Mikkola et al. (2008) reported that in Phase 12 “all remaining roof structures now collapsed, forming the lowest layer of stone tumble,” which included “remains of wooden roof beams, branches and clayey soil from the structures of the Phase 9 roofs.” The thickness of this stone tumble “varied significantly from one trench to another,” but averaged c. 1.5 m in the church and reached as much as 1.8 m in the chapel. Although “gradual decay and periodic earthquakes” continued to cause additional stone fall and soil accumulation, the overlying tumble layers were “less intensive.” Several rows of voussoirs from fallen arches were identified within the Phase 12 deposits in both the church and the chapel.

Beneath the “lowermost deposits of stone tumble” lay a “hard-packed, clayey soil” containing relatively few finds, which “probably represents material fallen from the structures of the roof.” Wooden roof beams recovered from this deposit were interpreted as forming “the main part of the roof construction,” while branches, covered by a thick layer of clayey soil, filled gaps and created an even roof surface. These roof materials appear to have fallen first, as they were overlain by “the arches and other stone elements of the walls.” The beams and branches were in a poor state of preservation and were heavily carbonized, apparently as a result of natural decay rather than burning.

Mikkola et al. (2008) also documented the remains of fallen arches. In the eastern part of the nave, two collapsed arches were identified in the stone tumble (loci F.04, F.09, F.10, F.11; Fig. 114): one running north–south between the pilasters F.07 and F.05d, and another apparently running east–west between the same pilaster (F.05d) and pilaster F.06 (Fig. 115). Clear remains of fallen arches were also observed in the stone tumble (loci T.05, T.08, T.10) in the western part of the aisle.

Phase 12 stone tumble and destruction layers in different parts of the church and chapel are described in detail by Mikkola et al. (2008). This Phase 12 destruction—interpreted as the final major destructive event—was dated on the basis of ceramics to the very late 9th or 10th centuries CE.