PROBLEMS OF THE LIBYAN DESERT

John Ball, O.B.E., D.Sc., M.Inst.C.E., F.G.S., Director of Desert Surveys, Egypt

Map following p. 96.

THERE can be few tracts of the Earth’s surface which present such a number and variety of interesting geographical problems as the Libyan Desert. The following are some of the more puzzling of the questions which have been asked at various times since I made my first acquaintance with the desert thirty years ago:

1. Did the Nile, or a branch of it, ever flow through the Libyan Desert to the west of the present Nile Valley?

2. By what natural agency were the great depressions in the Libyan Desert excavated?

3. Can other depressions in this region, besides the much-discussed Wadi Rayan, be considered possible of utilization in connection with irrigation, flood-protection, or drainage of the Nile Delta, or as sources of water-power for Egypt?

4. Whence comes the artesian water of the oases?

5. Why have not certain lakes, such as Sittra and Bahrein, situated as they are in an almost rainless region, long since dried up?

6. Can the present scanty supplies of drinkable water derived from local rainfall along the Egyptian Mediterranean Littoral be supplemented by artesian borings?

7. Are the artesian water-supplies of the oases gradually diminishing?

8. Where shall we look for the mysterious “Zerzura,” or “Oasis of the Blacks”? Are there any other “lost” oases remaining to be discovered?

9. Can the present difficulties of travel in the Libyan Desert, which are chiefly due to scarcity of water, be alleviated by the sinking of new wells?

10. Where are the “Tortoise Marshes” of Ptolemy?

11. What determined the peculiar distribution of the sand-dunes, some of which extend in straight lines for hundreds of kilometres?

12. What is the full extent of the distribution of flint implements and pottery, of which accumulations have been found in what are now among the most desolate parts of the desert, and what light does this throw on human history and climatic changes?

To most of these questions no completely satisfactory answers have as yet been given, though several of them have been the subject of painstaking investigations by many workers. Our knowledge of the facts bearing on them has, however, increased considerably during the past few years; and in the present paper I propose briefly to review the various questions in the light of the latest available data concerning them.

The problems, though they concern a variety of subjects, are very much interwoven with each other, and they are mostly alike in that the first difficulty in any attempt to solve them has always been the incompleteness of our geographical knowledge concerning the Libyan Desert as a whole. If we possessed a series of detailed contour-maps covering the whole of the desert, most of the problems would be in a fair way of solution. But the surveys which would be necessary for such a result are impracticable at present, mainly on the score of expense.

Some fifteen years ago I essayed the construction of a contoured map of the Libyan Desert, in connection with the International “Million” map of the world. But I had to give up the task, finding that there were vast areas without a single observation for altitude, while such altitude-data as did exist were open to errors of a magnitude intolerable for even a preliminary contouring of the areas with which they were concerned. The attempt however was not without its value, because it drew attention to the great lack of reliable levels in the Libyan Desert, and to the necessity of remedying this defect before even an approximately true picture of the general relief could be obtained.

In the interval which has elapsed since my former attempt, a great deal of surveying has been carried out in the Libyan Desert, and in most of this work special attention has been paid to the matter of levels. Two long chains of triangulation have been run westward from the Nile to Siwa Oasis, one viâ the Wadi Natrun and the other viâ Baharia Oasis; another chain has been completed along the coast westward from Alexandria to Sollum, and a triangulation is now being carried out which will eventually connect the oases of Kharga and Dakhla with the Nile Valley. Small local triangulations have also been made in certain areas, such as Dakhla Oasis and in the neighbourhood of Owenat. In all these triangulations vertical as well as horizontal angles have been observed, so that reliable trigonometric altitudes are now available for almost all the points occupied or sighted. In the southern and central parts of the desert, we are still dependent on barometric levels; but the old determinations have mostly been replaced by later and better ones; and the use of motor-car transport, by facilitating exploratory journeys in hitherto untrodden regions, has permitted of a large number of additional determinations being made in tracts where no previous observations of any kind existed.[1]

Owing to the methods employed in the latest barometric determinations, the resulting altitudes are of a much higher order of accuracy than could be hoped for in previous measurements of the kind; in fact, it is believed that they are not so very far behind trigonometric levels in precision. This increased accuracy has been rendered possible by the collaboration of the Egyptian Meteorological Service, which, thanks to the abundant weather-data now received and dealt with by it, is in a position to furnish a close approximation to the sea-level pressure at any point in the deserts at any given instant. The aneroids used in the desert explorations are compared with the standard barometer of the Meteorological Service immediately before the start and immediately after the return of each expedition. At each observation in the desert, the date and time, and the temperature of the air, are noted, as well as the approximate latitude and longitude of the place. On the return of the expedition, the recorded reading is corrected for the error of the aneroid on the standard barometer, and the Meteorological Service is asked to supply the sea-level pressure at the particular place and time of the observation. From the difference of the two pressures and the air temperature, the height of the place above sea-level is then worked out directly with the aid of Jordan’s table of “Barometrische Höhenstufen.” If the place of observation lies considerably above sea-level, a small correction is applied to the observed air-temperature to allow for the temperature-gradient in the air; for of course it is the mean temperature of the air column between the place and sea-level which should be employed in the reduction.

It is unlikely that with this procedure either the sea-level pressures or those observed at the places can be in error by much more than half a millimetre; and consequently, even with some slight uncertainties in the air-temperatures, the resulting altitudes may be expected to be accurate to within about 15 metres when only a single observation has been made, and to within much less than this when, as is usually the case at important points, the altitudes have been calculated from a series of readings spread over several days. That this degree of accuracy is actually attained seems evident on comparison of the levels obtained for the same places on different dates. Thus, for instance, at a camp close to Bir Terfawi, during a stay of five days in January and two days in March 1925, I made in all eighteen aneroid observations. The mean altitude found was 244 metres; the average deviation of a single observation from this mean was 8 metres, the maximum being 16. On the assumption that all sources of error are accidental, the probable error of the mean resulting altitude works out at less than 2 metres, which is about the same as might be expected in trigonometric levelling to the place from the Nile Valley.

The points in the interior of the Libyan Desert of which the levels are fairly accurately known now number many hundreds, and with the exception of a large blank in the unexplored west, and another between the Nile and Merga in the Northern Sudan, they are fairly well distributed over the country. It therefore seemed to me to be worth while to make another attempt at a contoured map; for the levels now known (though still far too few for contours to be drawn with any great precision) might suffice for the construction of a map on a comparatively small scale and with contours at fairly wide intervals; it was felt that a provisional map of this kind, besides affording at least an approximate general view of the relief, might help towards the solution of one or more of the problems which I have mentioned above. I commenced by sketching in the contours at 100-metre intervals on the original sheets of the 1:500,000 map of Egypt, a revised edition of which is in preparation. They were then reduced to the million scale, and afterwards still further reduced, with additions from my own recent observations around Owenat, Gebel Kissu, and Merga, from Hassanein Bey’s records of his journey between Jalo and El Fasher, from the altitude-data obtained by the Anglo-French Boundary Commission in the Sudan,[2] and from the maps of Prince Kemal el Din’s expedition of 1926 to Owenat and Sarra Well,[3] to form the map following p. 96. The attempt has, I think, fully justified itself, and more than fulfilled the hopes that were entertained concerning it.

Probably the feature that will first strike the attention on glancing at the contoured map is the series of depressions below sea-level stretching westward from near Cairo to Jarabub. The following table gives a list of the principal of these depressions, with the approximate extent of the areas lying below sea-level, and also (where known) the depths below sea-level of their deepest parts. Where the lowest part is covered by a lake, the depth shown is that of the water-surface, not that of the lake-bottom:

Depression.	Approximate area below sea-level in sq. kiloms.	Approximate depth of deepest part below sea in metres.
Wadi Natrun	220	23
Faiyum	700	45
Wadi Rayan	280	43
Qattara (including Moghara, Qara, and Tebaghbagh)	18,000	134
Sittra (including Timata)	300	33
Bahrein (including Nuemisa)	150	15 (?)
Watiya	70	15
Areg	60	25
Siwa (including Maasir and Maraqi)	800	17
Girba	20	(?)
Kheiba	40	(?)
Shiyata	20	(?)
Melfa (including Shebat, Exabia, and Gagub)	260	18 (?)

As will be seen from the figures in the table, by far the largest and deepest depression in this series—seven times as large, in fact, as all the others put together—is that of Qattara, which has not been shown on any previous map, and of which the existence was proved only last year. That so deep a depression should exist comparatively near to the Mediterranean and to Cairo, and have remained undiscovered until now, is striking evidence of how much has still to be learned concerning the configuration of the desert surface. The manner in which the discovery was made is perhaps worth placing on record. Hearing early in 1917 that a military patrol was about to operate in the Qattara region, I lent a small aneroid to the officer in charge of the patrol, and asked him to take readings with it at various points along his route, with the object of getting some idea of the altitudes in that region. It was, of course, certain that low-lying country existed along the foot of the Qattara escarpment, but no idea was entertained that any great part of it might be below the level of the sea. The officer brought back aneroid-readings which seemed to indicate that the spring at the foot of the scarp at Qattara was about 60 metres below sea-level; but the aneroid itself was unfortunately lost during the return journey. The result was too surprising for it to be accepted without evidence that the readings of the instrument had not been vitiated by accidental rough treatment, and consequently I refrained at the time from placing the levels found on the maps. But I resolved to confirm them or otherwise on the first opportunity.

That opportunity came last year, when it was possible to send a survey-party to triangulate westwards from the Wadi Natrun so as to cross the place where the aneroid readings had been taken. The work was entrusted to Mr. G. F. Walpole, of the Survey of Egypt, who had already distinguished himself by successfully carrying out a difficult triangulation from the Nile across some 500 kilometres of the Libyan Desert viâ Baharia to Siwa. The result of Mr. Walpole’s work was not only to confirm the substantial accuracy of the previous estimation of the level of the Qattara spring, but to bring to light the existence of a vast hollow, thousands of square kilometres of whose floor lie at even lower levels, and which at one place descends to a depth of no less than 134 metres below the level of the sea.[4] This last-mentioned spot, some 100 kilometres south-south-west of the Qattara spring, is probably the lowest-lying point on the land-surface of the African continent.

A scarcely less remarkable feature is the series of depressions comprising the oases of Kharga, Dakhla, Abu Mungar, and Farafra; while Baharia occupies a more isolated position about midway between Farafra and the Faiyum. In contrast with those of the northern series, these southern depressions do not penetrate to sea-level. The lowest points in Kharga Oasis are probably at or just below the level of the sea, but the general level of the Kharga floor is about 70 metres above that datum. The lowest point of Farafra is somewhat higher than the general level of Kharga; those of Dakhla and Abu Mungar are higher still, while in Baharia, though it lies farther to the north, the floor-level averages some 130 metres above sea.

A noteworthy thing about the larger depressions is that their northern boundaries are formed by steep escarpments, while to the south the slope is more gradual. This is the case, for instance, with the Faiyum, Qattara, Siwa, Kharga, Dakhla, Abu Mungar, and Farafra. Baharia is unique in being entirely surrounded by escarpments, as also in containing a large number of hills approximating in height to that of the bounding scarps. In the majority of cases the greater steepness of the northern boundaries can be correlated with geological structure; the northern walls of the Qattara and Siwa depressions, for instance, mark the southern limit of certain Miocene strata, while those of Kharga and Dakhla coincide with the southern limit of the Eocene limestones.

To the south-west of Dakhla Oasis there stretches a broad tract of rising ground, with the Gilf Kebir and the peaks of Arkenu, Owenat, and Kissu as conspicuous features. On either side of this tract the general level falls, on the one hand towards Kufra and Cyrenaica, and on the other hand towards the Nile. The Gilf Kebir is a great broken plateau of sandstone, rising very nearly to the 1000-metre contour and stretching in a direction a little west of south for over 100 kilometres. I saw the south end of this plateau from several points on the way from Terfawi to Owenat when travelling with Prince Kemal el Din in 1925, and fixed its position by compass-bearings. As seen from the south, the plateau appeared to be merely a flat-topped hill; its true form and extent were discovered by the Prince in the following year, when he passed along its eastern foot on his way from Pottery Hill to Owenat. The highest peak of Arkenu, in about lat. 22° 17′, long. 24° 46′, I estimated to rise to approximately 1410 metres above sea. The altitude of Gebel Owenat far transcends Hassanein Bey’s previous estimate of 1100 metres; a trigonometric measurement which I made in 1925 gave its summit as 1907 metres above sea. The general level of the ground at the foot of Owenat I found to average about 600 metres, thus agreeing with Hassanein Bey’s figures. Gebel Kissu, though lower than Gebel Owenat, is considerably higher than Gebel Arkenu, and being an isolated mountain with a single well-marked peak, it forms a very conspicuous landmark; my trigonometric determination gave 1726 metres for the altitude of the peak above sea.

To the south-west of Kissu there are a few widely scattered isolated hills, but the country in general forms a rolling plain of sand extending to the north-east corner of French Equatorial Africa with a nearly uniform level of about 700 metres. Beyond the French boundary the ground rises towards the Erdi Hills, which attain over 1000 metres.

From the corner of French Equatorial Africa to Merga, a distance of 250 kilometres, the country consists of alternations of sand-plains and flat stony tracts, with a very gradual fall to a level of about 560 metres at the edge of the Merga depression.

Between Merga and Owenat the country consists of sandy plains alternating with broken stony ground and occasional hills. The general level rises from about 570 metres on the northern side of the Merga depression to nearly 800 metres on the parallel of 20° 30′, then gradually falls to about 600 metres at the foot of Gebel Owenat.

1. Did the Nile, or a Branch of it, ever flow through the Libyan Desert to the Mediterranean?

The idea that a “dry river” exists in the Libyan Desert is a very old one, and seems to have had its origin in the fact that barren depressions in that region are sometimes called by the local Arabs Bahr bela ma, that is, “sea without water.” But the Arabic word bahr, which properly means a sea or a lake, as in Bahr Lot (the Dead Sea), is also applied in Egypt and the Sudan to a river, as in Bahr el Nil (the Nile) and Bahr el Azrak (the Blue Nile).[5]

The earliest depression to be discovered bearing the name of Bahr bela ma was the one now known as the Wadi Faregh (the “empty wadi”), situated immediately south of the Wadi Natrun. Father Sicard, who visited the place in 1712, correctly interpreted the Arabic name to mean literally “sea without water,” and having discovered fossil trees there, which he imagined to be the petrified remains of masts of ships, he inferred that a narrow arm of the sea formerly extended from the Mediterranean into the locality.[6] At a later date Sicard’s view that the petrified trees were the masts of ships was recognized to be a mistaken one; but a greater error was committed by changing the translation of Bahr bela ma from “sea without water” into “river without water,” as was done by D’Anville in his ‘Mémoires sur l’Egypte,’ published at Paris in 1766. D’Anville was a great authority in his day, and his translation was adopted by subsequent writers for over a hundred years. General Andreossi, who commanded Napoleon’s artillery in the expedition of 1798 and 1799, made a map of the “Dry River” and the “Valley of the Natron Lakes,” in which both depressions are shown open at both ends, instead of being closed in as they really are. Andreossi considered that the “Bahr bela ma” was the dry bed of an ancient branch of the Nile, which left the present Nile Valley somewhere in Middle Egypt and entered the sea to the west of Alexandria; he also thought that the ancient Lake Moeris was probably formed by the damming of the “dry river” near its supposed offtake from the Nile.[7] Later on, after Cailliaud and other early nineteenth-century travellers had discovered that other depressions existed in widely separated localities and bore the same Arabic name of Bahr bela ma, it seems to have been hastily assumed by the cartographers of the time that all the depressions bearing the same name were parts of a single dried-up river-channel. Thus, on a large map of Egypt compiled by Colonel Lapin and published at Paris in 1856, the “Bahar Belah-mah ou Fleuve sans eau” is stated to come from the Congo and is depicted as coursing through the Oasis of Dakhla, thence passing to the east of Baharia Oasis and on to the Mediterranean. On another large map of Egypt, compiled by Muzzi Bey, the then Director-General of Egyptian Posts, and published in Florence in 1876, a continuous valley is boldly shown leading from the mountains of Darfur to the Mediterranean, while cross-valleys, labelled “Old Bed of the Nile,” are depicted as connecting the main valley with that of the Nile at Korosko and Dongola.

The Rohlfs expedition of 1874 proved that there was no such channel running northwards through Dakhla Oasis; that the “Bahr bela ma” which had been crossed by Cailliaud between Siwa and Baharia was merely a closed-in local depression, and that the continuous empty river-beds which were shown on the maps of that period had no real existence.[8]

But though the continuous “Bahr bela ma” channel of the old cartographers has been shown to be purely imaginary, and has in consequence disappeared from our modern maps, the belief that an old dry river-bed may exist somewhere in the Libyan Desert has apparently persisted in many minds even to our own day. Only a few years ago I was called upon by the Government to discuss a suggestion, made in all seriousness by a person of considerable eminence, that the Nile, or a branch of it, must at one time have followed a course from somewhere near Dongola through some of the Egyptian oases to the Mediterranean; and that by tracing out the old channel, and deepening it artificially where necessary, a part of the river might be taken off along this path and its water utilized to irrigate the desert country on either side of it. To the few scientific travellers who have journeyed extensively in the Libyan Desert, it will doubtless appear incredible that such a suggestion as this could be seriously made. But old traditions concerning the geography of little-known regions die hard, and this particular one has probably been fostered by the circumstance that on most small-scale maps the oasis-depressions have been shown without any precise indication of the altitude of the intervening ground, as well as by the speculations of geologists as to the existence of a river in the region in past geological ages. Blanckenhorn, for instance, published in 1902 a series of small-scale maps depicting the course of a hypothetical river, the “Libyan Ur-Nile,” running northwards through the desert in the Eocene and Oligocene periods, and ceasing to exist in the Pliocene.[9]

That a river flowed during late Eocene or early Oligocene times in a north-easterly direction through what is now the Libyan Desert, and entered a sea near what is now the Faiyum, can scarcely be questioned; for it is only by the existence of such a river that we can account for the assemblage of fossil trees and curious remains of Eocene animals (including the ancestors of the elephant) discovered by Mr. Beadnell north of the Birket el Qarun and described by the late Dr. Andrews.[10] As to the size, length, and exact position of this river of the geological past there is very great doubt; but neither its situation nor even its existence can be considered material to our present inquiry. No one conversant with the known facts bearing on the general geological history of Egypt during late Tertiary and Quaternary times could make the mistake of thinking that the path of an Eocene river might be traceable in the present desert relief. The evidences of such path must be purely of a geological nature. In his endeavours to suggest what might possibly have been the course of the river on whose banks the Eocene animals lived, Mr. Beadnell found in two localities deposits which might have been formed along the ancient drainage-line; and in both these localities the deposits now occupy the summits of hills. There is insufficient evidence of any kind to justify so exact a location of the Eocene river as Blanckenhorn’s maps would suggest, and none whatever of any relationship, either between the present surface-configuration of the desert and Eocene or Oligocene drainage-lines, or between those drainage-lines and the River Nile. The appellation “Ur-Nile” of Blanckenhorn is thus a misleading one. If we are to look for traces of a dried-up river in the surface-forms of the Libyan Desert of to-day, it must be a river which, like the streams which eroded the great wadis of the Eastern Desert, existed in geologically recent times, and which became dry either through the continual deepening of the present Nile-channel, with consequent capturing of the former drainage, or through climatic changes causing a diminution of rainfall in the collecting area.

A study of the contoured map seems to me to be absolutely conclusive on the question at issue, for it indicates clearly that at no time within the geologically recent period can there have been any such continuous channel as has been supposed. Let us try and trace, with the aid of the map, the most likely course along which the ancient Libyan Nile-branch, if ever one existed, must have flowed. Starting from the Nile in the neighbourhood of Dongola, we observe that the lowest possible track is through Sheb towards Kharga. But Sheb can only be reached after passing over some 350 kilometres of desert lying well above the level of the Nile at Dongola, which is about 230 metres above sea. It is, of course, quite conceivable that even within geologically recent times the Nile at Dongola may have been considerably above its present level; so that we may pass over this first difficulty as not being definitely conclusive against the hypothesis we are pursuing. From Sheb to the south end of the Kharga depression there is a drop of about 160 metres, so that up to this point a channel may have been possible. But from here onwards, difficulties come in. If we turn northwards, we soon encounter the north wall of Kharga, rising to 300 metres above sea-level; so there is no continuing that way. Instead, we turn westwards towards Dakhla. But from Kharga we have to pass over ground reaching very nearly to the 200-metre contour for some 80 kilometres to enter the Dakhla depression, where we drop to a level of 119 metres. We have to traverse some 15 kilometres at altitudes above 200 metres to reach Abu Mungar, where the level is 117 metres; then a further 12 kilometres lying above the 200-metre contour to get to Farafra, where the level descends to about 90 metres. To get out of Farafra, we have again to traverse country lying over 200 metres above sea, whether we choose a north-westerly route viâ Sittra to the Qattara depression, or a north-westerly one viâ Baharia to the Wadi Rayan and the Faiyum. Taking the first of these alternatives as being the shorter, we cross some 30 kilometres of ground above the 200-metre level, and a further 120 kilometres lying between the 200- and 100-metre contours, before reaching the Qattara depression. And once we are in the Qattara depression (the lowest point of which is 134 metres below sea), there is no possibility of reaching the Mediterranean except by crossing a considerable tract where the ground-level is over 100 metres above sea-level.

A section of the route described, with a comparison-line passing through the present Nile-levels in corresponding latitudes, is shown in the figure below:

Can any one believe that a course of this kind was ever that of a river? To make a through channel would entail deep cutting for more than half the total distance of 1700 kilometres, and would consequently involve excavating hundreds of millions of cubic metres of rock, much of it of considerable hardness. Moreover, it cannot be supposed, in the present state of our knowledge of the region, that more favourable lines of communication exist and have been overlooked. Many depressions doubtless still remain to be discovered; but these, like those already known, will surely prove to be merely local basins. So many cross-country journeys have now been made in various directions through the desert, that it is inconceivable that any continuous channel of the kind we are discussing could possibly have escaped notice. Nor can we reasonably suppose that recent earth-movements have produced the present severance of the various depressions; such movements might perhaps be invoked to account for one or two of the separating elevations, but certainly not for all. We can therefore be absolutely certain that neither the Nile nor any branch of it ever passed through the Libyan Desert to the sea.[11]

Even channels of tributary streams to the Nile, such as abundantly survive in the great wadis which enter the Nile valley on its eastern side, are markedly absent in the Libyan Desert. The contours on the map suggest that possibly one such tributary stream may formerly have drained the south-western faces of the great plateau in which the depressions of Dakhla and Kharga form bays, and have entered the Nile somewhere between Aswan and Halfa; but if such a channel ever did exist, its actual traces have long since been obliterated by the smoothing action of subsequent denudation, and only the great escarpment of the plateau which it half encircled remains to tell the tale.

Photographs by Prince Kemal el Din, Dr. Ball, and Mr. Walpole

2. Origin of the Depressions.

It is a marked characteristic of almost all true deserts, that they are regions of internal drainage; and the Libyan Desert is no exception to the rule. There are a few gullies draining from the plateau to the sea along its northern edge, and a few others draining into the Nile Valley along its eastern border. But none of these external drainage-lines extends for more than an insignificant distance backward into the interior of the country. There is, in fact, in the Libyan Desert (especially in its north-eastern portion) a marked absence of any distinct drainage-lines at all; in their place we find extensive stony plains, which either slope gradually down towards the various inland depressions, or terminate in an abrupt fall where the depressions are cut back into them. Even in the Pluvial epoch, when Europe was glaciated and the deep wadis of the Eastern Desert on either side of the Nile were being cut out by great streams, there cannot have been any external drainage from a large part of the Libyan Desert; for if there had been, we should find the Libyan plateau far more intensely dissected than is actually the case. The rainfall in this region, even during the Pluvial period, must have been far smaller than in the neighbouring tracts of the Eastern Desert and Sinai; such rainfall as did occur was drained into the series of depressions already described, and there either formed lakes or was evaporated. But the action of inflowing drainage waters, carrying sand and mud in suspension and salts in solution, is not to deepen the basins into which they flow, but rather to fill them up. Hence we must look to some other cause for the formation of the basins themselves. A possible cause is to be found in earth-movements—either a local down-folding of the crust, or an upraising by faulting of the surrounding tracts. It is not unlikely that such earth-movements have to some extent conditioned the formation of the depressions; but that the depressions are not simply faulted-down areas or subsidences is abundantly clear from an examination of the bounding scarps and the floors of the larger oases. Baharia is wholly, and Kharga and Dakhla are partially, bounded by high escarpments, in many places hundreds of metres high, and the floors of these oases are composed of the same rocks as are exposed at the base of the bounding escarpments. It is, in fact, obvious that these great hollows are natural excavations, not subsidences. What was the excavating agency?

To the south-west of the limestone plateau which lies west of the Nile and extends northwards from lat. 23° 30′ there is fairly open sandstone country, and the oases of Dakhla and Kharga occupy great bays in the escarpment which marks the south-west termination of the plateau. It seems reasonable to suppose that the Eocene and Cretaceous seas had their southern shores somewhere about the latitude of Wadi Halfa, for the limestones do not extend far beyond the tropic; and to the south the hard limestones of the plateau probably passed gradually into softer shallow-water deposits, which would lend themselves easily to degradation and to removal by the agency of streams whose traces have long since disappeared. We can thus possibly invoke water-action to account for the primitive formation of the “bays” in which Kharga and Dakhla are situated. But we cannot in this way explain either the further deepening of these oases, which converted them into basins, or the excavation of the more northerly depressions such as Baharia and Qattara.

The problem as to how the excavation of the northern depressions, and the deepening of the southern ones, were accomplished is not altogether an easy one. There is, of course, no doubt that wind has been the main excavating agency. All the depressions occur in areas where soft rocks are overlain by hard ones, and once the hard overlying rock was removed at any place, the action of the wind on the softer beds would soon excavate a hollow. The difficulty is only as to how the hard overlying rock was first removed. To some who have studied the question, a sufficient explanation is to be found in the general degradation, combined with the tectonic disturbances—folds and faults—which are known to have affected some of the oasis-areas, as, for instance, that of Baharia. Given an anticlinal fold, a general lowering of the surface by wind-degradation would eventually result in the exposure of the soft underlying beds at the highest point of the fold, and subsequently continued wind-erosion would proceed most rapidly on these soft underlying rocks. Others, while admitting this explanation as partly accounting for the formation of the depressions, consider that other factors must have assisted to cause penetration of the hard overlying beds, especially in the areas where folding is not very marked.

Of such other factors, I can only suggest what appears to me to be a possible one. About twenty years ago I was mapping the Mediterranean coastal area in the neighbourhood of Mersa Matruh, some 160 miles to the west of Alexandria. In this area, which has about 6 inches of rainfall in an average year, there are numerous closed-in basins a few miles long and a few hundred yards in width, descending to a depth of 10 metres or more in the rather soft limestone which forms the principal rock of the locality. These basins, though called wadis on the maps, are more usually known as ’ebs or “bosoms” by the local Arabs. I was at first much puzzled to account for them; but eventually I arrived at what I believe to be the true explanation of their formation. The basins, of course, collect the surrounding rainfall in the winter, and the drainage carries with it into the hollows a certain amount of fine loam and sand from the disintegration of the surrounding surfaces. After the rains, the bottoms of the basins become covered with vegetation, while the surrounding rocky surfaces are left bare. In the hot dry summer, however, the vegetation perishes, and much of the loam is removed by the wind. We have only to assume (what seems indeed likely) that the rootlets of the vegetation disintegrate the limestone floors of the hollows to form fresh soil, and that more of this is removed each year by the wind than is brought into the hollows by the drainage, to find an adequate cause for a slow but progressive deepening of the basins. The primitive hollow may well be a very slight depression of the surface, such as might arise from greater local softness of the rock and consequent more rapid weathering. Once vegetation has commenced to grow in a slight hollow of this kind, the action would lead to a slow yet steadily progressive deepening year by year, and sand-erosion would tend towards an elongation of the depression in the direction of the prevailing wind. Can such a process as this, continued through vast ages during which the climate was somewhat less dry than at present, have been active in the formation of Baharia and similar depressions?

3. The Possibility of Utilization of Depressions in the Libyan Desert for Irrigation, Drainage, or Hydraulic Power.

The depression known as the Wadi Rayan, discovered by Mr. Cope Whitehouse about 1883, situated some 40 kilometres west of the Nile Valley and about 100 kilometres south of the latitude of Cairo, has been so frequently described and considered in regard to its utilization as a reservoir or as a flood-protection for Lower Egypt, that I need say little about it beyond a reference to the principal published works dealing with it.[12] There is no other depression in the Libyan Desert which has received so much attention from the economic standpoint. That the Wadi Rayan has not already been utilized as a reservoir has depended on a variety of considerations, of which one of the principal has been the doubts always entertained as to whether there might not be an underground leakage from it, resulting not only in a loss of part of the stored water, but also in damage to the adjacent flourishing province of the Faiyum. These doubts will not be lessened by the converging lines of evidence which I shall bring forward in dealing with the artesian water-supplies, all leading to the conclusion that there is an underground water-connection between the various depressions of the Egyptian Libyan Desert. Of late years, however, it has been proposed to make use of the Wadi Rayan, not as a reservoir, but as a sump for disposing of the drainage water of Upper Egypt by evaporation; and to this, in consequence of the lower level at which the water would stand in the depression, there is far less objection from the point of view of possible leakage or damage to the Faiyum.

The Wadi Natrun, which, though considerably less deep, is not greatly less in extent than the Wadi Rayan, has received some consideration as a possible sump for disposing of some of the drainage water of the Behera Province of the Delta. So long ago as 1895 the Public Works Department studied the possibility of draining Lake Mariut into the Wadi Natrun, and thus doing away with the increasing cost of keeping down the level of the lake by pumping water from it into the sea.[13] The Wadi Natrun has an area, even at the low contour of 10 metres below sea-level, of nearly 100 square kilometres, and the average rate of evaporation from an open water-surface within it could not well be much less than 3 mm. per day; the wadi could therefore dispose annually of at least 100 million cubic metres of drainage water by evaporation alone, and probably of a considerable additional quantity by seepage. But the high cost of cutting a drainage channel from Lake Mariut to the wadi, and the heavy maintenance charges which would be incurred in keeping the drain open, were formidable objections to the scheme. From a reconnaissance carried out in 1896, Mr. Verschoyle found that the length of the proposed drain would have to be almost 80 kilometres, and for the greater part of the way it would have to be a cutting 20 metres in depth through desert and ridges of drifting sand; he remarked that if it were an easy matter to make the connection, it would be no easy matter to maintain it; and he concluded that the scheme was an impracticable one.[14] A more recent proposal has been to construct a partly tunnelled channel instead of an open drain; but this has likewise been found to be impracticable, as involving too heavy an outlay for the benefits which would result.[15]