¹ I am most grateful to Dr James Diggle for valuable criticism.

² The main literature on the hypotheses of Aristophanes of Byzantium is Nauck, A., Aristophanis Byzantini grammatici Alexandrini fragmenta (Halle, 1848)Google Scholar; Schneidewin, F. W., De hypothesibus tragoediarum graecarum Aristophani Byzantio vindicandis (Gottingen, 1853); Raddatz in REix. 1.415–21;Google ScholarT. O. H. Achelis, Philol. 72 (1913), 414–41, 518; 73 (1914), 122–53Google Scholar;D. L. Page, Euripides: Medea (Oxford, 1938)Google Scholar, liii-lv;G. Zuntz, The Political Plays of Euripides (Manchester, 1955), 131, 139–41Google Scholar;Pfeiffer, R., History of Classical Scholarship i (Oxford, 1968), 192–6Google Scholar; Budé, A. W. A. M., De hypotheseis Griekse tragedies en komedies (The Hague, 1977), 33–9.Google Scholar

³ I am following Page in using ‘synopsis’ to mean a summary of a play's plot, with any attendant mythological material (i.e. a in the usual Greek sense), and ‘hypothesis’ to mean a whole preface to a play, which may include a synopsis as well as didascalic material etc.

⁴ Ten out of 14 are so attributed by at least one MS. Exceptions: Phil., Clouds, Lys., Hero.

⁵ An exception is Roster, W. J. W. in Charisteria F. Novotnýoblata (Prague, 1962), 43–50Google Scholar; Budé (n. 2), 40–7, is agnostic. At any rate, if the attribution were right, we should no longer be able to think of Aristophanes as a businesslike or reliable scholar.

⁶ A prose hypothesis to Birds is so attributed in MS U, but this is considered to be a mere slip by Tzetzes (Koster [n. 5], 44).

⁷ For the hypotheses of comedy see L. Radermacher, Aristophanes' ‘Frösche’² (Vienna, 1954), 74–85, as well as Achelis and Budé (n. 2).

⁸ E.g. Eur. Hipp., Andr.; and we shall see that I.T. really belongs in the same category. Editors are not consistent on whether to regard a typically Byzantine synopsis followed by ‘Aristophanic’ material as one hypothesis or two.

⁹ This list corresponds with that of Page (n. 2), liii, except that I have added O.C. and I.T. and omitted Eur. Supp., for which we have a scrap of ‘Aristophanic’ material but no synopsis of any kind. Except for O.C, I am referring to hypotheses printed in the Oxford Classical Texts of the tragedians (for many plays there also exist late hypotheses which these editions ignore), and using the numeration given or implied there.

¹⁰ Achelis 1914 (n. 2), 146, and Zuntz (n. 2), 131, claim that Aristophanic synopses do not normally consist of more than two ‘enuntiata’ or ‘sentences’, but do not explain how they define these terms. Several synopses which they seem to regard as authentic contain more than two points at which a full stop could be placed.

¹¹ The dubious grammar of could easily be remedied, if necessary, by (Wecklein) or

¹² For the use of there are adequate parallels (Kühner-Gerth' ii.201).

¹³ Moulton, J. H., A Grammar of New Testament Greek (Edinburgh, 1908), i³87–90; Mayser, Grammatik der gr. Papyri i2(2).65, ii(2).73–4.Google Scholar

¹⁴ Unless Nicolaus Com. fr. 1.35 is earlier; but it need not be.

¹⁵ might look odd for a sea-battle, but is in fact correct Greek (e.g. Isoc. 5.147).

¹⁶ Achelis 1914 (n. 2), 140, considers that this is not a true synopsis but a relic of a

¹⁷ Wecklein's conjecture, now found in I¹ according to Smith, O. L., Scholia Graeca in Aeschylum ii(2) (Leipzig, 1982), 1. It is of no consequence that M readsGoogle Scholar

¹⁸ The papyrus hypothesis of Laius or Oedipus (TrGF i DID C 4[a]), which duplicates part of hyp. Sept., contains no synopsis.

¹⁹ I might, however, have noted that the worst fault in this synopsis could be alleviated by simply changing to since the passive use of the aorist middle of extends to Attic prose (cf. Barrett on Eur. Hipp. 27).

²⁰ Apart from this hypothesis, TGL s.v. cites Clement of Alexandria, Nicostratus, and Julius Africanus.

²¹ The Antigone³ 3–4.

²² The v.l. would be Attic.

²³ Moore (n. 2), 292–3; Achelis 1913 (n. 2), 434.

²⁴ In fact Diggle, besides changing to inserts the name after perhaps feeling that it should have come from somewhere in this context, or on the analogy of hyp. I. But it would do the writer more credit to suppose that he did not supply the name Glauce at all (cf. n. 27); and the variety of MS readings here could well point to an intrusive gloss (cf. Murray's apparatus).

²⁵ Dr Diggle points out that this use of is affected by the author of the ‘Dicaearchan’ hypotheses, being found at I Rhes. 13 and (in a variant which Dr Diggle favours) at I Phoen. 10 (Murray's line-numbers). recalls in hyp. II Alc, though there may be more emphasis in this case: ‘her own children’.

²⁶ I am prepared to believe, with Diggle and other editors, that the tautologies and are due to textual corruption.

²⁷ ‘Numquam Aristophanes nomen proprium, quod in fabula deest, addidit de suo’, declares Achelis(1914[n. 2], 148 n. 338), to the approval of Zuntz(n. 2), 140 n. 1. These scholars consider that hyp. Rhes. has suffered interpolation or rewriting, but they do not explain how they dispose of Thebans in hyp. Sept., Eumenides in hyp. Eum., Calchas in hyp. Phil., or Glauce in hyp. Med.

²⁸ Cf. Pfeiffer (n. 2), 193.

²⁹ J. Rusten, GRBS 23 (1982), 357–67.