^{page 113 note 1} J.H.S. XV. (1895), p. 164.

^{page 113 note 2} Some unimportant omissions, occurring only in ABCK, have been dealt with already (C.Q.,1920, pp. 14–15).

^{page 114 note 1} See Allen, T. W. in J.H.S. XV. (1895), pp. 161–4Google Scholar.

^{page 114 note 2} For ‘Homer’ A and Q (our E and e) are among the ‘Parisini,’ while T, II, and L (our S, II, D) belong to MrAllen's, x group (J.H.S. XV. 1895, pp. 174–7 and 146 sqq.)Google Scholar.

^{page 115 note 1} Except where the contrary is stated, all manuscripts, in the passages cited below, in dicate their omissions by leaving a blank space. Politian's readings are given as in his first edition (1489).

^{page 115 note 2} Where ed. pr. agrees with II I call it D; where it follows I, or gives a reading not found in either n or II, I style it ‘Lascaris.‘

^{page 116 note 1} The initial lacuna of y here is answered by a final one in VI. 15 (ify ⁹ took -χορον there from the scholium), at an interval of 31 lines ( = 29 of text + 2 for the title of VI.). From the arche type's having had defects, at the same interval from each other, in V. 136 and VI. 23 (if it lacked ίδέσθαι), we could infer that its Pages were of 31 lines each, and that z, whose descendands are complete at V. 128 and VI. 15, was in-curred by the archetype—an inference which witz (preaf., P.15: ‘si quis stature.’ etc.). The archetype also showed correspondence in defect at an inteerval of 23 lines (IV. 177–8 and 200.i )—perhaps an inheritance from an ancestor which had had 23 lines to the page.

For 31 lines in each page of y ³, compare V. 108 with 139; and for 30 in each ofx, IV. 224–5 with V. 107.8 (209 lines of text + 1 for the title of VI. =210=30 × 7). This last case points to Iv. 224–5 stood on the recto of the first folio the last, and that the stain reached the last through this outside sheet's being at some time folded the reverse way.

^{page 117 note 1} S's habit is either to reject an incomplete word (here, VI. 13, 18, 22), or resort to conjecture word (here, VI. 17, 21, 119) or borrwing from politian (eg. V. 139); at VI. 86 he blends rejection and conjecture. x, on the other hand, cherishes fragments (VI. 13, 15, 17, 18,22, 23, 86, 119), and conjecture, though he completes a word at IV. 225.

^{page 117 note 2} Only Q Ee have this scholium, which Ee present thus: καλλχορον φρÉαρ ⋯καλεἷτο έν ⋯λενσἷι Ȅστι δέ καÌ (e omits καÌ) σ⋯ος. Q ends it with έκαλεȋ, perhaps owing to clipping on the edge of his source's page: and q, embodying Q's form of it in his text, corrects to ⋯καλεἷτο

If y ⁹ inherited only Τρìς δπ⋯ καλλι from y ², we have here for y a rement more nearly equivalent to its putative text in V. 128 (see footnote on P. 116).

^{page 118 note 1} See C.Q. 1920, p. 68.

^{page 118 note 2} Qqe alone have it. Corrected, it runs thus: τò έξ⋯ς τà δρáγματα ἓδεɩν, ἳνα κνΡ τɩς ⋯περβασίας ⋯λÉηταɩ Q must have got it from Ee's lost kinsman. As ὡς (l, 22) should be followed, as in 18–20, by naration, τà δρáγματα ἓδεɩν cannot possibly conceal any of the words lost from 23. I can only suggest that the commentator took ὡς as ὣς ‘thus ’ (or found the latter in his text). and referred it to δράγματα of 19–20 connecting πατῇσαɩ with πατεῇσθαα, ‘devour.’

^{page 121 note 1} If so, this passage is an exception to my earlier statement that interaction between the z group and the other extant MSS. does not exist.

^{page 121 note 2} We might go further and ascripe to contamination IID'sΦαμένα (V. 131) and ⋯λολυγαυς(V. 139).

^{page 121 note 3} I have already excluded θέμɩδες (V. 78); see above, p. 115).

^{page 122 note 1} If V. 138, the corresponding line of the source's intervening recto, be superimposed directly on 109, measurements taken in four photographs show that ίδέσθαɩ. (of log) lies beyond the end ofμέλεταɩ (llq), or at least beyond μέλετ- (Q) or μέλε- (B). Consequently, with the source's second folio torn, and 138 lying, of course, in verted on 109, the latter's ίδέσθαɩ may readily have shown through to VI. 23, without 138's having lost anything ad init. True, the recto lines of a manuscript often start nearer the left side of their page than do those of the verso; but the ‘even’ line 138 may have been, as the ‘evens’ in Hymn V. are in so many extant manuscripts, not flush with the adjacent ‘odds.’ Anyhow, y^4' initial integrity in 138 is derived from y^l; and an inherited defect there in the contaminator would not invalidate my position.

^{page 122 note 2} θέμɩδες (V. 78) I have explained above as a conjecture of y⁴ or the contaminator; and thus IID's advantage there is unreal. At VI. 17 II's δɩ⋯ if it arose from δηῖ (of the contaminator, I have suggested above), may be left to stand or fall with V. 128 and 136 (ad fin.).

^{page 123 note 1} See C.Q. 1920, pp. 119–121.

^{page 123 note 2} I.e. even granting that έαδò and ύπαρβασɭας are in no measure due to contamination.

^{page 123 note 3} Also, we must remember that y^2' descendants, E and e, stand in the Homeric Hymns quite apart from S {AC lack ‘Homer’) and II.D, which are there rather closely akin to each other (Allen, cited on p. 114, footnote 2).

Moreover, the contamination of y⁴ with a lost z MS. will explain why II, alone among the y group, has the poem 'ϒμν⋯ τ⋯ν ύψɭζγνλ;γιν κτλ., of which 11. 1–4 are found in some z MSS. (see C.Q. 1920, p. 105).

^{page 124 note 1} C discards the source's citation of doublets (see C.Q. 1920, pp. 12–13), adopting the suprascripts only in IV. 208, 318 of these five passages.

^{page 124 note 2} II has also three unshared doublets (III, 180

ωɩ

αι

θ

⋯κεῖνον: V. 93 àμΦοτέρησɩ: VI. 114 τότ') and two marginal variants (IV. 150. text είσόκεν but marg. είσόσα which latter = the text of ABCK ed. pr. V. 25, text ῖλαβοσα, marg. γṔ βαλοῖσα). At IV. 264 it gives αύτᾐ (αύτᾐ cett.), to suit εϊλετο (sic IIxz²), of which 'Iνωπòς was seemingly taken as the subject. At V. 73 II's εσαν results from correction; the underlying form seems to be εύσαν which ed. pr. gives (from D).

^{page 125 note 1} I must express my gratitude to those who have so kindly helped me to obtain photographsof MSS. My friend Mr. J. A. Twemlow, Associate Professor of Palaeography and Diplomatics in the University of Liverpool, took much trouble over the MSS. in Italian Libraries, for many of which Mgr. A. Ratti gave valuable advice. I am indebted for aid in connexion with S to the Librarian of the Biblioteca Nacionale, Madrid, and as regards H to Dr. P. J. Enk, of Leiden, ever generous in assisting British scholars.