^{page 243 note 1} Quatremère, , Mémoire sur la ville ďEcbatane, in Mem. de ľ Ac. des Inscr., 1851, xix/1, pp. 419–456Google Scholar

^{page 243 note 2} Kiepert, , Atlas Antiqwus: Ganzak at Leylān. Nöldeke, Geschichte der Sasaniden, 1879, p. 100: “Ueberhaupt enthält diese Abhandlung Rawlinsons bei aller Verdienstlichkeit sehr viel verfehltes.”Google Scholar, Herzfeld, , Arch. Mitt., II/2, 1930, p. 72,Google Scholar places the temple at 6 fare, from Maragha “in Richtung Zinjan ” [?]. Marquart, , A Catalogue of the provincial capitals, 1931, 109 (Ganzak = Laylan)Google Scholar

^{page 243 note 3} Hoffmann, G., Auszüge aus d. syrischen Akten, 1880, p. 252;Google ScholarFabricius, , Theophanes von Mytilene, Strassburg, 1898, p. 228Google Scholar (the author winds up by surrendering his lucid arguments to the authority of the “Orientalists” ); Marquart, ĒrānŠahr, p. 108 (but see corrections in his later A Catalogue, p. 109); Pernice, A., L'Imperatore Eraclio, Florence, 1905, p. 125Google Scholar, still found “le raggioni del Rawlinson convincentissime”; Schwarz, P., Iran im Mittelalter, viii, 1932, pp. 1099, 1454;Google Scholar, SirStein, A., Old routes in Western Iran, 1940, p. 341Google Scholar (with some hesitation).

^{page 244 note 1} See Rawlinson, passim; Hoffmann, Index; W. Fabricius, pp. 227–231; Weissbach, Gazaca in Pauly-Wissowa, Seal-Lexicon; A. V. W. Jackson, Persia Past and Present, pp. 124–143; Mary Crane in Bull. Amer. Inst. Iran. Art., December, 1937, pp. 84–9.

^{page 244 note 2} According to Muralt, this happened in A.D. 591.

^{page 244 note 3} I feel pretty certain that the Kurdish tribe called in Arabic Ḥumaidī is connected with Hnāithā, just as the Hadbhbānī Kurds have been surnamed after Hedhayeb (Adiabene).

^{page 245 note 1} Doubtful. Marquart, Südarmenien, 1930, p. 337, identifies Zerbis with Bohtān-su, and places the Azoni near Arzūn (in the neighbourhood of Se'ert). Instead of Azoni, one MS. has Aloni, which name may be represented by that of the district Alān (in the gorge of the Lesser Zab).

^{page 245 note 2} More doubtful is his further identification of it with Σιθκαρ, which, following Ptolemy's co-ordinates, lay considerably further east (27–2 miles east of Δαριασρ=Daryāz on the Sauch-bulaq river).

^{page 245 note 3} “In the enemy's territory” dva rmt άθά τό θ πολέειοθ Χωροθ. The text suggests that the victors were collecting the booty. “The stench” also is a certain hint at the nearness of the battlefield.

^{page 246 note 1} Possibly the Miyān-du-āb, the strip of territory between the rivers Jaghatu and Tatavu with its excellent grazing grounds. Even now it is occupied by the royal studs.

^{page 247 note 1} On Firdausi, vide infra, p. 255.

^{page 247 note 2} In this connection one inigbt remember (1) that the River Sārfīq, which is the south-eastern affluent of the Jaghatī, might have been taken for the chief head-water of the latter river, (2) that in the neighbourhood of the southern bank of the Sārīq are situated the famous caves of Kerefto bearing a Greek dedication to Heracles (see now Sir A. Stein, 324–346), and (3) that Heracles is the Greek equivalent of the Iranian Varatraghna > Varahrān > Varārān.

^{page 247 note 3} To the six nāḥya of Marāgha the Nuzhat al-qulīb adds two dependencies (tavābi'): . The latter might correspond to Valārān, but it has numerous variants, etc., which would suggest *Qizī-uzan.

^{page 248 note 1} E., Gerland, Die pers. Feldzilge des Kaisers Heraeleios, in Byz. Zeitschrift, iii, 1894, pp. 330–373Google Scholar; A., Pernice, L'Imperatore Eraclio, Florence, 1905, (compilative).Google Scholar

^{page 249 note 1} This fantastic detail should be compared with Ṭabarī, I/2, 866, where it is reported that, after his victorious campaign against the Khāqān, Bahrāin Gīr presented the jewels of his booty to the fire-temple of Shīz, as he also attached to it the Khāqān's wife as a maid-servant.

^{page 249 note 2} Vide infra, p. 255. The editor of Mas'īdī, Barbier de Meynard, brought in a new confusion by declaring that al-Birlca (which he apparently mistook for Forg) was a town in Fars!

^{page 251 note 1} Kele-shīn—to be distinguished from the famous Kele-shin, lying much more north-west between Ushnī and Ravānduz (probably used by the Byzantine troops in A.D. 591).

^{page 251 note 2} Bull, of the American Inst. for Iranian Art, December, 1937, pp. 71–105.

^{page 251 note 3} A. F. Stahl, Peterm. Mitt., 1905, p. 32: “Nichts deutet darauf hin dass hier einst eine grŏssere Stadt stand”.

^{page 253 note 1} Under Malik-shāh the length of a farsakh was ascertained to be: 6,000 paces in 'Irāq, Kurdistan, etc., but 10,000 paces in Azarbayjān and Armenia; see Nuzhai al-Qulīb, 164 (transl. 161).

^{page 253 note 2} BS0A8, XI/1, p. 87Google Scholar.

^{page 253 note 3} I. Khurd., 121, mentions a stage Shiẓ (without article!) at 4 fars. from Dīnavar, on the road to Sīsar. V. infra, p. 264, n. 2.

^{page 253 note 4} The element vān, “a place” (Arm. avan), is frequent in North-Western Iran: Jāb.r-vān, Bājer-vān, etc., as probably also Shirvān, Gurzivān, etc.

^{page 253 note 5} Perhaps identical with the place called in Assyrian sources Ṣiṣṣirtu; see Minorsky, Senna and Sīsar in E.I.

^{page 254 note 1} In fact Muqaddasī, 382, describes a direct road Marāgha-Shahrazīr making no detour via Barza (6 marhalas plus 30 farsakhs).

^{page 254 note 2} See also Sir A. Stein, op. cit., 349–351.

^{page 254 note 3} See also below, p. 265, Yāqīt's description of Kaznā and Jaznaq.

^{page 254 note 4} JBGS., 1832, pp. 5–6.

^{page 255 note 1} Cedrenus uses the same terms as Theophanes in describing the temple, the treasure of Croesus and “the charcoal trick”, but he adds a detailed and interesting description of a statue of Chosroes (cf. Mas'īdī, iv, 74). He may have found it in some other source. In any case his location of the temple in Ganzak has no decisive importance.

^{page 255 note 2} Mis'ar has been known through the quotations found in Yāqīt's Mu'jam al-buldān, and. in Zakariyā. Qazvīnī. Marquart did not live to fulfil his promise (Festschrift Sachau, p. 292) I to study the problem of the authenticity of what goes by the name of Mis'ar. Meanwhile two lf risāla of Mis'ar's travels have been found in the Mashhad MS. containing alṣo a part of I. Faqīh and an important copy of I. Faḍlān (incomplete).

^{page 256 note 1} All the quotations in Marquart, Catalogue, pp. 108–9; but H. W. Bailey, Zoroastrian problems, 230, suggests the reading *Ganj-i šasapīkān.

^{page 256 note 2} Nat. hist., vi, 43. In another place he speaks, vi, 42, of the capital of Atropatene: “oppidum ejus Gazae (var. Gaze, Gazzea), ab Artaxatis C CCCL passuum, totidem ab Ecbatanis Medorum, quorum pars sunt Atropateni,” ed. Detlefsen, 1904, p. 136. This indication suits Leylān but not Takht-i Sulaymān.

^{page 256 note 3} At the most one might recognize Šiz in the last element of Dar-arta-sis, but such a hypothesis does not solve the difficulty in the beginning of the name.

^{page 256 note 4} As suggested by the editor Quercius, ed. Migne, p. 1329.

^{page 257 note 1} Rawlinson, 81, refers to the book “Tebektegin” from which Mas'ūdī borrowed his information on Persian antiquities. In the printed edition of the Murīj, ii, 118, 120, the name is given as al-S.kīsaran (according to Christensen, Les Kayanides, 143: “the chiefs of the Saka ”). In Tanbīh, 96, a different (?) book is quoted: Baykār (apparently identical with Murūj, ii, 44: al-Bnlcs). Marquart, Streifzüge, 166, restored it as *Paykār, “Book of wars.” Unfortunately Mas'udī's quotations contain nothing on fire-temples.

^{page 258 note 1} See D. N.Wilber in Bull. Am. Inst. Pers. AH, V/2, p. 102.

^{page 258 note 2} Mong. soqur “blind, a blind man” + Turk, suffix -luq, perhaps meaning “ a blind alley ” (?).

^{page 259 note 1} In his thesis on Theophanes of Mytilene, Strassburg, 1888, W. Fabricins studies also the fragments of Dellius.

^{page 259 note 2} I leave the problem of the temple of Baris over which so much ink has been spilled; see H., Stephanns, Thesaurus, Eng. ed., 1816–1818, ii, cccxxii-v:Google Scholar L. C. Valckenaer, Dissertatio de tocabulo Bαρις. See also Pauly-Wissowa sub verbo. J. Schrader's restoration *Mams (in Armenian: Ararat) for Bαρις is still very tempting in view of the quotation from Nicolaus Damaseenus inJosepmis, Ant. Jud., i, iv, on the mountain Baris, situated towards Armenia, on which many people took shelter during the Flood.

^{page 259 note 3} Now Ala-dagh, forming the north-eastern barrier of the Van basin.

^{page 259 note 4} This form is repeated in the compilation called Parthica and falsely ascribed to Appian, ed. Schweighăuser (1785), p. 77. The other variants are ραάρτ, ραάοθ see Plutarch, VII/1 (Teubner, 1915), p. 113.

^{page 260 note 1} In which Strabo agrees with Theophylaotus, vide supra, pp. 245, 254.

^{page 260 note 2} Cf. Fabricius, 227: Gazaea— “die Unterstadt”; Vera, “wie schon der Name andeutet (er soil von pers. var- saepes, arx kommen, Kramer) die Burg bezeiehnete”.

^{page 260 note 3} The emendation of Γάζα καί έθ to Γάζακα αϋθ is due to FalfiKa aim is due to Fabricius.

^{page 261 note 1} Rawlinson, 120, wrongly compares Gazaca not with Zazaca, but with Azaga, which must have lain in the region of Mākī

^{page 261 note 2} Rawlinson, op. cit., 121: “from some cause… there is a greater tendency to exaggeration in Ptolemy's latitudinal measurements of Western Persia than in those of any of the contiguous countries.”

^{page 261 note 3} Wrongly dotted by the editor Agrāh-rūdh. See Minārsky, Maragha in E.I., in which the suggestion of Phraata = Marāgha was first made. On a similar name of a Sistan river: Avestan Fradaθā, now Parah-rūd, see Marquart, Wehrot, 1938, p. 22. Marquart disregards the form Phraata and explains Phraaspa as *frāδah-aspa “fostering horses”, Ērānšahr, 108; Catalogue, 109.

^{page 261 note 4} Cf. Mecquenem in Annales de Géographie, 1908, 128–144.

^{page 262 note 1} Another important ancient site in the same region is the castle known now as Qal'a-Zohāk, on the Qaranghu; see Monteith, op. cit., 4. It lies some 52 miles east of Marāgha, and tentatively might be identified with Ptolemy's Phanaspa (?). Its distance from Zanjān is circa 100 miles. Thus in fact it lies at one-third of the distance Maragha-Zanjān, and the general direction of the road (south-east) corresponds to Ptolemy's co-ordinates.

^{page 262 note 2} Plutarch, cf. 41: “Antonius was intending to lead his troops back by the same road, which was through a plain country without trees,” but a Mard guide “advised him in his flight to keep to the mountains on his right”, and took him by a “shorter road” along inhabited villages. The way there may have been via Sofyan—west of Tabriz—eastern bank of Lake Urmia, or alternatively, Sofyan-Tabriz and round the north-eastern side of Sahand. The retreat must have been by some shorter cut of the eastern Sahand, and more to the north-east of the former road (i.e. hugging the western hills of Qaraja-dagh which overlook the Tabriz plain).

^{page 263 note 1} This point of view is not contradicted by a one day's raid and the foraging expeditions which Antonius launched from Phraata (Plutarch, ch. 39–40).

^{page 263 note 2} I doubt Herzfeld's statement about the data of the Kay Khusrau cycle being “wirkliche Geographie”, Arch. Milt., II/2, p. 72.

^{page 264 note 1} Nasawī, the biographer of the Khwārazm-shāh Jalāl al-Dīn (p. 225), boldly indicates the exact place of the event in the highlands of Barda'a. [In the Bundahishn, xxii, 8, the “ lake of Khusrau ” is placed at 50 farsakhs from čēčast. Even reckoning 1 far. = 3 miles, the distance would take us beyond Zanjān. E. W. West identified Khusrau's lake either with that of Van, or with Sevan (GOkche). The latter is preferable, as Van has no connection with Khusrau.]

^{page 264 note 2} Two other Shīz (without the article) are known, Schwarz, op. cit., 703, 917, but their names are doubtful: one of them (perhaps *Bir) in Shahrazūr, and the other (var. Sīr) at 1 fare. north of Dīnavar. F. supra, p. 253, n. 3.

^{page 264 note 3} There was of course a temptation to take Čēčast for Čēč-ast “ it is Čēč”.

^{page 264 note 4} The restoration of this name with -khōst, “a dam, a road rammed down,” is unlikely. On khōst see Herafeld, Arch. Mitt., 11/2, 80–3.

^{page 265 note 1} See, Wustenfeld, Jâcût's Reisen, in ZDMG, xviii, 1864, p. 441Google Scholar. Yāqūt definitely says that he visited Baswē which lies to the south-west of Lake Urmia, but his road to Marāgha must have left Ganzak considerably to the south-east.

^{page 265 note 2} In theBundahishn, xxii, 8Google Scholar.