^{page 183 note 1} First published in Berliner Klassikertexte, Vol. V. 2. Chief editions Diehl, E., Anthologia Lyrica Graeca (Teubner, 1925)Google Scholar fr. 96, Lobel, E., Σαπϕο⋯ς Μέλη (Oxford, 1925)Google Scholar ἒ. 3, Lavagnini, B., Nuova Antologia dei frammenti deila Lirica Greca (Paravia, 1932) No. XIGoogle Scholar; cf. Bowra, C. M. in Powell, J. U., New Chapters in the History of Greek Literature, Third Series (Oxford, 1933) pp. 4–5Google Scholar, 8–9.

^{page 183 note 2} Except where otherwise stated, all fragments are quoted by the numeration of E. Diehl op. cit.

^{page 183 note 3} See Hephaistion x. 6 (ed. M. Consbruch—Teubner, 1906). The numeration of this edition is used in all references to Hephaistion.

^{page 183 note 4} The names given to these metres by Hephaistion vii. 5–7 are bracketed to show that Lavagnini, their use is not intended to prejudge the issue.

^{page 183 note 5} So Ahrens for MSS χρύσειοι, which is not a Lesbian form (cf. Bechtel, F., Die griech. Diahkte, Vol. 1, p. 51, § 46)Google Scholar. Lobel, however, in discussing Sa. 55 (The Wedding of Hector and Andromache) writes ‘The neuter plurals πορϕύρᾱ⋯ργύρα (. 2. i. 9–10), on the one hand, and the for χρύσειοι (which appears to be demanded by the metre at inc. lib. 28 [i.e. Diehl 119]) on the other, are found varying with the form χρυσία (. 2. i. 8).(Ἀλκαίον Μέλη P. xiv).

^{page 184 note 1} Ox. Pap. 1232 and 2076, which contain parts of Sa. 55 (which is written in this metre), both have a colophon Σαπϕο⋯ς Μελ⋯ν β.

^{page 184 note 2} τ⋯ δ⋯ Αίολικ⋯ καλούμενα [δακτυλικά] τ⋯ν μ⋯ν πρ⋯τον ἒχει πάντως ἒνα τ⋯ν δισυλλάβων άδιάϕορον, πτοι σπονδεῖον ἤ ἴαμβον ἤ τροχαῖον ἤ πυρριχιον τοὺς δ⋯ έν μέοψ δακτύλους πάντας τ⋯ν δ⋯ τελευταῖον… δάκτυλον μέν ἥ κρητικ⋯ν δι⋯ τ⋯ς τελευταίας άδιάϕορον, έ⋯ν ⋯κατάληκτον ἥ, ⋯⋯ν δ⋯καταληκτικόν, κα⋯ τ⋯ ⋯π⋯ τούτου μεμειωμένα είς δισύλλαβον κα⋯ συλλάβην. It may be noted that H. gives no examples of the type which is καταληκτικ⋯ν είς συλλάβην.

^{page 184 note 3} It may be noted that Hephaistion includes this type under μέτρα μονοειδ⋯ κα⋯ όμοιοειδ⋯, and that he does not regard it as having any affinity with the μέτρα άσυνάρτητα or ⋯πισύνθετα.

^{page 185 note 1} In an article ‘Neuere Forschungen auf dem Gebiet der griech. Metrik’ (Neue Wegi zur Antike viii. pp. 35–54, Leipzig, 1929)Google Scholar. The greatness of my debt to Professor Körte will be visible to all who have read his article, but to give references in detail would have been cumbersome and often impossible, owing to the completely different plan of the present article. I should like to take this opportunity of expressing my gratitude to Professor Körte for the kindness with which he answered a number of questions which I addressed to him on points arising out of his article.

^{page 185 note 2} I refer especially to ProfessorHardie's, W. R. discussion of the matter in Res Metrica (Oxford, 1919) pp. 129–143Google Scholar, which seems to me definitely unfair and a serious blot on a valuable book.

^{page 185 note 3} In discussing Alkaios's metres in New Chapters p. 20 Mr. Bowra makes no mention of this crucial passage.

^{page 185 note 4} Cf. Wilamowitz-Moellendorff, , Gr. Verkunst p. 235Google Scholar n. 2.

^{page 185 note 5} See, for example, Thomson, G., Greek Lyric greatness Metre (Cambridge, 1929) p. 153Google Scholar.

^{page 185 note 6} Those who doubt the reasonableness of this scansion need only look at Soph. O.C. 670–672, where the Glyconics must be so read, however they are scanned.

^{page 185 note 7} This description is now apparently accepted by . MrBowra, (New Chapters pp. 4–5)Google Scholar; but I had already arrived at the view here expressed before the publication of his article.

^{page 185 note 8} Cases are also found where the quadriarticle. syllable is added at the end (e.g. Alk. 45, 46), or where a cretic ( = acephalous iambic dipody) is Metrica added at the beginning (e.g. the first line of the stanza in Sa. 97, 98) or a bacchius (= catalectic iambic dipody) is added at the end (e.g. the third line of the stanza in Sa. 97, 98): but these types are not here in question, though their of existence strengthens the argument for the quadrisyllabic scansion of the Glyconic.

^{page 186 note 1} The fact that 97. 9, whatever we may call.it, must have been scanned – o – u – u [ u – ] renders impossible the restoration [θέαν] proposed by Lavagnini, op. cit. pp. 167Google Scholar, 184.

^{page 186 note 2} To judge, at any rate, by the passages in Wilamowitz, , Gr. Verkunst, which deal with ‘Freie Responsion’ in Glyconics, (i.e. pp. 235, 238, 259, 510, 511, 550, 552)Google Scholar.

^{page 186 note 3} For the real meaning of βάσις and the only sense in which the name ‘Aeolic Basis’ is intelligible see Schroeder, O., Nomenclator Metricus (Heidelberg, 1929)Google Scholar s.v. β⋯σις. But even in Schroeder's sense the term is not apparently necessary.

^{page 186 note 4} Normally, because if we take account of the variations in Sa, 97 and 98 the scheme, even of the literary Glyconic, might be xxxx|xxu-, which is identical with that for Korinna's dimeter.

^{page 186 note 5} These references are taken from . MrBowra's, Tradition and Design in the Iliad, p. 56Google Scholar.

^{page 187 note 1} The obvious explanation of this line as anapaests is barred, not only by Hephaistion's description (which is not in itself conclusive) but also by the metre of the next line χρ⋯ συνμποσίαις ⋯πʹ ὂνασιν ⋯μοί γε γενέσθαι which is quoted with it.

^{page 187 note 2} Cf. Heph. xi. 3–5. It is open to those who prefer the spondee dactyl trochee scheme for the Glyconic to analyse the dimeter here as spondee, anapaest, catalectic iambic dipody, and to account for the longer lines by assuming an inserted Ionic a minore—but apart from the clumsiness of this, the fact that it disregards Hephaistion's evidence and the existence of type (2a) both seem to raise serious objections to any such description.

^{page 187 note 3} This is the Alcaic dodecasyllable of Heph. xiv. 4. For a more fanciful derivation cf. Headlam, W. in J.H.S. xxii. pp. 226–227Google Scholar.

^{page 188 note 1} This derivation of the Aristopbanean, if accepted, destroys the theory of the Sapphic stanza put forward by . MrThomson, , op. cit. pp. 18–19Google Scholar, which rests on the expressed assumption that the Aristophanean is a Pherecratean with the dactyl in the first place, and the tacit assumption that the Adonius has no independent existence. That this latter assumption is incorrect is proved by Sa. 126 (e.g. σκιδναμ⋯νας ⋯νστήθεσιν ὃργας, i.e. two Adonii). . MrThomson's, ‘Aeolic Tripody’ (p. 154Google Scholar, cf. p. 10, where Soph. El. 245 εί γ⋯ρ ⋯ μ⋯ν θανών is quoted as an example) is perhaps a catalectic Aristophanean.

^{page 188 note 2} vii. 8. The shakiness of the argument here cannot be denied. Of the seven stanzas of Alkm. I whose last words survive, four end with lines of the form ἒργα πάσον κακ⋯ μησαμένοι (35) and three with lines of the from ⋯υειρομέναι μάχονται (63), which looks like an early case of the choriamb-trochaic substitution allowed both in Pindar (e.g. Ol. vi, 76=100 Schroeder) and Bakchylides (e.g. V. 71, III = 31. 151,: 191 Blass-Süss). It is not, however, necessary to the argument to assume that double scansion originated with Sappho or Alkaios; and it may well be that the evolution of the ‘Alcaic decasyllable’ (an obviously effective clausula for a dactylic system) is the cause and not the effect of choriamb-trochaic substitution (which only seems to occur at the ends of lines in Alkman). The appearance of Praxilla in the series is no bar to this view, as the fact that her dithyrambs were written in heroic hexameters (cf. frs. 1, 2) indicates that she can hardly be later than Stesichoros.

^{page 188 note 3} No early Greek example of the ‘Greater Sapphic’ is extant, but I see no reason to assume that it is Horace's own invention.

^{page 188 note 4} I draw attention to this point because Professor Körte in dealing with this line (op. cit. p. 43) abandons his principle that the greatest respect should be paid to the ancient metrical writers, and especially to Hephaistion (p. 40), and describes it as an acephalous Glyconic, i.e.!=; – u [ o – u –.

^{page 189 note 1} Wilamowitz, (Sappho und Simonides p. 20)Google Scholar tried to torture these two lines into Asclepiads. I follow the metre of Edmonds, J. M. (Lyra Graeca, Vol. 1, Sa. fr. 108)Google Scholar, which involves only slight corrections of the wording of the quotation in Plutarch. There 1. I ends μοισοπόλωνοίκία (corrected by Edmonds), and the second word in 1. 2. is εῖναι (corrected by Wilamowitz). See Diehl and Edmonds ad loc.