Dakhla Oasis, prehistoric sites
Dakhla Oasis (centered on 25°30′ N, 29°00′ E) is located in the Egyptian Western Desert, halfway between the Nile Valley and the Libyan border, at roughly the latitude of Luxor.
The largest of several Western Desert oases, Dakhla is a depression 70km long (east-west) by 20km wide. The oasis, bounded on the north by a 300m high plateau, is divisible into three zones north to south. The piedmont zone slopes southward from the base of the plateau to the central lowland, which is marked by a discontinuous belt of cultivation fed by artesian wells (the only water available today in this hyperarid area). South again, the third zone, with fossil spring terraces and spring mounds, old playas (ancient lakes) and sandstone ridging, slopes upward to the desert plain beyond.
Aside from the mention of a few stone tools in a 1936 publication by H.E.Winlock, the study of Dakhla prehistory began only in the 1970s. In 1972 members of the Combined Prehistoric Expedition (CPE), led by Fred Wendorf of Southern Methodist University and Polish archaeologist Romuald Schild, visited Dakhla as part of an archaeological reconnaissance of the southern half of the Western Desert. While in Dakhla they excavated two Pleistocene spring vents in the eastern lowlands. Then in 1978, the Dakhleh Oasis Project (DOP), with Canadian archaeologist A.J.Mills as field director, began its investigation of human adaptations to changing environmental conditions within the oasis throughout prehistoric and historic times. The DOP divides the prehistoric sequence into Pleistocene and Holocene portions, with M.R. Kleindienst responsible for the former, and M.M.A.McDonald for the latter portion.
Dakhla Pleistocene prehistory begins with the appearance of the first hominids in the area over a quarter of a million years ago and persists until the end of the last Ice Age, about 10,000 BC. Holocene prehistory runs from that date to about 2200 BC, when immigrants from the Nile Valley brought elements of late Old Kingdom civilization to Dakhla.
A problem shared by Pleistocene and Holocene prehistorians in Dakhla and elsewhere in the Western Desert is that sites are usually severely deflated. In these arid areas, the wind over time removes all but the most consolidated of deposits, plus most organic material including food remains and datable remains such as charcoal. Often all that remains are surface scatters of stone tools and occasional hearth stones.
The problem is particularly severe at sites of Pleistocene age, where the sometimes extensive scatters can be redistributed or mixed with later material. Accordingly, the focus in Dakhla, with some exceptions, has been less on finding localized "sites" than on mapping the distribution of artifacts across the landscape, relating this to geomorphic units, changing paleo-climates and potential resources. The Pleistocene geomorphological sequence, defined by DOP geographer I.A.Brookes, includes erosional episodes which left three gravel-bearing pediment remnants in the piedmont zone, labeled, from oldest to youngest, P-I, P-II and P-III. A sequence of lacustrine laminated sediments falls between P-II and P-III in time, while several episodes of artesian spring activity, within and just south of the central lowlands, have left behind extensive sheets of water-deposited sediments, as well as spring mounds or vents at points where the water surfaced. Kleindienst has been running a series of archaeological survey transects north-south across these geomorphic regions and into the desert beyond, sampling lithic artifact distributions on each, in order to determine human land-use patterns and changes in those patterns through the Pleistocene. In the absence of chronometric dates, artifacts are dated from their association with units of the geomorphic sequence, and through comparisons with archaeological sequences elsewhere in this part of Africa, notably that worked out by Gertrude Caton Thompson for nearby Kharga Oasis.
So far, several Pleistocene cultural units have been identified from analysis of the stone tools. They can be classified as either Early Stone Age (ESA), traditionally characterized by the Acheulian handax or biface, or Middle Stone Age (MSA), with its Levallois or specialized core preparation technique. The earliest materials identified so far in Dakhla are a few distinctive handaxes found on P-II gravel surfaces and the flanks of a spring mound. The handaxes are large, usually of quartzite rather than flint, and worked around their entire circumference. Typologically they are "Upper Acheulian," and might be 400,000 years old.
The next well-defined unit, called the "Balat," also features bifacial tools, but of a very different kind. Mostly of chert, they are small (less than 160mm long, mean circa 100mm), with thick unworked butts and trimming confined to the tip and one or both side edges. There is little evidence for the use of the Levallois technique or core preparation. It is Balat unit material that the CPE excavated in 1972, recovering hundreds of bifaces and other tools at two spring mounds. While Balat unit artifacts are commonly found on pediment surfaces and elsewhere in the oasis, the only other in situ finds are from river gravels of probable P-II age. The Balat, on analogy with East African material, might be very late Early Stone Age or, more likely, a transitional ESA/ MSA industry, and appears to be well over 100,000 years old.
For the Middle Stone Age, several units have been defined, distinguishable in part by the size of artifacts and by site locations within the oasis. Present as well are two specialized groupings of stone tools, the "Aterian" or "Dakhla unit," and the "Khargan." As before, the evidence is largely from surface scatters, but now specialized sites/workshops, living sites and lookout points can sometimes be detected. A "large-size" MSA unit, featuring specialized cores averaging 90mm in length, and long, lanceolate bifacial tools is, on analogy with the Khargan sequence, early MSA. A probably younger "medium-size" MSA unit (mean artifact sizes 70-75mm) is, like the large-size unit, found mostly on P-II and other northern gravel surfaces.
The Aterian is a distinctive North African stone tool industry featuring tanged implements of various kinds as well as bifacial lanceolates and specialized cores. In Dakhla, the Aterian or Dakhla unit is divisible into at least two variants, based in part on artifact size. The larger variant, featuring implements up to 150mm long, has been found on the piedmont, associated with post P-II sediments and P-III gravels, and on an occupation site in the desert well south of the oasis. Similar material occurs at Adrar Bous, 2,100km to the west in the central Sahara. The smaller variant, with flakes ranging to 110mm, occurs as knapping sites on P-II gravels and as scatters in central and southern Dakhla Oasis. It resembles the Aterian of Kharga Oasis, and may be less than 50,000 years old. A still smaller MSA unit, found in Dakhla on younger surfaces and spring deposits, and perhaps the equivalent of Caton Thompson’s "Khargan," has yielded a date of 23,000 BP (years before present) at Dungul Oasis in southern Egypt.
One intriguing finding at Dakhla Oasis is the still somewhat fragmentary evidence for continued occupation of the oasis throughout the late Pleistocene; studies elsewhere in the area have suggested abandonment of the desert in the hyperarid period 50,000-12,000 BP.
The last three prehistoric cultural units identified in Dakhla Oasis are of Holocene age. These sites are also severely deflated but, due to late prehistoric cultural innovations, more categories of evidence are now available, including grinding equipment, small finds of stone and shell, stone shelters, pottery and rock art. Also, fragmentary in situ deposits yield such organic material as bone, plant remains and charcoal for radiocarbon dating. Moreover, the climate can be reconstructed with some accuracy: generally, the Sahara was more humid than it is today through the Holocene, until about 3000 BC.
The three Holocene prehistoric cultural units in Dakhla are the "Masara," dated 72006500 BC, the "Bashendi," 5700-3250 BC and the "Sheikh Muftah," which begins during the Bashendi and survives to overlap with the Old Kingdom occupation in the oasis.
"Masara" is the local name for a cultural unit elsewhere called the "Epi-paleolithic." Epipaleolithic sites, scattered from the Nile Valley westward across the Sahara, tend to be little more than sparse clusters of lithics, the products of small, highly mobile groups of hunter-gatherers. Similar small sites are found in Dakhla as well, where they are labeled "Masara A." In Dakhla, however, the picture is complicated by the presence of contemporaneous sites of another kind. "Masara C" sites, in addition to lithic artifacts, feature clusters of stone rings, anywhere from 2-20 per site. These stone rings, 3-4m in diameter, oval or bi-lobed, are interpreted as bases of hut structures. They suggest somewhat more settled groups than at other Epi-paleolithic sites, an impression reinforced by the evidence for a wide variety of activities performed at these sites, from storage to bead making, and by their reliance on inferior but locally abundant lithic raw material. While Masara A sites are found across the oasis and even atop the northern plateau, Masara C sites are confined to one well-watered spot on the sandstone ridging in the southeastern corner of the oasis, an area that was also heavily settled by later Bashendi groups. The Dakhla Masara C sites seem unique within the eastern Sahara for that time: it is another 500 years before the next group of relatively settled sites appears, at Nabta Playa in southern Egypt.
The next Dakhla cultural unit, the Bashendi, is divisible into two phases, A and B, on the basis of site location, artifact inventories, subsistence and age. While Bashendi sites occur throughout the oasis, the fullest record comes from the large basin and ridges in the south-eastern corner of Dakhla in the vicinity of the Masara C sites. Bashendi phase A sites consist of extensive scatters of hearths and artifacts eroding out of playa silts in the basin floor. Artifacts include fine bifacial knives, a variety of arrowheads (including hollow-based, leaf-shaped and tanged forms), grinding stones, abundant ostrich eggshell beads, lip-plugs of barite and rare pottery. While the assumption was that these were the campsites of pastoral nomads, in fact, all animal bones identified so far are of wild species. Radiocarbon dates are from 5700 to 5000 BC.
One anomalous kind of site dates to the very end of the Bashendi A sequence. A group of stone ring sites, one consisting of 200 structures, occurs on the ridge adjacent to the large basin. In addition to hunting, people on these sites seem to have herded goats. Phase B campsites are found on the basin edge, above silt level. Characteristic artifacts include, besides knives and arrowheads, (side-blow) flakes, planes, small polished axes, amazonite beads, and marine shell pendants and bracelets. Faunal remains, mostly of cattle and goat, suggest a heavy reliance on domesticated animals. Phase B spans a millennium, starting at 4550 BC.
Many of the characteristic artifacts of both phases A and B, including knives, arrowheads and many of the small finds, are shared with Neolithic and Predynastic sites in the Nile Valley, from Khartoum to the Delta, and also with Neolithic sites far to the west across the Sahara. Interestingly, though, the Dakhla occurrences are older than dated examples from either of the other two regions.
Sites of the third Holocene unit, the Sheikh Muftah, are located much closer to the oasis central lowlands, where they are often obscured by later cultural material. There is still no evidence of permanent settlement, although pottery is abundant and copper was used. The unit survived to overlap with the Old Kingdom presence in the oasis after 2200 BC.
The picture emerging from the study of Dakhla prehistory is not so much that of an oasis isolated within a vast desert, as one with at least occasional far-flung contacts: with neighboring oases and the Nile Valley, with sites westward across the Sahara and with sub-Saharan Africa. Apparently large enough to support life even during a hyperarid period when the rest of the eastern Sahara was deserted, Dakhla Oasis seems to have served sometimes as a node on communication lines crossing the desert, sometimes as a meeting point for desert-adapted cultural traditions, and occasionally, as in mid-Holocene times, as a center for cultural innovation. In this last role, as cultural innovator, the Dakhla Bashendi unit, through its contact with the Nile Valley, appears to have contributed to the early stages of the development of Egyptian Neolithic and Predynastic cultures.
Dating, pharaonic
The chronology of pharaonic Egypt is based on a sequence of thirty-one dynasties, or ruling families, as defined by Manetho, an Egyptian priest who compiled a history of Egypt in the third century BC. While modern study has shown that Manetho’s work is incorrect at many points, his basic dynastic structure with the appropriate changes is still used today. Manetho’s dynasties and lists of kings echo those of earlier times. The earliest king-list we now possess is the fragmentary Palermo Stone which, in its original state, named the kings of Egypt up to the 5th Dynasty and important events that took place during their reigns. Another king-list is found on a fragmentary papyrus in Turin, originally a catalog of Egyptian kings up to the later 19th Dynasty with the regnal years of each. Other king-lists were drawn up for various reasons. The best known is the long roster of kings receiving offerings inscribed in the Abydos temples of Seti I and Ramesses II. Other lengthy lists come from private tombs, again for cultic purposes or as footnotes to long genealogies of high officials. Scores of shorter sequences of kings are found in tomb inscriptions and administrative documents.
The sum result is that, except for the more obscure periods of Egyptian history, we have a workable list of families of kings for the entire thirty-one dynasties of pharaonic history. Thousands of religious, administrative and private documents dated to a specific year of a given king have helped fill in the lengths of reigns. While this element in the dynastic structure is still far from perfect, the chronological skeleton is there. The next step is to translate the dynasties, the lengths of royal reigns and the multitude of documents dated to these reigns into an absolute chronology in terms of dates BC.
The background for such an absolute chronology is the Egyptian calendrical system. As any society must, the Egyptians kept track of units of time—days, years, seasons and the like—for the requirements of both religion and administration. For this, they created what at first sight appears to be a conflicting pair of calendars, lunar and civil. That these were never in synchronism presented no problem since the two calendars served different purposes.
The most obvious method of gauging time, dating back to prehistoric times, was the simple observation of the seasons created by the annual phases of the Nile River. A period of inundation of the valley was followed by a growing season, in turn followed by a dry period when the Nile was low. But the onset of the inundation which began the agricultural year could occur at any time within a period of several weeks. The length of time from inundation to inundation therefore fluctuated, and any given agricultural year could be longer or shorter than the one before and after. Such a time frame was sufficient for agricultural purposes, but for nothing else.
The more precise measurement of time required for religious festivals was accomplished by observing the phases of the moon, also a very early development. This lunar calendar was divided according to the three agricultural seasons, each of which lasted approximately four lunar months. The resulting twelve-month lunar year averaged 354 days, as each lunar month is 29 or 30 days long. The names of the agricultural seasons—inundation, growing, dry—were retained and the months were named after the most important feast that took place in each. But this lunar year was also tied to the sidereal year in which the heliacal rising of the star Sirius, or Sothis, played a major role. Each year for a period of seventy days, Sirius is hidden by sunlight. The day when the star can again be seen in the eastern horizon just before sunrise is its heliacal rising, called "the coming forth of Sirius" by the Egyptians. New Year’s Day in the lunar calendar was the first day of the lunar month following the annual heliacal rising of the star.
It seems likely that the reappearance of Sirius was chosen as the herald of the New Year because this event took place about the time the inundation of the Nile began each year. Since the length of the lunar year was shorter than the sidereal year of 365.25 days, a thirteenth lunar month was added every three or four years which kept the two in general synchronism with each other and with the agricultural seasons. Such a method of reckoning time served the needs of religion, though it was too flexible for administrative requirements.
To fill the latter need, a civil calendar with a fixed length of 365 days was introduced shortly after Egypt was first united under a central government. Various theories suggest the 365 days arose from the average length of a series of lunar years, or the average length of a series of agricultural years, or simply the period between heliacal risings of Sirius. Whatever its origin, the civil calendar adopted the three seasons and the twelve months of the lunar year, each month now fixed at thirty days. Five extra, or epagomenal, days were added at the end of the year to fill out the 365-day total. This provided a calendar that was perfectly regular and without the fluctuations of the lunar calendar.
Dates were recorded as "Year 2, month 3 of Inundation, day 16 (of King X)." It did not trouble the Egyptians that this "month 3 of Inundation" could occur during the dry season of the natural year for they understood this simply as "month 3 of season 1." In the civil calendar, the three "seasons" were only traditional names for three segments of the civil year which, from its inception, had nothing to do with agriculture. The civil calendar became the medium by which all documents and events were dated and provided a simple and uniform method for keeping administrative records.
It must be emphasized that while the civil calendar of fixed length and the lunar calendar of variable length were used concurrently, they were not opposed or in competition with each other, but were used for entirely different purposes. The lunar calendar established religious events such as feast days and sacrifices. The civil calendar was for the ordering and recording of daily life. Judaism and Islam still use both a lunar and a civil calendar for the same reasons.
From our viewpoint, there is a major flaw in the civil calendar. Its 365-day year fell just short of the sidereal year of 365.25 days. This means that every four years the civil calendar fell one more day behind the sidereal year. Dubbed by modern scholars "the wandering year," the civil year regularly progressed backward so that its first day eventually fell on every day of the sidereal year. The resulting period of 1,460 years (365×4) is called the "Sothic Cycle," that is, the length of time between concurrences of New Year’s Day in both the civil and sidereal years. But this is a modern measurement of time, unknown to the Egyptians who always knew their civil year did not correspond to either the lunar year or the annual appearance of Sirius. Since this was not a problem to them, they never took steps to bring the two calendars into synchronism.
It does present a problem to modern historians, for the documents they must use are dated by the civil calendar with its slightly shorter years, whereas our own absolute chronology of Egypt must be expressed in terms of the sidereal year if that chronology is to make sense to us. Synchronizing the Egyptian calendars is therefore a primary task of present-day scholarship. One important help in creating that absolute chronology are the rare instances in which the Egyptians recorded a heliacal rising of Sirius on a particular day of the civil calendar. There are only five such references known in all of Egyptian history, the first in the 12th Dynasty, the last in the Roman period. Using somewhat complicated astronomical arguments, modern scholars are able to use these as fixed chronological points. For example, a heliacal rising noted for the seventh year of Senusret III of the 12th Dynasty can be placed circa 1872 BC; another for the ninth year of Amenhotep I of the 18th Dynasty indicates circa 1541 BC. Adding the substantial information from king-lists, dated documents, and other material, a chronology can be worked out for the 12th and 18th Dynasties, then for the Middle and New Kingdoms to which these dynasties belong, and finally the dynasties before and after those kingdoms. The result is only an approximate chronology in terms of years BC, not a precise one. As there are so many variables involved, it is doubtful that precision can ever be achieved.
A dynastic chronology such as that of Egypt depends heavily on two factors: the lengths of the reigns of individual kings and the number and length of coregencies. In themselves, individual discrepancies may seem relatively unimportant: Merenptah ruled ten years rather than the traditional nineteen; the coregency between Tuthmose III and Amenhotep II lasted from none at all to three years, according to different scholars. But when such minor differences occur frequently over three thousand years of history, their collective impact is a serious obstacle to reliable absolute dates.
Even the astronomical testimony—records of lunar months, heliacal risings of Sirius, and the like—is plagued with variables. There is, for example, the arcus visionis, the angle between Sirius and the sun when the star is first observed in its heliacal rising. Modern studies have fixed this angle at 7.5°, but variations in the arcus visionis change the chronological calculations based on it and we have no way of determining what this angle was for any ancient observation.
Another variable is the point in Egypt where an ancient observation took place. Memphis, Thebes and Elephantine have been defended as the site of a "national observatory" where official sightings of lunar phases and heliacal risings were made. The problem here is that a heliacal rising, for example, is observed one day earlier for every degree of latitude as one moves south along the Nile. Translating this into absolute dates, the heliacal rising recorded in the ninth year of Amenhotep I would have occurred around 1521 BC, 1523 BC or 1519 BC, depending on whether the observation was made at Memphis, Thebes or Elephantine.
The alternative is to see the whole matter in terms of the ancient setting. We have no reason to suppose that the Egyptians expected the new year or the new month of the lunar calendar to begin simultaneously throughout the country. To them, it was a matter of a few days at most and it did not really matter if the same religious festival was celebrated a few days earlier at Elephantine than at Memphis. What did matter was that any religious festival should occur on the proper day of the lunar year. This suggests that astronomical observations were made in each locality which kept its own lunar calendar. The lunar month or new year thus began at this or that town when the appropriate observations were made locally, allowing each district to adhere to the strict pattern of festivals and ceremonies required by religion. The civil calendar, which had none of the drawbacks of the lunar calendar, could be used throughout the country. It had a uniform meaning everywhere that never changed. A date such as "Year 2, month 3 of Inundation, day 16 (of King X)" meant exactly the same day—the 76th day of the king’s second regnal year—whether it was used to date a document at Memphis, Thebes or Elephantine.
In spite of all the problems involved, with a judicious use of all the sources noted above it is possible to present an absolute chronology, though one with a margin for error that expands the farther one moves back in time. The earliest fixed date in Egyptian history on which all agree is 664 BC, the beginning of the 26th Dynasty. Moving back from that year, through historical synchronisms with Assyrian chronology, the beginning of the 22nd Dynasty fell in the period 947 to 940 BC, so there is already a small margin for error. The beginning of the 19th Dynasty is calculated as anywhere from 1320 to 1295 BC, the margin for error now a quarter-century. This remains about the same for the beginning of the 18th Dynasty, said to be from 1570 to 1540 BC.
It is with the beginning of the 12th Dynasty that a really serious discrepancy in current chronological studies begins; the dates currently defended range from 1994 to 1938 BC. This much larger margin for error results from very different interpretations of the astronomical evidence, in particular, the location of an assumed national observatory where "official" observations of the heliacal rising of Sirius were recorded. This six-decade margin for error remains fairly constant back through the Old Kingdom, but looms larger for the earliest dynasties; dates from 3100 to 2950 BC are currently proposed for the unification of Egypt at the beginning of the 1st Dynasty.
Dating techniques, prehistory
Before the advent of chronometric dating methods, archaeologists of prehistoric sites in Egypt relied on a relative chronology based on the sequence of riverine terraces bordering the Nile. This sequence was first established at the beginning of this century by geologists K.S. Sandford and W.J.Arkell, who correlated Nile terraces with circum-Mediterranean marine terraces. Today this approach has been abandoned in favor of a relative chronology based on lithostratigraphic units belonging to successive stages in the evolution of the Egyptian landscape.
So far, radiocarbon dating has been the most widely used chronometric technique. Recently, thermoluminescence, optical, electron spin resonance, amino acid racemization and uranium series dating techniques have been applied to a series of sites predating the range of radiocarbon age determination (approximately 60,000-40,000 years ago).
Radiocarbon dating has been extensively used for sites ranging from Middle Paleolithic to Predynastic sites (as well as Dynastic sites). Carbon-14 is a carbon isotope formed from nitrogen in the atmosphere. Plants and animals receive Carbon-14 during their lifetimes. Carbon-14 begins to decay after the death of organisms. Until the 1970s, age determination was based on measurement of the radiation resulting from the decay of Carbon-14, which required relatively large samples of organic materials (usually charcoal). Today, the concentrations of Carbon-14 in very small samples can be measured directly using accelerator mass spectrometry (AMS).
Thermoluminescence dating (TL) can be used on a number of materials, but has a much wider range of error than radiocarbon dating. The method is based on a measurement of the emission of light upon heating the sample. The amount of TL is proportional to the amount of radiation the material was exposed to after a certain event, such as the firing of clay. Optical (stimulated) dating (OSL) is based on the luminescence resulting from the eviction of electrons from traps in the material by the action of light.
Electron spin resonance dating also depends on the nuclear radiation that has been experienced by a sample. However, in this method the electrons are not excited. Their signal is detected by their response to high-frequency electromagnetic radiation. The method is suitable for tooth enamel, mollusk shells, calcite and quartz.
Amino-acid racemization dating depends on changes in the molecular structure of amino acids, the building blocks of protein. The changes hypothetically occur at a constant rate and produce mirror images of amino acids. The ratio of isomers of aspartic acid (one of the amino acids) can be used for dating samples a few thousand to several million years old. The technique is of limited use and is subject to errors due to the susceptibility of racemization to temperature.
Uranium series dating is based essentially on the decay of thorium-230 into a series of uranium radio-isotopes. Calcite samples can provide age estimates in the range of 5,000350,000 years.
Thermoluminescence, optical, electron spin resonance, amino-acid racemization and uranium series dating techniques have been used to date Lower and Middle Paleolithic sites and climatic events. Middle Paleolithic sites range from 175,000-70,000 years ago. Using a combination of radiocarbon dating and TL dating, late Quaternary sites in the Nile Valley and the Western Desert have been dated and assigned to arid climatic conditions from 65,000-12,500 BP (before present in radiocarbon years). Upper Paleolithic sites in Egypt date from 33,000-30,000 BP, Late Paleolithic sites date from 20,000-12,000 BP, and Epipaleolithic sites date from circa 11,000-7,000 BP.
Neolithic and Predynastic sites (excluding Nagada III sites) in the Nile Valley date from 5,900-4,600 BP. Since radiocarbon years do not correspond to calendric years, radiocarbon dates may be calibrated using measurements on samples dated both by tree-ring and radiocarbon dating techniques to obtain calendric years (BC).
