Chapter VI: Superficial Accumulations

The general absence of a deeply decomposed crust upon the rocks of the Sahara and the Sudan. The drift of Hausaland and Borgu; kankar; origin of the plains of Hausaland; the surface ironstones; origin of the drift; Mounts Dala and Kogon Dutsi; the surface drift of the Bauchi plateau; conditions of formation of the surface ironstones; periods of maximum and minimum erosion; periods of elevation and depression of the crust; the recent arid period. The drift of Bornu and Katagum; the absence of surface ironstones; the “firki”-land of eastern Bornu; the shelly marls of Kukawa; the conditions of formation of the drift of Bornu; the early existence of swampy and lacustrine conditions in Bornu; regional oscillations of level in the Chad area; erosion of the floor of the present lake; the cuvettes of Manga. The drift of the Gongola valley. The drift of the Benue valley. The ferruginous crusts of the flat-topped hills of the Lower Niger and Benue.

No fossiliferous deposits of later age than Eocene have been discovered within the Protectorate. M. Chudeau[102] quotes the occurrence of Miocene shales at Boutoutou[103] to the north of Gober and concludes that the sea did not finally quit Central Africa until the close of the Miocene. If, however, any deposits of similar age were ever accumulated within the Protectorate, they appear to have been subsequently removed and replaced by unconsolidated accumulations, most probably of late Pliocene age, which now in places cover Eocene, Cretaceous, and crystalline rocks alike. These accumulations, which vary considerably in composition in different parts of the Protectorate, are of continental and in part of lacustrine origin and represent the products of weathering and disintegration of the crystalline and sedimentary rocks, sorted and rearranged by the action of rain, river, and wind. Their superficial portion now forms the surface soil which except in swampy localities is characterised by the small quantity of humus which it contains. M. Chudeau[104] has remarked upon the absence of any deep decomposition of the rocks in situ in the Sahara and Sudan to the north of lat. 11°, and his remarks may with justice be extended to the greater part of Northern Nigeria, where the original products of decomposition have rarely remained in place but have travelled outwards for a variable distance from the parent rocks and in places accumulated to very considerable depths. This drifted or travelled character of the surface alluvium may be typically observed round the margins of the larger groups of granite hills. The disintegrated debris rapidly accumulated during torrential rains at the mouths of the narrow valleys or at the base of the weathered slopes is gradually broken up and spread out upon the plain during succeeding periods of less intense fluviatile action.

1. _Hausaland and Borgu._

The red sandy loam which covers the greater part of the crystalline plains of central Hausaland and Borgu, while predominantly of local derivation, possesses for the most part a drifted or travelled character, and in composition presents relatively little affinity with the rocks which anywhere immediately underlie it. A rude stratification may at times be observed within the drift, but for the most part it belongs to the category of ill-defined alluvium and surface wash.[105] Red earth and subangular quartz grains in varying proportions naturally make up the bulk of the deposit, but it contains also more or less worn and weathered crystals of all the common rock-forming minerals together with pebbles and fragments of quartz, quartzite, granite, gneiss, amphibolite, and other rocks which are naturally most abundant where the drift is thinnest. Calcareous concretions similar to the kankar of India are common in places where the drift is predominantly earthy. They have, however, nowhere been observed forming any continuous sheet of sub- surface limestone, while, contrary to M. Hubert’s suggestion,[106] their distribution appears to have little or no relation to the occurrence of crystalline limestones within the schistose series. This composite drift, while evidently resulting from the rearrangement of weathered and disintegrated material, can frequently be seen covering up and wrapping round smooth and little weathered surfaces of rock in the midst of the open plain. The smaller inselberge moreover appear to have little influence upon the composition of the drift, which preserves its heterogeneous character up to the base of the dome, and even encloses on the margin sub-angular blocks and boulders of granite detached at an earlier period from the surface of the dome itself. In places, as on the high plains of Zaria and Bauchi, a very considerable thickness of drift has accumulated, and where locally it has been in part removed the rocky floor below is hummocky and irregular and as a rule remarkably free from any evident decomposition. It is highly probable, indeed, that the sandy loam with its heterogeneous composition and its subangular pebbles of various rocks was slowly accumulated upon an irregular surface and spread out and levelled by wind and rain.[107] That it was accumulated, moreover, under very different conditions from those which now prevail is evident from the fact that it is nowhere still in process of deposition upon the open plains of Hausaland. On the contrary it is now undergoing erosion, rearrangement, and gradual removal by the rivers and streams which in many parts of their course have not yet succeeded in exposing the rock below. Its presence also to a remarkable depth upon the watersheds and the loftiest plains makes it highly improbable that such regional accumulation could have taken place at the present elevation of the central provinces. It would seem rather that there has been within comparatively recent times a rearrangement of the drainage system of the Protectorate, brought about by a crustal movement of sufficient intensity to transform the plains of Hausaland from low-lying areas of accumulation into high-lying areas of erosion.

Such an hypothesis serves also to explain the present typical aspect of the plains themselves when it is remembered that their gently undulating character is due not so much to the smooth and even surface of the underlying rocks as to the covering of drift which has obscured all the minor irregularities of the crystalline floor below. There can be little doubt that the plains of Hausaland arose primarily as approximately base-levelled plains of erosion. There is, however, no reason to believe that the actual surface of the underlying rocks coincides in more than a general way with that of the open rolling plains of to-day. All the indications indeed point to quite the opposite conclusion. The base- levelled plains appear to have suffered some slight regional elevation and erosion, and the consequent hummocky and irregular surface has been in turn levelled up and buried in drift. The plains as we know them, therefore, while primarily plains of erosion, are in their more obvious characteristics to some considerable extent plains of accumulation.[108] The processes of erosion and accumulation moreover, to which they owe their present aspect, are for the most part the same as those which are now active in these regions. The isolated hills and the bare surfaces of rock frequently show evidence of abundant pot-holing and of extensive fluviatile action. Messrs. Andrews and Bailey[109] have suggested that a sub-aerial plain of erosion may in time, by means of dry weathering and run-off, replace a tract of hilly country at an altitude of even more than 3,000 ft., without the help of river erosion or planation. Whatever local influence these processes may have, however, the possibility of such a result upon a regional scale is exceedingly doubtful. The efficacy of dry weathering indeed, as an agent of extensive erosion, is, in spite of the able championship of Professor Walther,[110] becoming more and more discounted with the progress of research in the Sahara.[111]

A characteristic feature of the red sandy loam on the high plains of Borgu, Zaria, and Bauchi is the occurrence in places of a ferruginous cap or crust, which frequently forms a continuous floor many acres in extent. The ferruginous cap has arisen through the cementation of the quartz grains and pebbles into a hard resistant mass by means of hydrated ores of iron originally deposited in the subsoil above ground water level. The characteristic cellular or vesicular character and hard lustrous surface of the ironstone are due to the leaching out of the uncemented material and the rearrangement and partial dehydration of the iron ores upon exposure at the surface. Wherever, therefore, the ferruginous cap is now exposed, it may be assumed that the soil originally overlying it has been removed in the course of denudation. In southern Bauchi and Zaria[112] the vesicular ironstone may at times be found forming low platforms raised a few feet above the general level of the plain, which is itself in places provided with a similar floor of ironstone. At other times where erosion has been more intense it caps low flat-topped hillocks of drift whose slopes are littered with fallen blocks and boulders of gritty ironstone. Rolled pebbles of concretionary limonite and subangular fragments of an earlier ferruginous crust are frequently also characteristic constituents of the surface soil. It is evident, therefore, that in addition to the period of formation of the ferruginous crust which now in places forms the floor of the plain, there has been at least one earlier period of formation of gritty ironstone within the drift. It may therefore be presumed that the drift was originally of sufficient thickness to cover the highest platform of ironstone of which traces still remain, and that the ironstone platforms of lower levels were formed during subsequent intervals in the later erosion of the drift. In places fragments or masses of pebbly ironstone of heterogeneous composition can be seen adhering to bare and unweathered surfaces of granite, sandstone, and other rocks. Such occurrences are to be regarded as relics of earlier sheets of ironstone formed, it may have been only locally, when the drift was thicker and the ground-water at a higher level than it is at present.

The clue to the origin of the drift is found within the walls of Kano. The two flat-topped hills, Dala and Kogon Dutsi, each 150 feet in height, are composed almost entirely of soft and thoroughly decomposed rock, stained red towards the summit and capped by a layer of vesicular ironstone (p. 90). The presence of rolled fragments of an earlier ironstone in the ferruginous cap indicates that the flattened summits once formed the floor of the plain and were themselves probably covered by an earlier sheet of drifted alluvium. These hills bear witness, however, not only to the extensive decomposition _in situ_ to which the crystalline rocks were once exposed, but also to the intensity of the subsequent erosion which has left them as solitary relics of the earlier surface of the plain. It is possible, indeed, that the formation of such a sheet of weathered rock possessed a regional character, and that in its later erosion is to be found the explanation of the general absence of any deep decomposition of the rocks _in situ_ not only within the Protectorate but also over the whole surface of the central Sahara and Sudan. To the erosion of this weathered crust is also to be ascribed the formation of the hummocky and irregular surface upon which the drift was afterwards spread out, and it is probable that the greater part of the drift itself was derived from the further erosion of those portions of the original weathered surface which still remained above the level of accumulation. No facts have been ascertained with regard to the occurrence of true lateritic material either in the original sheet of weathered rock or in the drifted alluvium. There is some reason to believe, however, that the surface ironstones in places contain considerable quantities of free alumina.

It is possible that before accumulation began the weathered rock had not been entirely removed from the surface of the plains, and that portions of the original sheet of decomposed material may therefore be preserved in places underneath the drift. Moreover worn surfaces of the original weathered rock may have been left practically devoid of any covering of drift, and these surfaces, as around Aribi and Gantam on the Nassarawa tableland, may have been in course of time provided with a typical crust of gritty and vesicular ironstone. The floor of the plain of Kano around the two prominent hills probably represents in this manner a platform cut in the earlier sheet of weathered rock and covered with a ferruginous crust. A similar origin should perhaps be ascribed to the plains of Rauni which are characterised by the great depth of their surface ironstones and the presence of hillocks of similar composition which recall the buttes of Djougou in Dahomey.[113] Between Kano and Katsina, moreover, the amount of drift varies much from place to place, and appears to have been locally very largely removed and replaced by recent products of weathering and considerable accumulations of blown sand and clay which have probably arisen to some extent through the rearrangement of the earlier drift under the influence of the winds. These recent accumulations are sometimes found resting upon older surfaces of cellular ironstone, but are never themselves covered with a ferruginous crust.

The red earthy drift which forms the detached flat-topped hills upon the summit of the Bauchi plateau (Plate XI, Figs. 1 and 2), while located for the most part upon a crystalline floor, may possibly have originated to a considerable extent through the disintegration of the fragmental material of the volcanic cones in the neighbourhood, and through the rearrangement of the original weathering products which may be presumed to have formed upon the lavas of the Morrua plateau. There are, however, few remaining traces of any earlier weathering of the crystalline and igneous rocks preserved underneath the superficial accumulations. For the most part the latter rest upon smooth and comparatively unweathered surfaces of rock (Plate XV, Fig. 2), while their travelled origin is indicated not only by the presence of quartz grains and granitic debris in the ferruginous cap, but by the occurrence of tinstone in the drift throughout the whole area. The tinstone is naturally most abundant around Ngell in the neighbourhood of its centre of dispersion, but traces of it are to be found even on the hills of Rukuba from which the drift itself has now been almost entirely removed. There is good reason to believe, indeed, that the Bauchi plateau up to a comparatively recent date was continuous with the central plains of Hausaland, and that it participated in all the oscillations of conditions to which the latter were subjected. It is probable, therefore, that the plateau was at first covered with a sheet of decomposed rock of variable thickness, by the subsequent erosion of which the irregular surface was produced upon which the drift was accumulated and spread out. It is obvious, however, that the drift itself with its crust of compact or vesicular ironstone is now undergoing erosion, and that the present tin-bearing alluvium is derived for the most part from the rearrangement of the drift and only to a small extent from the denudation of the underlying rocks which is now again in process. It is probable that, during the earlier stages in the erosion of the plateau drift, accumulation continued upon the plains of Zaria and Nassarawa round the base of the plateau, but any such action is practically negligible under the present _régime_.

M. Hubert[114] in Dahomey, and Dr. Passarge[115] in Adamaua, assign only a very subordinate place to the formation of surface ironstones upon drifted alluvium. They believe that the ironstones have almost always formed within material which has arisen as the result of the decomposition _in situ_ of the rocks immediately underlying.[116] M. Hubert’s insistence upon this point of view has led him to suggest that the crystalline rocks of the plains of Borgu were levelled before being decomposed and covered with surface ironstones, and that the plains themselves have preserved their present character and elevation since the close of the Eocene. It may be granted that the crystalline rocks of Nigeria were at one time deeply decomposed, but Mounts Dala and Kogon Dutsi bear unmistakable testimony to the subsequent intense erosion of this former crust of weathered rock. In southern Bauchi and Zaria, moreover, extensive sheets of surface ironstone can frequently be seen capping typical gritty alluvium, and there can be little doubt that within the Protectorate as a whole, the formation of surface ironstone upon material actually weathered _in situ_ is the rarer phenomenon. For the most part the ironstones rest upon and have been formed within drifted and travelled superficial material which has been accumulated, near or at base-level, upon an irregular and hummocky surface, and spread out in such a manner as to obscure all the minor irregularities and to produce the present smooth and even character of the plains of Hausaland. M. Chautard[117] has noted the similar extensive formation of surface ironstones upon the ancient alluvial plains of Upper Guinea and the Futa Djalon.

The actual conditions of formation of the surface ironstones are very problematical. Their association with the older alluvium affords some reason for believing that their extensive formation in the past has been accomplished under somewhat different conditions from those which now prevail. It seems probable also that the presence of surface ironstone as a ferruginous cap upon drifted alluvium is an analogous phenomenon to its occurrence upon material which has arisen, as in Mounts Dala and Kogon Dutsi, through the decomposition _in situ_ of the rocks below. Drs. Hatch and Corstorphine[118] believe that the surface ironstones of South Africa have been formed very largely as bog iron ores by the oxidation of ferruginous solutions or by the deposition of the iron salts during the evaporation of pools of stagnant water. Captain Freydenberg[119] has noted the accumulation of ferruginous clay on the northern shores of Chad, but extensive sheets of ferruginous material comparable to the platforms of surface ironstone have nowhere been observed in process of formation upon the surface of recent alluvium or of æolian accumulations under the present _régime_ of periodic rains. The surface waters accumulated during the summer rains are for the most part rapidly and completely drained, and any probability of the formation at high altitudes within the Protectorate of pan or bog iron ores upon an impervious subsoil would appear to be prohibited by the general open and porous character of the drift. Moreover wherever true swamps, subject to partial or complete annual desiccation, are to be found, surface ironstones are typically absent, a waterlogged condition of the subsoil being apparently unfavourable to the deposition of the iron ores. M. Lacoin[120] believes that in alluvial deposits such as those of the Shari and Oubangui an enrichment in iron takes place in the lower beds at the expense of the upper, and that the banks of ferruginous material, formed by sub-surface evaporation and oxidation after the manner of the “alios” of Landes,[121] are afterwards exposed by the removal of the overlying accumulations. Such an hypothesis, however, is equally open to the objection that the surface ironstones of Nigeria are most typically found where the sub-surface drainage is most perfect. Professor Walther[122] believes that the surface ironstones are entirely the product of arid conditions, and that their origin is similar to that of the protective crust of the Libyan desert. While, however, there can be little doubt that the hardening of the surface of the ironstone and the production of the characteristic varnish is the result of the partial dehydration of the iron ores upon exposure to conditions of excessive evaporation, it is difficult to believe that the occasional rains and dews of an arid climate could suffice to produce any extensive superficial accumulation of the iron ores. MM. Chautard[123] and Hubert,[124] in view of the sporadic distribution of the ironstones, would extend to their formation Sir T. H. Holland’s hypothesis[125] of the bacterial production of laterite under humid conditions, an hypothesis, however, which still awaits experimental verification. Dr. Passarge[126] is of opinion that surface ironstones are formed under semi-arid conditions where the aridity causes the rapid evaporation of the occasional rains, while the high temperature promotes rapid chemical weathering of the rocks, and the excess of ozone and oxygen compounds in the rains, together with the absence of reducing substances in the soil, brings about the rapid oxidation and precipitation of the iron ores. Mr. Maufe[127] believes that the formation of surface ironstones (muram) in East Africa is an immediate consequence of deforestation. There is some reason to believe, however, that even underneath a covering of vegetation the slow precipitation of iron ores in the subsoil may give rise to the formation of ferruginous concretions which may in time coincide to form a continuous layer. Where the streams have trenched the deeply decomposed surface of certain belts of schists in Kabba and Nassarawa which support an abundant vegetation, it may be observed that the iron shows a decided tendency to accumulate in the upper portion. The solution of the iron is brought about by the descent of water charged with carbon dioxide and with organic acids, arising primarily from the decomposition of the superficial vegetation, while the subsequent rise and redeposition of the iron ores above ground water level is accomplished during periods of less relative humidity. It is conceivable that a similar action may take place within alluvial deposits when covered with a rich vegetation, and that a subsequent change in conditions may bring about the more rapid evaporation of the ferruginous solutions, the removal of the vegetation and of the surface soil, and the hardening and polishing of the crust into a typical surface ironstone. The present irregular distribution of the ironstones may have been conditioned by the amount of subsequent erosion as well as by the original composition of the drift, the content in iron of the circulating surface waters, the abundance of the vegetation and the character of the drainage, the free access of oxygen to the subsoil being a necessary condition of their formation.

Dr. J. M. Maclaren,[128] Mr. F. P. Mennell[129] and Mr. J. M. Campbell[130] have suggested an analogous origin for certain laterites and ironstones of India, Rhodesia and Upper Guinea respectively, and postulate for their formation an alternation of wet and dry seasons similar to the present _régime_. Messrs. Mennell and Campbell are of opinion, moreover, that the concentration of the ores upon definite horizons to form the floors of successive platforms or terraces may be explained as the result of a spasmodic lowering of the ground water level under climatic conditions similar to those which now prevail. While, however, the erosion of the platforms themselves within the drift may certainly be ascribed to this cause, it is much more probable that, at least in the case of Nigeria, the spasmodic activity in the deposition of the iron ores has been brought about by an alternation of humid and arid conditions, the formation of ironstone being most marked during periods of gradual desiccation. It is well known that the present _régime_ of periodic rains within the Sudan represents a return to more humid conditions after a period of aridity, and there is some reason to believe that there have been many such climatic oscillations in the past. Nothing is more probable, therefore, than that certain periods of desiccation should immediately follow the formation of certain platforms of erosion within the drift, and that to the incidence of these periods of desiccation is to be ascribed not only the maximum deposition of the iron ores but also the removal of the overlying alluvium and the hardening of the surface of the ironstones. It is probably correct to assume that an annual alternation of rains and droughts similar to the present _régime_ is favourable to the deposition of iron ores in the subsoil, and there is every reason to believe that their accumulation is still going on where conditions favourable to their formation are to be found. Under present conditions, however, the amount of iron salts dissolved in the ground water must be relatively less than it would be at the close of a period of maximum humidity, and it may therefore be presumed that in Northern Nigeria the present is a period of comparatively slow accumulation of iron ores. There is indeed, as already remarked, little or no evidence of any extensive formation of surface ironstone, since the beginning of the present _régime_.

The ironstones must be exposed actually at the surface before they can assume their characteristic lustrous carapace, and the removal of the vegetation and of the surface soil can be accomplished only by a climatic change. Mr. Campbell[131] in Upper Guinea, and Mr. Maufe[132] in East Africa, would assign the deforestation very largely to the influence of man, and the subsequent removal of the soil to the agency of the tropical rains.[133] It is difficult to believe, however, that the characteristic ironstone platforms of Hausaland can possess such a comparatively recent origin, and there can be little doubt that the highest platforms of which traces still remain should be referred to at least the later Tertiary period. Once formed, the hard resistant crust may be covered by later alluvial or æolian accumulations or by the products of disintegration and re-arrangement of the ironstone itself, upon the return of more humid conditions. Wherever covered by later deposits, however, the ironstone is softer and lacks its characteristic varnish. Moreover, where a scattered vegetation is now found creeping over a surface of lustrous ironstone, it is frequently a difficult matter to decide how much of the scanty soil is a recent accumulation, and how much of it represents the original covering of alluvium which had not been entirely removed.

PLATE XIV

The formation of the surface ironstones of the plains of Hausaland is thus by no means readily explicable. Their origin is probably due, in part, to many different causes, and their fragmental occurrence at various levels, sometimes upon weathered rock, sometimes upon unweathered rock, and sometimes upon drifted alluvium, points to repeated formation and repeated erosion at several different periods. It is believed that, if thoroughly investigated, their distribution and the distribution of the surface drift would afford some clue to those repeated climatic and crustal changes to which within comparatively recent times the surface of the central Sudan has been subjected. The evidence already detailed appears to be sufficient to establish the following succession of events upon the plains of Hausaland:—

(1) A period of minimum erosion during which the superficial rocks were deeply decomposed to a variable depth and covered with a ferruginous cap.

(2) A period of maximum erosion during which the earlier sheet of decomposed rock was largely removed, with the exception of such relics as Mounts Dala and Kogon Dutsi.

(3) A period of minimum erosion and accumulation during which the surface drift was spread out upon the plains.

(4) The present period of erosion of the drift during which, in places, successive platforms have been cut in the drift and floored with concretionary ironstone.

The cause of the alternation of these periods of accumulation and erosion is probably to be sought in crustal changes of a regional character. If a condition of low relief be granted to begin with, a general depression of the surface to the neighbourhood of base level may be postulated during the periods of accumulation, and a corresponding elevation of the surface above base level during the periods of erosion. The various periods indicated above may therefore be termed respectively:—

(1) The first period of depression.

(2) The first period of elevation.

(3) The second period of depression.

(4) The second period of elevation.

The history of the earlier periods is somewhat obscure and it is possible that there may have been many minor oscillations of the crust during the first period of depression and the first period of elevation. It is possible also that the first period of depression may have been represented merely by a general base levelling of the surface through erosion, and that the original sheet of weathered rock may have been formed in part underneath sedimentary rocks which were afterwards removed. The flat-topped hillocks of consolidated granitic grit and clay between Kano and Kazaure may perhaps represent contemporaneous accumulations during the first period of erosion, which were themselves subsequently exposed to denudation. There is some reason to believe, also, that there has been within the various periods a minor oscillation in climatic conditions. These minor oscillations would naturally be most clearly recognisable within the second period of elevation which is continuous with and includes the present. There is little doubt that the formation of the surface ironstones was favoured by the successive occurrence of periods of maximum evaporation, and it is highly probable that these periods assumed at times a typically desert character, and that the partially eroded surface of the drift was covered up at intervals by æolian accumulations. One such period at least immediately preceded the present more humid conditions, and is clearly recognisable in the worn and rounded dunes of the dead _erg_ of northern Hausaland. The dunes are now grassed over and covered with scattered trees, but their origin is still sufficiently indicated in their composition and general outlines as well as in the characteristic banking of the sand against rocky exposures. The surface soil and the recent alluvium of the plains of northern Hausaland probably contain much wind-blown material, and the swampy character of much of the surface of the plains during the rainy season is probably to be in part explained as the result of the interference of the accumulations of this recent arid period with the earlier system of drainage. At the same time, however, it should be remembered that the erosion of the drift has in most places not yet become very intense and that the average slope of the plains of Hausaland is as a rule so gentle that a considerable volume of water must accumulate before a perceptible flow can be established. Much swampland is found on the high plains around and to the east of Zaria, where Dr. Baikie[134] first recognised the drifted character of the alluvium.

2. _Bornu and Katagum._

A striking feature of the plains of Bornu and Katagum and of the Gongola valley between Nafada and the Yola border is the absence of any superficial crust of vesicular ironstone. The drift which may reach in Bornu a depth of more than 300 feet differs considerably in composition from the characteristic drift of Hausaland. Upon the open plains of Bornu the drift is composed for the most part of a mixture of white or yellow sand and light or dark grey clay. A red ferruginous staining of the drift is comparatively rare to the east and north-east of Gujba. The clays in places yield irregular concretions of kankar and small nodules of concretionary ironstone are dug from the surface sands in the neighbourhood of Kukawa, but no evidence has been anywhere obtained of the formation or earlier existence of any continuous sheet of vesicular ironstone. Well sections exhibit below the surface alternations of sand and clay in beds of very varying thickness.[135] The sand varies much in coarseness of grain and is in part angular and of drifted fluviatile origin and in part well rounded and of æolian origin, the latter apparently predominating largely towards the north-east. The clays are finely divided and flocculent in water and mixed with minute grains of blown sand. In places they become darker coloured from the inclusion of organic matter and pass into the “firki” (p. 52) or so-called black cotton soil of Chad and the lower Shari. Where exposed at the surface, the clay becomes plastic and very tenacious in the rains, but hard and brittle and much cracked and fissured in the dry season. These clays are rich in azote which contributes much to their fertility.[136] This advantage, however, is counterbalanced by the almost complete absence of lime, potash, magnesia, and phosphoric acid, and by the presence of large quantities of silicate of alumina and of salts of soda. Fragments of broken shells of land snails and lacustrine molluscs are common in the clays of eastern Bornu, while thin beds of calcareous marl[137] with abundant remains of recent molluscs are intercalated among the sands and clays of Mongonu and Kukawa. It is evident that the conditions of accumulation of the drift in Bornu were different from those under which the drift of Hausaland was accumulated. The latter was preeminently a terrestrial or sub-aerial accumulation while the former in the finer- grained character of the sand, the abundance of clay and the absence of pebbles and of any intense ferruginous staining affords indications of having been deposited to a great extent under water in a shallow lacustrine or swampy region. It is probable, however, that the accumulation of the drift of Bornu was, at least in the earlier stages, contemporaneous with the accumulation of the drift of the plains of Hausaland and may therefore be referred to a considerable extent to the second period of depression. No traces are left, however, upon the plains of Bornu of the earlier periods of depression and elevation, and it may therefore be presumed that the surface upon which the drift was deposited was entirely cleared of earlier accumulations during the first period of erosion and that its minor irregularities have been completely obscured by the later drift. The second period of elevation which brought about the erosion of the drift in Hausaland appears to have resulted in comparatively little erosive action upon the drift of Bornu. It is highly probable indeed that Bornu retained its swampy and lacustrine and therefore receptive character to a comparatively recent period and that the oscillations of less and more humid conditions which probably prevailed during the accumulation and erosion of the drift of Hausaland are reflected in Bornu in the intermingling of æolian and fluviatile and lacustrine material. The local abundance of alkaline salts in the drift of Bornu may also be taken as indicating an occasional and local desiccation.

The presence of swampy and lacustrine conditions in Bornu dates, therefore, from the second period of depression and there is no reason to believe that there has been any marked break in continuity between these later Tertiary swamps and the present Lake Chad. There is, however, no evidence that the present lake or the former more extensive swamps in any way represent the shrunken relics of the early Tertiary seas. The later swamps indeed, with which the lake is more or less continuous in point of time, appear never to have reached any great depth and to have been liable to repeated desiccation. Desert periods may have changed them partially or wholly into sand fields, but the return of humid conditions constantly re-established their former character.[138] It may be presumed also that during periods of exceptional rainfall the swamps became united into a shallow lake of considerable extent which may even at times have covered the greater part of central and Eastern Bornu. There is no reason to believe, however, that this was to any extent a normal condition of things, or that the drift was primarily accumulated in a large permanent lake. So far as known, marly beds occur in Bornu only in the neighbourhood of the present lake, and whatever the extent of the marginal swamps may have been, the deepest and clearest water seems always to have been confined to eastern Bornu. The thinness and local character of the marls would seem to imply, however, that the requisite conditions for their formation were by no means common or universal. Captain Freydenberg has described in sufficient detail the former extension of the swamps to the north, east, and south of Chad.

The character and thickness of the drift deposits appear to indicate that Bornu has in the past formed part of an area of predominant depression. The shelly marl of Kukawa which lies at least fifty feet above the present level of Chad, undoubtedly indicates that the floor of the lake was at one time at a higher level than it is at present. The long-continued accumulation was probably conditioned at first by the second period of depression and later by the gradual uprise of the Congo watershed during the latest period of elevation. In the neighbourhood of Kukawa, however, the low hillocks of drift show evident signs of erosion, and it may therefore be presumed that the present floor of the lake has been excavated out of the earlier lacustrine and drift deposits. This erosion which M. Garde[139] would presumably ascribe entirely to æolian activity, was probably in great part conditioned by the continued relative subsidence of the Bodele depression, the consequent eastward draining of the lacustrine region and the establishment of the Bahr el Ghazal as a waterway from west to east. The lower relative level of Bodele has now been definitely established, and M. Chudeau’s[140] arguments in favour of the eastward flow of the Bahr el Ghazal may be provisionally accepted until fuller information has confirmed or contradicted them. Desert conditions, however, followed close upon erosion. The worn and rounded hillocks of drift around Kukawa were partially buried in blown sand and the Bahr el Ghazal was blocked by the action of the wind. On the establishment of the present conditions the lake as we now know it was formed and limited towards the west by the line of sand dunes, now covered with abundant vegetation, which stretches from Kukawa to the river Yo along the margin of the lake. There is no reason, however, to postulate the presence of an earlier lake actually on the site of the present Chad at a period immediately anterior to the coming of the desert.

It may be presumed that during the period of erosion which preceded the advent of the desert, shallow watercourses were cut in the drift of Bornu and that they were in part dismembered and filled up again during the succeeding period of desiccation. One such watercourse may be clearly traced from Murguba to Maigumeri in central Bornu. The present channels of the Yo, the Maidugari river, the Yedseram, and the Shari were probably also established at the same period. The cuvettes of Manga probably originated at the same epoch and represent portions of earlier watercourses detached and isolated by æolian accumulations. The sites of the cuvettes may possibly have been determined in part by the presence of springs in the earlier river beds. While, therefore, there is little reason to believe with Captain Freydenberg[141] that the bottoms of the cuvettes represent detached portions of the original floor of an extensive lake almost entirely filled up with æolian accumulations, there is equally little reason to believe with M. Chudeau[142] that the cuvettes have arisen through the local subsidence of the underlying rocks brought about by the dissolution of beds of soluble salts, hypothetically contained within the sandstones of the Tegama below. M. Chudeau’s objection that an approximately level surface such as that of the plains of Bornu cannot be produced under the influence of the wind is clearly untenable in view of the known occurrence of extensive level sandfields in modern deserts. Moreover, in Katagum solitary dunes or isolated hummocks of blown sand formed during the recent arid period may be found alone in the midst of an open undulating plain. There is some reason also to believe that the winds which never attained any great intensity varied in direction in different parts of the Protectorate. In Bornu the dunes are elongated in a N.–S. direction while in Katagum they are more generally E.–W. and it is possible that a successive variation in the direction of the winds over the same area may have been responsible for much of the levelling of the present surface of the plain.

The lacustrine conditions which prevailed over central and eastern Bornu during the deposition of the drift extended also for a considerable distance westward through Katagum. The drift of Katagum is very similar in character to that of Bornu and the abundance in it of wind-blown material, as in the drift of Bornu, bears witness to the constant struggle between humid and desert conditions during the period of its accumulation. The predominance of the latter at a period immediately preceding the present is indicated as in Bornu by worn and rounded dunes and banks of sand. There is every reason to believe that Bornu and Katagum have been for long on the margin of the desert and that the history of the repeated advance and retreat of the desert can be clearly read in the varied character and composition of the drift.

3. _The Gongola Valley._

To the west of Gujba, thick accumulations of red sandy loam effectually obscure the irregularities of the topography of eastern Kerri Kerri. The drift has evidently been largely derived from the disintegration of the red sandstones and ironstones of Kerri Kerri, but the surface of the original deposits has been in great part rearranged by subsequent æolian action. Upon the plateau itself the æolian accumulations reach in places a very considerable thickness. In the valley of the middle Gongola to the south of Kerri Kerri there is as in Bornu and Katagum, a remarkable absence of surface ironstone. Around Nafada and southward upon the narrow plain to the west of the Gongola the drift is predominantly a red sandy loam. To the east of Ashaka, however, the drift consists very largely of a stiff tenacious black or dark grey clay, very similar in appearance to the “firki” of eastern Bornu. This material also covers large areas upon the central plain of Bauchi to the south and south-east of Deba Habe. It is clear from the manner in which the red loam and the dark coloured clay obscure the irregularities of the surface and creep up the flanks of the sandstone hills that the superficial accumulations are here also, as on the plains of Hausaland, of drifted or travelled origin and not primarily of the nature of material weathered and disintegrated in situ. The local origin of the drift may however be presumed. The red sandy loam of the northern and western plains has been derived very largely from the disintegration of the Kerri Kerri and Gombe sandstones, while the clays of the east and south have probably arisen in great part from the decomposition of the lavas of the Burra plateau. The original depth of the drift is unknown, but must have been very considerable. The Gongola in places has not yet succeeded in cutting its way through the drift to the rock below, while the highest portion of the plain which is still buried in drift lies about 300 feet above the present river bed. The total thickness, however, may well have been much greater and there is every reason to believe that what is now the valley of the middle Gongola between Nafada and Gasi formed, along with that portion of the central plain of Bauchi which lies between Deba Habe and Putu, an area of special accumulation throughout the second period of depression and that the activity of the river and its tributaries during the most recent period of elevation has been directed mainly towards the removal of the drift which had accumulated in the earlier valley.

4. _The Benue Valley._

The alternation of the major periods of accumulation and of erosion as established for the plains of Hausaland was probably to some extent of a regional character, although the relative amount of respective depression or elevation may have varied much in different parts of the Protectorate. The irregular movements to which the Benue valley as a whole was subjected during the latest period of elevation renders the correlation of the superficial deposits with those of the northern States a matter of considerable difficulty. There is good reason to believe, however, that the lower Benue valley and the greater part of the provinces of Kabba and Ilorin were exposed during the same period to very intense erosion, in consequence of which the original drift deposits have been almost entirely removed and the underlying Eocene sandstones deeply trenched and worn. The middle Benue, on the other hand, appears to have been long a region of relative depression and accumulation, and although platforms of surface ironstone may be found in places even on the plains of Muri,[143] the requisite conditions for its formation appear to have been for the most part wanting. It is possible that its absence may be due, as in Bornu and the Gongola valley, to a prolonged water-logged condition of the subsoil. Wherever, on the other hand, the surface of a hillock of red sandstone was exposed above the drift, there seems to have been a rearrangement of the ferruginous material and a concentration of the iron ores in the upper part which has given rise upon exposure to a typical honeycombed limonitic cap. The ferruginous crusts of the flat-topped hills and plateaux of the lower Benue and the Niger have a somewhat similar origin and have arisen in part through the rearrangement and deposition in the subsoil of the original iron of the sandstones and oolites and in part, like the ferruginous conglomerate on the summit of Mount Patti, through the cementation of travelled material of a similar composition. Where the summits are bare and lustrous it may be presumed, as in the case of the ironstone platforms of Hausaland, that the surface soil, originally overlying, has been removed in the course of denudation. At Awe, on the middle Benue, a sheet of sintery calcareous material appears to have been deposited upon the upturned edges of the Cretaceous before the accumulation of the drift.

PLATE XV

Attention has been already directed (pp. 27, 77) to the occurrence in Kabba, Nassarawa, and Southern Zaria of belts of felspathic muscovite magnetite schists whose deeply decomposed crusts support a luxuriant vegetation. Where partially eroded, it may be observed that the iron ores have a tendency to accumulate in the subsoil, and it is possible that if the overlying soil and vegetation were removed, the subsoil might harden into a typical vesicular ironstone. It is difficult to decide, however, to what extent the decomposed crusts may be considered the result of weathering under present conditions. It seems probable in view of the absence of any deep decomposition of the adjoining rocks that they may to a large extent represent the irregular base of the earlier sheet of decomposed rock to which the presence of Mounts Dala and Kogon Dutsi bears witness upon the plains of Hausaland and which may possibly have extended at the same time over the remainder of the Protectorate. In some places, moreover, as at Kanna, there is evidence that the crystalline floor had been locally and deeply decomposed as the result of differential weathering underneath the Eocene sandstones. A similar occurrence has been noted by M. Hubert[144] at Youri in the Niger Valley.

Little consolidated accumulations of late Tertiary age, frequently provided with a ferruginous cap like the drift of Northern Nigeria, have been described under the designation of “Benin Sands” by Mr. Parkinson[145] from Southern Nigeria, of “Ossa Sandstones” by Dr. Esch[146] from the Kameruns, of “terre de barre” by M. Hubert[147] from Dahomey, and of _reg_ from the Northern Sahara by numerous French authorities.[148] It is yet premature, however, to attempt any correlation of these deposits with the drift of Northern Nigeria.

[Footnote 102: Chudeau, _Sah. Soud._, 1909, p. 97.]

[Footnote 103: De Lapparent, _C. Rd. Ac. Sc._, 26 Dec., 1904.]

[Footnote 104: Chudeau, _op. cit._, p. 273.]

[Footnote 105: See Chamberlain and Salisbury’s _Geology_, 1905, p. 174.]

[Footnote 106: Hubert, _Miss. Sc. au Dahomey_, 1908, p. 102.]

[Footnote 107: _Cf._ Lenz, _G. M._, 1879, p. 172.]

[Footnote 108: _Cf._ Passarge, Rumpffläche und Inselberge (Kordofan- type), _Zeit. Deut. Geol. Gesell._, 56, 1904, p. 193.]

[Footnote 109: Andrews and Bailey, _Q. J. G. S._, Vol. LXVI, 1910, p. 228.]

[Footnote 110: Walther, _Das Gesetz der Wüstenbildung_, 1900.]

[Footnote 111: _Cf._ Chudeau, _Sah. Soud._, 1909, p. 285.]

[Footnote 112: Rohlfs evidently mistook the surface ironstones of Bauchi and Zaria for rocks of sedimentary origin. _Pet. Mitt. Erg._, VII, 1872, p. 64.]

[Footnote 113: Hubert, _Miss. Sci. au Dahomey._, p. 160, Plate VIII.]

[Footnote 114: Hubert, _Miss. Sci. au Dahomey_, 1908, pp. 109, 123.]

[Footnote 115: Passarge, _Adamaua_, 1895, p. 400.]

[Footnote 116: _Cf._ also Mennell, _G. M._, VI, 1909, p. 350.]

[Footnote 117: Chautard, _Le Fouta-Djallon_, 1905, p. 126.]

[Footnote 118: Hatch and Corstorphine, _The Geology of South Africa_, 2nd Ed., 1909, p. 330. _Cf._ also Hatch, _Mineral Resources of Natal_, 1910, p. 73; Rogers and Du Toit, _Geology of Cape Colony_, 1909, p. 390.]

[Footnote 119: Freydenberg, _Tchad et Chari_, 1908, p. 21.]

[Footnote 120: Lacoin, _Bull. Soc. Geol. Fr._, 1903, p. 484. _Cf._ Foureau, _Miss. Sah._, p. 674 _et seq._]

[Footnote 121: De Lapparent, _Traité de Géologie_, 4th Ed., p. 333.]

[Footnote 122: Walther, _Das Gesetz der Wüstenbildung_, 1900, p. 19.]

[Footnote 123: Chautard, _Le Fouta Djallon_, 1905, p. 137; _C. Rd. Ac. Sc._, 146, p. 239.]

[Footnote 124: Hubert, _Miss. Sc. au Dahomey_, 1908, p. 199.]

[Footnote 125: Holland, _Geol. Mag._, X, 1903, p. 59.]

[Footnote 126: Passarge, _Adamaua_, 1895, p. 396.]

[Footnote 127: Maufe, _The Geology of East Africa_, Col. Reps. (Miscell.), No. 45, 1908, p. 53.]

[Footnote 128: Maclaren, _Geol. Mag._, 3, 1906, p. 536.]

[Footnote 129: Mennell, _Geol. Mag._, 6, 1909, p. 350.]

[Footnote 130: Campbell, _Tr. Inst. M.M._ 19, 1910, p. 432.]

[Footnote 131: Campbell, _op. cit._, p. 209.]

[Footnote 132: Maufe, _op. cit._, p. 208.]

[Footnote 133: _Cf._ also Mennell, _op. cit._, p. 209.]

[Footnote 134: Baikie, _A Journey from Bida to Kano_, 1866.]

[Footnote 135: _Cf._ Freydenberg, _Tchad et Chari_, 1908, p. 1.]

[Footnote 136: Courtet, _C. Rd. Ac. Sc._, 140, p. 163.]

[Footnote 137: _Cf._ Courtet, _C. Rd. Ac. Sc._, 140, p. 160; Garde, _C. Rd. Ac. Sc._, 148, pp. 1616, 1698.]

[Footnote 138: Freydenberg, _Tchad et Chari_, 1908, p. 63.]

[Footnote 139: Garde, _C. Rd. Ac. Sc._, 148, pp. 1698, 1616.]

[Footnote 140: Chudeau, _Sah. Soud._, 1909, p. 232.]

[Footnote 141: Freydenberg, _Tchad et Chari_, 1908, p. 61.]

[Footnote 142: Chudeau, _Sah. Soud._, p. 84.]

[Footnote 143: Passarge, _Adamaua_, 1895, p. 400.]

[Footnote 144: Hubert, _Miss. Sc. au Dahomey_, 1908, p. 368.]

[Footnote 145: Parkinson, _Q. J. G. S._, 63, 1907, p. 309.]

[Footnote 146: Esch, _Geol. von Kamerun_, 1904, p. 14.]

[Footnote 147: Hubert, _Miss. Sc. au Dahomey_, 1908, p. 219.]

[Footnote 148: See Gautier, _Sahara Algerien_, 1908.]