The river Nile flows about 6,800 kilometers (4,200 miles) from equatorial Africa to the Mediterranean Sea, across thirty-five degrees of latitude. In the far south, several large East African lakes temporarily collect and store runoff and direct precipitation from the early summer and autumn rainy seasons, discharge that is then delayed several months as it slowly moves through the Sudd swamps. Relatively impoverished in sediment and nutrients, these White Nile waters pass the central Sudan after the peak discharge of the summer monsoon over Ethiopia. It is the runoff from Ethiopia, mainly the Blue Nile system, that determines the duration and level of the annual flood in Egypt, and that contributes its increment of silt and nutrients. The blending of these two distinct sources of water during the span of the hydrological year provides a comparatively reliable river that has served as a lifeline for Egypt, allowing the emergence of stable adaptations in late prehistoric times.

The strong seasonal flux of water volume is fundamental to the spilling of water out of the channel and onto the flood and delta plains late each summer. The annual fusing of staggered rainfall maxima and the temporary storage of water volume along this longitudinal waterway, in turn, prolong peak discharge. The Nile, therefore, guarantees a substantial energy pulse, but one that is beneficial and not destructive. Temporary storage in Ethiopia, and especially East Africa, also moderates year-to-year fluctuations of rainfall, while a primary catchment greater than 1.5 million square kilometers (930 thousand square miles) averages out at least some of the interannual deviations of rainfall among the many contributing watersheds.

Nonetheless, there are and always have been substantial positive or negative trends in the composite volume of water passing through Egypt. For example, the mean for 1870–1898 was 16 percent higher than for 1899–1971, a trend spanning a century. For shorter time spans, the discharge for 1954–1967 was 11 percent above, but for 1972–1986 almost 13 percent below the mean for 1899–1971; these represent irregular cycles with a wavelength of nine to seventeen years, each characterized by different amplitudes and year-to-year variability. Efforts to explain such fluctuations using modern records since the 1870s, or flawed notations since 622 CE, have included solar output (sunspot) cycles, lunar tidal forcing, and ENSO (El Niño), but all have been inconclusive. Each of these quasi-periodic factors may well have played a part in a multicausal spectrum effecting short-term variation, but each such anomaly during the twentieth century seems to play out differently among the many tropical watersheds, and correlation of rainfall trends over Ethiopia with the equatorial lake regions is in fact random. For longer-term trends, of a century or more, there can be little more than speculation, and for changes spanning centuries to millennia, systemic interactions among atmosphere, oceans, and surfaces, and biota must assume a much greater role.


Nile. Diagram of the Nile floodplain, c.3500 bce, and a plan of an early floodplain settlement. (Courtesy Karl W. Butzer)

Whatever their explanations, and the uncertain progression of change in different segments of the tropical watershed, the variations of the Nile flood volume over decades, centuries, or millennia have had economic, biotic, and even physical repercussions or impacts in Egypt. Such changes in equilibrium or vulnerability are outlined below, beginning with the evolution of the flood and delta plain environments, and concluding with questions of economic and possibly social implications in the course of Egyptian history.

Geological and Archaeological Records prior to the Predynastic.

Most major rivers have long geological histories, and the Nile is no exception. Even its existing valley within Egypt had been cut down to well below modern sea level by five million years ago, and its delta also had begun to subside tectonically. In part this was a response to the drying out of the Mediterranean Sea about seven to five million years ago (late Miocene). Subsequently, this deep canyon was submerged when the sea rose again (Pliocene), flooding its northern end and filling in the overdeepened valley with lagoonal, lake, and stream deposits. During the two million years of Pleistocene time, the axis of the river was already well defined, as it underwent repeated cycles of river cutting and filling that nonetheless left much of the Pliocene sequence of sediments in place, in part near the modern valley margins, but mainly buried underneath thick bodies of sands and gravels. Minerals of both clay and sand size leave no doubt that Ethiopian floodwaters were flowing to the sea throughout this time, but the geological record is dominated by coarse sweeps of material derived from Saharan tributaries that responded to episodic heavy winter rains in an environment periodically less arid than today.

In most parts of Africa, this time range would have seen Acheulian occupation on at least a sporadic basis. Yet in or on the Nile “terrace” deposits of one million to 150,000 years ago, Acheulian materials are only locally common, and such occurrences date mainly to the later part of the time interval. From late during this time range, there also are geological deposits in which Nilotic flood silts, of Ethiopian flood origin, are interbedded with thick sandy gravels carried into the Nile Valley by then-active desert watercourses in the form of alluvial fans. This documents a transition of the Nile to a river resembling that of today. The braiding and unstable channels of the mid—Pleistocene probably were less productive or dependable than the Saharan spring-fed oases, some of which were attractive to early prehistoric settlement; in fact, Mousterian sites of perhaps 150,000 to 40,000 years ago are better documented in Saharan oases than in the Nile Valley.

About 25,000 years ago, the transition to a silt floodplain was completed, when summer flood silts accumulated in and around the valley—no longer to be eroded or overwhelmed during the cooler half year by torrential, tributary runoff. Although minor moist intervals in the Egyptian deserts and Sahara promoted some wadi activity or higher water tables in the oases, the climate of Egypt has been too dry during the last 50,000 years or more to dominate the geomorphic processes of the Nile Valley. The Nile had become an exotic river, supplied by a distant watershed, and responding mainly to environmental changes in Ethiopia and East Africa.

By 20,000 years ago, Late Paleolithic peoples were continually settled along the shifting banks of the Nile, even as the river dwindled c. 17,000–15,000 BCE, or as it swelled during a period of exceptionally high floods, the “Wild Nile” of 11,500–11,000 BCE. The basic continuity of site location, economic utilization, and stone tool assemblages across at least eight millennia implies a fairly stable adaptation to this “tropical” river regime: riverine food resources were exploited during an annual cycle of flood, post-flood abundance, and pre-flood shortage.

The picture was different in the Nile Delta, reflecting the changing level of the Mediterranean Sea. While ocean waters were bound in the great continental glaciers during the ice ages, the world sea level was much lower than now—as much as 100 meters (325 feet). Thus, in about 33,000–14,000 BCE, the coastline was located far out on the modern continental shelf, so that the several branches of the Nile responded by cutting deeper channels into the older delta surface, eventually eroding a fairly level topography—which is now buried under 30 to 50 meters (100 to 165 feet) of younger deposits. When the sea level once again rose rapidly after 14,000 BCE, seasonal flood deposits began to accumulate rapidly across the exposed delta surface. About 9000–5000 BCE, the sea returned to near its present level, flooding the edge of a long-exposed plain, consisting of multiple, sandy channels, with zones of (summer) flood silts in between. The distinctive marine sands and muds of this transgression in the northernmost Nile Delta form a critical marker horizon. This shows that the Nile branches had already been actively building up sediment for many millennia, and it explains the development of poorly drained tracts near the shore that supported lagoons or marshland. Such organic beds are dated as early as 6500 BCE, and were most extensive about 2,500 years later, at about 3750 BCE.

Although the basic configuration of the modern Delta landscape was falling into place during the time range of the Levantine Neolithic, whatever late Paleolithic or Neolithic record there was within the immediate Delta is now buried under many meters of sediment. Visible or accessible are only those sites, such as Merimde and Maadi, that were at the edge of the desert, or those occupation traces of the fourth millennium BCE recovered underneath younger settlements.

A part of the missing occupation sequence in Lower Egypt is recorded in the Faiyum Depression, connected to the Nile Valley across a buried threshold at 10 meters (33 feet) below sea level. This medium-sized basin was fed by overflow from the Nile whenever the floodplain was higher than this bedrock sill, bringing in silt and supporting a non-outlet lake. But it was dry and subject to wind erosion until about 8500 BCE, evidence that the lowermost Nile was also undercut in response to the sea level regression that affected the Delta. At that point the Faiyum was abruptly flooded, and within a millennium it was settled by Epipaleolithic people who exploited the aquatic resources of the lakeshores. About 5500 BCE, other Neolithic people, practicing a little agriculture, replaced the fishers and gatherers already around the lake.

This late appearance of farming is similar to the archaeological record of the Nile Valley farther south. There, a major period of river adjustment led to some five hundred years of floodplain entrenchment (about 9500 BCE), during which resources appear to have been unusually scarce. When the floodplain began to build up and again to expand laterally, there were far fewer sites in the valley, implying a strong reduction of population. The stone tool industries had also changed, and there was a new emphasis on fishing, in addition to hunting and gathering. That suggests a turnover in social identity, accompanying a shift in economic and ecological adaptation. Evidence for farming and herding in the main valley of Egypt begins only during the fifth millennium BCE, assuming importance after 4000 BCE. With the archaeological bias of a record based on sites limited to the floodplain-desert margin, and heavily represented by cemeteries rather than habitation sites, it is still uncertain whether, in addition to residual fisher and gathering groups, there may also have been distinct groups favoring farming activities or pastoralism.

In both the Faiyum and Nile Valley, therefore, it is plausible that there was no “Neolithic Revolution.” Instead, it is possible that several groups of people with different ecological adaptations utilized the larger system represented by the Egyptian Nile, emphasizing complementary or only partially overlapping econiches. Alternatively, their economies may have been uncharacteristically flexible and fluid—from season to season, year to year, or place to place—in response to fluctuating and variable resource types. That is what may well emerge from contemporaneous archaeological residues in the Saharan oases or in the central Nile reaches of Sudan. In any event, there is growing reason to suspect that the long prehistoric millennia preceding full-fledged farming, pastoralism, or metallurgy in Africa were a great deal more fluid and complex than the conventional archaeological sequence of the Near East or southern Europe.

An Environmental Model for the Late Prehistoric Nile.

A river floodplain is primarily a zone of sediment accumulation rather than erosion. The subtle but significant surface features are determined by the prevailing processes of alluviation. These reflect the turbulence of flow in the channel during times of peak discharge, and the concentration of suspended sediment in the floodwaters.

A turbulent stream, with a dominant “bed load” of sand and gravel, tends to have a broad, shallow channel; the rapid accumulation of channel bars forces the weaving stream axis to shift repeatedly, so that the floodplain is built up of lenticles of sand and gravel. As the river tops its low banks during the flood peak, a gentler flow deposits a thin mantle of overbank silt and clay across the flat floodplain. This cover of overbank sediments tends to be ephemeral, much of it swept away the next time flood waters spill over; when they abate, they deposit fresh or reworked overbank silts. This pattern is typical of arid zone watercourses and can be observed in some wadis of the Eastern Desert.

On the other extreme, a major river with abundant suspended sediment has little channel turbulence and moves only finer grades of sand. The channel is deep and comparatively narrow, flowing through fairly stable banks of silt and sand, so that the excess water and energy is dispersed across the floodplain, rather than being concentrated on temporarily deepening and then refilling a channel. Sediment builds up fastest on the inside bends (point bars) of the river, and directly where water spills out of the channel during flood stage. This is where energy is greatest. As a result the perimeters of the channel—the levees—tend to rise up to a meter or two above the alluvial flats, where flood energy is rapidly reduced and particle sizes in transport are much smaller. Here, the annual increments of silt and clay build up somewhat more slowly, but provide the most fertile soils. Because the perimeter of the channel is slightly higher than the seasonally flooded plain, this landform complex is known as a convex floodplain. The Nile and Mississippi are two of the best known examples of the kind, but the floodplains of the middle Niger in Africa, or the Ganges and Mekong in Asia, are very similar.

On convex floodplains, settlements are preferably located on the levees that form higher ground near the channel. The levee crests do not flood regularly, but only during the highest flood events and, since mud-brick structures rise through countless collapses and rebuilding, villages and towns grow above the levees on which they are situated. A second feature of convex floodplains occurs when a period of waxing floods may carry enough water through levee breaches to create diverging, secondary channels, such as the Bahr Jussef on the western side of the floodplain in Middle Egypt. Although smaller and lower, the levees of such branches will eventually offer advantageous settlement sites, with permanent water and navigability. Finally, whereas the river channels of convex floodplains do not migrate rapidly, as those of flat floodplains do, increasingly convoluted meanders eventually become unstable. At that point, an exceptionally strong flood may cut off a meander loop or even bypass a longer channel stretch. This “jumping” of the main channel, known as an avulsion, will leave a cut-off lake or even a stretch of river in a state of atrophy, but the levees remain, so that there now will be several ridges of higher ground within the floodplain.

In these various ways, a convex floodplain creates multiple, linear environments of elevated ground, suitable for settlement location, and that outline distinct flood basins, which will eventually lend themselves to compartmentalized irrigation. That is the environmental model for the prehistoric Nile and for dynastic Egypt. It presupposes that convex floodplains can be settled permanently with minimal technological application, by either fisher-hunter-foragers or farmer-herders. The floodplain will be underwater for four to six weeks annually, even as the majority of levee-top sites remain dry in most years. Livestock, like the native fauna, move to higher ground or the desert edge while the water is up, and then return to the emerging pastures when the waters drain back into the main channels, as the flood recedes. Since the floodwaters rise and move slowly, there are few losses among such animals, and even so, water depths are mainly in the range of only 1.0 to 1.5 meters (3 to 5 feet). For agriculture, the natural hydraulic system is even more advantageous, since seed can be broadcast on the wet increment of overbank silt as the waters recede; furthermore, the clayey soil retains abundant moisture that, in damper areas, is adequate to the maturation of a crop.

The features describe a model of a Nile floodplain where the raised river banks and old levee ridges have always been favored as settlement sites, even as the flood-basins fill and empty when the annual flood rises and recedes. It is a free draining, not a marshy floodplain, and one that invites riverine settlement. The model is supported by settlement patterns and ethnographic data from other convex floodplains on several continents. Most compellingly, however, it can be tested archaeologically.

Late Prehistoric Settlement Mosaics.

Egyptological archaeologists and historians tend to be occupied with reconstructing the evolution of a stratified society in the Nile Valley, the unification of Upper and Lower Egypt, and the emergence of the Egyptian state—primarily in the context of other complex societies in the Near East. Africanist archaeologists, on the other hand, prefer to examine a broader, African canvas, to identify prehistoric processes in the deserts, as well as in the Nile Valley. They seek to understand how that long transition—from hunting, gathering, and fishing, to food production and, ultimately, a complex society—played out, when, and where. Such research clusters are not monolithic, and the dichotomy is overdrawn, but it draws attention to different questions, methodologies, and conceptual modes. Largely as a result of overemphasis on Predynastic social evolution and state formation, some significant implications of the Paleolithic and Neolithic records have been overlooked.

The great majority of Late Paleolithic sites in Southern Egypt, dating c.19,000–11,000 BCE, were situated on point bars and levees of the Late Pleistocene Nile, for example, on the Kom Ombo Plain. Some of these sites extend for hundreds of meters along the former riverbank, and seasonality studies suggest that they were occupied as soon as the floodwaters began to recede until well into the post-flood season, or beyond. Others were located around ponds fed by flood seepage into depressions between levees and dune ridges, or within a dune field, such as between Esna and Edfu, or in the embouchure of Wadi Kubanniya. The faunas represent a riparian ecotone: wild cattle (favoring brush or woodland), hartebeest (requiring a grassy groundcover), gazelles (adapted to semidesert settings), an occasional hippo (aquatic), a variety and abundance of fish remains (some of which spawn on flooded surfaces), occasional river clams, and a good range of waterfowl. The hartebeest and fish remains are the most common and ubiquitous. Fishing was apparently done with fish-gorges, baskets, and nets. Study of plant remains has begun, and indicate the presence of aquatic plants (Cyperus and Scirpus), ferns, dom palm, and tamarisk at a limited number of sites. But the residues indicate food processing of aquatic tubers, perhaps during the pre-flood season.

Such Late Paleolithic sites extend in large numbers from Qena to Wadi Halfa, a 650-kilometer (403-mile) stretch where fluvial deposits of the period are exposed above the modern floodplain along the desert edge. Hunter-gatherers are mobile, of course, so that such sites would not be permanent. But key base-camps were reutilized indefinitely, and the tentative bioarcheological inference is that they were occupied for one or more seasons, perhaps even most of the year. That is supported by the apparent absence of sites on the former desert margins: all are water-edge locations, regardless of differences in stone tool assemblages.

Similar riparian adaptations are evident after the floodplain had readjusted to deep entrenchment of the river, but there are far fewer sites pertaining to Epipaleolithic industries such as the Shamarkian and Elkabian, which span the interval from 10,500 to about 6000 BCE. The faunal palimpsests are sparse but indicate a consistent dependence on hartebeest or wild cattle, as well as fish. Such sites are also found on former riverbank locations. A special case is the Faiyum, where Qarunian sites (c.7300–6200 BCE) are situated along the former fluctuating lake shores. By weight, the Qarunian fauna consists overwhelmingly of fish bone, with subsidiary wild cattle and hartebeest; bone points and harpoons underscore the aquatic adaptation.

Insufficient attention has been paid to the Epipaleolithic sites, but the key problem is visibility. In the Nile Valley, contemporary sediment exposures are limited to the reach south of Esna, and the deposits are commonly thin or somewhat eroded. Except for a desert pediment west of Luxor, where similar stone artifacts (Tarifian) are found at the surface, no sites younger than 6000 BCE have been identified. After a sharp recession of the Faiyum lake, that area was recolonized by settlers with a Neolithic tool inventory, and who practiced some cultivation of grains and herding of animals, primarily sheep or goat. But most of the food remains are fish and soft-shelled turtle, and the seasonality of fishing activities was unchanged from the Qarunian. Nonetheless, there is only minimal evidence of big-game hunting. Known as the Faiyum A (or Faiyumian), this occupation of c.5500–4400 BCE was not associated with permanent settlements. A few large sites were repeatedly reoccupied as base-camps, above the fluctuating shoreline, but consist primarily of large, superimposed hearths and storage pits for grain. More common are intermediate-sized occupation areas, with large concentrations of stone artifacts and potsherds, repeatedly used on a perhaps seasonal basis on tracts that were periodically flooded. Other sites are small and were only occasionally used in the course of seasonal rounds. The implied mobility patterns are those of hunter-gatherers, rather than farmers.

The Faiyum A was succeeded by what is now called the Moerian (c.4400–3800 BCE). This has been labeled Neolithic by virtue of bifacial and blade tools, as well as pottery, in an otherwise Epipaleolithic artifact assemblage. Fish bone is abundant, but domesticates are not evident. A Predynastic occurrence, based on its pottery, is found in the same setting, but it, too, is a fishing site.

The Faiyum settlements are highly informative for several reasons: (1) They show that similar economic adaptations could be pursued by social groups with different material cultures. Designations such as Neolithic, based on a handful of diagnostic artifacts, do not necessarily identify land use or settlement type. (2) The Moerian indicates persistence of a food-collecting social group, possibly in the process of assimilation, within the lower Nile ecosystem until the early fifth millennium BCE. That raises the reasonable hypothesis that two or more internested socio-cultural groups occupied the Nile Valley during the previous millennium; these probably exploited complementary ecological niches, but with diverging spatial expression. (3) Economic modes in the Nile ecosystem, from the Faiyum to the Central Sudan, were in a state of flux during the fifth, and perhaps even the fourth millennium BCE. The Faiyum A suggests that opportunistic exploitation of highly productive natural resources was as important as available agro-pastoral experience, which would then affect site location and mobility patterns.

Merimde, situated on the southwestern edge of the Delta, was a very different matter. Five periods of occupation are identified, spanning perhaps five hundred years (c.4750–4250 BCE), the last two covering an area of 25 hectares (62 acres). There is good evidence of house structures, and the economy was predominately agro-pastoral.

But it was located directly above the bank of a minor Nile channel, and fish, as well as other aquatic resources, were intensively used. In the southern suburbs of Cairo, Omari (c.4600–4400 BCE) was situated on the banks of a wadi, next to the edge of the floodplain. Initially a fishing encampment, Omari developed into a small farming village, in which living areas continued to shift. Although pig, cattle, and some sheep/goat were kept, most of the excavated bone belongs to fish. The small settlement sites near Badari (c.4500–4000 BCE) were not much different—on the desert edge but close to a Nile branch, so that fish were a prominent economic element. These late fifth millennium BCE sites were effectively riverine sites, even if on the desert margin, and they combined older gathering pursuits with agro-pastoral activities. In the case of Merimde and the Badarian sites, there also were floodplain and desert game animals, suggesting that the wildlife was not yet hunted to extinction, or that the adjacent deserts still supported some game, or both.

The fourth millennium BCE saw an ecological shift. Near Cairo, the desert edge site of Maadi (c.3900–3500 BCE) has minimal evidence for fish and game, as do the early Predynastic sites of Armant and Hu. The desert edge settlement of Hierakonpolis (c.3700–3400 BCE) however, has considerable fish bone and a stronger herding component. Its counterparts at Nagada (c.3600–3300 BCE) and the slightly younger Delta site of Tell Ibrahim Awad also have abundant fish bone. Yet none of these has evidence for more than incidental game. They were preeminently agro-pastoral sites, some of which had convenient access to water bodies, others not.

Paleobotanical studies at Nagada and Ibrahim Awad, as well as at the Old Kingdom delta site of Kom el-Hisn, focused on animal dung, mainly of small livestock (sheep/goat). They reveal that mostly wetland plants were fed to the animals, presumably kept in enclosures. That shows that there was no desert pasturage, and that pastoralism did not extend into the desert in any significant way. The inference is that domesticated stock were grazed in the floodplain during the post- and pre-flood seasons, but that they were removed to higher ground during the flood months, to be fed with cut vegetation from wetlands, presumably gathered previously. In other words, fourth millennium small-stock pastoralism was not only fairly intensive, but it was very much tied to the floodplain. Whatever the importance of cattle, they were managed differently, in keeping with their different ecology. In East Africa cattle can be observed to their dewlaps in water, consuming emergent aquatic plants, and much the same can be seen on Old Kingdom tomb reliefs. Presumably Egyptian cattle in the valley and the Delta remained in the shallower fringes of the floodbasins during the annual inundation.


Nile. The Nile at Philae, before the inundation. (Courtesy David P.Silverman)

This confirms that Predynastic (and probably earlier) pastoralism in the lower Nile ecosystem was essentially limited to riverine pasturage. The desert edge was used like the levee environments, to keep small stock during the flood season, with the help of cut feed. There were no desert pastoralists, even if pastoralism had once been introduced from the desert oases. Whatever desert game may have been available during the fifth millennium BCE had disappeared by Predynastic times.

That supports the geoarchaeological model of an “accessible” floodplain mosaic described above. The agropastoral economy of the Egyptian Nile ecosystem, even during its initial stages of complementary or supplementary fishing and hunting, was focused on one and the same riverine environment. When the desert edge settlements disappear during late Predynastic times, that does not indicate a settlement shift “into” the floodplain, but rather a further intensification of subsistence activities with respect to the linear distribution of higher ground, provided by active or abandoned levees within the floodplain and Delta.

Nile Flood History.

The changing behavior of the Nile River becomes tangible from the geological record in southern Egypt after about 19,000 BCE, when high-level flood silts interfinger with mobile dunes on the Western Desert edge. About 16,000 BCE, flood volume declined and the river incised its channel, before achieving a more vigorous flow about a millennium later. Increased turbulance allowed the transport of pebbles in the stream bed over great distances, with some channel erosion into the hard rocks of the cataracts. In wide floodplain sectors, such as at Kom Ombo, the Nile formed multiple, shifting channels, and flood waters seeped into old dune fields to the west of the valley. About 11,500 BCE, the river entered a “Wild Nile” stage, with repeated “catastrophic” floods to 5 or even 10 meters (16 to 33 feet) above the floodplain, which signalled a major change of climate in East Africa. By 11,000 BCE, the floods dwindled rapidly, causing the channel to downcut its bed by as much as 25 meters (82 feet). At that time, only a very narrow floodplain remained inundated, and the number and size of population groups was greatly diminished.

A flood regime, broadly similar to that of today but a little more vigorous, resumed about 10,500–6000 BCE, interrupted by two or more intervals of weaker floods. This is the time of dispersed fisher-hunter-gatherer settlements (Epipaleolithic), and evidence for Nile behavior increasingly comes from the Faiyum and the Delta. About 6000–5800 BCE there was another hydrological readjustment, possibly marked by repeated failures of the Blue Nile floods, that left the shores of the Faiyum lake dry. Epipaleolithic sites seem to disappear, but the Faiyum A Neolithic appears shortly thereafter, to be followed by the Neolithic of Merimde and Omari, the Badarian, and the enigmatic Moerian of the Faiyum. The next Nile “crisis” is dated about 4000 BCE, and again lasted perhaps two centuries. The Faiyum lake level fell abruptly and in the Delta, peat formation was possibly interrupted and channels deepened, perhaps in response to erratic low and very high floods, as the White Nile failed repeatedly for much of the fourth millennium BCE. But the Blue Nile and Atbara floods appear to have been strong during most of the Predynastic period, which would imply greater amplitude between high and low water, a stressful pattern for agricultural land use.

A more equitable picture is suggested for the time of the first dynasty, but during the second dynasty sporadic historical observations point to a 1 meter (3 foot) decline in flood height and a 25 to 30 percent decrease in Ethiopian discharge. Toward 2800 BCE, there was recurrent flood failure, but the geochemical record suggests that it was the White Nile discharge that was reduced throughout the Old Kingdom period. If there was a failure of the Nile floods during the sixth dynasty, it was very brief, and by 2150 BCE, the Blue and White Niles were both strong. By 2050 BCE, the lower valley was experiencing a repetition of the “Wild Nile,” with four periods of very high flooding in the Faiyum. In the Delta, the surge of flood waters required the river to cut deeper channels to accommodate the phenomenal discharges. Indeed, such flood episodes are verified by more than twenty high-water marks, dating about 1840–1770 BCE in the Semna cataracts. They would indicate that two years out of every five would create flood levels 4 to 7 meters (13 to 23 feet) higher than those of the late nineteenth century at Aswan, and 3 to 5 meters (10 to 16 feet) higher across the broader floodplain near Cairo.

These flood perturbations of the Middle Kingdom now are well documented by convergent lines of evidence, and spanned up to 350 years. Their implications for Egyptian agriculture would be negative, since such catastrophic floods would wash out dikes, endanger settlements, and rapidly slit up irrigation canals. In addition to repeatedly destroying Egypt's infrastructure, very high floods imply a much longer flood season, so that crops could only be planted many weeks later, maturing early in the season of hot khamsin winds, with increasing drought stress. Long-term waterlogging of soils also increases soil parasites and endangers crops through rot, vernim, and blight. The Middle Kingdom was, in other words, a period of great environmental stress.

In Ramessid times, the Blue Nile floods began to fail, and much of the floodplain in Nubia was no longer inundated. Under Ramesses III (1198–1166 BCE), the Egyptian records first inform us about food shortfalls on a serious scale. Wildly fluctuating food prices (relative to other prices) argue that shortfalls or famines were common in about 1170–1100 BCE. They probably contributed to the destabilization of the New Kingdom.

The floods appear to have regained a more “normal” level during the tenth century BCE, and were comparatively high at the time of Herodotus's visit (c.450 BCE), judging by water levels in the Faiyum. The divergence of the Nile over five or more tributaries in the Delta, according to Strabo (c.25 BCE) and Ptolemy (c.160 BCE), also suggests a “strong” Nile, but in Nubia flood levels only returned to those of 1250 BCE during the period of about 600–1000 CE. By the time of Idrisi (1154 CE) however, the Delta tributaries were reduced to two, with the failure of the western and easternmost Canopic and Pelusiac branches. This suggests a lower discharge norm, comparable with that of the twentieth century.

Egyptian Names and Symbols.

Egyptologists have adopted an Egyptocentric nomenclature for the cataracts that occur along the Nile's course, so that numbering begins with the First Cataract at Aswan and proceeds south to the Sixth Cataract not far north of Khartoum. The First Cataract had to be periodically cleared of rocks that impeded its navigability; the Second Cataract (the Batn el-Hagar, “Belly of Stone”) was navigable only by very small boats, making it necessary to portage around it. It was there that the twelfth dynasty kings established a line of fortresses to keep Nubians from penetrating farther north.

During those times of the year when the Nile was not in flood, the Egyptians called the river “Iteru” (itrw), the term's origin is unclear, but it has been suggested that it means “the Seasonal One.” The Nile in inundation was designated as Hap or Hapy (Hʿpy), and was thus deified as the god of the Nile and father of all beings (not to be confused with Hapy, one of the Four sons of Horus). The form of this hypostasis of the Nile's fertilizing bounty of water and silt was represented most commonly by an androgynous figure: bearded, with pendulous breasts and prominent belly, but with no visible genitalia. In some representations, the body is covered with wavy blue lines evoking the river's waters. Frequently, the Hapy figures are seen in pairs flanking the emblem for the unification of Egypt (smʒ tʒwy), signifying the Nile as a basic unifying feature of Egypt. In other representations, processions of Hapy figures bear on their heads the names of the nomes (the districts of Egypt) or depictions of the characteristic riverine plants, the papyrus and lotus. They are often shown bearing offering jars for libations, plants, trays of offerings, or simply the hieroglyph for offering. The south-to-north flow of the Nile formed the basic idea of a river for the Egyptians, so that the rivers of the Near East, notably the Euphrates, were puzzling; they referred to the latter as “the river that goes backwards.”

See also HYMNS, article on Nile Hymns; and IRRIGATION.


  • Andres, W., and J. Wunderlich. “Late Pleistocene and Holocene Evolution in the Eastern Nile Delta and Comparisons with the Western Delta.” In From the North Sea to the Indian Ocean, edited by H. Brückner and U. Radtke, pp. 120–130. Stuttgart, 1991. Synthesizes important observations on buried sites and their environmental context.
  • Bell, Barbara. “The Oldest Records of the Nile Floods.” Geographical Journal 136 (1970), 569–573. Plots the surviving flood records from the first to fifth dynasties.
  • Bell, Barbara. “Climate and the History of Egypt: The Middle Kingdom.” American Journal of Archaeology 79 (1975), 223–269. Analyzes the exceptionally high flood levels recorded at Semna.
  • Brewer, Douglas J. Fishermen, Hunters and Herders: Zooarchaeology in the Fayum, Egypt (ca. 8, 200–5,000 B.P.) Oxford, 1989. A key resource on the fishing economy of the prehistoric Faiyum.
  • Butzer, Karl W. Recent History of an Ethiopian Delta: The Omo Delta and the Level of Lake Rudolf. Chicago, 1971. Documents a modern delta and floodplain, with its settlement patterns, including aerial photographs of characteristic features.
  • Butzer, Karl W. Early Hydraulic Civilization in Egypt: A Study in Cultural Ecology. Chicago, 1976. Examines the implications of the floodplain environment and its changes for the evolution of irrigation. In part, now dated.
  • Butzer, Karl W. “Long-term Nile Flood Variation and Political Discontinuities in Pharaonic Egypt.” In From Hunters to Farmers: Causes and Consequences of Food Production in Africa, edited by J. D. Clark and S. A. Brandt, pp. 102–112. Berkeley, 1984.
  • Butzer, Karl W. “Sociopolitical Discontinuity in the Near East c 2200 B.C.E.: Scenarios from Palestine and Egypt.” In Third Millennium b.c. Climate Change and Old World Collapse, edited by H. N. Dalfes et al., pp. 245–296. Berlin, 1997. Reexamines and rejects the First Intermediate Period “low Nile” as a factor in Old Kingdom collapse.
  • Butzer, Karl W. “Late Quaternary Problems of the Egyptian Nile: Stratigraphy, Environments, Prehistory.” Paleorient 23/2 (1998), 151–173. Revises the history of the Nile floods and floodplain from 20,000 BP to the end of the Old Kingdom, based in good part on recent evidence from the Faiyum and Delta, with critical overview of the literature.
  • Gautier, Achilles, and Willem van Neer. “Animal Remains from the Late Paleolithic Sequence at Wadi Kubbaniya.” In The Prehistory of Wadi Kubbaniya, edited by F. Wendorf and R. Schild, vol. 2, pp. 119–151. Dallas, 1989. The most comprehensive examination of animal remains available.
  • Krzyzaniak, Lech, M. Kobusiewicz, and J. Alexander, eds. Environmental Change and Human Culture in the Nile Basin and Northern Africa until the Second Millennium B.C. Poznan, 1993.
  • Krzyzaniak, Lech, and M. Kobusiewicz, eds. Late Prehistory of the Nile Basin and the Sahara. Poznan, 1989. Like its 1993 counterpart, and van den Brink's edited volume, a rich repository of current research relevant to this article.
  • Shahin, Mohammed. Hydrology of the Nile Basin. Amsterdam, 1985. A fine, reasonably up-to-date guide.
  • van den Brink, Edward, ed. The Nile Delta in Transition. Amsterdam, 1993.
  • Wetterstrom, Wilma. “Foraging and Farming in Egypt: The Transition from Hunting and Gathering to Horticulture in the Nile Valley.” In The Archaeology of Africa: Food, Metals, and Towns, edited by T. Shaw et al., pp. 165–226. London, 1993. A fresh and revolutionary monographic treatment of the paleobotanical record, in an excellent volume presenting the Africanist perspective.

Karl W. Butzer