^{page 33 note 1} Snell, B., Pindari Carmina cum Fragmentis ii ³ (Leipzig, 1964), 94.Google Scholar The following cor-rections and additions to Snell's apparatus should be noted: v. I codd. , v.κєλάρυζєν, v.2 codd. κρηνάων: Bergk, v. 3 codd. θήλєον (not ): Page; (not ) Naber, Schroeder. I thank A. T. Cole, Warren Cowgill, and C. M. Dawson for their helpful comments on a previous version of this argument.

^{page 33 note 2} ZfA i. 433 f.; cf. Pindari Carmina (Gotha, 1843)Google Scholar, fr. 182^d, 330.

^{page 33 note 3} ‘Lesefrüchte XXXII’, Hermes xxxiv (1899),Google Scholar 223 f. (repr. Kleine Schriften iv [Berlin, 1962], 64–6Google Scholar).

^{page 33 note 4} The Works of Pindar ii: Critical Com-mentary (London, 1932), 426.Google Scholar Cf. Schroeder, O., Pindari Carmina ( = Poetae Lyrici Graeci⁵, Leipzig, 1900),Google Scholar 423 f., Sandys, J., The Odes of Pindar (London, 1915),Google Scholar 562 f., Puech, A., Pindare iv (Collection Budé, Paris, 1923) 235,Google ScholarBowra, C.M., Pindari Carmina (Oxford, 1947),Google Scholar fr. 93, Turyn, A., Pindari Carmina (Cracow, 1948),Google Scholar fr. 108. Not all editors express Farnell's assurance. Puech, for instance, writes: ‘je crois prudent de rester dans le doute’, and the fragment is asterisked by Schroeder and Snell. None of these editors, however, relegates the fragment ment to a section of dubia.

^{page 33 note 5} Anacreontis Carminum Reliquias edidit T.B. (Leipzig, 1834),Google Scholar vii f.

^{page 33 note 6} Leipzig, 1843, pp. 266–7 (hereafter abbr. PLG).

^{page 33 note 7} Beiträge zur Kritik der Poetae Lyrici Graeci, edidit Theodorus Bergk (Göttingen, 1844), 8.Google Scholar

^{page 33 note 8} Leipzig, 1853, pp. viif. The motto to Bergk's Praefatio reads:

^{page 33 note 9} In PLG ³ (Leipzig, 1866; cf. PLG ⁴, 1878, the edition commonly cited), it became fr. adesp. no. 90.

^{page 33 note 10} Loc. cit.: ‘Pindarus si carminis auctor est …’; cf. Forssman, B., Untersuchungen zur Sprache Pindars [ = Kl.–Phil. Stud. 33] (Wiesbaden, 1966), 160 n. 4.Google Scholar Schroeder's partial uncertainty is also reflected in his 1930 Teubner edition, which presents the passage between fragments 99 and 104c without specific enumeration.

^{page 34 note 1} Poetae Melici Graeci (Oxford, 1962), 528,Google Scholar fr. adesp. 79 (cited, but presumably not accepted, by Snell). Cf. Edmonds, J.M., Lyra Graeca iii (London, 1940), anon. fr. 74, 450–1Google Scholar, who cites the Pindaric ascription (giving Schroeder the credit).

^{page 34 note 2} Cf. Schneidewin (Pindari Carmina, loc. cit.): ‘referenda autem haec videntur ad carmen daphnephoricum Apollini Galaxio certo tempore a Boeotis oblatum’, and Wilamowitz, Farnell, Schroeder, and Turyn locc. citt.

^{page 34 note 3} The uniqueness of Plutarch's testimony especiraises a familiar problem. If we had to rely exclusively on Pausanias, for example, we would have no way of knowing that the Spartan inscription which he reports at 5. 24. 3 (= IG v. I. 1562 Olympia, V B.C. init.) contains the form and not the expected Doric form, recorded by Pausanias (cf. Francis, E.D., ‘Greek Disyllabic Roots’, Yale diss. 1970,Google Scholar 266 f.).

^{page 34 note 4} While in a discourse on the Pythian oracle, is likely to refer to Apollo, Bergk (PLG ¹ 267) suggests Dionysus as equally possible and Sandys accepts this 528, latter identification without comment. The fact that Apollo and Dionysus shared Delphi may have encouraged this suggestion. More-over, Dionysus is known as a milker of lions (cf. Alcm. 56 ap. Athen. 11. 498 f., also Aristides, Or. 41. 7) but emendation of codd. is less attractive than Leonicus in turn seems to exclude a reference to lions. The gender of can hardly designate the Bacchantes associated with the Dionysian ritual. Farnell is therefore probably correct in rejecting a Bacchic interpretation especiraises ally since Photius may offer corroborative evidence for the cult of Apollo Galaxios.

^{page 34 note 5} Drachmann, A.B., Scholia Vetera in Pindari Carmina i (Leipzig, 1903), 3.Google Scholar

^{page 34 note 6} Ap. Phot. Bibl. 321 b 29 f.: [Marcianus: A],

^{page 34 note 1} Euripides: Hippolytos (Berlin, 1891), 230–1Google Scholar ad v. 1103.

^{page 34 note 2} Euripides: Hippolytos (Oxford, 1964), 366–8,Google Scholar with extensive bibliography; cf. Fraenkel, E., Aeschylus: Agamemnon (Oxford, 1950), ad v. 562.Google Scholar Puech prints which seems totally meaningless.

^{page 34 note 3} Sappho and Alcaeus (Oxford, 1955), 244–52,Google Scholar esp. pp. 248 f.

^{page 34 note 4} T. Bergk, Anacr. Carm. Reliq., viii. Page's presentedition reads:

The sequence is [pher.], glyc., pher., pher., glyc., glyc., pher. (on the scansion of see bolow), pher., pher.

^{page 35 note 5} Schneidewing 1834, id. 1843, Bergk, 1843, id 1853. On Schroeder' see below.

^{page 35 note 6} The regular quantity of-υ-in the presentedition imperfect stem of is ambiguous but was probably short (cf. Schulze, W., Quaestiones Epicae, [Gütersloh, 1892], 343Google Scholar). Out-sdis side the present system, the stem was at Ap. Rh . I. 589 beside impf. (ibid., v. 862) could be exiriined as due to metrical lengthening. These considerations may slightly favour a scansion of

^{page 36 note 1} Schneidewin's of 1834 is an implausible hybrid with a Doric stem and Ionic suffix.

^{page 36 note 2} Emendation may be unnecessary since, as Chantraine observes (Grammaire homérique, i [Paris, 1958], 201Google Scholar), ‘la forme en -ων qui est le type morphologique des vieilles formules apparaît prépondérante. Elle est sentie comme archaïque par les aèdes.’

^{page 36 note 3} -ξєν Schneidewin, Bergk, Bowra, Turyn, Page, Snell; -¶єν codd., Wilamowitz, Schroeder, Sandys, Puech. It may be argued that the impf. κєλάρυζєν is stylistically more appropriate than the aor. -ξєν.

^{page 36 note 4} Cf. Gow, A.S.F., Theocritus ii ² (Cambridge, 1952), 164,Google Scholar 166 f. on vv. 135, 137, I44f.

^{page 36 note 5} Op. cit., 437 f. (cf. Chantraine, op. cit., 104). Hamm, E. (Grammatik zu Sappho und Alkaios [Berlin, 1957], 41)Google Scholar accounts for occurrences of in Aeolic lyric as borrowas ings from Epic : ‘keine metrische Notwendigi keit liegt vor bei .… und … dieses in der 4. Zeile der sapphischen Strophe, die sich besonders oft an Hexameterschlüße anlehnt’ (cf. Risch, E., ‘Sprachliche Be-merkungen zu Alkaios’, Mus. Helv. iii [1946],Google Scholar esp. 254f.). The metrical restriction is valid since in Sappho 2. 5 (the first line of a stanza) is iambic. As Page remarks (Sapph. and Alc., 287), is ‘alien to the Lesbian dialect’.

^{page 36 note 6} Index Verborum zur frühgriechischen Lyrik (Heidelberg, 1966), 171.Google Scholar

^{page 37 note 1} Frisk, Hj., Griechisches etymologisches Wörterbuch i (Heidelberg, 1960), 671Google Scholar (cf. θήλєιαν Pi. O. 3. 29).

^{page 37 note 2} Pfeiffer (ad loc.) also quotes v.l. in Schol. AT ad Il. 21. 454, Et. gen. B. v. ( = EM 450. 56) … and Il. 9. 577: (cf. Callimachus'

^{page 37 note 3} On the etymological function of in this formation see Benveniste, E., Noms d' agent et noms d' action en indo-européen (Paris, 1948),Google Scholar 117 f.

^{page 38 note 1} On the more frequent Homeric occur-rence of θήλєια etc., see Chantraine, op. cit. 252.

^{page 38 note 2} Cf. Pi. N. 4. 88 : θάλησє Κορινθίοις σєλίνοις, Il I. 234–6: Od. 5. 73–4: Sa. 2. 9–10: but also Pi. N. 10. 41– (of Argos):

^{page 39 note 1} Mnem. N.S. xii (1884), 40.

^{page 39 note 2} ‘'Eπίμπλαν certe legisse Plutarchus videtur', a statement for which Schroeder offers no corroboration. Cf. the editions of Sandys, Puech, Bowra, Turyn, Snell, and Pohlenz-Sieveking's Plutarch text; note, however, that Page reads unaugmented πίμπλων with the codical suffix -ων.

^{page 39 note 3} This form is queried by Schwyzer, , Griechische Grammatik i (Munich, 1939), 689;Google Scholar cj. πίμπλη LSJ 1405.

^{page 39 note 4} Cf. LSJ 1405, Bechtel, , Griechische Dialekte i (Berlin, 1921), 28;Google ScholarThumb-Scherer, , Griechische Dialekte ii (Heidelberg, 1959), 101.Google Scholar

^{page 39 note 5} Op. cit. 124 n. 293.

^{page 40 note 1} As Lobel remarks (AAKAIOY MEAH [Oxford, 1927], 27Google Scholar), ‘lectio nonnullis locis incertissima’. The text is printed with more confidence in Lobel-Page, Poetarum Lesbi-orum Fragmenta ad loc. and Page (Sapph. and Ale., 171 f.) offers a brief commentary. His translation of πίμπλєισιν as 3 sg. ‘[he] fills’ (cf. Edmonds, J.M., Lyra Graeca i [London, 1928], 367Google Scholar) is, however, not likely to be correct (cf. at Ale. 36. 23 L = Berl. Kl. Texte v (2). xii. 2).

^{page 40 note 2} Schwyzer (op. cit. 798) cites a Lesbian impv. πίμπλη for which there is no clear textual support.

^{page 40 note 3} Cf. Buck, C.D., The Greek Dialects ² (Chicago, 1955), 37;Google ScholarLejeune, M., Traité de phonétique grecque ² (Paris, 1955), 235–6.Google Scholar

^{page 40 note 4} One may possibly compare the un-usually abundant use of Epic dialect in Alcaeus' Hymn to the Dioskuroi and Page's comments ad loc. (Sapph. and Alc., 266). On the function of Epic in Lesbian poetry see Page (op. cit. 65 f. on Sa. 44) and the references to earlier discussions which he cites (esp. Lobel, Σμ. xxvf, Aμ. xf.).

^{page 40 note 5} As de Saussure already asserted in Mémoire sur le systèms primitif des voyelles dans les langues indo-européennes (Leipzig, 1878), 14.Google Scholar

^{page 40 note 6} Francis, op. cit., 260 f., esp. 271, also Kuryiowicz, J., L'Apophonie en indo-européen (Wroclaw, 1956), 205 n. 49,Google Scholar and Beekes, R.S.P., The Development of Proto-Indo-European Laryngeals in Greek, (The Hague, 1969) 246.CrossRefGoogle Scholar Attempts to justify Naber's as a poetic Doricism based on the model:

(Dor.) 3 pi. impf. : ptc. :: (as in ) where (: Attic-Ionic X. An. 1. 5. 10), or Wilamowitz's (leg. ) after (Pi. P. 3. 65) are rendered unnecessary if one disaccepts that the grounds for attributing this fragment to a poet writing in Doric (or Aeolic) are insufficient. On the other hand, if emendation is required to accommodate a Doric (or Aeolic) provenance, the other-wise unattested represents the crrect dialect form.

^{page 41 note 1} e g S. Ant 121, E. Hec. 1072 (cf. E. Or. 1363, πλήσαιμι E. IA 234, S. OC 528).

^{page 41 note 2} The extensive Homeric evidence is disaccepts ussed by Chantraine (op. cit . 472 f.).