Demographics of Berenike: a cautionary tale
by: William E. Gordon

The first issue that must be addressed in the pursuit of the demographics of Berenike is establishing a definition of demographics. The Webster’s Ninth New Collegiate Dictionary defines the word demography as "the statistical study of human population especially with reference to size and density, distribution, and vital statistics". Seeing the field of demographics as an attempt to define the demography of an area, we can simplify the study of demographics into three separate, but not necessarily, mutually exclusive, groupings of: size and density; distribution and vital statistics.

Size and Density
The characterization of the size and density of the population at Berenike has been attempted in the past. These early estimates were based mainly on the physical size of the town. The earliest population estimate for Berenike was by Belzoni. He estimated that the town had about 2000 houses, inhabited by an average of 5 persons per house, resulting in a population of approximately 10,000 persons. A later estimate by Wellsted, utilized the same methodology with a slightly different result. Wellsted approximated 1000 houses, inhabited by an average of 5 persons per house, resulting in an estimated population figure of 5000 persons (Meredith 1957). These two early estimates are difficult to understand. Even if one is extremely generous in an approximation of houses from Wilkinson’s map of 1826, there seems to have been only an estimated 284 houses in the main settlement area of Berenike. This methodology entailed counting every structure, large or small as a house, ignoring the possibility of warehouses and other large non-domestic architecture at Berenike.  This clearly cannot be the case.

The present-day remains of the main town of Berenike cover an area of approximately 300 by 350 meters (Wendrich 1998). The size of the area waxed and waned over the course of the occupation period of the town and at the time of its abandonment, in the early sixth century AD, the town apparently occupied an area of approximately 200 by 250 meters (Wendrich 1998). These estimates are hampered by the fact that much of the upper 30 cm of the center of the town is gone due to widespread leveling by bulldozer in 1973 (Wendrich 1998). The number of relatively small buildings that do not appear to be warehouses, temples or public buildings number dating to the latest period of occupation equals approximately 100. Given Belzoni and Wellsted's estimate of 5 persons per house, the population can then calculated to have been approximately 500 persons (Wendrich 1998). The times of growth at Berenike, however, seem to have occurred during the first and second centuries AD as well as the fourth and fifth centuries AD. Somewhere within these periods of growth the population would have reached its zenith.

Estimates of population size made by the ecological archaeology of the 1960's and 1970's were based on the much firmer foundation of environmental factors. According to these “at peak times of optimal climate and groundwater supply the desert population of dynastic Egypt was probably substantially less than 50,000” (Butzer 1976) and "assuming maximum productivity of the Graeco-Roman period was equal to 1882 productivity…population density…can be estimated from agricultural yield" (Baer, 1962). Many of the population estimates made by ecological archaeology are linked to the work by both Renfrew and Angel in 1972 on the population of Ancient Greece. Renfrew estimates the carrying capacity of the land in Ancient Greece to be between 4.5 and 63.5 persons/km2, while Angel assumes a general population value of 30 persons per km2. Both of these figures are premised on the idea that 60-80% of the surface land in Ancient Greece was not cultivated. To employ these figures in Egypt, where approximately 75% of the accessible flood plain was cultivated, the figures from Ancient Greece would have to be multiplied by 3, to achieve population densities of about 90 persons/km2. Moreover, in post-Ramesside times it is necessary to take into account the increase in cultivable land and advances in irrigation that would raise the population estimate from about 1.6 million in Ramesside times to perhaps 2.6 million in Ptolemaic times. Although more scientific and very detailed, the limitations of this approach become painfully apparent when applied to Berenike.

Berenike seems to have had no cultivable land or agricultural yield. If the ecological archaeological models are blindly applied to Berenike, the only logical conclusion that can be drawn is that none but a few Bedouin lived in Berenike. This is clearly not the case. According to evidence at hand, the site was occupied for approximately 800 years. Given the environmental situation, water and food must have been brought into Berenike from far afield. The only possible exclusion would have been the fish that could have been caught in the Red Sea. As ship travel at the time was slow, and mostly seasonal, it is unlikely that Berenike was resupplied by boat and goods must have been transported over land, from the Nile Valley. As there does not appear to be a trail of broken pots leading towards Berenike from the wells about 10 km. inland, which were present in other settlements such as Amarna or Deir al-Medina, it seems reasonable to assume that the water was transported in leather bags, as it is still done by the Bedouin in the area. These bags may not show up as clearly in the archaeological record as potsherds would.

Based on these environmental factors, we can deduce that the estimates given for the population of Ancient Greece or even the Nile Valley proper are irrelevant to the demographics at Berenike. In contradiction to previous assertions, it appears that the oldest and simplest method seems to be the most appropriate here. It is, however, important to keep the following limitation of this method in mind. First, it is not clear how many of the small structures on site are actually single family dwellings and how many did serve another function (workshop, store room, animal pen). Second, the houses may not have housed families but large numbers of boarders. Third, it is not certain that there was a relatively fixed number of inhabitants in Berenike all year round. In the summer the population probably ballooned due to the trade activity and shipping that occurred in Berenike during the summer season, due to the trade winds in the Indian Ocean. How to exactly quantify the drop in population during the winter is still a topic of debate and there presently does not seem to be any credible way to determine this figure with any degree of certainty.

The next topic of inquiry in the definition of demography is the characteristic of distribution. Due to the leveling of a significant piece of the eastern central portion of the site in 1973 it is difficult to do anything more than address the situation in a cursory manner. One thing that can still be gathered about this area of the site is that it contained what appears to be the largest temple in the town, the Temple of Serapis. Since the site of the Temple of Serapis has not been examined by the present expedition, due to the already greatly damaged limestone walls and inscriptions, all information has to be inferred from the records of early travelers to the site. In an article on the Roman Remains in the Eastern Desert of Egypt by David Meredith the plans of Purdy and Golenisheff were reproduced (Meredith 1953). The inscriptions that they recorded only mention first and second century AD Roman emperors. In 1826, however, a closer examination by J.G. Wilkinson brought to light a textual mention of 'the land of Wawat', which was later declared to be of Ptolemaic date. Wilkinson also found a Greek text that contained a dedication to Serapis and was sent a text (by Wellsted) that contained a dedication of Ptolemy VII and his queen, Cleopatra. As an interesting side note, wooden clamps were allegedly found on the site, one with the cartouche of Seti I. They are now in the British Museum in London (Meredith 1957). Taking a cue from other contemporary temples in other locales in Ptolemaic Egypt, one can propose two things despite the bulldozer damage. First, the temple was probably surrounded by a protective wall that served the double function of defining the liminal area between the holy and the secular and providing protection in times of instability. This second protective function was probably very important in the uncertain environment of a frontier town. A second feature that was almost certainly present within this area defined by the temple wall was a storage or magazine complex that housed all of the goods necessary to conduct the everyday business of the temple. This magazine complex was probably even more important in an area of limited resources such as Berenike. In addition to this magazine complex, the temple precinct will have contained industrial and residential areas.

A magazine complex surrounding the temple in the central area of Berenike may have been found Trench 96-10 (phase IV) where the remains of an industrial were unearthed, a notion that has parallels in most Egyptian temple complexes. The remains recovered include a stone installation, crucible fragments and a large terracotta vessel imbedded in an extensive ashy layer. It lined up with evidence from Trench 94/95-1, which yielded a number of crucible fragments suggesting industrial activity in this area as well. All of these industrial areas, even if not all located within the boundaries of the temple complex, would still fit the Egyptian model of the industrial and magazine complex surrounding the temple (Sidebotham 1998). Additional industrial areas were grouped to the northwest, west and southwest of the main town. Trench 96-11, towards the west, contained the remains of Ptolemaic industrial activity . Two separate phases can be differentiated in this trench. During phase I, the area was a dumping ground for charcoal and burnt soil debris from a nearby industrial area, possibly producing fired bricks. At a certain time the burnt debris was covered by a special riverine stone fill, possibly to keep the dust down. Between phase I and II the area was abandoned which is evident from two layers of windblown sand. During phase II the area was used as a burial ground, as shown by the discovery of a very shallow inhumation burial of a 40-50 year old shroud-wrapped female. This burial contained nothing that would allow firm dating. However, the remains of the shroud seem to indicate an ancient burial. Scattered throughout the later layers of Trench 96-11 were human bones from a number of individuals that seem to come from disturbed burials in the area. All of this evidence supports the idea of the area's later use as a burial ground (Sidebotham 1998).

Another characteristic of the site are middens or trash mounds. As expected, these mounds are scattered around the outskirts of the city where they would have been of least disturbance. Three of the middens lie to the northeast of the site, immediately adjacent to a number of large buildings. Another trash dump concentration at the site in the Roman period occurs in the area to the northwest of the Temple of Serapis. In this second group of middens a wealth of garbage was found, including ostraka, cordage, leather, glass, textiles, metal fragments and botanical remains.

From these characteristics one can begin to understand the development of Berenike. It is reasonable to assume that the town began in the central portion of the site around and west of the Temple of Serapis. The dynamics of the temple establishment in Ptolemaic Egypt, as well as the grouping of the industrial areas to the west of this central area seem to reinforce this impression. We also know that the sea receded to the east during the occupation of the site and that the buildings and remains in the east of the site all date to the Roman Period. The areas to the northwest of the Temple of Serapis, as well as the areas around the periphery of the eastern portion of the site, were used to dump trash during the Roman Period. These circumstances point to the development of a Roman Period town due east of the Temple of Serapis.

Vital Statistics
The third and final area of demographic inquiry is the vital statistics of the population of Berenike. The traditional manner of defining culture in the archaeological field has been through the creation of archaeological assemblages. In this method, the archaeologist links the makeup of a certain archaeological corpus with a particular group of actual human beings in the past. Once these characteristics are defined, the encounter of another assemblage sufficiently similar to the archetype assemblage is seen as proof of the existence of that particular cultural group at that time and place. This practice has proven itself useful in the past but the methodology appeared rife with too many disadvantages to apply it the archaeological corpora without strict critical oversight. Most archaeologists will now realize that an assemblage does not automatically equal a cultural group and that the concept of cultural identity is multiply defined. Each member of any society has multiple identities that are evident at different times in different situations. In addition, one must realize that there is no such concept as linear acculturation. At any given point in time, the members of a certain cultural group can retain some aspect or aspects of a previous cultural identity in complete defiance of total enculturation. These points are particularly salient in the Hellenized and Roman world. The existence of an overarching cultural identity during the Graeco-Roman period makes it very difficult to tease the cultural or ethnic identity of the residents of Berenike out of the archaeological record. It is very difficult to even figure out what is occurring in the mortuary sphere and there does not seem to be enough evidence to make an unqualified judgment. This does not mean that the residents had no other identity than that of the Graeco-Roman nor does it mean that this cultural identity cannot by defined. It merely means that the approach to the archaeological material must be more nuanced and tentative.

There are a number of different tools that one can utilize in this quest of a more nuanced understanding. The first category of artifacts that may be of assistance are the texts written by the individuals of a particular culture. At Berenike, the majority of the texts that so far came to light are records of business transactions at the customs gate (Bagnall, Helms and Verhoogt 2000). These form a useful source of information about the inhabitants of Berenike because of the personal names mentioned. For instance, the name of Marcus Aurelius occurs more than once. He was the Roman emperor that made Roman citizenship possible for a large number of inhabitants of the Roman Empire and many of these new citizens added his names to theirs when they had to chose their 'patronymic'. The occurrence of this name in Berenike suggests that some of the inhabitants were most likely such new Roman citizens. A Palmyran inscription from Berenike does not only provide specific names of some of the Palmyran individuals at Berenike, but also links those individuals to a cultic practice (Dijkstra and Verhoogt 1999). The question remains, however, whether these individuals considered themselves primarily Roman, Palmyran or a combination.

A wide range of remains with a religious significance were excavated at Berenike. These could not always be linked to a specific group of people or even a specific cult, partly because of the eclectic nature of the religious sentiments at the time. It is clear, however, that at least a few individuals were worshipping the Palmyran god Hierobol. There is also abundant evidence for a cult which involved the offering of incense in wooden bowls (Sidebotham, Wendrich and Hense 1996).

One last category that could provide information to address this subject are domestic artifacts and especially the ceramics. Traditionally, it is within this category that one observes the most use of culturally defining material. There were a number of different types of ceramics found at the site, ranging from Indian to Egyptian and from Axumite to Italian. One type of pottery might be produced by the 'local' population (Rose 1995). This pottery may turn out to be very significant, based on the evidence that it was intentionally different from other pottery at the site, both in term of shape (rather small, carefully handmade vessels) and decoration (there is is wide variety of impressed and incised decorational patterns, some remarkably asymmetric). In addition to this 'local' ware, the excavators found Indian ware at the site (Begley and Tomber 1999). This points to extensive contacts with portions of the Indian subcontinent and the possibility of an ethnic East Indian presence at Berenike. A third type of ceramics recovered at Berenike are amphorae from Italy (Hayes 1996). It is possible these made their way to Berenike as a trade good but it is equally possible that these amphorae and the wine that they held were accompanied by merchants from Rome. At the least, it can be surmised that there was regular contact between Berenike and Roman Italy. Other than wine from Italy, numerous other trade goods are represented at Berenike. For instance, a pot full of peppercorns was found at the site. Was this an isolated trade good guided by a middleman, or had it been accompanied by an East Indian trader? Although, the present material does not allow the forceful assertion of a particular viewpoint, one should at least keep all of these possibilities in mind.

Miscellaneous artifacts found at the site include a gold earring recovered in Trench BE 97-17, glass beads from around the site, resist-dyed cloth, and teak wood found re-used in the walls of numerous structures. The earring and the glass beads, along with the child found in Trench BE 99-11, point with certainty to the presence of women at Berenike (Francis 2000). The teak wood, not attested elsewhere in the Roman world, and the resist-dyed cloth once again point to a close relationship between Berenike and South Asia (Van Waveren and Wendrich 1995).

The population of Berenike is difficult to determine due to the destruction of the central portion of the site and the need for more information about the occupants of the site. However, the distribution of the inhabitants seem easier to determine. It is altogether reasonable that the settlement began in the central portion of the site around and to the northwest of the Temple of Serapis. The dynamics of the temple establishment in Ptolemaic Egypt, as well as the grouping of the industrial areas to the west of this area reinforce this impression. One knows that the sea receded to the east during the occupation life of the site and that the buildings and remains in the eastern portion of the site all date to the Roman period. It is also certain that the areas to the northwest of the Temple of Serapis were used extensively as trash dumps during the Roman period. These circumstances point to the development and movement of the town due east of the Temple of Serapis during the Roman period. The vital statistics of the site can again only be established more tentatively. From the goods recovered, one can strongly postulate the presence of foreigners at Berenike. The fact that a great deal of foreign trade was routed through Berenike is clear from the trade goods recovered at the site. The presence of women and children is also certain from the remains.

"Berenice Troglodytica"; by: D. Meredith. In: The Journal of Egyptian Archaeology 1957; 43: pp. 56-70.

Early hydraulic civilization in Egypt: a study in cultural ecology; by: K.W. Butzer. Chicago, University of Chicago Press, 1976.

"Report on the handmade sherds"; by: P.J. Rose. In: Berenike 1994: Preliminary report of the excavations at Berenike (Egyptian Red Sea Coast) and the survey of the Eastern Desert. Leiden, Research School CNWS, 1995.

"Textiles"; by A.M.I. van Waveren and W.Z. Wendrich. In: Berenike 1994: Preliminary report of the excavations at Berenike (Egyptian Red Sea Coast) and the survey of the Eastern Desert. Leiden, Research School CNWS, 1995.

"The Pottery"; by: J.W. Hayes. In: Berenike 1995: Preliminary report of the excavations at Berenike (Egyptian Red Sea Coast) and the survey of the Eastern Desert. Leiden, Research School CNWS, 1996.

"Statuary and Cult Objects"; by: S.E. Sidebotham, W.Z. Wendrich and A.M.Hense. In: Berenike 1995: Preliminary report of the excavations at Berenike (Egyptian Red Sea Coast) and the survey of the Eastern Desert. Leiden, Research School CNWS, 1996.

"The Excavations"; by: S.E. Sidebotham. In: Berenike 1996: Report of the excavations at Berenike (Egyptian Red Sea Coast) and the survey of the Eastern Desert. Leiden, Research School CNWS, 1998.

"Fringes are anchored in warp and weft"; by: W.Z. Wendrich. In: Life on the Fringe. Leiden, Research School CNWS, 1998.

"Indian Pottery Sherds"; by: V. Begley and R.S. Tomber. In: Berenike 1997: Report of the excavations at Berenike and the survey of the Egyptian Eastern Desert, including excavations at Shenshef. Leiden, Research School CNWS, 1999.

"The Greek-Palmyrene Inscription"; by: M. Dijkstra and A.M.F.W. Verhoogt. In: Berenike 1997: Report of the excavations at Berenike and the survey of the Egyptian Eastern Desert, including Excavations at Shenshef. Leiden, Research School CNWS, 1999.

"Preliminary Report on Textual Finds"; by: R.S. Bagnall, C.C. Helms and A.M.F.W. Verhoogt. In: Berenike 1998: Report of the excavations at Berenike and the survey of the Egyptian Eastern Desert, including excavations in Wadi Kalalat. Leiden, Research School CNWS, 2000.

"Human Ornaments"; by: P. Francis. In: Berenike 1998: Report of the excavations at Berenike and the survey of the Egyptian Eastern Desert, including Excavations in Wadi Kalalat. Leiden, Research School CNWS, 2000.

Back to Introduction
Back to Contents