¹ Other evidence will be considered in due course. The fragments of the lost play are too few to 'permit a strong metrical argument. Although the rates of resolution in the trimeters are quite 'close, the slightly higher rate for the lost play would be consistent with a later date. Of greater 'interest is the fact that both rates are considerably lower than those of Alcestis (438) and Medea; (431), so that on this basis either Hippolytus play could be dated arbitrarily early. See M. Cropp, and G. Fick, Resolutions and Chronology in Euripides: The Fragmentary Tragedies (BICS supplement 43,1985).

² First prize was also awarded to the posthumous production that included Bacchae and Iphigenia in Aulis: E Ar. Ran. 67 combined with Sud. e 3695 (= TrGF Did. C 22). It is not known which plays were performed on the occasion of Euripides' three other victories.

³ Unless Euripides' two Phrixus plays provide another example, but that remains disputed.

⁴ The priority of the lost Hippolytus is to my knowledge universally accepted. We shall see that 1W. S. Barrett described the evidential value of the hypothesis accurately, but once he accepted the (judgement it offers, he never turned back (Euripides' Hippolytus, edited with introduction and icommentary [Oxford, 1964], p. 29; all references to Barrett are to this work). I know of only one scholar who expresses skepticism concerning the notice: J. Griffin, 'Characterization in Euripides: Hippolytus and Iphigenia in Aulis', in C. Pelling (ed.), Characterization and Individuality in Greek Literature (Oxford, 1990), pp. 128–49, at 130 n. 7. In the end, however, Griffin accepts the standard chronology and merely shifts Euripides' motive for revision to an aesthetic plane. Moreover, he mentions only the order of production; he does not draw the necessary conclusion that a skeptical reading of the notice calls into question the connection of the extant play with the year 428 and thus its status as prize–winner as well. (My thanks to Micala Root for helping me put this matter in focus.)

⁵ T. B. L. Webster's use of this term has been influential, perhaps because his reconstructions of the lost plays, reckless though they often are, are easily accessible in The Tragedies of Euripides (London, 1967). Note, however, that in his scheme both Phaedras are 'bad women'; moreover, his distinction between bad women and unhappy women is not always happily drawn (e.g. Pasiphae in Cretans is an 'unhappy woman', but her famous self–justification was 'one of the great dangerous speeches', pp. 86, 148; Pasiphae's speech may be now consulted in C. Collard et al., Euripides: Selected Fragmentary Plays [Warminster, 1995], vol. 1, pp. 62–7). I use the term to mean female characters in tragedy whose words and behaviour Old Comic poets could exploit, usually with a suggestion of immorality.

⁶ I have also not accepted the burden of refuting every claim that a feature of the surviving play can only be understood as a later variation of something that was in the lost one, though I return to one such consideration at the end of this paper. Because no one has doubted the hypothesis, most arguments along these lines have been made quite casually. The present study should considerably raise the threshold for considering them compelling.

⁷ See Pfeiffer, R., History of Classical Scholarship (Oxford, 1968), vol. 1, pp. 192–6. A noteworthy study was made by G. Zuntz, The Political Plays of Euripides (Manchester, 1955), pp. 129–52, and the remarks of Wilamowitz have been influential (Einleitung in die griechische Tragodie = Euripides' Herakl^s, vol. 1, pp. 146–8, cited from the 1959 Darmstadt reprint). Among commentators, see especially D. L. Page on Euripides' Medea (Oxford, 1938), pp. liii–lv; A. M. Dale on Alcestis (Oxford, 1954), xxxviii–xl; and now N. Dunbar on Aristophanes' Birds (Oxford, 1995), pp. 31–5 and 127.Google Scholar

⁸ The other tragic examples occur in the hypotheses to Euripides' Alcestis, Andromache, and Orestes; cf. Suppliant Women. If the remark TO Spafta TOVTO, Euripides Phoenissae (Cambridge, 1994), p. 168. 2.

⁹ On the fragments, numbered 618–30 in V. Rose's edition (Aristotelis Fragment^ [Leipzig, 1886]; cf. TrGFDID C), see especially E. Reisch in RE 5 (1905), 394–401 (s.v. Didaskaliai); A. C. Pickard–Cambridge, The Dramatic Festivals of Athens, 2nd edn rev. J. Gould and D. M. Lewis (Oxford, 1968), pp. 70–4, 80; and R. Blum, 'Kallimachos und die Literaturverzeichnung bei den Griechen', Archivfiir Geschichte des Buchwesens 18 (1977), 1–360, esp. 50–91 (translated by H. Wellisch as Kallimachos: the Alexandrian Library and the Origins of Bibliography [Madison, 1991]; I cite from the German). A. E. Haigh, The Tragic Drama of the Greeks (Oxford, 1896), pp. 395–402, though cited by O. Taplin as the best work on tragic titles, is misleading on the subject of distinguishing epithets. Against Haigh, Taplin concludes that they were never assigned by the poet and never used in the archives ('The Tide of Prometheus Desmotes', JHS 95 [1975], 184–6, at 185 n. 4).Google Scholar

¹⁰ Barrett saw this (10 n. 1) but did not pursue the implication that Aristophanes of Byzantium was familiar with at least one previous attempt to distinguish the plays. I add that, given the casual manner of citation, it is not likely that it was Aristotle who added the epithet.

¹¹ Fr. 629 (o Se ApiaroTeXrjc ev rate SiSacKaXiaic Svo [sc. poets named Kinesias] iprjcl ycyovivai) may mean only that there were two entries and that they assigned different patronymics or other incompatible characteristics to the poets in question. I see no evidence of similar notes. Fr. 630 (= E Ar. Ran. 404), on joint choregia beginning in 406/5, is not comparable since it was indispensable to explain the form of successive entries.

¹² Thus fr. 623 on Ajax: iv Se. Similarly, Eratosthenes speculated on whether the later of two entries recording a Peace of Aristophanes ofuovvfuoc, i.e. without epithet, referred to a revival or a second play (Arist. fr. 622). The author of the hypothesis to the surviving play thinks this issue resolved by the occurrence of the adjective erepa in a comment by the Pergamene scholar Crates. Clearly, then, the AiSacKaXiai did not even distinguish these plays as e.g. 'first' and 'second'. To take another example, IF Ar. Nub. 552 (= Callimachus, fr. 454 Pf.) shows that no designation irporepai or irpwrai was found in the entry for that play (otherwise Callimachus would not have made the mistake corrected by Eratosthenes). Someone, however, has inserted npSirai in the hypothesis (II in K. J. Dover's edition of Clouds [Oxford, 1968], pp. 1–2), in what otherwise looks exactly like other excerpts derived ultimately from Aristotle's work. (Naturally there can have been no distinction made in the archives since Clouds II was never produced [cf. n. 49 below]. But two texts were in circulation, and if we wish to evaluate the suggestion that Aristotle took an interest in such problems in notes to his AiSaacaXlai, then this example is important as a counter–argument.) We read Trpwroc and Sevrepoc in the papyrus hypotheses (of the 'Tales of Euripides' variety) to Euripides' two Phrixus plays (P. Oxy. 2455 fir. 14 and 17), and similar designations (e.g. A' and B') have been restored by conjecture elsewhere. But even if Dicaearchus of Messene wrote these plot summaries (which I doubt) and distinguished between Hippolytus plays, this still does not take us back to Aristotle, and the impact on the present argument is accordingly nil. On the possibility that the anonymous Life of Euripides used irporepoc of a Hippolytus play, see below.

¹³ On the sources and methods of ancient biography, see Lefkowitz, M. R., The Lives of the Greek Poets (Baltimore, 1981), passim, and for Euripides, D. Kovacs, Euripides: Cyclops, j Alcestis, Medea (vol. 1 of his edition and translation for the Loeb Classical Library j [Harvard, 1994]), pp. 1–36. I emphasize that it is not only its similarity to suspect sources that discredits hyp. Hipp., but this fact combined with the absence of any trace of a trustworthy source or reason to believe that trustworthy sources regularly existed for the kind of information it offers.Google Scholar

¹⁴ My thanks to Albert Henrichs for this suggestion. Like Griffin (n. 4), p. 130 n. 7,1 take the meaning of ifupaiverai here to be an emphatic 'it is clear', like yaiverai and Sourer in notices to, be considered below (nn. 15 and 16). We know of course that what is said to be clear often is or may be disputed.

¹⁵ Because of yap, we must keep TO yap airpeirec KT4. together with iiupalveTat Se vcrepoc i ytypafiiiivoc, but this comment as a whole could have entered the hypothesis tradition at almost any point, as indeed could any of the words from ten Se OSTOC on. Such an insertion has probably been made in the portion of hyp. O.T. headed Sid ri rvpawoc kinyiypa–mai; As W. Luppe shows, a–rravrec in the second sentence and eici Se ot in the third are inconsistent ('Dikaiarchos und der OIAinOYC TYPANNOC, Hermes 119 [1991], 467–9; the rest of Luppe's argument does not concern us here). Professor Luppe informs me per litteras that he in fact believes that the statements ecn Se K–TC. and epxpaivfrai. Se KT4. belong to separate authors. The wording of our notice is closely paralleled by the end of E Eur. Andr. 445, where after comments indicating that two scholars had consulted Aristotle's AiSacKaXiai directly, someone wrote. While Th. Bergk attributed the unhelpful addition to an anonymous imperial exegete, 'ein beschrankter Kopf ('Philologische Paralipomena Theodor Bergks', Hermes 18 [1883], 487–510, at 489–91), P. T. Stevens felt compelled to admit that it may have some foundation (in his commentary on Andromache [Oxford, 1971], p. 16 n. 2). Cf. n. 16.

¹⁶ see n. 14 above. In hyp. Eur. Med., for example, the phrase is (cf. n. 23). In E Eur. Andr. 445 (previous n.).

¹⁷ Sommerstein, A. H., Kleophon and the Restaging of Frogs, in Sommerstein, Tragedy, Comedy and the Polls (Bari, 1993), pp. 461–76, at 462. Sommerstein acknowledges G. Kaibel (ap. Kassel–Austin, PCG III.2, 2, on line 35 of the Life) for the idea of the decree, which is also accepted by K. J. Dover in his commentary on Frogs (Oxford, 1993), pp. 73–4.Google Scholar

¹⁸ Sommerstein also refers to the olive wreath attested in a related contemporary (403/2) context. As for his third argument, mention of the parabasis (the very feature that leads to the conclusion that the two notices derive from a common source), I am less sure that it would not have been invented if, as the opponents of Kaibel's idea propose, Dicaearchus merely had a didascalic entry attesting a second performance and made the rest up.

¹⁹ This is true even if one believes, against the arguments I shall offer below, that the mentions of an unspecified Phaedra in Aristophanes' comedies refer to the lost play.

²⁰ Aristotle does relate that one Hygiaenon challenged Euripides to an antidosis and, in the course of the proceedings, quoted Hipp. 612 against him (Rhet. 1416a28–35; cf. Ar. Thesm. 275; Ran. 101, 1471). Despite what Kovacs ([n. 13], 19) writes, the story does not imply a first prize for the extant Hippolytus. If anything, it suggests failure, though the point should not be pressed. Euripides says Hygiaenon should not bring theatrical decisions into court. If he won with the play his opponent uses to mock him, a more obvious retort would be, 'The Athenians approved of my Hippolytus'. This is not implied by the neutral SfSwKevat Xoyov, which is not even applied specifically to Hippolytus. More generally, whether the incident is historical or not, there is no reason to believe that there ever circulated an account of it with just the characteristics needed to pin down a date for Hippolytus. While Aristotle's use of the verb Karryopiv here (cf. Karrjyopiac agiov in the hypothesis) is not technical (or at most blends technical and general meanings), we do find in Satyrus mention of a prosecution of Euripides by Cleon on a charge of acefieia (fr. 39, col. x). A third–century C.E. list of rhetorical exercises on papyrus is the only other source to mention an impiety trial of Euripides, but as Kovacs notes, this does not imply knowledge of Satyrus' story (Euripidea [Leiden, 1994], pp. 62–3), nor would it make the story any more likely to be true if it did. In sum, to believe in a record of a public event that provided sound evidence for the date of Hippolytus is wishful thinking.

²¹ As Harvey Yunis and Walter Stockert point out to me, either a mere mention of companion plays or arrangement of texts by trilogy (or tetralogy) would have put the matter beyond doubt.

²² Dicaearchus wrote irepi AiowciaKibv aywvwv, reflected in several learned (if not always credible) notices in hypotheses and scholia (frr. 73–89 Wehrli; cf. Pfeiffer [n. 7], 193). Next to nothing is known of the Euripidean researches of other Peripatetics, but they are certainly or probably attested for Theophrastus, Heraclides Ponticus, and Aristoxenus (on whom cf. n. 25). Callimachus wrote, in addition to his comprehensive nivaKec, a TTIVO.% KO.1 avayparTCOV card Xpovovc Kat an (frr. 454–6 Pf.; cf. Blum [n. 7], 198–208). Eratosthenes' research on comedy naturally required occasional work on tragic SiSacKaXteu (e.g. 27 Ran. 1028; cf. Pfeiffer 162). As for Aristophanes of Byzantium, a late source, Choeroboscus ap. Et. Mag. 672.27 (included in Call. fr. 456 Pf.), reports that he wrote his dramatic hypotheses after chancing upon the irlvaicec (sic) of Callimachus. It is unclear to me why anyone would think that Aristophanes did not consult Aristotle's work directly, and the participle and the plural strike me as more than a little suspicious, but Pfeiffer accepts the report (193; cf. his comment on Call. fr. 453 dub.), as do Haigh, The Attic Theatre (Oxford, 1907), pp. 47–8, and tale (n. 7), p. xxxix. It is rejected without argument by W. J. Slater on his frag. spur. 434 (Aristophanis Byzantii Fragmenta [Berlin, 1986], p. 172). Blum (55–6) reasonably suggests that ^ristophanes may sometimes have used Callimachus' –nival; as an index to Aristotle's 'AiSacKaXicu when its different arrangement was better suited to his purposes.

²³ Hyp. Eur. Med. = Arist. fr. 635 Rose = Dicaearchus, fr. 63 Wehrli. Obviously the information in Aristotle's AiSacxaXiat was not sufficient to settle the claim, which the hypothesis introduces with a non–committal. After giving the story careful consideration, Page ([n. 7], xxx–xxxvi) concludes that incorrect attribution of a later play to the Neophron who preceded Euripides ITrGF 15) probably lies behind the story. Malice towards Euripides is naturally not excluded. For different view, see E. Stemplinger, Das Plagiat in der griechischen Literatur (Leipzig, 1912; repr. flildesheim, 1990), pp. 20–1. Nothing is known of the Glaucus who made the similar claim that teschylus' Persians was plagiarized (irapanerroirjcdai) from the Phoenissae of Phrynichus (hyp. ers.).

²⁴ For reasons including independent evidence for Sophocles' generalship, many scholars accept the story about Antigone, and it must at least have been chronologically consistent with the AiSacxaXiai. But it is apparent that it would have been difficult to demonstrate it, and Aristophanes, citing no authority, merely says (pad.

²⁵ Diog. Laert. 3.37 = Aristoxenus, fr. 67 Wehrli = Protagoras, 80 B 5 DK. Aristoxenus wrote at least two books nepi frr. 113–16 Wehrli). His claim (fr. 115 = Vita Soph. 1) that Sophocles' father was a carpenter or smith is unlikely to be well founded. If his name is not merely a mistake in fr. 116 (see Wehrli, ad Ioc, and Radermacher and Dover on Ar. Ran. 1400), he solved a Euripidean zetema by blithely assuming a Siopdatcic of Telephus. (Someone, at any rate, made this suggestion.)

²⁶ For reasons including independent evidence for Sophocles' generalship, many scholars accept the story about Antigone, and it must at least have been chronologically consistent with the AiSacxaXiai. But it is apparent that it would have been difficult to demonstrate it, and Aristophanes, citing no authority, merely says (pad.

²⁷ For simplicity's sake, I now speak only of Aristophanes of Byzantium. The name of any ' earlier scholar could be substituted without altering the conclusions.

²⁸ This explanation strongly implies that Aristophanes (1) consulted two didascalic records of productions that included a Hippolytus and (2) found a victory in 429/8 and a poorer showing in the other year, which was earlier. Although neither point is quite susceptible of proof, both derive from reasonable assumptions and may be conceded here, for the theory of 'correction' remains speculative even in these circumstances. It must also be insisted that the other steps in the theory gain no support from the assumed second didascalic entry.

²⁹ To my knowledge, no one has ever doubted that what the ancient scholar deemed 'unseemly and worthy of accusation' was the character and behaviour of Phaedra (except that Griffin, while conceding that this was the Athenian verdict, attributes a rather different view to Euripides; above, n. 4). It is not a necessary assumption, of course.

³⁰ Consider, for example, what happens when we apply to the production of 438 the assumption that one aspect of one play determined the outcome of the contest. We know from hyp. Ale. that Euripides came in second to Sophocles on this occasion. Aerope of Cretan Women was a 'bad woman'. Was Euripides to conclude that she accounted for his 'failure'? If placing second to Sophocles needs to be explained at all, even our scanty information permits any number of alternative theories that are at least as convincing (Telephus degraded the art and became 'notorious' with the comedians, Alcestis was unconventionally substituted for a satyr play, and while abounding in unfilial, unpaternal, and unheroic sentiments, yet ended happily, etc.).

³¹ Amator. 756BC. My analysis follows Wilalmowitz on Eur. Her. 1263; cf. id., 'Melanippe, SPAW, 63–80 (= KS 1,440–60), at 71 (= 449): the verse Zevc, Adyon was a comic conflation of Euripidean phrases, misidentified by someone as the first line of Melanippe Sophe. When a controversy arose because it was not found there, the story, with its manifest anti–Euripidean slant, was invented. I am not convinced by the arguments of W. Luppe, who tries to salvage the verse as the genuine opening of Melanippe Desmotis ('Plutarch fiber den Anfangsvers der 'Melanippe' des Euripides', WJA 9 [1983], 53–6); see now M. J. Cropp in Collard et al. (n. 5), 266. Strangely, P. Oxy. 2455 fr. 1 (first published in 1962) seems to have had a third version of the opening line. For the whole sequence as Wilamowitz reconstructs it, compare the theory attributed to Aristarchus or Aristoxenus concerning the mystery line Ar. Ran. 1400: Euripides revised Telephus to remove a scene that had given offence (above, n. 25).

³² I print Sauppes correction of the MSS tacitly accepted by Cropp in Collard et al. (n. 5), 266. H. J. Mette emphatically rejects this solution but offers no alternative ('Perithoos–Theseus–Herakles bei Euripides', ZPE 50 [1983] 13–19, at 15 n. 2).

³³ Compare the tales of audience reaction to Euripides' Aeolus (fr. 19; references to fragments of Euripides are to A. Nauck, Tragicorum Graecorum Fragmented [Leipzig, 1889]), Danae (fr. 324), Ixion (p. 490 Nauck), and Auge (Diog. Laert. 2.33; cf. p. 437 Nauck). Two of these come from Plutarch. Aristarchus was willing to consider revision as an explanation of the controversy surrounding Ar. Ran. 1206–8 (said by some to be the opening of Archelaus, but not found there or anywhere in Euripides by Aristarchus), but he himself had no stake in the theory, and it receives no elaboration in our source (i7ad loc).

³⁴ Aristophanes' Frogs itself testifies to Euripides' popularity, if not his success at winning the prize, but during the contest with Aeschylus his confidence, e.g. in his prologues, is always comically misplaced. The biographical tradition makes much of the supposed fact that Euripides was better appreciated by Macedonians and Syracusans than Athenians, and someone in Satyrus explicitly makes him resent his rate of competitive failure (fr. 39, col. xv). A counter–tradition asserts that he was indifferent to his fellow citizens' faulty judgement. See the anonymous Life, 118–20 in the lineation of L. Meridier (Euripide: Tome I [Paris, 1926], pp. 1–5 = p. 3, 18–21 Schwartz {Scholia in Ewipidem [Berlin, 1887–91], vol. 1, pp. 1–6). The point, of course, is that wedo not know how Euripides responded to failure, and neither did the ancient scholars and biographers. I merely register here my opinion (shared with Griffin [n. 4], 130–1) that failure in competition would not have moved him to treat the same story a second time, something he apparently did on no other occasion.

³⁵ Fr. 39, cols, x and xi; two abbreviated versions are found in the anonymous Life, 77–8,100–9 Meridier (= p. 5, 11–12; p. 6, 7–13 Schwartz). Explicit connection of the quoted lines with a pact between Euripides and the women depends on the Life alone, since it falls in the gap between columns in Satyrus. See G. Arrighetti, Satiro: Vita di Euripide (Pisa, 1964), pp. 126–9, and cf. Lefkowitz (n. 13), 33–4, 89–90.

³⁶ A clarification is necessary. Knowing that the story derives ultimately from comedy, we naturally say that it aims to get laughs, not to explain anything. If we had only the anonymous Life, we would assume that some dull–witted epigone misunderstood a reference to the comedy. The larger context in Satyrus, however, makes it virtually certain that the plot of the women was retailed as fact even by some who knew Aristophanes' play: so Wilamowitz (n. 31), 71–2 (= KSI 450–1), approved by Arrighetti (n. 35), 128. But in that case the narrative followed its own logic and pretended to explain something. Of course it was a pretence: the real aim was slander.

³⁷ 68–72 Meridier (= p. 5,4–7 Schwartz).

³⁸ Schwartz is followed by Barrett (26 with n. 2), Kirchhoff by Nauck (p. 491) and Meridier, whose text Kovacs now prints ([n. 20], 6) and approves ([n. 13], 3 and 19).

³⁹ So Wilamowitz (n. 7), 7 n. 12 (and already at Analecta Euripidea [Berlin, 1875], p. 148 n. 3), approved by e.g. Kovacs (n. 13), 19

⁴⁰ Fr. 39 col. xii; cf. Arrighetti (n. 35), 130.

⁴¹ See n. 44 below. Both Barrett (31 n. 3) and Nauck (491) connect the story in the Life with the lost play, though they differ on the restoration of the text (above, n. 38). Nauck also links it to Ar. Ran. 1043 (see below).

⁴² Without questioning that this is the meaning of the hypothesis, some scholars wonder whether Phaedra was really so bad: thus O. Zwierlein, Senecas Phaedra und ihre Vorbilder (Akad. d. Wiss. Mainz, 1987.5), pp. 24–5, noted by Griffin (n. 4), 129 n. 5, and B. Snell, Scenes from Greek Drama (Berkeley, 1964), pp. 33–46, among others. The question is not without interest, for we are forced to balance Aristophanes'judgement (since he could read the entire lost play) against the possibility that his theory derives from disreputable sources. I mention in passing that the epithet of the lost play (KaXvirTo/xevoc or KaraKaXxmTop^voc) can no longer be taken without argument to show that Phaedra's behaviour shocked the Athenians, or even the Alexandrians. According to a once nearly universal view, it referred to a scene in which Hippolytus veiled his head to avoid contact with Phaedra and her shameful proposals (e.g. Barrett 37), but this has now been questioned by W Luppe, 'Die Hypothesis zum ersten Hippolytos', ZPE102 (1994), 23–39, at 29 n. 5, 37–8. (Doubtless Hippolytus disapproved of Phaedra; at issue is the epithet as evidence for a particular and powerful expression of 'shock'.) The strongest element in Luppe's argument is the existence in the papyrus of an aorist participial form of KaX–point of speculation as to the meaning of the epithet, then one no longer has any reason to believe in the scene of Hippolytus veiling himself.

⁴³ Perhaps Agathon portrayed her as well: cf. Thesm. 153, where she is merely the type of the bad (desiring) woman.

⁴⁴ In the extant play, cf. especially Hippolytus' famous diatribe at 616–68; for the lost one, frr. 436,439, and 440 possibly contain 'reproach'; frr. 429 and 430 could imply it.

⁴⁵ E.g. Barrett 30–1, Webster (n. 5), 65, but not Stanford or Dover ad loc. As far as I am aware, this use of nopvat for adulteresses is unique in classical literature. Much later, iropveta becomes 'fornication' and is interchangeable with fioixela in many contexts, a s seen for example in several scholia to the extant Hippolytus. Ar. Ran. 849, explaining the phrase, says that Apollonius asserted a possible reference to Aerope of Cretan Women, (sc. EvpiirtSrjc) nopvevovcav. What Apollonius (of Rhodes?) actually wrote is uncertain, since the active participle reveals the comment in this form to be late. This seems to be, as Webster says, the only other reference to a Euripidean heroine as Tropvrj.

⁴⁶ It is usually assumed that Thesm. 546–7 alludes to t he rhesis in Melanippe Sophe in which Melanippe, who bore twin sons to Poseidon and concealed them in a cattle shed, tried, after they were discovered a n d threatened with death, to defend them without revealing that she was their mother. Aristotle is likewise thought to have had this notoriously clever speech (parodied at Ar. Lys. 1125) in mind when he cited ' t h e speech of Melanippe' as a n example of tragic f)60c that is (Poet. 15.1454a30–l). That is, if these assumptions are correct, Melanippe was thought too clever for a woman, and that is what the woman in Thesmophoriazusae means when she says that Euripides portrayed her as novqpa.

⁴⁷ I hope to demonstrate this point more fully in the course of a reassessment of Aristophanes' use of Euripidean 'bad women'. For the present, a list of passages must suffice: Hipp. 219–22 (Ar. Anagyros F 53 KA), Hipp. 375–6 (Ar. Eg. 1290–4, Ran. 930–2), Hipp. 415–16 (Xenarch. Pentathlos F 4.21–2 KA), Hipp. 345 (Ar. Eg. 18), Hipp. 675–7 (Ar. Thesm. 715–16). P. Rau believes that Ar. Vesp. reflects Hipp. 215–16,219, and 230, along with other Euripidean passages (Paratragodia: Untersuchung einer komischen Form des Aristophanes [Zetemata 45 Munich, 1967], 153–4). Finally, it is often thought that the npoaycoyoi alluded to at Ar. Ran. 1079 include the Nurses of both Stheneboea and the extant Hippolytus (the same pair, then, as at 1043). It is not known (pace Webster [n. 5], 71) whether the Nurse played a comparable role in the lost Hippolytus.

⁴⁸ Barrett 9, 11, and passim. Most often mentioned in this connection is the extant play's lack of the two 'high points' of the (simple) story: Phaedra attempting to seduce Hippolytus, and Phaedra denouncing Hippolytus to Theseus (e.g. Griffin [n. 4], 131–2).

⁴⁹ Dover (n. 12), lxxx–lxxxi, shows that the proper inference from the remarks of Eratosthenes ap. 552 is that Clouds II was not produced in Athens, and Aristophanes may never have completed the work of revision. For explanations along the lines indicated in the text, though with differences of detail, see Dover lxxx–xcviii; T. K. Hubbard, The Mask of Comedy: Aristophanes and the Intertextual Parabasis (Cornell, 1991), pp. 96–106; D. O'Regan, Rhetoric, Comedy, and the Violence of Language in Aristophanes' Clouds (Oxford, 1992), pp. 67–79,133.

⁵⁰ 50 Phoenix, Stheneboea, and Peleus. The statement of Webster (n. 5), 77, that 'all belong t o the first [i.e. Zielinski's "severe"] period on metrical grounds' is roughly borne out by Cropp and Fick (above, n. 1), but it has no bearing on our argument since that period extends down to the surviving Hippolytus. It is not even necessary to conclude that the Hippolytus produced in 428 came last in the series. Aristophanic parodies fix Phoenix before Achamians, Stheneboea before Wasps, and Peleus before Clouds

⁵¹ Cf. above, n. 2. Webster (n. 5), 32, does not hesitate to assign both Phoenix and Cretans with their notorious women to the winning production (Euripides' first) of 441, though that is admittedly nothing but a guess.

⁵² Besides those mentioned already, Michael Halleran and the editors and anonymous referee for CQ made helpful comments on this article. I thank them all, and especially Harvey Yunis.