Chapter XXX: Appendix: I (1)

THE GEOGRAPHY AND WINDS OF THE LIBYAN DESERT

THE views on the geography of the Libyan Desert, current at the time of my visit to the country, are summarised by Mr. F. R. Cana in the valuable paper and map that he contributed to the “Geographical Journal.”[19] Writing of this desert he says: “Some knowledge of its character has been obtained where it borders the Nile, and to the north along the edge of the Cyrenaican plateau . . . the general character of the desert is that of rocky wastes in the north and east, and a vast sea of sand in the centre.”

The rocky wastes begin on the north, immediately to the south of the cultivable belt along the North Egyptian coast; the desert here rising to form a plateau. This tableland is succeeded by a vast depression, portions of which lie below sea-level. This hollow runs from Siwa Oasis in an easterly direction towards the Nile. Reference will be made to this huge valley later on.

The southern side of this depression has never, I believe, been mapped, but, beyond it, the level of the desert rises again to a limestone-capped plateau, that in places is over a thousand feet above sea-level. The western limit of this has not, I believe, been ascertained in its northern part, but towards the east it is bounded by the Valley of the Nile, where it forms the towering cliff on its western side that is familiar to all visitors to Egypt. About the latitude of Qena the plateau narrows down to a width of about a hundred miles, its western limit being the great escarpment that bounds the oasis of Kharga on its eastern side.

The geography of this limestone tableland, and of the desert beyond it up to the Egyptian-Tripolitan frontier, was tolerably well known at the time of my arrival in Egypt; but, with the exception of Kufara and the other oases in the same group, practically the whole desert beyond it was a _terra incognita_, the domain of the Senussi, and, as stated in Mr. F. R. Cana’s paper, considered to consist of an impassable sea of enormous sand dunes.

I made altogether eight journeys into the desert to the south-west of Dakhla, only some of which have been described. In addition to these, and to the others into Farafra and elsewhere already referred to, I made some to the north of ’Ain Amur, where I found a perfect network of small depressions.

These little hollows, which were only about a hundred and fifty feet in depth, mostly opened out of each other, and fairly honeycombed the limestone plateau. Entrance to them was gained by a gap in the cliff on the northern side of ’Ain Amur Valley, from which a small belt of sand dunes issued. They were mostly only a mile or two in width, though some of them were of considerable length, one in fact stretched for some thirty-four miles from north-east to south-west. There was one extraordinary little hollow that formed as it were an isolated pot-hole, nearly circular in shape, in the limestone plateau, about a hundred and fifty feet in depth with almost vertical sides. It was difficult to see how it could have been formed. Judging from the jagged skyline, shown by some of the cliffs surrounding these hollows, there were probably several other depressions that time did not allow me to explore.

A curious fact in connection with these little hollows was that, although many of them were floored with clay, they contained practically no vegetation. On the limestone plateau that surrounded them, however, bushes and even small patches of scrub were not infrequently seen.

The origin of the depressions in this limestone plateau has been the subject of some dispute. The action of water, sand erosion and folding of the strata have all been put forward to account for them by various writers. It may be interesting in this connection to mention that the southernmost of these little hollows to the north of ’Ain Amur contained the bed of a perfectly distinct watercourse, containing well-rounded shingle, running towards the entrance into the ’Ain Amur Valley.

In addition to these journeys I made one for three days to the south of Belat, in Dakhla Oasis. On the edge of the oasis we got into some very rough salt-encrusted ground, containing some patches of rock salt.

Beyond this lay a large scrub-covered area known as Dhayat en Neml, or sometimes as El Girgof, which was fairly thickly overgrown with bushes.

On leaving this we got into the open desert, whose level here rose fairly rapidly towards the south. Here we found traces of an old road, which, however, we lost on the third day after leaving Belat. It was probably only a branch of the Derb et Terfawi.

Early that afternoon our road led us to the top of a cliff about two hundred feet in height. A few hours’ journey farther on and a second cliff, two hundred and forty feet in height, was reached, beyond which lay an expanse of level sandy desert, dotted here and there with a few rocky hills. This second scarp was apparently the eastern continuation of the cliff that forms the southern limit of the sandstone plateau to the south-west of the oasis of Dakhla, which, however, breaks down into a gradual slope to the south of the town of Mut, on the western side of Dakhla, where the road from Mut runs to Selima Oasis.

I endeavoured to make the most of the few opportunities that presented themselves of gathering such information as I could of the unknown areas beyond the Egyptian frontier. But opportunities of this kind unfortunately were few. These parts were regarded by the Senussi as being their particular reserve, and they did their best to keep them closed to outsiders. It was consequently only members of this sect and their friends who had any knowledge of the district, and the Senussi were so extraordinary secretive about their country that it was with the greatest difficulty that I was able to extract any information concerning it. Enquiries, too, had to be conducted with caution, for collecting data of this kind was apt to prove an unhealthy occupation.

Part of the map I was able to compile from the data given to me has since been verified from other sources; as much of it has thus been found to be correct, the remainder of it is presumably equally reliable. The absolute positions, so far as latitude and longitude are concerned, are, of course, considerably in error in most cases, but the relative bearings and distances of each place to those surrounding it are, in most instances, represented with reasonable accuracy.

A map based on native information can, of course, never hope to compete in precision with one constructed by the methods of a modern survey; but it need not be very far behind those produced by the rough-and-ready methods of the geographers of a hundred, or even fifty, years ago. Its object is to give a general idea of the district it covers, and, more especially, to give future travellers an objective, and sufficiently accurate information as to its position, to enable them to find it.

No definite system of spelling was adopted in my map. Many of the names are clearly not Arabic, but either those of the Tibbus or Bedayat, whose alphabets—if they have any—have not, so far as I know, been reduced to any system of transliteration. It was impossible to get the spelling of even the Arabic names, as all my informants were illiterate _bedawin_, so I have spelt them as nearly as I could phonetically. Probably the pronunciation that I heard differs from that in vogue in Sudan, so the comparison may be of use.

The information was collected as far as possible in the form of through routes, joining places that had already been more or less accurately fixed by previous travellers. The remainder of the data was then fitted into these routes, or plotted from more or less reliably fixed points on other maps.

The three main routes on which the map was based were as follows:[20]

Route I. From Tollab, in Kufara Oasis, to Bidau.

Three days south to the well of Bushara.

Four days south to Asara, or Sarra (a well only).

Six days from Asara, S.S.W., to Tikeru.

Half a day west from Tikeru to Erwully, a well.

Three days west to Guru.

Three days south-west from Guru to Ungoury.

One day S.S.W. to the village of Ertha.

One day west to Bidau.

Route II. From Tikeru to Abesher.

Three days south to Wanjunga Kebir. Another Wanjunga, known as Wanjunga Sgheir—little Wanjunga—lies one day’s journey to the east. Both are inhabited. This district is sometimes called Wanjungat.

Three days south to Bedadi, a well belonging to the Bedayat.

Three days south to a well called Funfun, belonging to the Bedayat tribe.

One and a half (or two) days south to Wayta Sgheir.

A short day south to Wayta Kebir.

Five days south—I was also told four—to Mushaluba (Um Shaloba), which has recently been fixed by the French. The route then continued via Lughad and Aratha to Abesher.

Route III. From Wayta Kebir to El Fasher.

One day east to Um el Atham, a Bedayat well.

One day south to Baky.

One day south to the well of El Guttara.

Two days east to ’Ain el Baytha, a Bedayat well.

Two days east to Baou, a Bedayat well, or pool, in a fertile valley inhabited by Bedayat.

One long day, or a day and a half, south to Kuffara, a valley with plenty of water.

One day east to Medjoures, a valley containing many wells.

Three days and two “hours” east to Wady Howar. The road crosses the Howar Wady, and, after leading for a two days’ journey further east-south-east, it enters one of its tributaries—the Wady Faruwiah. After crossing the wady, the road leads for two days south-east to Musbut.

Musbut and Faruwiah are both said to be large wadies containing many wells. Musbut is said to be on the Derb el Arbain, the old “forty days” caravan road from the Sudan by which slaves used to be taken to Egypt.

Two “hours” south from Musbut to Buhuruz, two days south-east from Buhuruz, the road reaches Formah, a Zaghawa well.

Two days farther east, and it arrives at Kafut, a village in a wady of the same name, which, half a day to the south, receives a tributary known as the Wady Kobay. Wady Kafut itself, some seventy miles farther north, unites with the Wady Kuttum to form the Wady Meleeat, itself a tributary of the great Wady Howar.

Mr. Boyce’s survey shows a short section of the Wady Kuttum running east and west. Mr. Sarsfield-Hall shows Kuttum on his map as a village, situated upon an unnamed wady—presumably the Wady Kuttum—of which only a short section, running north-east to south-west is shown. But farther north he shows a large unnamed wady, discharging to the north, that corresponds well with the description I was given of the Kuttum-Meleeat Wady.

Three days farther east the road ends at El Fasher.

These routes were all taken as starting from Tollab, and when plotted on the basis of a day’s journey of twenty miles in a straight line, closed with the following errors, from the positions of their termini as shown on the maps: Bidau, 95 miles 34° from the true position, on a route from Tollab of about 430 miles as plotted—in other words, with an error of about 22 per cent of the total route. Abesher, 95 miles 98° from the true position, on a route of about 680 miles, i.e. about 14 per cent of the total distance; El Fasher, 160 miles 59° from the map position, on a fifty days’ journey of 1,000 miles, or 16 per cent of this distance. Considering the nature of the material upon which the routes were mapped, this compares not unfavourably—except in the case of Bidau—with a probable error in a prismatic compass traverse of about 10 per cent.

But this data may perhaps be misleading, for Colonel Tilho—a very accurate observer and equipped with a wireless installation for getting longitude by means of the signals from the Eiffel Tower—found that Nachtigal’s positions for some of the places in Tibesti were as much as fifty miles in error; so presumably his position for Bidau is not to be greatly relied upon. Rohlfs’ positions for Tollab, and other parts of Kufara Oasis, also remain to be checked by modern and more accurate methods than were available at the time of his journey.[21]

In constructing the maps, the routes were treated first as being three separate roads, joining Tollab to Bidau, Abesher and El Fasher respectively, and were adjusted separately by the ordinary graphic method, to their respective termini. This gave three separate positions for Bushara, Asara and Tikeru. These three positions in each case were meaned and accepted as correct, and from the position of Tikeru thus found the route to Bidau was plotted afresh and adjusted to Nachtigal’s position.

The routes south to Abesher and El Fasher were again plotted from the new position thus found for Tikeru, and were again adjusted to the map positions of Abesher and El Fasher, as two separate routes running to those places from Tikeru. This gave two separate positions for Wanjunga Kebir, Bedadi, Funfun, Wayta Sogheir and Wayta Kebir.

The mean of these two positions was taken as correct, and from this mean position for Wayta Kebir, the roads to Abesher and El Fasher were plotted and finally adjusted to their termini. Many of the places on these routes were consequently adjusted five times.

Route IV. Dongola to the Howash Valley.

From Dongola a road runs five days south-west to Bu Senata—a well belonging to the Kebabish tribe.

Four days farther west-south-west, and Jebel Maydob is reached. From there, five days due west, and the road comes to Bu Zibad, and, three days farther west, reaches the Howash Valley.

It was not until long after I had been given this route that I heard of the existence of old Dongola, which lies some sixty-five miles about south-south-east of the better-known new Dongola.

When plotted on a twenty-mile-a-day basis from either of these places, the road is considerably in error, if the position of Maydob can be relied upon. The route starting from old Dongola closing, however, much better on to Maydob than the one from new Dongola. The positions on the map are those found by plotting the information, without adjustment from new Dongola. Upon this route, too, depends the portion of the Wady Howash, lying to the north of Kowora (q.v.).

The description given to me of the Wady Howash was extremely interesting. It was said that the sides of the wady in places were covered with coloured paintings, and that it contained numerous ruins of burnt brick, probably Meroitic in origin, and in addition many statues—colossi apparently—and pits in the ground containing ashes covered with stone slabs—possibly these were funerary pits containing human remains. The Bedayat country generally was reported as containing many stone-built ruins (_ders_), and to have numerous rock inscriptions and “Roman”—i.e. artesian—wells, like those in the western oases of Egypt, and seems likely to prove a valuable field for future archæologists.

Among the remaining places shown on the map the following may be mentioned. My informants were all Arabs, or Sudanese living in Egypt, so the names are those in use among the Arabic-speaking _bedawin_ and may differ from those used by tribes speaking another language, such as the Tibbus, Bedayat and other Sudanese races.

Dendura: This is Rohlfs’ alternative name Zerzura—“the oasis of the blacks.” I concluded that Zerzura, if it exists at all, is a different place from Dendura. The latter was described as being as large as Dakhla Oasis, and as lying just to the west of an enormous longitudinal dune that is almost impassable, seven days due west from Bu Mungar (see also Zerzura).

Dunes: All the dune belts of the Libyan Desert are said to run from north to south. In the neighbourhood of Dakhla Oasis they were found to run 352° mag., and if, as I heard, they run parallel with the Tollab-Tikeru road, the belts appear in reality to converge slightly towards the north. Possibly the prevailing wind blows from a more easterly quarter as one proceeds westwards from Egypt, as the influence of the hot air rising from the Arabian deserts would be less felt in the western part of the Libyan Desert. Towards the Central Sudan it appears to approximate to the direction of the north-east trades normally found in these latitudes, for Commandant Tilho found the prevailing wind in Borku to blow from this direction. A dune belt that is easy to cross is said to commence about two hours to the west of Erbayana, near Kufara Oasis, and to extend for three days—I was told four—to the westward, and to die out before reaching the latitude of Bushara. A belt was also reported to exist about two hours to the west of Erwully. This belt, though about the same width as the one farther north, lying to the west of Erbayana, is perhaps only sand banked up by the Tibesti hills, though from its position it looks like the continuation of the belt near Kufara, with which it is exactly in line; Tilho’s paper confirms the existence of a great dune field as far as Ertha and Borku, and also to the south of Wady Dom. The line of dunes following the Tollab-Kufara road is said to start close to the west of Kebabo, and to be about a day’s journey in width from east to west. The belt is said to go to the Sudan, and to die out in the vegetation of Wanjungat. Between Wanjunga and Demi, Tilho found a “little chain of sand dunes, about fifty feet high, stretching from north-east to south-west, and extending from five to six miles in breadth,” which appears to be the end of the belt described by my informant.

The dune belts form a useful check upon the accuracy of the data given me. One I saw at the end of our journey to the south-west of Dakhla, near Jebel Abdulla, is apparently the northern continuation of that which runs from El Atrun to “the Egyptian Oasis.” The southern portion of the line of dunes lies farther west than the part that I saw; but a map constructed from native intelligence can hardly be expected to be very accurate, and, assuming the prevailing wind in the neighbourhood of El Atrun to blow from a direction approximating to north-east, as found by Tilho in Borku, the line of dunes would be certain to curve round somewhat towards the west at its southern extremity. The belt near Dendura is almost exactly in line with the one reported in the neighbourhood of Owana; so, too, is the wide belt to the west of Erbayana with the similar dune-field west of Erwully. Sand-free intervals in these sand belts are not uncommon, and the dying out of this belt about the latitude of Bushara that was reported to take place, may only be the commencement of one of these gaps, the line of dunes becoming continuous again farther to the south.

“Egyptian Oasis”: An Ebday (i.e. one of the Bedayat tribe) who was a friend of one of my guides, told him he had once ridden from Merga (q.v.) with two _hagins_ (i.e. riding camels) for five days north, following the dune belt (I was also told that the distance was ten long days from Merga with an ordinary caravan). He had then climbed a very high black hill lying in the dune belt and had seen in the distance under a cliff a huge oasis, containing a number of olive trees and much _terfa_. He was too far off to see if it were inhabited, and was afraid to go in, because he said it was an “Egyptian Oasis,” and he feared that he would be killed if he did so. Another Arab told me that a cousin of his was riding along the top of a scarp about eight days somewhere to the south of Dakhla, when he saw below him a very large oasis, containing a number of olive trees, palms and wells. There was one very big ruined town that did not seem to be inhabited and a few _ezbas_ (i.e. hamlets), in which a few people were to be seen. I was also told that this place was from seven to ten days from Dakhla Oasis.

Ershay: A lake of sweet water, variously described as being three miles across and five or six _feddans_ (i.e. acres) in extent; also a wady of the same name that runs into it. Entering the wady near El Guttara, the lake lies three days’ journey along the wady to the north. The lake is said to contain crocodiles, which seize camels when they come down to drink. There is nothing very improbable in this, for crocodiles have actually been found in the middle of the Western Sahara, in the pools of the Wad Mihero, that leads into the Wad Ighargharen, itself a tributary of the great Wad Igharghar, a valley that in prehistoric times must have contained an immense river. Presumably, as this part of the desert dried up, the crocodiles became cut off, and now exist only in the pools of the river bed—a similar state of affairs probably accounting for their presence in the Ershay Lake.

Fardy, Wady el: Another name for the Wady Tibbu, or Bahr el Ghazel, that connects with Lake Chad. Barth calls it the Barrum and the Fede. The latter name may perhaps be a corruption of the Arabic word Fardy.

There is also another Wady el Fardy that is said to run east of Jebel Kusu, through Guru, Erbayana, Buseima and Taiserbo. It crosses the Jalo-Kufara road four days south of Jalo, and then runs past Jarabub, Siwa and Bahrein to join the Nile. One account said that it also ran through the Fayum.

A large number of branch wadies are said to discharge into the Wady el Fardy from the Tibesti range. These valleys support a large population of Tibbus; but as these natives appear to be among the chief followers of the Senussi, I could not induce my informants to give me any information about this district, which appears to be one of the chief strongholds of the sect. Several wadies, however, are shown on existing maps, I do not know on what authority, as starting in Tibesti and running in this direction.

There are a number of native reports of dried-up river beds in various parts of the desert, that have never, I believe, been investigated, and may be without foundation, and information of this kind must be regarded with suspicion. But the Wady el Fardy sounds authentic, for wadies have already been reported to exist at all the points mentioned by my informants in its course. In the Tibesti range there must even now be a considerable rainfall in the rainy season, and probably in former times it was still heavier. The ultimate destination of the water that falls on the northern and eastern sides of Tibesti is still unknown. It is hardly likely that it can break back through the range and discharge towards the south, so apparently it must flow somewhere towards the north.

A plentiful supply of surface water exists in the Kufara group of oases—more than the extremely slight rainfall of the desert could possibly supply. In the Western Sahara the large oasis groups are fed by wadies of this description—the Wad Saura, for instance, brings down the rainfall from the Atlas Mountains to supply the oases of the Twat depression, while the Wad Ghirh group between Tuggurt and Biskra is supplied by the great Wad Igharghar, which takes its rise in the Central Sahara. It seems quite possible that the Kufara Oases and the Wady el Fardy are their counterparts in the Libyan Desert.

Not only are wadies known to exist at all the points along this river bed, where reported to me, but there is also known to be a large depression between Siwa Oasis and the Nile, portions of which, such as Sitra Lake, contain water and are below sea-level. The northern boundary of this great hollow has been surveyed almost throughout its entire length, the cliff running by Jebel el Ghazalat, Jebel Tarfaia, Jebel Dakar, Garet el Leben, Jebel Somara and Jebel Hashem el Gud, to the Wady Natrun. On the southern side of this great valley a well-marked boundary has been found to exist in the neighbourhood of Siwa, extending eastwards towards the Nile as far as Araj Oasis and Lake Sitra. Beyond that point it does not seem to have been surveyed. If this great depression forms part of the Wady el Fardy, it must have widened out to the east of Siwa Oasis into something like an estuary. The Wady el Fardy was described to me as being as big as the Nile Valley.

In addition to the Wady el Fardy, I heard of another great river bed known as the Wady Howar that in places was also said to be as deep and wide as the Nile Valley. The bottom of the valley was said to consist of clay and to contain much water after the rains—presumably in wells and pools—but to dry up in the hot season.

Hurry: The name of a Bedayat tribe. Also of the district that they inhabit, which contains a lake of good water, more than an Arab gunshot across from north to south, and an hour’s journey from east to west. There are a few trees and palms round the lake and some cultivation. The lake lies three days east from Wanjungat, the road lying all over sand and rock, it is seven days due north from Ershay Lake, the road being all over rock with a great deal of water in the rains, but none in the dry weather. There are settlements at the east and west ends of Hurry Lake.

Iddaila: Three days to the south-west of Iddaila is a large oasis, and two days to the west of Iddaila is a large _hattia_. A road runs from Iddaila to Kufara and another runs west, hidden under the dunes.

Ko’or Wady: Lies six days west of Erbayana.

Kowora: An important Bedayat district. To the east of Kowora, after one day’s journey the road is all level sand, west of which it is all rock.

Merga: The so-called “_hattia_ of the Bedayat,” of the _bedawin_ of the Egyptian deserts. It was described as being about the size of the Tenida-Belat district in Dakhla Oasis—say ten miles each way. It contains a pool, about an acre (_feddan_) in extent, fed by a spring, or _’ain_ (artesian well), which is surrounded by many palms, around which again is a belt of scrub—_argul_ and _terfa_. It lies two and a half days north-west from El Atrun and three to three and a half—I have also heard four—days west from Lagia. Two and a half days to the south-west from Merga there is a high cliff with a _negeb_ (i.e. pass) leading up on to the plateau on the top.

After rather more than half a day’s journey over the plateau another pass is reached, leading down into the “Valley of the Bedayat” (q.v.). Merga is not regularly inhabited, but the Bedayat come there in the season to gather the dates. They also use it as a base from which to raid the caravans of Egyptian _bedawin_ who go down to El Atrun to gather natron. A road runs from Merga to Kufara, Owanat (q.v.) being half-way along it. There are no Bedayat to the north of Merga. This is perhaps the place that Miani calls Ptolemy’s “lake of the mud tortoises.”

Colonel Tilho’s native information on the question of this _hattia_ agrees very closely with that I received. He, too, was told of the road from Merga to Kufara via Owanat, and also heard the place described as a pool surrounded by palms. He estimates the position as being “between the 25th and 26th degrees of longitude east and 18th and 19th degrees of latitude north.” This estimate agrees well with my information so far as the latitude is concerned, but, according to my intelligence, it should be at least a degree farther east than he puts it.

No’on Lake: Belongs to the Bedayat and is about three miles across. It lies close to the hill of the same name. There is cultivation on the south-east and south-west sides. It lies one day west from Jebel Kuttum. A road runs south and another east from the lake (? destinations).

Owana: A place half-way along a road from Merga to Kufara, consisting of a well, with no vegetation in the immediate neighbourhood, but much green grass in the district after rain, on which wild asses and “bekker el wahash” (probably Barbary sheep) feed. North of the well is a cliff with a pass that takes two hours’ easy travelling to traverse, and on the top is a high level oasis, the landmark for which is two rocky hills that look like one from the north; the scarp runs north-north-west and south-east from the well. Jebel Abdulla is reported to be visible from near here. The whole of this district is known as Owanat (plural of Owana). A dune belt comes down close to it from the north and dies out about two days farther to the south.[22]

Zerzura: Possibly only a generic name applied to any mythical or undiscovered oasis. I have heard it applied to Rohlfs’ Zerzura (Dendura?) (q.v.)—also to the “Egyptian Oasis” (q.v.), to an oasis supposed to exist eight days somewhere to the south of Dakhla, and to a stone temple I heard of about eighteen hours’ journey west of Jedida in Dakhla Oasis. The map on native information here reproduced originally appeared in the R.G.S.J., Sept., 1913. But then it was somewhat altered and a considerable amount of material was introduced into it, for which I was in no way responsible. The map is here reproduced in its original form.

The following are the principal roads that traverse the Libyan Desert: One starting from Giza, near Cairo, runs in a westerly direction through the Wady Natrun and the Wady Moghara, follows the course of the reported Wady el Fardy to El Qara, and then proceeds via Siwa, Jaabub, Jalo and Aujila to the oasis of Abu Naim and on into Tripoli. Farther to the south a caravan route, starting from the Fayum, crosses the desert to the northern end of Baharia, and just before entering the oasis is joined by roads from Wady Natrun and from Maghagha. Other routes, leaving the Nile Valley at Bahnessa and Minia, enter the oasis farther to the south, and combine to form a caravan route leaving the western side of the oasis and going, via Lake Sitra and Araj, to Siwa Oasis. A road also runs northward from Baharia, via Wady Moghara, to Alexandria, while another leads in a south-westerly direction, by way of ’Ain el Wady, to Qasr Farafra.

Farafra is connected with the Nile Valley by a road that runs through Kairowin _hattia_ to Beni Adi, near Assiut. A route that was used by the Senussi runs towards the south-west, past ’Ain Sheykh Murzuk and to the _hattia_ of Bu Mungar and thence, across a field of big dunes, to Kufara. It is said to pass on its way through a large oasis. To the north-west there is a road from Farafra, crossing a small detached limestone plateau, known as El Guss Abu Said, to a well called Bir Labiyat, and then proceeding via the little oases of Iddaila and Bahrain to Siwa. Another road leads to the south from Iddaila, and passing through a little oasis, known as Nesla, leads to Bu Mungar, and on to Dakhla Oasis, running along the foot of the cliff immediately to the east of Bu Mungar. A second leads from Iddaila, over El Guss Abu Said, direct to Dakhla, along the top of the cliff above Bu Mungar.

The usual road, however, between Farafra and Dakhla runs south-east from Qasr Farafra to Bir Dikker, and then follows what is known as a _gassi_, or path swept by the wind free from sand, through a field of biggish dunes, and enters Dakhla Oasis in its north-western corner. This road emerges from the southern side of Dakhla Oasis at the town of Mut, in the form of the Derb el Terfawi, leading to Bir Terfawi and Selima Oases.

Two other roads enter Dakhla Oasis from the north—the Derb el Tawil, or “long road,” that leaves the Nile Valley at Beni Adi, and, after crossing the limestone plateau, descends the cliff that bounds the oasis on the north by a pass known as El Agaba, in the eastern part of the oasis, and leads to Tenida; the other is a little-used road, called the Derb el Khashabi, that runs direct from the north-western corner of the oasis to some point in the Nile Valley to the south of Beni Adi, or perhaps goes direct to Assiut.

There are two roads in regular use between Dakhla and Kharga Oases—the northernmost leads to the well ’Ain Amur, high up on the northern side of a small detached limestone plateau, and the other, the Gubary road, runs along the foot of the cliff on the southern side of this plateau. The latter is waterless and rather longer than the ’Ain Amur road, but runs over the level, and is consequently the one most used, as the ’Ain Amur road necessitates a laborious climb up to the plateau, with a correspondingly difficult descent. It has, however, the advantage of having a well of rather bitter water about midway along it, and in consequence is occasionally used by small parties, especially in the hot weather. Several branch roads split off from the Gubary road, and lead to Beris and the places in the south of Kharga Oasis; but they are hardly ever used.

Dakhla Oasis is bounded on its western side by a field of dunes. The traces of several old roads are to be seen leading in this direction and, before the dunes came down to this part, were probably used by caravans proceeding to Kufara; but these are now, I believe, never followed. Traces of the disused road from Mut, via Owanat, to the Central Sudan has already been mentioned.

Several roads lead over the eastern cliff of Kharga depression; from the southern end of the oasis one goes, via the little oasis of Dungun, to Tumas, near Derr; a second goes, via the small oasis of Kurkur, to Aswan; a third leads in a north-easterly direction to Esna; while a fourth makes across the desert to Farshut. Esna can also be reached from the centre of the oasis by a road starting from Qasr Zaiyan. From Kharga village, in the north of the oasis, roads run to Sohag, Girga, Farshut, Qena, Luxor and Esna. But the main road running to Kharga is the Derb el Arbain, or “Forty days” route of the old slave traders, which leaves the Nile Valley near Assiut and enters the oasis at its northern end, runs through it from north to south, and then proceeds, via the wells of Shebb, Selima Oasis, Lagia and Bir Natrun, to Darfur.

The other great caravan road traversing the Libyan Desert from north to south is that starting from Benghazi, and proceeding, via Jedabya, Aujila and Jalo, to the Kufara group of oases—this was practically the route followed by Rohlfs’ and the Forbes-Hassanein expeditions. From Kufara it continues, via Bushara, Asara and Tikeru, into Wanjunga and on into Central Sudan.

It will be noted that most of the roads traversing the Libyan Desert run roughly from east to west. This is unfortunate, as it necessitates their crossing the sand belts which run approximately at right angles to them, the direction of these belts being due to the strongly predominant character of the northerly winds.

The winds of the Libyan Desert are a rather interesting problem. The prevailing direction from which they blow lies slightly to the west of north, which is not quite the quarter to be expected.

All along the North African coast lies the Mediterranean. The temperature of an area of water of this size and depth changes very slowly, and would not be appreciably affected by one or two days’ sunshine. In the absence of wind, the heat of the atmosphere lying over the sea is largely controlled by that of the water itself.

On the other hand, the Libyan Desert, owing to its lying nearer the equator, is subjected to a much hotter sun. Its intensely arid character prevents it from being cooled in any way by evaporation, and it is entirely unshaded by any vegetation. Under the heat of the tropical sun, the stones and sand of which it is composed very rapidly heat in the sun’s rays. The air overhanging it consequently gets greatly heated in its turn, becomes lighter and so rises, and the colder, denser air from the Mediterranean has to rush in and take its place.

A very similar state of things occurs in the Atlantic and Pacific oceans, where the cold air rushes in towards the equator from both north and south to form the trade winds. These winds, however, do not blow at right angles to the equator, for the earth in its rotation slips away, as it were, from under the surrounding atmosphere, with the result that the trade winds blow from the north-east and south-east, in the northern and southern hemispheres respectively.

But instead of blowing from the north-east, as would be expected from the character of the trade winds, the wind in the Libyan Desert comes from slightly to the west of north, which at first sight appears rather puzzling.

The explanation of this peculiarity lies, I believe, in the fact that to the east of Egypt—across the Red Sea—lies a like track of country to the Libyan Desert in the shape of the Arabian Desert, where very similar conditions prevail. The Red Sea, which separates the two, being too narrow to greatly affect the situation, the air overhanging the eastern part of the Mediterranean has also to take the place of the heated atmosphere rising from Palestine and a considerable part of Arabia, and consequently tends to draw in that direction.

Various phenomena of light and electricity, seldom or never seen in other climates, are to be found in the desert. I several times saw what appeared to be a large star suddenly flash up and disappear immediately. This I believe to have been not a meteor, but some form of electrical phenomenon, as I usually saw it after a sandstorm, when the air was full of fine electrified sand.

Solar phenomena are sometimes very conspicuous in the desert, the zodiacal light that follows the sunset being often very clearly seen, in the form of a semi-elliptical portion of the sky showing brightly above the horizon, after the sun had disappeared, and shining with sufficient strength to give quite an appreciable light.

There was another phenomenon that I saw in this desert that I have never seen alluded to by other travellers. Sometimes at sunset luminous lines would appear in the eastern sky radiating from a point opposite to the setting sun. These lines were never very pronounced, but at the same time were quite clearly visible. On one occasion, instead of luminous lines, dark ones appeared instead.

Another peculiarity of this part of the desert was the way in which a sandstorm would sometimes spring up without any appreciable change in the strength or direction of the wind. I concluded this to be due to a change in the electrical condition of the atmosphere.

The connection between the winds of this desert and its geography is so close that I have included them both in this section. The chief effect of the wind on the desert sands is, of course, to be seen in the dunes. In addition to those mentioned when dealing with the native information I collected, there is a large area in the Farafra depression covered with sand hills and those in the neighbourhood of Bu Mungar which have been previously mentioned.

Several narrow belts of dunes are also to be found starting in the great depression between Siwa and Cairo. These belts, too, run roughly from north to south; but with one exception they all die out before reaching the latitude of Farafra Oasis.

The effects of the violent desert gales was everywhere to be seen. On the limestone plateau, on the way to Kharga Oasis, and on the Derb el Tawil very fine examples of sand erosion are visible.

Mr. H. Ll. Beadnell,[23] who carried out a series of weather observations over a considerable time in Kharga Oasis, estimated that the sand blew from a northerly direction for five days out of six in this district. Violent sandstorms are unpleasantly common, by far the greatest number of them coming from the predominating northerly direction. The result of this uniformity in the direction of the winds is clearly seen on the plateau. The swirling clouds of sand, driven by the furious desert gales, have in many places not only shaped the hard limestone into long ridges running up and down wind, but have even cut grooves in the ridges themselves, as though with a gigantic gouge, forming a type of desert known to the natives as _kharashef_. These ridges rise sometimes twenty feet above the level of the plateau, but more usually are only three or four feet in height, producing a surface somewhat resembling a rough sea—a very uncomfortable type of desert to have to cross with a caravan. A diminutive form of this—_kharafish_—in the shape of small ridges, often with a cutting edge, is also frequently seen, and works havoc with the soles of the soft-footed camels. Besides these forms a flat rock surface with very shallow groves, known as _rusuf_, is occasionally met with.

Some of the limestone boulders lying on the surface of the plateau are perforated in the most extraordinary way. The driving sand apparently eats its way into the softer portions of the stone, boring holes into its surface. Small pebbles are often to be seen which have been blown into these holes. These fly round and round in the excavation under the influence of a strong wind, and presumably continue the erosion of the sand blast, in the same way as a stone wears a pot-hole in a stream. In course of time the whole boulder becomes so riddled with holes as to resemble a gigantic sponge.

In several places are large patches of desert more or less closely covered with round boulders up to a foot in diameter, a type of erosion known to the natives as _battikh_, or water-melon desert.

In other places in the desert perforated rocks and small natural arches are to be seen; while near Farafra village were a number of fine “mushrooms” and table stones cut out of the chalk by wind-driven sand. Similar “mushrooms” of sandstone were, moreover, met with near the centre of the desert.

In Kharga Oasis there is an area several square miles in extent covered by curious clay ridges. These, which seemed all to be under twenty feet high, were evidently formed by the erosion of the earth by the wind-driven sand, for they all ran roughly from north to south, in the direction of the prevailing wind.

Apparently, as the sand wore away and lowered the surface of the desert, it encountered here and there harder portions of the clay which resisted its erosive action. These consequently remained protruding above the surface of the desert as the surrounding clay was eaten away by the sand blast, and consequently acted as a protection to the earth immediately to leeward of them, which remained intact above the level of the desert in the form of a ridge running in the direction of the prevailing northerly wind. I have found similar forms to the west of Dakhla and in that oasis itself.

While in the central part of the desert in my first season, I found embedded in a sand dune two short pieces of dried grass much frayed and battered;[24] so, as has already been mentioned, on leaving the camp next day, we followed the line of the sand belt, to the north as showing the direction of the prevailing wind, and so found the place from which the dried grass embedded in the dune had come.

The occurrence of this grass so far to windward of the piece I had picked up among the dunes is only another illustration of the great

part the strongly predominant character of the northerly wind plays in this desert. Had we continued marching towards the north, along the same bearing as we had during the day, we should have found the oases, or _hattias_, of Bu Mungar, Iddaila and Sitra, all on the same line. The original seed, of which the grass we found were the descendants, were probably specimens carried by the wind from Sitra to Iddaila, where they took root and produced seed that was similarly carried to Bu Mungar and from thence—perhaps through another _hattia_, or oasis—to the place where we found it growing. Very probably the line of plantations of the grass may even reach to the Sudan, should there be any places along it where the seed could germinate.

It is not only in the distribution of the vegetation that the agency of this predominant wind is apparent. The sand belts follow the same direction, so, too, do the clay ridges in Kharga and elsewhere; even the hills in the desert show in many cases a distinct tendency to have their longest diameter pointing up and down this north to south line. The predominant character of the northerly winds even has an effect upon the caravan roads; that, for instance, that runs from Farafra Oasis to Dakhla lies along a hollow among the sand dunes that occupy a large area in the centre of Farafra depression, that is kept permanently free of sand by the northerly wind that continually sweeps along it, so affording an easy path through what would otherwise be a most difficult dune-field to cross.

Once at ’Ain Amur I experienced a curious storm. It was a windy morning and I had been out after gazelle. On returning towards the camp, I saw in the distance what looked like very heavy rain falling in the ’Ain Amur Wady, so hurried back to the well to gain the shelter of the ruins in its neighbourhood.

The rain, however, turned out to be only a fog—or perhaps, since the well is 1,680 feet above sea-level, it would be more correct to say cloud—only a few drops fell; but for nearly an hour the whole district was enveloped in a thick white fog, while a furious wind sprung up and blew with great violence. At the end of that time the fog cleared off, the wind suddenly dropped and a sunny and unusually hot day followed. Fog, it is needless to say, is not a common occurrence in the desert.

The word oasis is unfortunately a very vague one, to which varying meanings are applied in different parts of the world. The dictionaries usually define it as a “fertile spot in a desert.” So far so good. The difficulty comes in when one attempts to apply this definition, for one is at once faced with two problems: what constitutes “fertility” and what is the size of a “spot”?

The Wad Ghirh district, for instance, in the south of Algeria, runs for some ninety-five miles from north to south, contains about forty villages, with their palm plantations, and a population of some fifteen thousand inhabitants. Each of these villages, with its palm plantations, is regarded as one oasis. The whole Wad Ghirh district being considered a group of oases.

Kharga and Dakhla Oases, which are separated from each other by about seventy-five miles of waterless desert, were considered by the ancient writers to form one oasis, known as the “great oasis.”

Kharga Oasis, according to Dr. John Ball, who surveyed it for the Egyptian Government, is about a hundred and forty miles long from north to south, contains some fifteen villages and has a population of seven thousand eight hundred odd. All this is considered even now to form one oasis.

To the north-west of Kharga lies a little place known as ’Ain Um Debadib, where there is a considerable area of scrub, with a few palms and _sunt_ (i.e. acacia) trees. A certain amount of ground is cultivated here by a family from Kharga village, who, however, do not always reside there. This is usually considered a separate oasis.

In the eastern part of the Farafra depression is an area, some sixteen miles by twelve, covered with scrub, among which a few date palms, which seemed to be to some extent cultivated, are also to be seen. This place is known as Kairowin. There are no permanent houses or inhabitants, though frequently _bedawin_ living in tents, or brushwood huts, are to be found there with their camels. The place contains several wells, and is sometimes described as an oasis, though the natives more usually allude to it as a _hattia_.

At ’Ain Amur, half-way between Kharga and Dakhla, is a well, some ruins, a patch of scrub and a palm or two. This, however, never has any residents—the well being only used by travellers between Kharga and Dakhla. This place is usually described as a _hattia_, though I have heard even this called an oasis.

From places such as this there is a regular gradation through scrub-covered patches containing water, but no palms, scrub-covered areas containing no water, and mere patches of half a dozen bushes clustered together in the desert, called by the Arabs _roadhs_, to places that remain sufficiently damp after one of the rare desert showers to allow a few widely scattered blades of grass to grow, which the _bedawin_ call _redirs_.

It is very difficult to see where to draw the line. The system adopted in Algeria of describing places like Kharga and Wad Ghirh as groups of oases—or oasis-archipelagos, as they are sometimes called—seems preferable to the Egyptian plan of alluding to them as a single oasis. It would seem advisable to confine the term oasis to a place which is actually cultivated, whether continuously inhabited or not, and to use the word _hattia_ for “fertile spots” in the desert where no cultivation exists. The dividing-line, however, between a _hattia_ and a deserted oasis would not be very sharply drawn.

The western oases of Egypt were known to the ancients, for some reason not very apparent, as “the Islands of the Blest”—a name that to a modern visitor has an air of being somewhat ironical.

They are usually said to lie in depressions of the plateau. So far as Baharia is concerned this is an accurate description, for it is almost entirely surrounded by cliffs; but in the case of Farafra and Kharga, the scarps of the limestone plateau only partly hem them in, and they would be more correctly said to be situated at their base. Dakhla, in the same way, lies at the southern foot of the chalk plateau that forms the floor of Farafra Oasis, and gradually rises in level as it proceeds southwards from Qasr Farafra, until it breaks down in the huge cliff that bounds Dakhla Oasis on its north and east. The little oasis of ’Ain Um Debadib, lying slightly to the west of Kharga, is also only bounded on the north by the cliff. Kharga towards the east, north and north-west is hemmed in by cliffs and hills. The cliff in places, on the eastern side, is nearly eight hundred feet in height, but is considerably lower at the northern end.

Starting from the north, on the western side of the oasis, a jumbled mass of hills, cliffs and sand dunes extend for about thirty miles towards the south to form its boundary. South of this, a long north and south belt of dunes runs parallel with the eastern scarp, and cuts off the oasis from the open desert lying to the west. At its southern end, the oasis merges into the desert without any well-defined boundary.

To anyone standing, say, on the eastern scarp and looking down on to the floor of the oasis, it would appear as flat and almost as featureless as the surface of the sea when viewed from the summit of a high cliff. Two huge flat-topped hills rise abruptly from the level oasis floor, near the eastern cliff, to the height of the plateau above it; they are known as Jebel Ghennihma and Jebel Um el Ghenneiem, and stand about thirty-five and forty miles respectively from the northern end of the oasis. Almost facing them on its western side stand Jebel Taaref and Jebel Ter. With the exception of these four hills and the smaller conical peak known as Gorn el Genna jutting up from the centre of the oasis floor, there are no conspicuous elevations to break its level monotony.