Chapter II: Part 2

[Footnote 65: Oppenheim, p. 121. [The present Shehu of British Bornu. There is a photograph of him in Macleod, p. 244. Ex-Shehu Omar Sanda, surnamed Kura (the elder) and Kori (the short), is now Ajia of the British district of Gujba.]]

[Footnote 66: [June, 1901. In the person of the late Major A. McClintock, D.S.O., Seaforth Highlanders, formerly commanding the 1st Batt. Northern Nigeria Regt., and until his recent death second senior Political Officer of Bornu Province. His escort was commanded by Sergt. Moman Machena (Auta), now a Preventive Service Agent in Bornu. Cf. Decorse, _Rabah_, &c., p. 18.]]

[Footnote 67: Oppenheim, p. 123; Möckler-Ferryman, p. 142. Fadel Allah was in fact actually recognized as Sultan of Bornu. [This is hardly correct. Sir F. D. Lugard in his _Report on Northern Nigeria for 1901_ mentions that Major McClintock presented Fadel Allah with a shot gun of his own. This led to a report, believed by the French, that the British authorities were supplying Fadel Allah with rifles. Major McClintock formed a high opinion of Fadel Allah, who appears to have been a most gallant soldier and a capable and determined ruler. Major McClintock’s report was wholly in favour of recognizing Fadel Allah as Emir of Bornu. After the French had killed Fadel Allah at Gujba, they demanded from Shehu Garbai an indemnity of $80,000. They managed to wring $73,500 from the exhausted country before the arrival of Col. Morland and a British column. Captain McCarthy Morrogh with a company of the West African Frontier Force was left at Maiduguri (Maifoni) as temporary military resident until the arrival of Mr. W. P. Hewby, who was appointed civil Resident of Bornu. Vid. Lugard, _Report on N. Nigeria for 1902_.]]

[Footnote 68: Moisel, p. 184. [Captain McCarthy Morrogh escorted Garbai in from Dikoa.]]

[Footnote 69: [Thanks! P. A. B.]]

III

SITUATION AND PHYSICAL CONFORMATION

Bornu is a political not a geographical entity. It is difficult, owing to the continuous changes to which the political expansion of Bornu was subjected in the course of centuries, to lay down its outline with even approximate correctness, especially as natural boundaries are conspicuous by their absence. If we leave almost entirely out of consideration the arbitrary divisions which resulted from the partition of the country between the three colonial powers, England, France, and Germany, we cannot do better on the whole even to-day than follow the line left behind by Nachtigal as the result of his conscientious investigations.[70] The extreme limits of Bornu practically correspond to a rhombus, whose longer diagonal runs from north-west to south-east. The tenth and fourteenth degrees of north latitude and the tenth and sixteenth degrees east of Greenwich denote in general the line beyond which districts belonging to Bornu do not, or do not appreciably, extend.[71] Bornu is clearly defined only on the east, where Lake Chad and its affluent the Shari form a natural boundary; it is tolerably clear also in part on the south, where indeed the Mandara highlands belonging to Adamawa form a natural boundary-wall. All other boundaries are more or less undetermined. Between the Shari and the highlands, the southern boundary of Bornu is practically coterminous with the southern boundary of the Musgu tribe—whose territory long furnished Bornu with the majority of its slaves—and almost coincides with the tenth degree of north latitude.[72] East of the Mandara highlands the boundary does not reach quite so far south, it touches here the country of the Marghi pagans between Bornu and Adamawa, and goes from about Kofa on the River Yedseram in a westerly direction till it reaches the River Gongola, follows this upwards for a short distance, and then becomes a western boundary bending round, leaving Katagum to the west, and runs north-west to Zinder.[73] From the point of intersection of this line with the tenth meridian, the Bornu boundary then runs north of Lake Chad through the country of the Tuareg, whose forays indeed were the cause of constant alterations of the border. The boundaries given above embrace a territory measuring in round figures 140,000 square kilometres.[74]

This great territory was first partitioned among the three colonial powers only on the map;[75] and it was not till the year 1903 that the settlement of their boundaries was commenced, a work which has only quite recently been concluded. The northern and smallest part, stretching from the Komadugu Yo and lying north of it, belongs to the French Sahara. The largest and central part consists of British Bornu, the north-easterly province of Northern Nigeria, and faces on the east the German Chad Territory belonging to the German Kameruns, and extends along the River Yedseram and the water-courses which establish a connexion in the rains between this river and Lake Chad. Only to the west of Dikoa does the boundary make a slight deviation over the Yedseram towards the north-west so as to include the arable land belonging to Dikoa.

Bornu is extraordinarily uniform in its physical conformation, perhaps more uniform than any other country of similar size in tropical Africa. It consists of a single vast plain covered with alluvial soil, with an average height above sea-level of between 300 and 450 metres, sloping almost imperceptibly to the lowest level of the Central Sudan, i.e. the shallow depression of Lake Chad, and never reaching a considerable elevation, save in the border districts, if we except the sand-dunes and the Wasa rocks which jut out of the plain. Although the question of mountain formation lies outside Bornu proper, yet it deserves notice, for where the mountains appear they mark the border districts of the otherwise flat country in conspicuous fashion. For not only the mountains in themselves, but also in places the flora which is dependent on them, give the landscape an appearance varying from that of the rest of Bornu. But since the way of life of the ruling tribes of the country is bound up with the plain and its products, these districts coincide naturally with the political boundaries.

The mountains, when they do appear, are of massive granite formation and are often of very peculiar shape; and they are only found in the extreme north-west and in the south of Bornu. The mountains in the north-west include the hill country of Zinder and Munio. They are remarkable on account of their composition. Immediately east of Zinder stretches north and south a mass of quartz rich in natural cisterns, which is bordered on the west, south, and east by masses of granite, which partly in a continuous chain and partly in the form of a single cone form the connexion with the highlands of Munio.[76] The granite mountains of Munio, which reach a height of 600 metres, are continued towards the south-south-east by single ‘kopjes’ of the same rock, which get lower and lower until they finally quite disappear. Peculiar to all this district, which has many of the characteristics of the desert, are the numerous lakes and pools very strongly impregnated with natron, which extend, moreover, as far as Chad and in places attain considerable dimensions, such as the twin lakes of Badamuni or Gadabuni described by Barth.[77]

Far grander than the above-mentioned mountain districts are the granite masses which form the boundary of Bornu on the south for a long distance. The highest elevation reached is in the highlands of Mandara belonging to Adamawa, which push out their northern spurs, such as the Seledeba Range, which is over 1,300 metres high, like rocky peninsulas far into the Bornu plains. The peculiar wild beauty of these mountains has already been aptly described as follows by Denham: ‘Though not to be compared with the higher Alps, the Apennines, the Jura, or even the Sierra Morena in magnitude, yet by none of these are they surpassed in picturesque interest.’[78] This mountain mass is flanked both east and west by peculiar rock formations, which may best be compared to greater or lesser rocky islets in the ocean, for they jut up quite promiscuously from the alluvial plain. If one stands on an elevated point in the country round Issege and lets one’s glance wander southwards, one has a surprising panorama in front of one, provided the weather is clear. The whole plain seems studded with shapes of various sizes, which assume the most bizarre forms and show the most diverse stages of disintegration; one sees hemispherical, cone or needle-shaped hills, even regular mountain ranges, such as that of Uba, some of them of considerable height, all over the plain, which are completely separated from each other and thus give the impression of insularity. The country on the north and east flanks of the Mandara Range is of similar formation, and among them is that of the well-known double cone of the Mendif. In some instances these granite formations hardly rise above the level of the surrounding country; such masses of rock in the form of giant tables of stone embedded in the soil are found between Uba and Issege on the left bank of the River Yedseram. The already mentioned Wasa[79] rocks are an offshoot of the extreme limit of this mountain country: they are in about 11° 30′ north latitude and their contours are a welcome change in the monotony of the plain.[80] Otherwise Bornu is entirely free from rock formations, and it is only in the country near Gujba that Rohlfs mentions the appearance of red sandstone.[81]

The low-lying plain itself consists entirely of a sandy argillaceous soil, the composition of which, however, is not uniform but contains sometimes a larger proportion of loam, sometimes of sand. The appearances which are most characteristic of this plain, viz. sand-dunes and the ‘firki’ soil peculiar to Bornu, are only intelligible by studying simultaneously its hydrographic conditions, which are indeed of supreme economic importance in this Central African country.

Bornu is, hydrographically considered, one of the most interesting districts in Africa, for its character is essentially influenced by the peculiar lake known as Lake Chad and by its affluents, which form its boundaries for long stretches or flow through it. A study of this important Central African inland lagoon and its water-system, whose investigation is amongst the most disputed of geographical problems, is absolutely necessary for the understanding of many important questions. Lake Chad, whose average water-level according to Tilho is 283 metres above the sea,[82] fills the tray-shaped depression, to which the Bornu plain slopes gradually from south-west to north-east. While it is itself extraordinarily shallow—in the deepest part the lead only marks fourteen metres—no perceptible effluent can be traced, and for a great part of its surface, owing to the fact that its banks are flat and nowhere sharply defined, and owing to a constantly varying volume of water, it is subject to continuous fluctuations. These fluctuations do not only depend on the seasons, they appear also in the course of long periods during which the shore-line may shift inside a zone of ten or more kilometres broad. A French writer, Lieut. Freydenberg,[83] relying on native information, calculates shorter periods of twenty years, which bring slight fluctuations in the level of the lake, and longer ones of about seventy years, during which it is said that there is an alternation between a nearly complete disappearance of water and a very considerable inundation. This theory, however, contradicts the observations of Barth, Rohlfs, and Nachtigal. When Barth on his exploring expedition visited Lake Chad for the second time (in 1855), the places along the western shore of the lake were threatened by an extensive inundation, and the town of Ngornu was actually completely destroyed.[84] Eleven years later, on Rohlfs’ visit, the shore of the lake was quite within normal limits.[85] According to Freydenberg’s theory a steady retrogression of the water-level of the lake ought to have followed until it had nearly completely dried up; instead of which Nachtigal in 1871 found an inundation which probably surpassed that of 1855, because it actually threatened Kukawa, which lies far from the lake.[86] A similar high-water period has not appeared again since Nachtigal’s time, on the contrary a continual shrinkage of the shore of the lake from year to year is to be registered; this shrinkage has been under scientific observation, for since 1900 the lake has been subjected by the French to careful inquiry in every direction.[87] Freydenberg, who in the summer of 1905 traversed the northern part of Chad, found that the retreat of the water had proceeded so far, that between Barua and Ngigmi on the west bank a broad ridge of completely dried-up country had pushed itself eastwards right into the middle of the lake. As long as the laws which govern the periodically recurring inundations are still unknown, the yearly recurring alterations in the level of the lake can only be attributed to the rise and fall of the rivers, to which the Chad, in addition to its own relatively small rainfall, owes its supply of water. To its most important affluent, the Shari, the lake owes practically two-thirds of the volume of water discharged into it; on this river, as well as on the season, depends its average depth. At the end of December, that is in the middle of the dry season, when all the other affluents are falling and the majority of the river-courses in the southern bay of Chad are already dried up, the lake is actually still rising, for at that time the Shari is still discharging into it a large quantity of water. From the middle of December to the end of January the Chad has reached its high-water mark and then occupies an area of about 20,000 square kilometres. Soon afterwards, in consequence of enormous evaporation, the lake begins to fall rapidly and continues falling even when the rainy season has already set in. Countless mud- or sand-banks then emerge, which have hitherto been covered with only a few inches of water, and now in some parts put on a green carpet of vegetation: owing to these the lake’s navigability, always of a very limited nature, is rendered still more difficult, if not entirely suspended.[88] In the end, apart from the often canal-like lagoons, which surround the countless islands on the Kanem shore of the lake, there remain only two comparatively small areas of open water, and they lie at the mouths of the two chief rivers, but have so little connexion with each other, that their water shows a different chemical composition.[89]

The regrettably small proportion of water in the lake during the dry season, which was not inquired into till within the last ten years, has no doubt contributed to render it possible that individual French travellers can express themselves so extremely pessimistically about the lake’s future. Audoin even calculates that according to his observations there is in normal years a sinking of about 0·15 of a metre[90] in the level of the lake. Only the future can teach us whether the retreat of the lake—apart from the loss of water due to natural evaporation—is connected with a subterraneous outlet[91] (perhaps into the bed of the Bahr-el-Ghazal) which it has constructed for itself in the course of years, with alterations in meteorological conditions at the sources of its chief affluents, or simply with tectonic displacement. Perhaps, as a matter of fact, it will again enter on a high-water period, similar to that observed by Barth and Nachtigal. At any rate, the information now available with regard to the lake itself, comprising only the results of inquiries of the last ten years,[92] is not sufficient to enable us to form a definite judgement.

Lastly, one must also, with regard to the alterations in the level of the lake, take account of those phenomena recurring daily during the dry season, which certainly have a similarity with the periodical movements of this inland sea, but are attributable solely to the prevailing winds. The north-east trade wind produces, quite mechanically and always according to its strength, a slight rise or fall of the water on the flat south-western shore of the lake. D’Huart’s assumption that the greater evaporation caused by the wind also influences this phenomenon seems to rest on slender foundations, since during the time when the wind drops, i.e. the morning hours, insolation, operating in the same sense but more intensely, sets in.[93]

Moreover, since evaporation or percolation happens to be exerted over so extraordinarily shallow a basin as that of Lake Chad, alterations are inevitable owing to the deposit of the silt brought down by the rivers. The question is, what becomes of the quantities of water from the Shari and other rivers when the basin of the lake has once been filled up with the silt? This leads to another interesting problem, viz. the connexion between Lake Chad and the Bahr-el-Ghazal, which in a way is a continuation of the lake towards the east. The views on this question, even of recent geographers, differ widely as to whether the Bahr-el- Ghazal is a former affluent or an effluent of the lake. Although Nachtigal tries to produce proof that this valley was formerly a channel of effluence of Chad leading to the depression of Bodele, which lies 100 metres nearer sea-level,[94] yet some of the travellers who have latterly examined these regions have set forth the opposite assertion.[95] Freydenberg cites as a proof of this the delta-like passage from the Bahr-el-Ghazal to the lake, as well as the appearance of stones in this channel, which can only have originated from the highlands of Tibesti in the Sahara.[96] The delta-like formation at the mouth of the Bahr-el-Ghazal is at least not a sufficient argument, for it resembles in many respects the mass of islands stretching north- westwards along the Kanem shore of Chad.[97] This archipelago is of enormous extent, and the number of its islands and islets is legion. All these islands are of sandy formation and run longitudinally from north- west to south-east, decreasing in height towards the middle of the lake.[98] Where these rows of islands cut the Kanem shore they are continued by sand-dunes; when the level of the lake is low, those dunes which are nearest to the land assume the form of peninsulas; on the other hand, canals in the form of countless lagoons running between the islands penetrate the mainland and lend an extraordinarily jagged appearance to the coast there.

Even the island, or more correctly the peninsula, of Seyorum on the Bornu side has the same longitudinal direction from north-west to south- east; the same applies to the dune formations which continue along the Bornu shore roughly parallel with it. This remarkable identity of direction between the dunes and the islands on the Kanem shore on the one hand, and the dunes on the Bornu shore on the other, is at the root of the conjecture in which Marquardsen indulges when he says, ‘the proof that the islands belong geologically to the mainland would as a consequence lead us to the conclusion that before the lake was here there existed a desert which created the dunes.’[99] As a matter of fact, such a conclusion would not be incorrect. How otherwise could the existence of dunes on the Bornu side, in spite of the thickly overgrown bank, be explained, unless in earlier times there was dry land where now is the bed of the lake? And the whole behaviour of the lake speaks in favour of such a theory. At all events, there can to-day be scarcely any doubt that the islands are the ridge-tops of dunes which were cut off from the land by water, but which formerly originated, like the still existing dunes, under the influence of the strong north-east trade wind prevailing in Bornu.[100] In what manner this collection of water, to which the Chad owes its origin, has found room for itself, whether through the blocking up of the Bahr-el-Ghazal, as Nachtigal assumes,[101] or through tectonic movement, this indeed has still to be definitely settled. Nachtigal’s assumption has in its favour the greater probability; if one accepts his theory, one can also assume as causes of the blocking up, owing to the slightness in depth of the depression, the formation of dunes as well as the copious deposit of sediment by the Shari. Audoin’s opinion, that volcanic disturbance has had a say in the matter, would presuppose a high antiquity for the lake, which is scarcely to be assumed, since neither the configuration of the shore nor its vegetation warrants such a conclusion.[102]

A further point, which has called forth much discussion, concerns the saltness of the lake. Barth, as well as Nachtigal, lays special stress on the fact that Chad is a fresh-water lake;[103] Barth does so, although he adds that the soda-impregnated soil on the north shore of the lake ‘communicates this quality to the water’. Nay, he even says further on (German edition, iii, p. 237): ‘It is very remarkable that, while the water of Chad itself is fresh, the greater part of the water found in the region quite close to the edge of the shore of the lake contains natron.’ It was at this place that ‘the water was so strongly impregnated with that mineral as to be scarcely drinkable’. Nowadays it can hardly be any longer disputed that the water of Lake Chad is salt—though certainly in a different sense to that of other well-known salt inland seas—even if the actual quantity of salt present is so small as to be generally not perceptible to the taste, at least at the period of high water.[104] The saltness of the water varies not only according to the time of year, but also according to its situation. The places which are least salt are those which lie at the mouths of the rivers and which are distinguished by the free flow of water. One may well distinguish the water here at all times of the year as fresh or quite fresh.[105] According to Tilho’s map of Lake Chad,[106] the saltness is greater to the north of the two stretches of open water than to the south of them. This tallies with the fact that those pools which are partly separated from the lake, and which contain a more or less strong solution of salt according to their volume of water, leaving behind when evaporated a visible layer of this mineral, are also to be found along the banks of the northern half of the lake, just where Denham was struck by the taste of natron in the water.[107] These salt pools form the continuation of that strip of country already mentioned, dotted with salt lakes, which extends from Munio to the north-west corner of Lake Chad and forms the southern boundary of the Sahara. The soil here contains everywhere a considerable amount of natron, which gradually disappears as one proceeds southwards. It is only natural that the water of Chad should penetrate the natron-impregnated soil and communicate traces of this mineral to the whole basin of the lake, traces which will be perceptible in a greater or less degree according to the hydrographic relations already described.

Of the rivers which mingle their waters in Lake Chad, the Shari with its affluents and side streams is by far the most considerable and also the most important for Bornu, although it is only in its lower reaches that it flows through this country. Rising outside the Sudan proper, with its furthest southerly sources lying somewhat below 6° of north latitude, in a district whose climatic conditions show a strong resemblance to the damp tropical basin of the Congo, it conveys a powerful volume of water, and, including the Logone, which debouches into it, is the only affluent of Lake Chad, which contains continuously flowing water at all times of the year. In its lower reaches, which form the eastern boundary of Bornu, the river has an average breadth of from 2,000 to 4,000 metres, but below Kusseri it has dug itself a bed contracted to 1,000 metres, and its banks have often a sheer drop of as much as ten metres. This bed is, of course, not always full. Towards the end of the dry season it is beset or constricted by multitudinous sand-banks, and the river itself is generally not more than 400 metres broad and often less, but, nevertheless, even then retains its navigability, and, with the exception of a few fords, cannot be crossed except in canoes.[108] In June the river begins to rise and reaches its highest point in October, when its level has risen from four to seven metres; it then completely fills its bed, and under certain circumstances even overflows its banks far and wide. Conditions similar to those of the Shari prevail with its most important affluent, the Logone. This river also takes its rise in districts which are lower-lying than Bornu itself, though its sources do not reach quite so far south; it also conveys a constant stream of water, and in the rains shows a still greater inclination to overflow its banks even than the Shari. Besides, in the rains it conveys to its lower reaches, through numerous streams and rivulets, a large part of the heavy rainfall from the Mandara highlands. The Logone, which rises from three to four metres in the rains, shows in places a canal-like character; its course is very tortuous, for with a very small fall it apparently has great difficulty in forcing its way through the clayey soil of the absolutely level plain. To the same reason is attributed the endeavour of the river to build up a system of steep-banked side streams, such as the deep and canal-like Lage-Matia (Laho-Matia) below Musgu—a sort of discharge-valve for the pent-up masses of water. To this also may be traced the origin of the numerous natural canals which unite the Logone and the Shari, and which caused Barth to call this region ‘the African Holland’.[109] Many of these canals appear to have no current whatever, and are designated in the language of the country ‘ngaldjam’[110] according to Barth, and are called by the same traveller ‘meadow-water’ or ‘flooded meadows’. This extensive network of natural canals is nevertheless not sufficient to drain off the volume of water brought down by the river, and so from about the end of September to a time when the waters of Lake Chad are still far from being at their highest level, widespread floods occur, which place a large part of Bornu under water. Every year from August to January the whole country between Logone and Matia is one huge lake which, during the time when the water reaches its highest level, may even extend to Mandara, and out of which the towns of the Musgu stick up like islands; the right bank of the Logone till it reaches the Shari is also then far under water. Communication between village and village is then only possible in canoes, and even the harvest has often to be brought in in this fashion.[111] At the beginning of the dry season the water in the canals very quickly runs off and only remains in isolated spots—in fact, it often dries up so completely that the canals appear as hollow roads, and can be used as avenues of communication by people on foot.[112] To the Shari-Logone system belongs also the Ba-Ili, the ‘second great artery of the Musgu country’,[113] which has just the same characteristics as the Logone, with which it is also connected through the various ‘ngaldjam’. It has this difference, however, from the latter, apart from its size, in that during the dry season it contains no flowing water, but only exhibits larger or smaller stagnant lagoons. Moreover, before the Logone, reinforced by the Ba-Ili, discharges itself into the Shari, it sends towards Lake Chad a regular network of waterways of its own, which are again themselves connected with the Shari, and thus give the impression of a wide-branching delta. The many-armed delta proper at the mouth of the Shari is, however, not in connexion with the network of canals. In their condition during the dry season these water-courses resemble the canals connecting the Shari and Logone; only a part of them then have water communication with the lake. Before the many-limbed river and canal system of the Shari-Logone reaches Lake Chad, it occasions in places the formation of a very remarkable kind of soil especially characteristic of these regions, a formation which is common also to the other rivers flowing into Chad from the south of Bornu. This formation is known in the country under the name of ‘firki’, a peculiar boggy loam—very expressively called ‘terre cassée’ by Lenfant[114]—which is saturated with water in the rainy season and forms a gigantic sponge, while the humidity thus absorbed is again returned to Chad in the form of little rivulets. Though entirely covered with water in the rainy season, the ‘firki’ dries up within a few days, the deposits disappear, and the rivulets dwindle, so that the soil as far as the eye can reach is split open by gaping cracks over a yard deep, which divide the surface into larger and smaller many-cornered lumps, and render the crossing of the marsh over these stretches extraordinarily arduous both for man and beast. Only isolated spots still remain covered with water, and form during the dry weather perennial pools. Marquardsen’s theory[115] that the ‘firki’ is a loamy deposit above a sandy subsoil has the most probability: the sand allows the water to sink through and withdraws it from the upper layer of loam, which, receiving no new accessions of humidity, dries up completely under the intensive insolation and splits asunder into big lumps. It is only where the subsoil also is loamy that it is possible for the water to lie and form those lakes that continue even during the dry weather. The layer of ‘firki’, quite different in its strength in different localities, is only interrupted by the higher dune formations, it submerges the more insignificant sand outcrops.

In many more striking ways the presence of ‘firki’ soil on the lower course of the Yedseram, the third largest of the rivers flowing towards the Chad basin, makes itself felt, as also on the smaller rivers Ngadda, Goma, and Ngua, which happen to flow parallel to it. All these do not reach Chad itself, but are in communication with the lake only during the rains by means of larger stretches of flood water, in which the current of the river is still perceptible, whilst the ‘firki’ soil underneath, probably of considerable thickness here, absorbs vast masses of the water.[116] The ‘firki’ is here sometimes piled up into regular dikes—probably built up on sand-dunes underneath—as one can clearly perceive in the dry weather at certain spots between Ulugo and Bornuski. At this time of the year there remain also channels leading from these ‘firki’ masses to Chad, such as the Mbulu, in the form of stagnant shallow canals consisting of extremely dirty water, whilst the streams ending in the ‘firki’ district have long ago dwindled away. The Yedseram is remarkable in that its hard gravelly bed contains in its upper courses even in March, i.e. at the end of the dry season, a continuous flow of water; and above all, in that its sources springing from the Mandara Highlands form even then mountain brooks of crystal clearness, and in some cases of considerable size. But below Mutube the water dwindles and becomes more and more exiguous among the pebbles, and then enters the dried-up river-bed consisting of pools more or less deep, whose water becomes more and more dull-coloured as the loamy low-lying plain is reached. At the end of September it has reached its highest level, and is then impassable except at the fords even above Issege, but even by October the running water has quickly fallen again, and recedes further and further towards the source.

Much more considerable than this river, both as regards its length[117] as well as its volume of water, is the river of Yo or the Komadugu-Yobe. It takes its rise from quite a number of streams, which unite into one river below the ruins of the ancient capital, Ghasr Eggomo, after forming large and numerous fertile islands in a rich low-lying plain. The main stream proper comes from the Hausa States and is known as the river of Katagum, that on the left, which is not much smaller, as the river of Hadeija, and that on the right coming from southernmost Bornu and rising south-east of Gujba is called Ansei. The streams coming from the north-west, such as the river flowing from Yamia, are mere ‘wadis’, as might be expected from the desert districts where they rise, and seem to contain running water only intermittently. The bed of the Komadugu- Yobe is scarcely fifty metres broad at the mouth, and the river discharges itself into Lake Chad without any delta being formed; during a great part of the year it contains no running water, and at the beginning of the rains has nothing to show but a few larger or smaller pools. At the beginning of July the river begins to contain running water, and by the end of November, when it reaches its highest level and overflows its banks, it can only be crossed by artificial means; but it falls again very quickly, and by the end of January it is again fordable, so that it can only fulfil the duty entrusted to it of forming a bulwark on the border against the hordes of robber Tuaregs in a very imperfect fashion and for a very short time.

[Footnote 70: Nachtigal, ii. 380.]

[Footnote 71: Denham gives the respective degrees as 10° and 15° N., and 12° and 18° E. (Denham, ii. 138). The difference is accounted for by the fact that Denham regarded Kanem as still forming part of Bornu.]

[Footnote 72: Along the tenth degree of north latitude runs also a part of the present boundary-line between the German Kameruns and the French Congo.]

[Footnote 73: Cf. also the map annexed to Lugard’s ‘Northern Nigeria’ in _Colonial Reports_, 1907. Even this boundary-line is somewhat arbitrary, since the districts in question belonged sometimes to Bornu and sometimes to the Hausa States on its western border. But the spheres of influence of the adjoining Sultanates are not exactly determined.]

[Footnote 74: Nachtigal puts it at 150,000 square kilometres. Nachtigal, ii. 382.]

[Footnote 75: Cf. also below.]

[Footnote 76: Chudeau, p. 334.]

[Footnote 77: Barth, iv. 68. One-half of the lake contains fresh water.]

[Footnote 78: Denham, i. 174.]

[Footnote 79: Barth, iii. 255.]

[Footnote 80: Even immediately to the south of Chad shore, but outside Bornu, in Bagirmi country, there is a similar rock formation, Hadjer el Hamis (called Hadjer Teous by Denham) composed, according to Foureau, of rhyolite. Foureau, _Doc. sc._, p. 728. Cf. also Denham, ii. 52, and the plate in Lenfant, p. 203. [Cf. p. 311 of the present work.]]

[Footnote 81: Rohlfs, ii. 120.]

[Footnote 82: _La Géographie_, xiii (1906), p. 203.]

[Footnote 83: _La Géographie_, xv (1907), p. 169.]

[Footnote 84: Barth, ii. 322; cf. v. 396. Even Denham mentions a considerable rise of Chad (at the beginning of February), owing to which the cotton plants at Wudi were under water. Denham, i. 73.]

[Footnote 85: Marquardsen attributes the high water mentioned by Rohlfs (ii. 87) to the flooding of the River Wobe (_Mitteil. aus d. deutschen Schutzgeb._ xviii. 328).]

[Footnote 86: Nachtigal, ii. 9.]

[Footnote 87: M. Audoin, _La Géographie_, xii. 308.]

[Footnote 88: Freydenberg, _La Géographie_, xv. (1907), 166.]

[Footnote 89: Tilho, _La Géographie_, xiii. 205; cf. also the map annexed to his article.]

[Footnote 90: _La Géographie_, xii. 310.]

[Footnote 91: Nachtigal, ii, p. 357; Tilho, _La Géographie_, xiii. 204.]

[Footnote 92: Information obtained prior to this only concerns the banks of the lake.]

[Footnote 93: D’Huart, _La Géographie_, ix. 164; M. Audoin, _idem_, xii. 318.]

[Footnote 94: Nachtigal, ii. 116 and 120. Cf. also Denham, ii. 57.]

[Footnote 95: Foureau, _D’Alger au Congo_, p. 670, and _Doc. sc._, p. 278.]

[Footnote 96: _La Géographie_, xv. 162.]

[Footnote 97: [Vid. Appendix IX and note [343].]]

[Footnote 98: Audoin, _La Géographie_, xii. 305.]

[Footnote 99: _Mitteilungen a. d. deutsch. Schutzgeb._, xviii. 339.]

[Footnote 100: Audoin, _La Géographie_, xii. 307.]

[Footnote 101: Nachtigal, ii. 359.]

[Footnote 102: _Mitteilungen a. d. deutsch. Schutzgeb._, xviii. 339.]

[Footnote 103: Barth, ii. 325; iii. 53.]

[Footnote 104: Tilho, _La Géographie_, xiii. 204.]

[Footnote 105: In February 1904 I camped at Ulugo on the southern shore of Chad, and for days I had to use the lake water for cooking and drinking. Although the pools, from which I took the water, were cut off from circulation with the open water owing to thick ambach and papyrus vegetation, yet I could not perceive the slightest taste of salt.]

[Footnote 106: _La Géographie_, xiii, no. 3.]

[Footnote 107: Denham, ii. 92; Nachtigal, ii. 328; Foureau, _Doc. sc._, ii. 665; Destenave, _La Géographie_, vii. 425.]

[Footnote 108: Cf. Nachtigal, ii. 549; Foureau, _D’Alger au Congo_, p. 712.]

[Footnote 109: Barth, iii. 241.]

[Footnote 110: Kund, as well as Dominik, indeed affirms that ‘ngaldjam’ means a rhinoceros in the Musgu language, so there may be some misunderstanding here. _Mitteilungen a. d. deutsch. Schutzgeb._, xix. 18; Dominik, p. 239. [‘Ngaljam’ is a Kanuri word equivalent to the Hausa ‘faddama’, meaning a shallow water-course.—P. A. B.]]

[Footnote 111: Stieber, _Kolonialblatt_, xvi. Jahrg., p. 83.]

[Footnote 112: Cf. plate, Dominik, p. 188.]

[Footnote 113: Stieber, _Kolonialblatt_, xvi. Jahrg., p. 117.]

[Footnote 114: Lenfant, _La grande route du Tchad_, chap. vi, viii.]

[Footnote 115: _Mitteilungen a. d. deutsch. Schutzgeb._, xviii. 342.]

[Footnote 116: Cf. Rohlfs, ii. 66 and 72.]

[Footnote 117: Nachtigal gives the length as about 600 kilometres. Nachtigal, ii. 353.]

IV

CLIMATE

Climatically, as well as geographically, Bornu belongs completely to the tropic zone; although cut off from the warmth of the Equator, its average temperature throughout the year of 29° to 30° centigrade makes it not only one of the hottest countries in Africa, but even of the whole world.[118] Though the climate of Bornu may on the whole be called regular, considerable fluctuations in temperature nevertheless take place, as is proved by the records of the last sixty years.

Of all the travellers who have spent any length of time in Bornu, the only observations that extend over a detailed period are those of Barth and Nachtigal; moreover, those of later travellers, which embrace a few months at the most, in the main and on the whole, in so far as they are not of small value, confirm the former results. Barth’s observations have the advantage of extending over several years, of embracing all parts of the country which he personally visited, and of having been manifestly taken at places whose situation could not influence the result one way or other. But their reliability is somewhat prejudiced by the fact that the necessary thermometrical readings are sometimes discontinued for weeks at a time, and were resumed on odd days at quite different times. These irregularities happen to be of especial importance in the thermometrical readings, which were taken after midday or during the afternoon, when according to experience quite considerable increases of temperature may occur in a short time. Owing to the fact that the traveller, especially at the beginning of his journey, was not always in a position to be able to take observations during the hottest part of the day, many high maximum temperatures no doubt escaped him. Barth himself was well aware of the insufficiency of his data, for he refers to them throughout only as ‘Fragments of a Meteorological Register’.

Nachtigal’s observations are extraordinarily conscientious and arranged with greater regularity and include also barometrical readings. But they extend, as far as Bornu is concerned, only over seven months and omit precisely the period when the highest temperatures occur. Moreover, the place where the observations were taken, viz. the court-yard of Nachtigal’s house in Kuka, was such that it must have influenced at any rate the thermometrical readings. Inside the walls of the town, daily baking in the sun, the nightly radiation from the soil does not nearly reach the same intensity as in the open plain; this is strikingly apparent in Nachtigal’s thermometrical readings. There is an absence throughout the winter months of the low temperatures, shortly before or after sunrise, which have struck every one travelling in Bornu at this season.[119] Moreover it is not to be forgotten that the shade-tree, under which Nachtigal took his observations, influenced the result both of the maximum and minimum readings.[120]

The table given below makes it apparent that April, with an average temperature of 34° centigrade, is the hottest month of the year and December the coldest; further, that the chief rainy months, July, August, and September, show the least fluctuations in daily temperature, while the coolest months, November, December, and January, show the greatest. The reasons are not far to seek, they depend on the position of the sun together with the degree of intensity of its rays resulting from its position. At the end of March the sun approaches its zenith for the first time—i.e. as regards Central Bornu—the atmosphere begins to be saturated with moisture and this lessens the fall of temperature at night, which becomes still less marked in April, in which month prevails the most oppressive heat, not yet mitigated by the rains.[121] The first tornadoes now set in, and become more and more frequent, while their ever-increasing volume lessens the nightly fall of temperature, though it affects the daily temperature in the opposite direction, so that the daily fluctuation of temperature in July and August only reaches 9°. At the beginning of the month of October the tornadoes cease, and consequently there is an ever-increasing rise of temperature to correspond, which, however, very soon drops in the following months.

[Approx. avg.: Approximate average reckoned from the observations of
both (an addition of 3° Centigrade should be made to the temperature
recorded by Barth at midday from May to August).

Mon. avg.: Monthly Average.

Avg. snr.: Average of readings at sunrise.;

Avg. 2 p.m.: Average of readings at 2 p.m.;

Hst. fl.: Highest fluctuation in daily temperature.

Avg. 12-1 p.m.: Average of readings from 12-1 p.m.

Avg. 1-2.30 p.m.: Average of readings from 1-2.30 p.m.]

+------+------------------+------------------------+-------------+----+
| | Nachtigal’s | Barth’s Observations. | | |
| | Observations. | Degrees in Centigrade. | | |
| | Degrees in | (Those given in | Approx. avg.| |
| | Centigrade. | Fahrenheit reduced | |Mon.|
|Month.| | to Centigrade.) | |avg.|
| +-----+------+-----+-----+-----+------+-----+------+------+ |
| |Avg. | Avg. |Hst. |Avg. |Avg. | Avg. |Hst. | Avg. | Avg. | |
| |snr. |2 p.m.| fl. |snr. |12-1 |1-2.30| fl. | snr. |1-2.30| |
| | | | | |p.m. | p.m. | | | p.m. | |
+------+-----+------+-----+-----+-----+------+-----+------+------+----+
|Jan |18·9 | 30 |16·2 | 15 | 35 |33 (?)| 24 | 17 | 32 |24·5|
| | | | | | | | | | | |
|Feb | 19 | 30·2 |16·1 | 15 | 31 | 35 | 19 | 17 | 33 | 25 |
| | | | | | | | | | | |
|Mar | | | | 21 | 37 | 36 | 16 | 21 | 36 |28·5|
| | | | | | | | | | | |
|Apr | | | | 25 | 37 | 43 | 18 | 25 | 43 | 34 |
| | | | | | | | | | | |
|May | | | | 23 | 40 | | 17 | 23 | 43 | 33 |
| | | | | | | | | | | |
|Jun | | | | 24 | 37 | | 20 | 24 | 40 | 32 |
| | | | | | | | (?) | | | |
| | | | | | | | | | | |
|Jul | | | | 24 | 31 | | 9 | 24 | 34 | 29 |
| | | | | | | | | | | |
|Aug |23·3 | 29·3 | 9·2 | 22 | 30 | | | 23 | 31 | 27 |
| | | | | | | | | | | |
|Sep |24·1 | 31·4 |10·8 | 22 | 31 | 36 | 11 | 23 | 34 |28·5|
| | | | | | | | | | | |
|Oct |24·5 | 35·5 |16·1 | 24 | | 36 | 15 | 24 | 36 | 30 |
| | | | | | | | | | | |
|Nov | 20 | 33 |16·5 | 16 | | 32 | 27 | 18 | 32 | 25 |
| | | | | | | | | | | |
|Dec |18·5 | 30·4 |16·3 | 12 | | 31 | 26 | 15 | 31 | 23 |
+------+-----+------+-----+-----+-----+------+-----+------+------+ |
| Yearly average of temperature measured at | | | |
| sunrise and at 2 p.m. | 21 | 35 | |
+--------------------------------------------------+------+------+----+
| Total yearly average | 28°|
+----------------------------------------------------------------+----+

Naturally the conditions for the northern and southern parts of Bornu vary somewhat, and local circumstances may also cause quite considerable fluctuations of temperature. Thus I noticed at Idjege in Southern Bornu, at the beginning of October, that the granite masses thickly studded with trees, on the slopes of which the place lies, had the result of causing an unusual increase in heat, which close to the surface of the soil measured 51° centigrade in the shade. While the nights here were even oppressively hot, a few days later on the plain at Dissa, not far from Idjege, the thermometer only showed 10° centigrade in the morning before sunrise, and by so doing registered the lowest degree of temperature, whose fluctuations far exceeded 30°. In general the heat reaches its highest point about 2 p.m. and then only declines very slowly, so that even in the evening abnormally high temperatures may be registered,[122] in fact the interesting phenomenon may occasionally be observed of a higher temperature at sunset than at midday.[123] If one takes the yearly average temperature according to Barth’s and Nachtigal’s tables, one cannot but notice that Nachtigal’s tables leave out of consideration just the hottest months. Having regard to this circumstance, one may reckon without much inaccuracy the yearly average temperature at from 29° to 30° centigrade.[124]

With regard to the second most important factor in the climate, viz. the winds, there exist no such detailed observations as with regard to the temperature; still the facts communicated on this subject by Nachtigal and recent French explorers, especially Foureau,[125] are sufficient to show that just as the atmospheric pressure is extremely regular so likewise are the winds. During the greatest part of the year, under the influence of the north-east trade-wind, north-easterly or easterly winds predominate, but south of 12° north latitude, where the wind comes mostly from the south with a decided inclination towards the south- east,[126] it is less constant, a fact which Foureau attributes to the counter-influence of the south-east trade-wind which reaches as far as here.[127] According to the universal verdict of all observers, easterly and north-easterly winds predominate especially in the dry weather, while in the rains, especially in July and August, westerly winds gain the upper hand, only to surrender their predominance to the easterly winds in October.

Views as to the daily fluctuation in the strength and direction of the wind appear to vary. Unfortunately Nachtigal’s tables[128] do not inform us whether observations with regard to the wind were taken also at night-time. The French naval officer Audoin, who personally took observations of this sort on Lake Chad, where the conduct of the watery plain is bound to give indications conformable to every movement of the wind, lays stress on the predominance of the north-east wind, especially from 11 p.m. to 9 a.m., while it is replaced by the south-west wind for the rest of the day.[129] The north-east trade-wind brings with it from the Sahara a great volume of dust, which often renders the atmosphere so opaque that the sun only appears through it like a dark-red disk. This phenomenon, which causes a peculiarly oppressive sensation,[130] is known throughout the whole of Western Tropical Africa under the name of Harmattan, but is especially characteristic of the Central Sudan.

Apart from the violent atmospheric disturbances which the tropical thunderstorms bring with them, the cyclones or tornadoes, coming sometimes from the south-east, sometimes from the north-east or south- west, especially during the times of the change in the wind, affect Bornu as well as the whole tropical district of the Western and Central Sudan.

In Bornu, where over wide stretches of country the weakening effect of dense forests is lacking, they attain an unparalleled violence. They give notice from afar of their arrival by thick clouds and whirling dust, and sweep everything off the earth that has not been made trim and taut. Sometimes cornfields which have been overtaken by them look as if a roller had passed over them. During the autumn of 1903 the roofs of the houses at Dikoa were blown off by a violent tornado. In general the cyclones are not preceded by discharges of electricity, but there are exceptions to this rule.

The tornadoes may be matched in violence by the thunderstorms which occur in Bornu. As is often the case in the tropics, they are heralded beforehand by the discharge of a large amount of electricity, a phenomenon which it is especially fine to observe at night, when the lightning flashes following each other uninterruptedly render it possible to read a book.[131]

Whilst on the subject of hydrometry, the question of cloud formation has next to be considered. The sky in Bornu, quite apart from the Harmattan, is very seldom quite clear; it is only in winter that there may be several successive days free from clouds; at other times ‘cirrus’ clouds at any rate are perceptible. The most wonderful cloud formations can be observed at the transition period shortly before the bursting of a thunderstorm. The results are analogous to those prevailing in Europe under similar circumstances; only in the Central Sudan the ‘cumulus’ formations differ in being grander, and they discharge themselves more quickly.

Fogs occur, but they are chiefly confined to the cooler parts of the year, and usually, as in northern latitudes, are connected with the presence of swampy districts, such as the damp low-lying shores of Chad. Besides it would be difficult, especially in the cooler morning hours, to distinguish sharply between fog and Harmattan.

Bornu lies exactly in the zone of the tropical summer rains, whose northerly boundary, naturally not to be sharply defined, coincides roughly with the fifteenth parallel of latitude, and thus includes also the most southerly part of the Sahara.[132] The moisture that falls there is only in the form of dew and rain. The mountains which border the country are not high enough to favour the formation of hail, which, however, is often enough to be seen in the neighbouring highlands of Central Adamawa and the grass-lands of the Kameruns that lie to the south of them. Dew appears—though only seldom—chiefly in September, October, and March; but then usually in profuse volume. Thus Denham in March and Barth in September experienced a quite extraordinarily heavy dew on the banks of Lake Chad, which they thought worthy of particular mention.[133]

The weather conditions accompanying the beginning or end of the rains usually occur with great regularity in Bornu. In this respect the country shows the typical Sudan climate, which is generally entirely without rain during the months of November, December, January, and February. The only noteworthy exception appears to be the actual borders of Chad. On the evening of the 22nd of January, an unusually cool day, and on the following morning, in the immediate vicinity of Chad, Foureau notes violent showers of rain, which were preceded in the afternoon by slight showers with some thunder and lightning.[134]

The rainfall is influenced in two different ways by the mountains on the southern border. Firstly, the volume of rain that falls here—though proper measurements of it have hitherto not been forthcoming—appears to be more intense than in the plains; and, secondly, the rains here set in earlier and do not end till later.[135] A greater or lesser distance from the Equator operates in a similar way. For the further south one goes, the longer is the duration of the rainy season and the greater the volume of the rainfall.

Apart from the above-mentioned exceptions, the first drops of rain do not fall in Bornu before the end of March. The downpour is heaviest in July and August. Though many of the tornadoes are extraordinarily severe and long-continued, yet the rain never lasts without intermission for a whole day or even a week, as it may do in the primaeval forest region of the Kameruns.[136]