pursuit of game, and therefore that they may have possessed palaces outside Kûfah to which they were accustomed to resort. Manṣûr, the second of the line, founded Baghdâd in A.D. 762, and removed the offices of government thither from Hâshimiyyeh near Kûfah in 763. His predecessor Ṣaffâh had lived at Hâshimiyyeh near Anbâr: it was he who had transferred the capital from Damascus to ‘Irâq. Previous to 750, when the last Umayyad khalif, Marwân II, was deposed and slain, the eastern provinces of the empire were governed by powerful viceroys, and if Ukhaiḍir is to be regarded as pre-Abbâsid it is to one of these that it must be attributed. Men like Ziyâd ibn Abîhi or Ḥadjdjâdj, who controlled the riches of ‘Irâq and Persia, were scarcely second in wealth and power to the khalifs themselves. Ziyâd’s personal austerity is attested by historians who had no desire to depict the character of Mu’âwiyah’s vicegerent in a favourable light, but his architectural activity is shown not only by the number of mosques which he founded or rebuilt, but also by the erection of palaces at Baṣrah.[448] He died in A.D. 673 after holding his high office under ‘Ali and Mu’âwiyah for a period of nearly fifteen years. Ḥadjdjâdj was governor of ‘Irâq from A.D. 695 to 713. In the khalifate of Ḥishâm, Khâlid ibn ‘Abdallâh ruled over ‘Irâq for thirteen years (724-737), and Yûsuf ibn ‘Umar, who succeeded to the post, held it for seven years. Any of these men might have built and occupied palaces in the wilderness, imitating the practice of their Umayyad masters, and also of their Nu’mânid predecessors in the very region in which the Umayyad viceroys wielded in their turn an authority far greater than that to which the Arab princes of Ḥîrah could lay claim. But the existence of a miḥrâb in the mosque fixes a date before which it is unlikely that Ukhaiḍir could have been built. According to Mohammadan writers, the first miḥrâb was that which was constructed in the mosque of Medînah between A.D. 709 and 711, and if that be so Ukhaiḍir cannot be placed earlier than the last years of Ḥadjdjâdj. I take the years 709-711 as the earliest possible date and the khalifate of Hârûn al-Rashîd as the latest possible date, and with due regard to the probable age of the Syrian palaces on the one hand, and to the architectural features of Ukhaiḍir as compared with those of Raqqah and Sâmarrâ on the other, I conclude that Ukhaiḍir must have been built towards the middle of the eighth century.