T HE inscriptions turned up by the Italian excavations at Cyrene have given us a new picture of the dialect, making the accounts in the earlier handbooks quite obsolete. The only comprehensive presentations based on the new material are those of Devoto, Riv. fil., LVI (1928), 365 ff., and (with the dialects of Thera and Melos) Kieckers in Thumb-Kieckers, (Griechische Dialekte, I (1932), 170 ff. The excuse for adding another is that several important matters merit further discussion, and especially that in the citation of inscriptions the cumbersome references to scattered articles in journals and proceedings may now be replaced by references to the Supplementum epigraphicum Graecuin, Vol. IX, fase. 1, where the new material is conveniently assembled. So in the following I shall give simply the numbers, e.g., 72. 85 = Suppl. ep. Graec., Vol. IX, fasc. 1, No. 72, line 85. Numbers in bold-faced type refer to the sections of my Greek Dialects, where the corresponding phenomenon is discussed and to which the Cyrenean evidence should be added in any future revision. In comparing the dialects of Cyrene and Thera, one must bear in mind the different status of the material. From Thera we have a great number of short archaic inscriptions, a few short ones from the fourth century B.C., then only from the period after KOLV1 influence had begun. So the long Will of Epicteta, ca. 200 B.C., preserves most of the native forms-in fact, is our principal evidence for thembut also shows many KOLV' forms, as the hybrid EC' Ka (so always) vs. Dor. al Ka; the numerals in -KO'LOl VS. -KaTlOl, -XlXlol vs. -XwXtot, Mei-a vs. earlier -reba. From Cyrene, on the other hand, we have few archaic inscriptions, but three fourthcentury inscriptions of considerable length (Nos. 2, 3, 72, of 60, 51, 141 lines respectively) and in the pure dialect, except that No. 3 has several KOLV7 forms, as ol' beside Tol, ace. pl. -ovs beside -os, and once infin. -vat beside -,uc-v. The edicts of the Ptolemies and the Roman emperors are, of course, in the KOLV', e.g., Nos. 1, 7, 8, etc., and so the letter and edict of Ptolemy X and Cleopatra in No. 5. 46 ff., while the preceding decree of Cyrene in their honor is in the dialect with some KOLV' mixture. The decree in honor of Barcaeus, priest of Augustus (No. 4, 16/5 B.C.) is an almost faithful representation of the native dialect, with all its important characteristics, but with the wellknown late confusion of w and wl in the gen. and dat. sg. (38b) and the intrusion of a few KOLV' forms and expressions. There are numerous lists of officials, priests, soldiers, market prices, as well as dedications, honorary decrees, and epitaphs in the dialect, but mostly only confirming what is known from the longer inscriptions. The passage 3. 23 if., the "oath of the colonists," is taken by some as really