^{page 102 note 1} By ‘horizontal equation’ I mean the certainty that two stones contain parts of the same horizontal line: for example, the top edge of the stele, or the prescript of a list. If such stones belong to groups which also contain part of the vertical margins (and in fact they do in all cases) their relative positions are fixed. See A.J.A. xxxiii 4–6Google Scholar.

^{page 103 note 1} See below, p. 106.

^{page 104 note 1} Yet we shall see, this could not be done without increasing the 453/2 list by 9 or 18 lines, which is not quite easy.—It is possible that the numerous επιφοραι in the Ionic List of S.E.G. V 15Google Scholar, though not actual arrears, may yet point to some irregularity in the Ionic province in the year before.

^{page 104 note 2} Lines, not names. I do not try to estimate for the cases of two names on a line, or one name in two lines, though both occur. S.E.G. V 2Google Scholar appears to avoid giving 2 lines to a name, but (with S.E.G. V 7Google Scholar added) has a few repetitions.

^{page 106 note 1} 21 letters is difficult in line 12: we should have to supply [στστερ] εσκυ[ζικενοι εβδομεκοντ]α, omitting the aspirate. 22 letters would allow ηεβδομεκοντα οΓ τετταρακοντα.

^{page 106 note 2} Nor do I know any reasonable explanation (supposing that the two portions of S.E.G. V 7Google Scholar go together) of the large uninscribed space above columns 2 and 3.

^{page 106 note 3} It is interesting to note that some, presumably of the Propontis cities, paid in gold.

^{page 106 note 4} See p. 102 note 1.

^{page 106 note 5} The left-hand column has used lateral crowding for some names, which changes the general impression; and the letters there seem to me taller than in the right-hand columns, though the vertical spacing is, of course, the same. Distinctive letters are, sloping nu, angular rho (with the top stroke rather flat), circular phi: the lists in S.E.G. V 7Google Scholar have no phi but agree in the other letters; and a curious lambda, with the diagonal slightly curved, occurs in both S.E.G. V 2 and 7Google Scholar and not (I think) elsewhere. They also both have an unusual ligature for 50 drachmai (in the quotas of Αμυνανδες Πελειαται).

^{page 107 note 1} By ‘one line’ I mean, one line on face B.

^{page 108 note 1} See p. 103. It would have 153 lines only, compared to 157 in S.E.G. V 9 and 163Google ScholarS.E.G. V 11Google Scholar.

^{page 110 note 1} Of the decisive letter (the P of φοροσ) I believe the upright stroke is still recognisable as a cut; but the loop appears merely as sound surface amongst the corrosion round it, somewhat like the lower letters in frag. 18 of I.G. I ²62Google Scholar (photograph in Meritt and West Ath. Assessment fig. 9). My photograph shews the loop more clearly than the upright.

^{page 110 note 2} See also Mr. McGregor's own statement on pp. 135–6 below.

^{page 111 note 1} Note that the spellings Λεφσιμανιοι and Σερμυλιεσ, both correct for the first period (i.e. 453/2) but incorrect for the second (i.e. 448/7), are certain in S.E.G. V 7Google Scholar (see fig. 6): although S.E.G. V gives (wrongly) the second-period forms Λεφσιμαν[δ]οι and [Σερβ]υλιεσ.

^{page 112 note 1} Ηεβδομες fits the stoichedon order.

^{page 112 note 2} The evidence of the sigma is not conclusive. I note here that the prescript of S.E.G. V 2Google Scholar has a quite certain 4-bar sigma (see the facsimile), though the list of names has the 3-bar form; and there is an exactly similar mixture from the same year in I.G. I ²20Google Scholar.

^{page 112 note 3} See Lombardo, G.Cimone (Rome, 1934) 127Google Scholar f. and 142 note 7: J.H.S. lii 223Google Scholar.

^{page 113 note 1} Line 41 in I.G. I¹. But I use my own line numbers (as attached to Pls. 15, 16, and the transcription on pp. 120–1), since some of the vertical intervals were wrongly transcribed in I.G. I¹ and consequently the line numbers are wrong in both editions of I.G.

^{page 113 note 2} The relative position of the two main fragments, which they have, in fact, occupied in the Museum since Lolling's time, was first published (I believe) in Cl.Qu. xxiv 116Google Scholar.

^{page 113 note 3} A variation of one letter, caused by irregularity of the right-hand margin, is always possible: in this case, a variation between 66 and 67.

^{page 113 note 4} Wrongly published previously as I.G. I ²49Google Scholarc. The other fragment which Wilhelm ibid, ascribes to our document does not belong: it is correctly published in I.G. I ² as 97Google Scholarc. The spacing of the letters is distinct from that of our document and identical with that of I.G. I ²97Google Scholar (identical also with 71; but, as between these two, the character of the Σ is decisive for 97). The spacing in our document is: vertical 0·0136 m., lateral 0·0105 m.: the same as in I.G. I ²98/99Google Scholar, though the letter forms here (especially the phi) are different from ours. I know no other example of this spacing. For the system of measurement, see below, p. 123, note 1.

^{page 115 note 1} This removes, I think, any lingering doubts whether the last letter space in line 40 might not have been (exceptionally) inscribed. See note 3, p. 113; and for examples of irregular right-hand margins, Pl. XI–XIII and XVII in Meritt's Ath. Fin. Doc. (In Pl. XVII there is a variation of two letters: lines 5, 7, 8, 19, are one longer, 6, 33, 34, are one shorter, than the normal, 17, 18, 20–22, 28–32, 55–62.)

^{page 117 note 1} See the photograph (fig. 7). The iota stands rather to the left, and a hard escarpment in the marble has not been properly smoothed (there is another flaw in the surface in the space of Υ in the next line): Hiller von Gaertringen has taken the escarpment for the second vertical stroke of a Γ.—There may perhaps have been an interpunct after the iota.

^{page 117 note 2} Or some part of it.

^{page 118 note 1} Possibly [τιμωρηοομαι αυτον], or some first person singular, is more suitable to an oath: the text, which I have not ventured to change, was established before the discovery of the Syme fragment, whilst it was still believed that the clause was to be added, not to an oath (ορκον) but to a bill (γυωμην).

^{page 119 note 1} It is to be administered jointly by the Horkotai and an embassy coming from Chalkis (lines 16 sqq.). This could hardly be every year; and the aorists ομοξαι (line 3) hορκοcαι (16) ομοcοcιν (19) imply that it is done once for all.

^{page 119 note 2} In that case it should probably include frag. 5.

^{page 122 note 1} But by no means all: and I fancy one of the first principles of distinction among other hands will be to separate those which are from those which are not. In the British Museum (for instance), No. 6 ( = I.G. I ²66Google Scholara) certainly is not: No. 5 ( = I.G. I ²51Google Scholar) lines 9 sqq., probably is. I note incidentally that I.G. I ²66Google Scholara has its right- and not itsleft-hand margin intact.

^{page 123 note 1} The vertical spacing is especially common; Dow points out to me that it is quite probably one Attic dactyl (0·0184 m., if the Attic foot was 0·294 m.): and the combined vertical and lateral is not uncommon. I measure, vertically, from base-line to base-line, and where possible 10 letters at once, and then divide by 10: laterally, from centre to centre, also over 10 letters if possible. This reveals certain variations (0·0178 to 0·0187 m., vertically: 0·0130 to 0·0132 m. laterally).

^{page 123 note 2} In I.G. I² they are duly recorded as standing above ΟΙ (since one more letter is supplied in each line). The two traces over the εν are also correctly (though not quite completely) recorded; but four spaces are allowed between them and επεστατε instead of three (i.e. the spaces over ονμ). This slight error (which results in allowing 46 letters to line 3) can be easily corrected from the photograph in fig. 11.

^{page 125 note 1} The other decrees which can be ascribed to Perikles are, the Congress decree of c. 449 (Plut. Per. 17), the decree of possibly the same date (J.H.S. lii 223Google Scholar) or else 431/0, touching the disposal of 5000 Talents (Anon. Arg. lines 5–8, quoted I.G. I ² p. 287Google Scholar), the Franchise decree of 451/0 (᾿Αθ.Πολ xxvi 3), and perhaps the Megarian decree (Arist, . Ach. 532Google Scholar).

^{page 126 note 1} Hiller von Gaertringen is certainly right to read τα[δ]εδου[ενα]: neither τα [δ]εδογ|[μενα] nor τα[τ]ε νομ[ιζομενα] is possible.

^{page 126 note 2} Kahrstedt has rightly seen that hοσ in line 10 must be accusative plural (cf. αυτοισι in 11), and proposes some part of ἀνελεῖν, Staatsgebiet und Staatsangehörige in Athen (1934) 335Google Scholar. He appears to accept the rest of Hiller von Gaertringen's restorations, though I do not know how he gets out of them the meaning which he does.

^{page 126 note 3} I believe a trace of this Ν can be read on the stone.

^{page 126 note 4} Νῦν with aorist, cf. Iliad III 439Google Scholar; and often.

^{page 127 note 1} Casualties [εν Σιν]οπει I cannot illustrate it, and I know of no photograph. The letters are so magnificent that a squeeze of it should be in every squeeze library.

^{page 127 note 2} For possible types of postponement, see Denniston, Gk. Particles 188Google Scholar. I see no way of avoiding this, with Kahrstedt's 36-letter line. The asyndeta in lines 4 and 12 are not very nice: nor the ace. and inf. after εχcεcτο in line 12 (followed by a change of subject); and is not the αυτομολοσ allowed (by implication) great impunity? An αυτομολοσ is ενοχος τοις μεγιστοισ anyhow: the special sanction here must be against someone who abets him. Objection might be taken to εσιεναι Ελευσινι in 10–11 (see p. 84) and 12–13, and to Ελευσινιος δ εναι hοστισ in 8-9. No doubt some of these could be avoided: e.g. we might write Ελευσινιος δ εστο for the last (see p. 81); but line 10 is past mending.

^{page 127 note 3} [Ηοπο hεκαστος ετυγχανε π] ρο το πολεμο οικον in line 9, καθαπερ προ το [πολεμο] in 11/12.

^{page 130 note 1} For a further example of this irregularity, compare the placing of the letters ΤΟΣ in lines 3 and 7. Note too the rather steep Χ in line 5: cf. line 13 of the large fragment, and line 2 of fig. 11. Many, indeed most, of these observations I owe to Mr. Meiggs in Oxford and Mr. Dow in Athens.

^{page 131 note 1} See p. 102 note 1. The equation looks good when the stones are put in their places: it brings the fractures (at the top of 61 and foot of 169) in a good line with each other.

^{page 132 note 1} There is an extra letter in I.G. I ²77Google Scholar (= fig. 11) line 12. I think there has been an erasure here (the surface is badly gone, and was gone before ΜΟΙ was inscribed). What was written before? Just without Η ? It was a very elaborate correction for so slight a matter (the crowding begins with Ι of ).