^{page 46 note 1} In the Dyskolos (879) Menander still has a flute-player for his iambic tetrameters catalectic.

^{page 46 note 2} Doubtless the delivery of other forms of stichic verse fell short of what could properly be called singing; we usually conceal our ignorance of what it really sounded like by using the term ‘recitative’, a rough translation of Plutarch's but this certainly implies some sort of instrumental accompaniment.

^{page 46 note 3} For instance, the ease with which Greek syllables, aided by the absence of stressaccent, lend themselves to precise conven tional distinction of long and short, the abundance of short syllables, especially in the earlier stages of the language, the flexibility conferred by its many inflected forms and rich choice of particles. The device of contraction is essential to the concept, for rhythmical variation and much more for word-coverage. (Imagine an epic unable to speak of or The various licences through which it enabled itself, always under strict convention, to enlarge its vocabulary further, or adapted its formulae to linguistic change, have been strangely used as an argument for foreign derivation, but that is to pick up the problem from the wrong end altogether.

^{page 47 note 1} I find myself quite unable to take seriously the claim that the Margites is older than Archilochus. Bergk's brilliant conjecture (Iambog. 153, on which see Rzach-Rader- macher s.v. Margites in R.E.) disposes of the one apparent piece of evidence. Certainly the perpetrator of that farrago was himself no metrical in ventor—the whole joke assumes familiarity with the iambic line and its normal uses— and we should be driven to postulate a considerable body of iambic poetry already in existence before the seventh century and a creative genius whose poetry and whose name have vanished without trace.

^{page 47 note 2} I do not feel that the explanation given by Snell, Gr. Met. ³, p. 11, of a time-count in iambic and trochaic is quite sufficient here. He thinks the fractional pause at word-end added to a long syllable would distort the time-relation too much; but if, as I believe, long anceps was in any case perceptibly shorter than true long, no such distortion would be felt. It is also noteworthy that Alcman in the trochaics of his Parthenion, both dimeters catalectic and acatalectic and trimeters, avoids as carefully as I except where the anceps, long or short, is a monosyllable (enclitic, proclitic, or neither), which suggests again that it was the rhythmic effect rather than the time-relation to which he was sensitive. (Snell does himself also speak (ibid., p. 6, n. 1) of the ‘rhythmische Bedeutung’ of Porson's Law.)

^{page 48 note 1} Plut. De Mus. 1141, in referring to Ar.'s ‘part song, part accompanied speech, in which he was followed by the tragedians’, is anything but clear.

^{page 48 note 2} Aristotelian fundamentalists in the ‘his tory of Greek drama’ should pause to reflect what our ‘histories of Greek metric’ would look like if we had no external evidence by which to check his obiter dicta. For him Greek drama was a thing of long, long slow growth (indeed there seems no obvious reason why on his theories tragedy did not gradually develop over the whole of literate Greece, instead of springing suddenly to life in sixth-century Attica), and the ‘coming of speech’ into it, which prompted ‘nature herself’ to invent the appropriate metre, only makes sense if dramatic dialogue ante dated all other uses of the trimeter ( cannot mean ‘discovered which was’ the appropriate metre). The fact that he has shortly before been speaking of an era in which poets were either writers of heroic hexameters or of iambic lampoons (for which the trimeter duly made its appropriate appearance), this being apparently antecedent to tragedy and comedy, only shows how mistaken it is to try to reduce his various ‘constructions’ to a single homogeneous and logical, or historical, line of development. The organic growth he is immediately concerned with here is that of drama itself, from a capering, satyric, trochaic-singing to a dignified, human, iambic-speaking performance. The incompatibility of these two ‘constructions’ so close together within this one chapter raises doubts about the shape of this part of our extant Poetics which are another matter again.

^{page 48 note 3} This might account for the Suda's statement that Phrynichus was the ‘inventor’ of the tetrameter.

^{page 49 note 1} For the form cf. Chantraine-Irigoin, , R.É.G. lxiv (1951), 1.Google Scholar

^{page 49 note 2} Nor should the three-line stanza (or part-stanza) be run together as an ‘iambic hexameter’. Alcman composes in cola, and , with its long an- ceps and rhetorical pause, is an unmistakable colon. The smoothness of the follow-throu from one line to the next is characteristic— as throughout in the Parthenion.

^{page 49 note 3} Hermann's emendation is essential, pace Wilamowitz and others, not only for the improved sense but because with the line is not only ametric but non-metrical.