^{page 101 note 1} The strictness must not be exaggerated: the preserved part of the stone shows irregularities which could not have been exactly predicted. I incline to think that line 4 had 34 letters, for the reason given in note 4.

^{page 101 note 2} I do not question Tod's and Meritt's supplement in line 3 τ[α δ ε Σαλαμινι μ]ε.Hiller's τ[οcδε κλεροc μ]ε makes good sense: κλεροc need not refer specially to Klerouchoi since it is a common word for ‘property’ (cf. the titles of Isaios’ speeches; and Hesiod, , Erga 341)Google Scholar. But it is one letter short and I think it unlikely that this could be due to the interpunct just before (see n. 4); moreover, if we used the singular (as I would prefer) this would probably shorten it by. another letter still: τ[ον δε κλερο(μ)μ]ε.

^{page 101 note 3} Cf. Tod, , SGHI. 33Google Scholar, lines 23–4, Εο Κνοcιο[c] | εν Τυλιcοι Εο δε Τυλιcιοc εν κνοcοι; ib. 34, lines 1-2, τον ξενον με Εαγεν: ε ταc Χαλειδοc: τον Οιανθεαμ | εδε τον Χαλειεα: ε ταc Οιανθιδοc.

^{page 101 note 4} The six main clauses of the decree are probably separated from each other, and from the concluding formula, by six main interpuncts, in lines 3, 4, 7, 8, 10, 12. (Sc. 3-point interpuncts: the 2-point stop between [τ]ριακοντα and δρ[αχμον] in line 10 is another matter; it is probably intended to isolate the numeral, and in the draft it was perhaps written ‘:ΔΔΔ’:.) Only the first and last of these six are on the stone. In 4, 8, 10 the interpunct seems to be demanded by the δε which in each place introduces the clause following: in 7 both interpunct and δε are lost. In none of these cases does the interpunct appear to affect the count of letters, unless in the case before us, line 4. In both extant cases, however (sc. 3 and 12), it more or less upsets the stoichedon order; and in 12 it actually occupies a letter-space. Both these derangements are soon compensatedby some slight crowding. (We can observe this on the stone in 12: in 3 the derangement is much smaller and the compensation may be safely presumed. Cf. p. 101, n. 2, above.) It is possible that in line 4 the cutter felt himself too near the end of the line to compensate, and therefore lost one letter-space. It should be noted that four interpuhcts out of the six fall in lines 7–12, which are so widely spaced that compensation offered no problem at all.

^{page 102 note 1} The same unknown condition perhaps applied to Lemnos in 386 B.C.: IG. ii². 30b 4 μητεμιсθωсαι πλην[… … ], cf. ibid. line 7 [καθαπ]ερ τοιсεс Σαλαμ[ινα…]. Cf. SEG. iii. 73.

^{page 103 note 1} If this use of εсτιν be questioned, the same idea could no doubt be phrased differently. E.g. [Εα νομιμ(α) αсτοι]сι, [Εα νομοс Αθενε]сι, possibly [Εα κεται αсτοι]сι, But εсτιν seems to me clear, and to express what is incumbent in the same way as ὐπάρχειν in IG. ii². 43 A 54–5 [υ] | παρχετωμ[εν] αυτωι ατιμωι ειναι.

^{page 103 note 2} Most probably so. But conceivably τελεν is here equivalent to τελήτω, as μcθον in 4 is equivalent to μισθούτω.

^{page 104 note 1} It may be worth remarking (though very little weight should attach to it) that the current supplement restores εαν[δε με - - ] and before such a δή clause one might expect an interpunct (see p. 101, n. 4 above).

^{page 104 note 2} Line 4 has only 34 letters: the interpunct is taken as occupying a letter-space. See p. 101, n. 4 above. Alternatives for line 4 are suggested on pp. 101–2 and for line 2 in note 1, p. 103: and a more different alternative for line 2 on p. 103. In line 12 perhaps [επ]ι τεc β[ο]λε[c προεβολευθε]: for these opening and concluding formulas we may perhaps compare IvO. 30 εδοξεν Αλειοιcι… εν Ολυμπιαι εδοξεν (punctuating as I have proposed, CQ. xxxix. 23, n. 2). The alternative, [επ]ι τεc β[ο]λε[c… ειπε] might be preferred as giving more substance to the record: we should need a name of seven letters.

Additional Note to p. 102. Line 4: εα με οικ[εν εκει τ]ο[ι μιcθομενοι] with the finite verb (μιcθοcει” understood? I had assumed that after εα(μ) με the finite verb must be expressed, but Meritt or Raubitschek will soon publish a contrary instance in a new fragment of IG. i². 6 (cf. Hesp. xiv. 61 ff.).