^{page 414 note 1} He was certainly not a libertus: the cognomina disprove that idea.

^{page 415 note 1} Among the cities which boast their participation in the Koinon only Savatra has left coins of the time of Pius; several cities began to coin under Marcus, but some only in the third century (so far as known). Cana (or Kanna) has not left any coins, but was certainly one of the Koinon cities, and probably its coinage may yet be discovered among the incerta of Museums or by new purchases.

^{page 415 note 2} Isaura Palaia must have been part of Isanria, as defined about A.D. 130–5.

^{page 415 note 3} Laodiceia must, of course, go with Iconium, as part of Galatia. Parlais is not so certain: if Ptolemy be right when he puts both Misthia and Vasada in Galatia (the latter might seem doubtful, were it not confirmed by Acta S. Eustochii, see Histor. Geogr. p. 333), Parlais also must have been in Galatia. On the position see Pisidia § 18, Annual Br. Sch. Athens 1902–3 pp. 261 ff.

^{page 415 note 4} See his Epigraphic Journey pp. 226–7.

^{page 415 note 5} Kyros le Jeune en Asie Mineure p. 401 ff. The real Khan Zindjirli (Djindjirli) is the one which M. Cousin passed (p. 403) at 8 p.m., about 4 hrs. after leaving Kutu-Delik-Khan; see Cronin in J.H.S. 1902 p. 368. M. Cousin's march from Kutu-Delik to Konia was performed on foot, a remarkable feat of endurance. His times everywhere are slow; probably his hour means usually about 2½ to 3 miles (here 2½ or less). Kutu-Delik-Khan was destroyed in great part between 1898, when M. Cousin saw it, and 1904, when we examined it. His first inscription is now built into the kitchen fire-place at Ak-Bash-Khan and the letters are mostly destroyed, tie suggests that here stood the city Amandra; but Amandra was the old name of Iconium itself (according to Malalas p. 36). A village of the Iconian territory stood here: probably Salarama (J.H.S. 1902 p. 368: Expositor 1905 Oct.).

^{page 415 note 6} It is impossible to render the character of these badly formed letters by types.

^{page 416 note 1} I verified this point carefully. The text is quite certain. M. Cousin agrees.

^{page 416 note 2} M. Cousin correctly indicates this in his copy, but does not draw the right inference.

^{page 417 note 1} The termination does not permit such a meaning.

^{page 417 note 2} The gap is due to the double column arrangement, which causes several difficulties in interpretation. My reading confirms St. except that I read ϲ for ϵ.

^{page 417 note 3} On the Ormelian estates (called Hadriana), the πραγματευτα⋯ are often mentioned: in Histor. Geogr. p. 173 and Cities and Bish. of Phr. i. p. 281, I have given negotiator as the proper equivalent. Professor Pelham (footnote Cit. and Bish. loc. cit.) and Dr. Schulten Röm. Mitth. 1898, p. 225, prefer actor. Professor O. Hirschfeld, I think, has approved of negotiator, but I cannot find the reference.

^{page 418 note 1} See my Histor. Commentary on Galatians p. 211.

^{page 418 note 2} See Pisidia and the Lycaonian Frontier in the Annual Brit. Sch. Ath. 1902–3 pp. 248 f., 253 f.

^{page 418 note 3} Histor. Commentary on Gal. p. 211, Cities and Bish. of Phrygia i. pp. 10 f.

^{page 418 note 4} Annual Br. Sch. Athens 1902–3, pp. 261 f.

^{page 418 note 5} C.I.L. x. 5853, Mommsen Hermes xii p. 123: described in my paper on the ‘Permanent Attachment of Religious Veneration to Sites in Asia Minor,’ in Proceedings of the Oriental Congress in London 1892, pp. 390 f.Google Scholar

^{page 418 note 6} In Histor. Geogr. p. 410 I restored [πραγματευ]-το⋯ instead of [μιςθω]το⋯ (which also ought to have given the clue to the Imperial estates), but the former restoration requires a slave or freedman of the Emperor, whereas [μισθω]το⋯ suits the name of the free citizen which follows.

^{page 418 note 7} See § viii, opening of R. I.

^{page 418 note 8} Buresch, aus Lydien pp. 10, 41, 130.Google Scholar

^{page 420 note 1}

^{page 420 note 2} Perhaps read .

^{page 421 note 1} He omits by a slip to state that I recopied no. 369, 370; but mentions this in his Appendix p. 430. Among my additions was no. 371 (which is equivalent to R. II), also the small fragments nos. 383, 384.

^{page 421 note 2} Owing to a mistake which he explains in his Appendix, p. 428.

^{page 421 note 3} The name occurs in E. II=St. 371, and in St. 369 f., 372, also in E.V. above. I formerly accented on the supposition that ει was a mere variety for ι: now I think it was intended and marks the connexion with τεκμορε⋯ω.

^{page 421 note 4} Both R. I. and St. 366 had read ΡϵΙΟΝ: in 1886 I saw that N was a mistake of copyists, caused by a fault in the stone: ΡϵΙΟΙ is certain.

^{page 421 note 5} St. 374 is an improved copy of St. 373; two antigrapha were kept at different places, perhaps; but they differ in order of names.

^{page 421 note 6} Gissa or Gisza, a village near Ak-Sheher (Philomelion), Anderson in J.H.S. 1898, p. 113: Carian γ⋯σσα ‘stone.’

^{page 422 note 1} Another possible restoration, perhaps preferable, is given below.

^{page 422 note 2} is higher than , but they form part of one line, and are so interpreted in R. I., though numbered as two lines, 7, 8.

^{page 422 note 3} δόντος omitted, probably by a slip. The numerous faults of engraver and of composer add difficulty to the recovery of the true text: e.g. in 11, ΤΥ for ΤΟΥ.

^{page 423 note 1} On Manes see Cities and Bish. ii. p. 566. inscription is in J.H.S. 1899 p. 84 (Anderson), where Mr. Anderson shows that it belongs to an Imperial estate.

^{page 423 note 2} Possibly ΙΜΑ.

^{page 424 note 1} My copy does not give the personal names in this part; but I conjecture ΚΛΑΥΔΙΟϹ for ΚΑΝΛΚϹ (with last Κ marked doubtful) in Sterrett's copy, which he reads Κανλας. Κ and ΙϹ are often confused in difficult inscriptions.

^{page 424 note 2} ϹΥΡϹΗΝΟϹ, slip of engraver.

^{page 425 note 1} Hence I prefer to regard the Pesceniate Menneas Imenos as the one whose epitaph Anderson publishes, rather than the Oikeênian, St. 373, 30.

^{page 425 note 2} The engraving of the letters is so loose and sketchy, that it is often difficult to attain certainty, as here between Υ and Ϲ.

^{page 425 note 3} This, however, is very improbable, for the reason stated above.

^{page 425 note 4} This observation about the use of Aur. as praenomen was, I believe, used there for the first time as a proof of date. It is now abundantly justified; yet quite a number of writers since have slated it wrongly. The use of Aurelius as a nomen implies only a date after the middle of the second century; it is only the strictly non-Latin and incorrect use as praenomen that proves the date after 211.

^{page 426 note 1} The fault may be mine: my copy has ΑϹΤΙ ΒΙΑ, with note ‘room for letter between Ι and Β. but no trace.’ The fact that I saw no trace is inconclusive: I have same note in other cases, where Sterrett could read a letter: my eyes are not sensitive to faint effects, and I have always to compensate the delect by studious care and accuracy.

^{page 426 note 2} Called Yenije by Mr. Cronin in J.H.S. 1902 p. 106 (on my authority). In 1905 I thought that Genlije was the true form. Those who have tried to get the true form of Turkish village names will know how difficult it often is.

^{page 427 note 1} There is an error in that passage regarding Justinianopolia, corrected in ii. 578 and 787.

^{page 427 note 2} The copies agree.

^{page 427 note 3} So perhaps read in R. I. 57.

^{page 428 note 1} Compare e.g. Todd in Annual Brit. Sch. Ath. viii. 198.

^{page 429 note 1} On this genitive ending, see above. The accent of this non-Greek name is quite uncertain: perhaps Κωκουτ⋯ς would be better.