Stratum II
The eastern structure was built on a high hill that overlooked the city gate. Remains of two walls (W102, W111; Fig. 2), which formed the southwest corner of a building and continued east and north beyond the limits of the excavation, were discovered. Wall 102, aligned north–south, was built of basalt fieldstones and its core consisted of small flat fieldstones with gray soil (length 4.9 m, width 0.9 m). It was preserved two courses high (0.45 m) and founded on a layer of gray sediment (L112), which contained small lumps of travertine and a few potsherds from the Roman period (second century CE). Wall 111, oriented east–west, was built of basalt fieldstones (length 2.4 m, width 0.75 m) and preserved a single course high. The building’s floor that did not survive was probably destroyed in the Byzantine period (Stratum I), when the western building was constructed. 
Stratum I
A built drainage channel, dug into a layer of travertine (L105) was discovered. It crossed the area in a southwest-northeast direction and continued beyond the excavation area. Its northern part overlaid the top of Walls 102 and 111 of Stratum II (Fig. 3). The channel (width 0.2 m, depth 0.3 m) was built of limestone and basalt stones in secondary use. Its sides (W103, W104, width 0.20–0.35 m) were built of a single row of stones bonded with small fieldstones and cement. The inner faces of the channel’s sides had been coated with plaster in the past; however, only the negative imprints of the plaster were preserved on the travertine accumulation. The floor of the channel was coated with hydraulic plaster and covered with an accumulated thick layer of sediment, which made it narrower (width 0.1 m). The large amount of travertine lumps attests to the prolonged use of the channel.
The channel was closed with a layer of silt that contained a few potsherds from the Late Roman period and a coin from the years 351–361 CE (IAA 112104). On both sides of the channel, a layer of gray soil fill that contained a few potsherds from the Late Roman period overlaid the travertine deposit (L107, L114).
Remains of the western building were discovered in the northwestern part of the excavation. It was built on a slope, which was part of a stepped residential quarter, descending toward Nahal Harod. A cellar room and an open courtyard to its south and west were exposed (see Fig. 1). The cellar room (exposed length 3 m, width 3 m) was dug into the travertine and its floor was c. 1 m lower than the floor of the open courtyard (Fig. 4). The eastern (W115), southern (W118) and western (W122) walls of the cellar room were built of small basalt fieldstones and chunks of travertine (width 0.4 m), with a core of small stones bonded with cement and lime. The outer sides of the walls adjoined the bedrock, whereas their inside was lined with potsherds embedded in a lime-based plaster layer. The walls were founded on the bedrock, as was the cellar floor (L124), which abutted them. The lime-based plaster floor was composed of three layers: a bottom layer of small stones (thickness 0.1 m), a middle layer of potsherds and cement (3–5 cm) and an upper layer of lime-based plaster (thickness 1 cm). The plaster layer of the floor also scaled the cellar walls.
Three semicircular basalt steps (diam. bottom step 1.2 m, diam. middle step 1 m, diam. top step 0.6 m, height of each 0.22 m) that ascended from the basement to the open courtyard were discovered on the floor in the southwestern corner of the cellar room. 
A destruction layer, which had overlain the floor of the cellar, contained gray soil, small fieldstones and lumps of plaster, potsherds and a city coin from the second–third centuries CE, possibly from the Caesarea mint (IAA 112108).
The pottery assemblage, dating to the second–third centuries CE, included a plain krater (Fig. 5:1), a krater with a thickened and rounded rim and a ridge below the rim (Fig. 5:2), a krater with a ridged rim and plastic thumb-indented ornamentations on its sides (Fig. 5:3), a cooking krater that has ribbed sides with two horizontal handles (Fig. 5:4), spherical cooking pots that have a plain rim with a round or triangular cross-section and an everted neck (Fig. 5:5–8) and jars that have a plain, rounded or triangular rims (Fig. 5:9–13).
The boundaries of the open courtyard to the south and west of the cellar room are unclear. Two installations were exposed on the travertine floor of the courtyard. A rectangular installation (L120; 1.24 × 1.30 m), located just south of the cellar, was delimited on all four sides by a row of sedimentary rocks and preserved a single course high (width 0.15, height 0.15 m). The floor of the installation consisted of basalt debesh and was founded on the bedrock. A second installation (L121) was discovered in the northwestern corner of the courtyard; it extended beyond the excavation limits. The installation had a thin mud-brick wall (length 1 m, width and height 9 cm), with a foundation of small fieldstones flanking its sides. The installations and courtyard were covered with a layer that contained potsherds from the second–fourth centuries CE.
A tamped floor (L106, L110) that covered the top of W102 from Stratum II was exposed east of the cellar. It consisted of gray soil, small fieldstones, roof-tile fragments and potsherds and was founded on a sterile layer of sediment (L108). The floor was overlain with a loom weight, three coins (IAA 112105–107) from the second–third centuries CE, and potsherds from the second–fourth centuries CE , including cooking kraters with a sharp carination in the middle of the body and an overhanging ledge rim (Fig. 6:1–3), a cooking krater with two horizontal handles (Fig. 6:4), a ribbed cooking krater lid (Fig. 6:5), cooking pots with a triangular rim and a short neck (Fig. 6:6, 7), a cooking pot with an overhanging rim (Fig. 6:8) and jars with a rounded rim and a ridge at the base of the neck (Fig. 6:9–12).