This is documented in the Periplus by several examples and concerns food items , textiles and clothing, lac dye , and slaves . After the downfall of the Roman Empire and the arrival of the Arabs in Egypt in AD 640, Berenike lost its transfer function and was at the mercy of the destructive forces of the desert. Both sandstorms and supply of sand by the wadi branches north and south of Berenike buried the city in the course of time, leaving only a vague contour of the walls of the latest buildings and a variety of densely scattered trash. The lagoon south of Berenike has been filled by wadi fl ood sediments and wind borne sediments .Roman activity in the Eastern Desert lasted from 30 BC to AD 638 and is still reflected by many archaeological installations, including settlements, watering stations , forts, quarries, roads, graves, and cemeteries. This is also true for the environment of Berenike. To optimize a comparison of the sub-fossil plant remains of this harbor site with the present-day situation, it was decided to make an inventory of the present vegetation of the immediate surroundings of Berenike, the nearby watering station Kalalat, and the mountain settlements of Khesm Umm Kabu, Hitan Rayan, Shenshef, and Qariya Mustafa ‘Amr Gama . These sites will be briefl y described, starting with the northernmost site and ending with the most southerly one. Khesm Umm Kabu is located in the upstream part of Wadi Gimal . In this part of the wadi, the Romans exploited several beryl mines in the mountains. These mines are located at different branches of this wadi and are referred to as Mons Smaragdus. Surface finds indicate that the occupation period of the mining settlements lasted from the first to the fifth centuries AD . Contrary to the nearby mining settlements Sikait and Nugrus,plastic potting pots no obvious remnants of Roman buildings are present these days at Khesm Umm Kabu.

The only two ruins and an associated well are of recent date. Berenike is located on the seaward edge of the coastal plain between the branches of Wadi Mandit and Wadi Umm el-Mandit. The Ptolemaic buildings were founded on an elevated limestone outcrop that originally had a maximum elevation of well over 3 m above the current high tide . Ptolemaic as well as Roman buildings have been completely covered by wind-blown sand, which has accumulated during the course of time. With the exception of the remains of the temple of Serapis, located on the highest point of the town center, only the contours of the most elevated walls are visible on the surface . Photographs taken during an expedition in 1897 conducted by Raimondi clearly show that this temple has suffered much from its exposure during the twentieth century. Therefore, the Berenike project decided to rebury the temple. Although detailed maps are now available presenting the buildings visible on the surface and despite eight years of excavation, it is still difficult to make a reliable estimate of the total number of buildings. Demographic data, such as the size of the population in a particular period, are equally unclear. This is especially true for earlier habitation periods, as these have been built over or destroyed by subsequent activities in the town. An estimate for the last habitation phase, dated to the fifth to the early sixth centuries AD, is also difficult to make, since part of the topmost portion of the surface on the west side has been leveled by bulldozing in about 1973, when military installations were built in this area. Estimating the original size of the settlement is further complicated by the fact that it is likely that only a part of Berenike consisted of stone structures. Locally available fossilized coral was the most commonly used building material .

Temporary residence, using tents in a similar way as was practiced during the excavation seasons, could have been established perhaps west of the town center on a fl at area that covers about 7 ha. At its maximal extension, Berenike may have comprised an area of about 2 km². This area not only included living quarters, but there is also evidence of some industrial activity, including the production of metal, rope, and perhaps glass and beads . A conservative estimate of 500 persons is given by Wendrich , taking into consideration only the small buildings visible in the still-intact town center and supposing an occupation per house of an average of five persons. The location of the harbor, whether it was a real harbor or some convenient landing place, as described by Strabo , is not yet known for certain. Possibly its location shifted from north to south of the town in the course of time, as did the extension of the town center. Although some burials were found in the built-up area of Berenike, most bodies seem to have been buried outside the town. Scattered over the higher levels of the coastal plain, dozens of graves are still present. Here they are well protected against flooding, although this location could not have prevented plundering. In Wadi Kalalat, close to the mountains, a small fort and a large watering station are present at a distance of about 1 km from each other. The function of the small fort is still unclear. The watering station has a rectangular outside wall and a large, walled, circular depression on the inside. Measuring approximately 80 x 90 m, it is measured up to the largest forts of the Eastern Desert . The watering station is located on the floor of the most active part of Wadi Kalalat. At this latitude most of the water drains off underground to the Red Sea, but occasionally serious flooding may occur after heavy rainfall. Probably, the thick outside walls were built to protect the inner construction, in which the latter probably functioned as a well .

To ensure the indispensable supply of water throughout the year, Harrell suggests that a large diameter of the well should have been necessary because of the low groundwater-fl ow velocity. It is very likely that this watering station played an import role in the supply of fresh water to Berenike. Although the principal drainage course of Wadi Kalalat flows into the Red Sea just south of Berenike, the watering station is located 8.5 km southwest of the site. This western location was probably chosen to avoid the effect of saltwater penetration nearer to the sea. Partly, salt water penetrates through the soil and partly salt crystals are blown inland from dried-up land during low tide . Excavations have, so far, produced no evidence of the existence of a pipeline to transport water from Kalalat to Berenike. Because the watering station is located in the active part of the wadi, where a direct water pipeline between Kalalat and Berenike could easily be damaged by erosive fl oodwaters, it seems more likely that the water was carried to Berenike by animal transport . Hitan Rayan and Shenshef are located in Wadi al-Rayan and Wadi Shenshef, respectively, at a distance of 8.5 km in a straight line from each other. Obviously, both settlements are located west of an outcrop of the impermeable bedrock, in very shallow sand. As a result, the underground water fl ow is pushed up and slowly running fresh surface water is present . Probably because the drainage basin of Wadi Shenshef is larger than that of Wadi al-Rayan, the amount of surface water in the latter is more substantial and has never been seen dried up over the last six years, as has been the case in Wadi al-Rayan. The long-term character of surface water is supported by the presence of bulrush near Hitan Rayan and common reed and Cyperus laevigatus near Shenshef.The function of Hitan Rayan is enigmatic. Shenshef was, most likely, a satellite settlement of Berenike and was occupied during the fifth to early sixth centuries AD. The settlement area of Hitan Rayan, which is mainly concentrated on the terraces on the north side of the wadi, comprises remnants of 141 separate structures and 31 burial cairns,raspberry container growing partly concentrated in a cemetery on a terrace at the east side of the settlement and partly at isolated locations on the mountains. An analysis based on variables, such as the method of wall construction in relation to roofi ng; the number of enclosed spaces in one building, varying from one to three; and the position of components such as ovens and storage vessels to each other, suggest that the settlement was primarily used for living . Despite the abundance of quartz veins in the rocks around Hitan Rayan, there is, with the exception of one locality close to the ruins, no indication of gold mining as is evidenced more to the southwest . The Hitan Rayan settlement might have been a gold exploration camp. Unlike Hitan Rayan, there is an obvious plan to the settlement of Shenshef. There, also, the buildings are Figure 1.7. Surface water in Wadi Shenshef near a narrow gorge westward of the settlement . See Color Plates section, page 204. of a better quality, due to the drystone construction with laminar-formed stones.

The ruins of houses are present over a distance of about 800 meters and occur on both elevated terraces of the wadi that become only occasionally flooded . In total, 332 structures are mapped, representing four distinct phases of construction. A comparison with Murray’s plan of the site, reflecting the situation in the 1920s, shows that in the last decades several buildings have been destroyed . In addition to the houses, there is an enclosure, and many partly clustered graves are scattered over the area. Finally, a large hilltop fort on the east side of the settlement stretches over two adjacent hilltops from which the Red Sea can be seen. Ceramic finds date this fort to the first century AD, possibly predating the settlement. Shenshef is not located along a major roadway or are quarries or mines present in the surroundings of Shenshef, which might have explained its former function. The quarries north and south of Shenshef follow aplite dikes and were only used for the production of local building material. It is also from this construction phase that some of the houses still have walls about 2.5 to 3 m high standing, with complete doorways, windows, and internal niches. The last Roman installation to be mentioned here is Qariya Mustafa ‘Amr Gama, located in Wadi Umm Athl . S. Sidebotham named the site in honor of the Ababda, who showed them the site, whereas the wadi itself bears the vernacular name of the leafless tamarix , which grows in this Figure 1.8. Shenshef with remains of buildings along both sides of the wadi, looking northwest . See Color Plates section, page 204. wadi at the site proper. This tree, which is also present near the former settlement of Khesm Umm Kabu, is sensitive to water shortage and its presence may be indicative of water accumulation or crack seepage. The site dates to the late fifth century, possibly early sixth century, and is, thus, contemporary with Hitan Rayan and Shenshef. A nearby shady rock pool contained a reasonable amount of water during a visit in early February 1998.Rainfall in the Eastern Desert is scanty, unpredictable, and concentrated in the winter period. Short, heavy showers may produce huge quantities of water, but these cloudbursts are quite rare and localized. After such heavy rains, silty, fl at depressions are filled for weeks or even months with water and can be used for cultivating plants that are adapted to this kind of water supply, such as sorghum . Especially in the Red Sea coastal strip, where a well-developed seed bank is present owing to the atmospheric humidity, which is responsible for a more regular water supply, such erratic cloudbursts will result in a considerable increase in the vegetation cover . In his description of the Eastern Desert, Theophrastus , states that it never rains except at intervals of four or five years, and then the rain comes down heavily and is soon over. According to Hobbs , rainfall is sometimes especially abundant in a particular year or even for several consecutive years.