ABSTRACT
Vérin recognised that the mosques with elaborate carved coral decoration at Tsingoni on Maore (detailed below) and Sima and Domoni on Nzwani implied occupations of the fourteenth century or before at these sites. We had some idea of what the ceramics would be from his excavations at the site of Mahilaka on the northwest coast of Madagascar (Vérin 1975: 618–39). Based on the similarities between the ceramics of several sites known from surface survey, Kus and Wright defined a ‘Hanyundru (then spelled ‘Agnundro’, and later ‘Hagnoundrou’) Phase’ (Kus and Wright 1976). In 1980, Wright excavated several soundings in a wave-cut cliff at the modern village of Hanyundru on the Bay of Boeny in southwestern Maore (Wright 1983). This was a coastal village of 2–3 hectares. These efforts yielded shells and animal bones, a range of local ceramics and a sherd of late Persian Gulf sgraffiato with matte green glaze suggesting a thirteenth-century date. Wright also observed, but could not excavate, the footings of a masonry building oriented east–west, probably a mosque, and two burials on their right side with face oriented north, the prescribed Islamic rite. No absolute dates were possible. Beginning in 1981, Claude Allibert began his long-term study of the sites of Bagamoyo and Mirandolé on the small volcanic island of Pamanzi off the northeast coast of Maore, a site partly damaged by wave erosion (Allibert et al. 1983; Allibert 1992). Here his team found the remains of domestic features, iron-working installations, and a range of burials, with a radiocarbon date calibrated to the twelfth century. The excavated ceramics are different from those of Hanyundru, suggesting that it might be possible to define several cultural phases within the period 1000 to 1400 ce.
