During June 2006, a fourth season of excavations was conducted in the Late Roman fort at Yotvata (License No. G-2/2006; map ref. NIG 2043/4217; OIG 1543/9217). The excavation was directed by G. Davies, on behalf of Florida International University and J. Magness, on behalf of the University of North Carolina at Chapel Hill and was funded by the Elot Regional Council, with special thanks to D. Banet, the Toronto Jewish Community and the Foundation for Biblical Archaeology (Sheila Bishop). The excavation staff included R. Darby, C. Duncan, C. Fenwick, D. Nelson, A. Ratzlaff (area supervision), J.S. Bucko (surveying), A. Zadok (administration), J. Haberman (photography), E. Stegmaier (drawing), N. Elkins (numismatics) and M. Robinson-Mohr (registration). It is with great regret that we report the untimely passing of D. Nelson in June 2006. We are grateful to the Elot Regional Council and Qibbuz Lotan for their support and assistance at that difficult time.
The pit (L2069; Fig. 2), which was discovered last season, was defined and excavated.
Work stopped (at elevation 74.9 m) after the pit’s fill had become entirely sterile. The pit appears to be a natural solution hollow that was widened during the Roman occupation, perhaps for use as a latrine.
Work resumed inside the main, east gate to the fort (L5013) and to its north (L5012; Fig. 3). The gate consisted of a passage (L5005) through the fort wall (W517 to the north and W510 to the south). The first meter or so of this passage was constricted by opposing piers, which presumably supported an arch.
A large semicircular niche (L5015; width 1.4 m, depth 0.7 m) was cut into the core of W517 (Fig. 4). The fill of the niche yielded a large Greek ostracon (at elevation 76.93 m).
Excavations in front of the niche revealed a built, rectangular stone platform (L5025; 1.19 × 1.61 m) that incorporated re-used Nabataean ashlars. A step was built into the middle of the platform’s western front side. A series of four plastered floors, terminating with the lowest, earliest Roman floor (L5024; closing elevation 76.21–76.18 m) and covered with layers of ash, was in front of the platform and below L5012. Two piers or benches, the north one of stone and the south one of mud brick, were found next to W523 and W525, flanking the front of the platform on either side. Excavation in this area stopped at L5024, since further expansion of this area to the west was hindered by the standing British Mandatory police station.
Area 6000 (Fig. 3)
At the end of 2005, this area consisted of a square room bound on the east by the fort wall (W517), on the north by W524, on the west by W530 and on the south by W523. The continuation of the excavation in this area revealed a succession of three lime-flecked dirt floors with large burnt lenses that represented hearths, ending with floor L6034 (closing elevation 76.21 m). Two doorways in W530 and W523 were excavated. The successive floors yielded rich assemblages of finds, including 25 coins, an intact pottery flask filled with seeds, an intact oil lamp and a bone needle.
A pit (L6042) was cut into the natural concretion in the center of the room, below the lowest Roman floor (L6034). On the east side the pit was connected to a stone-covered drain (L6040/L6043). Soil samples were taken from the pit and drain to determine their use, which may have been related to the stabling of horses, mules, or camels.
Area 7000 was opened during the 2005 season in the southwest corner of the fort (Fig. 5).
The work in the single square, which was adjacent to the south fort’s wall (W519) ended with the discovery of the make-up (L7021; elevation 76.89 m) for the uppermost Roman floor in this area.
At some point during this year the site was vandalized. Due to this contamination, the area was extended to the east by 1.5 m (L7022). A mud-brick partition wall (W535), perpendicular to W519, was discovered. It defined the eastern side of an area that appears to have been an open courtyard in the Roman occupation phase, with the other sides delineated by W519, the staircase on the west (W534), and W533 on the north.
Wall 535 was constructed from two distinct types of mud bricks: those of the upper courses were rectangular in shape, softer, with less lime in their composition, and the bricks of the lower courses were thick, square, and contained more lime. The lower brick courses were set on raised stone foundations and outlined the sides of a doorway.
In the southeast corner of the courtyard, an almost complete pottery flask was uncovered in association with a floor (L7026; closing elevation 76.76 m). Another floor that consisted of a layer of thin, uneven plaster was found below it (L7027; closing elevation 76.74 m). A large hearth (L7028) was in front of the doorway in W535 and appears to continue east through it. The hearth was associated with L7027 and the next floor below, which was a thin plastered surface (L7029; closing elevation 76.62 m). Several small stake holes (diam. 2–3 cm), perhaps related to a cooking installation, were visible along the hearth’s eastern edge and inside it. Below L7029 was a layer of gravel, overlying the natural concretion on which the fort was built.
Area 7000 was extended northward, to the south balk of Area 8000. The extension (2 × 5 m) along the inside of the fort’s west wall (W518) revealed several walls. Wall 521 was perpendicular to W518 and had cut through the earlier W527. Another mud-brick wall, W540, was perpendicular to W518 and W527 and did not extend as far as the east balk. Area 7000 was also extended southward, beyond the collapsed stairs (W534) by the southwest corner tower and postern gate. Several large cut stones lined the postern gate entrance. The doorway to the southwest corner tower lined up directly with the lower stone steps in an east–west orientation.
Area 9000 (Fig. 6)
Area 9000 (4 × 7 m) is located along W518, south of the northwest corner tower. Excavation began on the west side of the area, working in terraced steps due to the sloping ground. Evidence of Bedouin occupation was found just below surface, as elsewhere in the fort. The tops of mud-brick walls began to appear at elevation 77.96 m and the area was subdivided accordingly. Wall 536 was a north–south wall east of W518; Wall 537 was an east–west wall perpendicular to W518 and W536 and Wall 538 abutted W518 and ran east–west along the south balk. The southwest section of Area 9000 (Loci 9014, 9020, 9021, 9024; closing elevation 77.21 m) yielded a rich assemblage of pottery, including an intact flask, glass, animal bones, and coins, which apparently represented a dump.
Area 1000 (Bathhouse) (Figs. 8)
Earlier excavations by Z. Meshel in the bath house, to the north of the fort, brought to light portions of the caldarium and tepidarium. The goal of the renewed work is to clarify the plan of the bath house and conserve the site for public display (Fig. 7).
Three areas were opened: an area (3.5 × 5.0 m) to the west of and partially overlapping the area of the caldarium; an area (length 8.5 m, width 6 m) to the east of and partially overlapping the tepidarium and a third area (length 4.45 m, width 2.2 m) that exposed the full extent of the tepidarium’s south wall (W542).
Inserted through the west wall of the caldarium (W543) were the badly damaged remains of the praefurnium flue, clearly identifiable by its vaulted tile construction. On either side of the flue were the remains of the base for a hot tub. Immediately west of the praefurnium was a large, semicircular, plastered pit (L1011), filled with ash, apparently from the furnace.
Excavations east of the tepidarium revealed the inner face of the bath house north wall (W545), which was covered with a thick layer of white plaster. Six square piers constructed from mortared rubble were aligned east–west in two parallel rows. Two of the three piers in the north row were plastered on their north sides. A thickly plastered surface (L1008) covered the northern half of the area and continued up the inner face of W545. It seems likely that this area represented the cold plunge bath (Fig. 8). How far it extended southward is still unclear. However, the presence of thick plaster only on the north face of the two north piers may indicate that the cold plunge bath occupied only the area between those piers and W545. A plastered step led from the tepidarium to the cold plunge bath. The outer face of W542 was thickly plastered, perhaps indicating the presence of another cold plunge bath in this area.
Our excavations indicate that the bathhouse does not conform to the military reihentyp that consisted of a single row of rooms, but forms a larger complex, the excavation of which will continue in 2007.
Three major occupation phases occur in the Yotvata fort: Late Roman (first half of the fourth century); Early Islamic (late sixth or early seventh century to eighth century) and a recent phase (perhaps Ottoman). The late Roman phase has four successive occupation levels, which yielded dozens of identifiable bronze coins, dating, so far, to the third century and first half of the fourth century, with the exception of a single Nabataean coin of Aretas IV from Area 2000.
Perhaps the most important and intriguing find of the 2006 season were the semicircular niche and platform in Area 5000. The niche’s orientation to the east and the fact that the Greek ostracon found in it appears to include a Chi-Rho symbol suggests that this room might have been used as a chapel. The identification of the Chi-Rho symbol was made by L. di Segni, who is studying the ostracon together with H. Cotton. The niche and platform were not originally part of this room, but were added in the second Roman phase.
The Early Islamic occupation phase included three levels, which could not always be distinguished in the course of excavation but were clearly visible in some of the balks. These levels consisted of relatively ephemeral remains of dirt floors, hearths and thin mud-brick walls. Some of the best evidence for the succession of Early Islamic levels can be seen in the south balk of Area 3000 and in the doorway of W530 in Area 6000.
The dating of the latest uppermost occupation phase has not yet been firmly established; it is hoped that radiocarbon samples will provide an answer to this question.