Three excavation areas (A–C; Fig. 2) were opened along Yitzhak Rabin Blvd. Three squares (A1–A3; Fig. 3) were opened in Area A, and two squares (C1, C2; Fig. 4)—in Area C. The excavation in Area B was halted due to safety reasons. The remains in Areas A and C sustained damage when drainage pipelines were installed some four decades ago. Three strata (III–I) that date from the end of the Roman period to the Early Islamic period were exposed yielding remains of a refuse pit with pottery-kiln debris from the Roman period, remains of a glass workshop from the Late Roman – early Byzantine periods and building remains ascribed to the Early Islamic period (eighth–tenth centuries CE), particularly the Abbasid period (mid-eighth century – ninth century CE).
 
Stratum III (Roman period)
In Area A, remains of a refuse pit (L11, L17, L18, L25; Fig. 5) were exposed at a depth of 1.8 m below the surface. The pit contained accumulated debris from a pottery kiln. These include domestic kraters (Fig. 6:1, 2); open cooking casseroles with a horizontal handle (ilpas; Fig. 6:3–6), one of which has a distorted rim (Fig. 6:6); cooking casseroles with a ledge rim (Fig. 6:7, 8); and cooking pots (Fig. 6:9–12), including one with extremely distorted walls and rim (Figs. 6:11; 7). In addition, Ashqelon-Gaza jars (Fig. 6:13–15), bag-shaped jars (Fig. 6:16), a juglet (Fig. 6:17) and a disc lamp (Fig. 6:18) were discovered. Although this assemblage included jar fragments, it was largely made up of fragments of cooking vessels, cooking pots and kraters—all dating to the second–third centuries CE or the early fourth century CE at the latest. Given the fragments of flawed vessels, it is obvious that a pottery kiln operated at nearby, and that some of the debris was discarded in the refuse pit. The pit was partly sealed by a floor of soft plaster, of which several sections survived (F20; thickness 5 cm; Fig. 8). The floor was cut on the north and east along straight lines, probably evidence of walls that were situated to its north and east, but did not survive.
 
Stratum II (Late Roman and Byzantine periods)
Most of the architectural remains, including the remains of an installation for glass production (L10) and a wall stump (W104), and the small finds from this stratum were exposed in Area A. The installation was square, and its walls were built of baked mud-bricks (Fig. 9). Its entire southern wall (W101) was exposed. It is bonded with two walls, on the west (W102) and east (W103), that run in a general north–south direction; their height and manner of construction suggest that they were contemporaneous. Three to four of their courses (height c 0.5 m) were identified with certainty. Due to the poor state of preservation, the installation’s opening was not exposed. Baked mud-bricks that collapsed in disarray (L26, L33; Fig. 10) covered the walls and filled the installation. These probably collapsed from the roof of the installation; however, the plan of the upper part of the installation and its roof remain unclear. Among the toppled mud-bricks were lumps of greenish blue raw glass and glass-blowing debris, indicating that the installation served for glass production (see Gorin-Rosen, below). The collapse also contained a fragment of Byzantine-period bag-shaped jar (Fig. 6:21). Under the toppled mud-bricks and above a hard accumulation at the bottom of the installation (L39) was a fragment of of bag-shaped jars (Fig. 6:22). The accumulation contained porous, lightweight lumps of baked clay, probably mud-brick chips that fell from the roof and walls of the installation due to exposure to high temperatures and baked over again after their collapse (Fig. 11). This accumulation contained two fragments of bag-shaped jars (Fig. 6:23, 24) typical of the early Byzantine period (late fourth – early fifth centuries CE), which were obviously exposed to the high temperatures of the installation, and a Gaza jar (Fig. 6:25). Once the hard accumulation was removed, the installation’s floor was exposed (L42; Fig. 12). On the floor was a folles dating to 294–298 CE (IAA 143496), which was struck at Antioch. The burn marks on the coin indicate that it too was exposed to high temperature within the installation. It thus seems that the construction of the installation should be dated to late in the third or early in the fourth century CE at the earliest. The glass fragments found within the installation have been dated similarly (see Gorin-Rosen, below). Nevertheless, since Jars 23 and 24 were obviously inside the installation while it was in use, a fourth century CE date for its construction and use is more plausible.
West of the installation was a leveled bedding (L38) that contained mud-brick debris similar to that found in the accumulation covering the installation’s floor (L39); this might be evidence of another installation that did not survive. Sandy earthen fill (L43) devoid of finds was discovered in a probe excavated below the leveled bedding. Wall 104 resembles the construction manner of the installation’ walls. Near it was a pile of toppled mud-bricks that apparently collapsed from east to west — probably the remains of the upper courses of a similar installation that collapsed.
Scant finds from this stratum were revealed in Area C. A layer of sandy soil devoid of any sherds, was found below a courtyard floor belonging to Stratum I (see below; Fig. 4). It is ascribed to the period when the settlement was abandoned in the late Byzantine period. Underneath this soil was a fill (L65) containing sherds from the Byzantine period, including a bowl (Fig. 6:19) and a cooking-pot lid (Fig. 6:20).
                            
Stratum I (Early Islamic period)
Fragmentary remains of two medium-sized buildings were uncovered (I, II). Their southern side was completely exposed, while their northern part apparently extended beyond the excavation area. The building remains were rather damaged as a result of modern construction. Building I, unearthed in Sq C2 (Fig. 2), sustained significant damage during the installation of a water pipe. Evidence of additional intrusions, which probably occurred during the Ottoman period, was found near the walls: a horseshoe (Fig. 13:1) and a key for opening a horse’s bridle (Fig. 13:2). The southern part of Building II, exposed in Sq C1, was damaged as the result of a lime pit (L54), that was set at the top of a wall from the Abbasid period (Fig. 14). A tin soldier discovered at the bottom of the lime pit (Fig. 13:3) indicates that the pit was dug at the earliest during the British Mandate. The soldier is depicted in a charging position. He wears a typical Adrian-type helmet and the uniform of an officer with a cross belt, a map case and feet leggings. The details of his clothing, particularly the helmet, indicate that the toy depicts a French officer from the period between 1915 and 1945.
Building I. Remains of two walls (W105, W107) that formed a corner were exposed. All that survived of them was the foundation of W105 and a single course of W107. It seems that most of the walls’ stones had been robbed after the building was abandoned or destroyed. The wall foundations were built of a row of dressed stones. The spaces between the stones in the wall were filled with a bit of gray mud-based mortar. A threshold stone installed in the middle of W105 was exposed in situ, indicating that the direction of entry was from the south (Fig. 15). A construction seam of fieldstones and mortar bonded to W107 was visible at the western end of W105. A floor (F63) partly built of medium–large fieldstones abutted the walls from the east. The stones were embedded in the ground and roughly hewn on the top. The floor was installed on a thin, compact bedding of small stones overlying natural sand, similar to the bedding unearthed inside the building (L64). Collapsed stones (L53) that had probably fallen from the walls of the building were found on the floor. The collapse was exposed mainly in the southern square. Below it were pottery sherds dating to the Early Islamic period, including vessels characteristic of the Abbasid period (ninth–tenth centuries CE): kraters (Fig. 16:1, 2) and a complete jug (Fig. 16:3). No paving stones were discovered inside the building; they were probably robbed like the stones courses of the walls.
Building II. Remains of a wall (W106) and sections of wall foundations (W108, W109; Fig. 17) that formed a square outline were exposed. Wall 106 was built of two rows of stones consisting of an inner face of large dressed stones and an outer face of medium-sized fieldstones. The spaces between the stones were filled with small amounts of gray mud-based mortar, as in the walls of Building I. The wall was preserved to a maximum height of one course (height 0.33 m; Fig. 18), and its foundation was set in virgin soil at a depth of 1.05 m below the surface. The foundations of W108 and W109 survived. They were built of mortar and fieldstones, as was W107, and were bonded with W106. It seems that as in Building I, here too most of the stones were robbed from the walls. On both sides of W109 were collapsed stones (L51, L60; Fig. 19) that probably fell from the walls, possibly when the walls were plundered. Pottery sherds dating from the Abbasid period were discovered. These included bowls (Fig. 16:3–7); a pale yellow-green glazed bowl that is characteristic of the period (Fig. 16:8); a deep bowl (Fig. 16:9); deep bowls with decorative band slips that is characteristic of the period (Fig. 16:10, 11); a krater with decorative band slips (Fig. 16:12); a krater (Fig. 16:13); a fragment of a storage jar decorated with opposing diagonal band slips (Fig. 16:14); and a jug (Fig. 16:15), which is identical to the one exposed on Floor 63 in Building I. A fragment of a marble plaque (Fig. 20:1) was found not in situ, in the soil fill in the middle of the building, below the elevation of the wall foundations. A fragment of a grinding stone (Fig. 20:2) and a stone pounding vessel (Fig. 20:3) were discovered below the fill in the lime pit, next to several fragments of glass vessels and bronze items that could not be identified.
 
 
Waste from a Glass Workshop
Yael Gorin-Rosen
 
Remains of an installation and waste from a glass kiln were discovered in the excavation. The processing of glass was done in two stages: in the first stage the raw glass was prepared from the raw materials, and in the second stage lumps of raw glass were melted in order to create vessels and useable objects. The kilns differ from one another in structure, size and location, as did the waste that remained after the kilns were dismantled or abandoned. The two separate processing stages are also apparent in the location of the kilns (Gorin-Rosen 2000; Gorin-Rosen and Katsnelson 2007:145–147). The remains are clearly indicative of glassware production. The size of the unearthed kiln and the amount and types of waste are more in keeping with a workshop kiln used to manufacture vessels rather than with a kiln used to prepare raw glass. Nevertheless, several lumps of debris may indicate that raw glass was produced nearby, although the amount of waste found in the excavation is insufficient for reaching a definite conclusion.
 
Area A
The Debris. The largest amount of debris that was found in the installation consists of brick fragments from the walls and ceiling of the installation (Fig 21). This debris characterizes the finds discovered in Sq A2. Inside the installation were fragments of red bricks (Fig. 21:1–3) similar to those used in its construction. A shaped brick with protruding edges that was baked in the high temperature of the kiln was also found (Fig. 21:4a, 4b). One of the brick fragments in the debris belonged to a convex brick that was presumably part of the kiln’s roof (Fig. 21:5). Chunks of bricks baked to various degrees (Fig. 21:6–10) were also found; some turned dark gray (Fig. 21:7, 9); others became porous, soft and sponge-like (Fig. 21:10); and one brick fragment (Fig. 21:11) melted due to the high temperature. Most of the fragments carry no remains of glass or vitrification. Another type of brick debris, which bears a vitrified layer, reflects the inner part of the brick, which came in contact with the high temperature and with the fumes emitted during the process of melting glass. The vitrified layer is either light or dark greenish in color, with white and gray spots. The layer is uneven; sometimes there is a trickle of vitrified material in the form of a drop that may indicate that the brick was part of the kiln’s ceiling. Chunks such as this were found among the debris fragments (Fig. 21:12, 13). Their most striking feature is the outer layer of vitrification (Fig. 21:13a). The brick’s surface is burnt or red; the cavities left by organic material that was added to the mud-mix and then burnt and consumed can still be discerned (Fig. 21:13b). Examples of bricks bearing a vitrified layer were found in excavations of glass kilns at other sites, usually in association with the production of raw glass, as in Bet Eliezer or Apollonia. However, the relatively small amount of these bricks and their shape may associate them with a smaller kiln that was used to melt glass for the production of vessels.
Another type of waste, which was found in smaller quantities, mostly in the squares adjacent to the kiln (A1 and A3), originally came from inside the kiln—glass mixed with waste. One such fragment is from the hard accumulation at the bottom of the kiln; it is covered with a layer of clean, greenish blue glass of fairly good quality (Fig. 22:1a). The bottom part of the fragment is flat and consists of a large amount of waste (Fig. 22:1b). The greenish blue color of the glass is predominant in most of the lumps that were mixed with the kiln’s waste (Fig. 22:3–5), although several lumps are of an olive-green shade (Fig. 22:6). These chips are remains of the walls and floor of the installation that came in contact with the molten glass. Some of the glass penetrated and filled the cracks in the bricks (Fig. 22:7). Most of the lumps could not be recycled nor could the glass in them be used, and hence they were found in the waste alongside the kiln. Oftentimes they were found at a distance from the kiln, as for example at Apollonia, where they were used as building material, but in our excavation the waste remained near the kiln. The quantity of this sort of waste fits a relatively small installation rather than a larger facility where raw glass was produced.
Moil fragments are the most distinctive examples of waste that can be ascribed to the production of glassware. These are glass debris that remain after vessels are detached from the blow pipes; the cut separating the vessel from the glass at the end of the pipe is visible on one side of the moil, and the uneven break that occurred when the waste was detached from the blow pipe is visible on its other side. Since this type of debris could not be recycled, as it was contaminated with metal from the blow pipe, it usually remained near the workshop, indicating that glass vessels were blown there. The moil in Fig. 23:1 was found in the balk between Sqs A2 and A3, next to a base of a cup and a fragment of a vial that was distorted by the heat (Fig. 24:7). Another fragment distorted by the heat (Fig. 23:2), probably part of a neck or a small pipe, and a fragment of a distorted pendant (Fig. 23:3) were also found.
Several flakes of raw glass were discovered among the waste fragments. Most of the flakes are relatively small (max. 5 cm); they have a triangular cross-section and are bluish green or yellowish green in shade (Fig. 23:4–6). These flakes were ready for re-melting in the kiln for preparing new vessels.
Given the size of the installation and the amount and quality of waste that was found inside and around it, it seems that its walls and ceiling collapsed inward. The amount of waste is not large, and is exactly in keeping with a medium-sized kiln. It is important to note that the amounts of both the waste that contained glass and the clean, raw glass chunks, which were found throughout the excavation area, were relatively small. The installation may have been abandoned after the craftsmen removed from it all of the material that could be re-used. It should be remembered that the excavated area was small and the area was damaged in the past; therefore, we should be cautious when drawing any conclusions. Judging by the burnt coin discovered on the installation’s floor, which dates to 294–298 CE—a date consistent with the dates of the glass vessels found in its vicinity—it is reasonable to assume that the kiln operated in the early fourth century CE. Thus, it is the earliest in-situ kiln to have been unearthed in the country, and is earlier than the kiln found at Jalame (Weinberg 1988). Waste dating to the Late Roman and early Byzantine periods from glassware workshops was found at many sites in Israel, some not far from ‘Aqir, such as Khirbat el-Fatuna (Jackson-Tal 2007) and Khirbat el-Ni‘ana (Gorin- Rosen and Katsnleson 2007:145–147, and see therein the comprehensive discussion on the subject), and waste from the Byzantine period and the Early Islamic period was found at Horbat Harmas (Elisha 2007; Gorin-Rosen 2005; 2006) and Ramla (Tal, Jackson-Tal and Freestone 2008).
 
Glassware. A few fragments of glass vessels were found. These include vessels dated to the Roman period (second–third centuries CE) and several that are presumably from the fourth century CE. Most of the glassware fragments were exposed in Sq A1. The vessels are arranged chronologically, with an emphasis on their provenance within the excavated area.
A deep bowl (Fig. 24:1) found in Sq A2 was made of translucent glass and was covered with a layer of white creamy weathering. Even though the shape of the bowl, which has a round, everted rim, appeared already at the beginning of the Roman period and continued until the Byzantine period, the quality and wear of the material are characteristic of glass from the second–third centuries CE. This glass is known from the caravan sites in the Negev, from excavations in Jerusalem and elsewhere. Vessels from this time period were found in the balk between Sqs A2 and A3 as well. One is a delicate everted cup rim made of light bluish glass and covered with golden-silvery creamy weathering (Fig. 24:4). The rim seems to have belonged to a light greenish bag-shaped cup that had a concave or delicate ring base, as that in Fig. 24:5, which is also covered with golden-silvery creamy weathering and was found in the same locus. This cup is typical of assemblages from the third century CE and is known from many sites, particularly in burial assemblages from the Galilee.
Other vessels that are dated to the early stratum were found in Sq A1, and include a bowl, a small bottle, a base of a bowl and a bottle rim. The bowl has a broad out-folded rim, is translucent and is covered with black weathering (Fig. 24:2). Bowls of this type first appeared in the Early Roman period and continued until the Umayyad period. However, the features of Bowl 2 point to a date in the Roman period, and it greatly resembles bowls found in second century CE assemblages from the Jerusalem region and the Judean Desert. The bottle, which has a light bluish hue, is decorated with very thin turquoise trails that were carefully applied (Fig. 24:10); it dates to the end of the Roman period. The bowl’s base is a low, hollow ring of bluish green glass and is covered with black weathering (Fig. 24:3). The bottle rim is rounded, has a bluish green hue and is covered with black weathering (Fig. 24:9). These vessels were found together with the industrial waste.
The base of a cup and a bottle were found together with the glass blowing waste in the balk between Sqs A2 and A3. The base is small and solid, made of greenish blue glass and is covered with silvery and black weathering (Fig. 24:6). It belongs to a relatively small cup, referred to as a ‘cup with a solid base’, that is well known from assemblages dating from the fourth century CE The bottle is plain, made of light bluish glass, with a rim that is folded slightly inward; it has a thin neck and its body was distorted by the heat (Fig. 24:7). In addition, several fragments that belonged to the upper part of a bottle were found in the balk between the squares. The bottle has a very delicate fold below the rim, is made of greenish blue glass and is covered with black pitted weathering (Fig. 24:8). The bottle was restored, and it probably dates from the end of the Roman period. This vessel was found together with waste from the kiln that contained glass remains.
Three fragments of vessels typical of the Byzantine period were found on the surface: a rim of a bowl-shaped lamp, a distorted fragment and a handle. The lamp rim (Fig. 24:12) is out-folded and has a characteristic handle. The heat-distorted fragment is industrial debris; it was found together with small lumps of kiln waste. The handle is broken; it also belongs to a bowl-shaped lamp, but of a sub-type that is characterized a\mainly by the design of the handle (Fig. 24:13).  
 
Area C
Sq C1 yieldd fragments of modern glass, including a medicine bottle and a flat liquor bottle designed to be carried in a pocket; a fragment of a pale blue glass bottle decorated with turquoise trails (Fig. 24:11) that dates to the Byzantine period; and a small lump of raw glass was also found.
Fragments of glass vessels found in Sq C2 include very small bowls, cups and beverage bottles dating from the time of the British Mandate; glass industrial waste, including two chunks of kiln waste that had glass within them, one large and one small, and a brick bearing light green-blue glass (Fig. 22:3); and two flakes of raw glass and glass working debris that looks like a distorted glass trail (Fig. 23:3–6).
 
As this was the first archaeological excavation to be conducted in ‘Aqir, the excavation is very important. Fragmentary remains of a settlement dating to three main periods were unearthed: Roman, Byzantine and Early Islamic. A pottery workshop apparently operated there during the Roman period as attested to by the distorted sherds in the refuse pit (Stratum III). In the early Byzantine period (fourth century CE), a glass workshop was built of brick walls. It seems that its location near the highest point in the settlement was no coincidence, as it is indicative of planning meant to take advantage of the wind’s direction and intensity. The workshop ceased to operate during the Byzantine period (Stratum II). Fragmentary remains of two buildings ascribed to the Early Islamic period were found (Stratum I). Their similarity in both dimensions and general orientation suggests that they belonged to one complex, which according to the ceramic finds was constructed in the Abbasid period (ninth–tenth centuries CE); the same method of construction is apparent at other sites of the Abbasid period (Arbel 2005; Haddad 2010). Further salvage excavations may further uncover the workshop that operated on the site as, well as other remains from the Early Islamic settlement.