[19] Dr. Reuther observes here the funnel leading from the bottom of the niche to the top of the arch which had been described in the outer gates. [20] The decoration as well as the funnel had escaped my notice, but when Dr. Reuther called my attention to the former I was able to verify the correctness of his observation on one of my own photographs. [21] _Journal of the Hellenic Society_, vol. xxx, 1910, p. 77. [22] In the spring of 1910, I asked M. Viollet, who was then on his way to Mesopotamia, to clear away the ruins from the middle of the south wall and ascertain whether there were any sign of a miḥrâb. Upon his return he informed me that he had discovered the niche at the point which I had indicated and that he felt no hesitation as to its being in fact the miḥrâb. When I was at Ukhaiḍir in 1911, I uncovered the niche still further and photographed it carefully. Two of these photographs I sent to Dr. Wetzel for publication in the German work, and they are there reproduced, _Ocheïdir_, Figs. 22 and 23. Professor Brünnow has suggested that since prayer niches with flanking colonnettes were known to the Nabataeans, the Mohammadan niche, with its non-Arabic name, was certainly derived from pre-Mohammadan usage. (’Zur neuesten Entwicklung der Meschetta-Frage,’ _Zeitschrift für Assyriologie_, August 1912, p. 129.) This view is not likely to find acceptance. It is expressly stated that the miḥrâb was a feature of the mosque which was borrowed from the Christian cult and that it was not adopted until the beginning of the second century of Islâm. (See Lammens, Ziâd ibn Abîhi, _Rivista degli Studi Orientali_, vol. iv, 1911, p. 246 (94), note 1, and Becker, ‘Zur Geschichte des islamischen Kultus,’ _Der Islam_, vol. iii, 1912, p. 392.) I continue, therefore, to regard the niche at Ukhaiḍir as a clear proof that the building was originally intended for a mosque. [23] _Ocheïdir_, p. 24. [24] There seems to me to be an error in the reconstruction of the north façade given in _Ocheïdir_, Plate 24. Dr. Reuther makes the wall of the chemin de ronde, immediately to the west of the gate-house, stand flush with the outer edge of the vault between the gate-house and the tower. I do not think that this is correct. The chemin de ronde projected no further here than it projected between the other towers, i.e. it was flush with the face of the pilasters, and in my Plate 11, Fig. 1, its windows can be seen behind the balcony. If the wall had been flush with the edge of the balcony vault, the fall of that vault, partial to the west of the gate-house, total to the east, must have entailed the fall of the wall also. But this is not the case; the chemin de ronde is intact on either side. [25] Aiwân is the Persian form, very commonly used in the Shâhnâmah. It has become lîwân in Arabic by the incorporation of the article al-Aiwân. (Note by Sir Charles Lyall.) [26] _Ocheïdir_, p. 5.