^{page 81 note 1} The obligations expressed to Prof. Ramsay in Part I. have to be repeated here.

^{page 81 note 2} This is the more striking when we consider that Prof. Ramsay had explored only the fringes of this district.

^{page 81 note 3} We may take as an example the case of Mysotimolos. Buresch at first rejected Ramsay, view in the strongest terms, but recently he has come to regard it as ‘auf richtigem Wege’ (Reiseber., p. 103Google Scholar = Aus Lydien,āp. 201): only he would place the town nearer Blaundos, while Ramsay in the meantime has also found reason to bring it nearer Blaundos (CB. vol. ii.). Again, in regard to Apollonos Hieron their earliest views differed greatly, while they are now (as we shall see) practically agreed.

^{page 82 note 1} The Seven Churches, p. 227.

^{page 82 note 2} Researches i. p. 526.

^{page 82 note 3} Not Dede keui, as in CB. i. p. 194.

^{page 82 note 4} At Geune I heard of the previous visit of an archaeologist whom I inferred to be Dr. Buresch, but unfortunately I had not seen his Reisebericht, which appeared in Ber. der Kgl. Sächs. Ges. d. Wiss., 1894, p. 88–128, and has been reprinted in the full account of his researches, Aus Lydien, epigraphisch-geographische Reisefrüchte hinterlassen von Karl Buresch (published just as this paper was going to press). I have added references to the latter.

^{page 82 note 5} Aber ich glaube versichern zu dürfen, dass sie wie ungezählte andre vom Erdboden verschwanden ist, p. 108 (= Aus Lydien, p. 205).

^{page 83 note 1} Some evidence might be found if all the blocks could be turned over and examined: but our efforts were unrewarded.

^{page 83 note 2} In Anbetracht ihrer Bedeutung bin ich … geneigt in ihr Sala zu erkennen (Reiseb. p. 108 = Aus Lyd. p. 205).

^{page 83 note 3} The Epirote name ΜΟΛΟƵΟΣ occurs on its coins (cp. Part I. p. 408). Hamilton could find no traces of fortification (II. p. 371).

^{page 83 note 4} Cp. Hist. Geog. p. 124, CB. I. pp. 194, 197, etc.

^{page 83 note 5} Nic. Chon. p. 539 ed. Bonn: which means that Frederick marched through its territory (not necessarily through the town).

^{page 83 note 6} Not. X. 232 and XIII. 92, ed. Parthey and Pinder.

^{page 83 note 7} Tomaschek, , Zur hist. Topogr. von Klein Asien im Mittelalter, p. 98Google Scholar. He falls into error, however, when he says ‘der antike Name von Ἀετὸς war Ἀπόλλως ἱερὸν (Not. ep.),’ misunderstanding the meaning of ἤτοι Aïdos is now corruptly pronounced Aïdas. Tomaschek';s identification of Aetos with Aïdos is accepted both by Prof. Ramsay and by Prof. Radet.

^{page 84 note 1} So far, of course, as one may judge from a small fragment.

^{page 84 note 2} So called in Kiepert's map, but, I think, not rightly (though I am not sure of the correct name).

^{page 84 note 3} We followed this route in going from Aïdos to Baharlar: it is an easy road.

^{page 84 note 4} Or Assar.

^{page 84 note 5} I have corrected the oversight vom Hermosthale nach der Maiandrosebene.

^{page 84 note 6} Reisebericht, p. 111 (= Aus Lydien, p. 207).

^{page 85 note 1} My note says, ‘a fine space on the top.’

^{page 85 note 2} We may add (1) traces of steps up one side (2) quite near the foot remains of building in much better style than those on the summit. The cemeteries of the village contain a very few old stones.

^{page 85 note 3} Τραλεῖς Inschr. Perg. no. 13: Hesych. s.v. Τραλλεῖς (cp. Tomaschek, Die alten Thraker ii. p. 44Google Scholar).

^{page 85 note 4} His map (vol. i.) rightly indioates Tralla somewhere in the Devrent Boghaz (in accordance with the view expressed on p. 180).

^{page 86 note 1} He ingeniously explains S[ocra]tu as a Latin corruption of ʼς χωρίον Ἄτυος through the form s-cor-atu. The reading is not certain, as may be seen from the photographic reproduction (Vienna, 1888).

^{page 86 note 2} So on a coin (CB. p. 196 n. 3); Cogamus in Pliny.

^{page 87 note 1} Radet, , B.C.H. 1891, p. 373 ff.Google Scholar; cp. CB., i. p. 199 f. Buresch, , Reisebericht, p. 112Google Scholar f. (Aus. Lyd. p. 208) restores the first three lines.

^{page 87 note 2} Athenaeum, Oct. 23, 1897, p. 566.

^{page 88 note 1} This nearly agrees with Buresch: ‘4 um fangreiche (als ΜπΗ) oder zur Noth 4 Buchstaben von Durchschnittsbreite und 2 schmale (d. h. Ι) verschlungen haben. My views were formed at Baharlar before I knew of his paper.

^{page 88 note 2} Allowing for two narrow letters (like Ι or Ρ) and considering that the letters are not always of quite the same breadth, nor equally spaced, I said in Athenaeum that ‘the space cannot contain above seven letters’: but the above statement is more exact.

^{page 88 note 3} Eusebius, , Hist. Eccl. v. 16Google Scholar (reference to Montanus).

^{page 88 note 4} I made a note to the effect that the letter before Β is apparently Λ, although it might possibly have been Α (considering the way in which parts of letters have disappeared).

^{page 88 note 5} But I did not think they extended nearly so far as Bahadyr keui.

^{page 89 note 1} We cannot therefore approve Buresch's idea (quoted above, under Aetos) that Kallat. might be placed at Devrent Kalessi. Speaking of the site near Baharlar, he asks Könnte man nun etwa diese für die Stelle des herodoteischen Καλλάτηβος, der unumgänglichen Wegestation, ansprechen? Gewiss ist, dass man jener Stätte schon weit eher als Ine Gjöl die Eigenschaften eines als ‘unumgänglich’ gekennzeichneten Punktes zusprechen könnte (p. 117=Aus. Lyd. 212). This entirely misses the point of Herod's words, which simply state the fact that the only available road was that on which Kallat. lay. The Baharlar site is, therefore, not a whit more ‘unumgänglich’ than Ine Giöl or any other point on the road.

^{page 89 note 2} For the loss of the a, see CB. p. 195 note 1. Compare also Bulawôdin or Bulawâdin, the modem name of Polybôtos. The sense in which a modern town may be said to represent an ancient one is clearly defined in Hist. Geog. p. 83.

^{page 91 note 1} The lettera have apices, and in l. 2 the stone has ἑατοῦ.

^{page 91 note 2} Cp. ἐν τẽς ἀμπέλοις, J.H.S. 1887, p. 393.

^{page 91 note 3} Compare ἰσελθῖν at Eumeneia, acc. to Prof.Ramsay, reading, CB. ii. p. 525, no. 367Google Scholar, where Paris, M. read εἰσ[βα]λεῖν (B.C.H. 1884, p. 251Google Scholar).

^{page 92 note 1} That the name (Hierocles) which M. Radet places at Basmaktchi (map in En Phrygie) is a corrupt form of as is of was shown long ago by Prof. Ramsay.

^{page 93 note 1} Yet it may be held that this fact is not inconsistent with the want of a real πόλις; for though the relation of the villages in a κωμόπολις is an unsolved problem, we must infer that one κώμη was more important than the others and formed a sort of centre.

^{page 93 note 2} It has been carried from one of the villages in the plain, no great distance. See the map in CB. vol. i.

^{page 94 note 1} Another kalè was reported further to S.E. (nearer Dinêr-Apameia).

^{page 94 note 2} See CB. p. 227 f., 347 etc.

^{page 95 note 1} The koppa is slightly blurred on the stone, but it seems quite certain.

^{page 95 note 2} Inscr. of Apollonia C.I.G. 3969 ff.; Le Bas-Wadd. 1192 ff.; Sterrett, , W.E. 517–554Google Scholar; B.C.H. xvii. pp. 255–59.

^{page 96 note 1} C.I.G. 3969 etc.; Sterrett no. 517: Hist. Geog. p. 172.

^{page 96 note 2} From Antoninus Pius to Gallienus (Head, Hist. Num.).

^{page 96 note 3} Cp. Wadd. on 1195; G. F. Hill, B. M. Catalogue of Lycia etc., p. civ.

^{page 96 note 4} As Prof. Hirschfeld, G. thought (Gött. Gel. Anz. 1888, p. 592)Google Scholar: so Ramsay, , Hist. Geog., p. 400Google Scholar, who shows that means (as given in an inscr. ib. 172).

^{page 96 note 5} Cp., for example, Sterr. W.E. no. 352, E.J. 97 etc.

^{page 96 note 6} Possibly is not an engraver's error but a form really nearer the original.

^{page 96 note 7} Skr. bala, ‘power,’ balin ‘powerful’: cp. Tomaschek, , Die alien Thraker ii. p. 41Google Scholar, (in Sitzungsber. Wien. Akad. 1894), where this is given as one of several explanations of the name ‘Βαλίας oder Βαλιός’ which ‘soll Dionysos bei den Thraken geheissen haben EM.’ Gaisford's edition of Et. Mag., however, gives neither Βαλίας nor Βαλιός, but only Βάλιν (without any variant), which is clearly the same as Phryg. βαλήν.

^{page 97 note 1} Die griech, götterbeinamen in Bezzen. Beiträge xx. p. 160.

^{page 97 note 2} Wadd.'s reading is the correct one: in l. 4 I read (complete).

^{page 100 note 1} ll.3–4 in Sterrett are wrongly spaced. Read

^{page 100 note 2} My copy has in l. 8 ΓΟΝΩΝΑΞΙΑ ΤΡΤΟΓ; but καὶ is necessary, whether actually on the stone or not. The inscription lies on its side at the very top of the Kalè wall, and I read it with a glass from the top of the wall opposite.

^{page 100 note 3} In B.C.H. no. 36, l. 6–7, read and in l. 9 In no. 35, l. 8, I seemed to see etc.

^{page 101 note 1} Strabo, p. 663.

^{page 101 note 2} Strabo, p. 577,

^{page 101 note 3} J. H. S. 1887, p. 481.

^{page 101 note 4} Hist. Geogr. p. 170.

^{page 102 note 1} En Phrygie (1895), p. 123.

^{page 102 note 2} From this point the village is perhaps about an hour distant. Uzun Bunar seems to be M. Radet's Fontaine (see his route-map); but if so, it does not seem to be correctly placed, as the above description shows. Cf. Ramsay, , CB. p. 752Google Scholarn. 4.

^{page 102 note 3} Here is the itinerary from Baljik-Hissar to this point. It will show the erroneousness of M. Radet's description. [It will be noted that this route cannot pass Alaka.] At the village the aneroid read 4,700 ft. The road winds by a fine curve round the Kalè and then over the ridge.

9.33 A.M. Leave the Kalè.

9.42 „ Brow of the ridge (5,200 ft.). Reach the summit (5,300) at

9.49 „ Begin to descend towards the broad ravine, opposite the mouth of which lies Atli-Hissar. The descent is rather steep towards the foot, which is reached (4,800) at

10.8 „ Cross the ravine at its extreme limit, and begin the ascent over the next ridge, reaching the top (5,200) at

10.32 „ Thence an easy descent, reaching the foot (5,100) at

10.52 „ Ascend again till

11.5 „ (summit 5,350). Thence a winding, undulating path, and a final descent to the Uzun Dere, which is reached at

11.25 „ Level is 5,200. Here we join the Alaka-Ginik road. At 11.5 we were still passing fields belonging to Baljik-Hissar.

^{page 103 note 1} On the rounded hill above the village, round which the road runs, a very few traces of the old settlement remain. The natives call it the ‘Castle’ (Kalè). The highest point of the kalè is 375 ft. above the village.

^{page 103 note 2} Drexler, , Num. Zft. 1889 p. 177Google Scholar; Imhoof-Blumer, , Griech. Münzen in Abhandl. d. Bayer. Akad. d. Wiss. (I. Cl.) 1890 p. 748Google Scholar.

^{page 103 note 3} So it is rightly explained by Ramsay, in Hist. Geog. p. 36Google Scholar note.

^{page 103 note 4} It may be Melissa (see below).

^{page 103 note 5} Sterr. W.E. no. 366, 20 and 376, 33, 42, etc. See Hirschfeld, G., Gött. Gelehrte Anzeigen 1888Google Scholar, followed by Ramsay, H. G. p. 409Google Scholar. We cannot at all accept M. Radet'a explanation that ce dème [de Synnada] s' appelait Thynnara et qu'il honorait le héros Thynnaros comme éponyme (p. 123). [I have since come across an excellent parallel in a metrical inscription of Temenothyrai, where the inhabitants of that city are called i.e. descendants of the hero Temenos (mentioned on their coins), as the late Dr. Buresch rightly explains it, quoting as a parallel (Aus Lydien, p. 164)].

^{page 103 note 6} The road is still used by the villagers of Baljik-Hissar in going to Tchul Ova (Camp. Metropol.). At some future time the boundary-stone mentioned in CB. ii. p. 752 will be discovered to give us a fixed point, but the peasants will have to re-discover it first. Only one man seems to have seen it: he guided us to the spot,—but it could not be found!

^{page 104 note 1} The I perhaps belongs only to the word PHRYGIA written across this route.

^{page 104 note 2} Whence it follows that ‘the sum of separate distances is decidedly greater than the whole ength of the road,’ J.H.S. 1887 p. 463, where the principle is stated.

^{page 104 note 3} They are much worn, but one is the head of an ox.

^{page 104 note 4} For the form see no. 36 above. In Sarre no. 8 is exceedingly common in such Byzantine inscriptions: read also with a twirl].

^{page 104 note 5} So identified by Ramsay, , Hist. Geog. pp. 142Google Scholar and 171 and Radet, Map in En Phrygie (where it is placed at Kara-dilli).

^{page 105 note 1} Excepting the corruption Σικνόδον (κ for β).

^{page 105 note 2} Not. Basilii [=Not. i. in Parthey-Pinder] and Not. Leonis edited by Gelzer (Georg. Cyprii Descriptio Orb. Rom.). De Boor's, Not. in Zft. f. Kirchengesch. xii. p. 528Google Scholar.

^{page 105 note 3} The intrusion of Meros is probably to be explained by the circumstance that it was formerly subject to Prymnessos (see Part i. p. 424).

^{page 105 note 4} ‘Boudeia and Phyteia are perhaps other forms of the name Beudos Vetus’ (H. G. p. 143). This is clearly right.

^{page 105 note 5} Inscriptions are published by Ramsay, , B.C.H. 1883, p. 299Google Scholar, by Legrand, MM. and Chamonard, , B.C.H. 1893, p. 289–90Google Scholar, and by Radet, MM. and Ouvré, , B.C.H. 1896, p. 109 fGoogle Scholar.

^{page 105 note 6} We might compare, for example, Hierocles' order Sinethandos, Laodikeia Katakekaumenè, Tyriaïon; or Homonadeis, Ilistra (Ilisra), Laranda (Karaman), Derbe (Gudelissin). A protest must be made against the exaggerated stress that is often laid on the precise order of names in Hierocles (or the Notitt.) We have a good specimen in Radet's, M. reasoning about Κλῆρος Ὀρίνης, Κλῆρος Πολιτικῆς, Δ༵βαλικία, Λυσίας, Σύναδα (Hier. 677, 3–7)Google Scholar. Synnada is fixed. A town ‘Oriné’ is placed at Tchukurdja in a mountainous district (on evidence for a criticism of which it is sufficient to refer to CB. ii. pp. 635, 687, etc.), whence it follows that Κλῆρ Πολιτ. ‘le Domaine Urbain’ designates the neighbouring plain Kutchuk Sitchanli Ova. Le classement du Synecdème exige qu'on y place non seulment Κλῆρ. Πολ, mais encore ľèvêché qui lui fait suite dans le catalogue Δ༵βαλικία [= Sibidounda]…qui se place tout naturellement à Karadirek, dans ľangle oriental du Kutchuk-Sitchanly-Ova. Lysias, qui vient ensuite, tombe à É fek-keui [Effe Sultan]. The topography of Asia Minor would soon be settled, if such reasoning were admitted!

^{page 106 note 1} I examined the stone on two occasions. CIB is certain (the first letter is not so rounded as the C above and may be a mis-cut for S): OS is also certain. There are remains of a letter below B (in the line above), which on the first examination I took for a badly-formed I, but the second time it seemed to be the tail of an S. There is room for other four letters, for the space between LIB and PROC is greater than Prof. Ramsay's copy shows. SIB[INDEN]OS seems quite probable.

^{page 106 note 2} Which agrees with the insignificant character of the ruins there.

^{page 106 note 3} CB. p. 755 NOTE 1; Imhoof-Blumer, , Monn. grecq. p. 345Google Scholar; Reisen in Lykien etc. ii. p. 168 ff. (reliefs): G. F. Hill, B.M. Catal. of Lycia, Pamph., and Pisid. p. lvii.

^{page 106 note 4} For example, coins of Ariassos, Andeda, etc. A cursory glance at Hierocles'list of towns in Pamphylia might lead the reader to think that Sibidounda is really to be found there under the form But if he takes the trouble to compare the Notitiae, he will see that corresponds to or i.e. Andeda.

^{page 106 note 5} Here is the whole passage:

^{page 106 note 6} May there not be something more than a mere coincidence in the occurrence of the name Baljik-Hissar ‘Honey Castle’ at a site which, as we have seen, may be Μίλισσα? The suggestion (which occurred to me independently) was made in Hist. Geog. p. 36 n. I do not, however, mean that the Greek name was translated by the Turks. Radet, M. asserts (En Phryg. 123Google Scholar) that rien n'est plus fréquent, dans ľonomastique de ľ Anatolie, que ce remplacement du terme grec ancien par un mot turc offrant une signification identique (La Lydie etc. p. 36 n. 2): but he shows no proof. There is some doubt as to whether the real name is Baljik or Baghtche (‘garden’) Hissar [or Assar]. From the rude pronunciation of the peasants which runs the words together, it is impossible to decide. If you ask them which is the correct form, they answer ‘The two are one’! One of the Hodjas assured me that the former was more correct, but they do not seem really to know. W. M. R. was corrected for saying Baljik and assured that Baghtche was the real name.

^{page 107 note 1} The name Lysias is connected with Seleucid history: Lysias, a general of Seleucus Nikator in 216 B.C. (Polyaen. iv. 9, 5), may have been the founder of the city.

^{page 107 note 2} Cp. also J.H.S. 1887, p. 497.

^{page 107 note 3} See CB. ch. xvii. p. 714; Church in R.E. p. 436 n., ‘At the source of a stream among the mountains between Synnada and Hieropolis was a place called Gonyklisia—i.e. where the early rite γονάτων κλίσις was held. This remote place was clearly a secret meeting-place; and after the meetings had ceased, and the archaic term was no longer understood, a foolish legend grew up to explain the name; see Expositor 1889, p. 262.’

^{page 107 note 4} The name itself is sufficient proof, for it states in so many words that this is a ‘non-Turkish fountain’: compare, for example, Giaour Ören, the name of the ruins of Trajanopolis. The Turkish number which is scratched on one of the fallen blocks is a mere recent graffito: I saw it again on a teheshme at Utch Eyuk, in the valley of the Tembrogios. It means 1871 A.D.

^{page 108 note 1} Sterrett W.E. no. 366, 33 and 46 (=J.H.S. 1883, p. 23 ff.); no. 374, 2 and 15, no. 378, 5 and 9; no. 366, 32.

^{page 109 note 1} τὸν Κηναβορίου Not. vii. 170, ix. 353; 6 ὁ Κινναβωρίου viii. 443; ὁ Κιναβωρίου Not. Basil. 389 (ed. Gelzer); ὁ Κινν༵βορίου Not. De Boor.

^{page 109 note 2} A site ‘perhaps near Geneli’ is proposed in J.H.S., but the few remains that the village contains have probably been carried.

^{page 109 note 3} For the significance of this fact cp. above, Part I, Vol. xvii. p. 400.

^{page 109 note 4} Einleitung in d. Gesch d. Griech. Spr. p. 307. He wrongly calls it Κινναβορα, not observing the occurrence of the city name in the Notitiae.

^{page 109 note 5} Cp. Kinnaborion (above).

^{page 109 note 6} The basement of the fine mosque-minaret is composed entirely of old blocks re-faced: a small ‘door-stone’ may be seen built in near the top.

^{page 110 note 1} Some inscr. from the north side of Paroreios (along Emir Dagh) are published by Hogarth, , J.H.S. 1890, p. 158Google Scholar ff.

^{page 110 note 2} The site at Karadja-ören, however, was probably the chief quarry for the buildings at Tchai.

^{page 110 note 3} The Peut. Tab. gives no help, as one of its two numbers is wrong.

^{page 110 note 4} The water in the plain is not good for drinking and there is no other copious stream flowing down from the Sultan Dagh nearer than Deresinek, a village lying up a glen far off the road.

^{page 110 note 5} Synnada, Prymnessos (Kara Hissar,) and others might be added.

^{page 110 note 6} Amongst these is a large building of Byzantine work. At Tchai, apart from the Seljuk ruins, there are only some Byzantine blocks in fountains, etc.

^{page 111 note 1} Inscr. of Philom., and vicinity, C.I.G. 3982 ff.Google Scholar; Le Bas-Wadd. 1704–5; Sterrett E.J. No. 155 and perhaps 163–4; Ramsay, in Kuhn's Zft. f. vergl. Sprachf. viii. p. 391 (Phrygian)Google Scholar; Heberdey-Wilhelm, , Reisen p. 163Google Scholar, no. 271. In C. I. G. 3984 (= Wadd. 1704) Hamilton's copy of ll. 7–11 is correct and is wrongly altered by the editors. The inscr. was apparently not completed.

^{page 113 note 1} Steph. Byz. s.v. Monogissa, . Cp. Hist. Geogr. p. 412Google Scholar.

^{page 114 note 1} So called, at least in its lower course, from a Circassian village on its banks.

^{page 114 note 2} I enquired particularly at different places about the course of these streams and all accounts agreed in saying that the Doghan Assar stream flows past Arkut Khan into Ilghin Lake, while the Ayaslar-Urus-Yendin (miscalled Kendil) stream falls into Balki Deressi a little to S. of Ilghin. We crossed the latter about three-quarters of an hour after leaving Ilghin. There is no stream from Doghan Assar joining the Ilan Yusuf Tchai at Kotchash (so far as I heard or could see).

^{page 114 note 3} There is also a hill-path from Kunderaz, described by Hogarth, l.c. p, 153 f.

^{page 115 note 1} They were turned up by a villager while ploughing his field behind the village: the place was shown us by this man's brother.

^{page 116 note 1} Anatolian pronunciation tends to convert g between two vowels into a y sound and finally to let it drop altogether: cp. Tchigil, now pronounced Tchiyil (below). Regiz is a different village from Egrigioz, which is quite close to Ak Sheher.

^{page 116 note 2} The correct reading is

^{page 116 note 3} The correction Silindo-kome, there suggested in a footnote in place of Silindiconense, is proved to be right by comparison of the Greek original, published by Joannes Theophili in his

^{page 116 note 4} J.H.S. l.c., Ath. Mitth. 1889, p. 180–1.

Xen., Anab. i. 2Google Scholar: (from Kaystron Pedion)

^{page 117 note 1} The lettering is poor and late.

^{page 117 note 2} ‘Agha-üt, , Herrenpassage’ Tomaschek p. 103Google Scholar. Gaita is also the name of a village near Nicaea, , H.G. p. 201Google Scholar.

^{page 117 note 3} There is hardly space for more than one letter. The last three words are under the relief.

^{page 118 note 1} So Ἐνᾶς for Νᾶς, etc.

^{page 119 note 1} Cp. CB. no. 91, p. 269, where is rightly read, Rev. Univ. Midi, 1895, p. 362.

^{page 120 note 1} The ordinary road keeps along the Dere.

^{page 120 note 2} Hist. Geog. p. 359.

^{page 120 note 3} Mich. Glykas p. 528.

^{page 120 note 4} Tomaschek, op. cit. p. 103.

^{page 121 note 1} Cp. Nicetas, p. 72, ed. Bonn.

^{page 121 note 2} p. 46.

^{page 121 note 3} p. 47.

^{page 121 note 4} Cp. Hist. Geog. p. 359 note +.

^{page 121 note 5} In no. 13, 1. 2 fin. I read ΓΑΙΟΥ; no. 14, the epigraphic text is correct. One inscr. of Tyriaïon in Heberdey-Wilhelm, , Reisen, p. 162Google Scholar, no. 270.

^{page 122 note 1} Except the two fragments added by Sterr, . E.J. 174Google Scholar (see above) and 186; three added by Hogarth l.c. nos. 1·3; and one by Legrand, MM. and Chamonard, , B.C.H. 1893, p. 289Google Scholar (better in B.C.H. 1896, p. 111): a copy of this inscription which I made in 1896 reads ΑΔΔΑΚΕ ΤΟΡΔΕΩ⊏ etc., the last letter being slightly blurred in the inside. In B.C.H. 1896 l.c. MM. Radet and Ouvré collect a few of those previously published, in ignorance, apparently, of the articles quoted above.

^{page 127 note 1} Cp. CB. ii. p. 750–1.