The excavation unearthed part of a peristyle courtyard (L110; c. 3 × 3 m), paved with a well-preserved white mosaic. The mosaic floor was composed of coarse white tesserae (c. 2 × 2 cm; average density 25 per sq dm), set in a thick bedding of whitish-gray mortar. Two stylobates, a western (W111) and a northern one (W112), delineated the mosaic-pavement. The northern stylobate was constructed of a single course of elongated ashlars, overlying a foundation of fieldstones and bonding material, with an ashlar pedestal at each end of the wall. The eastern pedestal was square; the western pedestal was L-shaped (Fig. 4) and may have also served as part of the threshold that was installed in the western stylobate. The pedestals functioned as pillar bases, indicating the original presence of pillars in the corners of the courtyard. The eastern stylobate was not preserved, but a robbers’ trench (L109) marked its location, retaining negative marks of ashlars with similar dimensions to those of the ashlars in the northern stylobate. To the north and west of the peristyle courtyard, white or whitish-gray mortar surfaces (L105, L106) were probably the bedding of mosaic floors that were not preserved.
A rock-hewn water cistern (L123) was uncovered below the mosaic pavement. The cistern was almost entirely preserved (Fig. 5), but was only partially excavated, without reaching the floor, due to safety considerations. The cistern’s opening (L124) in the center of the courtyard, had a recessed groove cut around it for a stone lid; a square shaft led down into the cistern. The cistern had a barrel vault ceiling, and dimensions similar to those of the overlying mosaic floor (c. 3 × 3 m, 2.2 m excavated depth from the top of shaft). The walls were coated with pinkish gray hydraulic plaster containing a mixture of grog and fine gravel. Drainage channels directed water into the cistern via opposite openings at the top of the northern and southern walls. Two channels with square cross-sections (L107, L121) converged near the opening in the northern wall. Remains of a shallow, probably open channel (W122) and a gutter (L119) on the surface next to the western pedestal, fed water into these channels. The opening in the southern wall was only partially excavated, as it lay beyond the excavation limits. The similarity of the openings in the cistern’s northern and southern walls, suggests that a similar system of water channels may have directed water into the cistern via the southern wall.  
Remains of a large round oven (tabun; L117; 0.9 m diam.; Fig. 6) were discovered to the east of the peristyle courtyard. Traces of ash from the original phase of the oven’s use were found at the bottom of the tabun. The ash remains were sealed beneath a tamped layer of leveled, light-colored soil (c. 5 cm thick). Above the light-colored soil, a layer of dark soil, containing large fragments of pottery and numerous other finds (see below), filled the interior of the oven.
 
The Finds. The latest finds retrieved on the mortar surfaces (L105, L106) and around the gutter (L119) date the final use of the peristyle courtyard. The finds include pottery, two coins, animal bones, c. 20 roof tiles, a few hard limestone grinding stones (Fig. 7), several shells, small iron nails and a bronze button-shaped object (Fig. 8). The pottery assemblage, including bowls, some imported (Fig. 9:1, 2), cooking ware (Fig. 9:3, 4) and a jug (Fig. 9:5), dates from the fourth to the first half of the fifth century CE. The absence of later Byzantine-period pottery indicates that the building was abandoned before the sixth century CE. The two coins retrieved from mortar surface L105 were identified as a coin of a Roman governor in the reign of Augustus (IAA 152897), and a coin from the time of Septimus Severus minted at Eleutheropolis (Bet Guvrin; IAA 152895). Most of the faunal assemblage was retrieved from surfaces L105 and L106, and consisted mainly of sheep/goat bones. The bones are mainly lower limb bones or teeth, consistent with butchery byproducts, although due to the small size of the assemblage, this cannot be categorically ascertained. The fill overlying the gutter (L119) yielded a sheep/goat hip bone bearing a single cut mark; a chicken bone was also retrieved in the building.
The finds from Oven 117 include numerous pottery sherds, glass vessels, three coins, a few animal bones (including a chicken bone), a lead tablet, a single olive stone and a few pieces of charcoal. The pottery dates from the late fourth–early fifth centuries CE and includes imported bowls from Cyprus and North Africa (Fig. 10:1, 2), a mortarium (Fig. 10:3), a cooking pot (Fig. 10:4), amphorae (Fig. 10:5, 6), numerous jars (Fig. 10:7–10), and an intact Bet Natif oil lamp decorated with vine tendrils on the shoulder and a geometric motif on its nozzle (Fig. 11). Two holes drilled on either side of the filling hole, were probably for hanging it up. The glass assemblage is contemporary to the pottery repertoire, including numerous body fragments, an everted fire-rounded rim, an out-folded hollow rim and handle of a lamp bowl, assorted bottle fragments and two solid bases of beakers, the latter retrieved from the dark soil (L116) sealed the southeastern corner of the building and the oven area. Solid-based beakers were not manufactured after the early fifth century CE. The three worn bronze coins date from the last quarter of the fourth century CE.
The lead tablet (c. 3 cm wide; Fig. 12) was discovered rolled up in the oven, in the upper soil fill that overlay the ashy burnt residues; its poor state of preservation precluded unrolling it in the laboratory. The tablet may have had a magical purpose; the use of lead tablets for magical purposes, is known specifically in association with aggressive forms of magic. The Book of Mysteries (‘Sefer HaRazim’) describes a magical practice that involves passing talismans or written curses through a fire-related context, and also mentions the use of lead in association with aggressive magic (Margaliot 1966:82–85, note 63). A similar phenomenon, in which a folded curse tablet was discovered buried in an oven, is known from a Roman site at Kelvedon in Essex, England (Gager 1992:195, No. 97). Curse tablets are also known from closer sites, including a lead tablet discovered in a Late Roman building in Jerusalem (Ben-Ami, Tchekhanovets and Daniel 2013). Most of the finds from the oven date from the late fourth–early fifth centuries CE, further consolidating the date of the final use of the building.
 
The finds retrieved in the excavation date the building’s final occupation to the late fourth–early fifth centuries CE. The numerous assorted finds in the oven probably indicate its secondary use as a refuse pit, or possibly reflect the intentional burial of artifacts. If the lead tablet had magical significance, the oven may have been sealed as part of a magical ritual. Due to the small size of the excavation, it is not possible to ascertain whether there was a connection between this building and the adjacent, previously excavated Late Roman villa.