The site of Avdat, situated in the central Negev highlands, was occupied from the Late Hellenistic until the Late Byzantine periods.  During four months in 1993 we excavated two areas of the site. The main goal of the dig was to conduct a detailed study of  the destruction processes in order to differentiate between dynamic and static events of collapse, and to determine the extent of the influence of  seismic events on  the settlement history of Avdat in the mid-fourth and the beginning of the seventh centuries AD.

During the excavation we were able to establish at least ten criteria that are indicative of seismic events. The main criteria we will be present here are related to: 1) roof and arch collapse, 2) walls breaching and pilaster twisting, 3) Deformation of doorposts and doorsteps and knocked down lintels.

During the excavation it became clear that the collapse pattern in the structure was uniform and was expressed in the systematic collapse of the arches and the roof slabs directly on the floors of the larger rooms (rooms 2, 4 and 7).  In the smaller rooms (rooms 5 and 6) some arches remained in situ but it was possible to detect  the beginning of slippage of  some arch stones that nonetheless did not result in collapse.  In our opinion the arches in the smaller rooms escaped damage due to the fact that the force exerted upon them was not powerful enough to stretch the arch in a way that  would free the individual stones and allow their collapse,  in contrast to the fate of  the longer arches in the larger rooms.  In several rooms (rooms 4, 5 and 7) it was possible to detect  movement and twisting of the pilasters.

Differences in the level of damage that was uncovered in the smaller and larger rooms were also evident in the longer and shorter walls. Long walls situated on the eastern side of the structure (rooms 4 and 7) were breached and collapsed on top of the roof slabs that fell probably seconds earlier. It is interesting that the western walls in which the entrances to the rooms were situated were less affected. It is possible to explain the difference in the extent of the damage between the two as being due to a lack of flexibility of the eastern walls that contained no openings. Additional support for this assumption may be found in the inclination and twisting of the eastern wall of the courtyard in which the stones of the eastern face of this wall were peeled away and fell, face down, into the courtyard. 

Further evidence of the direction and force of the earthquake could be detected in the manner in which the lintels of  entrances located on the eastern wall of the courtyard were thrown several meters northwest into the courtyard as well as twisting of the doorposts and cracks in the doorsteps of these entrances. 

Supporting evidence for the direction of the quake was found in room 3. This is the only room in which the arches were perpendicular to the other arches in the building. It was therefore expected that the destruction pattern for the arches in this room would be different and confirm our working hypothesis. And indeed, this was found to be the case:  whereas in the other rooms the arches fell directly onto the floors in the case of room 3 the roof slabs and arches fell at a 45 degrees angle concentrated up against its northern wall.

On the basis of this and additional evidence found throughout the building we suggest that the destruction of the building was a single, sudden event and not a result of a slow continuous process.

Excavations along the Byzantine city wall uncovered underlying structures dated to the Late Roman period that offered evidence of an additional seismic event, probably the well known earthquake of 363 AD. This important evidence will not be discussed here for lack of time. However it may be concluded that Avdat was hit by at least two major earthquakes the first of which, in 363 AD. terminated the Late Roman settlement of the city. Several years later, a new town was rebuilt on the ruins of the earlier one that is reflected in the stratigraphy described above. The new Byzantine town existed until the beginning of the seventh century AD, probably 633 AD, and was then totally destroyed by an earthquake never to be rebuilt.

This may explain the absence of any Early Muslim period finds at the site in spite of the continued occupation of other Negev sites such as Nessana and Shivta that existed until the mid-ninth century AD. These towns, located west of Avdat and further from the Syro-African Rift valley, were probably less affected by the earthquake.