^{page 49 note 1} We are glad, however, to be able to say that by walking warily and using all tact, and especially by the assistance of H.B.M. Consul at Angora (Mr. H. S. Shipley), we met with uniform courtesy from the local governors. Perhaps also we owed something to the political mood of the Turks, which is very variable; but a great deal always depends on the temper of the Vali Pasha and his individual subordinates in the provinces.

^{page 50 note 1} They are published in J.H.S. 1898, Pt. ii. p. 342 ff.

^{page 50 note 2} Even in the recent map published by Major von Diest in Petermann's Mittheilungen, Ergänzungsheft, No. 125, this village is still placed on the south of the river. It is strange that such a mistake should be made when the survey of the Chemin defer Ottoman if Anatolie was used by the compiler of the map. (Probably the survey is not very careful.)

^{page 50 note 3} So named on Von Diest's map.

^{page 51 note 1} Mansi, , Sacr. concil. nova coll. vi. pp. 57, 94, 1085Google Scholar; vii. pp. 41, 158.

^{page 51 note 2} Op. cit. vi. p. 1172; vii. p. 125.

^{page 51 note 3} Op. cit. vii. p. 437.

^{page 51 note 4} vii. p. 407.

^{page 51 note 5} Hist. Geog. p. 143, No. 75; J.H.S. 1887, p. 492.

^{page 51 note 6} CB. ii. p. 753. I use the abbreviation CB to denote ProfRamsay's, work, The Cities and Bishoprics of Phrygia (Oxford, 1895–1897).Google Scholar

^{page 51 note 7} On the office of tabularius (or accountant) which existed side by side with that of dispensator in the various bureaux for the administration of the Imperial patrimonium (res privata), compare Rostowzew, M. in Römische Mittheil. 1898, p. 111 ff.Google Scholar

^{page 52 note 1} CB. ii. p. 499.

^{page 52 note 2} This cemetery stands by the side of the great road from East Phrygia to Kotiaion and the coast.

^{page 52 note 3} The higher slopes of Murad Dagh, along which passes the direct road from Kara Hissar to Gediz (Kadoi), were explored by W. von Diest in 1886 (see Petermann's Mittheil., Ergänzungsheft, No. 94); and we did not visit this district again.

^{page 53 note 1} The altitudes in this district are reckoned by aneroid, which we compared at Tchigurler with the altitude assigned by the Railway survey. They are not likely, therefore, to be far wrong.

^{page 55 note 1} There is a longer road, to S. of this path, just practicable for an araba.

^{page 55 note 2} x, 32, 3.

^{page 55 note 3} viii. 4, 3. The older and more strictly correct form of the name of the Phrygian city was doubtless, as Buresch remarks, 'Αζανοι which is used by Strabo, Ptolemy, Hermogenes apud Steph. Byz, Hierocles, and the episcopal lists. The explanation of the name quoted by Steph. (see Hist. Geog. p. 147) looks like a piece of popular etymology, and the story doubtless originated with the priest-kings who in early times ruled the city and perhaps bore the name Euphorbidai, like the ᾿Ανθεάδαι at Halicarnassus, etc.

^{page 55 note 4} Aus Lydien, p. 159.

^{page 56 note 1} The stream which appears in the photograph is merely a mill-stream.

^{page 57 note 1} Arch.-epigr. Mitth. aus Österreich, xviii. (1895) p. 3. The rock-‘altars’ (as they have generally been called) on the acropolis of the Midas-city seem to me quite different, and I cannot agree with Dr. Körte's views about them (Es sind Throne für die zinsichtbare Gottheit, Athen. Mitth. 1898, p. 119). Whether the altar grew out of the throne is a different question.

^{page 57 note 2} The other has no doorway (if my memory serves me rightly)

^{page 59 note 1} The west side of the Lake is too marshy for a road to pass over it.

Route from Ak Sheher to Durgut Ova:—Melles keui ( h.), Kurd keui ( h.), Tuzlukjü (1 h. 55 m.; before this village the road to Piribeyli diverges), foot of the hills (1 h. 45 m. ), summit (1 h. 15 m.), foot of Bagirsak dere ( h.), Khursunlu (ca. 1 h). The time-distances added after a place are always reckoned from the previous station; square brackets denote a détour and return to previous station. “H.” signifies hour, and “m.” minutes. Ak Sheher Lake is made too large in recent maps; our route did not pass near it.

^{page 59 note 2} Haïmane (i.e. Waste) is the name of the whole country S. and S.W. ot Angora (Ankyra) between the Sangarios and the vicinity of the Halys.

^{page 60 note 1} The men know Turkish, but Kurdish is the language of domestic life. So the ancient Galatians spoke Celtic in their own homes, using Greek only as a medium of intercourse with the outside world, until Christianity enforced its language.

^{page 61 note 1} Italicised g is used to denote that the letter is softened down in pronunciation so as to become almost silent.

^{page 61 note 2} Details of route from Durgut Ova to Piribeyli:—[From Khursunlu to] Kale (20 m.), Harranlar (15 m.), Durgut (1 h. 20 m.), Geuz Ören (2 h.), Kizil Kuyu Yaila (deserted village at W, extremity of Ova), Yunak (2 h.), Agz Atchik Yokaru (1 h. 45 m.), Agz Atchik Ashagha (55 m.). Ören keui (1 h. 15 m., N.W.), Ashagha Piribeyli ( h.).

^{page 61 note 3} So called from a small Ruinenstätte on its slopes further East (infra).

^{page 61 note 4} Route from Piribeyli to Ak Giöl:— By Yokaru Piribeyli to Kurdushan: thence Tcheltik (3 h.), Kutchuk Hassan (1 h. 25 m.), Oda-bashi (50 m.), over undulating country to Hadji Fakir (1 h. 35 m.); thence through villages near E. end of Ak Giöl: Imam Oglu (20 m.), Serai k. (10 m.), and across the lake to Sakushagh (1 h. 5 m.).

^{page 62 note 1} At Geuk Bunar W.S.W, rises another stream which passes Renkoglu [derived from routes of Prof. Ramsay] and pours a large volume of water into the Sangarios just below (North of) Elles Pasha, 3 hrs. from Tcheltik. We found it impossible to ford this river an hour above its mouth (July 28).

^{page 62 note 2} The road turns up the valley of the Ilidja Su, which comes down from the east of Hammam and falls into the Sangarios about hour below Kabak.

^{page 63 note 1} Route from Sakushagh to Angora:—Yuzuk-bashi (2 h. 5 m.), Siman (1 h. 20 m.), Toprak Bunar on edge of marshy valley running towards Sakaria (10 m.), Turk Tahajir at foot of Sakaria cañon (fully 2 h.), opposite Kabak (45 m.), turn E. up a dere for h. and over plateau down to Katrandji Inler in about hours more, thence Kizil Hissar ( h.), [Tabur-oglu ( hr. to left)], opposite Kiraz-oglu on right (1 h. 50 m.), opposite Baghtchejik lying 20 m. up hill side on left (42 m.), go N.N.E. in direction of 30° for 1 h. 12 m., thence due E. for 25 m., whence Katrandji keui reads 150° (24 m.). Thence to Yamak (reading 330°) on opposite side of the valley (1 h. 10 m.), and over hills to Hammam (ca. h.), Kadi keui (50 m.), Erif (1 h.), Hüyük (55 m.), Gerrim (20 m.), Sari-giöl (ca. 1 h.), Mandra keui (29 m.), Giaour Kalesí (37 m.), Dere keui (20 m.), Oyadja (40 m.), Topakli (1 h. 40 m.), Hadji Muradli h. to left of road (57 m.), opposite Yawrudjik [Perrot's “Giaurtchik”] (28 m.), opposite Deli Hümmetli which lies ca. 1 h. up hills on left (17 m.), junction of alternative road by Giaour Kalesi with ours (48 m.), Hadjilar (27 m.), brow of hills on south of Angora (1 h. 15 m.), thence by Yalandjak, Balghat, etc. to Angora.

^{page 64 note 1} In precisely the same way Bolegasgos is given on the direct road from Angora to Tavium, whereas it is in Tchibuk Ova (Balikassat keui), north-east of Angora.

^{page 65 note 1} Route from Angora to Istanos:—Evedik ( hr.), Yuva (1 hr.), Tchakirlar tchiftlik (57 m.), Emir Yaman (1 hr.), Kebir Giöl ( mile), Serai k. (ca. 2 hrs.), by Teshrek (Circassian) to Meranos (ca. 3 miles), foot of dere (1 hr.), Aïdin (40 m.), Mekhti (15 m.), Halkali (10 m., across river), Yassi ören (1 hr.), Soguljak (30 m., N.W.), Karalar (1 hr.), Shimshit (3 miles), Emir Ghazi (9 m.), Bitdik (37 m.), Girindos (50 m.), Mülk, Il-agut (pronounced Ilaῦt), Akja ören (ca. 20 m.), Istanos ( hr.).

^{page 65 note 2} Hist. Geog. p. 242.

^{page 65 note 3} Ibid, p. 20.

^{page 66 note 1} The name sounded like Tchaï or Tchal; but when I asked the people to pronounce it slowly, it seemed to be Tchagha.

^{page 66 note 2} Route from Istanos to the R. Siberis:—Kaye [not Kaya] keui (ca. 1 h.), Irkek-su tchifilik ( h.), over Ayash Bel, joining chaussée east of Bash Ayash (1 h. 8 m.), Ayash (57 m.), [Mr. Crowfoot went by the chaussée and visited Bash Ayash.], Kassaba (55 m.), Bairam (2 h.), Tizke (1 mile N.N.W.), back to Bairam, Ilidja keui (1 h., W.); Mallal (2 h. 5 m., ca. h. N. of Tizke), Ukhan keui (2 h. S.W., on right bank of river which joins Ayash water; 1 h. from Ilidja), Geunidje tchiftlik ( h.), Tchagha (? Tchai) keui (1 h. 20 m.); Ak kaya (1 h.), Kale (i.e. Petobriga, 1 h. 20 m.), cross Kirmir Tchai and up left bank to Ada ören (1 h. 20 m.), Indje Pellit ( h. in airline), Oyun Pellit (1 h.; N.W. of Ada ören), Kizil Sqfut (“red willow,” h. S.W.), Bey-bazâr ( h.). Back to Devrent on Kirmir Tchai (time forgotten: 3 h. Perrot), along Kirmir Tchai by tchiftlik I. (1 h. 20 m.). tchiftlik II. (20 m.) Bey-bazâr stream (1 h. 10 m.) to Fazil tchiftlik; Mal Tepe (10 m.), Mukhalitch Keuprü (3 h. 15 m., very fast riding), down Sakaria to Yardibi tchiftlik (1 h.), Eski-sheher (=Sykeon) and Bridge on Ala Dagh Su (=Siberis, 10 m.).

^{page 67 note 1} Holder, Sprachschatz, s.v.; Ramsay, CB. ii. p. 382.

^{page 67 note 2} The Greek original is published in Μνημεῖα ῾Αγιολικὰ ἐκδιδόμενα ὑπὸ Θεοφίλου Ιωάννο(BENETIAi, 1884).

^{page 68 note 1} Route from Bridge on Siberis to Juliopolis:—Sarilar (under h.), Ak Yar tchift. ( h.), small eyuk (20 m.), Gemi Bashi tchift. beside Baluk keuprü ( h.), turn N.W. and W. to Aiman bashi (1 h. 58 m.), Emrem Sultan (22 m.), IULIOPOLIS ( h.). Direct road from Juliopolis to Aiman bashi is 35 m. (travelled by servant). From Aiman bashi we returned to Baluk keuprü.

^{page 68 note 2} Called also Kuyudjak, but not in common parlance. It is a poor village and apparently not very old.

^{page 69 note 1} Routes in Mukhalitch district:—From Baluk keuprü to Iki Kilisse (2 h. 5 m.) [whence we visited Gozlu, on top of ridge just over the stream]; Tchardak (1 h., S.W.), Mukhalitch (2 h. S.S.E.). [Mr. Crowfoot went by Bey keui, etc.] Thence to Kaye (mudurluk, I h.), village ten minutes to right (57 m.), Geudje (18m.), Kizil Böyukli (or? Beyikli, 43 m.), back to Geudje, whence Kara Geyikler (30 m.), Tut-agatch (30 m.), Doghan-oglu (40 m.), eastward again to Mohajir keui (20 m.), Igde agatch yokaru ( h.) [ashagha ( h.)], Geuktche aiva (or? Geuktchai) (over h., but ca. 20 m. in air-line), Yaila (1 h. 15 m.), Yarikdji (2 h. 40 m.). Hence I went to Sari keui and by rail to Angora, returning to Bitcher; while Mr. C. visited some villages in the hilly country E. and N.E. of Mukhalitch, rejoining me at Bitcher.

^{page 69 note 2} Μνημεῖα ῾Αγιολογικά p. 434. The Latin translation in Acta Sanctorum wrongly gives it as Ambrena (Hist. Geog. p. 246).

^{page 69 note 3} On ‘Michael’ in modern Turkish names, see CB. i. p. 31.

^{page 69 note 4} For ancient royal estates still remaining as such in Byzantine and Turkish times, cf. CB. I p. 10–11 etc.

^{page 70 note 1} DrKörte's, identification with Pebi is probably correct (Athen. Mittheil. 1897, p. 19 ff.).Google Scholar

^{page 70 note 2} Hist. Geog, pp. 224, 237.

^{page 70 note 3} In a paper published in Globus, vol. lii. (1887).

^{page 70 note 4} Route as far as Pessinus:—From Bitcher to Karadat ( h., in S. direction except for the last h, when we turned sharply W.), Idedjik (55 m.), Soman Ilissar (19 m.) [Karadja Kaya (10 m.)], Dumrek (32 m.), Masud keui (over h.), Babadat ( h., E.); back to Masud, whence Eldjik ( h.), Sivri Hissar (1 h. 35 m.), Tek ören ( h.), Siliba (30 m. ?), over the hills to Arslanli (2 h. 9 m.), Kotchash (23 m. S.E.), Kadynjik ( h. S.S.W.), back over the hills to Bala-hissar (Pessimis, 2 h. 27 m.).

^{page 71 note 1} We several times heard this name applied to the ruins. Hamilton's “Hergan Kale” is unknown. Is it not a misunderstanding of “Djirgin Kale” (Djirgin being the former name of Azizie, the principal town of the district, re-named after Abd-ul-Aziz)?

^{page 71 note 2} On Abrostola, Hist. Geog. p. 236.

^{page 71 note 3} Route from Pessinus back to Angora:—Fet-oglu (2 h. 27 m.), Hadji-Ali-oglu (39 m.), Veledler (35 m.), Bostanlar (15 m., just over the Sakaria, where it bends to S.), Ak Keuprü just below junction of Bunarbashi Su with Sakaria (45 m.), Kutchuk Buzludja (37 m.), Böyuk Buzl. (18 m.), Ab-ed-din tchiftlik (43 m.), Waïsal (28 m.), Bunarbashi (1 h. 35 m.), Hamza Hadji (20 m.), AMORION (25 m.), by Bunarbashi to Zonk (53 m.), Yeni keui (1 h. 55 m.), Lower Tuluk (15 m.), Kaldirim (55 m.), Geuk Tepe (55 m.), Elles Pasha (i h. 49 m.), Topalli on Sangarios [ca. 1 h.; we made circuit by Bolat Hissar (1 h.) to Topalli (63 m.)], Tchakmak (1 h. 45 m.), Kabak (2 h. 56 m.), Bash-keuprü keui (1 h.), Etrek (1 h.), Kavundji keuprii (1 h. 20 m.), Yarre (1 h. 25 m.), Merdjan (31 m), Bedin (13 m.), Aïwalli (7 m.), Yürme (28 m.). Back to Kav. Keuprii, whence Tchanakdjik ( h.), Bolatli station (3 h.), Basra ( h.), Beyjaz (55 m.), Kara-eyuk ( h.), Hadji Toghrul ( h. in direction 325°), Tekke on hills (40 m.), thence straight to Ala Geuz S.E. of Malli keui (4 h. 15 m. fast riding, about 18 miles); Assarli kaya (1 h.), Assarli kaya keui (ca. 7 m.), Balik-koyundji (1 h.), Aladja-atli (2 h. 20 m.), Angora (2 h. 58 m.).

^{page 72 note 1} Hist. Geog. pp. 257, 260.

^{page 72 note 2} H. G. ll. cc.

^{page 73 note 1} Route from Angora to Tcheshnir keuprtt:—Kayash (40 m.), opposite Kizildja keui ten minutes to left (45 m.), Orta keui (25 m.), ruined Khan (1 h. 45 m.), small tumulus Tashli ören [Hassan-oglu (52 m.)], Task Bunar beside the direct road from Hassan-oglu to Kaledjik (1 h.; reckoned 1 h. from Hassan), Yenishan (35 m.) in same dere as Assi Yuzgat (5 m.). Cross watershed of Sakaria and Halys to Kilidjar (1 h. 56 m.), along right bank of Tabanli Su for 35 m., then turn up dere and over hills and across the Halys by a ford to Yakshi Khan (3 h.); [Kürüghin Kale (Eccobriga, 55 m.)], along Halys by Haidarlar (2 h. 20 m.), Kara Ahmedli (1 h. 20 m.). Tcheshnir keuprü (1 h. 30 m.).

^{page 73 note 2} Route from Tcheshnir Keuprü as far as Shedit Hüyük:—Karaketchili (39 m.), Utchemli (2 h. 48 m.=Alibey k.), Kartal (kaimmakamlik, 1 h. 45 m.), Tol ( h.), Beinam (under 3 h.), TchaKal (ca. 1 h.), Kara-oglan (55 m.), Örendjik (53 m.), Dikmen k. (1 h. 55 m.) and back by Roman road to Örendjik; along the post-road to Tchakal (2 h. 35 m.), Aghaboz (1 h. 33 m.), Hadji Izzet Bey tchiftlik (2 h. 14 m.), [hüyük (19 m. N.W.), Karali (56 m.)], Abbasli (1 h. 3 m.), over hilly country to Tchidemli in a dere (2 h. 15 m.; Baghtche Karadallar is beyond ridge on left); S.W. to Adji öz stream (1 h. 44 m.) on direct road from Abbasli, Avshar (37 m.), S.E. again to Adji öz stream (2 h. 5 m.), Karaburun (20 m.), Shedit Hüyük (1 h. 40 m.).

^{page 73 note 3} Ramsay places it near Beinam, a village about an hour east (Hist. Geog. p. 256, etc.), which comes to the same thing.

^{page 75 note 1} Hist. Geog. p. 287.

^{page 75 note 2} Route from Shedit Hüyük to Kotch Hissar: Over low slopes of Pasha Dagh and rolling country to Hadji Tahir k., E. of direct road (3 h. 18 m.), Fezulagh tchiftlik on direct road ( h.), junction of roads from Tchikin-agil and Parlassan (35 m.), Mulkus Obasi (45 m.), Parlassan (48 m.), Palas (25 m.), Ishekli (37 m.), Haidarli (15 m.), Deliler (23 m.), Obasi Veni Yapan (1 h. 8 m.), Demirji (20 m.), Hüyük below Tchikin-agil (1 h. 35 m.), return to Demirji, whence Boghaz keui at foot of the plateau (2 h.), Bazirgian Hüyük (fully 1 h.), [Harmandal (35 m.)], Deve-dami (1 h. 15 m.) Maradit hüyük (45 m.), Tchaili hüyük (55 m.), back to Harmandal, whence we go towards the lake by Kutukli k. (23 m.), Shabanli (1 h. 3 m.), Abbali (1 h. 10 m.), Marnali (45 m.), Devekowan (I h. 5 m.) on road from Parlassan to Ak Serai, reported 5 h. from former; whence by circuit over hills to Kotch Hissar (3 h.).

^{page 76 note 1} See his own account in a paper entitled “Asiana” which has just appeared in the last number of the Bulletin de Corr. Hell. pp. 234–5.

^{page 76 note 2} Route from Kotch llissar to Konia:—Khüiidil hüyük near Lake (45 m.), Bash Khan (3 h. 17 m. very fast, reckoned 6 h. from Kotch Hissar), head of Lake (25 m.), Ak-in hüyük (45 m.), Akardja (1 h. 10 m.), [Böyuk Kishla (37 m.)], Shekerli keui (Tatar, 25 m.), Kulu keui (2 h.), Mandra k. (1 h. 40 m.), Yarashli (1 h. 5 m., reckoned 3 h. from Kulu), Kostangil (35 m.), Arsindji (30 m.), Kozanli (1 h. 35 m.), [visit various old cemeteries]; At-Kafasi (reckoned ca. 2 hrs.; we lose the way), opposite Böyuk Beshkavak (1 h. 39 m.), Kutchuk Beshkavak (35 m.) at foot of Kara Oagh, détour to Kutchuk (or? Kutuk) Ushak (1 h. 35 m.), Yenidjoba ( h., but reckoned 2 h.), Kushdjali or Hadjilar (1 h. 52 m.), southwards to Bunarbashi (1 h. 40 m.), Bunarb. keui (6 m.), Tcheshmeli Zebir k. (2 h. 50 m.), Kuyuli Zebir k. (35 m. fast), [Lek k. ( h. fast)], Insuyu ( h., reckoned 4), Tchorgia hüyük (2 h. 53 m.), Kara Tepe ( h.), Inevi (1 h.), Herkenli or Dongdurma (4 h. 10 m. going 4 miles per hour), Kski-il (4 h. 15 m.), Timm hüyük (reckoned 1 h.), Tash Bunar (2 h. 15 m.), Hadji Bey Yailasi ( h.), Toprak kale (32 m.), Tekir k. (1 h. 20 m.), Hoïdus Yaila (1 h. 25 m.), Suwarek ( h.), Yaila (3 h. fast, reckoned 4), Kayadjik yaila ( h. very fast), Konia (2 h.).

^{page 76 note 3} See Hist. Geog. p. 227.

^{page 77 note 1} Hist. Geog. p. 344.

^{page 77 note 2} Ibid.

^{page 77 note 3} Fragments of early painted pottery, ornamented merely with lines, were found here.

^{page 77 note 4} Ibid.

^{page 77 note 5} Arch.-Epigr. Mitth. 1896, p. 33: Hier fanden wir keine Spuren einer antiken Ansiedlung. But in his volume “Reise in Kleinasien” (p. 98 f.), he says that Perta is perhaps to be sought here. Die kolossale Ausdehnung des Gräberfeldes zeugte voti der Grösse dieser Ansiedlung, die, wie aus den vorhandenen spätrömischen oder byzantinischen Skulpturfragmetiten hervorgeht, in frühe Zeit zurückreicht.

^{page 77 note 6} It cannot, however, lie so far to the south as he puts it (“4 hours S.W. from Eskil,’ P· 343) it is more west than south of Eski-il. He does not name the site, but the identity is certain.