Five levels of stone collapse were encountered. A partial excavation of Sq D revealed a Mamluk aqueduct, running parallel to the southern face of Wall 3.
The fifth and fourth levels of stone collapse had been excavated in previous seasons. The third level in Sq B yielded a rough stone, inscribed with the Roman numeral 56 between two slanting lines (LVI/; Fig. 1). No other part of the stone had been dressed and it was probably in secondary use. Other finds among the stones were a small silver ring and a silver earring, as well as several well-cut voussoirs.
It is known from a 1685 CE drawing (Z. Vilnay, The Holy Land in Old Prints and Maps, 1965:126) that the Crusader building at Moza stood at least partly to its second story roof. The collapse of the building occurred all at once, with rows of stones jammed tightly into the ground and a large area of plaster lying intact, face down. Such a collapse can only occur as a result of a serious earthquake and two major earthquakes are known in the area during the eighteenth century CE. The first was in 1752 and the second in 1759 (D. Amiran, A Revised Earthquake-Catalogue of Palestine, IEJ 1, 2). It is assumed that one of these toppled the building.
At the end of the 2003 season, it was thought that the original floor of the Crusader building had been found. Yet, further excavations in Sq B showed that Mamluk squatters created several floors and installations in the Crusader building whose original floor lay far below. The Mamluk floor sequence, which was bordered by the northern wall of the Crusader building (Wall 1; Fig. 2) and by Wall 60 to the south, was composed of a white lime floor (L64), an accumulation of plaster, mud and dark ash lenses (L48) and a beaten-earth floor full of flint debitage (L47). A beaten-earth floor (L61) to the south of W60 extended southward and bore an installation that probably served as a manger (L59; length 1.5 m, width 0.4 m). It used W60 as its northern side and a carelessly built row of stones as its southern boundary, leaving the east end open.
The excavation of the Mamluk aqueduct was continued in 2004 (L35; Fig. 3). Its dismantling revealed a deeper layer, which was a deliberate fill of compact brown soil with gravel and small stones (L65). Higher up, it changed to a hard and dark stony soil (L55). These two loci (depth 0.7 m) from the Mamluk period predated the aqueduct. It proved that the original surface outside the Crusader building had not been reached and that its exterior wall stood high above the ground.