^{page 269 note 1} I am most grateful to Prof. H. Lloyd-Jones, Prof. D. L. Page, Mr. P. J. Parsons, and Mr. Thomas Gelzer, who read through this article in manuscript and made a number of valuable suggestions.

^{page 269 note 2} In Antinoopolis Papyri, vol. iii, verse 6 was mistakenly printed in the secondary transcript as:

^{page 269 note 3} Reported in Antinoop. Pap. iii.

^{page 270 note 1} pp. 258 ff. above.

^{page 270 note 2} Barber's change oftoin v. 7 needs no detailed defence of course. Callimachus is using an established Homeric expression: Il. 10. 2610. 91, Pind. N. 8. 2Critias 4. 10 D², Mosch. 2. 3, Call. fr. 21. 2 ]in the same sedes. Cf. Alcman 3. 72, Eupolis 94. 5 K.

^{page 270 note 3} Although Musaeus 33 readsFor further material relevant toandand alsocf. Thuc. 1. 84. 3 (and scholia ad loc.), Plato, Charm. 160 e and 161 a, Laws 6. 772 a and 2. 671 d.

^{page 271 note 1} For other examples cf. Sappho fr. 137. 5, Theogn. 85 f., ? Aesch. fr. 355. 21 ff. M, Aristot. apud Athen. 13. 564 b (fr. 96 Rose), Orph. Arg. 933, Longin. De Sub. 4. 4, Isid. Pelus. Ep. 5. 28, proverb apud Suda s.v.(Adler, ii. 163). Call. fr. 80. 10–11 probably reads:

in which Pfeiffer supplementsA.R. 1. 790 ff. supports him:

^{page 271 note 2} For Barber's suggestioncf. A. Otto, Die Sprichwörter und sprichwörtlichen napcids Redensarten der Römer, 171: the phrase is usuallybut that is only a minor drawback. Perhaps read

^{page 272 note 1} Mr. P. J. Parsons has very kindly re-examined this part of the papyrus and writes that ‘suits the traces, as well as’. I had at first thought to readbut Prof. D. L. Page pointed out to me that thecontract form is much commoner. This is one of those parts of the Greek language where opinions differ as to how many verbs can be extracted from the evidence (see, e.g., LSJ s,v.Curtius, G., The Greek Verb (1880), 104Google Scholar, Tucker ad Aesch. Sept. 836, Schwyzer, Griech. Gramm. i. 679), and MSS. are confused (cf., e.g., A.R. 2. 296, 1010; 3. 307). Homer never contractsthough the shortenedform is common: in fact there are only two certain examples ofcontracts in the MSS. of the main Greek authors―Soph. Trach. 645Parthenius xxvi Mcinekc (Anal. Al. p. 279 = Diehl Anth. Lyr. Gr. ii. 244, fr. 15)Elmsley's emendationis accepted by most editors for the Sophoclean passage,on the other hand has the following post-Homeric backing: Aesch. Pers. 25, Sept. 31, 836, 842, Soph. Aj. 1414, Aristoph. Vesp. 209, 458, Epilycus fr. 3 K (i, p. 803) which may, however, be corrupt, and Plut. Mor. 2. 362 c. In Callimachusoccurs at v. 4, and perhaps again at fr. 7. 31–2 (see Barber, and Maas, , CQ [1950], 168)Google Scholar. The actual formis not as yet attested anywhere, though Prof. Page notes Hesychius(‘where Latte suggestsbut that has its own entry below in Hesych.’).

^{page 272 note 2} In v. 4 ifbeforeis correct then it must not be the final syllable of a word-unit: in Callimachus the eighth element is not long if word-end follows (cf. Maas, § 92). Thus(see app. crit.), orwould be suitable.

^{page 272 note 3} According to Tzetzesoccurred in Hipponax fr. 65 C; Masson excludedas a gloss and restored metrical sense to the line. Extra support for this may be found in Phrynichus Soph, who underin Praep. Soph. p. 87 B comments:also occurs in the Corp. Gloss. Lat. = ‘morsus’.

^{page 273 note 1} The only salient exception to this is: of pain, once each in Aesch., Soph.,and Lucian.is used once by Euripides to mean ‘torturing’.

^{page 274 note 1} This is of course only a supplement, but it is very difficult to see what else thecould be doing to.The other obvious possible reconstruction istakingas nominative: e.g. ‘another song became necessary for the food-jar’, but this founders onsince the tense would preferably be imperfect. Onsee now Masson, E., Recherches sur les plus anciens emprunts sémitiques en grec (Paris, 1967), 44–5.Google Scholar

^{page 274 note 2} Aesch. Cho. 1024 f. has an abstract subject and is not very relevant:.

^{page 274 note 3} Cf. Eustath. 746. 3

^{page 274 note 4} Ter. Phorm. 695. Cf. Cic. Ad Att. 1. 19. 8 at crebro mihi … insusurret… cantilenam illam uam, 13. 34 haec decantata erat fabula; Sen. Ep. 24. 6 decantatae … in omnibus scholis fabulae istae sunt; Macrob. 5. 2. 6 talia ut pueris decantata praetereo. Commentators sometimes explain this idiom as a Greek one: this is incorrect.

^{page 274 note 5} Documentation for this idiom may be useful since it does not appear in the paroemio-graphers. Examples are: Aristoph. Av. 39–41, fr. 7 Dem., Eupolis fr. 2 Dem., Men. Epitrep. 408–9. Otherwise it is documented only by the lexicographers: Photius, s.v.(Reitzenstein, Anfang … 48. 7 ff.) explains, and so too Phrynichus Soph. Praep. Soph. 21. 1. Orion Thcb. Etym. 23. 1 gives the origin:. The full range of the expression can be seen in Plato, Lysis 204 d (of Hippothales the lover)which leads to 205 b–cand 205 d. A pleonastic version of this expression is found in the paroemiographers,: Zenob. 1. 72, Diogen. 2. 19, and Greg. Cypr. (Cod. Lied.) i. 47 (ii. 60) (the phrase is used literally with a possible pun on the idiom in Aristaen. 1. 27). Cf.═ ‘proclaim’ in Plato, Laws 854 cand 854 d …is the verb which is more usual in this sense of ‘harp upon, repeat’ (see LSJ s.v. II).

The only example of ‘sing the same song’ is Theophil. Com. fr. 7 (Kock ii. 475):but the metaphor here is not an independent one, it is an extension ofin the previous line.

^{page 275 note 1} This scarcely needs illustration: cf., for example, frr. i, 2, 112, 203, 228.

^{page 275 note 2} Thereafter the expression appears occasionally in the Christian writers, and viain the commentators, grammarians, and lexicographers—see Blomfield, loc. cit., A. C. Pearson on Sophocles fr. 885, F. H. M. Blaydes on Aristoph. Eq. 1068.

^{page 276 note 1} : it may be worth remembering that in fr. i. 7 Callimachus addresses the Tel-chines as: here too perhaps were mentioned the uninspired poets whose Muse was not.

^{page 276 note 2} Cf., of course, fr. I. 37–8. Mr. Thomas Gelzer suggests to me that the ‘gods’ in this passage might be Callimachus' patrons the Ptolemies. This is a very attractive suggestion, but one which needs caution, I think. Callimachus is very flattering about the Ptolemaic family, but only in his panegyric poetry does he actually identify them with the gods; in his less heightened poetry his royal patrons are usually only juxtaposed with the immortals.