¹ Barbicans are likewise to be found at three other Ismāʿīlī castles, namely Maṣyāf, al-‘Ullayqa and al-Qadmūs.

² Berchem, Max Van, ’Épigraphie des Assassins de Syrie’, in J.A., 9th ser., IX (1897), 494.Google Scholar

³ Firās, Abū, Manāqib al-Mawlā. Rāshid al-Dīn, ed. with French trans, by Guyard, S., ‘Un grand maître des Assassins au temps de Saladin’, in J.A., 7th ser., IX (1877), trans., 416, text, 469.Google Scholar

⁴ Max Van Berchem, ‘Épigraphie …’, 487–94 with pl. 4.

⁵ Qur'ān, xv. 46. The relevance of this verse in relation to the mood that was sought in the ḥammām is extended by the two verses 47 and 48 which follow but which do not appear on the inscription: ‘We shall remove all ill-will from their hearts. They will be as brothers resting on couches facing one another. They shall not be wearied by toil, nor shall they be expelled.’

⁶ Qur'ān, III. 122.

⁷ In the republication of the inscription's text in the Répertoire chronologique d'épigraphie Arabe (ed. Combe, Ét., Sauvaget, J. and Wiet, G., XI (1941–1942), p. 93, no. 4143)Google Scholar the word, ‘the Imāms’, has been erroneously omitted.

⁸ For possible alternative translations for the equivocal phrase see Max Van Berchem, ‘Épigraphie …’, 489, 494 with n. 1.

⁹ Ibid., 489.

¹⁰ al-Nasawī, Muḥammad, Histoire du Sultan Djelal ed-Dīn Manḳobirti, trans. Houdas, O. (Paris, 1891–1895), IX, 167–8Google Scholar: x, 280–1.

¹¹ al-ḥamawī, Muhammad, Al-Ta'ríkh al- Manṣūrī, ed. Gryazneviĉ, P. A. (Moscow, 1960), fos. 164a–164b.Google Scholar

¹² Ibn Wāṣil, Mufarrij al-kurūb, anno 637, MS. Paris 1702, 333b–334a; MS. 1703, 23b; Lewis, B., ‘Assassins of Syria and Ismā‘īlīs of Persia’, in La Persia nel Medioevo, Accademia Nazionale dei Lincei (Rome, 1971), p. 579 n. 24.Google Scholar

¹³ Max Van Berchem, ‘Épigraphie …’, 490.

¹⁴ See especially Hodgson, M. G. S., The Order of Assassins (The Hague, 1955), pp. 246–7Google Scholar with n. 18; N. A. Mirza, ‘The Syrian Ismā‘īlīs at the time of the Crusades’, unpub. Ph.D. thesis, Durham, 1963, pp. 84–5, 100 with n. 2.Google Scholar

¹⁵ At this point I would like to offer my thanks to Dr. A. Bahnassi and his staff at the Department of Museums and Antiquities in Damascus for smoothing the way to my fieldwork in Syria; to Mr. R. Heikal of the Museum of ṬarṬūs whose informed company on our journeys was always appreciated; and to Mr. P. Guedes of the Architectural Association School of Architecture in London, a colleague in this and other surveys carried out in the Jabal Anṣarīya and with whom I was able to discuss, on site, some of the ideas expressed in this article. The cost of travel to and from Syria was defrayed by grants offered by the University of London through the Central Research Fund and the School of Oriental and African Studies.

¹⁶ Through the process of soil erosion several lengths of piping are now exposed on the surface.

¹⁷ For a concise description of Arabic terminology as it applies to the medieval and modern ḥammām, see J. Sourdel-Thomine, ‘ḥammām’, in Enc. Islam, 2nd. edn., III, 141.

¹⁸ Johns, C. N., ‘Excavations at Pilgrims’ Castle (‘Atlīth)’, in Quarterly Dept. Antiq. Palestine, I (1932), 34, pl. XLV, fig. 2.Google Scholar

¹⁹ This type of vault is not to be confused with the barrel vault which can be open at the ends or closed with flat lunettes.

²⁰ J. Sourdel-Thomine, ‘ḥammām’, 141.

²¹ Through a small gap at the top of the plug it was possible to determine that the masonry was 20 cm. thick.

²² It is to be added here that, in contrast to the barrel vaults which are surmounted by a flat roof, the dome and cradle vaults rise above this roof as independent forms. The same obtains in the rest of the ḥammām.

²³ By an oversight these light apertures were not checked to ascertain if they still contained the bottle-end glazing familiar to better-preserved ḥammāms.

²⁴ What cannot be avoided here is the possibility that there is a second set of pipes circulating in the walls below debris level. This at least would bring the ḥammām into line with many other ḥammāms of similar and later age, where the usual practice is to incorporate two systems of piping, one for hot water, the other for cold.

²⁵ This type of mortar, which is to be equated with the opus signinum of the Romans, is given its pink colouration by the addition of crushed earthenware.

²⁶ Calculated on the basis of an average width of 2.35 m. length of 2.45 m. and depth of 1.08 m.

²⁷ J. Sourdel-Thomine, ‘ḥammām’, 144.

²⁸ ‘Excavations at Sīrāf: Fourth Interim Report’, in Iran, IX (1971), 10–12Google Scholar; Rogers, Michael, The Spread of Islam (1976), p. 65.Google Scholar

²⁹ ‘Excavations at Pilgrims’ Castle (‘Atlīth)’ in Quarterly Dept. Antiq. Palestine, I (1932), 124–9Google Scholar, with pls. LI and LII.

³⁰ Burckhardt, J. L., Travels in Syria and the Holy Land (London, 1822), p. 151Google Scholar; Meryon, C. L., Travels of Lady Hester Stanhope, 3 vols. (1846), III, pp. 334–7Google Scholar; Rey, E. G., ‘Reconnaissance de la montagne des Ansariés’, in Bull. de la Soc. de Géogr. de Paris, XI (1866), 5th ser., 461Google Scholar; Boulanger, R., The Middle East, Hachette World Guides (Paris, 1966), p. 450.Google Scholar

³¹ al-Ukhūwa, Ibn, Ma‘alitn al-qurba fī aḥkām al-ḥisba, ed. and trans, by Levy, Reuben (Cambridge, 1938), trans., 52; text, 100.Google Scholar

³² Such a reservoir is termed ḥawḍ al-nawba, ‘reservoir of rotation’ by Ibn al-Ukhūwa (ibid., trans., 52; text, 100).

³³ It follows, of course, that should other varieties of masons' marks come to light this hypothesis would have to be revised.

³⁴ Écochard, M., ‘Trois bains ayyoubides de Damas’, in Les monuments ayyoubides de Damas (Paris, 1940), pp. 103–6Google Scholar; Écochard, M. and Coeur, C. Le, Les bains de Damas, 2 vols. (Beirut, 1942–1943), II, p. 19.Google Scholar

³⁵ Ibid., II, pp. 20–2.

³⁶ Sauvaget, J., ‘Un bain damasquin du XIII^e siècle’, in Syria, XI (1930), 376–80Google Scholar; M. Écochard, ‘Trois bains …’, pp. 105–6.