² e.g. Wilamowitz-Moellendorff, U. v., Griechische Verskunst (Berlin, 1921Google Scholar, repr. Darmstadt, 1962), Schroeder, O., Aeschyli cantica (Leipzig, 1916 ²)Google Scholar, Sophoclis cantica (Leipzig, 1923 ²)Google Scholar, Euripidis cantica (Leipzig, 1928 ²)Google Scholar, P. Maas, Greek Metre (transl. by H. Lloyd-Jones, Oxford, 1962)Google Scholar, Snell, B., Griechische Metrik (Gottingen, 1962 ³)Google Scholar, Dain, A., Traite de metrique grecque (Paris, 1965)Google Scholar, , A. M.Dale, The Lyric Metres of Greek Drama (Cambridge, 1968 ²)Google Scholar, Raven, D. S., Greek Metre: an Introduction (London, 1968 ²)Google Scholar, Korzeniewski, D., Griechische Metrik (Darmstadt, 1968)Google Scholar. Though I am indebted to these scholars, I do not refer to them nor to the differences between their interpretations except in some special cases.

³ Maas describes the (Lesbian) glyconic as . He uses the symbol to note anaclasis. Therefore, when the form is given to the glyconic in tragedy, is excluded by definition. Snell, however, uses the symbol oo a little more loosely. It stands for ‘2 ancipitia, wo selten Doppelkurze erscheint’. When I employ the symbol for ‘aeolic base’ in this paper, it covers and and for convenience) but not However, I hesitate to take it as anaclasis. For the question whether ‘aeolic base’ should be of the same nature as other cases of anaclasis cited by Maas, see p. 80 below.

⁴ Dale takes E. IT 1126/1141, , as an example of contraction of glyconic (or wilamowitzian; for this nomenclature, see n. 5 below). But both 1126 and 1141 are easily changed into normal wilamowitzian by transposingκάλαμος and πτέρʋуας to the end of the verse (as Diggle does in the new OCT).

⁵ So-called ‘choriambic dimeter’. I tried to show the inappropriateness of this name in CQ n.s. 32 (1982), 59–74CrossRefGoogle Scholar. ‘Wilamowitzianus’ is Maas's nomenclature, here anglicized.

⁶ But is extremely rare even in Lesbian metres. As to glyconics, is found only at Sappho 94 LP 22, ?96. 4,98(a)8, ?(b)9. Also, as Maas notices (§33. 3), is much rarer than is imagined. Page, Cf. D., Sappho and Alcaeus (Oxford, 1955), pp. 80–1Google Scholar.

⁷ In MSS readings, corresponds with at E. Ba. 404/419 and 406/421. They can easily be emended as in Murray's text: 404 ĭʋα: ĭνʹ οί Heath, 421 ĭσα: ĭσαν 1(= Triclinius (?)).

⁸ If there were an example of antistrophic responsion between and with unambiguous demonstration of colon-beginning, this colon would be identified beyond doubt as a different colon from glyconic. Dale mentions E. Hyps, fr 1 ii 23/iii 26 (LM ² 134 n. I), but the colometry represented in the papyrus is not reliable. See p. 69 below.

⁹ S.Aj. 1190/1197 and Ant. 106/123 are included. Once the responsion between gl and wil is accepted in pre-Euripidean work, ἂν (Wilamowitz) and πφφΑρуόθєν (Erfurdt) seem to me the easiest solution. However, lines with textual corruption in the part occupying aeolic base are, as a rule, excluded from my figures.

¹⁰ (cr+gl), at Med. 155f./18Of., is not included. For this compound, see p. 79 below.

¹¹ The figure will be different to some extent according to which definition is given to synaphea, especially to ‘word-group’. This question is important, but it is dispensable for our current purpose.

¹² Tragic ‘ibycean’ is used after ‘enoplian’ or ‘hemiepes pendant’:

, E.Andr. 827, HF 1030Google Scholar;

, E.HF 1033, Or. 1257/1277Google Scholar;

, E.Hec. 1069Google Scholar (κσαι’ κσαιο τνφλν, Αλιε φγγος παλλας).

E. El. 701/715;

, E.Tro. 258Google Scholar.

Other examples are: A. Th. 222/229 (this is the unique case of antistrophic responsion of with – at the penultimate element), E. Ale. 244/248 (); , S.OC 239, 1245Google Scholar, , E.Or. 1381Google Scholar (). For A. Cho. 315/332, , E.El. 151, 155Google Scholar, see below. ? Tro. 248 (ἒννєπє τλάμοναΚασάνδραν) ?Tro. 269 (ἇρά μοɩ λύσσοɩ) ? HF 1205 (πέπλον πόδɩ)κє, ῥθος єλίῳδєîξον). Also cf. . , E.Andr. 831Google Scholar (?/827 above), , E.Tro. 267Google Scholar; E. Ion 685/704, E. Hec. 647, 649, HF. 1184, 1186, 1188, Ion 717, 1487, Pho. 121.

¹³ S. OC 1564/1575, E. Ale. 594/602, Andr. 841.

¹⁴ E. Andr. 296/304, 298/306, Hel. 640, Hyps, fr 64 77, 81.

¹⁵ , E.El. 586, 588Google Scholar, Ion 1448, 1479, 1482, 1484, 1486, HF 1196, Hel. 657, 680, 681, Rh. 459/825, Phaeth. 276, Hyps, fr 64 94, S. Tra. 647/655.

¹⁶ My interpretation of these verses is new. A full treatment of so-called ‘prosodiac-enoplian’ is given in Part II of my thesis (see n. 1). I am preparing to publish this part. Here, however, without exact definition of ‘ibycean’, the current argument about distinction of glyconic with the base from ‘ibycean’ is valid.

¹⁷ Ba. 410/425 Πɩєρία μούσєɩος ἕδρα κατ ϕάος τє ϕίλας IA 553/568 ὦ Κύπρɩκαλλίστα θαλάμων φφμέуα τɩ θηρєύєɩν ρєтάν (ὦ Κύπρɩ should be preferred to Κύπρɩ L²P² even metrically because the responsion between and is far more irregular), IA 764/753, Τρєς ὂταν χάφΑρης;λκασπɩς Έλλάνων στρατίας, IA 765/754 πόντɩος є ὐπρῴροɩο πλάτας φ νά τє ν;σὺν1;υσìν καì ὂπλοɩς, Hel. 1347/1363 τύμπανά τφ ἒλαβє βʋρσοτєν φ κύκλɩος ἒνοσɩς αἰθєρία (τύμπανα need not be emended into τύπανα as far as this passage is concerned; if κύκλɩοςς is scanned as is obtained).

¹⁸ I take this opportunity to correct an error in my previous article CQ n.s. 32 (1982), 64Google Scholar. , E.Suppl. 999Google Scholar χαλκєοτєʋχος το Καπανέως should be scanned as , which correct sponds with the identical form.

¹⁹ IT 1098/1115, IA 169/190, 759/770 are discerned as glyconics comparatively easily. Hel. 525 (παντοδαπς πì уᾱς πόδα) and Or. 831 may be glyconics.

²⁰ Dale stresses that different branches of Greek lyric odes should not be mingled in analysis (“The Metrical Units of Greek Lyric Verse, I”;, CQ 44, 1950, 138 ffCrossRefGoogle Scholar. = Collected Papers, pp. 41 ff.). I agree with her in principle.

²¹ The colon is used also as the second colon of another dicolon by Eupolis (Heph. Ench. ch. 15, 22): (or, with Hermann's conjecture, .

²² GrenfellandHunt(followingWilamowitz), H. v. Arnim.O. Schroeder.D. PageandG. Bond.

²³ Overlapping pherecratean (4, Bond's metrical numeration), free responsion between (4 and 5), two emendations metri causa (κορτάλων str. 9, ποταμοîο ant. 6), lacuna (str. 10).

²⁴ It should be remembered that Alexandrian colometry is rather indifferent to dividing , Cf. Pindar fr. 169a 13 Sn, or Stesichorus Geryoneis.

²⁵ The term ‘unreiner SchluB’ was used by Wilamowitz and it is translated into English ‘impure ending’. Dale means by ‘drag’ the phenomenon that a short between two longs has occasional licence to lengthen. Her usage is not necessarily restricted to the endings. Even limited to endings, it is still questionable whether dochmiacs (e.g. ), ibyceans (), glyconics (Dale rightly rejects the idea of ‘iambic with impure ending’ which Denniston applies to or (‘Lyric Iambics in Greek Drama’ in Greek Poetry and Life, Oxford, 1936, pp. 141–2)Google Scholar) should be called by the same name.

²⁶ Cited by Pearson, A. C., CQ 23 (1929), 173CrossRefGoogle Scholar.

²⁷ The scheme is found, for example, in Platnauer, M., Iphigenia in Tauris (Oxford, 1938), p. 184Google Scholar, or Rupprecht, K., Einfiihrung in diegriechische Metrik (Munchen, 1950 ³), p. 52Google Scholar. The idea of ‘choriambic nucleus’ shifting its position seems to me still dominant; but I cannot accept it. The hypothetical colon ‘choriambic dimeter’ is actually to be divided into two different cola, namely wilamowitzian and iambo-choriambic dimeter and . Cf. n. 5 above.

²⁸ Strictly speaking, Wilamowitz supposes that the penultimate element of glyconic, as well as the second element, is treated as Senkung by Euripides, while Schroeder takes both and as a kind of Vierheber (enoplia) in his sense, a different verse from glyconic.

²⁹ Euripides, Hippolytos (Oxford, 1964), appendix IGoogle Scholar.

³⁰ Ba. 867/887 turns to if overlapping is rejected at 865–6/885–6 (p. 75 above).

³¹ I include S.Aj. 697 in the examples of gl + ba provisionally. This line has ὧ at the end. This is avoided by adopting another colometry Schroeder, , Dale LM ²151Google Scholar) which follows rhetorical division better, but needs to admit an unfamiliar colon repeated twice. Perhaps this colon is analysable as ba + dodr Α cf. Sophoclean examples of ba + gl, p. 80.

³² Maas does not describe anaclasis as the inversion of the positions of two elements. He may have been cautious so as to avoid any implication about the origin or the historical process of evolution of the cola, which are identical except for the two elements in question.

³³ Of 24 examples of wilamowitzian corresponding with glyconic (including wil/ gl), 20 are and 4 are . The fourth element of wilamowitzian in general is occupied by long more frequently than by short.

³⁴ Jebb, , at S. Phil. 1082/1102Google Scholar.

³⁵ Ench. ch. 10. Notice that he analyses the telesillean as ‘major ionic hephthemimer’ (ch. 11.2; this analysis is also found in Heliodorus. See his scholia on , Ar.Pax 1329 ffGoogle Scholar. (White, p. 440)) or as anapaestic (ch. 4.4) too.

³⁶ Dale adds a note to Barrett's scheme that –x may be inverted in the ‘left’ of ‘choriambic nucleus’ (LM2 153). Most likely this idea has its origin in the notion of anaclasis. But the tendency above observed about aeolic base and the fact that the third element of wilamowitzian is always long are not well explained.