fGoogle Scholar.

⁴ As does Farnell, L. R., Greek Hero Cults and Ideas of Immortality (Oxford, 1921), 333Google Scholar.

⁵ E.g. by Dietrich, , op. cit. (n. 1), 98Google Scholarf.

⁶ It is doubtful whether there was, or was believed to be, a cave near the grove at Colonus; seeRobert, C., Oidipus i (Berlin, 1915), 23–31Google Scholar.Julius Obsequens 56 mentions a ‘lucus Furiarum’ (apparently in Asia Minor or the Aegean, if not fictitious), but we cannot tell the name of these ‘Furies]; perhaps they too were Eumenides. Plut. C. Gracch. 17. 2 mentions an ἂλσος 'Eριν⋯ων at Rome, but appears to mean the grove of Furina (cf. Cic. ND 3. 46).

⁷ IG iv. 571, 574–5, 668; SEG 11. 368;Milchhoefer, A., Mittheilungen des deutschen archdologischen Institutes in Athen 4 (1879), 474–6Google Scholar; Harrison, , op. cit. (n. 1), 255 f.Google Scholar ; Vollgraff, W., BCH 68–9 (1944–5), 395 f.Google Scholar; Nilsson, , Geschichte dergr. Rel. i.³101Google Scholar. Three of the tablets have commonly been seen as evidence for a cult further west, near Argos, but Vollgraff argues reasonably against this.

⁸ Cf. Orph. fr. 197 K, where the Eumenides are ⋯νθεσιουρο⋯ and daughters of Pluto and Persephone; Euphorion fr. 94 P (quoted at n. 117 below). These sources must be influenced by cult.

⁹ SEG 9. 325–46; ibid. 20. 723; Forbes, K., Philol. 100 (1956), 235–52CrossRefGoogle Scholar.

¹⁰ SEG 17. 441 a; Forbes, , art. cit. (n. 9), 242–5Google Scholar.

¹¹ Note also P. Derv. col. ii (see n. 45 below): worship of Eumenides, with wineless libations and sacrificial cakes, is common among μ⋯σται.

¹² The word εὐμεν⋯ς, which it presupposes, occurs first at Anacr. 357. 6 P, Sim. 519 fr. 35 (b) 4 P. Horn. Hymn 22. 7 need not be earlier; but εὐμεν⋯της at Od. 6. 185 may indicate that the absence of εὐμεν⋯ς from early poetry is deceptive.

¹³ In general the doctrine of ‘antiphrasis’ looks rather artificial. The four examples constantly cited – Eὐζειν⋯δες, Eὔζεινος, εὐøρ⋯νο, εὐώνυμος – belong to four different semantic fields, and may well have diverse origins.

¹⁴ See Wüst, , art. cit. (n. 1), 86–91Google Scholar.

¹⁵ For the Potniai see nn. 53, 128, below. The Maniai of Arcadia (Paus. 8. 34. 1–3), who pursued Orestes until he placated them by biting off his finger, have clear affinities with the Erinyes and with the Eumenides of Cerynea (see p. 271 below). At one point, indeed, Pausanias calls them Eumenides, but this probably reflects his memories of Cerynea or the fact that Orestes' pursuers were commonly so called in his day, and does not mean that Eumenides was a local cult title. Certainly the name Erinyes, which he also uses here, is due merely to ‘writers on Peloponnesian antiquities'.

¹⁶ Modern scholars tend to call them simply Semnae. The ellipse of the noun does occur in Greek (e.g. Diog. Laert. 1. 112), but is far from common; see Harrison, , op. cit. (n. 1), 239 fGoogle Scholar.

¹⁷ See also scholia on Aeschin. 1. 188, Soph. O.C. 489.

¹⁸ Dinarch. 1. 47; cf. Dem. 23. 67 f.

¹⁹ Aeschin. 1. 188; line 9 of IG ii.² 112 = SIG 181 = Tod, , GHI ii. 144Google Scholar(362/1 B.C.); hence the restoration in line 7 f. of IG ii.² 114 = SIG 180 = Tod, , GHI ii. 146Google Scholar.

²⁰ Diggle's text of El. 1272, where I take ϰρηστ⋯ριον in apposition to the sentence rather than to π⋯γον. The word is problematic, since oracles at this site are otherwise unknown, and seem unlikely. If the text is sound, Eur. is perhaps influenced by legends of other figures, such as Amphiaraus, who plunged beneath the earth to become an oracular hero.

²¹ Also schol. on Thuc. 1. 126 (τ⋯ν σεμν⋯ν θε⋯ν): τ⋯ν' Eριν⋯ων, κατ⋯ ⋯ντ⋯øρασιν. ἂς μετ⋯τ⋯ν 'Oρ⋯στην ο⋯ 'Aθηναῖοι πλησ⋯ον το⋯ 'Aρε⋯ου π⋯γου ⋯δρ⋯σαντο, ἴνα πολλ⋯ς τιμ⋯ς τϰωσιν. Schol. (‘Ulpian’) on Dem. 21. 115: κα⋯ οἱ μ⋯ν øασιν ὃτι δι⋯ τ⋯ν 'Oρ⋯στην ⋯κεῖ καθιδρ⋯θη…, sc. the shrine of the Eumenides (sic) beside the Areopagus.

²² A different account is given by Diog. Laert. 1. 112: the cult was founded by Epimenides the Cretan.

²³ Some early examples: Ar. Knights 1312, Thesm. 224; Thuc. 1. 126; the inscriptions cited in n. 19 above; Dem. 21. 115; Aeschin. 1. 188; Dinarch. 1. 47, 64, 87, fr. VIII. 2 Conomis. Note also Lucian, Bis Ace. 4.

²⁴ Schol. on Soph. O.C. 42 = Philemon fr. 217 K.

²⁵ Apart from various scholia and entries in lexica note Cic. ND 3. 46; Plut. Thes. 27.

²⁶ The presence of a tomb of Oedipus (Val. Max. 5. 3. 3 as well as Paus. loc. cit.) clearly provides a link with the Eumenides of Colonus, but this evidence is hard to evaluate. Scholars are divided on whether the tomb is earlier than the legend of Oedipus' death at Colonus, and whether either is earlier than Soph. O.C: see e.g. Jacoby, locc. cit. (n. 3); Rosenmeyer, T. G., Phoenix 6 (1952), 99 n. 30CrossRefGoogle Scholar.

²⁷ The Oracles of Zeus (Oxford, 1967), 131–4Google Scholar. The argument is very speculative, and, if it were correct, I should expect to find some mention of this oracle, or of the associated legend, in other Sources.

²⁸ I am concerned only with cults explicitly attested. Others have been postulated on the basis of myths linking Orestes, Oedipus or the horse Arion with particular sites, but, even if this is a valid procedure, we cannot tell what name the Furies were given at these sites, nor at what period they were worshipped.

²⁹ For the evidence for the Boeotian cult see Schachter, A., Cults of Boiotia l (BICS Suppl. 38. 1, 1981), 164Google Scholar. But Schachter's tentative suggestion that it is ‘only a literary invention based on a confusion between the Boiotian Telphousa and the Arcadian Thelpousa’ is doubtless correct. The names of Tilphusa and Thelpusa are both found in various forms, and either, it seems, can become T⋯λøουσα (Pfeiffer on Call. fr. 652). The schol. on Lycophron 1225 certainly does not distinguish between them, and the only source to mention Boeotia explicitly is schol. MV on Il. 23. 346, which is clearly concerned with literature, not cult. Demeter Erinys, then, will be a purely Arcadian deity, like many who seem eccentric from an Athenian viewpoint.

³⁰ If we knew anything about the Erinys detected on two Linear B tablets (KN 200, 208), she would doubtless prove equally unhelpful.

³¹ The Theran cult is sometimes said to receive confirmation from IG xii. 3. 367, but the reading looks very uncertain.

³² Some believe in a cult on Mount Cithaeron, but this rests purely on ps.-Plut. de Fluviis 2. 3, for the authority of which see Page, D. L., Corinna (London, 1953), 22Google Scholar.

³³ The famous saying of Heraclitus (fr. 94) that, if the sun transgresses his limits, the Erinyes will find him out, is probably not evidence for any literal (still less popular) belief. P. Derv. frr. A and B (see n. 45 below) seem to deal in esoteric doctrines derived from Heraclitus and the poets.

³⁴ Cf. Timaeus, , FGrHist 566 F 55Google Scholar; Plut. Dion 55. 1; Prop. 2. 20. 29; Harrison, , op. cit. (n. 1), 231Google ScholarMoulinier, L., Le pur el Vimpur dans la pensee des Grecs (Paris, 1952), 259–70Google ScholarParker, R., Miasma (Oxford, 1983), 14, 104–9Google Scholar. Perhaps superstition declined from Antiphon's time to that of Aeschines; but the world of the Tetralogies is fairly artificial, and even there Antiphon avoids the word Erinyes (probably too literary or melodramatic) and does not suggest that the avenging spirits have a very solid existence.

³⁵ I could have missed instances, but the oniy one known to me is IG iii. 3 Suppl. 108 (3rd or 2nd century B.C.). And this is no true exception, since the inscription is in verse, with literary pretensions (the same is true of Alcaeus' prayer, but the early date of this makes mere literary convention less likely).

³⁶ Audollent, A., Defixionum tabellae (Paris, 1904), 462Google Scholar.

³⁷ I have argued at JHS 103 (1983), 26, that some aspects, at least, of Aeschylus' presentation of the Erinyes reflect dramatic needs rather than popular belief.

³⁸ Thus, of the scholars cited in n. 1, Harrison and Farnell are right to stress the differences between the various species, while Gruppe, Wiist and Dietrich are wrong to play them down.

³⁹ I have counted only instances in the plural because the singular of Eὐμεν⋯δες seems not to occur in Greek. Examples in Latin, such as Lucan 1. 576, will be mere solecisms.

⁴⁰ It does not occur in Bacch. and I.A., the two plays later than Or.

⁴¹ Dale, A. M., Euripides: Helen (Oxford, 1967), xxiv–xxviiiGoogle Scholar.

⁴² I.T. 79, 294, 299, 931, 941, 963, 970, 1439, 1456. Eur. seems particularly in need of a synonym for 'Eρινᾷες at 944, where, having just used the word at 941 and feeling that variatio is required, he incongruously resorts to calling them ταῖς ⋯νων⋯μοις θεαῖς.

⁴³ Some other examples: Hel. 1301–68 (see Dale and Kannicht ad loc), Phoen. 109 f., 685 f., Bacch. 275 f., Soph. Phil. 391–402.

⁴⁴ Cf. e.g. Virgil, , Aen. 6. 280, 375Google Scholar.

⁴⁵ ZPE 47 (1982), following p. 300; brought to my notice at a late stage by West, M. L., The Orphic Poems (Oxford, 1983)Google Scholar . West, who dates the text to the early fourth century (pp. 81 f.), assumes (p. 78) that the Erinyes (frr. A and B) are the same as the Eumenides (col. ii), though the published fragments, at least, do not make this certain. It must anyway be significant that the name is Erinyes in the context of crime and punishment, Eumenides in the context of cult.

⁴⁶ Cèbe, J.-P., Varron, Satires ménippées iv (Rome, 1977), 545Google Scholar. In Varro's satire Eumenides the Eumenides were evidently Furiae causing madness.

⁴⁷ The equation does not become invariable in Greek, however; note Orph. Hymns 69 (to the Erinyes) and 70 (to the Eumenides).

⁴⁸ Eumenides at Georg. 1. 278 (cf. 'Eριν⋯ες at Hes. Op. 803), 4. 483, Aen. 4. 469 (Pentheus haunted by Eumenides, Orestes by Dirae), 6. 250, 280, 375.

⁴⁹ It is likely (though we lack direct evidence) that the Erinyes were called Eumenides in Ennius' Eumenides (based on Aeschylus'). For Varro see n. 46 above. See also Lucil. 172, Cat. 64. 193, Cic. ND 3. 46.

⁵⁰ The analogy of τ⋯ μητρ⋯ς αἶμα, not μαν⋯αισιν, that has replaced a word for ‘Furies’.

⁵¹ See also Wedd ad loc. At Eubulus fr. 64 K ⋯νομ⋯ζειν γ⋯ρ αἰδο⋯μαι θε⋯ν is doubtless a complete sentence, but this proves little.

⁵² There is worse inconsistency in a similar context at I.T. 941–4; see n. 42 above.

⁵³ I lack space to discuss the puzzling word ποτνι⋯δες, applied to the Eumenides at Or. 318, to frenzied mares at Phoen. 1124, and to Bacchants at Bacch. 664. I doubt, however, whether it justifies the claim sometimes made, that Potniades or Potniai was a cult title of Erinyes or Eumenides (cf. n. 128 below).

⁵⁴ Besides his edition of Aesch. see his Opuscula II (Leipzig, 1827), 132–4Google Scholaribid, vi (1835), 2. 117–23.

⁵⁵ Aischylos: Interpretationen (Berlin, 1914), 228 fGoogle Scholar.

⁵⁶ As well as several editors (e.g. Paley, Verrall, Sidgwick, Blass, Smyth, Headlam-Thomson) see Müller, C. O. [K. O.], Dissertations on the Eumenides of Aeschylus ² (trans., London, 1853), 174 n. 3Google Scholarid., Anhang (Gottingen, 1834), 26–9; Robert, C., Hermes 38 (1903), 634–7Google ScholarHarrison, , op. cit. (n. 1), 252 f.Google ScholarMacleod, C. W., Maia 27 (1975), 201Google Scholar = Collected Essays (Oxford, 1983), 41Google ScholarTaplin, O., The Stagecraft of Aeschylus (Oxford, 1977), 412Google Scholarn. 2. See also n. 86 below.

⁵⁷ As he claimed at Opusc. vi. 2. 121.

⁵⁸ Nachmanson, E., Der griechische Buchtitel (Goteborg, 1941), 6 f.Google ScholarHunter, R. L., Eubulus: The Fragments (Cambridge, 1983), 146–8Google Scholar.

⁵⁹ Aesch. is told to recite the prologue from the 'ρ⋯στια, and replies unhesitatingly with the beginning of Cho.; so 'Oρ⋯pστεια should mean Cho., and not, as the schol. thinks, the trilogy or tetralogy. Then at Ar. Thesm. 135 Aυκουργ⋯α will mean simply Edonoi (‘Mnesilochus’ is about to quote a single passage from a single play); in IG ii.² 3091 (= TrGF i DID B 5) the Tηλ⋯øεια of Soph, will be simply Mysoi or Telephus (no tetralogy by Soph, is otherwise attested); and in Aristotle's edition of the Didascaliae the Oἰδιπ⋯δεια of Meletus (TrGFi 47 MELETUS I 1) will perhaps be a single play also. Alexandrian scholars, however, mistook 'ρ⋯στεια and Aυκοργ⋯α in Ar. for names of trilogies or tetralogies, and also gave the name A⋯κο⋯ργεια to a tetralogy by Polyphrasmon (TrGF i DID C 4) on the analogy of these. Or so it seems to me; cf. now Knox, B. M. W. et al. in Sophocle (Entretiens Fond. Hardt 29, 1983), 224 fGoogle Scholar.

⁶⁰ Verrall, A. W., The Seven Against Thebes of Aeschylus (London, 1887), xvi–xviii, xxxiiiGoogle Scholar. Verrall's argument seems conclusive on this point, whatever Aeschylus' reason may have been for omitting the name from the text.

⁶¹ Taplin, , op. cit. (n. 56), 12–18Google Scholar.

⁶² Haigh, A. E., The Tragic Drama of the Greeks (Oxford, 1896), 395–402Google Scholar.

⁶³ Admittedly there is no consistent principle at work here. The different plays about Prometheus were distinguished by epithets (West, M. L., JHS 99 [1979], 131)CrossRefGoogle Scholar, and each of the titles Δαναΐδες and Aυκο⋯ργος was arbitrarily assigned to one of at least two plays which could have borne it. But, if titles were assigned as haphazardly as I have argued, no consistent principle is to be expected.

⁶⁴ Macleod, loc. cit. (n. 56), while rejecting Hermann's theory, claims that the title ‘indicates the meaning of the whole conclusion of the trilogy, the reconciliation of the Erinyes’. We can be sure that no one would have conferred such a title for such a purpose – least of all Aesch., who presumably felt that the play's meaning was sufficiently indicated in the play itself.

⁶⁵ Schol. M on Eum. 761 also refers to the Chorus as Eumenides (while other scholia refer correctly to Erinyes).

⁶⁶ For the date see Turner, E. G., JEA 38 (1952), 92Google ScholarHemmerdinger, B., REG 72 (1959), 107–9Google Scholar. For other information about him see Schultz, H. in RE vii. 2. 2412–16Google Scholar.

⁸⁷ He is cited as a source for six fragments by Mette, H. J., Die Fragmente der Tragodien des Aischylos (Berlin, 1959), 276 fGoogle Scholar.

⁶⁸ It is uncertain whether Didymus wrote a commentary on Aesch.; see (in favour) Wartelle, A., Histoire du texte d'Eschyle dans Vantiquite (Paris, 1971), 185–95Google Scholar. In any case he did not always make good use of his reading; see West, S., CQ n.s. 20 (1970), 288–96CrossRefGoogle Scholar.

⁶⁹ Dem. 23. 66.

⁷⁰ For some reason the entry appears in Porson's and Hermann's editions but not in Naber's.

⁷¹ Not Eὐμεν⋯δες πραΰνας, as is claimed by Smith, O., Scholia in Aeschylum 1 (Leipzig, 1976), 42Google Scholar(I have checked the facsimile). I have relied, however, on Smith's reports of the other manuscripts.

⁷² As do Verrall, A. W., The ‘Eumenides’ of Aeschylus (London, 1908), xxxvi f.Google ScholarHarrison, , op. cit. (n. 1), 253Google Scholar. It is true that the schol. on Soph. O.C. 42 puts the placating and renaming of the Erinyes after Orestes' return to the Peloponnese (see p. 271 below), and in theory M's text of hyp. Eum. could be referring to this, in complete defiance of the play. But it is incredible that (as Harrison seems to believe) a misunderstanding of the hypothesis could have fortuitously led Harpocration to the correct statement that the Erinyes were placated by Athena.

⁷³ So Wilamowitz, loc. cit. (n. 55), as well as Harrison.

⁷⁴ Cf. Verrall, loc. cit. (n. 72). There seems to be no authority for Page's omission of πρ⋯ςτ⋯ν 'Oρ⋯στην in his app. crit. on Eum. 1027. Against the extraordinary claim of Robert, art. cit. (n. 56), that Harpocration accurately reports the contents of a lacuna in Eum., see Hense, O., RhMn.F. 59 (1904), 176 fGoogle Scholar.

⁷⁵ The argument is not impregnable, since our sentence could have replaced one which merely said ‘Athena placated the Erinyes’. And cf. the ‘Dicaearchan’ hyp. I Eur. Andr., which ends with a false statement generally considered spurious.

⁷⁶ Eur. II Ale, II Med., II Or., II Bacch.

⁷⁷ Cf. the convincing demonstration by Rusten, J., GRBS 23 (1982), 357–67Google Scholar, that the ‘Dicaearchan’ hypotheses were falsely attributed in antiquity.

⁷⁸ This information is repeated in the Suda s.v. Eὐμεν⋯δες

⁷⁹ For similar conflations see Eur. El. 1254–75, I.T. 940–82, Or. 1643–52; Paus. 8. 34. 4.

⁸⁰ FGrHist III b (Suppl.) i. 24 f., ii. 20–9. Admittedly Jacoby's argument, if I rightly understand it, does not really amount to much, so it remains possible in theory that “ἒνιοι” preceded and influenced Aesch. But I think it most unlikely; the conflation of traditions looks Euripidean or later. The usage seen in Eur. Or. could be derived from vague memories of the Cerynean legend, but could equally be independent of it.

⁸¹ The only other text known to me which mentions the change of name is the schol. on Aeschin. 1. 188: κληθ⋯ναι δ⋯ Eὐμεν⋯δας ⋯π' 'Oρ⋯στου, πρ⋯τερον 'Eριν⋯ας καλο⋯μ⋯νασ.

⁸² ⋯νδυτ⋯ν Hartung, ⋯νδυτ⋯ς Murray? (Murray's app. crit. refers to Headlam, W., JHS 26 [1906], 268CrossRefGoogle Scholar, but the emendation does not appear there, nor in Headlam-Thomson).

⁸³ Taplin, , op. cit. (n. 56), 408–10Google Scholar, considers that the last ‘act’ runs from 778 to the end; this is sensible in itself, but does not account for these isolated trimeters. I cannot forbear to point out, however, that there is a formal parallel at Sept. 1005–25, if my account of those lines is correct; see CQ n.s. 26 (1976), 206–19CrossRefGoogle Scholar(I should have mentioned Eum. 1021–31 on p. 212).

⁸⁴ J. F. Davies's edition (Dublin, 1885) places the passage before 1003; this is no improvement.

⁸⁵ For references see n. 54 above.

⁸⁶ Wecklein and Weil placed the lacuna before, not after, 1027, but 1026 and 1027 cohere together too well for this; 1027 was certainly designed (whether by Aesch. or, as I shall argue, by someone else) to follow immediately after 1026.

⁸⁷ For Müller, Macleod and Robert see n. 56 above.

⁸⁸ But for 1026 f., one would suppose that the πρ⋯σπολοι comprised only the priestess of Athena Polias and her two (?) assistants. Even if we count in, with Wilamowitz (n. 55) and Groeneboom ad loc, such figures as the two Arrhephoroi, the dubious Errhephoroi, the Loutrides or Plyntrides, and the Ergastinai (and it is more than doubtful whether all these could be said to guard Athena's image), the numbers are still not impressive. For Athena's personnel see Jordan, B., Servants of the Gods (Hypomnemata 55, Göttingen, 1979), 28–36Google Scholar.

⁸⁹ In the 1st person an opt. with ἅν can be a polite, or emphatic, way of expressing intention, while in the 2nd or 3rd it can express a request or suggestion; see Kuhner-Gerth i. 233 f. Kühner-Gerth i. 235 lists some places where an opt. with ἅν is parallel to a future, but even here it seems to have potential force. I have not found an instance in the 2nd or 3rd person expressing a simple prediction.

⁹⁰ Taplin, , op. cit. (n. 56), 410–15Google Scholar. Certainly πολισσο⋯ϰοι παῖδεσ Kραναο⋯ (1010 f.) can refer to no one but the Areopagites. Several attendants must enter before 1003 with torches and sacrificial animals (the σø⋯για of 1007 are now protected; see JHS 102 [1982], 31 n. 31), and I take τ⋯νδε προπομπ⋯ν (1006) to refer to them (rather than the Areopagites, as Taplin thinks); they will retain their torches throughout. Areopagites and attendants are here for a reason, and in time for the start of Athena's instructions; it would be a different matter to have women and children drifting on belatedly, merely in order to exit in the procession.

⁹¹ The Areopagites have been silent since their arrival at 566 and are to all appearances extras, so it would be surprising if they suddenly turned out to be a chorus (the children in Eur. Supp. could be compared, but Euripides' boy singers seem in general exempt from normal rules of dramatic economy). Also, if they are the only Athenians present (apart from attendants), and if we rule out ‘audience participation’ (Taplin 129–34, 394 f.), who are the ϰωρῖται whom they address at 1035, 1039, 1042 and 1047?

⁹² Athena obviously has the authority to issue instructions to citizens (Areopagites) and Furies and to reveal the dispensations of IIαλλ⋯δος ⋯στος (1045) in her mouth-more forcible here than ‘my citizens’ (862) would be-cf. Sept. 7, P. V. 296, 506; Il 8. 22, 470; Eur. Hel. 1667. If we class the song as a monody, it will perhaps be the only authentic example in Aesch. (but note Hera's lyric hexameters in fr. 355 M); I wonder, however, whether a ‘monody’ is a more serious anomaly than a subsidiary chorus, for, given the uncertainties surrounding Supp. 1018–73, we now face the prospect of eliminating subsidiary choruses from the work of Aesch. (and Soph.) altogether.

⁹³ The Eumenides by Aeschylus (Englewood Cliffs, 1970), 72.

⁹⁴ Cf. Eur. Phoen. 802, ' Aρτ⋯μιδο ϰιονοτρ⋯øον ὃμμα Kιθαιρών.

⁹⁵ In fact the precinct of the Semnai Theai seems to have been between the two hills (cf. Paus. 1. 28. 7 with Val. Max. 5. 3. 3). Some claim that the exact spot can be identified, e.g. Robert, op. cit. (n. 6), 42; Judeich, W., Topographie von Athen ² (Munich, 1931), 300Google Scholar.

⁹⁶ Cf. 833 (ζυνοικ⋯τωρ ⋯μο⋯), 916. It is also necessary to believe that the goal of the procession is the Acropolis if we are to accept Headlam's suggestion of a parallel with the Panathenaic procession (art. cit. [n. 82], 268–77; Headlam-Thomson on 1028–32). This is certainly one way to account for the red robes of 1028 (about which I have no private theories).

⁹⁷ For λ⋯ϰος of the Erinyes cf. 46; for εὐκλε⋯ς cf. κλ⋯τ⋯ at Cho. 650.

⁹⁸ This reading is suggested by Lloyd-Jones's translation, loc. cit. (n. 93): ‘You shall come to the very eye Of Theseus' land, O honorable band of children…’. But I have not been able to obtain confirmation that ⋯ζ⋯κοισθ' ἂν is what Professor Lloyd-Jones in fact reads.

⁹⁹ This is pointed out by e.g. Müller and Robert (n. 56). Robert, rejecting the lacuna after 1027, wanted to place the renaming after 1031, but it is no better to separate the arrangements for the procession from the procession itself than to split up the arrangements; see also Hense, art. cit. (n. 74), 170–85.

¹⁰⁰ Hermann, (Opusc. vi. 2. 120)Google Scholarseems to have thought that the Athenians were told to honour the Furies now with the name Eὐμεν⋯δες, but the use of the name can hardly form part of the processional ritual.

¹⁰¹ It is no use inserting the bare word Eὐμεν⋯δες, or a synonym, into a corrupt line such as 1040; nor taking εὔøρων at 992,1030,1034 as hinting at Eumenides (cf. Winnington-Ingram, R. P., Studies in Aeschylus [Cambridge, 1983], 166 n. 39)Google Scholar . Such theories credit Aesch. with pointless mystification, while failing to square with the external evidence which prompts them.

¹⁰² Eὐμεν⋯δες in the two testimonia need not be a mistake for Eὐνεῖδαι: see Goossens, R. and Darquenne, E., Chronique d'Egypte 33 (1942), 127–32Google ScholarPieters, J. T. M. F. in Miscellanea tragica in honorem J. C. Kamerbeek (Amsterdam, 1976), 267 fGoogle Scholar .

¹⁰³ On the other hand the Eume[nides] of Teleclides, in which Edmonds does believe (FAC i. 182), rests on pure speculation by Koerte, A., RhM 60 (1905), 441Google Scholar .

¹⁰⁴ E.g. Edmonds, L., HSCP 85 (1981), 227 f.Google ScholarKamerbeek, J. C., Mnem. Ser. 4. 35 (1982), 42Google Scholar .

¹⁰⁶ See esp. Bowra, C. M., Sophoclean Tragedy (Oxford, 1944), 317–19Google ScholarWinnington-Ingram, R. P., JHS 74 (1954), 16–24CrossRefGoogle Scholarid., Sophocles: an Interpretation (Cambridge, 1980), 216, 264–75, 325 f.

¹⁰⁶ Sophocles the Playwright (Toronto, 1957), 165Google Scholar .

¹⁰⁷ For some counter-arguments see Knox, B. M. W., The Heroic Temper (Berkeley & Los Angeles, 1964), 194 n. 12Google Scholar .

¹⁰⁸ The fact that Soph, died with O.C. unperformed need not mean that he did not complete it until 406, for he may never have completed the three plays which should have accompanied it. The story in the Life of Eur., that Soph, dressed a chorus in black in mourning for his rival, should imply, if true, that he produced something at the Dionysia of 406, but little time would be left for the writing of O.C. between that date and his death. I am also unconvinced by the idea that the play was inspired by the Athenian victory mentioned at Diod. 13. 72 f. and there dated to 408/7. For other views see R. G. Tanner in For Service to Classical Studies: essays in honour of Francis Letters (Melbourne, 1966), 153–92Google ScholarStoessl, F., Dioniso 40 (1966), 5–12Google Scholar .

¹⁰⁹ Nor in that of any predecessor who may have influenced Eur. (p. 266 above).

¹¹⁰ The fact that Aesch. and Eur. had the Erinyes ‘converted’ (by implication) into Semnai Theai, and that Soph, equates the Semnai Theai with the Eumenides, clearly does not mean that Soph, equates the Eumenides with the Erinyes. However popular the Oresteia may have been, the Athenians could surely think of the familiar cult of the Semnai Theai without thinking of the myth of Orestes. Nor is the audience of O.C. given any reason to recall Eum.; if they recall any Aeschylean play, it will doubtless be Eleusinioi (the suppliant play which Aesch. set in his home deme).

¹¹¹ Winnington-Ingram, , Sophocles (n. 105), 267Google Scholar .

¹¹² The point was not lost on the scholiast on Aeschin. 1. 188, who, discussing the Semnai Theai of Athens, wrote øασ⋯ μ⋯ντοι αὐτ⋯ς οἱ μ⋯ν I⋯ς εἶναι κα⋯ ζκ⋯τους… (cf. O.C. 40).

¹¹³ It is not generally accepted that 1299 refers to a curse by Oedipus, but the point will be argued below.

¹¹⁴ The descriptions of the grove may have some sad and sinister undertones (Segal, C., Tragedy and Civilization [Cambridge, Mass., 1981], 371–6)Google Scholar , but the dominant impression is of tranquillity and beauty.

¹¹⁵ Linforth, I. M., UCPCP 14 (1950–1952), 92–7Google Scholar , denies that the Eumenides have any connection with Oedipus' death. He is right to point out that they are not mentioned at the end of the play, but they surely do more than ‘create…an atmosphere of solemnity’; we must assume that the prayer at 101–10 is answered, though in a mysterious and undefined way.

¹¹⁸ E.g. Whitman, C. H., Sophocles (Cambridge, Mass., 1951), 201 f.CrossRefGoogle ScholarGellie, G. H., Sophocles: a reading (Melbourne, 1972), 161Google Scholar . These scholars also claim that the Eumenides are identical, not with ‘normal’ Erinyes, but with the reformed Erinyes at the end of Aesch. Eum., but, as we have seen, this is difficult to reconcile with the fact that unreformed Erinyes are still at large (see also n. 110 above).

¹¹⁷ The beauties of the grove include the ν⋯ρκισσο, μεγ⋯λαιν θεαῖν ⋯ρϰαῖον στε⋯⋯νωμ', ὃτε ϰρυσαυγ⋯ς κρ⋯κος (O.C. 683–5). There can be little doubt that the dual is the right reading and that the Great Goddesses are Demeter and Kore, but the schol. on 681 reports a controversy in which one side claimed that these goddesses were the Eumenides (or ‘Erinyes’) and cited Euphorion fr. 94 P, where the Eumenides (as agents of punishment) are ναρκ⋯σσο⋯ ⋯πιστε⋯⋯ες πλοκαμῖδας. The ambivalence of the Eumenides, well illustrated by this quotation, does resemble that of Demeter and Kore, and this adds point to Sophocles' reference to the latter here.

¹¹⁸ Cf. Adams, loc. cit. (n. 106); Linforth, , art. cit. (n. 115), 93Google Scholar .

¹¹⁹ Loc. cit. (n. 107).

¹²⁰ Campbell and Jebb take ἒτι not with μ⋯ but with τ⋯σδε τ1FC6;ς ⋯ρ⋯ς; Jebb renders, ‘Nay, may the Powers of this place suffer me to utter yet this curse’, in addition to the curse on the sons at 421 ff. I am doubtful whether ετι can be used in this way without ὃλλος or the like, but my argument is little affected, for Jebb's rendering still implies that the Eumenides might not suffer Oedipus to utter a curse. Cf. Campbell ad loc. (‘Though they enjoin silence from cursing, let them not enforce it here’), and contrast Radermacher ad loc. (‘Die Eumeniden hieCen selbst 'Aραι…’) .

¹²¹ Art. cit. (n. 115), 93. Silence is appropriate in the Eumenides' presence, as Dr Richardson has pointed out to me, citing 128–32, 489 (with schol. ad loc).

¹²² Loc. cit. (n. 115).

¹²³ E.g. Winnington-Ingram, , Sophocles (n. 105), 326Google Scholar : ‘the generous reception… which is also, mysteriously and paradoxically, a reception by the Eumenides who, as Erinyes, have dogged his tragic career’.

¹²⁴ See Edmonds, , art. cit. (n. 104), 225 fGoogle Scholar .

¹²⁵ For the ‘Lille Stesichorus’ see Parsons, P., ZPE 26 (1977), 7–36Google Scholar . That Oedipus is dead or exiled here seems to follow from the division of his property; the awkward notion of preserving him as a skeleton in the cupboard is unlikely to predate Eur. That the mother of Eteocles and Polynices is Epikaste or Jocasta (i.e. the mother of Oedipus) and not Euryganeia cannot be proved (Parsons, p. 20; Gostoli, A., GRBS 19 [1978], 23–5)Google Scholar , but is suggested by the analogy of Aesch. Sept. (missed by Parsons and Gostoli) and Eur. Phoen.

¹²⁶ See in general Baldry, H. C., G&R Ser. 2. 3 (1956), 24–37Google Scholar .

¹²⁷ That this is true of Aesch. has not often been noticed, but is shown by Sept. 926—31. If the death of Eteocles and Polynices is one of the factors making Jocasta the most unfortunate of mothers, she must have lived to see it.

¹²⁸ Some see significance in the fact that Aesch. had Laius killed at Potniae (fr. 172 M), but the eponymous Potniai seem to have been Demeter and Kore (Paus. 9. 8. 1; cf. Ar. Thesm. 1149, 1156, Soph. O.C. 1050), not the Erinyes (cf. n. 53 above).

¹²⁹ Commentators hasten to warn us that hysteron can be proteron, erreira need not be temporal, and the lines need not mean what they appear to say. One wonders who warned the audience in the theatre – an audience which, we must remember, could have known the work of Stesichorus and Aesch. but could not have known O. T. It is true that we do not elsewhere find a version in which Jocasta survives Oedipus and then hangs herself, but such a version is quite intelligible as a compromise between Stesichorus and Homer. And, if Soph, is hinting that she hanged herself because her son and husband was dead, has this not some thematic relevance to a play that includes the suicides of Haemon and Eurydice?

¹³⁰ Parker, , op. cit. (n. 34), 109Google Scholar , thinks that Laius' Erinyes ‘remarkably’ spare Oedipus because he acted in ignorance; but Laius' ignorance seems a more likely explanation.

¹³¹ Apollo is aided by Keresin a highly figurative passage at O.T. 469–72. O.T. 417 f. (Oedipus will be exiled by the δειν⋯πους ⋯ρ⋯ of his parents) must also be understood figuratively.

¹³² Admittedly he is not consistent; see 1132–8.

¹³³ //. 21. 412; Od. 11. 280, 17. 475; Hes. Theog. 472; Thebais fr. 2. 8; Ale. fr. 129. 14 L-P; Aesch. Sept. 70, 723, 887 (cf. Ag. 1432 f.); Hdt. 4. 149. 2; Soph. Ant. 1075, O.C. 1434; Eur. Med. 13S9, hoen. 624. For some later examples see Rohde, E., RhM 50 (1896), 10 f.Google Scholar = Kl. Schr. 233–5.

¹³⁴ Op. cit. (n. 107), 194 n. 14. The argument seems to me valid despite Winnington-Ingram, , Sophocles (n. 105), 266 f. n. 50Google Scholar ; anyway the combination of 1299 and 1375 appears conclusive.

¹³⁵ In fact, as is pointed out by Robert, op. cit. (n. 6), 479 f., Soph, is following the version of the epic Thebais, where the sons were condemned by one curse to quarrel, by another to die at each other's hands. The fact that at O.C. 369 the quarrel seems to have been caused by the ancient curse on the whole family is quite compatible with a further causal strand derived from a curse uttered by Oedipus (cf. Aesch. Sept. 720–91).