The excavation was located c. 2 km southeast of the ancient city of Ashqelon and the Mediterranean coast, and c. 2 km west of the course of the coastal road that linked Ashqelon to ancient Gaza (today’s Highway 4). The site is situated at the eastern edge of a narrow east–west depression, a seasonal pond known as Ha-Marzeva Pool, lying between Khirbat Khisas to its west and the former Arab village of Na‘iliya to its east (Fig. 1). The depression appears to have been part of an ancient streambed that was blocked at its western end either by sand dunes, still active today, or by the westernmost of the three north–south kurkar ridges traversing the coastal plain of Israel. When the pond filled up in the winter of 2019–2020, due to a series of exceptionally heavy rains, it became apparent that the site was located at an elevation near the maximum extent to which the depression could fill up. Probes mechanically dug to a depth of 3 m uncovered alluvial sediments of the ancient streambed containing scattered sherds from the Chalcolithic period.
The location of the excavation area is labelled “Er Rasm” on maps of the British Mandate period and appears within the declared boundaries of Khirbat Khisas. The site was identified in the survey map of Ziqim (Berman, Stark and Barda 2004:30*, Site 33), and it is mentioned in Y. Huster’s study of the hinterland of Ashqelon, in which the presence of a church was suggested based on the scattered marble fragments found on the surface (Huster 2015: Church No. 25).
Several small-scale excavations previously carried out in the vicinity of Khirbat Khisas, to the west and north of the present excavation site, revealed two Hellenistic winepresses (Lifshits 2015; Peretz 2019); four lead coffins (Richmond 1932); infant jar burials and a refuse pit from the Late Roman period, as well as Byzantine-period architectural remains (Nahshoni 2007; Eisenberg-Degen 2016); and remains of buildings from the Ottoman and British Mandate periods (Peretz 2015). A large Chalcolithic site, Agammim (East), was exposed c. 600 m northeast of the present excavation (Abadi-Reiss and Varga 2019; Varga et al. 2021).
 
Five excavation areas were opened: Areas D1–D4, forming the main part of the excavation (Fig. 2), and Area F, located c. 500 m to the southwest. The surface layer was removed by mechanical means in all areas but one, Area D1, where archaeological remains were discerned near the surface. The excavation exposed mainly architectural remains from the Roman, Byzantine and the beginning of the Umayyad periods, as well as scattered finds from the Hellenistic and Abbasid periods. Remains from the Early Roman period comprise a winepress, a fish pool, a well and a stone-built complex with facilities for producing fish sauce (garum), all in Area D4 (Fig. 3). Part of a necropolis of the Middle and Late Roman periods, and possibly still in use in the early Byzantine period was uncovered in Area D2. Part of a Byzantine-period monastic estate of the fifth–seventh centuries CE, comprising a church, three winepresses, storerooms and work surfaces, overlay part of the necropolis in Area D2 and extended into Area D4 (Fig. 3). Area D3, lying slightly southwest of Area D4, yielded only pottery kiln debris and wine-jar sherds of the middle Byzantine period, suggesting the presence of a nearby kiln. Remains of a large late Byzantine-period kiln complex for producing wine jars were uncovered in Area D1, c. 20 m southeast of Area D4. Area F, opened c. 500 m southwest of the other areas, yielded two floors with a few sherds of the Hellenistic and Byzantine periods. Additional archaeological finds were noted south of Area F but were not investigated, as they lie outside the area slated for development. Trees lining the southern side of Ha-Marzeva Pool likely mark the presence of an Ottoman-period, east–west dirt road that later turned northward and led to Ashqelon.
 
The Roman Period
Winepress and A Garum-Production Facility (Area D4). The Early Roman remains in Area D4 comprise a winepress and a stone-built complex that housed a garum-production installation and a fish pool.
The winepress (Fig. 4) had a plastered treading floor, most of which did not survive; it was laid on a bedding of interlocking ceramic pieces forming a hydra-like pattern. The settling and collection vats, both plastered, were located to its south. The settling vat was badly damaged, possibly as a result of the mechanical removal of the surface layer. The collection vat was preserved to its full height and was found completely filled with debris from the Early Roman period. The finds from the collection vat include fragments of luxury glassware; imported fine ware dating from the late first century BCE and the early first century CE; and fragments of Gaza wine jars belonging to the earliest forms of these jars produced in Ashqelon, from the first century CE. Several objects of Egyptian origin were found in the winepress: part of a zoomorphic figurine (Fig. 5:1) and a ceramic drinking vessel bearing a relief of the face of the Egyptian god Bes (Fig. 5:2) were found in the fill of the collection vat; a scarab as well as a bronze and lead pendant or weight in the form of a bull’s head (Fig. 5:3) were found near the treading floor. Two sherds of Nabatean Painted Ware bowls dating from the late first century BCE and early in the first century CE were discovered in the same area (Fig. 6). 
The stone-built complex consisting of plastered vats for garum production and a fish pool was uncovered a few meters to the southwest of the winepress (Fig. 7). Five plastered vats were uncovered: three—two small ones and one larger in size—were preserved intact, but only traces of plaster remained of the other two vats. Fragments of Gaza wine jars, of the same early form found in the winepress, were embedded in the plaster of two of the vats. Such sherds were also abundantly present in accumulations throughout the complex, with only rare finds of any other vessel type. The rim of a heavy dolium was discovered in the outlet of one of the small vats, and two parts of an imported, wide-mouthed amphora of Egyptian origin were each found in one of the two small vats. A raised, circular installation, possibly used to support a dolium, was discovered near the center of the complex. A heavy ceramic basin sunk into the center of a square-shaped room with an earthen floor was exposed immediately to its northwest (Fig. 8); the basin had a heavily eroded interior, a burnished exterior and handles bearing thumb marks (Fig. 9). A rectangular surface made of large plain tesserae was uncovered outside the square-shaped room.
Of the fish pool, only the plastered surface survived, as the walls had been completely robbed out (Fig. 10). It could still be observed that the pool had wide foundations with projecting supports. Next to the northwestern side of the pool were the remains of a well. A complete imported fine-ware bowl (Fig. 11) was found at a depth of nearly a meter below its mouth.
The stone-built complex with its multiple vats is comparable to garum-production facilities discovered in the western Mediterranean, for example at Malaga (Corrales 2017). Other features of the complex with a known connection to garum production include the dolium and the raised platform, which may have been its support; the heavy basin; and the imported Egyptian amphora. The presence of a plastered pool, probably used for raising small fish, and the adjacent well, were also part of the production facility. These installations belong to an early phase of the tradition of fish breeding, which developed in the Southern Shephelah during the Late Roman and the Byzantine periods.
 
Necropolis (Area D2). Fragments of a marble sarcophagus of the Middle Roman period were found ex situ, just below topsoil, and seem to belong to the necropolis. A gabled tomb containing a lead coffin, glass vessels and other small finds was dated to the Late Roman or possibly the early Byzantine period (fourth century CE). Oil lamps, including a 'Southern Lamp' (Fig. 12:1), imported lamps, Roman lamps with a decorated discus (Fig. 12:2, 3), Egyptian 'Frog Lamp' (Fig. 12:4) and Beit Nattif lamps of the third and fourth centuries CE, as well as other lamp fragments dated to the second–fourth centuries CE, were discovered in the matrix below the Byzantine winepresses (see below) and attributed to the use of the necropolis.  
 
The Byzantine Period
Church and Winepresses (Areas D2 and D4). The scant remains of a church and the remains of three nearby winepresses were dated to the Byzantine period. The structure of the church was almost entirely robbed of its stones; the measurements of the church could not be determined with accuracy, as the eastern end was completely robbed out. The remains of a plastered ceramic pipe, found along the northeastern side of the structure, were either part of the church or belonged to an earlier phase of construction.
The identification of the church is based solely on the east–west orientation of the wall foundations and on the numerous finds retrieved from the collection vats of the nearby presses. These include parts of marble trays (patens; Fig. 13); marble fragments bearing partially preserved Greek inscriptions; marble architectural elements (Fig. 14:1), including fragments of chancel screens (Fig. 14:2), fragments of marble altar tables, such as legs (Fig. 14:3) and floor slabs; pieces of mosaic flooring (Fig. 14:4); glass tesserae (Fig. 14:5), which probably decorated the walls of the apse; and large quantities of glass stemmed lamps (Fig. 15:1) along with a metal lamp hanger (Fig. 15:2). Isolated stone tiles, some made of porphyry, were found throughout Areas D2 and D4; these were probably part of an opus sectile flooring arranged in a geometric pattern, but it was not possible to ascertain whether they belonged to the decoration of the church or to that of the Early Roman installations in Area D4.
The remains of three winepresses arranged in a line (Fig. 16) were uncovered slightly northwest of the church. While the vats were preserved, stones from the treading floors’ walls were mostly robbed out, and only traces of their plaster lining survives, along with scant remains of a mosaic surface made from ceramic tesserae near the northwestern edge of the middle press. Each of the presses comprised two collection vats with an oval settling vat between them, the southwestern and northeastern presses with nearly square collecting vats and the middle press with octagonal collecting vats. The walls of the vats were built from grayish-white concrete, and their floors were paved with stone (Fig. 17). The settling vat of the middle winepress was likely lined with marble slabs, but these were later robbed out. A heavy, stone-carved screw well is nearly all that remained from the treading floor of the southwestern press, slightly removed from its original position in the center of the floor; the screw rested on a capital that had been part of the floor of the installation, placed there in secondary use (Fig. 18). Remains of the pressing apparatus in the other two presses comprise only plaster from the screw wells in the center of their treading floors and channels leading to the settling vats.
The winepresses and the church were dismantled for secondary use of their building stones, and the collection vats were used as refuse pits during the Abbasid period. Finds of this period were found discarded in the vats of the Byzantine-period winepresses: an oil lamp (Fig. 19:1), a bone-carved kohl container (Fig. 19:2) and a shred of a molded Buff Ware vessel (Fig. 19:3).
 
Stacks of Jars and an Installation (Area D4). A series of poorly preserved plaster floors was exposed in Area D4. Some of these floors were covered with a concentration of Gaza wine jars that had been stacked upside-down (Fig. 20). As was observed regarding jar sherds from the kiln complex in Area D1, the jars from Area D4 are of the latest form of Gaza wine jars. These vessels were apparently prepared for firing and placed over the plastered surfaces to dry after their initial shaping. Parts of a marble altar table were found among the jars, although not directly overlying the surface. Part of a marble turning wheel (Fig. 21) for making the jars was found discarded in one of the Byzantine collection vats in Area D2.
An installation comprising a settling vat and a collection vat were found near the southern end of Area D4 (Fig. 22); they were connected by way of a lead pipe. While this installation may have been a winepress, no remains of a treading surface were uncovered.
 
Pottery Kiln (Area D1). Although surface finds such as marble fragments initially indicated the presence of a church in this area, the excavation uncovered a kiln complex (Fig. 23) with three firing chambers connected by a central heat-generating tunnel (Fig. 24). Thick accumulations of sherds, comprising mainly the latest form of the Gaza wine jar (sixth–seventh centuries CE; Fig. 25), but also a 'Sandal Lamp' (Fig. 26:1) were uncovered throughout the area; a female figurine (Fig. 26:2) and a figurine of a saddled equid (Fig. 26:3) were found indise the kiln. A decrease in the scale of firing operations at the kiln appears to have occurred in the late part of the late Byzantine period, as evident by the blockage of one of the firing chambers. The complex eventually went out of use sometime in the seventh century CE. The uppermost level of one of the firing chambers was used for dumping debris in the Abbasid period (late eighth–tenth centuries CE), as evidenced by an ibex figurine or aquamanile (Fig. 27) and part of a glass vessel decorated with a medallion depicting a winged horse (Fig. 28) of this period found in the refuse pit. It is likely that during this period an extensive stone robbing from site’s Byzantine structures took place.  
 
The excavation yielded important information concerning the hinterland of ancient Ashqelon, mainly in the Roman and Byzantine periods. The Early Roman winepress and the garum-production facility appear to have been part of a flourishing agricultural estate of the first century CE. Garum production was a malodorous activity that was normally situated outside and downwind from cities. While a location nearer to the sea would have been preferable for such an industry, this was prevented by the presence of high cliffs north and south of Ashqelon. The site of er-Rasm was a practical alternative, situated not far from a main road and where water was readily accessible, given the high water table at the site.
Only a few fish pools of the Late Roman and the Byzantine periods have been uncovered along the southern coastal plain of Israel (Israel and Erickson-Gini 2013:184–185, Plan 4; Varga and Kobrin 2018), but the association of the garum-production facility with a fish pool is a unique find. According to Pliny the Elder, writing in the first century CE, raising of small fish in pools filled with rainwater for the purpose of garum production was a common practice (Pliny, Naturalis Historia, XXXI:44). The estate near Ashqelon was abandoned before the end of the Early Roman period, possibly as a result of two successive military assaults on Ashqelon and the surrounding area by Jewish rebels at the beginning of the First Jewish Revolt (Josephus, Jewish War II:457–460).
During the Middle and Late Roman periods, the excavation area was used as a necropolis, and subsequently it was the site of a monastic complex with a church, three winepresses and pottery kilns that produced a late form of Gaza wine jars—a vessel that was used in the export of wine from the Ashqelon region throughout the Byzantine Empire and beyond. Although little remains of the Byzantine church, numerous small finds uncovered near the structure suggest that it was lavishly decorated. The kiln complex belongs to a type of pottery kiln of the late Byzantine period uncovered south and west of Tel Yavne (Yannai 2014). The establishment of monastic communities was a common occurrence in the region during the middle and late Byzantine periods, and the community founded at the site likely benefited from access to prime agricultural land and to an important trade route.
The monastic complex appears to have been abandoned in the late seventh century CE, following the Islamic conquest and the collapse of the international wine trade. Its church and agricultural installations were systemically stripped of building stones and pavers in the Abbasid period. This was possibly done in order to supply material for the construction of a mosque in Ashqelon, ordered by Caliph Mahdi around 722 CE, an event documented in an inscription published by Clermont-Ganneau (1887:485–491; Le Strange 1890:400–410), though the precise location of this mosque remains unknown.