¹ By ‘Kilwa’ in this paper is intended the site of the ancient town round the village now known as Kilwa Kisiwani (‘Kilwa on the island’) some 160 miles south of Dar es Salaam.Google Scholar

² These sources have been frequently reviewed, e.g. by Schacht, J., ‘An Unknown Type of Minbar and its Historical Significance’, Ars Orientalis, II, 165–70.Google ScholarAdditional to those there cited is ‘Kitãb al-Zanūj’ in Cerulli, Somalia, 1, 233–92. The earlier of two (Arabic) manuscripts of this dates from the end of the nineteenth century. The text, which was evidently compiled on the southern Somali coast, seems to be based on more than one document, one of which deals with early Arab settlements on the East African coast.Google Scholar See also Freeman-Grenville, G. S. P. in Discovering Africa's Past, Uganda Museum Occasional Paper 4, 8–10.Google Scholar

³ Tanganyika Notes and Records, 53 Oct. 1959, 161–75.

⁴ The supposed finds of pre-Muslim coins are examined by DrGrenville, D. S. P. Freeman, ‘East Africa Coin Finds’, Journal of African History, 1 (1960), 32–4.Google Scholar

⁵ Annual Report, Department of Antiquities, Tanganyika 1960, II, Plate II.

⁶ Ibid., 12.

⁷ Smolla, G., Actes du IVe Congrès Panafricain de Préhistoire et de L'Etude du Quaternaire, Section III, 243–7.Google Scholar

⁸ Harding, J. R., ‘Late Stone Age Sites on the Tanganyika Coast’, Man (1961), 221.Google Scholar

⁹ Dr L. S. B. Leakey, verbal information.Google Scholar

¹⁰ Variously equated with Madagascar, Zanzibar and Pemba. Zanzibar is perhaps the most probable.Google Scholar

¹¹ Op. cit., 167.

¹² Cerulli, op. Cit., 1, 238, 266.

¹³ Antiquaries' Journal, XXXVII, 23.

¹⁴ Annual Report 1960, Department of Antiquities, Tanganyika, Appendix II.

¹⁵ ‘Coinage in East Africa before Portuguese times’, Numismatic Chronicle, Sixth series, XVII, 170.

¹⁶ For which see the author's ‘Notes on Kilwa’, Tanganyika Notes and Records, 53, 179–203.

¹⁷ ‘The Kilwa Civilization’, Tanganyika Notes and Records, 6 1938, 67.

¹⁸ Surveyed and drawn by Mr P. S. Garlake.Google Scholar

¹⁹ It was certainly designed to hold water; and being shallow and unprotected from the sun is unlikely to have been a cistern.Google Scholar

²⁰ This is the opinion of Mr Basil Gray and Mr John Ayers, for whose advice the writer expresses his grateful thanks.Google Scholar

²¹ Until the middle of the twelfth century the angular Kufic script was used for inscriptions, being subsequently replaced by the rounded nathki.Google Scholar

²² The standard dictionary meaning of Husuni is ‘fort’, ‘castle’. Grenville, G. S. P. Freeman, ‘Husuni’, Tanganyika Notes and Records, 58 and 59 (1962), 227–30, points out that the Arabic word husn (from which the Swahili husuni is derived) denotes in Southern Arabia a fortified residence. Other words are usually used for ‘fort’ in Swahili.Google Scholar

²³ The Travels of Ibn Battuta, edited SirGibb, Hamilton (Hakluyt Society) 11, 381. Ibn Battüta's statement that Kilwa was built entirely of wood is hardly credible. Ignoring the buildings under discussion, there is convincing evidence in the Kilwa Chronicle that the Great Mosque was built in stone at an earlier date, and trial excavations in the town have disclosed stone walls almost certainly datable to before the beginning of the fourteenth century. It would seem that Ibn Battūta's memory was at fault (as it on occasion was: Suakin is described, for example, as six miles from the coast, whereas it is very close to the mainland; on the other hand no mention is made of Kilwa being an island).Google ScholarFreeman Grenville has suggested an alternative reading for the passage, which, however, GibbGoogle Scholar(Ibid., 380, note 62) rejects.

²⁴ SirGray, John, ‘The French at Kilwa’, Tanganyika Notes and Records, 44 (1956), 47.Google Scholar

²⁵ SirGray, John, ‘Fort Santiago at Kilwa’, Tanganyika Notes and Records, 58 and 59 (1962), 178, suggests that this building is that referred to by Vespucci who writes of a castle with four towers which was half built which the Portuguese saw in 1508. Considerations of date apart, this identification seems unlikely as, first, mention would surely have been made of the distance from the town if Husuni Ndogo was in point, and, second, one would expect the much more impressive Husuni Kubwa to be mentioned as well.Google Scholar

²⁶ Creswell, K. A. C., Early Muslim Architecture, Penguin Books (1958), 125, 142. He dates Mshatta to the mid-eighth century A.D.Google Scholar

²⁷ ., 188.Google ScholarIbid

²⁸ See above, p. 181. This is the only building the Umayyads are recorded as having erected; note also that the Arabic word employed, husn, used is the equivalent of the Swahili husuni (see note 22). De Barros, however, in his version of the Kilwa Chronicle, mentions the building of a fortress at Kilwa by ‘Soleiman Hacen’ (Sulaimān bin al-Hasan bin Daud, c. 1170–88).Google Scholar

²⁹ Creswell, op. cit., 286–8.Google Scholar

³⁰ Marçais, G., L'Architecture Musulman d'occident (Paris, 1954), 20–1.Google Scholar

³¹ ‘Shiraz’ is, in view of the paucity of Persian elements in the culture of the east coast both in the past and at present, taken to include the Arab-influenced coast of the province of Fars; and, indeed, if the tradition of the emigration of the seven brothers of Lasa is another version of this story, the whole of the northern end of the Persian Gulf may be involved.Google Scholar

³² Schacht, op. cit., 170 and passim maintains that a type of minbar common in mosques of the last two centuries is a survival from Umayyad times and is evidence of the spread of Islam to the coast as early as the eighth century A.D. But the apparent absence of this type of mihrab in the intervening centuries is unexplained.Google Scholar

³³ The archaeological examination of Islamic sites in the Gulf has hardly been begun, and we are even more ignorant of Oman. Investigations in these areas might well alter the present picture.Google Scholar

³⁴ Annual Report, Department of Antiquities, Tanganyika (1960), 6.Google Scholar

³⁵ Kirkman, op. cit., 20 and Plate IOC.Google Scholar