^{page 100 note 1} Herodian is wrong in saying αἱ δὲ παραθέσεις τούτων, ἡνίκα μὲν αὔξειν τοὺς ἀριθμοὐς δέῃ, ἐπὶ τὸ δεξιὸν μέρος γίνονται, ἡνίκα δὲ μειοῦν ἐπὶ τὸ ἔτερον. ἡ γὰρ παράθεσις ἐκεῖθεν σημαίνει ὄτι χρὴ τοῦτον τὸν ἐλάττω ἀριθμὸν ἀπ᾿ ἐκείνου τοῦπλείονος ἀφαιρεῖν (Appendix to Stephanus' Thesaurus, ed. Didot, viii. p. 345).

^{page 100 note 2} With the possible exception of Koutsopodi No. 4 (below).

^{page 101 note 1} Priscian gives and as alternative forms for 10,000, but neither occurs in any extant inscription.

^{page 101 note 2} The obol is represented by on two Attic abaci, Πρακτικά, 1884, 74Google Scholar and Δελτίον, 1888, 175Google Scholar.

^{page 101 note 3} See Keil, B., Hermes, xxvii. 643 ffGoogle Scholar.

^{page 102 note 1} But see Keil, B., Anon. Argent. 278, note 3Google Scholar.

^{page 102 note 2} Cf. Lindemann, F., Sitzungsberichte d. Akademie zu München. Math.-phys. Classe. 1896, p. 678Google Scholar.

^{page 102 note 3} Fourmont's copy has T I A A; the correction is due to Keil.

^{page 103 note 1} Michel 584 turns the numerals into their Attic equivalents and is therefore useless for our present study.

^{page 104 note 1} For further details see Keil, B., Ath. Mitt. xx. 61 ffGoogle Scholar.

^{page 104 note 2} Wrongly printed in Michel, , Recueil, 1336Google Scholar.

^{page 105 note 1} Freiherr Hiller von Gaertringen has kindly allowed me to use proof-sheets of I. G. v. 2, which is not yet published, for these inscriptions.

^{page 106 note 1} Le Bas-Foucart, ii. 341 e.

^{page 106 note 2} I owe this information to Freiherr Hiller von Gaertringen.

^{page 106 note 3} I owe my knowledge of this and the following inscription to the kindness of Prof. W. Kolbe, who has sent me proof-sheeis of I. G. v. I containing these two texts.

^{page 108 note 1} The only alternative, so far as I can see, is to regard it as a mark of punctuation, dividing the three separate amounts—1 m. 10 st., 10 st. and 5 st.—set against Hagion's name.

^{page 108 note 2} Boeckh, (C.I.G. 1570)Google Scholar dated this text ca. 71 B.C.: but see Keil, B., Hermes, xxv. 614Google Scholar.

^{page 110 note 1} For examples see B.C.H. xxiv. 463 ff.Google Scholar, xxv. 107 ff., xxvi. 5 ff., xxvii. 5 ff., etc.

^{page 111 note 1} For the dates of the Delphian archons see H. Pomtow, Pauly-Wissowa, iv. 2589 ff. The stroke over the Μ is uncertain in S.G.D.I. 1770, while in 1818 the first sign may well be a combination of and Μ (see Wescher-Foucart, No. 153).

^{page 111 note 2} The readings of I.G. ix. 1. 379–384Google Scholar must be corrected by Ath. Mitt. loc. cit. Nos. 4, 6, 7, 11, 12, 14.

^{page 112 note 1} I myself believe that it stands for six minas.

^{page 114 note 1} Michel 1352.

^{page 114 note 2} Michel 834.

^{page 114 note 3} Once .

^{page 114 note 4} Michel 402.

^{page 114 note 5} In l. 13 Μ Μ = 2 minas.

^{page 114 note 6} Michel 1341.

^{page 115 note 1} The material is not yet fully published: I base my judgment on twenty-four texts, B.C.H, viii. 283 ff.Google Scholar, x. 461 ff., xxix. 418 ff., xxxii. 5 ff., xxxv. 5 ff. The earliest member of the group is Hicks-Hill 76.

^{page 115 note 2} I cannot but wonder if the published text is accurate in this respect. The ‘Sandwich Marble’ (I.G. ii. 814Google Scholar; Ditt., Syll. ² 86: 377–4Google Scholar B.C.), discovered at Athens, uses Μ (not ΤΧΧΧΧ) to express 10,000 dr. seven times in ll. 136–9.

^{page 115 note 3} The talent and the mina were, however, still employed occasionally as units of weight, e.g. in I.G. xi. 162Google Scholar B, B.C.H. xxxv. 260Google Scholar No. 51 (ca. 190 B.C.), vi. 6 (180 B.C.), xxxiv. 143 (ca. 180 B.C.), 173 No. 40 (170 B.C.), where the talent is represented by Τ, and respectively.

^{page 116 note 1} I.G. xi. 220, 221, 224, 225Google Scholar, and ten other texts.

^{page 116 note 2} In B.C.H. xxxii. 493 No. 24Google Scholar, it apparently takes the form I.

^{page 116 note 3} The earliest cases, with the exception of a solitary example ca. 260 B.C. (I.G. xi. 214Google Scholar), are B.C.H. xxxv. 243 No. 48Google Scholar (189 B.C.), 260 No. 51 (ca. 190 B.C.)

^{page 118 note 1} The reading πόδας in J.H.S. vi. 255 No. IIGoogle Scholar, was corrected by Paton and Hicks to πόδαςιδ´ (Cos, No. 152). The Δ Δ restored by Herzog, R. in Berl. Sitzungsber. 1901, p. 473Google Scholar l. 19, was subsequently altered by the editor to Ν (Beiträge zur alten Geschichte, ii. 321Google Scholar).

^{page 118 note 2} Anc. Inscr. in the Brit. Mus. cccxliii; Michel 642.

^{page 118 note 3} The S.G.D.I. omits in error one of the four Δ's.

^{page 119 note 1} Michel 593.

^{page 119 note 2} Michel 831.

^{page 120 note 1} Michel has turned all the Chalcedonian into the corresponding Attic numerals.

^{page 121 note 1} So Waddington: possibly we should read δρα(Χμα) χΔΔΔΠ⊦ΙΙΙ. In S.G.D.I. the letters Δ Ρ Α χ are omitted in error.

^{page 123 note 1} In A l. 2 should plainly be ; in l. 3 what appears as must be a Χ and is so rendered in the transcription. The meaning of the term ἑκατώρυγος (B ll. 5 f., 15 f.) is discussed by Keil, B., Hermes, xxxviii. 140 ffGoogle Scholar.

^{page 124 note 1} It is surely impossible to take Δ Ο as a compound sign: true, Δ is also used for 10 silver drachmas, but in each case the context determines the value.

^{page 124 note 2} The engraver may have written in place of , or we might conjecture .

^{page 125 note 1} Head, B. V., Historia Numorum (1st ed.)Google Scholar, Index iii.

^{page 125 note 2} Imhoof-Blumer, , Griech. Münzen, 9Google Scholar.

^{page 125 note 3} E.g. by Gundermann, G., Die Zahlzeichen, p. 23Google Scholar.

^{page 126 note 1} Used by J. Woisin, De Graec. not. num., and others.

^{page 126 note 2} See the Appendix to the Didot edition of Stephanus' Thesaurus, viii. p. 345Google Scholar.

^{page 126 note 3} Hermes, xlix. 253Google Scholar, note.

^{page 126 note 4} Commentatio de figuris numerorum, i. 5Google Scholar, in Keil, H., Grammatici Latini, iii. 406Google Scholar.

^{page 126 note 5} The only exception known to me is G. Gundermann, op. cit. 23 ff., who argues that the similarity between the numerals and the Greek letters , Δ, etc. is accidental.

^{page 126 note 6} ἰῶτα ἕν ἐστιν. See Appendix to Stephanus' Thesaurus, ed. Didot, , viii. p. 344Google Scholar.

^{page 126 note 7} Loc. cit. τὸν μὲν οὖν ἕνα ἀριθμὸν ἓν ἰῶτα σημαίνει.

^{page 126 note 8} Griech. Epigr. 542, Handbuch, i. 416Google Scholar, and especially Bursian's, Jahresbericht, lxxxvii. 145 ffGoogle Scholar.

^{page 127 note 1} No. 3 b (above). It should be noted, however, that though · denotes 1 dr., the sign for the pure number 1 may well have been l, as at Epidaurus (No. 7b).

^{page 127 note 2} No. 8 a.

^{page 127 note 3} He definitely includes them within it (Handbuch, i. 417Google Scholar), but appears to have overlooked their peculiarity.

^{page 127 note 4} Griech. Epigr. 542, Handbuch, i. 416Google Scholar.

^{page 127 note 5} Evans, A. J., Scripta Minoa, i. 256 ffGoogle Scholar.

^{page 127 note 6} Gundermann, op. cit. 18.

^{page 128 note 1} Keil, B., Hermes, xxvii. 645Google Scholar.

^{page 128 note 2} I.G. ii. 985Google Scholar. See Larfeld, , Handbuch, i. 417, ii. 543 f.Google Scholar; Keil, B., Hermes, xxv. 319 f.Google Scholar, who confidently asserts that the introduction into Athens of the alphabetic system falls ‘some time before 50 B.C.’

^{page 129 note 1} This view is defended in Berl. Phil. Woch. x. 842 ff.Google Scholar, and was adopted by Wachsmuth, C. (Stadt Athen, ii. 12, note 1)Google Scholar. Lolling placed it later than 14 A.D., Toepffer, (Quaest. Pisistr. 21)Google Scholar at or after the close of the Roman republic.

^{page 129 note 2} Hermes, xxv. 317 ff.Google Scholar; Berl. Phil. Woch. x. 1258 fGoogle Scholar.

^{page 129 note 3} With the possible exceptions of S.G.D.I. 1770, 1818 (see No. 22c).

^{page 130 note 1} Hermes, xxv. 320Google Scholar, note. Cf. Ritschl, , Die Alexandrin. Bibliotheken 99, 100, 123Google Scholar, note.

^{page 130 note 2} Short Hist. of Gk. Maths. 40. Cf. Journ. Phil. xii. 278Google Scholar.

^{page 130 note 3} The Darius-vase (No. 62) may very well represent the same stage of development.

^{page 132 note 1} In this paragraph I do not take into account the variant forms of the same letter.

^{page 132 note 2} See Keil, B., Ath. Mitt. xx. 64, noteGoogle Scholar.