^{page 107 note 1} Add Ar. Av. 1313–22.

^{page 108 note 1} Eur. Or. 1122, Phoen. 1349, Cycl. 215, El. 666, Ar. Nub. 469.

^{page 109 note 1} Com. Frag, adesp. 353 Kock.

^{page 111 note 1} That ⋯νοιτο is the true text, and this the meaning of the line, is shown by the order of the words.

^{page 112 note 1} Similarly in Hesych. s.v. καυκαλ⋯ς ii p. 452 φυλ⋯κταινα is a mistake for φυλ⋯κταινα.

^{page 112 note 2} Lobeck Rhem. 277, Parall. 207.

^{page 112 note 3} Hesych. .

^{page 112 note 4} Anacr. fr. 87 Bergk: spelt κνυσα in Herodas vii. 95.

^{page 112 note 5} Kock is plainly mistaken in desiring to read .

^{page 113 note 1} Liddell and Scott s.v. quote Pind. 0. x. 41 , wrongly rendering ‘all too late for.’ But s.v. ⋯ντ⋯ω they rightly take ⋯λώσιος to depend on ⋯ντ⋯σαις.

^{page 113 note 2} From stems in α or η the formation may be called legitimate. In Soph. fr. 122. 1 (Hesych. ii. 526) should perhaps be or ‘chosen as an honourable sacrifice.’ It looks at any rate like a compound such as .

^{page 113 note 3} It gives also , the reason for which is that was a word in late use. The schol. had μ⋯τοικοι.

^{page 114 note 1} for had been suggested, I now see, by Dr. Verrall before I commended it a year ago: Dr. Wecklein had omitted it.

^{page 114 note 2} Tyrwhitt's reading in place of οὐ, the phrase being equivalent to the common . At this rate φθ⋯νος and ν⋯μεσις refer to the same thing. Those who retain the MS. are obliged to refer φθ⋯νου to jealousy of heaven for some presumption of Orestes, and ν⋯μεσις to jealousy of heaven for the presumptuous language of Aegisthus. One could not praise such writing.

^{page 114 note 3} Journal of Philology xxiii. p. 272: add Quint, vi. 30–43.

^{page 114 note 4} 211, 165. This is the explanation of other words, applied to the delivery of oracles, as and and those which are technical of them, does not mean ‘to say,’ or as Liddell and Scott suppose ‘to noise abroad,’ but ‘to utter with a wild, confused, and half-articulate cry’ such as comes from the victims of a nightmare. Compare for instance Cho. 35, 533, Ag. 287. Upon all this subject I shall have more another day.

^{page 115 note 1} στρατωθ⋯ν is an epithet ‘limiting’ the metaphor.

^{page 115 note 2} Perhaps , though one rather desiderates . That at any rate should be the metre. Καλ⋯, the well-known epithet of Artemis, is used here after the usual custom, to flatter and conciliate the goddess.

^{page 117 note 1} So I understand it; but this explanation does not appear to have occurred to editors.

^{page 117 note 2} See the Asiatic view of this very matter as represented by Herodotus i. 4; when women were carried off, it was folly to make exertions for revenge, .

^{page 118 note 1} This is the nearest equivalent of 1437, as in other places, e.g. Ag. 283, Cho. 770.—In 1654 where she implores him to refrain from bloodshed, the appeal is by her love for him, ; but that is a different thing from talking of her εὐν⋯ with him to the public.

^{page 118 note 2} When I was studying scholia first, and reading those on Sophocles, I came upon (without a lemma) on O. T. 1070, and turned at once to see whether the text was or . I found . It is against all probability that ξα⋯ρειν should have been the lemma; but of those are the proper explanations: e.g. Pind. O. x. 99 : schol. p. 256 . Nauck for the same reason had conjectured . It is possible, indeed, to conceive and argue that Sophocles might wish to suggest ‘let her go’; but no one ever saw that word so glossed; and ξλ⋯ειν is the most appropriate word in this connexion: e.g. Eur. fr. 986 , P.V. 918 .