¹ He was pursuing a theory about Aeschylean choruses, which was refuted by Hermann, G., Opuscula vi (Leipzig 1835) 2.127–46Google Scholar. Whallon, W., Problem and Spectacle (Heidelberg 1980) 88–99Google Scholar, has now attempted to revive Müller's theory.

² Wilamowitz-Moellendorff, U. von, Griechische Tragödien ii (Berlin 1919) 247Google Scholar; Dirksen, H. J., Die aischyleische Gestalt des Orest (Nuremberg 1965) passim, esp. 16–18Google Scholar.

³ Solmsen, F., Hesiod and Aeschylus (Ithaca N.Y. 1949) 186Google Scholar n. 34. Reinhardt, K., Aischylos als Regisseur und Theologe (Bern 1949) 135–40Google Scholar, describes the end of Cho. in terms of the appearance of forces that were previously hidden. For a particularly dogmatic denial of ‘psychology’ in this scene see Schmid-Stählin, , Geschichte der gr. Lit. i 2.242 n. 7Google Scholar.

⁴ Jones, J., On Aristotle and Greek Tragedy (London 1962) 104Google Scholar.

⁵ For Erinyes in general see Wüst, E., RE Suppl. viii 82–166Google Scholar; Dietrich, B. C., Death, Fate and the Gods (London 1965) 91–156Google Scholar.

⁶ Ag. 59, 463, 645, 749, 991, 1119, 1433, 1580, Cho. 402, 577, 652; cf. the δαίμων of Ag. 1468 ff., the Alastor of Ag. 1501 ff., the Ἀραί of Cho. 406.

⁷ The v.l. εἰαροπῶτις (for ἠεροφοῖτις) Ἐρινύς at Il. xix 87 is seized upon (with a wrong reference) by Harrison, J. E., Prolegomena to the Study of Greek Religion³ (Cambridge 1922) 215Google Scholar n. 2, but it is doubtless derived from the blood-drinking Furies of the Oresteia, since εἶαρ, ‘blood’, does not occur in early Greek, and Aeschylus is actually mentioned in the note (Schol. T) which alone supplies the variant. See also n. 65 below.

⁸ Cf. e.g. Lebeck, A., The Oresteia (Washington 1971 97 f., 108 f., 114–16, 200 f.Google Scholar, with the judicious comments of Taplin, O., The Stagecraft of Aeschylus (Oxford 1977) 359 f.Google Scholar

⁹ I must be brief and dogmatic here, since I do not wish to become involved in the complex controversy over Orestes' attitude to the matricide before he performs it. (Does the Kommos strengthen his resolve? How real is his hesitation at 899?) My own view is that he 15 presented as feeling an instinctive revulsion from the matricide, but never doubts the moral and practical necessity of overcoming this; and that, while the revulsion should evoke our sympathy, his ability to overcome it is to be seen as a mark of heroism.

¹⁰ Signs of madness were seen in this speech by Wilamowitz (p. 42 of his edition of Cho.; Aischylos: Interpretationen [Berlin 1914] 215Google Scholar) and several other commentators, but this idea is convincingly rejected by Fraenkel, , Agamemnon iii 810 f.Google Scholar

¹¹ See esp. Fraenkel loc. cit. and Lloyd-Jones, H., CQ xi (1961) 181–4Google Scholar. But perhaps after all the best solution is that of Scholefield, to place 997–1004 before 983. There is no reason why Orestes should not describe the robe to the Chorus (and the audience) before having it displayed to the Sun.

¹² A different meaning can be obtained here by emendation (so Blass) or by far-fetched interpretation (so Tucker), but the tone of these anapaests is in any case so manifestly different from that of the last ode that the attempt seems pointless. We must beware of assuming, however, that talk of wretched deeds and a vile death means that Orestes was actually wrong to kill Clytemnestra.

¹³ So Taplin (n. 8) 358.

¹⁴ Ag. 475–88 notoriously illustrates how a chorus's attitude can be changed in response to dramatic requirements without much regard for consistent motivation (though there, as here, there are factors which help to justify and mitigate the change). At the end of Eur. El. (see Denniston on 1147–1232) there are shifts in the Chorus's mood similar to those in Cho. and serving much the same purpose.

¹⁵ Reinhardt (n. 3) 138 f. does justice (or more than justice; see n. 56 below) to the importance of the transition that occurs here, but in talking of hidden forces revealing themselves he does not really interpret the passage in its own terms. Weil takes more trouble than most editors with the text of these lines, but his insertion of 997–1004 after 1013 compels him to rewrite 1014–17 in a way which, while certainly ingenious, takes him unduly far from the MS.

¹⁶ Gr. Trag. (n. 2) ii 152, Aisch. Int. (n. 10) 215.

¹⁷ This is further confirmed by the fact that μαρτυρεῖ μοι should, as can be seen from the lexicon, mean ‘testifies on my behalf’, not ‘testifies to me’. δέ, if correct, will be continuative, the preceding double question being tantamount to an emphatic statement; though I strongly suspect that it should be γέ (a ‘question-answering γέ’ not implying any doubt about the sufficiency of the robe's evidence).

¹⁸ So several commentators here; also Fraenkel, , Agamemnon iii 703Google Scholar n. 2. Among other interpretations Hermann's αὑτὸν (i.e. ἐμαυτόν) has been popular, but it makes nonsense of παρών and goes against the evidence of line 8.

¹⁹ Dr Garvie points out to me that other corrections are possible; but in a one-MS play lacunae are always among the first possibilities to consider.

²⁰ I see no good reason for the change of number and suspect that it should be πάθη (πάθας Weil, not an Aeschylean word). The corruption would be a natural one before γένος.

²¹ Cf. Smyth, H. W., Aeschylean Tragedy (Berkeley 1924) 203Google Scholar: ‘the acute consciousness of his unhappy state, produced by a deed of such frightful and unheard-of justice, rather than the agonies of a sin-stricken soul’. Such distinctions seem unknown to Class, M., Gewissensregungen in der griechischen Tragödie (Hildesheim 1964) 46–8Google Scholar.

²² For this aspect of the robe see Lebeck (n. 8) 67 f.

²³ This interpretation, however, seems to me to involve reading Weil's ἡνιοστροϕῶ rather than Stanley's ἡνιοστροϕῶ for ἡνιοστρόϕου in 1022. Stanley's reading virtually forces us to take ὡς ἃν εἰδῆτ with this verb, thus crediting Orestes with a strangely formal announcement of his impending madness.

²⁴ It is true that madness is described in similar imagery elsewhere in tragedy (PV 883 f., Eur. Ba. 853), but even if such imagery were already standard by 458 this would not remove the need to consider how it functions in this particular context. In any case, we may now take it that PV, as well as Ba., is post-Aeschylean; PV 878–86 may then be directly influenced by Cho. 1021–5.

²⁵ Reading Abresch's ἡδ' with most editors, rather than M's ἠδ', which Page accepts. It is surely more pointed for the heart to dance to the accompaniment of Fear's singing (cf. 167, ὀρχεῖται δὲ καρδία ϕόβῳ) than for Fear to dance to its own accompaniment.

²⁶ If κότῳ is right, I suppose the word must be extended to mean ‘frenzy’ (so Paley) rather than ‘anger’, This is not easy, but no emendation that I have seen (including Abresch's popular κρότῳ) carries much conviction.

²⁷ Aeschylean Metaphors for Intellectual Activity, Hermes Einzelschr. xxxv (1975) 72Google Scholar.

²⁸ Sept. 289; Ag. 179, 834, 976.

²⁹ For irrational fear and its significance in Aeschylus see Romilly, J. de, La crainte et l'angoisse dans le théâtre d'Eschyle (Paris 1958) 55–106Google Scholar.

³⁰ At Symb.Osl. xxxii (1956) 6 f.Google Scholar Rose places more emphasis on the psychological aspects of the scene, talking of a ‘compromise’ between real Erinyes and the subjective visions of a guilty conscience. But I do not find ‘compromise’ a very satisfactory word.

³¹ Embarrassment at the text of Cho. and desire to explain it away are overtly displayed by Solmsen loc. cit. (n. 3). He bases his argument not only on the reality of the Furies in Eum. but also on his account of the intervention of the Erinys in Sept., an account which in my view is mistaken; see Brown, , Phoenix xxxi (1977) 309–11Google Scholar. He does not attempt to explain how we can escape the apparent implications of the text of Cho.

³² In 1026 f. ‘I announce to my friends and say that …’ “is hardly attractive. Lobel reverses the order of 1027–8, but this weakens the effect of 1028; better to mark a lacuna after 1026, ‘I announce to my friends <that …> and I say that …’

³³ The text of 1041 is irrecoverable, but Orestes seems to be telling the Argives to bear witness on his behalf, so his speech continues to have the practical purpose of self-defence.

³⁴ The lacuna after 1042 or 1043 presumably contained not only the main verb of the sentence but also a reference for τάσδε κληδόνας, for otherwise the words must mean simply ‘the reputation of being a wanderer and an exile’, which is very weak. Wilamowitz's 〈ἄπειμι, μητρὸς αὐτοχειρία ϕονεύς〉 (after 1042) would be acceptable. It is unlikely that the Furies were mentioned in the lacuna: I would expect any reference to them in this speech to come before the reference to Delphi.

³⁵ Cf. Mattes, J., Der Wahnsinn im griechischen Mythos und in der Dichtung … (Heidelberg 1970) 78Google Scholar.

³⁶ E.g. Macbeth iii 4Google Scholar, where the Ghost of Banquo is visible to Macbeth and the audience but not to the other characters.

³⁷ There is also the evidence of Euripides, who clearly has Cho. in mind in El., Or. and IT. At El. 1342f. κύνας τάσδε briefly indicates that the Furies are present, but much more would have to be said if they were brought on stage; in Or., where Orestes' madness is heavily stressed, their presence on stage would be out of the question; and at IT 281 ff. we have a description of Orestes seeing Furies invisible to the speaker.

³⁸ It may therefore have displaced some quite dissimilar word. I find Lobel's δμοιαὶ, which Page reads (see Fraenkel, , Agamemnon ii 318Google Scholar n. 1), more ingenious than convincing.

³⁹ So Reinhardt (n. 3) 140; Groeneboom on 1051–8.

⁴⁰ Cf. Devereux, G., Dreams in Creek Tragedy (Oxford 1976) 157Google Scholar.

⁴¹ Burges' στάζουσι νᾶμα, favoured by some editors, gains support from Eum. 54 (where λίβα is also Burges' conjecture but generally accepted). But if Orestes is merely saying that the Furies' eyes run, this seems a trivial detail to be singled out for mention here.

⁴² It would be possible to account for both these echoes in psychological terms, by saying that Orestes is subconsciously picking up words that the Chorus–leader has used and incorporating them in the fantasy that is forming in his mind. But this is a kind of psychological subtlety that I should not expect to find in Aeschylus.

⁴³ For text and interpretation of 1059 f. see ‘Problems’ 31.

⁴⁴ Cf. also the madness of Cassandra in Tro. and Agaue in Ba., which is plausibly depicted although its causes are supernatural; and contrast that of Ajax in Aj. and Heracles in HF, which, for dramatic reasons, arises from arbitrary divine intervention to the virtual exclusion of psychological realism. See also n. 51.

⁴⁵ A paranoid person will naturally imagine himself possessed by whatever Furies, devils or Martians his society happens to believe in, and his account of them will reinforce the belief of others: cf. Dodds, E. R., The Greeks and the Irrational (Berkeley/L.A. 1951) 66Google Scholar f.

⁴⁶ G & R xx (1973) 5 f.: the whole of the paragraph is highly relevant here.

⁴⁷ One of the few things known about Stesichorus' Oresteia is that Apollo gave Orestes a bow with which to ward off the Erinyes (fr. 40 P). This suggests that he presented them as tangible demons, although we need not assume that he was consistent in this presentation or that, even if he was, his audience was unable to see a connection between these demons and the forces of madness that might pursue a matricide in real life (see also n. 49).

⁴⁸ They are twice associated with ἄτη (Il. xix 87 f., Od. xv 233 f), but in both cases this is simply folly of an everyday kind, far short of insanity.

⁴⁹ At Paus. viii 34, however, we find an aberrant and primitive-sounding Arcadian version of the Orestes story in which Orestes goes mad and is pursued by Μανίαι. These were actually worshipped in a local cult, and in Pausanias' opinion were identical with the Eumenides. Unless there is literary influence here (and it hardly seems likely), this is proof that at least one version of the non-literary myth made the madness explicit.

⁵⁰ There might be some temptation to regard Orestes' vision as prophetic: the Furies he sees in Cho. will later start pursuing him in reality. But there is no indication of this in the text, and the idea will not bear close examination.

⁵¹ Or. 37 f., 238 and esp. 408–12. This last passage comes shortly after Orestes' famous description of his disease as σύνεσις (396), but he (in a moment of sanity) and Menelaus can still agree, in language that cannot easily be seen as metaphorical, that the goddesses he sees when mad are actually pursuing him.

⁵² Note also PV 673 f., where the onset of Io's madness coincides with her actual change of form, and 878–86, where she describes her madness in graphic terms, recalling those used by Orestes (see n. 24), after a scene in which the reality of her visions was not doubted by anyone.

⁵³ I shall assume here that we do see them immediately after line 63, as was argued in ‘Problems’ 26–8.

⁵⁴ Cf. Dirksen (n. 2) 4 f.

⁵⁵ For yet another function see p. 30 below.

⁵⁶ Thus it is here that unseen forces break into the action of the trilogy, and the idea of Reinhardt (n. 3) that the major change comes around Cho. 1010–17 is, I believe, seriously misleading.

⁵⁷ Similarly Tucker on his Cho. 1046.

⁵⁸ That he is still polluted was argued in ‘Problems’ 30–2. Tha the is still mad is not obvious from the text of the scene at Delphi, but we are given no reason to think that he has become sane, and it may be partly to preserve ambiguity in this matter that Aeschylus gives him only three lines in the scene.

⁵⁹ Dirksen (n. 2) passim; Class (n. 21) 46–65. Wilamowitz (n. 2) actually seeks to explain the Furies' continued presence after the trial in terms of Orestes' conscience, but this seems evidently impossible, since they are there concerned not with Orestes but with Athens.

⁶⁰ For a slightly different argument see Rohde, E., RhM i (1895) 10Google Scholar ( = Kl. Schr. 233) n. 1.

⁶¹ In fact Taplin's ἀλήτορα, ‘wanderer’, ([n. 8] 379 n. 4) seems irresistible, even though Bain, D. is right to point out (JHS xcix [1979] 172Google Scholar) that Hesychius' ἀλήτωρ, ‘priest’, will be primarily connected with λήτωρ and not with ἀλᾶσθαι. For ἀλήτωρ beside ἀλήτης cf. ἁρμόστωρ (Eum. 456 ἃπ. λεγ.), ἀσπίστωρ (Ag. 403 ἃπ. λεγ.), γεννήτωρ (Supp. 206), δέκτωρ (Eum. 204 ἃπ. λεγ.), εὐνάτωρ (Supp. 665), κωλύτωρ (fr. 17.20 M ἃπ. λεγ.), (ξυν)οἰκήτωρ (Supp. 952, Eum. 833), πορθήτωρ (Ag. 907, Cho. 974), beside ἁρμοστής, ἀσπιστής, γεννητής, δέκτης, εὐνάτσς, κωλυτής, οἰκητής, πορθητής.

⁶² It is true that when we stand outside the play we a free to see the Furies, the gods and indeed Orestes himself as significant in more or less symbolic ways. Orestes, we may say, embodies a test case, showing how a man can be guilty by one standard and innocent by another, and the Furies embody that law, or human instinct, which would find such a man guilty. Hence perhaps, the parallels which Class finds, (n. 21) 58–65 between their operation and that of conscience. But when we ask whether Orestes feels guilt, we are asking a question about a character within the play and accepting the fiction of his existence, and the answer to such a question must depend on what is actually said by him or about him in the text.

⁶³ The point remains valid for my purpose even though the Furies will retain many, perhaps all, of their old functions (927–37). Some change must occur around 900, even if it is only a change of attitude enabling them to perform their old functions to the advantage of Athens.

⁶⁴ At 230 ἄγει γὰρ αἷμα μητρῷον need mean little more than ‘the crime of matricide leads me on’; cf. e.g. Cho. 1038, φεύγων τόδ᾿αἷμα κοινόν.

⁶⁵ The sole instance I have found is Soph. fr. 743 P/R, which Radt prints in the form Τεισὼ δ᾿ ἄνωθεν †εστινη† αἱματορρόφος. Τεισὼ is R. Pfeiffer's conjecture for τίσω (WS lxxix [1966] 63–5), and is supposed to be a hypocoristic form of Τεισιϕόνη): this is plausible, especially since Τεισιϕόνη is attested much earlier than Pfeiffer himself realised (see Dyer, R. R., JHS lxxxix [1959] 52Google Scholar). The epithet will then be a reminiscence of Aeschylus. For Homer see n. 7.

⁶⁶ Nilsson, M. P., Greek Popular Religion (New York 1940) 91Google Scholar; MacDowell, on Ar. Vesp. 1035Google Scholar.

⁶⁷ To the modern reader the picture of hideous old women dancing round their victim and chanting spells at once suggests witches. ϕαρμακίδες were known in the fifth century (Ar. Nub. 749), but I cannot find evidence that the Greeks pictured them in this way.

⁶⁸ See esp. Winnington-Ingram, R. P., Gnomon xxiii (1951) 417Google Scholar and JHS lxxiv (1954) 18–21Google Scholar.

⁶⁹ E.g. Solmsen (n. 3) 189: ‘Now the antagonism between Zeus and the Erinyes becomes acute.’ But he himself goes on to qualify this (197 ff.), making some of the points that I shall make here; cf. also Taplin (n. 8) 408 n. 2.

⁷⁰ Cf. Gagarin, M., Aeschylean Drama (Berkeley/L.A. 1976) 72 f.Google Scholar Note also Eum. 269 ff, where the Furies sing with relish of Hades punishing offences against god, guest or parents.

⁷¹ At the beginning of this antistrophe there is no agreement as to even the approximate sense. The view commonly held, that the Furies are claiming to relieve the other gods of the task of punishment, would favour my present argument; but τινα (360) cannot mean ‘the other gods’. More than one scholar has conjectured Δία, which could have become τινα by way of Ζῆνα: so e.g. σπευδομένα (M^ac) δ ἀϕελεῖν Δία (Heath) τάσδε (vulgo) μερίμνας, θεῶν [δ'] (Hermann) ἀτέλειαν ἐμαῖς μελέταις (Voss) ἐπικραίνω (Hartung), and then as Page. ‘We do Zeus' dirty work,’ the Chorus sing bitterly, seeing things from their own warped viewpoint, ‘but of course such vile creatures are beneath his notice.’ But one could hardly rely on this.

⁷² His subjection of Kronos is mentioned (640 ff), and the emphatic statement at 19 that Apollo is the prophet of Zeus might remind the audience that the older oracular powers were not. But these hints cannot suffice to convert the theology of Eum. into that of PV.

⁷³ For the function of PV 515–20 in its context, and the fact that even here the author is not really interested in a ‘theological hierarchy’ (not that Eum. would be much affected if he were), see Fraenkel, , Agamemnon iii 729 f.Google Scholar

⁷⁴ Eum. 1045 f., Ζεὺς παντόπτας οὕτω Μοῖρά τε συγκατέβα, is constantly taken in accounts of Aeschylean religion (e.g. recently Winnington-Ingram, R. P., Sophocles: an Interpretation [Cambridge 1980] 156–8CrossRefGoogle Scholar), and often also in translations and commentaries (see esp. Thomson ad loc.), to refer to a reconciliation between Zeus and Moira. But the singular verb does not encourage this interpretation (cf. the distinctions drawn in Kühner–Gerth i 77–9), and the -κατα- element practically rules it out. The only other instance of συγκαταβαίνειν in Aeschylus is Cho. 727, where Peitho is to ‘descend to the aid of’ or ‘enter the contest on the side of’ Orestes. Similarly here the powers which collectively determine the future (coupled for emphasis as at Soph., Phil. 1466–8Google Scholar, Eur. El. 1248) are now allied with Athens: ‘Thus hath all-seeing Zeus and Fate entered the contest on behalf of the citizens of Pallas’ (Paley ad loc.). The interpretation will stand whether or not it is right to punctuate after ἀστοῖς.

⁷⁵ Winnington-Ingram, R. P., CR xlvii (1933) 97–104Google Scholar; cf. Kitto, H. D. F., Form and Meaning in Drama (London 1956) 68Google Scholar. It is worse still to suppose, with Kitto, that, because the fate of Cassandra could be seen from one point of view as the work of Apollo and from another as the work of the Furies (those of the House of Atreus), Apollo was conceived as making use of the Furies in Ag.

⁷⁶ Cf. Solmsen (n. 3) 181 f.; Dawe, R. D., PCPS clxxxix (1963) 59Google Scholar.

⁷⁷ Cf. Kitto, , Greek Tragedy³ (London 1961) 91Google Scholar.

⁷⁸ I am largely in agreement with the professed attitude of Dawe (n. 76) 21–6. It is unfortunate that, by the time he reaches Eum. (58 f.), Dawe's good intentions have been overcome by the temptation to be clever at the dramatist's expense.

⁷⁹ But the change in Apollo's character since his punishment of Cassandra, which is seen here by Kitto loc. cit. (n. 75), seems to me illusory. Cassandra was punished for breaking her word to Apollo, whereas Orestes is rewarded for obeying and trusting him: in both cases Apollo preserves his τιμή. It is true that in Eum. we are shown a more benevolent aspect of Apollo, as of Zeus, but there is no indication in Ag. that he is never benevolent, nor in Eum. that he is now entirely civilised in his treatment of women.

⁸⁰ For some other instances of ‘false preparation’ (but not this one) see Taplin (n. 8) 94–6.

⁸¹ One reader who presumably took the scene in this way was Richard Wagner: the failure of the Norns' prophetic powers in the Prelude to Die Götterdämmerung similarly indicates that prophecy is at an end and that the future is no longer determined by the past (though for quite un-Aeschylean reasons). See now Ewans, M., Wagner and Aeschylus (London 1982) 213–18Google Scholar.

⁸² But if, as I have suggested (‘Problems’ 30), Athena, like Apollo, is first seen on the roof, then Aeschylus is very cautious in bringing each of these gods gradually into contact with the action in the orchestra, Orestes does not, in fact, meet an Olympian at ground level until 566; and even the Furies do not actively confront him until 244. Cf. Melchinger, S., Die Welt als Tragödie i (Munich 1979) 115 f.Google Scholar

⁸³ On Homeric dreams and their relation to dreams in real life see Kessels, A. H. M., Studies on the Dream in Greek Literature (Utrecht 1978) 155–62Google Scholar and passim. The extent and purpose of poetic stylisation in Homeric dreams have been much disputed, as Kessels' account shows, and I do not claim that the point I am making explains all the phenomena.

⁸⁴ Dodds (n. 45) 88 n. 46 cites Plato Apol. 22c, Meno 99c, Ion 534c. The Euripidean Cassandra, in the first part of her scene (Tro. 308–64), gives rather the impression of being drawn from life.

⁸⁵ Herodotus i 182 ascribes this belief, which he rejects, to certain Chaldean, Egyptian and Lycian cults.

⁸⁶ Cf. Brown (n. 31) 308 n. 25.

⁸⁷ Zeus' seduction of Io, however true and significant in the world of Supp., is sufficiently remote in time to count as myth from the point of view of the Danaids themselves rather than as part of the ‘offstage action’. I am similarly untroubled by incidental references to the Sphinx at Sept. 539 ff., 776 f.

⁸⁸ The scrap of learning ἐκ τῆς Μουσικῆς Ἱστορίας found at the end of the Life of Aeschylus remarks, for what it is worth, that some of his tragedies διὰ μόνων οἰκονομοῦνται θεῶν καθάπερ οἱ Προμηθεῖς. But one may doubt whether the author could have cited any examples besides οἱ Προμηθεῖς (to which the information as given does not strictly apply).

⁸⁹ The fact that Aphrodite was a character in Danaides (fr. 44 M) could well mean that in this tetralogy, as in the Oresteia, there was a shift to the supernatural level in the third play (even though here the Chorus evidently remained human). In the Theban tetralogy the one irreducibly supernatural feature of the story, the Sphinx, is carefully relegated to the satyr play (she must doubtless have been mentioned in Oedipus, but the mention could have been brief and incidental to the play's action, as in Sophocles' OT). Psychostasia, on the generally accepted view, was set partly at Troy and partly on Olympus, but the extreme scepticism of Taplin (n. 8) 431–3 may well be justified.

⁹⁰ The case would evidently be altered, though we cannot tell exactly how, in plays which involved a divinity going unrecognised in the human world (Hera in Semele, Dionysus in Edoni).

⁹¹ Pp. xiii–xvi; cf. Lloyd-Jones, H., JHS lxxvi (1956) 65Google Scholar. This position has not gone unchallenged: see e.g. Kitto, H. D. F., Poiesis (Berkeley/L.A. 1966) 38–74Google Scholar. The second chapter of Kranz, W., Stasimon (Berlin 1933)Google Scholar, despite the use that Page makes of it, is not tarred with the same brush: at one point (46) Kranz actually makes the slightly exaggerated claim ‘dass alle verschiedenen Erscheinungen des Über - und Unterirdischen in dieser Welt nur Ausdruck ist für eine gerade im Augenblick als wirksam empfundene Macht’.

⁹² Thus, if an expedition avenging a breach of hospitality is said to be sent by Zeus Xenios (Ag. 60 ff.), and the winds that delay it are attributed to Artemis (Ag. 140 ff, 201 f), this does not mean that we should concern ourselves with the theological implications of a breach between Artemis and Zeus (cf. Peradotto, J. J., Phoenix xxiii [1969] 251 f.Google Scholar). Indeed the whole question of Artemis' motive here should not be discussed as though she were a character on stage to whom the dramatist was obliged to ascribe plausible mental processes; she is revealed purely through the phenomena that she causes, like gods in real life. Again, if certain events can be seen from different viewpoints as the work of Apollo or the Furies (p. 28 above), this does not commit Aeschylus to a doctrine that Apollo and the Furies are in alliance.

⁹³ Contrast the Cyclic poets, who, being less interested in human realism, admit the purely miraculous much more freely: see Griffin, J., JHS xcvii (1977) 39–53CrossRefGoogle Scholar.

⁹⁴ It is far from certain, indeed, that the trial was itself traditional: see Jacoby, F., FGrH IIIb (Suppl.) i 22–5, ii 20–9. For Tyndareus and Erigone as prosecutors see ibid. ii 48 n. 8Google Scholar.