^{page 95 note 1} B 475, 479, 529; E 571; Z 2; K 317; 0 585; Π 143 = T 390; Φ 575, 604; ψ 760; ε 272, 400, 476; σ 323.

^{page 95 note 2} See Van Leeuwen, Ench.

^{page 95 note 3} The πóλoγoι are reckoned.

^{page 95 note 4} Not reckoned, A 446, ∑ 519, and Ψ 295.

^{page 95 note 5} Not reckoned, Π 521, P 689, Ω 384 ó ἔριστos.

^{page 96 note 1} The extreme importance of the elision will be clear later.

^{page 96 note 2} Much rarer than is usually supposed. For instance, it is suspicious that the famous ó γρων is nearly always preceded by elision, even in the strong caesura where elision is disliked. And many of the other occurrences are not reassuring; e.g. All might well conceal o*** ἔνεκα χρύσην or something of the kind.

^{page 98 note 1} This point is so important that the narr. elisions must be enumerated:—

Possibly removable:—A 71, 188, 501, 536, 609; Γ 453; Z 313; H 271; Λ 309, 484, 495, 621 (cf. P 745); M 101; N 15, 27; Ξ 515; 0 315; Π 190, 300, 317; P 112; Υ 279, 340; Φ 69, 182, 490, 595; X 77.

Not removable:—A 2, 492? Δ 222; E 851, 863, 899; H 65? N 8, 369, 417, 640; Ξ 361? 458 486; 0 120, 279, 676; P 210, 698; Σ 591? T 386; X 26, 94.

Not reckoned as elisions:—

Genit. in -ov; E 338; N 211; P 611, 697; Σ 575; T 384; also, perhaps, Ξ 427.

Plupf. in -ει, -η: B 38; Σ 557; Υ 466; and κυκα***, Λ 624.

Not reckoned as tritotrochaic:—B 222; Γ 452, Z 22; H 2; N 163; Π 390, 419 (cf. Ξ 440).

^{page 100 note 1} The figures in brackets give the number of instances that must be, or that might be, removed by slight changes.

^{page 101 note 1} If semi-tomic overlengths should seem at first sight rather plentiful, it must be kept in mind that all the connective particles begin with a consonant. Scansions, then, like Aἴας ῥεκτωρ δ, etc. may be called forced overlengths; in the early part of the line especially, they are almost inevitable. The treatment of the 2nd spondaic in narr. may also need explanation, since 56 overlengths look too many for the 214 breaks of small pause-value; but it will be found that, of the 56, 14 are μν ττρ***τoν (-·α)—a very mild instance.

^{page 102 note 1} Narr.— Γ 26; Λ 414; P 282. Speeches—Γ 179; E 442; N 733; Ξ 233; 0 66.

^{page 102 note 2} Ten of them are combinations like π κλισας, where the line may rest on the hephthem.; and 11 more are phrases like π κλισας, κa, where again the caesura is in doubt; 12 are formed by δ (read δ' ἄρ?), and 13 are made by ephelcystic endings.

^{page 103 note 1} Ignored—βo***πóτνια, A 551, 568; Δ 50; Ξ 159. 222, 263; 0 34; Π 439; Σ 239, 360; Υ 309; and πpòς μ***θoν ἔειπε, E 632; Z 381; H 46; Λ 429, 522; N 306; Ξ 189; 0 13; also ἔκ θoμòν ἓλoιτo and the like, Γ 325; E 317, 346; M 150; and βo***ν κπαρ λσθαι, Λ 550=ρ 659.

^{page 103 note 2} Final Wernecke—E 763, 811; Π 13; ϒ 212. Elided: Γ 105, 254; Z 284, 326; H 30; X 389.

^{page 103 note 3} Ignored—βo***πις πóτvια O 49; Σ 357; πpòς μ***θoνἔειπεv B 59; and μ***ν, ύμ***ν A 579; A 671.

^{page 104 note 1} Especially, of course, after the strong caesura.

^{page 105 note 1} The opt. 3rd sing, was originally -oιτ, and the accent of τσoι is here regarded as an irrelevant petrifaction; it is of course not intended that the fut. opt. is Homeric.

^{page 105 note 2} This order would later have been changed, to bring the apparent hiatus into the 3rd trochaic.

^{page 106 note 1} 2nd troch.—

Narrr.— Δ 277; E 122; M 422; N 61; 0 322, 447; T 420. (Neglected, A 45; Γ 375; M 200; 0 459; X 368; Σ 251.)

Speeches—A 83, 269, 356 = 507; B 27 = 64, 240, 290; Γ 167; E 404; 0 474; Π 89; P 470; X 108. (Neglected, 9 instances.)

3rd dact. —

Narr. — Only the resumptive formula ὢs εlπ***ν τoύs (τήv) μν λιπεv Δ 292, 364; Σ 468. The phrase is a shift from the 2nd dactylic, and is not an absolutely certain instance, even so. (Neglected, Z 118; N 157, 803; O 647; P89 (vια λθ'), 523; Π 717.)

Speeches—Σ 403 pεν; T 88 ϕρεσιv. -v in thesi. —

Narr. — Γ 16? M 116; N 16, 353; Π 220 (μ. πρ.)? P 396? 705; X 320, and Γ 348 = H 259 = P 44. (Neglected, A 450, 498, 571; B 411; Γ 275; B 777, 870; H 347; Π 310, 680; A 479; 0 701; Σ 569; Γ454; Δ 444; M 91.)

Speeches—A 388; B 347; Δ 289; A 319; N 634; Ξ 338; 0 197, 491; Σ 123; ϒ 183; X 497. (Neglected, 12 instances, including κευ, μîν, and ύμîν.)

^{page 107 note 1} The present-reference aorist (e.g.oĩov ἔειπεs) with the rarest exceptions always takes the aug. unless the form so resulting would be highly inconvenient— open antispast or tribrach—or nearly impossible— fourth paeon or closed cretic etc. The true present-aorist, as in similes or gnomes, absolutely must, as a matter of idiom, be augmented. (This last point was proved by Prof. Platt.)

^{page 108 note 1} For figures, see above: among the speeches, the most notable exception is ι, where the statistics of the augment and of the various scansional points come nearer to narrative than in the rest of the άσóλγoι or in any other Odyssean speech. This makes it plausible that ι was originally cast in the 3rd person.

^{page 108 note 2} The necessity of making the penthemimeral line diatomic would no doubt restrict the free use of the strong caesura.

^{page 109 note 1} All the statistics for Σ 368–end (speeches of course omitted) are quite unparalleled.