During November–December 2001 a salvage excavation was conducted at Horbat Shimshit (Permit No. A-3536*; map ref. NIG 2228–33/7384–9; OIG 1728–33/2384–9), in the wake of a footpath developed by the JNF. The excavation, on behalf of the Antiquities Authority, was directed by H. Barbé (photography) and assisted by Y. Dangor (administration), V. Essman and V. Pirsky (surveying), N. Zak (drafting) and C. Hersch and I. Lidsky-Resnikov (pottery drawing) and D.T. Ariel (numismatics).
The excavation area was on the eastern slope of the Shimshit hill, along Nahal Zippori. Three areas were opened (Fig. 1): Area 10 on a small terrace set at the entrance to a cave; Area 20 on a portion of the footpath, located north of the cave
and Area 30 on an elongated embankment, whose northern side abutted a low fieldstone wall at the bottom of the hill’s slope. Its proximity to the Meggido–Zippori Roman road prompted the anticipation of an ancient roadway.
Three soundings, spread south–north, were excavated in front of the cave’s entrance, on a small terrace delimited by a low fieldstone wall. A fourth sounding was dug inside the cave (Fig. 2).
The first sounding (No. 1000; 3 × 4 m, depth c. 3 m) was on the south, against the terrace wall. Three phases were distinguished (Figs. 2, section 1-1; 3). The first consisted of a fill, divided into three distinct layers. The upper layer was a relatively thin forest humus (L1000), resting on a homogeneous layer of white limestone (L1004), which may have derived from the extraction of chalk from the cave. The lowest layer was ashy earth with many charcoal fragments (L1005). The few uncovered potsherds may be assigned to the Mamluk period, including a fragment of a jug’s neck from the thirteenth century CE, Ayyubid or Mamluk (Fig. 4:1) and the base of a Mamluk jug (Fig. 4:2).
The intermediate phase consisted of a thick fieldstone layer, including various-sized stones and bedrock collapse, bound in light beige earth (L1007). The latest potsherds from this phase were dated no later than the Mamluk period.
The lower phase was light colored hard sediment fill, comprising many small fieldstones (L1010). This phase was the richest in finds of various periods, from the Iron Age to the Byzantine period, showing a chronological continuity. At the bottom of the fill were jar necks from the Persian period (Fig. 4:3, 4), a Persian mortar (Fig. 4:5), Early Roman cooking pot (Fig. 4:6) and a Roman-period krater (Fig. 4:7). The pottery in the upper part of the fill included primarily Roman-Byzantine types from the Bet Hananiya workshop, which cannot be dated later than the beginning of sixth century CE: a krater (Fig. 4:8); Bet Hananiya, Form 4B cooking pot (Fig. 4:9), dating from the middle of first–beginning of fourth centuries CE; Bet Hananiya, Form 1B Galilean Bowl (Fig. 4:10), dating from the beginning of second to the middle of fourth centuries CE; Bet Hananiya, Form 4C cooking pot (Fig. 4:11), dating second–beginning of fourth centuries CE; Bet Hananiya, Form 1E Galilean Bowl (Fig. 4:12), dating middle of third–beginning of fifth centuries CE.
Due to lack of time, technical difficulties and safety problems, the sounding was not excavated to natural bedrock.
The second sounding (No. 1001; 3 × 4 m; depth 2.2 m) was to the northwest of Sounding 1000 and close to the cave’s entrance (Fig. 2, Section 1-1). Although the sequence of layers was somewhat different from Sounding 1000, the latter’s phases were confirmed here. The top layer was forest humus (L1001) that contained pottery fragments from the Ottoman period, such as a jug (Fig. 5). It rested on a layer of fieldstones embedded in light colored hard sediment (L1009, L1011). The lowest layer below the stones was compacted clear earth, exposed on a limited surface (L1015), which contained pottery fragments assigned to the Persian, Roman and Byzantine periods.
The third sounding (No. 1002; 2.5 × 3.0 m, max. depth 1.1 m) was located at the northern end of the terrace. It revealed the corner of two fieldstone walls (W100, W101; Fig. 2), which were one-row thick and preserved two courses high. The higher fill was brown earth (L1002), resting on a beaten-earth floor (L1006) that abutted the walls, whose foundations consisted of a single double-lined course of limestone rubble for W101 and two courses for W100. The foundations were thicker than the walls and formed an inner step, over which the beaten-earth floor was set (Fig. 6). The upper fill and the dismantled floor yielded Ottoman pottery, such as a Gaza jar (Fig. 7:1), dating to the seventeenth–beginning of the twentieth centuries CE. The fill (L1008) below the floor and overlaying bedrock was a light colored sediment that contained potsherds from the Iron Age and the Persian, Roman and Byzantine periods, including an Iron Age cooking pot (Fig. 7:2); a Persian-period jar (Fig. 7:3) and a jar (Fig. 7:4), dating to the second–fourth centuries CE.
A small sounding (1.0 × 1.5 m) was opened against the exterior facing of W100. It confirmed the chronological sequence arrived at the excavation of the interior corner. The upper fill (L1003), excavated down to the bottom of W100, comprised Ottoman pottery among its latest dated artifacts. The fill below W100, on natural bedrock (L1012; Fig. 8), contained Byzantine pottery.
The last sounding in Area 10 (No. 1014; 1.2 × 3.0 m; depth 1.2 m) was located inside the cave. The top stratum was a succession of thin chalk layers and ash, which evidenced recent occupation (L1014; Ottoman, even modern). Although natural bedrock was soon reached, it was noted that the southwestern corner of the sounding had cut a depression filled with dark ashy and powdery sediment (L1016) that yielded some potsherds, dating primarily to the Iron Age and the Byzantine period.
The evidence gathered in the different soundings of Area 10 suggests that the cave was occupied at various, relatively continuous periods, between the Iron Age and the early Byzantine period. The late Byzantine, Early Islamic and Crusader periods were completely missing and a reoccupation was attested from the beginning of the Mamluk period onward. Part of the cave’s ceiling, close to the entrance, had collapsed between the two occupation phases, sealing the ancient occupation levels and preventing access nowadays.
Four soundings were opened in this area. The first two, immediately north of Area 10 and at a similar altitude of the slope, produced identical data.
The first (No. 2002; 2 × 3 m., max. depth 0.8 m) and second (No. 2006; 2.0 × 3.5 m, depth 0.5 m) soundings consisted of an alluvium layer that overlaid bedrock and contained many potsherds of various periods.
The third sounding (No. 2005; 2.0 × 2.5 m, max. depth 1.3 m), further to the north, was composed of a relatively homogeneous brown earth fill that contained some scattered potsherds, dated exclusively to the Roman and Byzantine periods. Natural bedrock was not reached.
The last sounding (No. 2000; 4 × 6 m, depth 2 m; Fig. 9) was the most significant. Its top layer was brown alluvium (L2000) that contained numerous potsherds of various periods (Roman, Byzantine and modern), associated with a Byzantine bronze coin, minted in Cyzicus (Constantius II, 355–361 CE; IAA 97481).
The layer below the top one was a beige earth mixed with many stones. To the north, this layer contained primarily Roman or Byzantine potsherds, as well as some Middle Bronze Age potsherds. To the south, with the exception of a single glazed potsherd that could not be earlier than the Middle Ages, the finds were assigned to the Middle Bronze Age. The third layer at the bottom of the sounding was light beige sediment mixed with chalk nodules, which contained an abundance of Middle Bronze Age potsherds, except for a single rim fragment of a bowl from the late Roman period. The lack of slip and burnish in the Bronze Age fragments point to a Middle Bronze II date for the assemblage, which included bowls (Fig. 10:1–3); carinated bowls (Fig. 10:4, 5); globular bowls (Fig. 10:6, 7); a handmade cooking pot (Fig. 10:8); juglets (Fig. 10:9, 10); a jug (Fig. 10:11); jars (Fig. 10:12–14), as well as concave disc and ring bases (Fig. 10:15, 16). This layer was dated to MB II on account of the pottery, which was not accompanied by architecture.
Natural bedrock was partially reached in the area and bedrock chunks were noted in the south side of the sounding.
Two rock-cut openings, possibly of tombs or burial caves, were recorded 25 m east of L2006, at a lower slope altitude.
A single sounding (No. 3000; 2 × 3 m, max. depth 0.7 m) was opened in this area, against a low wall, built of limestone boulders (W300; Fig. 11). The stratigraphy included a top layer of relatively thin brown forest humus with a few potsherds that could be assigned to the Roman or Byzantine periods and a homogeneous layer of reddish-orange earth, mixed with small chalk nodules and practically devoid of finds. Evidence of an ancient road was not detected in the sounding, which turned out to be a terrace (Fig. 12).
The salvage excavation at Horbat Shimshit shows for the first time a MB II occupation on the eastern slope of the site, albeit without architecture. The cave indicated occupation in the Iron Age and the Persian period, which continued after a gap into the Roman-Byzantine periods and ended in the Ottoman period, whose excavated remains included a floor and structures.