Area A is characterized by significant geomorphologic changes between the levels. Eleven squares (A1, A3–A12; Fig. 2) were excavated, exposing two levels: an upper level (thickness 2 m) characterized by black clayey soil (Grumosol) and a lower level characterized by a layer of very hard hamra (Ackerman, Roskin and Sapir, below). Three pottery clusters (L111, L119, L120; Figs. 3–5), mainly jars, were discovered in the upper part of the lower level. The vessels were arranged side by side, generally in groups of three. All of the vessels were placed upside-down in the hamra, their bottom portion missing. Fragments of the missing jar bases were found scattered beside some of the vessels (e.g., a jar that was restored [B1086, L120]; Fig. 6:8). Thus, it seems that some of the vessels were brought here intact; only later was their bottom portion removed, and the fragments were discarded nearby. The latest potsherds date from the Early Roman period, indicating that the vessels were broken during this period. It is unlikely that the vessels were broken due to modern activity (cultivation or plowing) or to the stratification of the site (i.e., the pressure exerted on them by the clayey layer covering them), mainly because of the manner in which they were broken. The jars were probably in secondary use. An archaeobotanical examination of samples taken from inside the jars did not yield clear results.
The vessels in these clusters are of fairly common types. They first appeared in the first century BCE and continued, with minor changes, until the beginning of the second century CE. The cluster 111 included a bag-shaped jar (Figs. 3; 6:5) whose base was found alongside it; body sherds of other vessels (not drawn), with ware in a variety of colors, were discovered nearby. This cluster was probably damaged as a result of modern activity. Cluster 119 included three bag-shaped jars (B1091–B1093; Fig. 4) arranged side by side, forming a triangle. Cluster 120 included two bag-shaped jars (B1086, B1087; Figs. 5; 6:7, 8), a cooking pot (B1088; Fig. 6:4) and body sherds of another jar. The vessels were arranged next to each other in a row. The cooking pot belongs to a type that first appeared in the Early Roman period, and continued at northern sites until the first century CE and in the Shephelah until the early second century CE, up to the Bar Kokhba rebellion. The three clusters were found far apart each other; no connection of any sort was detected between them.
Several explanations can be proposed for the phenomenon of the upside-down jars with removed bases. The first and most plausible possibility is that the jars were used for infant burials. Jar burials, sometimes in inverted jars (below), are known throughout Israel. Generally speaking, infants were interred in jars, although jars were at times used for secondary burial of adults. In these cases, the jar’s base or its upper part was removed to ensure a larger opening. A couple of techniques could have been employed in removing parts of jars. One technique was to drill numerous holes around the vessels’ wall; another technique involved sawing the jars. Since the body fragments discovered in the excavation were extremely worn and poorly preserved, it was impossible to identify the technique used in removing the bottom part of the jars. The straight edges noted on the three jars discovered in Cluster 119 would suggest that their bases were sawn.
Jar burials dating to the Roman period were revealed in excavations conducted at Yavne (Permit No. A-6025; Eli Yannai, pers. comm.), Ashqelon (Nahshoni 2007) and Kefar Saba (Ayalon 1995); the latter were found upside down, like the jars in our excavation. Although the soil found inside the jars was wet-sieved, no bone fragments were discovered (Nagar, below). Nevertheless, it is possibility that infants were interred in the jars, but their bones were disintegrated due to the climatic conditions and the extreme post-depositional processes that transpired at the site (Ackerman, Roskin and Sapir, below).
A second possibility is that the jars were part of an industrial installation. For example, inverted jars used in installations were found near a street and an installation on the Gaza shore (Ovadiah 1969:197), where they were arranged upside-down, in several tiers. In the excavator’s opinion the jars were meant to prevent sand from penetrating into the street and the installation. This explanation is incompatible with the remains discovered in the current excavation. Even if a nearby installation did not survive, we would expect to find a larger number of potsherds.
A third option is that the jars were used as planters or were placed in the area to protect new seedlings from the wind or animals, a phenomenon known in the modern era from northern Sinai and the Hebron highlands, where large, hollow, ceramic tubes or circles of stones were placed around each small seedling.
No architectural remains were discovered apart from several kurkar stones (L113) whose function and date are unknown. Except for the clusters of jars, the ceramic finds in the area were meager and mostly worn. A Seleucid coin dating to the second century BCE (IAA 138721) was also found.
This area, like Area A, is characterized by changes in the ground level. Twelve squares (B1–B12; Fig. 7) were excavated. The kurkar bedrock, sloping eastward, was exposed in the western part of the area (L216, L222, L223, L227, L233; Fig. 7: Section 9–9). The bedrock was covered with a layer of hamra, the width of which varied according to the depth of bedrock, so that in the region where the bedrock was lower, the hamra covering was thicker. Such areas were used for burial in the Early Roman period. Seven tombs (T205, T208, T217, T219, T226, T231, T232), all aligned on an east–west axis, were documented; of these, only Tomb 205 was excavated. All seven tombs are described below according to their location in the field, from east to west.
Tomb 217 (length 2 m, width 0.6 m; Fig. 8) was covered with five large well-dressed kurkar slabs, with no bonding mortar. The covering stones may have been placed on a cist tomb. The tomb was dug into the hamra; brown soil, disturbed by modern activity, had accumulated above it. No other finds were discovered above or alongside the tomb.
Tomb 226 (length 2.3 m, width 1 m; Fig. 9), a cist tomb, was covered with large, well-dressed kurkar stones set at an angle on their narrow side, thus forming a gable. The eastern and western parts of the tomb were lined with small- and medium-sized fieldstones. A concentration of potsherds (L218) that included body sherds of a jar and a bowl dating to the Early Roman period (not drawn) was discovered c. 0.2 m southwest of the tomb. The vessels might have been placed as offerings next to the tomb or might have been used in the funerary ceremony.
Tomb 208 (outer length 1.4 m, inner length 0.9 m, outer width 0.55 m, inner width 0.26 m; Fig. 10), a cist tomb, had walls were built of small- to medium-sized kurkar fieldstones laid without mortar. No covering was discovered. The size of the tomb suggests that this might have been the burial of a young individual. No finds were discovered above or alongside it.
Tomb 219 was partially exposed. One of its walls was made of tamped gray soil. Human bones were exposed on either side of the wall, perhaps representing two or more tombs. Restrictions imposed on the excavators by the Ministry of Religious Affairs prevented further exposure of the general outline of the tomb(s). A body sherd of a jar (not drawn) placed on its side was discovered in situ, c. 0.5 m north of the tomb. A rim of an Early Roman-period bag-shaped jar (Fig. 6:6), belonging to a type that continued until the early second century CE, was also discovered. The jar and body sherds were discovered at a slightly higher level than the tomb; therefore, it cannot be determined with certainty if they were directly related to the tomb.
Tomb 231 (outer length 1.73 m, inner length 1.2 m, outer width 1.0–1.6 m, inner width 0.58 m; Fig. 11), a cist tomb dug into the hamra, was partly exposed. No covering stones were discovered. Its walls were built of medium-sized dressed kurkar stones, with no mortar. The eastern wall of the tomb was built of a large kurkar stone (length 0.7 m, width 0.5 m, exposed depth 0.17 m). No finds were discovered above or alongside it.
Tomb 232 was exposed in the balk between squares, at a higher level than Tomb 231. Human bones were discovered but not analyzed. Nearby was a concentration of tamped gray earth, similar to the soil comprising one of the walls of Tomb 219 (above). A kurkar stone (L236) discovered 0.15 m north of the tomb, might have been related to the burial, possibly having been used to cover the tomb. Since the excavation in this area was not completed, there is no further information regarding the manner of burial.
Tomb 205 (length 2 m, outer width 0.8 m, inner width 0.46 m, depth 0.4 m; Figs. 12, 13), a cist tomb dug into hamra soil, had four walls built of small- to medium-sized kurkar stones with no mortar. The tomb, covered with large kurkar stones, contained a single individual of unknown sex, 15–20 years of age (Nagar, below). Two small hewn kurkar stones were discovered beside the individual’s limbs; these were probably used to stabilize the placement of the deceased. Several body sherds dating to the Early Roman period were found in the tomb. A rim fragment of a torpedo jar (Fig. 6:10) was also discovered. Jars of this type were found at sites dating from the end of the Persian period to the Early Hellenistic period. In all likelihood, this earlier body sherd probably came from the level into which the tomb was dug and is not an indicator of the tomb’s date. Sandy hamra fill was excavated outside the tomb. Two Early Roman period jars (B2045, B2046; Fig. 6:9), discovered c. 0.5 m south and c. 1 m southwest of the tomb, respectively, were laid on their side parallel to the tomb. This type of jar was fairly common, continuing to appear with minor changes until the beginning of the second century CE, up to the Bar Kokhba rebellion. Their proximity to the tomb suggests that they were placed there as offerings or were used in funerary ceremonies.
In addition to the tombs, a rectangular pit (L238; length 1.6 m, width 0.6 m, depth 0.4 m) dug in the hamra was exposed. The pit, devoid of ceramic finds, was found filled with brown soil. Its dimensions, alignment and proximity to the rest of the tombs suggest that it was most likely a tomb that was partially dug but never finished.
Three intact jars (L201: B2082, B2083; L230: B2048; Fig. 14) were discovered among the tombs, but were not directly associated with them. The jars, like those found near the tombs, were placed horizontally on the hamra. All three are bag-shaped jars dating to the Early Roman period (continuing until the early second century CE). Body sherds of other vessels were also found on the level where the jars were discovered. These included a bowl rim (Fig. 6:1), a ring base of a bowl or jug (Fig. 6:3) and a jar rim (Fig. 6:11). The bowl belongs to the most common type in the Late Hellenistic period, continuing until the Early Roman period. The bag-shaped jar sherd dates to the Early Roman period, continuing in use until the beginning of the second century CE. Additional tombs might have been located beside the jars, but limited excavation area and pressure exerted by the Ministry of Religious Affairs prevented expanding the excavation in this area. Even though the jars cannot be attributed to activity related to a specific tomb, their proximity to the rest of the tombs and their depositional similarity to the jars found alongside the tombs suggest that these were also related to the funerary customs. They may have been meant as offerings for the deceased or were vessels used in the ritual funerary meals that took place around the tombs.
A hearth (L212; Fig. 15) was discovered right below the surface (c. 0.5 m above the aforementioned tombs). It was built of large, well-dressed kurkar stones without mortar, placed on their narrow side in a small circle. Within the stone circle was an accumulation of brown soil, devoid of finds, that covered a thin level of soot, indicating it was used as a hearth. No diagnostic potsherds that could aid in dating the installation were discovered. The hearth postdates the Early Roman period (the burial level), and its proximity to the surface would suggest it was used in a late period, perhaps in the Ottoman period.
Three squares (C1–C3; Fig. 16) were excavated. A round cluster (L301; diam. 1.3–1.4 m, height 0.7 m; Fig. 16) of kurkar stones of various sizes and several potsherds were found in Sq C1, at a depth of 1.8 m below the surface. Only worn, non-diagnostic potsherds were found in the accumulation above the cluster. In Sqs C2 and C3 an accumulation of black and brown soil was excavated. It contained a meager amount of worn potsherds and a rim fragment (Fig. 6:2) belonging to a fairly common type of bowl that first appeared in the Late Hellenistic period and continued until the Early Roman period.
Oren Ackerman, Yoel Roskin and Yair Sapir
The site is located in the coastal plain of Philistia, on the northeastern flood plain above the streambed of Nahal Soreq. The surface in this region is characterized by a convergence of hamra and Grumosol soils in the west and sand dunes in the east. The climate is Mediterranean and semi-humid, hot and dry in the summer and cold and rainy in the winter. The average annual temperature is 20°C, with an average winter temperature (in January) of 12°C and an average summer temperature (in August) of 26°C. The average annual rainfall is 500–600 mm.
In order to describe the layers at the site, the ancient surface and the geomorphologic processes that transpired after it was abandoned, three representative sections were studied, one in each area—in Sqs A7 and B4 and in Area C (Fig. 17). The layers were defined in the field according to structure, texture and color.
Area A. Three main sedimentological layers were identified (Figs. 18, 19). The upper layer (minimum thickness 0.8 m) is composed of sand mixed with Grumosol. It may be an artificial disturbance resulting from the installation of an irrigation pipeline. The middle layer is dark gray-brown Grumosol (10YR4/2 dry; 0.90–2.95 m thick), becoming thicker in the east, on the slope of the site. The Grumosol layer is deposited atop a brown sandy layer (10YR5/3 dry). The occupation level was situated at the upper boundary of the sandy layer and inside it.
Area B. A yellow-red sandy layer (5YR5/6 dry; max. thickness 0.95m; Fig. 20); the underlying kurkar bedrock was exposed at the bottom of the layer.
Area C. Three sedimentological strata were identified: an upper covering layer of yellow-dark brown sand (10YR3/4 dry; thickness 0.9 m); a middle layer of multi-colored clayey hydromorphic Grumosol (minimum thickness 0.55 m); and a sandy lower layer of dark brown clay (7.5YR3/4 dry; thickness 0.2 m; Fig. 21). Ancient remains were found at the bottom of the section.
The location of the site on a kurkar ridge, on the bank of Nahal Soreq and near the beach, where these units converge, probably affected its physical qualities. The beach contributed the sandy foundation. The clayey covering is a possible result of Nahal Soreq having flooded. The archaeological finds were found at the interface of the clay soil and Grumosol; that is, the site was covered with clay sediments. A possible source of these sediments is the flooding of Nahal Soreq to the southeast, toward the site. Two main factors were probably responsible for the floods: sand dunes that covered and blocked the wadi, and an increased intensity of the wadi’s flow which contributed to the flooding, creating swamp-like conditions for clay sedimentation. In conclusion, it seems that processes involving the deposition of clay occurred at the site which in turn led to an increase in swampy conditions in the region, which resulted in it having been abandoned.
In Tomb 205, the excavated cist tomb, cranial fragments, teeth and postcranial bones were found in anatomical articulation, indicating a primary burial. The individual was placed in a supine position, in a general east–west direction, with the head in the west. Its arms were bent in the abdominal area. Presumably, the head originally faced upward, but had fallen, facing northward. The ends of the long bones disintegrated and the closure of the epiphyses could not be discerned. A mandible fragment included an eyetooth and premolar that exhibited enamel erosion; a first molar with dentine on three of its cusps; a second molar, where the root was closed and enamel erosion was evident; and a wide hole in the jaw where the third molar may not have yet erupted. On the upper teeth, slight exposure of the dentine was evident on the central incisors; enamel erosion was visible on an eyetooth and premolar (the root is closed); dentine was exposed on three cusps on a first molar; and enamel erosion was evident on a second molar. The individual’s age is estimated at 15–20 years, based on the rate of tooth wear; the individual’s sex could not be determined.
The sand found inside the jars near the burials was wet-sieved. No bones were found during the sieving. It is possible that infants were interred in the jars, but their bones disintegrated.
A cemetery and clusters of jars dating to the Early Roman period were exposed in the excavation. In addition, a hearth and a stone clearance heap that cannot be dated were discovered. Cist tombs, some with covering slabs, and pit graves were documented in the cemetery. One tomb was excavated in which a single individual was interred, with no artifacts inside. Intact jars and sherd clusters dating to the Early Roman period were discovered next to most of the tombs. These types of pottery vessels continued to be used until the beginning of the second century CE. It seems that the finds discovered outside the tombs were associated with them. The vessels might have been placed alongside the tombs as offerings that were used in ritual funerary meals or were used by those burying the deceased during the interment. The purpose of the clusters of upside-down jars discovered c. 70 m east of the tombs is unclear. The most plausible suggestion is that they were used for burial. The absence of any osteological remains in the jars can be explained by the geomorphologic processes that occurred at the site. Jar burials are known from other Roman period sites. Most of the pottery vessels discovered in the excavation were locally produced. Quartz, characteristic of a sandy beach, was discerned in the vessel’s clay.
The excavation finds are consistent with the results of the excavation conducted at the site in 2005. The results of the two excavations indicate the existence of an extensive necropolis. Cemeteries were usually established on the outskirts of settlements. Although no settlement from the Roman period has so far been discovered near the excavation area, it was probably located not far from the necropolis. In a survey conducted during the excavation, a concentration of potsherds was discerned on a hill, c. 150 m southwest of the necropolis, possibly the location of the ancient settlement (Fig. 1).
The cemetery ceased to be used after the early second century CE. Other than the hearth discovered on the upper level of the excavation area—the result of localized activity, possibly in the Ottoman period—no remains were detected. It is unclear whether the necropolis was removed to a new location in the mid-second century CE or the settlement ceased to exist altogether. The latter supposition seems more likely in the wake of the historical events that occurred at Yavneh—a major city in the area during the Roman period—following the Bar Kokhba revolt (Taxel 2005). However, climatic events, such as flooding, may have been a contributing factor in the abandonment of the site.
The excavation findings supplement our knowledge regarding the layout of the settlement both from an archaeological-historical aspect and a geomorphologic aspect. Future excavations in nearby areas may clarify the location of the ancient settlement as well as other issues raised here.