During June 2005, a third season of excavations was conducted in the Late Roman fort at Yotvata (License No. G-6/05; map ref. NIG 2043/4217; OIG 1543/9217; Fig. 1). The excavation was directed by G. Davies of
Florida International University and J. Magness of the University of North Carolina at Chapel Hill and was funded by the Elot Regional Council, with special thanks to Dorit Banet, the Toronto Jewish Community, and the Foundation for Biblical Archaeology (Sheila Bishop). The excavation staff included B. McCane, R. Darby, S. Pryor, A. Ratzlaff and S. Werlin (area supervision), J. S. Bucko (surveying), J. Haberman (photography), E. Stegmaier (drawing), R. Halbmaier (faunal analysis) and N. Elkins (numismatics).
The excavations concentrated in two areas (2000, 5000) that were opened in previous seasons and in three areas (6000, 7000, 8000), newly opened this season.
Area 2000 (Fig. 2)
Wall 502, discovered during the 2003 season, was removed and the area was opened up as far as W520 on the west. At the bottom of the trench, above the natural bedrock concretion on which the fort was established, a succession of floor levels with associated finds, including pottery and coins, was discovered. The lowest plastered floor level (L2067; elevation 76.08 m) was associated with a built feature, consisting of stone slabs laid upright, apparently to support a bench. A pit (L2069) was found in a natural bedrock depression between the stone slabs. Approximately 30 small bronze coins and a nearly intact ceramic flask were discovered in and around the pit. This floor (L2067) was covered with the same burnt layer, equated elsewhere in the fort with a destruction event (see Davies G. and Magness J. 2005. Yotvata – 2004. HA-ESI 117). As in the other areas, a new floor had been laid immediately over the destruction horizon. Excavations in this area will continue in 2006.
Area 5000 (Fig. 2)
The trench, which was opened in 2004 in front (to the east) of the British police station, was extended eastward, exposing a passage (L5005) through the east wall of the fort (W510). The ashlar facings along the sides of the passage had been robbed out. A small semi-circular stone feature (L5011), abutting the south side of the entrance, might have supported the springing of an arch. Two walls (W525, W532), abutting W510 at right angles, extended the entrance passage into the interior.
Area 6000 (Fig. 2)
This area (5 × 6 m) is located north of Area 5000, along Wall 517. Just below the surface, three mud-brick walls (W523, W524, W530) were discovered. Together with W517 they formed an almost square room (interior dimensions c. 4 × 4 m). Wall 523, the south wall of the room (c. 1 m thick at its widest), was substantially wider than other internal mud-brick walls in the fort, found to date. The bricks were laid in a distinctive manner, with each course consisting of a central row of large square bricks flanked by half-bricks on either side (Fig. 3).
A burnt timber (c. 2 m long; elevation 76.80 m) was lying in the center of the room (L6009), in situ, probably from the collapse of the roofing system or the upper story, after the abandonment of the fort (Fig. 4). The room was filled with layers of soil that contained numerous burnt patches (width 0.5–3.0 m; depth 1–3 cm), as well as soft, white lime flecks, but no plaster floors were identified. What appeared to be remnants of organic material, possibly seeds and coprolite, were discerned in some of the burnt patches, which were apparently associated with ephemeral Bedouin activity, following the abandonment of the fort.
Area 7000 (Figs. 5, 6)
This area consisted of a standard square just east of the southwest corner tower of the fort and the adjacent postern gate. The southern boundary was formed by the south wall of the fort (W519). The main feature in this area was the remains of a staircase, located by the entrance to the southwest corner tower and the adjacent postern gate. It was analogous to the staircase by the entrance to the southeast corner tower of the fort in Area 2000 (excavated in 2003). The lower portions of the staircases were built of well-cut stone blocks, whereas the upper steps were apparently of wood, which did not survive. The staircase in this area consisted of a mud-brick platform (L7017) on a stone socle, extending northward from W519. The lowest three steps were found in situ, with the bottom step turning at a right angle to the west. The next five stone steps were discovered as they had fallen.
A badly damaged mud-brick wall (W529) partly overlaid the platform for the staircase (L7017), which itself was built over an earlier mud-brick wall (W534). Another mud-brick wall (W533) in the northwest corner of the area was perpendicular to W529 and continued into the west balk.
Two coins were found inside a mud-brick semi-circular installation, located in the angle of W534 and W533 (L7020; elevation of coins 76.89 m). A granite grinding stone with a spout was in the southeast corner of the area (L7021), partially embedded in the east balk.
Area 8000 (Figs. 5, 6)
This area is a standard square, located to the north of Area 7000. It is bordered on the west by the fort’s wall (W518), which consisted of seven mud-brick courses on a stone socle. Directly below surface, two sets of mud-brick walls were uncovered. The first set of walls (W521, W522) ran roughly east–west and abutted W518; they postdated the Late Roman period. The second set of walls (W526, W527) was well-constructed and appeared to be Late Roman (early fourth century CE). Wall 526 ran parallel to W522 and W527 abutted W521 and W526. A large pit (L8014) filled with blackened soil was discovered on the north side of W522. It was cut to a level of 77.19 m, sloping down to the east and continued into the north and east balks.
The fort appears to have had three major occupation phases, each separated by accumulated debris (c. 0.5–1.0 m thick). The earliest occupation phase dated to the Late Roman period (early fourth century CE) and consisted of a series of living layers one above the other. The earliest living layer of the Late Roman occupation phase (at c. 76.10 m) showed signs of violent destruction, followed by an immediate reoccupation. The last living layer of the Late Roman occupation phase ended with the abandonment of the fort. It is not clear when the abandonment occurred, although the numismatic evidence points to a date in the mid-fourth century CE. Thereafter, the fort was filled with wind-deposited sand and dirt. Evidence for reoccupation emerged at an elevation of c. 76.50 m. The date of this horizon is uncertain but it may have occurred during the transitional Byzantine to Early Islamic period (late
sixth to eighth centuries CE). At the end of this phase the site was again abandoned and soon afterward the upper stories and roofs of the rooms collapsed. This was followed by a long period of abandonment, during which time the fort was filled with the collapse of the mud-brick walls. Hearths and burnt patches in the layers of fill attested to ephemeral and sporadic Bedouin occupation among the ruins. Finally, at a relatively high level (77.50 m) the fort was again reoccupied. Small cells delineated by thin mud-brick partition walls were built against the ruined south wall of the fort; evidence for an occupation horizon at this level was discerned in other areas as well. We are unable to determine the date of this latest phase; the nature of the pottery and the high elevation suggest a relatively recent occupation (Ottoman to late nineteenth or early twentieth centuries CE).