CHAPTER III

THE CITY OF AGADES

The Eghalgawen massif contains a number of watering-points. The pool of Eghalgawen is near the junction of a valley sloping down from the hills, the main valley here assuming the name of the watering-point. Abundant water exists all the year round under the sand in the bed near a low rock on the left bank. It has rather taken the place of Tergulawen well as a _point de passage_ for caravans on the Great South Road, and used in the past to be a favourite resort for caravan raiders. The neighbouring hill, like the one at Tergulawen, is a well-known watch-tower in times of trouble, since both of them command the approaches to a strategic point.[75] T’in Wana, Tarrajerat, Tebehic and some pools in the Isagelmas valley on the southern periphery of the Eghalgawen massif, are watering-points for the camels and flocks of the tribes which range over Azawagh, to-day the Ifadeyen. Their winter camping grounds can be seen all the way from Tagedufat to the River of Agades; they are readily distinguishable from the older permanent settlements of the original Kel Azawagh who grew millet in this area. Besides the Ifadeyen, the Kel Giga section of the Kel Tadek use the Eghalgawen hills and Azawagh pastures very considerably after the rains. The Ifoghas of Damergu rarely come so far north, since, having few camels, they lack incentive to seek these superlative desert pastures. Those members of this tribe whom I saw in Azawagh were typical in possessing only donkeys and goats, which of course will eat almost anything.

After a 560-mile excursion to Termit and Elakkos, I rejoined my travelling companions, whom I had forsaken at Tanut, in the little massif on the south side of the River of Agades. They were camped a short day’s march from Milen, at the famous permanent pool in the T’in Wana valley. Of all pools in Africa it is of T’in Wana that I shall keep the pleasantest recollections. I was greeted by a fusillade of welcome and immediately went for a swim in the deep pool that had recently been filled by the rains. The channel cut by the water in the rock was in places fifteen feet deep. The pool had a sandy bottom, with a rock four feet high at one end for a diving platform. A length of twenty yards was clear to swim in, and then came a succession of smaller pools beneath the arches and overhanging sides of red and black rock. The erosion of the sandstone was most remarkable. There were witches’ cauldrons and buttresses and enchanted caves, with deep crannies in the tall vertical sides. In the wide valley above, masses of green bushes and branching palms seemed to make the place a heaven-sent garden of rest in a hot land. We were all very happy, and the camels were improving fast. Our men were delighted to see the mountains of Air again. My guide from the south, Ishnegga, who was of the Ifadeyen, found relations in a neighbouring valley. There were acquaintances on the road to gossip with and discuss. Poor Ishnegga shot himself accidentally some months later, as I heard from his beautiful old mother, whom I had met at Hannekar and saw for a second time on my way home.

The sides of the T’in Wana ravine were covered with T’ifinagh inscriptions relating to the tribes that had pastured here in their time; they recorded the names of people, messages to and from their friends, and the professions of love of their men and women. The low hills behind were rough and without vegetation or soil; but some mountain sheep, gazelle and sand-grouse subsisted on the coarse grass in the ravines. The sandstone of the massif seemed to have been subjected to volcanic heat. A deposit of fossil trees among the rocks and boulders was found: a specimen piece picked up near Akaraq a few miles north-east had probably been brought from this deposit near T’in Wana. It was identified on my return as a Tertiary conifer, but the siliceous replacement had been too complete to permit of more detailed examination, except by microscope.

A very pleasant camp was eventually broken, and Tebehic, on the north-west side of the hills, with two watering-places, was reached after crossing the Isagelmas valley, a collector for several small rivulets draining the western side of the hills. In spite of an attack of malaria, which overcame me, Tebehic proved most interesting, for I made friends with a family of Ifadeyen who were camping there during the rains. The man had some cows and supplied me with fresh milk, a great luxury after camel’s milk and the condensed sort out of a tin. He was a widower with several children, and quite charming. One of the children was suffering from a severe abscess in the right ear. It had been “treated” by blocking the orifice with a paste made of fresh camel dung and wood ash mixed with pounded leaf of the pungent Abisgi (_Capparis sodata_) bush. I suppose the mixture was intended to act like a mustard poultice, but the discharge from the abscess being unable to escape had been causing the child acute pain, which it was easy to relieve by clearing out the mess and washing the ear. The abscess having previously opened of its own accord, the pain ceased almost as soon as the “remedy” had been removed. It was the first of my “cures” as a doctor among the Tuareg, and laid the foundations of a great reputation!

PLATE 8

[Illustration: TIN WANA POOL]

[Illustration: ROCK OF THE TWO SLAVES, AT THE JUNCTION OF THE TIN WANA AND EGHALGAWEN VALLEYS]

After a few days at Tebehic we proceeded to cross the broad plain of the River of Agades, whither one of my companions had preceded me. Memories of that plain are unpleasant. A day’s march from the shelter of the Tebehic valley we were overtaken by a violent thunderstorm right out in the open just south of T’in Taboraq. As a convalescent cure for malaria, designed to make any reputable European doctor shudder, I recommend getting up after three days in bed, marching six hours on a camel in the sun, and then spending two more holding up a tent in company with four other men in an eighty-mile-an-hour storm with a rainfall of three-quarters of an inch in about half an hour. The exertions of five of us were successful in keeping the tent up and the baggage dry, but proved tiring. As soon as the wind was over the five human tentpoles were turned on to canalisation, which soon became necessary to drain away the deluge. When this passed, a search over the countryside had to be instituted for articles of equipment carried away by the storm. The camp stove, an unwieldy cube of sheet-iron some fifteen inches each way, and weighing nine pounds, was found 3000 yards from the camp. But the storm had been magnificent. It had commenced at about 3 p.m. as a black cloud hanging over the Air mountains in the north. The wind, before it acquired full force, bore along a cloud of orange sand gleaming in the sun, which was still uncovered by the blue-black storm above. Suddenly everything seemed to be going on at once, sunshine, sand-storm, wind, purple squalls and a white uniformity of tearing, sweeping rain. By six o’clock it was all over. The sun set in a pale yellow sky behind the T’in Wana range. The northern hills grew slate-coloured and then black, and the storm went rolling on into Damergu, illuminating the night with lightning. Hitherto my worst experience of rain had been at Guliski in Damergu, when myself, three natives and our baggage lay in a hut nine feet in diameter; it rained all night, and slowly flooded us out. One felt the water rising among the blankets in an atmosphere of damp clamminess and native humanity. Then had come a hopeless dawn, but the air soon dried everything. Yet I had still to learn what storms in the mountains could be like.

The north side of the River of Agades opposite Tebehic has no definite bank. The mountains of Air slope gradually down to the valley; they are intersected by larger and smaller valleys, forming a series of roughly parallel right bank tributaries all in close proximity to one another. The widest of them are the Azanzara, Tureyet, Amidera, Teghazar and Telwa, most of which start north of the Taruaji mountains—the Tureyet and Telwa, in fact, have their head-waters in the Bagezan and Todra groups in Central Air. Some small villages lie among the foothills by these valleys, but it is dull country. A few small ill-grown trees and a little grass are all the vegetation on the succession of gravel patches which constitute the plain. The sight of the mountains of Air in front makes one want to hurry on.

South of one of these villages the opening tragedy of the 1917 revolution took place. A platoon of French Camel Corps, after completing their duties as escort to the Bilma salt caravan, had supervised the dispersal of the camels in their various tribal groups at Tabello, east of Bagezan, and were returning to Agades for a rest. They had been away perhaps a month and now were within a day’s march of the city. They knew nought of what had happened in Air, suspected absolutely nothing of the unfriendly disposition of the Tuareg. Near T’in Taboraq a large force of Tuareg, which had been lying in ambush behind a little hill on the northern edge of the plain, fell on the column as it was beginning its last day’s march into Agades post. A running fight ensued, in the course of which nearly the whole platoon of Camel Corps were destroyed. One officer, who was returning to France on leave, escaped southward, and a few wounded Senegalese “tirailleurs” found their way with difficulty into the fort at Agades, which had been attacked early one morning a day or two before while the garrison was out on parade. The revolution had been prepared for some time, with the connivance of the Sultan of Agades, by a Tuareg noble named Kaossen, an inveterate enemy of the French since 1900. The outbreak had been proposed by Kaossen and aided by the Senussiya and hostile elements in the Fezzan and Tripolitania as part of the anti-French and -British activities which continued in North Africa throughout the European war. The development in Air, however, came as a surprise to the French. All the Tuareg in the plateau rose, and although the garrison at Agades held out for over three months in doubt and in complete isolation, the revolt spread into Damergu and fears were even entertained for the safety of Northern Nigeria. The defence of Agades and the arrival of a column from Zinder, acting in conjunction with another column from the Niger, eventually saved the situation. The heroic resistance of the garrison at Agades and the magnificent work of the military organisation of French West Africa, over these huge expanses of country at the end of 1917 and early in 1918, have probably never even been heard of, still less recognised, in England, where events nearer home at a most critical period of the war obscured the issue of “another minor incident in the Sahara.” The column from Zinder, in spite of a severe check on the way, was the largest single body of men ever successfully sent over a desert against a nomadic people. It is my privilege to record in England, I think for the first time, the courage of those gallant French soldiers who indirectly defended Nigeria. Their efforts in Air saved a British colony from facing a situation which might have become serious owing to the general depletion of forces there, as elsewhere during those tragic months of the Great War. I am happy to make this acknowledgment, both as tribute to the French soldiers whom I had the pleasure of meeting during the Great War and in 1922, and particularly because even in Nigeria the gravity of the 1917 revolution has never been sufficiently recognised.

My route over the plain of the River of Agades lay in sight of Mount Gadé, a flat-topped hill standing alone to the south. The track used by parties from Sokoto passes this conspicuous landmark after crossing the Azawagh on the way to the rendezvous of the annual salt caravan at Tabello, under the eastern slopes of Bagezan in Central Air. After cutting this track we joined the Agades-Tabello road somewhat west of T’in Taboraq. East of this village the road passes through the other settlements which lie on the southern spurs of the Taruaji massif before it turns north to Tabello.

The track now entered and wound along a valley called the Teghazar.[76] The small torrent bed was very sodden after the rain of the previous day. On either side were low hills of bare gravel with some rock outcrop beginning to appear here and there. On the low scarps or patches of loose stones a few ragged acacias had secured an existence, marking the foot of the Air hills along the River of Agades. Eventually the track rose up to the level of the highest undulations and we came in sight of Agades. Almost simultaneously two Tuareg on camels appeared on the road. They had been sent out from Agades with an accumulation of letters, months overdue, and a message to say that we were the guests of the French officers in the fort, about a mile north of the city. As the last fold of ground was crossed, by a steep bluff where Kaossen had constructed a military work during the siege of the French garrison in 1917, the whole length of the city came in sight on a low ridge to the south-west. The far end was marked by the stately tower of the Great Mosque, unchanged since Barth saw it more than seventy years ago. Straight ahead lay the French post, surrounded by a defensive wall flanked by blockhouses and containing the tall masts of a wireless station, near the wells of T’in Shaman in a diminutive plain where the Foureau-Lamy Expedition had camped over twenty years earlier. In 1917 there was no W/T station and scarcely any fort; the buildings were all disconnected and scarcely defensible.

PLATE 9

[Illustration: AGADES]

The city of Agades used to be surrounded by mud walls, intended to baffle raiders rather than to withstand a siege. The distance along the line of their elliptical circumference, so far as it can still be traced, is a matter of three and a half to four miles. The wall has been much broken down and in some places is hard to find; its perimeter and plan seem to have varied from time to time according to the number of inhabitants. The best preserved parts are to the north-west beyond the Great Mosque, and to the north, where gates may be seen; there has evidently been considerable decay even since 1850.[77] At a distance the whole ridge on which the city stands appears covered with low, earth-coloured houses, for the most part without an upper storey. The regular sky-line is scarcely broken save by a few dûm palms and tenuous trees rising above the uniform level of the roofs. Only the tower of the mosque, like a finger pointing up to heaven, soars over the drab habitations. Their dull uniformity seems to enhance its dignity.

Agades is not a Tuareg city. Its foreign aspect is at once apparent. Although it also struck Barth immediately, he was, curiously enough, not so much concerned with what is really the most obvious feature of the alien atmosphere as he was with the foreign language and origin of most of the people he met there. His wanderings perhaps brought him less into contact with the permanent settlements of the Tuareg in Air than my good fortune did me; he could not otherwise have failed to remark that the houses in Agades are those of a Sudanese town and not those of the People of the Veil.

The most striking characteristic of the towns of the Sudan, of the immense walled cities of Kano and Zaria, as well as of the smaller places, is the mode of construction of the dwellings. There are two types of houses and in neither of them is stone used. The first type is the circular hut with a low vertical wall carrying a conical roof; the fashion extends throughout Central Africa. This abode is constructed of straw, or grass, or boughs, or of whatever material is readiest to hand. The ground plan is circular unless specific conditions have exerted a contrary influence, which occurs rather seldom. In the more advanced settlements of this sort in Northern Nigeria a development of the primitive form has taken place: it is a much larger structure with vertical mud walls which support the conical thatched roof, sometimes as much as twenty feet in diameter, standing within a compound. In many North Nigerian villages the dwellings consist exclusively of groups of such huts surrounded by low walls or enclosures.

The second type of house in the large towns of the Sudan is many-roomed and formless. The whole building, including the roof, is made of mud and often has one or more stories. The flat roof of mud and laths is carried on rafters of dûm palm wood which is one of the only available trees that resists the invasion of the white ant. Houses of this type often cover a considerable area, rambling aimlessly hither and thither in rooms, courts and alley-ways, according to the requirements and fancy of the owner or his descendants. The mud construction at times displays architectural features of real distinction. The thick tapering walls are wide and smooth. The doorways have a pylon-like appearance reminiscent of Egypt. The heavy squat façades are by no means unimposing: deep cold shadows cast by angles and buttresses break up the surface of the red walls. The broad panels around the doors are sometimes elaborated with decorative mouldings or with free arabesque designs in relief. The larger rooms which cannot be spanned by one length of rafter are vaulted inside with a false arch of mud, concealing cantilever timbering; the effect is that of a series of massive Gothic arches, plain but often of noble proportions. Technically, mud construction is easy, inexpensive and adequate in a climate where the rainy season is short and well defined. Balls of mud are dried in the sun and cemented together with wet mud. The outer and inner walls are faced with a plaster of earth and chopped straw. In the hot tropical sun the walls dry as hard as stone. The houses survive for an unlimited period of time if the outside surfaces are refaced every year after the torrential rains have washed away the stucco skin. Roofs, of course, have to be carefully levelled and drained to prevent the water accumulating in puddles and, in time, soaking through the ceilings. Gutters are provided with spouts projecting through the parapets of the roofs to prevent the water running down the sides.

The rambling mud house and the circular mud or straw and thatch huts, grouped in compounds, together make up the towns and villages of Northern Nigeria. The two types may be seen side by side, for instance in the country between Kano and Katsina, where the Fulani and Hausa population is mixed. It would be interesting to establish, as _prima facie_ seems to be the case, whether the circular houses were those of the sedentary Fulani, who are nearer the semi-nomadic state, and the more ambitious mud dwellings those of the Hausa. In neither of these two types of house is stone used, either as ashlar or as rough masonry. Nor do dry stone walls occur, for mud is more convenient even when stone is available.[78]

When the Tuareg, on the other hand, builds permanent or semi-permanent dwellings, he displays characteristics which at once differentiate him from the people of the south. His straw and matting huts are not of the Central African type; they have no vertical wall of reeds or grass and a separate conical roof; they are built in one piece as a parabolic dome. Another, movable, type of hut or tent consists of a leather roof arched over four vertical uprights surrounded by matting walls on a square plan. The appearance of these tents is that of a cube with a slightly domed top. The permanent houses in Air are regular, carefully built constructions of stone and cement. In them mud is not employed except where the fashion of the south has been directly copied in comparatively recent times. The rambling house plan of the south is almost unknown. The Tuareg dwelling has a definitely formal and rectangular character. It rarely consists of more than two rooms. Even the exceptions to this rule[79] display considerable differences from the southern type of house.

Both the temporary huts and the permanent dwellings of the People of the Veil, therefore, are intensely individual. They differentiate the Tuareg sharply from the southern peoples. But even a casual glance at the houses of Agades makes it obvious that they belong to a city of the south. There is plenty of stone all round the city which might have been used for building, yet nearly all the houses are rambling mud constructions like those of Kano or of any of the towns of Nigeria. The number of houses at Agades which reflect the formal Tuareg fashion of planning is small. The characteristics which one learns to associate with the truly Tuareg houses of Air are conspicuously difficult to find. When I was in Agades at the commencement of the rains before the annual refacing of the walls had been carried out, it was possible to observe the absence of stone building. An inspection of the broken walls of the many ruined houses confirmed this observation of the past. The number of pools in the town alone was evidence of the prevalence and antiquity of mud construction; Barth mentions the names of several of them. The borrow pits in the Sudanese towns, where water accumulates in the rainy season and rubbish is shot in the dry, are features which no one can escape, were it only on account of the smells which they exhale; for in the Sudan, even when stone is available as at Kano, it is not used. I have vivid recollections of Agades at this season and was particularly impressed by the efficiency of the spouts designed to carry the water off the roofs. Progress was necessarily circuitous in order to avoid drowning in the flooded holes and borrow pits, while distraction was afforded by a determined but usually unsuccessful effort to escape a series of shower-baths in the narrow streets.

The ridge on which the city stands is surrounded by several depressions where are the wells that supply the needs of the population. In addition to those outside the town there were formerly nine other wells within the walls, but, like the pools, they were nearly all adulterated by the saline impregnation of the ground.

I cannot here refrain from quoting Barth, whose capacity for meticulous observation depended on never missing an opportunity, however strange, of acquiring information. “The houses of Agades do not possess all the convenience which one would expect to find in houses in the north of Europe; but here, as in many Italian towns, the principle of _da per tutto_, which astonished Goethe so much at Rivoli on the Lago di Garda, is in full force, being greatly assisted by the many ruined houses which are to be found in every quarter of the town. But the free nomadic inhabitant of the wilderness does not like this custom, and rather chooses to retreat into the open spots outside the town. The insecurity of the country and the feuds generally raging oblige them still to congregate, even on such occasions. When they reach some conspicuous tree the spears are all stuck into the ground, and the party separates behind the bushes; after which they again meet under the tree, and return in solemn procession to the town. By making such little excursions I became acquainted with the shallow depressions which surround Agades. . . .” He then proceeds to enumerate them.[80] The plain where the French fort lies is called Tagurast, that to the S.W., Mermeru; to the S.E. is Ameluli, with Tisak n’Talle somewhat further away to the S.S.E.; Tara Bere lies to the west.

The city is divided into several quarters, the names of which are recorded on Barth’s plan. The only two I heard mentioned were Terjeman and Katanga, the former so called from the interpreters who used to live in the neighbourhood, the latter from the market where what Americans would term “dry goods” of the Air fashion are sold. Little seems to have changed in seventy-five years; necklaces, stone arm-rings, wooden spoons and cotton cloth can be bought, now as then. In the larger market near by, called by the Hausa name of Kaswa n’Rakumi (the Camel Market), live-stock of all sorts is sold. The vegetable market seems to be as ill furnished now as it was in 1850.

I visited two or three private houses. They were not imposing, lacking the architectural features of the better-class houses in the Sudan. The use of white and colour washes in the interiors and on the outside walls was interesting. This practice is the only feature in which the houses of Agades differed from those of the Sudan; it appears to be peculiar in this part of Africa to the Tuareg, the habit having, no doubt, been copied from the north. The pigment is made of a chalky substance found near Agades, or of ochreous earths occurring in various places in Air. One of the houses which I saw was that of the Añastafidet, the administrative head of the Kel Owi tribes. The rooms were small and ill-planned; there was no attempt at decoration. The technique of the south had evidently not flourished in the atmosphere of the Sahara. The two plans of private houses reproduced by Barth give an idea of the rambling and haphazard designing.

The most elaborate and well-kept house is the one which belongs to the Kadhi, near the Great Mosque. It must have been here that Barth attended several sittings of the Kadhi’s Court, adjudicating on inter-tribal matters which could not be settled by the tribal chiefs. It did not seem at all remarkable after the great houses of the Sudan, but was perhaps rather better kept than most of the other buildings in Agades. The people call it the House of Kaossen, and his family still live there. He carried on his intrigues from this place, and plotted with apparent impunity through 1917, until the time was ripe for open rebellion. He had returned from the Fezzan full of ambition to free his country from the white men whom he fought all his life. He had taken part in the operations against the French in Equatorial Africa, largely directed by the Senussiya from their “zawias” in Tibesti and Ennedi. When this period of hostility came to an end, but not before the French had sustained several severe reverses, notably during the fighting at Bir Alali (Fort Pradie), north-east of Lake Chad, Kaossen took refuge with the Azger Tuareg in the Eastern Fezzan, raiding and fighting with these lawless folk against their neighbours. Of his own initiative, but aided by the Senussiya and their Turkish and German advisers, Kaossen returned to his native country in 1917 with a small band of supporters to drive out the French, an effort in which he very nearly succeeded.

By far the most considerable monument of the city is the Great Mosque. I was unable to visit the interior, but from the general appearance of the building I am sure that I should have agreed with the description of Barth, who wrote: “The lowness of the structure had surprised me from without, but I was still more astonished when I entered the interior and saw that it consisted of low narrow naves divided by pillars of immense thickness, the reason of which it is not possible at present to understand, as they have nothing to support but a roof of dûm-tree boards, mats and a layer of clay.” He goes on to speculate on the superstructure which these “vaults or cellars” may have been designed to carry but which was never completed. I do not think such speculation is necessary. The description fits accurately every one of the seven or eight other mosques in Air which I saw within and without. In none of them were the walls ever meant to carry an upper storey. In all of them the ceiling was low and the roof flat, with rows of massive pillars and the naves running transversely from north to south across the buildings, which were usually far broader than they were deep.

The Great Mosque of Agades as it stands to-day was built in 1844.[81] It would hardly be remarkable were it not for the minaret, which was rebuilt by the Sultan Abd el Qader in 1847 to replace the one which had fallen. From a base thirty feet square resting on four massive pilasters in the interior of the mosque, this four-sided tower of mud and dûm-palm rafters rises to a height of between eighty and ninety feet, tapering from about one-third of its height to a narrow platform less than eight feet square at the top. Access is obtained by a spiral way between the solid core and the outer wall, which is pierced with small windows. From a little distance the foreshortening produced by the tapering faces gives the impression of immense height without accentuating the pyramidical form. The four-square, flat sides are bound together by transverse rafters projecting some three or four feet. These ends serve the purpose of scaffolding when refacing is necessary after the rains, an operation without which the tower would not have stood any length of time. Near the mosque is a heap of mud, the remains of an older tower called “Sofo,” presumably of the same type.[82]

The structure is properly speaking a minaret, but was used as a watch-tower in time of war. It is not now used for either purpose. The muezzin stands on the roof of the mosque below to call upon the Faithful at the prescribed hours to forsake their pursuits and turn to the only God. The Tower of Agades stands like a beacon, showing far over the monotonous plains. I remember this solitary pillar towering above a confused mass of low and ruinous buildings against the blood-red setting sun, which appeared and disappeared in the black clouds of an evening in the rains. The blue hills and sharp peaks of Air were distant in the north; to the south lay a drab plain, unbroken as far as eye could see in the gathering twilight. The Tower seemed like the lonely monument of a decaying civilisation.

There are said to have been as many as seventy mosques in and near the city, but only two, I think, are still used. Outside the walls to the S.W. there is a shrine known as Sidi Hamada, “My Lord of the Desert,” appropriately named considering the barren nature of the ground all round. It is an open place of prayer of much sanctity, and reputed to be the oldest Moslem place of worship in the neighbourhood. The Qibla is in a low bank, faced with a dry stone wall, which slopes down to the level of the surrounding ground a few feet on each side of the niche. On certain occasions prayers are said at Sidi Hamada, notably on the Feast of the Sheep, known to the Tuareg as Salla Laja, which I was fortunate enough to witness at Agades in June 1922.

PLATE 10

[Illustration: GATHERING AT SIDI HAMADA]

[Illustration: PRAYERS AT SIDI HAMADA]

It was made the occasion of much festivity. Every available camel in the vicinity was ridden by a Tuareg in the gayest saddle and bridle from the city to the shrine. These people do not feel that they are making the best of themselves unless they are mounted on a camel. A man and his camel are complementary and reciprocal to one another. When there is an occasion to celebrate they wear their best clothes and borrow any ornaments they can find to adorn their sombre garments. They are vain of their personal appearance and covetous of those pretty things which are considered in good taste, but their unselfishness is nevertheless remarkable. I have seen men forgo the real pleasure of wearing a silver ornament or a new face veil in order to lend them to a less fortunate companion whose general appearance was more ragged, or whose means and opportunities did not allow him to secure anything to smarten his turn-out. I had bought of the local jeweller-blacksmith in Agades a number of small silver ornaments of the sort which are affected by the Tuareg. All these, and even certain articles of clothing from our own scanty wardrobes, were borrowed for the day. It was curious to see that their sombre apparel was never lightened by any of the coloured materials so much in evidence in the Sudan. The best-dressed man is considered to be the one with the newest indigo-cotton robe and veil of the traditional plain design. At the most a red cloth is tied round the head over the face veil, or, in the case of the guides employed by the French, around the waist and shoulders: the robe must, however, always be plain white or dark indigo. The Tuareg of our own retinue picked out the best of our camels to ride. They turned out a very smart patrol, the camel men Elattu, Alwali and Mokhammed of noble caste, with two or three buzus or outdoor slaves, and Ali the son of Tama, the Arab from Ghat.

At an early hour the poorer people on foot began to stream over the tufted plain which lies between the place of prayer and the city. They were followed by little parties of men on camels, black figures on great dun-coloured or white riding beasts, girt about with their cross-hilted swords, and some also carrying a spear and oryx-hide shield. Finally, a larger group of men, preceded by three or four horsemen, was seen approaching. They were the Sultan of Agades, Omar, the Slave King of the Tuareg of Air, with his attendants, and the Añastafidet, a noble of the Kel Owi tribes, who, from the purely administrative point of view, is the most important man in the country. They were accompanied by the chief minister of the Sultan, the notables of the place, and other dignitaries. Among them was El Haj Saleh, the father of our camel man Elattu; he had performed the pilgrimage three times, in the course of which he had acquired the Arab fashion of dress used in the north. He wore the white woollen robe that is supposed to be descended from the Roman toga, with his head covered only by a fold of the cloth. El Haj Saleh has lived so long in foreign parts that he no longer veils his face and prefers speaking Arabic, but he is much respected as a learned and holy man; he is now employed by the French at the fort as Oriental Secretary and interpreter. With him were the Kadhi and the Imam, a solitary exception among the veiled Tuareg in the matter of display, for he had obtained from the south a buff-coloured silk robe embroidered with green. The Sariki n’Turawa, or chief minister of the Sultan, came next; near him gathered a number of Arab merchants from Ghat and Tuat in white robes; with one or two from the extreme west, there were a dozen or fifteen in all, who have the trade of Agades in their hands. Among them I perceived one Arab from Mauretania, a little man with delicate, sensitive features and a brown beard. He came straight up to where I was standing to repay me a debt of five silver francs which he had incurred some months before at Gangara in Damergu.

PLATE 11

[Illustration: PRAYERS AT SIDI HAMADA, NEAR AGADES]

When the crowd had collected, the men ranged themselves in rows facing east before the Qibla; the women stood together on one side. The Sultan and his party were immediately opposite the niche with the Imam facing them. He began to read the Quran and the multitude then prayed. On either side of the Sultan, as he knelt to make his prostrations, a Tuareg remained standing with his sword drawn, extended point downwards at arm’s length, in protection and salute. As the Sultan rose to his feet the guard sloped their swords, repeating the salute every time he bowed before the name of God. These two men are distinct from the officials in the local administration;[83] they are the personal body-guard of the Sultan, chosen among the “courtiers of the king,” who are young men selected in turn from the tribes in Air which owe allegiance direct to the Sultan.

After the prayers were over two sheep were slaughtered in the orthodox manner. Their throats were cut by the Imam, reciting the invocation of Islam, and the blood was wiped away with holy water to the accompaniment of suitable prayers.

The Sultan and the people then returned to the city, making a detour by the N.W. side through the ruined suburb outside the walls and past the Great Mosque to the present palace, an indifferent building, both tumbledown and dirty. The reigning Sultan, Omar, like all his predecessors, is of slave descent. He was chosen in 1920 by the tribes which have the right to elect him, from a collateral branch of the ruling family. He is a weak man, and too much in the hands either of interested advisers or of the French, which does not always mean the same thing. His predecessor, Tegama, on the other hand, was a remarkable man. His intrigues with Kaossen were successful in preparing the revolution in Air so quietly that practically nothing was suspected of his intentions until the fateful dawn when the black troops on parade at the post were fired upon from the outskirts of the city. After the French columns had relieved the besieged garrison, both Kaossen and Tegama fled east to Kawar, whence the former found his way to the Fezzan, only to be killed, so it is believed, in obscure circumstances north of Murzuk by some Arabs. The native accounts of the story cast some doubt on his actual death on the grounds that his body was never found among those of his massacred companions. It is further represented that the very Turks and Senussiya whom he had served put him to death for his failure in Air, but it appears more probable that on his way to seek refuge with the Senussiya in Cyrenaica, Kaossen and his friends had the misfortune to fall in with a band of Arabs whom he had raided in the olden days, and to have been killed by them.

The Sultan Tegama, on the other hand, betook himself to Tibesti, hoping to find sanctuary among the Tebu, who, though the hereditary enemies of the Tuareg of Air, were probably sufficiently hostile to the French to be counted on to harbour any prominent refugee from the wrath of the white man. By the influence of the Senussiya in these parts he expected to reach Kufra and so take up his residence among the malcontents who live in that remote land. Treacherous as ever and true to their reputation current all over North Africa, the Tebu entreated Tegama generously and took the first opportunity which presented itself to hand him over to a French camel patrol from Bilma. In the course of time he returned to Agades as a prisoner under an escort of negro Senegalese soldiers and was thrown into prison at the fort to await his trial by court-martial. He died suddenly one night in May 1922, by his own hand it is said, in the prison, while under the surveillance of the French, and he was buried. But one chief who was at the funeral told me that he looked under the mat which covered the alleged corpse and discovered that there was nothing there. The story spread that Tegama escaped and fled to the north, where he is still living. Perhaps it is better that this story should obtain credence than any other. Instead of Tegama, the French officer in charge of the post was court-martialled for the suicide of the king, but acquitted. The whole episode is curious, but the truth is perhaps rather unsavoury. It is another of the fierce tragedies of the Sahara.

Before Tegama, Osman Mikitan and Brahim (Ibrahim Dan Sugi) were Sultans. Mikitan was Sultan when the post was first established at the wells of T’in Shaman, but they changed places several times in the course of the intrigues which took place between the passage of the Foureau-Lamy Expedition in 1899 and the occupation of Air in 1904. In Barth’s day Abd el Qader, son of the Sultan Bakiri (Bekri), was on the throne. His tenure of office was as precarious as that of his successors, for he had been Sultan on a previous occasion before Barth reached Agades, only to be deposed in favour of Hamed el Rufai (Ahmed Rufaiyi), whom he again succeeded; they once more changed places some three years afterwards, Abd el Qader having reigned in all about thirty-two years, Hamed some twelve. The tenure of office of the Sultans of Agades during the last century has been as precarious as it was in Leo’s time, for we read in this authority[84] that the Tuareg “will sometime expel their king and choose another; so that he which pleaseth the inhabitants of the desert best is sure to be king of Agades.” Bello in his history says the same:[85] “whenever a prince displeased them, they dethroned him and appointed a different one.”

The installation of the Sultan with the customs that obtain is in the nature of a ceremonial recognition, by the representatives of the principal tribes of the Tuareg of Air, of his elevation to office. Taken in connection with the traditional mode of his selection, it throws an interesting light on relationships of the various groups of the Tuareg in Air. Barth, who was in Agades on such an occasion, wrote: “The ceremonial was gone through inside the _fada_ (palace); but this was the procedure. First of all Abd el Kader (Qader) was conducted from his private apartments to the public hall: the chiefs of the Itisan (Itesan) and Kel Geres who were in front begged him to sit down upon the _gado_, a sort of couch or divan, made of the leaves of the palm tree . . . similar to the _angarib_ used in Egypt and the lands of the Upper Nile, and covered with mats and carpets. Upon this the Sultan sat down, resting his feet on the ground, not being allowed to put them on the _gado_ and recline in the Oriental style until the Kel Owi had desired him to do so.[86] Such is the ceremony, symbolical of the combined participation of these different tribes in the investiture of their Sultan.”[87] The throne-room in the old palace seems to have been more imposing than any part of the royal dwelling of to-day. The present audience chamber is a low, arched room, with a small daïs or seat at one end near a narrow stairway leading up to three rooms in an upper storey, which is now not in use. These rooms are lighted by small windows looking over the outer court. I wandered at random in and out of the palace except that small part which is still used by Omar himself and his women-folk. The deserted rooms were deep in dust and fallen plaster. The courts were infested with dogs, children and chickens. The palace was far less magnificent and certainly less well kept than many other houses in the city. Even the small house of the Añastafidet, with its mats and solitary carpet of horrid colours on the floor of the guest-chamber, was more cleanly.

The present Sultan enjoys little or no authority; his predecessors, unless they were backed by the more important chiefs in Air, were almost equally powerless, for the position of the Sultan, or Amenokal, as he is called in Temajegh, is curious. It is said in the native tradition that in the early days there was no authority in the land other than that of the chiefs of the various groups of tribes, and these did not in any way acknowledge one another’s authority over affairs which interested the community at large. The groups and single tribes were constantly at war with one another, and there were then 70,000 people in the land, with no common ruler.[88] The more reasonable chiefs recognised that some figure-head at least was necessary, but they could not agree that he should be chosen from any of the principal groups of clans in Air. They therefore sent a deputation to Istambul or Santambul (Constantinople) to the Commander of the Faithful, asking him to appoint a Prince to come and rule over them. The Khalif called together the sons of his wives and offered them all the country from the land of the Aulimmiden in the west to Sokoto in the east (_sic_), and from Tadent in the north to the lands of the Negroes in the south. But Air was so far away that none of the sons of the Khalif was willing to leave the comforts of Stambul. The Embassy was kept waiting for three years. Finally the Commander of the Faithful, weakening before the tears of his legitimate wives, the mothers of his sons, selected the child of a concubine to rule over the Tuareg of the south. The candidate returned with the deputation to Air and from that day to this there are said to have been one hundred rulers in the land. This figure does not, of course, represent the exact number; it is only meant figuratively to indicate a long period of time.

From the original impressions I had received in Air I came to the conclusion that the installation of the first Sultan could be assigned to the beginning of the fifteenth century A.D., or, in other words, to a period prior to the capture of Constantinople by the Moslems. In the course of some research on the subject I discovered that 1420 A.D. had been suggested by one authority on the evidence of tradition, while the Agades Chronicle, independently of all this evidence, had recorded that the first Sultan, Yunis,[89] ascended the throne in 809 A.H., or about 1406 A.D.[90] The important thing in any case is that, if the story of his choice has any historical foundation whatsoever, it must be referred to a period when Christian emperors were still ruling in Constantinople. It is therefore all the more interesting to learn that the first Sultan was called Yunis, which means John, and that the wife of the first Sultan, a noble girl said to have been given to him in marriage by the Kel Ferwan tribe, was called Ibuzahil or Izubahil, a name bearing a curious resemblance to Isabel. It is a fitting name for the companion of John, the man from the distant land.

If a deputation went to the Mediterranean at all, it was natural at this period that it should go to Constantinople, still regarded as the capital of nations, with which no other city in the fifteenth century could compare for civilisation or splendour. But we shall probably never know whether a Byzantine prince came to Air in 1406 A.D. or whether the names and legend of John and Isabel are only coincidence. Yunis is described as the son of Tahanazeta, and I must leave for others to discover Byzantine resemblances to this name. For the name of one of his successors, Aliso, I suggest Louis may have been our equivalent, and regarding the latter’s brother, Amati, who followed, comment is hardly necessary.

Yunis reigned twenty years and was succeeded by Akasani,[91] who was the son of Yunis’s sister. Elsewhere El Haj Ebesan or Abeshan, a son of Yunis, and his son, El Haj Muhammad ben Ebesan, are said to have reigned respectively as second and third Sultans, but this is not substantiated by the Agades Chronicle, which mentions El Haj Ebesan only as the grandfather of the sixteenth Sultan, Yusif, who came to the throne about 1594. From this record there appear to have been some forty rulers, several of whom reigned more than once, but there are certain gaps in the series.[92]

After the very first ruler the reigning family divided into two branches, which keep on reappearing, many of the Sultans of one being deposed by powerful tribes like the Itesan in favour of candidates of the other line. The family of El Guddala or Ghodala figures prominently with several notable rulers like Muhammad Hammad, who was known as the Father of his People. From such records as are available I have tried to recover the genealogy of this stock; but the Agades Chronicle is neither accurate nor complete;[93] although it is almost the only detailed information which we possess for the present. One noteworthy fact accords well with Ibn Batutah’s observations and with certain matriarchal survivals which will be referred to in detail hereafter: there are repeated instances of descent being traced through the female line. Nevertheless, this was not an essential condition. The ruler to this day is elected by the same tribes originally responsible for the elevation of Yunis to the throne: he must be drawn from one of the two branches of the original family, and his heir, subject to due and proper election, is normally considered to be his sister’s son.

Being the son of a concubine or slave, the king, according to the rules of descent of all the Tuareg, was himself of slave caste, nor could he ever achieve the distinction of being ranked among the nobles. As it is the law among the People of the Veil that the child must follow the caste of the mother and not the father, whatever the latter’s claims, only the offspring of a noble Tuareg woman can be noble. In all other matrimonial combinations the child must be a serf or slave. A slight distinction is sometimes drawn if only the mother is inferior, but it has the effect, at the most, of creating a mixed caste, without admitting the possibility of the child becoming a noble. When the problem arose of finding a wife for the first ruler who had been selected by the Khalif, despite the pre-eminence of his sponsor, tradition prevailed, that he was to be given a slave woman for wife. The arrangement had the advantage of perpetuating the status of the original Amenokal, since his children perforce had to continue in the inferior caste. For political reasons certain exceptions seem to have been made, and the Amenokal, though a serf, was also allowed to marry a noble woman, but in that case her children were not eligible. The marriage of John and Isabel—if she came from the noble Kel Ferwan, and not from Constantinople, as I suspect—may be an instance of such political dispensation. The restriction of the choice of the Amenokal to one of the two branches of the original family, and the force of tradition in regard to his descent, have resulted in the apparent paradox that in order to be Sultan of Agades the candidate has to be a slave. These considerations duly influenced the choice of the present Amenokal, Omar.

Insignificant as his power nominally is, and unimportant as the office may practically be, many of the traditional stories which purport to explain the circumstances attending the Sultan’s elevation to the throne are probably fanciful. They may be accepted but still be fictions in the legal sense. Unless or until Byzantine researches can come to our assistance, the logical explanation, if there is one, must be sought. Shorn of romance, what appears most likely to have happened is that the Tuareg of Air at a certain stage were unable to reach any agreement regarding the selection of a head of the State. They were divided up into groups which their piecemeal immigration had accentuated. But the necessities of trade and caravan traffic made it essential for the common weal to have some sovereign or head, even if he were only a nominal ruler, to maintain foreign relations and transact political business on behalf of the inhabitants of Air generally with the Emirates and Empires of the Sudan. Since none of the principal tribes was willing to forgo the privilege of providing the ruler, the expedient was hit upon of appointing a man whose status would never conflict with the authority of the tribal chiefs within the borders of the country, but who could still be delegated to speak for the whole community with the rulers of the Southland. With all the jealousy that exists among the tribes on the question of relative nobility or antiquity, the only people fulfilling the essentials were of servile caste. The choice of such a man was nevertheless possible among the Tuareg, for neither “imghad” nor slaves are despised or regarded as mere animals. This, I think, is the only explanation of the usage which obtains, that whatever may be the caste of the Amenokal’s children, only the servile ones are eligible. Although the family of the Sultan may include noble persons, it is, as a whole, a servile group in both its branches; it seems that Barth is mistaken in regarding the group as noble. The family may, as he says, be called “Sherrifa,” but probably only on account of its reputed origin. It is not considered any the more noble in the Tuareg sense of the word for all that.[94]

This does not exclude the possibility of the Constantinople Embassy being true, but the explanation I have given of the slave kings of Air seems to be sufficient on its own merits and also reasonable. Every factor in the situation points to the care which was taken to eliminate all possible chances of dispute; even the relegation of the choice to one servile family singled out for the purpose would tend to diminish friction. On the whole the procedure may be said to provide a rational if cynical solution of what has always been a difficult problem in all countries.[95] Inasmuch as the explanation also serves to elucidate a number of other problems, it may be said to receive confirmation.

Thus, the principal Minister or Vizir of the Amenokal is the Sariki n’Turawa,[96] a Hausa term meaning the “Chief of the White People.” The White People are the Arab traders from the north, who themselves call this official the “Sheikh el Arab.” His functions are those of Minister for Foreign Affairs:[97] his duties are to regulate the foreign community of Agades and settle all questions of trade with the outside world. Though originally appointed to deal with the Arabs of the north, he came eventually to have more to do with the Southland. He used to collect the duties on merchandise in Agades and accompany the salt caravans to Bilma, a service for which he received an eighth part of an average camel load of salt. After the salt caravan returned, the Sariki n’Turawa proceeded south with the camels returning to Sokoto, and then went on to Kano. The latter part of his journey had already been discontinued in 1850, but he still accompanies the salt caravan as the representative of the Sultan and nominal leader of the enterprise. In addition to these duties involving foreign relations, he is the Amenokal’s chief adviser and “Master of the Interior of the Palace,” with the Songhai name of “Kokoi Geregeri.” He is also known as the “Wakili” or Chief Agent of the king. The reason for the Chief Minister in Agades being also Minister for Foreign Affairs needs no further comment after what has been said of the Sultan himself and his _raison d’être_.

Other officials and courtiers round the Amenokal include the Sariki n’Kaswa, or Chief of the Market Place, who collects the market dues and supervises the prices of commodities. There are, besides, police officials or policemen who are also the executioners, and a number of persons called after the class from whom they are chosen, the “magadeza.” The word seems to be a corruption of “Emagadezi,” meaning People of Agades, but has acquired a more restricted meaning, and is commonly applied to a number of rather fat men who are reputed to be the posterity of the attendants of the first Yunis who came from Constantinople.[98]

By virtue of his own position the Amenokal enjoys very little authority. He is used as an arbitrator and Judge of Appeal. In cases where the disputants are both from the same group of clans their quarrel would normally be referred to the head of their aggregation, except amongst the Kel Owi, over whom the Añastafidet is the administrative authority, or court of the second instance; in minor matters the tribal chief can, of course, decide on his own initiative. But in disputes between persons of different tribes who cannot agree on the finding of the chief of either of their factions, the case may be referred to the Sultan, on whose behalf the Kadhi renders judgment. Such functions as the Sultan performs are executed with the consent of the governed. Although all serious cases might be referred to him in theory, in practice his authority has never run in local tribal affairs. He has a common gaol for criminals, used in the first instance for those of the city, but also for such as cannot be satisfactorily punished under the tribal arrangements of a nomadic and semi-nomadic people. There were cases when chiefs of tribes might be, and were, imprisoned at Agades, but then it was because the power behind the throne had so desired it. The Sultan apparently at one time also had a dungeon with swords and spears fixed upright in the floor upon which criminal malefactors were thrown; but already in 1850 it was rarely used.

It cannot be too carefully emphasised that the rule of the Sultan as the elected head of the State of Agades was founded upon the consent of the governed. He is the figure-head of the community and performs the same useful duties which so many heads of more civilised States undertake. The Tuareg have probably never had occasion to discuss the social contract, and the works of J. S. Mill or Rousseau are not current in Air, but nowhere are these theories of government more meticulously carried into effect or do they assume the practical form which they have often lacked in Europe. With all their aristocratic traditions of caste and breeding, the Tuareg have never favoured an established or hereditary autocracy. The government they prefer seems to be a democratic monarchy. Their king is a slave elected by the representatives of certain, at one time doubtless the most important, tribes; he exists and carries out certain functions because the mass of the people desire it so. Authority is not inherited, and even men of inferior caste may become chieftains. The evolution of society has also inevitably rendered the king dependent for support upon the principal men of the country, and the latter upon the smaller chieftains. Where there is much rivalry or where the ruler is weaker than usual the frequent changes and inconsistency inherent in democratic government ensue. Equally the ascendancy of one man’s personality independently of his position may override the voice of the people, but in the absence of organisation or bureaucracy the conditioning factor is efficiency and competence. Tribal leaders are selected because they can lead; when they cease to lead they are deposed.

The unenviable position of the king and his dependence on the influence of the chiefs seem consequently to have been the same throughout the ages. Leo[99] refers to the practice of deposing one king and electing another from the same family who was more acceptable. Bello on the subject has also already been quoted. We have just seen how often and why Osman Mikitan and Brahim changed places. Barth recounts how in his day Abd el Qader was completely in the hands of the Kel Owi, who were represented by the dominant personality of their paramount chief, Annur. His own tribe was not even, as a matter of fact, among those responsible for the selection of the Sultan, but his personality was such that the Amenokal, at his request, or with his support, felt himself strong enough to imprison three turbulent chiefs of the Itesan who were stirring up the people in Agades in favour of a pretender. Yet the Itesan, a tribe of the southern Kel Geres, are the foremost of the tribes responsible for the Sultan’s very election and his maintenance in power. Without Annur’s support, Abd el Qader was powerless.

[Illustration: OMAR: AMENOKAL OF AIR]

I think that the persistence of tradition shows how essential the method devised for choosing the head of the community was, and is still considered to be among the Air Tuareg. Even to-day the Itesan retain their predominant voice in the election, though they live in the Sudan and are in part within the border of the country administered by the British Government, and though their king is in French territory hundreds of miles away. They were the deciding factor in the election, after the death of Tegama,[100] of Omar from the collateral branch which lives with them.

Only in rare cases was the Amenokal a leader in war. Muhammad Hammad is an instance in point, but it is clear he was an exceptional man. When raids had taken place or were threatening in such a manner as to affect the people of Air indiscriminately, or where individual tribes might not consider themselves sufficiently involved to occasion reprisals, the Sultan used to lead a counter-raid recruited from several clans and provisioned according to his direction from those groups most capable of supplying the needs. In no case could a Sultan lead a raid against an Air tribe, whether in the north or in the south, unless he had definitely thrown in his lot with a local intrigue, which theoretically would, and usually did, entail his eventual deposition. Within Air the Sultan was neutral, or as we should say “constitutional.” He could only take the field against people like the Aulimmiden of the west, or the Tebu of the east, or the Ahaggaren of the north beyond the borders of his country. As a general rule, however, leading in war was the task of tribal chieftains and not of the king.

The Amenokal does not seem to have had a fixed revenue. He lived principally on the presents given to him by the tribes on the occasion of his accession, and more especially by those tribes which owe allegiance directly to himself. He was entitled to collect a tax on foreign merchandise entering the city and a tithe from certain servile tribes in the southern parts of Air.[101] In addition he had certain perquisites in the shape of judicial fines imposed on individuals and tribes, and a revenue from legitimate trading with Bilma during the great salt caravans.

In considering the history of Agades one cannot fail to be struck by the peculiarity of the site.

Elsewhere in North Africa, where any of the great caravan roads pass through areas of fertility which break up the journeys into sections, towns and cities, in some cases of considerable magnitude, have grown up. Where these settlements are near the margin of belts of permanent sedentary inhabitation, they play the part of termini or ports for the trans-desert traffic. They have become markets and the seats of the transport and produce brokers, a development which has its parallels in Arabia and Central Asia. There are many instances in Northern Africa of such terminal points becoming large and important centres: some of the more active of these “ports,” as they may be called, in the north are Sijilmasa, Wargla, Ghadames, Tripoli, Orfella and Benghazi. Corresponding with them at the southern end of the various roads are Timbuctoo, Gao, Sokoto, Katsina and Kano.[102] In addition to that there are also the true Cities of the Desert. They have arisen in places where caravans can call a halt to rest and replenish food supplies, where water is plentiful, and sometimes also, where these requisites are present, at the intersection of important routes. These settlements are like island coaling stations in maritime navigation, but they are not termini; they are particularly interesting ethnologically, for they often mark the ends of stages where the transport of merchandise changes hands. At these points one tribe or race hands over its charge to another group of people. They are thus entrepôts where goods are discharged and reshipped—not markets, but broking centres where the transport contractors and merchants who live at either end of the routes have their agents. A money market often develops, but the local trade is small, for it is confined to the requirements of the place and immediate neighbourhood. At all costs, either by means of a strong local government or by mutual consent, tribes which elsewhere may be at war with one another must be compelled to meet in peace to pursue their lawful occasions. The essentials for the growth of such centres are invariably the presence of water, pasture and, to a lesser extent, food. Where these factors can be obtained at one definite point only, the centre is fixed, whereas if there are several places all more or less equally convenient for the traffic, the settlement has a tendency to move under the influence of political changes. In Tuggurt, Laghuat and Ghat may be found instances where the centre has been unable to shift on account of geographical conditions; but in the Tuat-Tidikelt area the most important town of In Salah has had many rivals, which have prevented it acquiring the same compactness or prominence as, for instance, the city of Ghat. At the latter place a large permanent water supply in an arid country practically limited the choice of sites to one spot. A commercial city of paramount importance, if of no great size, sprang up in the earliest times and continued uninfluenced by political vicissitudes. As an entrepôt of commerce where there was peace at all times among the local population, where feuds and racial hostility were set aside within its precincts, where free trade was the oldest tradition and where an efficient municipal organisation did not seek to extend its influence far beyond the walls, Ghat developed a government similar to that of an autonomous Hanseatic town. Ghat is the most interesting of all the cities of the desert, but the decline of caravan trade has brought ruin to its people and war among the tribes, which no longer have the material incentive of trade to refrain from fighting.

On the eastern of the two central roads across the Sahara there is a stage where one would expect to find a town like Ghat, for to the south on both these routes there is a tract of desert to be crossed before reaching Kawar or Air respectively. But in the Eastern Fezzan the choice of locality was not restricted by geographical and economic considerations, and Murzuk, as the counterpart in modern times of Ghat, has consequently not always been the most important centre of the area. In early classical times Garama, now known as Jerma, some sixty miles to the north of Murzuk, was the capital of the Garamantian kingdom. When Jerma was destroyed by the Arabs in the seventh century, Zuila, probably the Cillala of the Romans, became the capital of the Eastern Fezzan, maintaining its supremacy even after conquest by the Beni Khattab in the tenth century. When in the fourteenth century the Fezzan was overrun by the people of Kanem the capital again moved, this time to Traghen.

Air is the next stage on the road to the Sudan after crossing the desert to the south of Ghat. The requisites of water, pasture and food are found all over this vast oasis; the principal settlement might therefore be presumed to have changed its site under the influence of politics, and in a great measure this has happened, but the largest settlement in the country, the City of Agades, is comparatively modern and appears to owe its existence to political rather than to economic reasons.

Standing on the north side of the valley which is named after it, Agades is in one sense a City of the Desert, since it lies on the edge of a Saharan oasis. In so far as it is a true desert city at all, it is the greatest of them, but, as we shall see, it has not quite the same characteristics as its smaller rivals. Ghat, before the war, was said to number less than 4000 people, but may have attained double this figure at one time; the population of Murzuk was variously estimated at 2800 by Barth and at 6500 by Nachtigal; Ghadames is believed to have a population of about 7000. But Agades in the days of its prosperity must have contained not less than 30,000 inhabitants.[103] By 1850 the population had fallen to about 7000; ten years ago the number was estimated at 10,000. To-day there are not 3000 people in the half-ruined city, but the numbers are again increasing since the efflux of population after the 1917 revolution. These astonishing variations in population are a normal feature of desert cities, even as they are of harbours and seaport towns where the places are entirely dependent on conditions of trade, which is affected by political change; in the Sahara the mode of life of the surrounding nomads makes these fluctuations even more conspicuous.

None of the considerations governing the site of other desert cities applies to Agades. It lies on the southernmost foothills of the Air mountains, and in the history of the country there has never been any danger of invasion except from the south. Some of the Tuareg, it is true, gradually penetrated Air from the north, and pushed south by the progressive occupation of the northern mountains, which the original population may not have been sufficiently interested or numerous to occupy and defend. Small raiding parties can always enter the country, but it is certain that with even inconspicuous opposing forces the success of an invading army approaching Air from any direction except the south is out of the question, owing to the difficulties of moving large bodies of men over the appalling desert which separates the plateau from Ahaggar or the Fezzan. The same conditions obtain in the east, and to a great extent in the west also. On the south only is the position rather different. The steppe desert between Air and Damergu is neither so waterless nor so pastureless nor so deep as to preclude military operations from that direction. In point of fact Air was invaded on at least one occasion from that side with conspicuous success.[104] It is therefore anomalous that the capital of the country should have been located on the fringe of the mountains, where every road is defensible, in possibly the most vulnerable position which could have been chosen.

Nor is the explanation to be found in such economic necessity as has dictated the choice of site in other examples of desert cities. Agades is some distance from the great north-south road which runs, and always has run, east of the Central massif of Air, leaving the country on its way to the Sudan at the water of Eghalgawen or Tergulawen. An alternative route to the Sokoto area branching off the main road in Northern Air and descending by the Talak plain and In Gall passes some distance west of the city. No caravan road suitable for heavily-laden camels passes through Agades for the north, owing to the barrier of the Central massifs, through which the tracks are difficult even for mountain-bred camels. The old pilgrim road from Timbuctoo to Cairo enters the western side of the Air plateau at In Gall or further north, and passes to Iferuan and so to Ghat without touching Agades. Ibn Batutah’s route shows that this was so in his day, as it certainly has been the case since then. Caravans from the south crossing the Eastern Desert for Bilma pass across Azawagh to the eastern fringes of Air without going to Agades, which would involve a detour, as was explained in referring to the importance of the well of Masalet.[105]

While the trade routes of the country do not, therefore, provide an adequate justification for the choice of the site, climatic or geographic conditions have equally little bearing, for there are a number of points in Air where the pasture is good and where there is sufficient water to supply the needs of a large settlement. At Agades, as a matter of fact, the water is indifferent; while the surrounding gravelly plain, like the rest of the valley, is only covered with scanty vegetation, the neighbouring Telwa valley contains some pastures, but they are not abundant, and camels in the service of the local merchants have to be sent to feed as much as three or four days distant.

If the conditions which had led to the growth of a city in Air had been of a purely economic order, it might have been anticipated that it would have occupied the site of Iferuan, the first point south of Ghat where a permanent settlement with plentiful water, pasture and land fit for cultivation was possible. So convenient is the Iferuan valley that caravans, in fact, usually do rest there for long periods to allow both men and animals to recuperate after the difficult stage to the north has been negotiated. Or, again, a city might have stood at the eastern end of the River of Agades at the north end of the stage across the Azawagh, although this position would have been less dictated by necessity than the first alternative, since the steppe desert of the south cannot be compared for hardship with the northern waste. It would nevertheless have been convenient, if somewhat exposed to raiding parties, as a point for the concentration of caravans crossing the Eastern Desert to Bilma, or in other words at the branching of the Salt Road and the north-south route. On its present site, Agades is out of the way for travellers from any direction who may be bound beyond the city. Some other explanation must then be found, and it occurred to me only when I had reached the city itself.

The fact of the matter is that Agades is not the capital of Air at all. As we have seen, the city is not the seat of the central government because there is no real central government, and the King who lives there is not really King at all. Agades is only the seat of an administration set up in the first instance to deal with exterior affairs, and more especially those connected with the Southland. These affairs were in the charge of a figure-head ruler unconnected, except to a very minor degree, with the internal problems of the Tuareg tribes. When this is once grasped, Agades assumes a different position in the perspective of history and it becomes apparent that the site is really suited to the purpose for which it was intended. The place where the city lies is neutral as far as the tribes of Air are concerned; it has easy access to the Sudan yet is removed from the main roads, which are considered the property of certain groups of clans. But it follows that the character of the city must inevitably partake rather of the south than of the Sahara.

Finally, there is the most conclusive evidence of all; during the early part of the Tuareg occupation of Air, there was no city of Agades at all; it fulfilled no need despite the caravan traffic. It was presumably not founded when Ibn Batutah travelled through Air, for he makes no mention of the name; although this is negative evidence, it is valuable in the case of so observant a traveller. By 1515, when Askia conquered the Tuareg of Air, Agades, however, was certainly in existence, since it is on record that he occupied the city for a year, “sitting down north of the town,” possibly at T’in Shaman. Marmol, moreover, is quite definite on the subject, saying that the city was founded 160 years before he wrote, a date which has been reckoned at 1460 A.D.[106] We know that the first Sultans of Air did not live at Agades, but by inference it may be supposed that they soon came to do so, so that the date suggested is probably correct. With the advent of a figure-head king there sprang up a figure-head capital. The story of Agades is the story of its kings: the explanation of both is similar.

What seems to have struck Barth most about Agades was that the people spoke Songhai and not Temajegh; it was, in fact, one of the few places left where the language of the greatest Empire of the Niger still survived. There is reason to believe that most of the Emagadezi are not of Tuareg race at all. The Songhai element is probably preponderant even now, four hundred years after the conquest of Agades by the Songhai king, Muhammad Askia, who planted a colony there. The face veil has been adopted universally, but the physical type of the inhabitants is much more akin to that of people of the south than to that of true Tuareg. The descendants of the Songhai conquerors are coarse, broad-featured people with dark skins and untidy hair, which is an abomination among the noble Tuareg. The same characteristics reappear among the inhabitants of certain points west of Agades on the south-western outskirts of Air, where the Songhai element is also known to have become established and to have survived. The people of Agades are hardly even considered as natives of the country by the rest of the inhabitants of Air. They are not classed as a group, like the inhabitants of other settlements in the mountains. It is rarer to hear the “Kel Agades” mentioned than it is to hear such exotic compositions as “Kel es Sudan” or “Kel Katchena.” The people of Agades are more usually spoken of as the “Emagadezi,” in much the same way as the Kanuri in the Air dialect are called “Izghan” and the Tebu “Ikaradan.”

The family of the Sultan is foreign in appearance. The physiognomy of Abd el Qader, who wore the white face veil usually associated in the north with servile caste, was not, as far as could be seen by Barth, that of a Tuareg. His corpulence was equally a foreign peculiarity, despite which Barth considered him “a man of great worth though devoid of energy.” The personality of the present Sultan, Omar, has already been described; his dark skin and coarse features betray a very mixed ancestry. These peculiarities are not unexpected in a family descended through slave women, who may, of course, be of any race.

The different races and languages of Agades would be interesting to study in greater detail. The name Terjeman, given to one quarter of the town, is evidence in the estimation of its inhabitants of the Babel which has occurred. Temajegh, Hausa, Kanuri, Songhai and Arabic are spoken; even the more exceptional Fulani, Wolof and Tebu are heard, while the advent of the French garrison with its negro troops has introduced further linguistic complications, and will, of course, in time accentuate the Sudanese element in the racial composition, for at no time do the morals of the ladies of Agades appear to have been beyond reproach. The consequences of city life are felt even here in the Sahara. The forwardness of the ladies so moved Barth to indignation that he discoursed at considerable length on the standards of conduct which should be observed by Europeans in these far countries towards native women. He no doubt owed much of his success to the respect in which he held the feelings of the people among whom he travelled. Rather than provoke criticism, he recommends explorers to take their own wives with them. A few pages further on, describing his journey through the Azawagh, he is again referring to advances of the Tuareg women of the Tegama. One appreciates his resentment at these importunities, but is inclined to speculate on the true inwardness of his thoughts. On one occasion at least his artistic feelings rather than his sense of propriety seem to have been offended, for he writes: “It could scarcely be taken as a joke. Some of the women were immensely fat, particularly in the hinder regions, for which the Tawarek have a peculiar and expressive name—‘tebulloden.’”

[Footnote 75: Cf. Barth, _op. cit._, Vol. I. p. 523.]

[Footnote 76: Literally “a small river or torrent” in Temajegh.]

[Footnote 77: Cf. Barth, _op. cit._, Vol. I. p. 454.]

[Footnote 78: This generalisation is not intended to cover exceptional examples of stone construction such as those in Sokoto Province.]

[Footnote 79: For the houses of Air see Chap. VIII, where characteristic plans are given.]

[Footnote 80: Barth, _op. cit._, Vol. I. p. 477.]

[Footnote 81: According to Barth, _op. cit._, Vol. I. p. 451; but the minaret was built in 1847, according to the Agades Chronicle (_Journal of the African Society_, July 1910).]

[Footnote 82: This is the one to which Chudeau (_Missions au Sahara_, Vol. II, _Le Sahara Soudanais_, p. 64) refers as 980 years old according to tradition, presumably basing himself on the same information as Jean, _op. cit._, p. 86. The date is improbable, as Agades was not founded at that time.]

[Footnote 83: Cf. Leo Africanus, Vol. III. Bk. VII. p. 829: “The king of this citie hath alwaies a noble garde about him.” Cf. Plate 11.]

[Footnote 84: Leo Africanus, Vol. III. Bk. VII. p. 829.]

[Footnote 85: Denham and Clapperton, Vol. II. p. 397.]

[Footnote 86: The same procedure is indicated in the Agades Chronicle, which also states that the Kel Owi give him an ox (_Journal of the African Society_, _loc. cit._)]

[Footnote 87: Barth, _op. cit._, Vol. I. p. 422.]

[Footnote 88: Jean, _op. cit._, p. 89.]

[Footnote 89: Isuf or Yusuf according to Jean, who is certainly wrong in this respect. _Op. cit._, p. 89. Chudeau, _op. cit._, p. 70, gives his name as Yunis, as did my informants in Air.]

[Footnote 90: The date of the founding of Agades is a measure of confirmation: _vide infra_, Chap. XI.]

[Footnote 91: The second Sultan is given by Chudeau, _op. cit._, p. 64, as Almubari (El Mubaraki): a ruler of this name succeeded a Yusif whom he deposed in 1601; some confusion has probably arisen on account of Jean’s error in supposing that the first Sultan was called Yusuf instead of Yunis.]

[Footnote 92: See Appendix VI.]

[Footnote 93: See table in Appendix VI.]

[Footnote 94: Cf. Barth, _op. cit._, Vol. I. p. 468.]

[Footnote 95: See also the remarks made in Chap. XII regarding the tribes which elected the Sultan.]

[Footnote 96: For the explanation of the sense which these words have acquired, see second footnote, Barth, _op. cit._, Vol. I. p. 471.]

[Footnote 97: The Tuareg have forestalled many European Powers in making their Prime Minister also Secretary of State for Foreign Affairs.]

[Footnote 98: Jean, _op. cit._, p. 89.]

[Footnote 99: Leo Africanus, Vol. III. Bk. VII. p. 829.]

[Footnote 100: The influence of the Emir of Sokoto to which Barth has referred is exercised through the Itesan by virtue of their domicile near this city. Cf. Barth, _op. cit._, Vol. I. p. 468.]

[Footnote 101: Cf. Leo Africanus, Vol. III. p. 829.]

[Footnote 102: Cf. map in Chap II.]

[Footnote 103: My estimate of 30,000 inhabitants was arrived at locally without any books of reference. On my return I found that Barth had arrived at the same figure, with a possible maximum of 50,000 (_op. cit._, Vol. I. p. 472).]

[Footnote 104: The French operations of 1918 against Air, the occupation of the country from the south in 1904 and the passage of the Foureau-Lamy expedition are not considered, as the superiority of European weapons makes it impossible to compare these exploits with native enterprises, though the success of the first two and the appalling losses in camels and material of the last in a measure confirm the thesis.]

[Footnote 105: _Vide supra_, Chap. II.]

[Footnote 106: By Barth, _op. cit._, Vol. I. p. 459, and by Cooley, _Negroland of the Arabs_, p. 26, as 1438 A.D.]