^{page 22 note 1} As one reads these words it is as well to bear in mind that Zielinski had come in for some heavy criticism in Tycho Wilamowitz's chapter on the Trachiniae.

^{page 22 note 2} Sneering at Tycho Wilamowitz is still in fashion: Knox, 's jibe in Harvard Studies in Classical Philology (1961), p. 1CrossRefGoogle Scholar is among the most recent. Before pirouetting through the Ajax, Knox refers to ‘Kitto working on the unassailable basis that Sophocles knew more about dramaturgy than both Schlegels and Tycho von Wilamowitz rolled into one’. One might, however, think that a more relevant question was: ‘do Kitto and Knox rolled into one know more about dramaturgy than both Schlegels and Tycho Wilamowitz?’ My own adherence to Tycho Wilamowitz is based on the principle that where, as in the literary study of Greek Tragedy, die objective truth is unknowable, that theory is best which can most simply account for the maximum number of observed facts without forcing die evidence. That Tycho Wilamowitz was able to evolve a theory of Sophoclean drama without invoking thought patterns and symbolism every step of the way is one of his strongest recommendations.

^{page 23 note 1} From now on I often use the Christian name only: this is pardy the fellow-feeling of one Prussian for another, and partly a convenient way of distinguishing the son from die father, whose theories will also occupy us later in the article.

^{page 24 note 1} Tycho frequently uses the word ‘psychologisch’ with contemptuous overtones: and some critics may object that he begs the question when he rules out possibilities on the ground that they lay ‘ausserhalb des für Sophocles möglichen’. But from time to time he shows himself fully aware that psychology has its place in Sophocles, and in practice his formulae of dismissal were seldom used without justification.

^{page 24 note 2} A revealing comment was made by Wolff, Emily in Eranos (1958), p. 119.Google Scholar She defended the early dating of the play by protesting: ‘But if The Suppliants was written so late in Aeschylus’ career, our traditional view of his artistic development has been completely mistaken, and the standard criteria which have been used in tracing his development will be virtually meaningless.’ We may agree with her, while retaining a belief in the play's later date.

^{page 25 note 1} For example Böhme's law of ‘eine innere Gesetzmässigkeit tragischen Dichtens, dass nämlich die späteren Tragiker nur fortentwickelten was ihnen von ihren Vorgängern überkommen war'.

^{page 25 note 2} And how in any case is one to evaluate a play which contains, according to a combination of known facts and the conjectures of eminent scholars, (a) one dozen small winged chariots, (b) a four-legged bird which lives in a stable, (c) an earthquake, (d) pseudo-geography lessons in which the teacher is alleged to be a dummy and the pupil is a princess widi horns growing out of her head?

^{page 26 note 1} Already stressed by Waldock in the second chapter of his book on Sophocles.

^{page 27 note 1} But not, so far as I know, of the Zielinski variety. If anyone had maintained that Atossa and Xerxes were reunited in a scene of Phrynichus' Persians, a play to which Aeschylus makes allusion in the opening lines of his own, he would be voicing an improvable hypothesis, but one which, if true, would mitigate some of the difficuties with which we are faced. Atossa's fruitless obsession with clothes for Xerxes may have more to it than meets the eye.

^{page 27 note 2} The text is not certain, but this is as good as anything else.

^{page 27 note 3} Indeed one can go further: if Aeschylus had wanted to make the sentence refer to something not considered likely to happen, why did he abjure the use of si followed by an optative?

^{page 28 note 1} Sophoclean—though not of the specifically Tycho-Sophocles type—is the use of the long and tranquil choral ode which follows the Queen's words at 852ff. This stands in the greatest possible contrast to the excited lyric exchange which follows. This is not normal Aeschylean practice: I emphasize this only to illustrate the fact that the mere date of the Persians does not preclude it from the use of methods which we are perhaps accustomed to think of as more advanced. If the widely held opinion is right that the fragment of the waking-song after the wedding night in the Danaids belongs immediately before the discovery of the suitors' deaths, we would have an almost violent example of the same technique.

^{page 31 note 1} Already therefore in the Persians we can see an early example of the ‘Zerlegung einer Handlung in zwei Szenen’ which Tycho Wilamowitz established as a feature of Sophoclean technique. However, one could not properly claim this as a close link between Aeschylus and Sophocles, as opposed to most other Greek authors: Geffcken demonstrated in Hermes 1927 how long the ‘Zerteilung des Berichtes’ principle operated in Greek literature. He has, however, little to say on Aeschylus, and the examples which he chooses from diis author are by no means the most conspicuous.

^{page 32 note 1} The change in his character is not however total, for as Patzer, (in Harvard Studies in Classical Philology, 1958)Google Scholar has pointed out, Eteokles is still the ‘ideal commander’ in the second half of the play as he is in the first. But this is the only element in his character which remains constant. Patzer's article, and Wolff's which accompanies it, stimulating though they both are, need gentle correction in places, and this has been supplied by Lesky, in Wiener Studien (1961).Google Scholar More recently still Kurt von Fritz has published a study on the Seven in his Antike und Moderne Tragödie, in which he comments at some length on problems raised by Wolff and Patzer, and makes a fresh attempt to prove that the play is a unity. This attempt depends largely on positing not a real Opfertod, but a virtual one (p. 212). I discuss this further in the note on p. 41.

^{page 32 note 2} The Seven has been compared with inexplicable frequency to the Oedipus Rex, and commentators profess to find much in common between the heroes of the two plays (and indeed between them and certain fifth-century politicians). To my own way of thinking, the plays could hardly be more dissimilar, but the comparison can be made to yield one interesting result: Eteokles is (a) commander of the city, (b) son of Oedipus. He fulfils these roles in the play not simultaneously but one after the other. In the O.R. Oedipus is (a) the murderer of the last king of Thebes, (b) parricide-and-husband-of-his-own-mother. Sophocles still, from time to time, finds it necessary or desirable to treat his Oedipus under one aspect or the other, but in general he is able to keep both themes running at the same time.

^{page 32 note 3} Scholars intent on proving the unity of the Seven naturally fasten on to v. 70 as proving that Eteokles fulfills a dual role throughout the play. In a case like this everything depends on quantity. If Aeschylus wished to stress the Doppelheit surely he would have done more than allude to the fact in one line out of 653? Contrast the effort which Sophocles makes when he wishes to do this kind ofthing: ‘Besonders charakteristisch für the Technik des Sophokles scheint mir dann noch die Art zu sein, wie er das den zweiten Teil vorbereitende Hilfsmotiv in dem ersten eingeflochten hat’ (Tycho, p. 37). In a case like this it is always difficult to know what weight we dare allow to what may have been said in an earlier play of the trilology. v. 70 may, as Mr D. W. Lucas points out to me, simply be a reminder to the audience of the theme dealt with in the immediately preceding Oedipus. It remains none the less instructive to see how Aeschylus manages his themes within the limits of the single play. Already in 1848 Schneidewin was writing an article in Philologus in which he laid considerable weight on the duality inherent in the theme of the play and in the presentation of Eteokles.

^{page 33 note 1} Notice how Aeschylus has virtually anticipated Euripides' implied criticism by including in the messenger's first words (vv. 378–9) the sentence

^{page 34 note 1} Kurt von Fritz (pp. 203–4) seeks to explain καί δή πέπεμπται as instantaneous, on the same lines as Eum. 894 καὶ δή δέδεγμαι. The two uses of καί δή are well established and quite distinct: ‘Suppose I do accept’ is the sense at Eum. 894: there is nothing hypothetical about Sept. 472 (see Denniston, , G.P. pp. 249 and 253Google Scholar). Nor can I agree that ‘Der Aorist ήιρέθη in v. 505 macht ohnehin keine Schwierigkeit’, even though Lloyd-Jones, , Gnomon (1962), p. 741Google Scholar finds that one aorist and two perfects ‘are most naturally taken as referring to the immediate future’.

^{page 34 note 2} Or better, as Weil.

^{page 36 note 1} This point is properly emphasized by Wolff on p. 92 of his article. It applies also to another suggestion which has been made to me, namely that Eteokles could already have chosen six Theban champions to join with him in the city's defence, but that who should go to which gate was still left undetermined. Then as the main messenger describes the opponents, Eteokles could send off a series of minor messengers to despatch each champion to the appropriate gate. This would account for the mixture of tenses. But, as we have said, the process, which seems unnecessarily complicated, would have to be explained to the audience. There is no mention of any such series of messengers, and even if this hypothesis were correct, it would still not correspond with the exact wording of v. 284.

^{page 36 note 2} Lesky, , Wiener Studien (1961), pp. 8–9.Google Scholar

^{page 37 note 1} One demonstration in miniature of how Aeschylus manages to suggest that the defence of the city depends on Eteokles’ own prudent judgement as each new problem is presented to him, and at the same time that every move he makes is divinely pre-ordained, can be found in the same section of the same speech, at vv. 407–15. In v. 408 Eteokles says ‘I shall send the son of Astakus’, and at v. 415 ‘Justice is sending him forth’.

^{page 37 note 2} Similar to those of Solmsen, in T.A.P.A. (1937).Google Scholar

^{page 38 note 1} Kurt von Fritz, p. 211, points out how Euripides in the Phoenissae uses an Opfertod theme, applying it to Menoikeus, not Eteokles, because in Euripides’ treatment of the legend Eteokles is required to appear in a less favourable light. The suggestion that Menoikeus’ Opfertod in the Phoenissae was a development from one by Eteokles in the Seven—‘Es ist fast als ob Euripides the Auffassung der modernen Interpreten von dem Opfertod des Eteokles bei Aeschylus anticipiert hätte'—is too interesting to pass over in silence. The theme is, however, one dear to the heart of Euripides and one dares not make too much out of this.

^{page 38 note 2} v. 73 is doggerel, and I shall be giving fuller reasons for its deletion elsewhere. But for the moment those who share the interpolator's twin delusions that Theban houses were built on top of fireplaces, and that Argives did not speak Greek, should note that these curiosities of archaeology and linguistics do not actually aifect the present Opfertod arguments.

^{page 38 note 3} Though once again we must remember that we do not know what was said in the Oedipus: it may be—and this applies also to my third point—that the curse doomed the children and the city (cf. 764–5), and that Eteokles is simply saying ‘take my life, but spare the city’. In either case Eteokles' attitude is the same. We have no means of knowing whether Eteokles' death was the actual cause of the city's survival: but we do know that Eteokles was willing to die in the hope of the city's survival. This may not be an Opfertod in the strictest sense, but the difference, to borrow an expression of Professor Page's, ‘could scarcely be less without ceasing to exist’.

^{page 39 note 1} See also p. 42 below.

^{page 40 note 1} Aeschylus is much more definite than Solmsen, (T.A.P.A. 1937, p. 199)Google Scholar who writes: ‘The chorus tells us that by casting aside Apollo's warning and begetting a son Laios brought the curse upon himself and his descendants and possibly also Thebes.’

^{page 40 note 2} The argument to the Seven assumes the existence of this myth, as do the almost identical hexameters contained in the hypotheses to the Oed. Tyr. and the Phoenissae. The scholia on the Phoenissae, in particular the scholion on v. 1760, quoting Peisander, mention it several times, and other references occur in Apollodorus and Hyginus. But although the rape of Chrysippus is depicted on late fifth-century vases, our earliest literary evidence is Praxilla of Sikyon (no. 751 in Page's Poetae Melici Graeci) who ascribes the crime not to Laius but Zeus—though needless to say ύπὸ Διός has been emended to ύπ' Оіδίποδος, by Valckenaer, not that this is of much help. Welcker and G. Hermann assumed that Aeschylus knew the story, but Schneidewin, (Phihlogus, 1848, p. 351)Google Scholar and Robert are unwilling to believe in its existence before Euripides. The question is a complicated one, and is excellently treated by Lamer, in R.E. XII, I, pp. 476ff.Google Scholar, who, arguing against Robert, maintains that the Chrysippus story was associated with the Oedipus saga already in the ancient Oedipus epic. The one secure fact is that no other crime of Laius was known. It is not likely, but at the same time it is not impossible, that the παλαίφατοι άραί of v. 766 refer to the curse of Pelops, since an oracle, however menacing, is not quite the same as a curse. If this were so, the existence of the Chrysippus story for Aeschylus would be assured.

^{page 41 note 1} ProfessorManton, in the Institute of Classical Studies Bulletin, no. 8 (1961), p. 80.Google Scholar I use this quotation because it comes from a neutral source. Professor Manton's words do not in themselves make it clear whether or not he assents to the Opfertod theory.

A good illustration of how difficult it is to formulate precise views on Aeschylus can be found in the Schmid-Stählin discussion of this play. In their text they come out in favour of an Opfertod, and then in their footnote write: ‘Die Problemstellung war schon im Laios (Sept. 749) gegeben—was der Grossvater versäumt hat, leistet der Enkel, ohne sich persönlich vor den Wirkungen der Erbschuld retten zu können. Das ist grossartige Tragik.’ By denying Eteokles the theoretical possibility of not dying—a possibility which the chorus 686 ff. regard as quite open—they are in fact denying the very existence of an Opfertod. Nor does it need to be stressed that the Tragik would lose greatly in effect if Eteokles is treated simply as a plaything of the family curse, especially since it is nowhere suggested that Eteokles is himself guilty of any crime. As Lesky in his article properly insists, there is, as more than once in the Oresteia, a combination of free will and predestination which cannot be unravelled.

Von Fritz evolves a theory of a virtual Opfertod, which, he believes, unites the two halves of the play. ‘Die Situation aber hat sich so entwickelt, dass er dann dem Bruder im Kampf auf Leben und Tod gegenübertreten muss. Dass er, wenn er in diesem Kampf am Leben bleiben sollte, nicht fröhlich weiterleben geschweige denn weiterherrschen kann, ist auch deutlich genug. Damit ist auch sein eigener Tod in irgend einer Form gewiss. Das ist es, was Eteokles, als er sich nach dem Bericht des Spähers über das Verhalten seines Bruders als den “geeigneten” Verteidiger diesem am siebten Tore gegenüberstellt, vor Augen hat. Da bleibt ihm freilich kein freier Entschluss in dem Sinne, dass er den “freien Opfertod” für die Stadt wählen oder nicht wählen kann. Aber er ist doch “frei” in dem Sinne, dass er deshalb nicht anders handeln kann, als er handelt, weil er der ist, der er ist….’

So far from agreeing that all this is ‘deutlich genug’, I can find no trace in Aeschylus’ play of any such consideration. Nowhere does Eteokles (or the chorus) even raise the question of what would happen if Eteokles fought the duel but survived, and to imply that Eteokles makes his choice with this in the forefront of his mind is an interpretation which has no authority from Aeschylus, vv. 690f., 703 f., 732 f., and especially 734ff., make it clear that the only real alternatives are a duel with the death of both brothers, or, as the chorus urge (677 ff.), no duel at all. Defining a third possibility is merely an irrelevant intellectual exercise, von Fritz's contention that a genuine Opfertod is disproved by Eteokles' promise (271 ff.) to dedicate offerings to the gods in the event of his victory (not much of a victory, on von Fritz's argument!) is invalid. Until the Sieben Redepaare are over, Eteokles cannot definitely know that an Opfertod will be required of him. The sequence 271 ff. is a good illustration of the way in which the Seven divides. 271 ff. make excellent sense in the mouth of the ‘ideal commander’, but nonsense coming from the mouth of the doomed son of Oedipus, the role which he only fully assumes from v. 653 onward. von Fritz's description of Eteokles as the most ‘individualisiert’ character in the whole of Aeschylus (contrast Howald, who found that Eteokles was the embodiment of three quite separate personalities) is the natural counterpart of his other arguments. It is an extreme statement of the ‘Eteokles is the first tragic hero’ commonplace. Like many commonplaces and most extreme statements it is a view open to the strongest objections.

^{page 42 note 1} There is no need to refer v. 843 to apprehensions about the future: all the rest of the strophe and antistrophe refer to the brothers, and the news that there is lamentation for them in the city is (a) unsurprising, and (b) confirmed by vv. 900 ff. (άμφί πτόλιν = ‘in der Stadt umher’, to borrow K.–G.'s translation of Λ 706). There remains only the difficulty of vv. 902 f. At least two explanations are possible: (1) The meaning is posterity at large (cf. Homer's use of έπιγιγνόμενοι); (2) it is simply a momentary lapse, or an attempt to heighten the effect of irony, in eimer case a reminiscence to be explained on a Zielinski-type hypothesis. There is even a third explanation, Schneidewin's: ‘Das wahre ist, dass der chor mit bitterster ironie, wie in diesem päan des Hades oft, sagt: nun wird die habe — um welche sich der ganze fluch und Untergang der brüder drehte — bleiben für die nachkommen, nämlich die nicht vorhandenen.’ This explanation, absurd though it may seem at first sight, would not be discordant with what, for want of a better expression, we may call the ‘grim humour’ which Aeschylus permits himself from time to time.

Lloyd-Jones, , C.Q. (1959), p. 91Google Scholar, unwilling to believe that the poet carelessly combined incompatibles in a way scarcely paralleled in tragedy’ (sic), revives the old suggestion that ἀτέκνους is corrupt: as already Ahrens, , de causis Aesch. nondum satis emendati, p. 33Google Scholar, Hartung, Wecklein. At the very least ‘I think it not only possible but probable that ἀτέκνους does not mean what it is generally supposed to’. On the next page he writes: ‘Would the single word ἀτέκνους at l. 828 have made it clear to the audience that for the time being they were to forget about the Epigoni?’ But it is not a single word: in the part of the play which Lloyd-Jones wishes us to think is unjusdy suspected we read not only τετραμμένου παντρόπωι φυγᾶι γένους (954–5) but even clearer still, at vv. 1054–6,

Lloyd-Jones's article takes as its point of departure the ‘fact’ that the end of the Seven was only suspected after the discovery in 1848 that it was the last play of die trilogy. The truth is that already in 1848, after the discovery, Schneidewin was defending the end of the Seven against the doubts which had already been expressed before the discovery. ‘Und schon früher, ehe man die trilogische composition berücksichtigte, hat man anstoss genommen und den schluss, so gut es gehen wollte, zu rechtfertigen versucht, indem man wunderte, warum der Dichter nicht mit v. 995 das stück geschlossen habe’, Philologus (1848), p. 362. It is a pity that Lloyd-Jones did not include Schneidewin's article in the bibliography which accompanies his own.

^{page 44 note 1} Professor Fraenkel is of course, by comparison with many other scholars, commendably abstemious in the matter of psychological explanations. But it might help to illustrate my own prejudices if I take advantage of this point in the discussion to quote Tycho's words (p. 31) on Antigone's second burial of Polyneices. ‘Es ist natürlich leicht, sich das psychologisch irgendwie zurecht zu legen, aber nicht das ist die Aufgabe, sich auszudenken, warum unter den gegebenen Umständen ein Mensch etwa so handeln könnte — gerade als ob es sich um reale Verhältnisse handelte — sondern man hat allein zu fragen, ob und wie innerhalb des Stückes die Handlung vom Dichter motiviert ist.’ Again, when discussing the Elektra (p. 225) Tycho makes a similar observation, that the understanding of almost every scene in the play has been prejudiced by attempts to provide the events on stage with ‘einem den Gesetzen der Wirklichkeit und der Psychologie entsprechenden Unterbau’.

^{page 45 note 1} For example, Professor Winnington-Ingram, , C.Q. (1954), P. 30Google Scholar, concludes that Aeschylus, in order to meet a technical difficulty, has made brilliant use of a characterization of the Chorus already established to serve more fundamental dramatic purposes’. The chorus, we are told, are consistent throughout the play, one of their most important characteristics being evasiveness. Winnington-Ingram can even detect two separate degrees of evasiveness. One must hope that Aeschylus could count on Dikaiopolis in his audience being equally astute. When on p. 26 one reads that ‘in the Agamemnon it is the function of the Chorus (though not of the Chorus alone) to attempt, often desperately, to preserve a façade of good before the reality of evil’, one is almost paralysed before the choice of refuting evidence. What is especially evasive or façade-preserving about vv. 445–70, 975–1000 (in particular 990ff.) or 1017–33?

^{page 46 note 1} a description first used at v. 855, comes again, this time charged with irony, in the Klytaimestra speech which immediately follows their inglorious scene of indecision.

^{page 46 note 2} ‘Eine grössere Gegensätzlichkeit der Auffassung zwischen solchen Kennern der griechischen Tragödie lässt sich kaum denken.’ Gundert, Hermann, −ΕωΡΙΑ, Festschrift für W.-H. Schuchhardt, P. 73.Google Scholar

^{page 48 note 1} This kind of treatment suggests that Howald, , Griechische Tragödie, p. 89Google Scholar, may be nearer the truth than either Fraenkel or Denniston–Page when he speaks of an Agamemnon ‘der nichts als wahrhaft königliches Format repräsentieren soll’.

^{page 48 note 2} And, much more important, neither does the chorus. The carpet scene really requires an article to itself, but in this footnote I would simply like to challenge the assumption diat when Agamemnon treads the purple he is committing an act of extreme hybris. This assumption is universal, and on it depends not only the interpretation of Agamemnon's character, but to a considerable extent the interpretation of the whole play. Other considerations apart, the assumption seems to me disproved by two facts: the first that v. 933 ‘would you have vowed to the gods to do this act, in a moment of fear?’ is an idiotic question if ‘this act’ is one of extreme impiety: whoever prayed to the gods that, if rescued from danger, he would undertake to commit the most impious act he could think of?

The second reason lies in the choral comment on the scene. Elsewhere in Aeschylus, for example after the first news of the Persian disaster or after Eteokles’ decision to go out to fight his brother or after the successful completion of the revenge in the Choephoroe, the chorus comment directly on any critical act or decision. As Agamemnon goes into the house, treading on the purple, the chorus do not comment that he has been overcome by Ate: the general tenor of their remarks is that although things seem to be all right, they cannot rid themselves of a rather unpleasant feeling that things are not quite as they appear. Such a comment would be ludicrous if Agamemnon had committed an act of the most blatant impiety. Would it really be so remarkable if the most famous commander in Greek legend allowed himself, after his most famous exploit, to walk on a rather expensive piece of material? Let us not pretend that only the most sensuous oriental monarchs would do such a thing: our own Royal Family does it every day. The fact that Agamemnon himself demurs at the idea is, in my opinion, proof that he is a conventionally pious person, as unlike Klytaimestra, his opponent, as possible. In the carpet scene Klytaimestra defeats Agamemnon in the clash of minds: the fact that the minds are clashing over a carpet seems to me only of secondary importance (as John Jones has pointed out, they argue as much about the material extravagance of wasting a fine carpet as on the supposed impiety of treading on it). The poet wished to show King and Queen in conflict at a particular point in the play: in taking advantage of the physical requirements of the action, and elevating the question of the King's progress into the palace into a clash of wills, Aeschylus has shown one more sign of the accomplished skill which sets the Oresteia so far ahead of his earlier experiments in drama. That he did not wish to depict blind and insensate hybris on the part of the King seems confirmed by the fact that Agamemnon removes his shoes before walking on the purple. The prodigies of insight evidenced by commentators who try to reconcile this inconvenient fact with their view of the scene as a whole have little but entertainment value.

^{page 50 note 1} One person who would not recoil in horror is a distinguished member of the Music Faculty of this University, who retains a lively interest in the classics. In the course of a talk on a quite different topic, he referred to the ‘messenger's beacon speech in the Agamemnon’. His mind had seized on the really important fact, that this is properly a messenger speech. That it is delivered through the mask of Klytaimestra is largely a matter of dramatic convenience: we do not want two messengers scuttling on and off before the King's herald arrives.

^{page 51 note 1} It may be that even his claims (especially v. 1604) are meant to be taken quite seriously now that Klytaimestra has lost some of her nerve and the play has recently been dealing with Agamemnon's murder as but one more link, in the chain of death.

^{page 51 note 2} Perhaps if Aeschylus had not clung so tenaciously in the Agamemnon to the old technique whereby the actors tend to address the chorus rather than each other, this impression would not be so strong.

^{page 52 note 1} McDonald, W. A., Classical Journal (1960), p. 367Google Scholar, finds that “some sixteen editors favour Electra and approximately twenty-four champion Clytaemnestra’. McDonald establishes that the ascription of the speech to Electra was not first made by Turnèbe (1553) but appeared already in the Aldine edition of 1518. It is not clear who first gave the lines to Klytaimestra. ‘Porrus’ is usually given the credit for this, but although Franciscus Portus was an expert on Sophocles, and his son Aemilius on Euripides, neither of them appears to have edited Aeschylus, so if either of them did make the ascription, it must have been à propos of some other passage. The rarity of their books makes it difficult to check on this.

^{page 52 note 2} So thought Tycho's father: ‘Wer das verkennt, hat von der Tragik des Aischylos nichts begriffen’ (Hermes, 1927, p. 286). The inconsistency is less extreme for those who, like Fraenkel but unlike Page, believe that by the end of the Agamemnon Klytaimestra has already lost some of her iron nerve.

^{page 53 note 1} Of those who assign the lines to Klytaimestra, ‘practically all feel that the lament is ironical’, says McDonald, loc. cit. McDonald's own contention that the lines are simultaneously ironical and sincere is so fine and subtle that it passes without trace through the cruder mesh of my own thought processes.

^{page 53 note 2} Soph. El. 770.

^{page 54 note 1} Like ProfessorWinnington-Ingram, , C.Q. (1946), p. 59Google Scholar, who finds ‘subtle and intricate irony’ in this passage, because Electra thinks she is making up something entirely false, whereas ‘in fact, what she says is in every particular only too true’.

^{page 54 note 2} Contrast Soph. El. 1296–1300: though note that even here Tychoismus operates. We are led to expect that Electra will simulate grief before her mother, but in the event she only does so before Aigisthos. As devil's advocate I might perhaps say that the strongest argument in favour of ascribing Cho. 691 ff. to Electra is precisely the fact that since in almost every other particular the actual events fail to correspond with the proposed plan, Electra's failure to adhere to her instructions would be almost the most natural thing. Soph. El. 1293–1300 would then merely outline the fresh use of an old idea already found in the Choephoroe. One need not take very seriously the alleged mechanical difficulty of distributing the parts between the actors. McDonald's contention that the deuteragonist played both Electra and Klytaimestra, and that this one speech of Electra's would have to be delivered by the tritagonist, since both Electra and Klytaimestra would be on stage together, is based on a distribution of the roles which cannot be proved to be correct. In any case such an argument would be especially weak in a play which contains the single three-line speech of Pylades.

^{page 56 note 1} Notice how in the plan Klytaimestra is pushed so far into the background that she appears only as one of the two unexpressed subjects to the verb and participles in vv. 556–8: a fact not perhaps sufficiently appreciated by those who like to contrast the Choephoroe with Sophocles’ Electra.

^{page 56 note 2} The last half of v. 582 appears almost to contradict v. 581. Dindorf deleted v. 582, which Aulus Gellius quotes (with participles for infinitives) as coming from the Prometheus Πυρφόρος;. At Sept. 619 (with ή not καί) the tone is ‘if he speaks at all, it is to the point’. I interpret v. 582 to be an example of polar expression with the opposite emphasis, λέγειυ τά καίρια virtually corresponding to γλῶσσαν εὔφημον φέρειν. The use of καί not ἤ may support this. In any case the expression is clearly a conventional formula, and therefore not an adequate preparation for the chorus’ memorable intervention. Cf. Eur. fr. 413 and the other references there cited by Nauck.

^{page 56 note 3} Tragedy, even when representing Persian Queens or Phrygian slaves, appears never to make the same linguistic concessions as Comedy.

^{page 59 note 1} Having assailed the views of practically every other Aeschylean scholar, I felt it would be invidious to let these two distinguished names pass without mention.

^{page 61 note 1} If ever there was an ‘exotic and irrational entertainment’ it was Greek Tragedy, that curious amalgam compounded from diverse elements, some known and some unknown, in proportions that remain a matter for conjecture. The critics often remind us of its affinities with opera. Have they ever considered how absurd the techniques of modern Aeschylean criticism would appear if applied to, let us say, Don Giovanni?

Ignoring the Verschmelzungsprozess whereby the Masetto-Zerlina theme is inexperdy woven into the Commendatore-vengeance theme, we may begin with the Don's own character. Tradition gives us the outline, and tradition is instantly confirmed at the beginning of the opera when the Don is trying to rid himself of Donna Elvira. But thereafter he fails to seduce Donna Elvira's maid, and twice fails with Zerlina. (See Zielinski's chapters ‘de conduplicatione’ and ‘de inrito Consilio’.) There must have been a different tradition which depicted the Don as a sort of Margites. Donna Elvira (a transparent duplicate for Donna Anna) is herself the product of a dual tradition, being treated both as a figure of fun and as a tragic heroine. The plot itself is quite inadequately motivated; people come and go entirely at the author's whim, offering excuses which are entirely unsatisfactory—or more often offering no excuses at all. What are we to make of the crux involving Donna Anna? She sings an aria straight at Don Giovanni without misgiving, but at the end of the scene, when he is on the point of leaving the stage, she recognizes him as the murderer of her father. Deletion or ‘transposition are the only admissible solutions. It is clear also that one or other of the two final scenes must be an interpolation. The statue scene has no place in an Opera Buffa: but those scholars must be right who believe that this classification originally belonged to a different Don Giovanni, and choose therefore to delete the last scene of all, the Don's descent to hell being obviously the τέλος for which the piece was designed.

As opera-goers will know, that last remark was too near the truth to be funny, and I say no more on this score. But it is worth commenting that the inordinate number of inconsistencies in Don Giovanni should be considered in relation to the nature of the opera, filled with scenes both courtly and rustic, where we move from murder to expressions of love, and from ballrooms to cemeteries, with unexampled rapidity. If the reader remembers what was written earlier on the episodic nature of the Agamemnon, he may begin to wonder if, in art forms with conventions as strange and rigorous as those of Greek Tragedy and opera, inconsistency is not the natural partner of excitement in a piece where the author has aimed at providing his audience with a wide variety of swiftly moving scenes. As I say in the next page, Homeric epic is another obvious example.

^{page 62 note 1} An early draft of this article was seen by Professor D. L. Page, in whom it excited the emotions of pity and terror to such effect that I felt obliged to rewrite the whole thing from start to finish. For the criticisms which he made——I am more than ordinarily thankful. The second version was very kindly scrutinized by Mr D. W. Lucas, who made numerous suggestions, all of which I have gratefully adopted.