^{page 6 note 1} Above, p. 2. To his cases of хρν guaranteed by the metre (p. 4) add Soph. fr. 142. 13 P.; Eur. Alc. 633, 686, 709; Rhes. 752.

² p. 3: but Eur. fr. 736. 2 is not a good example, for can mean , and has no advantage over (By the way, seems to be a mistake for .) For χρή replacing δεî see Porson's note on Or. 667 (659), and add Rhes. 218, where the meaning seems to be ‘good luck is all that you need’ and δεîis preferred: but seeAlc. 653.

^{page 7 note 1} My paper was read to the Cambridge Philological Society on 16 Oct. 1941; see the Cambridge University Reporter of 11 Nov. MrJackson's, J. discussion of this question in the Classical Quarterly of October (xxxv. 170 f.)Google Scholar hadnot yet appeared, but I now refer to it in footnotes.

^{page 7 note 2} But I cannot find authority for Pearson's note ‘ἦ rec’on O.T. 1389.

^{page 7 note 3} For further detail, as also for ἧ in Plato, e.g. at Rep. 328 c, where Adam has ἧν without a note, see Rutherford's, W. G.The New Phrynichus, pp. 242 f.Google Scholar

^{page 7 note 1} The received conjecture ἢ λγοιοιν ἦ is dubious, and I suggest for consideration, though without confidence, γεμσι συν: ‘I consorted withleaders of men’. For a tribrach formed by a word of the measure of γεμσι see (e.g.) Hec. 1240 or Phoen. 79. But a better remedy, perhaps including neither ή nor any compound of ἦ, may yet be found.

^{page 7 note 2} These three passages were cited by Nauck, A. in his Euripideische Studien, i. 2Google Scholar. For his discussion of ἦν and ἦ see ii. 67 and 141.

^{page 7 note 3} When ἤμην was familiar, even ἦν was thought to need a note. See Murray's B atAlc. 655 and his M at Hec. 13 () and 15.

^{page 7 note 4} Who will may add fr. 953. 34 N.: but to me that passage is no more from Euripides than Rhes. Nauck's objection to is trivial, and his ἄζυγες is out of tone.

^{page 8 note 1} ‘Disconcerting’ to Jackson;neat, no doubt, ifἦν were wrong.

^{page 8 note 2} The κ is clear in Spranger's, J. A. photograph, on which see C.R. xlix (1935), 213.Google Scholar—Wecklein tells us that ἆρα is ‘for the metre's sake’ instead of ἆρα (as if the poet could not have written ἦν ἆρ').— Jackson says that, given those seven words, the poet's first impulse ‘must have been’ to write . ‘Must me no must, nor question me no question.’

^{page 8 note 3} Such are ἦμεν ἤσσονες, and Jackson's σπνιος ἦν ἤσςων νρ, ‘a worse man was scarce’

^{page 8 note 4} The syntactical connexion which it gives is superfluous in stichomythia.—Jackson's spoils the order of the words.

^{page 8 note 5} Wecklein' case against this line and its neighbours is coloured by his mistrust of the ἦν. Jackson regards thepassage, apart from 992, as ‘either Euripides or an imitation perfectto the last detail’. As for 992, he finds in Pearson's note ‘a pure fallacy’ but does not tell us what. Scaliger's troublemay have been due to the absence of the ἂν from some MSS. and editions: but the ἂν; is clear in L and is cited from P, and what is wrong with ‘If I were taking to tears (as I am not) I should be anobject of pity (as I am not) instead of a man of action (as I am)’ ? That is good enough for the Menelaus of this light-hearted play.

^{page 8 note 1} Without ἂρα the imperfectwould be strange. —Jackson dwells on the suspicions attaching to I.A., and to this passage in particular, and decides that ‘the passage is not evidence for the form—the form is evidence against the passage’ (not until the form has passed from the dock to the dungeon).

^{page 8 note 2} After η in general it is not at all rare.

^{page 8 note 3} For differences between one tragic poet and the others in such a matter compare μι and ὐμιν, in which Sophocles preferred ῐ.—No clear light is thrown on ἦ-ν by the Attic usage of other words in which a long vowel can have after it a ‘mobile’ consonant:οὐ, οὔτω, and pluperfects in -ει. They point to simple ἦ before a consonant if no pause intervenes; but οὐ and οὔτως before a pause point differentways. These words deserve further study from the present point of view. In Hipp., Hec, I.T., only at Hec. 296 and 785 do I find evidence for οὔτως before a consonant with no intervening pause. There is certainly in all our tragedies no trace of prodelision after οὔτω (any more than after οὐ); and at Ar. Lys. 816, where recent texts give 'κεῖνος perhaps κεῖνος should be read, though the context has no other un-Attic feature, and elsewhere in Ar. that word and its kin have except in un-Atticspeech.—Applying such analogies as best I can, I see nothing against the rule of thumb that Sophocles used ἦ but not ἦν. Tothetwo instances of the shorter form in Aeschylus a third would be added by Tucker's ἦ for ἦν at Supp. 344, a conjecture acclaimedas quite certain by Housman, in C.R. iv (1906), 106Google Scholar: but, if thefirstperson of the verb is right, conceivably ἦν may be right after all.—Jackson wonders why, if ἦν γὼ is right in our oldest play of E., ‘that useful cretic’ should not appear in ‘three myriads’ of later lines. Well, how often does the equally ‘useful’ cretic εἰμ' γὼ occur? In A. it seems to occur only twice, and in S. only thrice, though εἰμ is in A. and S. thrice as commonas ἦ. It is not before ἐγώ, usually superfluous, but before other words, that this ἦν would mostly be used by a poet who used it at all. Even if it could be shown to be relativelyrare in E., we need not be disconcerted, any more than by his one οἰσθα, or by the ᾔδειν of Ion 1187. Poets do not strictly ration themselves in the use of alternative forms.