Chapter II: The Southlands (2)

North of the broad Tekursat valley, with scarcely any marked channel and sparsely covered slopes, is a low plateau with three small valleys, rejoicing in the uncouth name of Teworshekaken. Beyond is the Inafagak valley, and finally the smaller and probably tributary valley of Keta. From here to the Taberghit valley the bush thins out more and more; patches of bare sand become frequent, and the trees are considerably smaller. In none of these valleys has the rain-water left a definite bed of flow, though dry pool bottoms and short sections of channel may be seen here and there. The valleys are sometimes several miles from side to side; they were probably in the first instance longitudinal depressions between heavy sand dunes formed along the direction of the prevalent wind; the sides are even now of too recent formation and too permeable to spill the rain-water into definite beds along the bottoms.

At the southern edge of the immense Taberghit valley the character of the country changes quite definitely. The surface becomes dotted with little hummocks where the sand has been washed against a small bush or piece of scrub; otherwise the ground is bare. The few trees are grouped in scattered clumps. The ground vegetation is no longer predominantly “Karengia,” but one of several kinds of less offensive and more useful desert grasses impregnated with salt. The best camel fodder, curiously enough, is the true desert vegetation. The animals eat it avidly on account of the salt it contains, and even long periods of drought do not conquer its obstinate greenness. Its nutritive power is greater and it is more wholesome than the luxuriant Southland fodder.

At Taberghit a track runs direct to Agades by way of Ihrayen spring. When both the eastern roads were in use, the Hannekar track was used by people going to Agades, while the more eastern Farak-In Asamed route by way of Tergulawen was frequented by caravans bound for Northern Air.

A day before reaching Milen well you feel very strongly that the Sudan lies behind. The last bush has been left near Taberghit. In front is an open depression perhaps five miles wide and not more than fifty feet deep: it contains no stream bed, but here and there patches of dry cracked mud indicate the formation of short-lived rain pools. East and west the same stark valley runs as far as eye can see. Its course is clearly defined and it is without intersecting basins or tributaries or curves. On the far crest are loose buff-coloured sand dunes and then a few small acacias. The levels gradually rise in a series of folds, one of which contains the closed basin and disused Anu n’Banka[59]; another forms a valley called Kaffardá, which is like Taberghit but on a smaller scale. The folds lie parallel to one another along the line of the prevalent E.N.E. wind which always blows in Azawagh. This wind is one of the peculiarities for which the country is notorious. Both times I crossed this region it was blowing with great violence. In June it was suffocatingly hot; I camped one noonday to rest out of sheer exhaustion in a group of trees on the northern side of Taberghit. There was practically no shade: the leaves of the stunted trees were too thin to shelter even three persons. The temperature was over 110° F. in the shade, and visibility did not exceed a quarter of a mile, owing to the blowing sand and dust. Six months later I returned the same way. The same wind was blowing, but it was so cold at midday that I was unable to keep warm, even walking, with two woollen shirts, a drill coat, a leather jerkin and a blanket over my shoulders. Where a bush or sand dune offered shelter from the wind the sun was quite hot, but that night the thermometer fell to 31° F., after having registered 92° F. at 3 p.m. in a sheltered spot in the shade. It was very unpleasant. Barth’s experience of the wind and cold of Azawagh was much the same as mine. He writes: “The wind which came down with a cold blast from the N.N.E. was so strong that we had difficulty in pitching our tent;”[60] it was responsible for the most “miserable Christmas” he had ever spent. I was there a few days before Christmas in 1922 and can vouch for the accuracy of his verdict. Even the blinding glare and heat of June were preferable to the bleak cold of the winter nights.

One effect of the constant wind is that the longitudinal dunes in Azawagh have retained their characteristic form more generally than further south. Their gentle rounded contours, which the wind tends to restore whenever the rain happens to have modified them, are characteristic. There is, of course, less precipitation here than further south, though it has been sufficient in Tagedufat to produce a considerable growth of desert vegetation along the bottom of the valley, where there are a number of small trees and an abundance of every conceivable type of salt bush and grass. It is said at certain seasons of the year to produce the finest camel fodder in this part of Africa.

All over Azawagh are numerous deserted sites where millet used to be grown on the sandy slopes. The people who cultivated this arid country lived in temporary tents and huts except further north between Tagedufat and Milen, and consequently no trace of their dwellings remains. The evidence, however, of cleared and levelled patches and of broken earthenware is as unmistakable here as in Damergu. Between Keta and Tagedufat there is a succession of such clearings. It is borne in upon one that this heavy buff-coloured sand country where only desert vegetation now appears to thrive is in reality quite fertile so long as it receives any rain at all. The climate has probably not altered enough in recent times to account for the desertion of Azawagh; it seems rather to have been due to a decrease of the population. The Kel Azawagh, according to tradition, were numerous at a time when Damergu was thickly peopled, and there was not enough land available there or in Air to satisfy the needs of a people squeezed between the south and the north, whence the population was constantly being driven into the Sudan. It is clear that the Kel Azawagh who made these millet cultivations in a zone of desert steppe must have been of a fairly sedentary disposition, for a nomad people would have contented itself, as the modern Tuareg inhabitants of Azawagh do, with grazing herds and flocks on the excellent pastures.

In referring to the Kel Tegama a plea was advanced that the name was primarily a geographical one, and one not properly appertaining to a single tribe. The name Kel Azawagh, to which the same considerations certainly apply, is found to some extent interchangeable with Kel Tegama. Now it will be shown later that the Tuareg of Air and Damergu only reached these lands comparatively late in history; consequently an allusion in Ptolemy to a Tegama people appears to refer to a non-Tuareg folk in this or some other area of the same name. I see no reason to doubt that it was these Tegama and Azawagh areas which were meant by Ptolemy, and therefore conclude that before the Tuareg arrived they were possessed by a people to whom the millet clearings and village sites are probably due. The later Tuareg Tegama, or Kel Tegama, as we should more properly say, as well as the Kel Azawagh, were merely a section of People of the Veil who later lived in the areas, and in the course of time were named after them, though it is possible that the name Azawagh was one given by the Tuareg to an area previously called Tegama by its former inhabitants.

We shall see[61] that among the ancient divisions of the People of the Veil in the Hawara group is a Kel Azawagh. The peculiarities of the Hawara clans would not connote any sedentary instinct in this tribe, whether it lived in this or in another area called Azawagh; but when we find in the Tetmokarak tribe of the Kel Geres group now living near Sokoto (whither they migrated from Air through this Azawagh area) a subsection called Tegama, and when we have learnt[62] that the Kel Geres are almost certainly a Hawara people, we can be even more inclined to the view just suggested regarding the use of the names Azawagh and Tegama and the origin of the people at various times living there. As a tribal name Kel Azawagh has now disappeared. The French 1/2,000,000 map displays it in the valley between Agades and the Tiggedi cliff, but out of place, for when still in use it was applicable to an area rather further east. Although it is no longer a proper name, it serves the Ifadeyen who now live in Azawagh for a descriptive term of themselves in accordance with the usual practice regarding local tribal nomenclature.

In the periods between the rains the village sites in the Taberghit or Tagedufat valleys watered at the deep wells of Tagedufat, Anu n’Banka, Aghmat, Taberghit and presumably Tateus, though I know nothing of the last named. All these wells have now become silted up by wind-borne sand, but could easily be cleared if the population returned, as the water has not disappeared.

The whole area between Taberghit and Tagedufat is covered with small mobile dunes; the two valleys themselves are, however, free of them. There is no loose sand at all in the Tagedufat valley, a curious phenomenon probably connected with the eddies formed by the prevalent wind in the channel of a depression between the higher banks. If this were true, the existence of dunes at Kaffarda would conversely point to its being an isolated basin, and this indeed is probably the case. Anu n’Banka is in a little hollow, the sides of which are also covered with small dunes. The bottom itself is clayey and free from blown sand, showing traces of having been a rain-pool at certain seasons. Surrounding the depression are millet clearings and a little rock outcrop. It is the most southerly point in Azawagh where stone occurs, and the outpost of the more conspicuous rock formations of the Tagedufat valley.

Although the first part of the descent into Tagedufat is imperceptible, the appearance of the ground has changed considerably on account of the small crescentic dunes of very fine white sand which overlie the heavier buff-coloured sand of the surface. The crescentic type is characteristic of young dunes in process of formation,[63] their last stage being the long whale-back down of heavy particles which tend to settle or become cemented and eventually to support some vegetation. The Azawagh valleys present a series of interesting examples of the youngest type of dunes, which are still moving rapidly, superimposed upon the oldest fixed dune formations oriented along the line of the prevalent wind. It is curious that at no point has the fine and very mobile sand which is continually being carried in from the great Eastern Desert collected in large masses: the small crescentic bodies, the horns of which, of course, lie down wind, or, in other words, point west to south-west, are neither continuous nor contiguous. The underlying buff-coloured surface is covered with a number of small trees and scattered scrub or grass in isolated clumps. This vegetation becomes covered by the crescent dunes and in time uncovered as the white sand moves westward. Where this vegetation can be seen emerging from the crescentic formations on the windward side it is still alive, pointing to a fairly rapid motion of the body of sand. It is true that some of this desert scrub is sufficiently hardy to withstand a period of, it is said, as much as four years without any rain, and even then it only requires very little moisture in the air or some dew; the numerous small acacias, however, if wholly engulfed for any length of time, would die. Yet at no point is there either a wake of dead vegetation behind the larger crescentic dunes or even an unduly large proportion of dead trees. The progress of the small dunes is therefore undoubtedly rapid, and is due to the constant wind, which should, however, have tended to create larger masses. The crescentic dunes are rarely more than twelve feet high at the most; their individual area is, of course, relatively large owing to the very flat slipping angle of the fine grains. Barth records dunes as far as Tergulawen; but there is no evidence regarding the country east of this point,[64] which is probably too far north of the dune belt on the Termit road to be connected with that zone.

The Tagedufat valley bottom, unlike the Milen and Taberghit valleys, is marked by a more continuous stream bed along which water flows every year for a short time during the rains. The most remarkable feature of the valley is a series of flat bare patches formed by the pools of rain-water; they are of no great size, but the surface is stained bluish-white by chemical incrustation. The Milen and Taberghit valleys, while possessing a few similar rain-pools, none of which survives for more than the briefest period, do not exhibit this complexion. The point is of particular interest in connection with a report given to me by my guide, Sidi, who was with me on the way south. He is a widely travelled and knowledgable man. He stated that the Tagedufat depression extended eastwards across the desert all the way to Fashi, and was marked along the whole of its course by such patches of chemical incrustation. My travelling companion, Buchanan, observed that the ground shortly before reaching Fashi was stained in the manner described. In the open desert, where in the immensity of space it is difficult to determine the direction of a very slightly accentuated valley, such noticeable features are valuable evidence.

Considering the size of the Tagedufat basin south of Milen, the valley shown as extending towards Termit on the French 1/2,000,000 map and called Tegemi (Téguémi), is perhaps a confluent, or even an inaccurate representation, of the main valley itself. A recent Camel Corps[65] reconnaissance from Talras to Eghalgawen possibly followed up one such affluent in the east bank of the main channel of Tagedufat. The importance of the Tagedufat valley from the hydrographic point of view cannot be over-stated.

Directly the Tagedufat valley is crossed the rock outcrop on the north bank becomes a striking feature. Increasing in size towards the west, it falls away below the surface to the east. Crescentic dunes reappear between the outcrops and continue almost all the way to Milen. On the north side of Tagedufat, near the track, for which it serves as a landmark, is a prominent mass of black rock called the Kashwar (Stone) n’Tawa or Tawar. Far away to the N.N.E. the relief becomes bolder, rising to a group of small summits clothed with loose sand, called the Rocks of Oghum. The remains of some stone houses, at one time the southernmost permanent settlement of Air, appear in the loose sand near the hills. North of Oghum in a little depression filled with acacias is Gharus n’Zurru.[66] After a further stretch of dunes a small valley running northwards diversifies the general lie of the ground. It is called Maisumo, and contains another deep well which is still in use. This valley after a short distance runs into the Milen depression, with the conical hill of Tergulawen visible to the east and the little massif of Teskokrit to the west. The northern part of the latter group extends eastwards from the main summits as a steep ridge forming the northern bank of the Milen valley itself.

East of Tergulawen again is a small and almost unknown group of hills called Masalet, where in recent years Kaossen, afterwards leader of the Air revolt in 1917, dug a well. It only yielded brackish water, which, though good enough for camels, proved too medicinal for the Tuareg, who filled it in again. It had been dug for political purposes largely in order to facilitate parties from and for the Southland participating in the yearly caravans which fetch salt from Bilma. Masalet was designed to obviate these parties making a detour along the River of Agades or via Eghalgawen: it provided an easterly watering-point in Azawagh corresponding with Tazizilet further north in Air itself. The unsatisfactory nature of the supply, especially for caravans engaged in crossing the eastern desert, did not, however, justify the risk of leaving so remote a watering-point available for Tebu raiding parties. The fact that Masalet was constructed in recent years is interesting, as showing that the Tuareg have not lost the art of locating deep water.

The western road from Tanut to Agades via Aderbissinat and Abellama runs over much the same sort of country as that which I have just described between Farak and Milen. Aderbissinat well, seventy-five miles from Tanut and ninety-three miles from Agades, is a point of such strategic importance that the French from Zinder built a fort there during the war in order to secure their communications with Air. It has not been garrisoned of late, but proved of paramount importance during the operations of the column which marched from the south to relieve Agades during the rebellion of 1917. With the exception of the deep but copious well of Abellama, there is no useful permanent watering-place between western Damergu and Agades, as the spring of Ihrayen in the Tiggedi cliffs has too small an output to provide for many animals. Nineteen miles north of Aderbissinat the bush ceases. As at Taberghit further east, the country rises some 200 feet to an average level of 1700-1800 feet above the sea. Beyond Timbulaga sand dunes appear on the level buff-coloured steppe, which is covered with the usual scanty vegetation of desert grass in tussocks.[67] The ground then slopes gradually down to the deep well of Abellama in Lat. 16° 16′ 30″ N. and Long. 7° 47′ 20″ E. G. Abellama as a stage corresponds with Milen on the other road.

On the easternmost or Tergulawen road Barth[68] shows that the country is again substantially the same. South of the “spacious” well, which is in a depression “ranging east and west,” with sand-hills on the south side bearing a sprinkling of desert herbage, the country is covered with small dunes on a “flat expanse of sand, mostly bare and clothed with trees only in favoured spots.” To the north is a great sandy plain running as far as the Ridge of Abadarjan, where the level descends to the upper basin of the River of Agades. The area is covered with “hád,” the most nutritious of desert plants and the most characteristic of the desert steppe of Africa. In all parts of the Sahara the distribution of the plant marks the division between the Desert and the Sown. This “hád” of the border line advances or recedes, sometimes from year to year, according to the rainfall. It is the tidal mark of the desert.

The northern part of Azawagh is geographically important, as it contains the transverse valleys which collect the southern rainfall of Air and carry it westwards into the Niger basin. The course of the Beughqot (Beurkot) and Azelik[69] valleys is wrongly shown on the French maps. They do not unite until they have reached a far more southerly point than where they are shown to do so on the Cortier map. Furthermore, when they have joined, they turn S.W. and not S.E. A recent reconnaissance as far as Masalet proved that after these two valleys meet they turn west into a large depression which is probably the same one as that in which the well of Milen is situated, though it might, on the other hand, be the Tagedufat basin; this is a point which must for the moment remain undecided. On a solution of this problem depends the answer to the question as to whether Milen or Tagedufat is the principal basin into which the Air valleys east of Beughqot as far as Tazizilet drain. All that is clear is that they turn southwards and then westwards to join one of the two systems in question, and do not peter out in the desert as Cortier’s map suggests.

West of Milen well the valley in which it is situated eventually joins the lower Tagedufat, which runs on S.W. or W. towards the Gulbi n’Kaba or the Tafassasset-T’immersoi basin. That the Tagedufat system does not enter the River of Agades over the Tiggedi cliff at some point near Ihrayen is probable owing to the fact that all this country has been subjected to a slight southerly tilt. The Tiggedi cliff, the Eghalgawen-T’in Wana massif, the cliff east of Akaraq and its continuation along the great valley, finally represented by the ridge of Abadarjan, as Barth rightly judged, are the northern boundary of this area, which slopes gently from north to south. The River of Agades receives hardly any left-bank tributaries.

Milen well could never be found without a guide. The wide valley, with sand dunes on the south side and a steep north bank where the now omnipresent rock of Air appears, is bare, dry and stony. It shimmers in the heat. Teskokrit appears as a black mass in the west on a bank of milk-white mirage set round a group of trees. The bottom of the valley is a gravel plain with a small patch of bare rock in it which an unwitting traveller would most probably pass unheeding. In this patch of rock is a small hole with a large circular stone near by. The hole, barely three feet across, is the mouth of a well driven through hard sandstone all the way down to the water-bearing stratum, seventy feet below the ground. The mouth can scarcely be seen fifty yards away. The rounded stone is several inches thick and was said to have been used to cover up the mouth of the well to prevent its becoming silted up with driving sand.

I came there in June, after more than forty hours’ march from Hannekar with four tired camels and two men, an Ifadeyen guide and an Arab of Ghat in the Fezzan. We had very little water left, so little, in fact, that it was all used in one pot to cook some rice for us three. The place was deserted and very lonely. The wind was driving the sand so hard that it stung the naked calves of my legs as I stood at the well with Ishnegga the guide, drawing water for the thirsty camels. Camels in hot weather drink a great deal, and hauling water in a two-gallon leather bucket from a seventy-foot well is hard work in a temperature of over 150° F. in the sun. The camels drank interminably. The last and best camel was still thirsty and remained to be watered. The beast was rather weak. It had a bad saddle sore, a hole about the size of a large man’s hand, in its back, and it was festering and full of maggots. We had all just done a journey of over 500 miles from Tanut to Termit and back, in thirty-five days, including nine days of halts, averaging, in other words, nearly twenty miles per marching day for twenty-six days. The camel had begun to drink. Then as we were drawing a full bucket the well rope broke six feet from my hand and fell to the bottom of the well with a splash. A vain hour was spent, while the rice cooked and got more and more full of sand, trying to fish up the rope and bucket with an iron hook made of the nose-piece of a camel bridle fastened to a knotted baggage rope. This too was lost after hooking the tangle, which it joined at the bottom of the well. Prospects looked gloomy as our thirst increased. I have distinct recollections of the sky and valley getting whiter and more metallic and the heat more intolerable. Finally, just enough rope was found by untying all the baggage to ladle up water a half-gallon at a time in a small canvas bucket. But the poor camel had to wait a long time to finish its drink, for the first of the supply to reach the top was used to refill the tanks.

As I was leaving the well two men with three camels came in from the south. They had started to return to their own country in the hills, after an enforced sojourn in the neighbourhood of the fort at Tanut on account of their rebellious propensities in 1917 and 1918. They had no possessions but three young camels, and had started with only enough water in one small skin for half their journey. The two men reached Milen, having drunk nothing for twenty-four hours. They were rather exhausted, but had fully expected to have to do another ten or twelve hours’ march the same night to the nearest water at T’in Wana, as they had only a calabash bottle and no rope with which to draw any more water. They had risked death sooner than stay a moment longer than was necessary in the south, even to collect enough well rope or equipment for a journey which most Europeans would consider difficult. It was very pleasant to give these two men, an old noble and his serf, some especially good cold water from a small canvas cooler which I had prepared. When the serf carried away a pan of icy water, he first offered it to his master, who drank it.

The second time I came to Milen was in December. There was such a crowd of people and of flocks belonging to the Ifadeyen watering that the supply was practically exhausted, and it took me five hours to get enough water for the return journey to Hannekar. But in June the camping grounds were deserted, for there was hardly any pasture during those last few days before the rains.

The deep wells of Azawagh fall into two categories. The narrow wells, like Milen, Aouror, higher up the Milen valley, and Maisumo, are intended primarily for watering flocks. Their output is copious but slow, and not unlimited. Not more than two buckets can draw water comfortably at the same time: for watering flocks where time is not important and the animals can be brought in from pasture in small batches, these wells are adequate. Tagedufat, like Tergulawen, on the other hand, was a caravan well; it was broad and capable of watering a whole caravan rapidly. It became silted up with drifting sand, like the pasture wells, Anu n’Banka and Gharus n’Zurru. Of Aghmat, Tateus and Taberghit I have no details, but when Barth passed this way no stop was made at either of the first two, which were on his road. The supposition is that unless these wells were dug since his day, which is not likely, as the population of Azawagh had by then already decreased, they also were intended for pastoral purposes. They are now all silted up.

The theory that the wells of Azawagh were made by the Ifadeyen, who have only recently come into this area for winter pasturage, was advanced to me, but my informant, who joined my caravan as an unbidden but welcome guest at Milen on my way south, was himself a member of this tribe, so the information is prejudiced. The wells are certainly very old and are probably the handiwork of the denser population which cultivated millet and had its permanent villages in the Taberghit and Tagedufat valleys. The pasture wells were regarded as the property of the tribe in the area, and now, therefore, of the Ifadeyen. The big caravan wells were under the tutelage of the keepers of the great highway to the south, the Kel Owi confederation, and before them, therefore, of their predecessors in Eastern Air. These big wells were always considered to be free for passing caravans to use without let or hindrance at any time, except in the event of a feud being in progress between the Kel Owi and the owners of the caravan. Caravans, on the other hand, using pasture wells, could only do so with the permission of the tribe pasturing in the area. The latter, conversely, had no rights over the great wells. The maintenance of these rights is the origin of confederations like the Kel Owi, for the freedom of the great wells is a vital necessity to a society of caravaneers, and has to be retained by force if necessary. It accounts for such raids as those conducted by Belkho on Gamram, where the Isherifan had interfered with passing caravans just once too often.

One of the Azawagh wells, Aouror, has been the object of much dispute among the Tuareg: there are inscriptions on a neighbouring rock recording the ownership and, to some extent, the history of the well. It would be attractive to think that “Aouror” meant the “Well of the Dawn.” It is not impossible, since Arorá or Aghorá[70] means “dawn” in Temajegh, and Aouror is almost the easternmost well of Azawagh. Like Milen, it is driven through the rock, but is only some four fathoms deep. Like Milen, too, its sides are scored by rope-marks which in places have cut deep into the hard sandstone. Wet ropes covered with sand of course cut into rock quite rapidly, but even so the antiquity of these wells must be considerable. The rock cutting, which no Tuareg to-day is capable of executing, is perfect; the walls are perpendicular and smooth; the plan is a perfect circle.

Abellama and Aderbissinat in the west of Azawagh are deep caravan wells with good water; the former is in friable soil, and has a tendency to fall in.[71] These two, with Aouror, Maisumo and Milen, are the only live wells in Azawagh to-day.

After a short gentle slope up, the ground descends from the ridge on the north side of the Milen valley in a series of long terraces to a basin, the lower part of which is known as the Eghalgawen valley. It joins either the River of Agades at the south-west corner of the T’in Wana massif, or turns south-west towards the Milen and Tagedufat basin; my own impression, based on native sources which are not wholly reliable, inclines to the first view. East of the watering-point of Eghalgawen, the valley runs in a fold, into which flows one of the Southern Air valleys. The actual stream bed is wide and well marked by the heavy annual flood which it carries away from the hills of Eghalgawen and T’in Wana. In character the lower part of the valley along the foot of the hills, with its short tributaries from this little massif, belongs to the Air plateau, and not to Azawagh. The vegetation in the bed is dense and heavy. Dûm palms (_Cucifera thebaica_) and large trees appear. Geographically and geologically the Air plateau has already commenced at the rocks of Tagedufat: actually, however, it is not reached till the River of Agades is crossed, for Eghalgawen is still held to be in Azawagh.

PLATE 6

The cliff of Tiggedi, with its continuation eastward for some way beyond the Eghalgawen hills, is the southern shore of a wide valley which serves as a catchment for all the waters of Southern Air that do not escape by the south-east corner of the plateau into the Azawagh valleys previously described. The cliff is a geological phenomenon of great interest. At the point where the Abellama road descends into the valley some forty miles south of Agades the cliff is sheer for a height of over 200 feet. The path down from the general level of the desert to the dry alluvial plain, which forms the bottom of the River of Agades, is steep and rough. Standing at the top and looking east and west, it seems like a cliff on the sea-shore broken by capes and small inlets; the illusion of maritime action is remarkable. Westwards at Marandet, though still a definite feature of the area, it is less abrupt; erosion has broken down the precipice, while the Marandet torrent has eaten away a ravine leading even more gradually up to the level of the desert. Eastwards, on the other hand, the cliff continues unbroken as far as the Eghalgawen and T’in Wana massif, where higher hills above the desert level take the place of the cliff itself. Though they form a salient in the line, their abrupt northern slopes continue the eastward trend until they come to an end near Akaraq, where the cliff reappears. Here again it is absolutely sheer, if somewhat less elevated; it is broken by a narrow inlet where the Akaraq valley, the only tributary[72] of any size on the south bank of the River of Agades, enters the main basin. At this point the cliff assumes the most fantastic form. The sandstone has been shaped by erosion into pinnacles and blocks of the strangest shapes. The Akaraq valley itself runs back like a cove in a cliffbound sea-coast; both banks are nearly vertical, decreasing in height as the level of the bottom gradually rises to the desert, where the bare rock has been deeply cut into by the water, lying in a semi-permanent pool in a very narrow gully. The bottom of the inlet is covered with luxuriant pasture and some fair-sized trees, while at the mouth, in the main valley, stands an island of rock with vertical sides to complete the illusion of a sea-coast.[73] From the top of the cliff you may look across the great broad valley toward the mountains of Air that are scarcely visible in the north. No defined bank appears to limit the far slope of the basin. There is deep green Alwat pasture[74] in the nearer distance, merging imperceptibly into yellow grass and bare sand further away. The blazing glare and shimmering heat wash the feet of the cliff where a shelving beach of loose white sand has been thrown up against the rocks. The plateau at the top of the cliff is quite flat, and covered with a layer of small hard gravel over the rock. It is without any vegetation.

The great valley bears several names. At the Akaraq inlet it is called Tezorigi. Opposite the Eghalgawen massif it is the T’in Dawin, and further west the Araten valley. West again it has no name, but where it finally leaves the mountains of Air for the Assawas swamp on the way to the T’immersoi basin, the natives call it the Ighazar n’Agades, or River of Agades, from the city which stands on its northern shore, and this is the name I have adopted for the whole. How far the cliffs extend eastward I do not know. A great fork in the valley is visible from Akaraq, the channel is divided by a bluff promontory, but the cliff continues along the southern bank of the southern branch until it is lost from sight. The ridge of Abadarjan which Barth crossed north of Tergulawen, I expect, is part of the same formation.

PLATE 7

Maritime action is highly improbable as the origin of the cliff. No traces of shells or beaches at different levels, to be accounted for by a receding sea, have been noticed. The supposition that all the Sahara was once a sea-bed is untenable, and in any case maritime action would hardly be limited to a few small areas such as this one. It seems easier to look for another explanation. The cliff and the Eghalgawen massif are a sandstone formation, but the Taruaji mountains of Air opposite the little Eghalgawen-T’in Wana massif are granitic. The cliff represents, I hazard, a fault north of which the igneous formation of the Air plateau has been extruded. The ground to the south slopes gradually away from the edge of the cliff, accounting for the virtual absence of any tributaries on the left bank of the River of Agades. There is apparently no igneous rock south of the basin, there is very little else to the north of it, with the exception of some Archean and very early rock. The fault, occasioned by the volcanic action which formed the massif of Central Air, erected a barrier to the southward drainage of the mountains, and the waters of Southern Air were diverted westward. A larger rainfall than now caused the gradual silting up of the area between the bottom of the fault and the southern part of the mountains. As the ground level rose and became an alluvial plain from which practically only Mount Gadé and the island off Akaraq emerge, the rain floods began to wash along the cliff and eroded the sandstone into the fantastic forms which are now seen. Wind-borne sand from the eastern desert completed the process of shaping the rocks. The accretion of alluvium diminished with a decreasing rainfall in Air, and the surface deposit of wind-borne sand formed what is now in dry weather a hard gravel-covered plain which, in the rainy season, turns into mud-flats and becomes almost impassable. The water flows aimlessly in the alluvium along deep-cut gullies with vertical sides that constantly change their course. The alluvial origin of the plain of the River of Agades is unmistakable.

[Footnote 44: That is, “The People of Katsina.”]

[Footnote 45: Chudeau has called this transitional area the Sahel Zone, but the name is borrowed from the north and does not seem to be used in the latitudes under discussion: cf. _Le Sahara Soudanais_, passim.]

[Footnote 46: Now called the “Colonie du Niger-Tchad.”]

[Footnote 47: The natives pronounce the name Tasawa, but “Tessawa” is consecrated by European usage since Barth’s day.]

[Footnote 48: The plural of “Akri” in Temajegh.]

[Footnote 49: Wrongly spelt Gum_rek_ by Barth, _op. cit._, Vol. I. chap. xxi.]

[Footnote 50: Jean: _Les Touareg du Sud-Est_, p. 15.]

[Footnote 51: Barth, _op. cit._, Vol. I. p. 36.]

[Footnote 52: _Ibid._, Vol. V. p. 554.]

[Footnote 53: _Ibid._, Vol. I. p. 529.]

[Footnote 54: _Vide_ Duveyrier, _op. cit._, pp. 328 and 359, _et infra_, Chap. XI.]

[Footnote 55: On the French 1/1,000,000 map. Cf. Appendix VII.]

[Footnote 56: Barth, _op. cit._, Vol. I. pp. 521-2.]

[Footnote 57: _Documents de la Mission Foureau-Lamy_, Fasc. II. p. 206.]

[Footnote 58: There are other small wells in the immediate vicinity of Milen: cf. _infra_.]

[Footnote 59: Anu (plural Unan) means “well” in Temajegh.]

[Footnote 60: Barth, _op. cit._, Vol. I. p. 523.]

[Footnote 61: _Infra_, Chap. X.]

[Footnote 62: _Infra_, Chap. XI.]

[Footnote 63: Cf. V. Cornish: _Waves of Sand and Snow_ (Unwin).]

[Footnote 64: Barth, _op. cit._, Vol. I. p. 523.]

[Footnote 65: _Vide_ Appendix III.]

[Footnote 66: “Gharus” means “deep” in Temajegh, and when thus used of places always signifies a “deep well.” This one, however, was silted up.]

[Footnote 67: Buchanan’s _Out of the World_, pp. 128-30.]

[Footnote 68: Barth, _op. cit._, Vol. I. p. 523.]

[Footnote 69: The indications on the Cortier map that the south-eastern and eastern valleys of the Air massif peter out into the desert in the direction of Termit are certainly inaccurate. Cf. 1/500,000 Carte de l’Air, 2 sheets, Service Géogr. des Col., 1912.]

[Footnote 70: This word is believed to have been borrowed by the Tuareg from the Latin. _Vide infra_, Chap. IX.]

[Footnote 71: The French are lining it with concrete.]

[Footnote 72: Unless, as has been mentioned, the Eghalgawen valley also joins the River of Agades, S.W. of T’in Wana.]

[Footnote 73: A similar island, but considerably larger, has been left isolated in the plain by the erosion of the water in the River of Agades; it is a low conical hill, rather similar in shape to the Tergulawen peak, called Mount Gadé, lying between the T’in Wana hills and Agades.]

[Footnote 74: A fleshy plant, growing about two feet high rather like a veitch, and containing as much moisture.]