This paper “The Geology Of North-Eastern British Somaliland”, deals with a strip of country lying along the 49th meridian east of Greenwich from the coast to about 80 miles inland, and with a belt 50 miles long-running westwards of that meridian between latitudes 11o 00′ and 11o 08′. The political boundary between British and Italian Somaliland coincides with this meridian.
The Geology Of North-Eastern British Somaliland
Part 2 Geology Of British Somaliland
By Charles Barrington Brown, M.C., M.A., F.G.S.
Quarterly Journal of the Geological Society, Volume 87, Issue 1-4, 259-280, 1 March 1931
[PLATES XX-XXV]
CONTENTS
I. Introduction
II. Topography
III. Geology
(a) Stratigraphical Succession
(b) Structure
IV. List of Works to which reference is made
I. INTRODUCTION
THIS paper deals with a strip of country lying along the 49th meridian east of Greenwich from the coast to about 80 miles inland, and with a belt 50 miles long-running westwards of that meridian between latitudes 11o 00′ and 11o 08′. The political boundary between British and Italian Somaliland coincides with this meridian.
I was fortunate in being provided with the topographical map, contoured at 250-foot intervals, which had recently been made by the officers of the Anglo-Italian Boundary Commission. Part of this was enlarged to the scale of 1 inch to a mile, and with it, I mapped geologically a considerable area.
North-eastern Somaliland has been visited previously by geologists on two occasions. R.A. Farquharson (1924), the Government Geologist of the Protectorate, made a reconnaissance from Las Khorai along the foot of the Al Hills for some 25 miles, and thence to Bunda Ziada, from whence he returned by a more northerly route. Prof. G. Stefanini (1925) journeyed southwards in 1924 from Bosaso on the coast, in Italian territory, and crossed into British Somaliland at Hegligab: thence he passed through Buran to the Nogal Valley. The greater part of the district I mapped had, therefore, been unvisited hitherto.
II. TOPOGRAPHY
The political boundary crosses four types of surface:–
(1) The Nogal Valley and Sorl Haud: drainage eastward.
(2) The Hormo Plain: drainage eastward.
(3) The Al Hills: drainage south to southeastward on one flank, and northward on the other.
(4) The coastal region: drainage northward.
(1) South of latitude 10o 10′ the Sorl Haud is the elevated plateau at 2000 to 3500 feet above sea-level, stretching from Buran to the Nogal Valley. Although the latter has no permanent streambeds (nor any features distinguishable as streambeds), it is plentifully supplied with wells, and forms a wide flat valley in the southern part of the Haud with scarps on either side: along these scarps, the ground falls from over 2500 to 1500 feet. This region was not examined geologically, but there is little doubt that it is formed of almost horizontal limestones of Middle to Upper Eocene age.
(2) The Hormo Plain is considered as being all that area along the meridian stretching from a point about east of Buran to the foot of the Al Hills at Damo. It consists of low, undulating hills interspersed with small mesas and areas of the river-plain, and is the result of the denudation of the same type of country as the Sorl Haud modified to some extent by folding. One expression of this folding is the peculiar and apparently unique structures described below as ‘synclinal basins’ or ‘sags’.
(3) The Al Hills rise somewhat sharply from the Hormo Plain, from which their slopes are separated by the Tug Darror, which runs eastwards. These slopes rise, with frequent interruptions by deep ravines and scarps, to the crest of the hills, which is sharply marked as a ridge running east and west with elevations along its course of from 4000 to 7000 feet. Immediately north of this ridge, the surface plunges, in the form of a steep scarp-face with narrow terraces on the upper parts, down to a belt of low rolling hills at between 700 and 1000 feet elevation. The topmost terrace is due to the denudation of a soft sandstone between two massive limestones. Three or four miles from the foot of the scarp is a line of limestone and conglomerate hills, attaining a height of 2100 feet in one place, with a general east-and-west direction. These hills mark the site of the Aden Rift Fault.
(4) North of the Rift Fault the ground slopes gently from about 800 feet to the seashore, forming a barren alluvial plain: it is interrupted at a short distance from the shore by a well-defined narrow ridge of hills, the highest point of which is 545 feet.
There are several interesting features in these hills. The drainage system of the coastal belt, considered as a whole, rises along the scarp of the Al Hills, flows through the Rift Fault ridge in narrow defiles, and after crossing the alluvial plain cuts through the shore ridge in similar defiles. At one point, however, 6 miles west of the Anglo-Italian boundary, there is a wind-gap. A streambed, which stretches no further back than the Rift Fault ridge, and which consequently has a small drainage basin, reaches the shore ridge at the entrance of this wind-gap and then turns sharply to the west along the southern slopes, eventually joining the stream which passes through a defile near Elayu. The shore ridge is an anticlinal fold, which rose so slowly that the northward flowing streams, with the exception just mentioned, kept their courses. That particular stream, because of its small drainage area, after a time was unable to keep its course and had to overflow to a flank.
The seaward margin of the shore ridge is a slightly modified sea-cliff, now 11 miles inland. The flat, gently sloping ground between the cliff and the seashore is a raised coral beach now 15 to 20 feet above sea-level, and which is being rapidly undercut. The top of the inland cliff is a narrow flat ledge, about 100 yards wide at an elevation of 200 feet, marking a stage in the uplift.
The drainage of the shore ridge is peculiar. Just north of the east-and-west median line, pairs of strike-streams unite at intervals and continue as dip-streams in narrow ravines through the ledge, debouching as deep V-valleys through the face of the cliff onto the coral beach. Immediately south of the median line a similar set of valley systems, T-shaped in plan, emerge on the southern flank and join the strike-streams of the main coastal plain.
The prevailing wind from September to January is due south. Moisture-laden air from the Gulf of Aden, passing over the barren and hot coastal plain, begins to ascend until, on reaching the scarp of the Al Hills, it is forced to rise rapidly and to flow over the crest. Cloud forms, and a good deal of rain falls. Consequently, the scarp face is clothed with vegetation and there are numerous springs near its base.
This water, however, disappears at once during most of the year, but reappears as pools and wells in river beds on the coastal plain. A short distance over the crest of the scarp the moving cloud begins to descend the southern slopes and rapidly disappears, leaving the slopes dry and barren.
III. GEOLOGY
The structural details will be given later. In summary, the structure is as follows. A mass of Jurassic, Cretaceous, and Eocene strata, resting on much older rocks, possibly of Palmozoie age, has been uplifted several thousands of feet along the south of the Gulf of Aden. Remote from the coast the elevation took place to a less extent and comparatively uniformly so that there is more marked tilting and far greater surface relief near the coast than inland. The Gulf branch of the African Rift-system has let down the seaward blocks, giving rise to the important scarp of the Al Hills. Hence arise the topographical divisions already noted–the sunken block of the coastal plain; the splintered mass of the Al Hills forming a lofty scarp facing seawards and (on account of the regional southern dip) its interrupted slopes down to the inland plateau; and the inland plateau now between 2500 and 3500 feet above sea-level.
(a) Stratigraphical Succession
A stratigraphical column of over 5900 feet has been worked out. The divisions represented are: Paleozoic, Jurassic, Cretaceous, Eocene, Miocene and Oligocene, Pleistocene and Recent.
Paleozoic
These are the Inda Ad slates and quartzites described by Farquharson (1924). They consist of slaty rocks, sandstones, quartzites, and limestones in varying degrees of metamorphism, and a mass of granite. Their outcrop is bounded on the north by the Aden Rift Fault, on the south by Jurassic beds on the scarp slope of the Al Hills, and on the west by similar beds on the scarp slopes of the Mana Hills. To the east, across the boundary, the fault converges towards the scarp and, as far as could be seen from a distance, this convergence causes ultimately the disappearance of the outcrop 5 miles or so beyond the boundary. The width of the outcrop is about 5 miles and the area 240 square miles.
Towards the western end of this outcrop is a large intrusion of granite covering 55 square miles: to the south, it passes under the Jurassic, to the east, near Tug Hauratiroh, it ends sharply against slates, to the west it interdigitates with slates, and to the north, its boundary is within the Paleozoic belt. It is a coarse-grained hornblende-granite with but few veins of either basic or acid later intrusions. Over most of the outcrop, it weathers to a white sand, but sporadically it forms conical and rounded hills up to 300 feet high of a bright red color–a marked feature of the landscape.
The sedimentary rocks have a general strike of a few degrees west of the north. In places, as at the foot of the Mana Hills, the strike swings more to the west, forming an arc of large radius. Belts of synclines alternate with belts of high dip where the beds are distinctly more slaty. East of the granite boss there are only a few rusty-brown limestones: west of it there are several beds up to 50 feet thick, which, in the areas of high dip (as near the Mana Hills) form marked ridges across the country. In the eastern half of the outcrop, sills of ophitic diorite, from a few inches up to 2 feet thick, are common: there are also some thin veins of pure quartz and of limonite and in one place a pegmatite vein with coarse white mica. These veins have no constant direction. West of the granite, quartz veins are more numerous, and on this side, there are hardly any diorite sills.
Jurassic
Resting on the upturned edges of the Inda Ad Series with its granite boss, at the 1400-foot contour, is a series of sandstones and limestones 1850 feet thick. At the base are 300 feet of sandstones and grits with one well-marked dark-red shale band: then 600 feet of limestone, stringed with quartz grains, containing two thick beds of sandstone. Here, at 885 feet above the base, is a bed of grey limestone extremely rich in fossils both in quantity and variety. The brachiopods (determined by Miss Muir-Wood) include Burmirhynchia subversabilis Weir, Charltonithyris africana Weir, C. bihinensis Weir, Cymatorhynchia (?) quadriplicata (Hartmann) Muir-Wood, Heimia jabbokensis Muir-Wood, and Ptyctothyris quillyensis Bayle, Weir. The echinoids (determined by Dr. E. D. Currie) include globose Cidaroid spines and Acrosalenia cf. wylliei Currie, A. somaliensis Currie, Hemipedina (Phalacropedina) somaliensis Currie and Recrosalenia somaliensis Currie. Among the lamellibranchs are Heligmus aff. humei Stefanini, H. integer Douville Weir, H. jabbokensis Cox, Modiolus blanfordi Rochebrune, Musculus sp. (: Modiola autissio dorensis Stefanini non Cotteau) and Ostrea costellata Douville. This assemblage shows the age to be Middle Jurassic (Bathonian-Bajocian), equivalent to the lowest part of the Bihin Limestone at Bihendula in Central Somaliland (Weir, 1929, 10.4).
Nearly 1000 feet more of limestones complete the column of Jurassic strata, and in the upper part occur fragments not complete enough for definite identification, but indistinguishable from Belemnopsis tanganensis Fiitterer, which, according to Weir (1929), ranges from the upper part of the Bihin Limestone (probably in part Corallian and in part Kimmeridgian) through the Meragalleh Limestone (Kimmeridgian).
Cretaceous
The top of the Jurassic has been put at the base of a massive, grey, cliff-forming limestone. From here upwards are 1350 feet of limestones, some massive and some thin-bedded and argillaceous, capped by a ferruginous sandstone 100 feet thick, making a total of 1450 feet of Cretaceous strata. There is no unconformity discernible throughout the Jurassic and Cretaceous column, and even where a change of dip occurs on the Mana Hills there appears to be continuous sedimentation.
The face of the Al scarp is precipitous, and the few routes across naturally take scree slopes through their greater part. Attempts were made to examine all the beds, both by descent from the crest and by ascents up several spurs, but sooner or later an unscaleable cliff was encountered. Consequently, the lower 900 feet, as well as the upper beds of the Jurassic above the one horizon with Belemnopsis tanganensis, may be fossiliferous, and the true base of the Cretaceous may be either below or above the massive limestone, the base of which has, for convenience, been taken as the dividing line.
At the top of this 900 feet of limestone, then, is the first fossiliferous horizon explored: a grey marl containing belemnites (Hilobites), Neithea, Exogyra and some ammonites including Holcodiscids which Dr. L. F. Spath considers of Barremian age (the highest division of the Neocomian, the Wealden). At a point 150 feet higher up comes in a thin-bedded limestone with Orbitolina, Dictyoconoides, and (two lamellibranchs which indicate a Senonian age) Ostrea (Lopha) dichotoma Bayle and Exogyra overwegi yon Buch. Orbitolina ranges upwards for nearly 300 feet to the base of the yellow ferruginous sandstone which completes the Cretaceous System. In this sandstone, the highest met within the 5900 feet of strata worked out, occurs Rudistes, and just below it are echinoids and oysters.
Eocene
These beds are chiefly limestones, but there is an important deposit of anhydrite and a number of thin clay bands. Sandstones, with the exception of some beds mentioned later and provisionally referred to this formation, are completely absent. The Eocene strata present crop out over the greater part of the area described and attain a thickness of over 2600 feet. There are no unconformities, and the beds appear to pass downwards into the Cretaceous without a break. The subdivisions are Middle and Lower Eocene.
The Lower Eocene.–These strata consist almost entirely of limestones. The lower beds are massive and form the lofty cliffs at the top of the scarp and the cliffs and deep ravines of the faulted areas. The upper beds are thin-bedded and contain several bands of yellow clay, shales, and chalky limestones, which give rise to a different type of topography. Their outcrops are entirely confined to the higher ground of the Al Hills since southwards they disappear under the Middle Eocene of the plateau region.
The group can be divided into the Damo Series (670 feet thick), and the Marojeh Limestone (800 feet thick).
The Marojeh Limestone is a well-defined unit. It is a grey rock weathering to a brownish color, and in many places gives rise to a very rough pitted surface. Along the crest of the scarp facing seawards the lower beds weather grey to white, but within the hills on cliff-faces, they have a reddish color and form caves and recesses. Some of these caves are regularly arched and are used by the Somalis as shelters for their flocks. On account of its resistance to weathering and the absence of jointing on a small scale, it gives rise to the most pronounced features of the landscape. These take the form of, first, lofty scarps on fault faces where the Marojeh Limestone abuts against the more easily eroded Damo Limestones, and secondly, deep steep-sided V-ravines and broad sloping plains, some of which, where the overlying beds have been cleaned off, are several square miles in extent.
The Damo Series is thin-bedded. White to pale-yellow chalky beds, harder bands with much concretionary chert, and yellow-green paper-shales follow in rotation. The basal beds are marked by numerous finger-like, dark-brown, ferruginous concretions. This series forms conspicuous flat-topped hills on the surface of the Marojeh Limestone, giving typical curves between miniature cliffs and ledges.
The description of the Eocene of Central Somaliland given by Wyllie and Smellie (1925, pp. 9-10), probably justifies the following correlation,[1] based on lithological considerations:–
Allakajid Limestone (thin-bedded, chalky) = Damo Series.
Auradli Limestone (massive) = Marojeh Limestone.
Fossils are fairly numerous at several horizons. Near the base of the Marojeh Limestone is a highly fossiliferous bed with many corals and a small nautilus, all being iron-stained. At higher horizons occur Siderolites miscella (d’Archiae & Haime) and Dictyoconoides newboldi (d’Archiac & Haime); whilst at the very top is a 30-foot bed which is full of Alveolina subpyrenaica Seymerie and Flosculina globosa (Seymerie). In the upper, shale-bearing horizon occur, for example, Linthia somaliensis Currie and Dictyoconoides tipperi (Davies).
The Middle Eocene.–Conformably upon the top of the Damo Series are two groups of strata, the upper of which is certainly of Middle Eocene age, whilst the lower, which contains no fossils, is included merely for convenience. They are the Buran Series (900 feet), and the Anhydrite and Gypsum Beds (respectively 200 and 200A- feet).
The Anhydrite Beds present two, and possibly three facies.
Facies 1.–Anhydrite Beds proper crop out as extensive flat tracts on the Hegligab and Hormo Plains, where denudation has removed the overlying Buran Series. Nowhere in the area examined was their base seen, though near Khumberl 150 feet of pure anhydrite is exposed. According to Wyllie and Smellie, in the Ali Wein Hills south-south-east of Berbera, gypsum-anhydrite beds are 2000 feet thick. Such beds are of great extent, covering a large part of eastern and south-eastern British Somaliland, though in the neighborhood of the Anglo-Italian boundary denudation has exposed them in only a few areas. They consist of anhydrite with a certain amount of gypsum, are white to pale-green in color, and have their origin in more or less isolated salt-water basins, as pointed out by Wyllie and Farquharson. At Hobat the pure anhydrite passes up into thin gypsiferous limestones with curious pottery-like concretions of chert.
Facies 2.–Traced northwards, these beds change their character completely, and at the foot of the southern slopes of the Al Hills they are represented by thin-bedded sintery limestones with but little gypsum. The total thickness here is 200 feet, whatever it may be under the Hormo Plain. Beds of this facies, with pottery-like concretions of chert as at Hobat, crop out also near Marojeh, where, lying below the Buran Series and above the Damo Series, they have been preserved by strong faulting.
Facies 3.–North of the Aden Rift Fault on the coastal plain there is a series of red and yellow coarse sandstones which appear to be referable to the horizon of the Anhydrite Beds since the limestones conformably above them contain fossils of Middle Eocene age. They are of unknown thickness and consist principally of sandstones with a marked concretionary structure which results, on weathering, in the production of a talus of small spheres from a quarter to 2 inches in diameter. A few limestone bands in which corals were found are also included.
The Buran Series lies conformably above the Anhydrite Beds. It consists of a series of limestones, thin-bedded and chalky at most horizons, with several bands of yellowish paper-shales. The series is over 900 feet thick and is richly fossiliferous throughout its lower half. Near the base the paper-shales are brittle and form two well-marked horizons. Dictyoconoides kohaticus (Davies) is plentiful at places in the upper horizon, and Nummulites discorbinus vat. major Rozlozsnik also occurs. Just above the upper shale there is a 30-foot bed of hard yellow limestone marked by the presence, amongst other fossils, of large quantities of N. discorbinus, which was rarely found outside this bed. The outcrops of this limestone extend over the greater part of the Hormo-Hegligab Plain and appear also at Marojeh. Above this limestone, there is a series of chalky limestones with hard bands, 170 feet thick, with a rich fauna. The lowest beds are marked by the presence of a large lamellibranch. Throughout the series, there are numerous large gastropods, such as Gisortia, 6 inches in diameter. It is remarkable that with few exceptions the Mollusca are represented by internal casts only, whereas the Echinoidea and foraminifera are well preserved. This may be said to be a general rule throughout the Kainozoic deposits in this region.
The fauna of the Buran Series includes
Nummulites discorbinus var. major Rozlozsnik
-
-
- somaliensis Nuttall & Brighton
- hormoensis Nuttall & Brighton
-
Dictyoconus cegyptiensis (Chapman)
Dictyoconoides kohaticus (Davies)
Linderina buranensis Nuttall & Brighton
Chama geslini d’Archiac & Haime
Corbula (Bicorbula) subexarata d’Archiac & Haime, cf. var. lituus Cotter
Lucina metableta Cossman
Pecten moelhensis Mayer-Eymar, Oppenheim
Spondylus radula Lamarck
Teredo (Kuphus) aft. polythalamia (Linne)
Vulsella cf. elongata v. Schauroth
Euphenax jamaicensis (Trechmann)[2]
Velates perversus (Gmelin)
Echinolampas cf. amygdala de Loriol, Currie
Euspatangus formosus de Loriol
Towards the top of the fossiliferous zone, harder bands become more numerous, and chert in tabular masses or as a filigree appears in quantity. As the series is ascended all fossils but a few nummulites disappear, and the uppermost beds, 200 feet thick, which are chalky limestones below a massive white limestone, are quite unfossiliferous. This white limestone, being almost horizontal, forms the surface of the plains around Buran, and doubtless also the greater part of the Sorl Haud to the south and south-west. It is the highest bed of the Middle Eocene discovered along this part of the boundary. The Buran Series crops out over the whole of the area under consideration, between Buran and the Tug Darror (which runs eastwards along the foot of the southern slopes of the Al Hills), except where denudation and folding has resulted in isolated exposures of the Anhydrite Beds. In the heart of the hills, near Marojeh, faulting has preserved a wedge-shaped outcrop of these beds and the underlying sintery limestone equivalent to the Anhydrite Beds. The coastal or shore hills at Elayu and Bunda Ziada and eastwards are also composed, at least in part, of the same beds, though their character is somewhat different since corals predominate there but are absent south of the scarp.
Oligocene and Miocene
Resting unconformably upon the red and yellow sandstones described above as the possible equivalents of the Anhydrite Beds (p. 266) is a thick series of micaceous sands and clays with gypsum lenses. These deposits are not fully consolidated and are brightly colored–green, pink, red, and white. No fossils were found. They crop out along the north of the Rift Fault, forming a gentle syncline just north of and parallel to it. Against the fault, they interdigitate with and pass into a thick mass of coarse conglomerate, which rests on the highly inclined masses and splinters of the Jurassic. Normally, there is a narrow belt of these rocks dividing the conglomerates from the Inda Ad Series, but in a few places to the west of the frontier, the conglomerates seem to rest on that series. The conglomerates were found to consist entirely of pebbles and boulders of Jurassic, Cretaceous, and Eocene age wherever they were examined in contact with the Jurassic.
The following is an analysis of part of the fauna collected on the southern flank of the coastal hills near Bunda Ziada.
Species 1-3 indicate a Miocene (Aquitanian-Burdigalian) age (see L. R. Cox in the discussion on this paper); species 6 indicates an Oligocene age; species 7, a Middle Eocene age. (The ages of these foraminifera are discussed by Nuttall and Brighton, 1931.)
The distances quoted in the localities are measured from Bunda Ziada Trigonometrical Station.
When I collected these fossils, together with numerous others which I considered to be of the Buran Series (Middle Eocene), I was satisfied that the whole mass of the hills was a continuous sequence of beds. There appears to be no doubt as to the later age of the fossils, and it may be that the Oligocene and Miocene beds rest in patches along the flank of the hills and were not differentiated by me. It is conceivable that they form part of the colored sands and clays cropping out not far away, though their lithological characters suggested no such connexion in the field.
Pleistocene and Recent
Referred to this group are the raised coral beach along the actual shoreline, some shell-banks (including oyster-banks with Ostrea cucullata Born) along the Tug Dugaan, and the alluvio-marine deposits of the coastal plain. These last-mentioned deposits have the nature of a composite flood-cone of detritus from the higher ground, laid down upon marine sands containing some recent shells (for example, Arca). They consist of sands, gravels, and boulders of rolled fragments of Jurassic, Cretaceous, and Eocene rocks; fragments of the Inda Ad Series are very uncommon. Consolidation of the gravel by calcareous cement is a noticeable feature along the ancient shore-line. It may be noted here that the complete cementation of modern scree and river-gravels is of common occurrence, and the resultant rock has a very deceptive appearance of considerable age.
The Quaternary is divisible into two distinct stages: an older, with a slope of 1o to the north, covering a wide area between the principal streams and forming long narrow tongues based on the foot of the slopes of the Pliocene conglomerates; and a younger, which forms tongues and terraces along the flanks of the streams at a level 40 to 50 feet below the older stage. The latter lies on a planed-down surface of the red and yellow sandstones and the Pliocene and ends abruptly to the south along the foot of the slopes of the conglomerate hills. This sudden change of slope denotes the margin of the shallow sea of the period. The elevation of this margin is now 450 feet above sea-level.
(b) Structure
As has been already noted, there are three distinct structural units:–
(1) The splintered mass of the Al Hills.
(2) The comparatively undisturbed plateau of the hinterland.
(3) The sunken block of the coastal region.
The Al Hills
These, with their extension the Aroru and Afar Hills, give rise to a lofty and precipitous scarp facing north for a length of over 100 miles, with a southern slope to the plateau.
About the 49th meridian, the mass is composed of Eocene limestones of a maximum thickness of 2100 feet, overlying Cretaceous and Jurassic limestones and sandstones 3300 feet thick with a regional dip of 8o in a direction S.-S. 10o E. The latter are exposed only along the face of the scarp, except where the Cretaceous Rudistes-sandstone is brought up by faulting in the floor of the valley at El Dibr. Along the scarp and the crest of the hills the dip is in many places greater than 8o in some reaching to 16o. Beneath are the Inda Ad slates.
FIG. 1
A number of tensional faults traverse the hills from east to west.
In the southern half of the area, the direction of the faults is west-north-west and east-south-east. Some hade northwards, some southwards, giving rise to both horsts and troughs. The value of the hade varies from 20o to 40o and the throw from 200 to nearly 1000 feet. The most noticeable fault is the one which runs just south of Marojeh, and which is responsible for the preservation of a wedge of the Buran Series. This fault is, in reality, composite, being made up of a number of faults in close echelon, each coming into existence and then dying out in turn, so that the throw varies from nothing to over 2000 feet. This splitting of the fault into echelon takes place where the fault changes direction from east-south-east to due east. To the east, it is clearly visible along the face of a scarp of Marojeh Limestone for at least 5 miles.
In the northern half of the hills, the course of the faults, which are less numerous, has a direction much more to the east and west; that is, from 5o to 10o north of west. The hade is greater, reaching as much as 52o, and is in all cases northwards. Still further north, a hade of about 45o is found in the Aden Rift Fault.
The Tug Galweineh lies in a trough a mile wide with precipitous walls on either side, the southern walls being far loftier than those on the north. This valley lies along the trace of a fault with considerable hade and a throw of 250 feet. This is proved by the occurrence of the Cretaceous ferruginous sandstone with Rudistes in the ravine-bed on the south side at El Dibr, a mile to the east of the political boundary.
FIG. 2
The bed of the Tug Darado has a like configuration on a smaller scale and is undoubtedly due to a fault with a hade of 52o northwards. The fault is clearly seen to the east of the valley and in the same line of strike.
The majority of these faults in the splintered zone are easily traceable, and form prominent scarps where the resistant Marojeh Limestone is brought up against the more easily eroded beds of the Buran and Damo Series.
The base of the Jurassic rocks is 1400 feet above sea level at the 49th meridian, and its beds dip due southwards at from 7o to 13o Twenty miles further west it stands at 2000 to 2200 feet, the increased elevation being due to the emergence of the Galweineh and Darado faults, or of others of that fault-system; while at Inda Ad it stands at 1880 feet with a small dip to the south. Here, the direction of the scarp takes a remarkable turn at right angles and faces eastward: the Jurassic and overlying beds are horizontal, and the scarp of the Mana Hills is therefore due to the lost dip and not to north-and-south faulting, as suggested by Farquharson. In fact, faulting in this direction is absent in the area.
The plateau of the hinterland
From the foot of the Al Hills to the Karkar Hills which form the northern edge of the Sorl Haud, the eastward-flowing Tug Darror, drawing its tributaries from both these ranges, has lowered the surface in its basin down to 2100 feet above sea-level.
Around Buran, at the northern limits of the Sorl Haud, the beds are practically horizontal and the surface coincides with the highest (unfossiliferous) beds found in the whole district. In the Buran canyon the lower (fossiliferous) beds are exposed in the walls and floor, and farther out into the open, successively lower beds are exposed down to the Anhydrite Beds, but nothing below the upper layers of that deposit is seen. Here, the elevation of the top of the Anhydrite Beds is 2600 feet, but at Sunto, the lowest point on the boundary (2100 feet), beds are exposed which at Buran stand at 3000 feet. Twenty miles west of Sunto the Anhydrite Beds cover a large area at about 2700 feet high. It is clear that there is a marked geological depression about Sunto, but its extension eastwards into Italian territory is unknown.
The edges of the Sorl Haud plain are, as might be expected, quite ragged, and outliers in the shape of flat-topped and rounded hills are numerous. Farther away from the plain these features are less pronounced, and the topographical relief is low. The particular limestone which forms the top of these hills around Hegligab, Hormo, and elsewhere, is a bed 30 feet thick near the base of the fossiliferous strata of the Buran Series, and marked by the presence of large numbers of Nummulites discorbinus var. major Rozlozsnik.
The most remarkable character of this part of the country, from Hegligab northwards to the foot of the Al Hills, is the presence of small synclinal basins, or sags, scattered sporadically over a large area, but confined certainly on the north by an east and west line along the foot of those hills. Many such basins were noted along the road running westwards from Buran to Erigavo, and again for several miles south of Erigavo. Apart from these regularly marked and symmetrical folds, there are many gentle undulations of larger extent and uneven character, interspersed with areas of almost horizontal and undisturbed strata. The surface might be described as rumpled by undulations and pitted with basins.
The basins are almost all elliptical in shape, with the longer axis directed about north-west to south-east. A few are apparently circular. The longer axis varies in length between a quarter to 1 1/2 miles, and the shorter axis, as a rule, is about half the length of the longer. The amplitude of folding is nowhere more than 400 feet and in general, is less—about 200 feet. There is no faulting of any kind in the whole area, with the exception of a very few quite trivial slips unconnected with any basin. The limestone beds, some 4 or 5 feet thick, have flowed perfectly, as though they had been cast in molds. The dips of the strata which flank the basins vary from a few degrees up to a maximum of 48o.
In general, the strata between basins are horizontal, and such blocks may be found dissected into flat-topped hills. Where two basins are close together the effect of an anticlinal fold is obtained, but there is no actual uplift into arches.
Most of the basins have given a trend to the drainage and have been hollowed out as deep ravines: some show very little relief. One of the largest, near Hegligab, has resulted in the preservation of an outlier of the higher fossiliferous beds. Another in the same district is almost a perfect circle, a mile in diameter with a small central hill. A very fine example occurs in the Tug Buran, 5 miles below the water-pool: it was surveyed with a plane-table. The strata curve down regularly from the flanking hills parallel to the longer axis, attaining a maximum dip of 48o, while at either end the outcrops of limestone beds curve sharply like the edges of a nest of pie dishes. The depth of the sag is 400 feet.
There seems to be no regularity of distribution (as, for example, in lines) other than that of the generally similar direction of the axes. Many of the basins were mapped, but more could be seen over a wider area than could be investigated. In the aerial survey photographs, the majority are easily detected. One feature which distinguishes them from fiat-topped hills of horizontal or non-folded strata is the direction of the drainage.
To give definition to a discussion of the origin of these basins, the following facts should be noted:-
The succession of beds is
The lower, fossiliferous, part of the Buran Series 400 feet
Anhydrite Beds 200+ feet
Lower Eocene 1470 feet
In the area where the basins were mapped, only the Anhydrite and Buran Beds are exposed, and all are affected; while outside that area, where basins were noted but not mapped, the same beds crop out, so far as could be determined by a cursory examination.
The basins are sporadic, numerous, and individually small. The general trend of their longer axes is north-west to southeast. Many are sharply defined between areas of horizontal strata. Some have steeply dipping sides and some have gentle dips, giving the impression of a downward limit to the curvature or the dying away at depth.
A satisfactory origin for these basins is hard to find. I put forward three possibilities:–
(i) Deposition of sediments on an uneven floor.
(ii) Regional folding.
(iii) Collapse along short lines and points.
(i) It is impossible to imagine sediments laid down in beds of uniform thickness on slopes of 48o, and even were it possible, the postulated depressions on the sea-floor would have to be accounted for. This theory can be definitely ruled out.
(ii) There are large areas where the undulations are gentle and without any perceptible directional design. These folds may have been caused by the movement responsible for the uplift of the region before the formation of the Aden Rift Fault. The basins with their orientation certainly appear to have some connexion with the north-west to south-east system of faulting, and with this movement. Nevertheless, the form of the basins, their sag-like appearance, and the absence of uplifted anticlinals between them make it difficult to assign their origin to simple folding.
(iii) Collapse, by the removal of supporting anhydrite strata, either by solution or by change of volume, might be considered possible and even probable, since the basins are confined to the plateau region where the anhydrite occurs, and are absent in the north where those beds have passed laterally into sintery limestones.
Such a collapse would necessarily have to be extremely slow, otherwise, fracture (which is absent) would result. If the removal of support were gradual and continuous during the deposition of the succeeding limestones, these beds, being still waterlogged and in a plastic state, might flow sufficiently to accommodate themselves to the change in shape and increase in area without breaking.
Gypsum has a specific gravity of 2.3, anhydrite 2.8 to 2.9. It is conceivable that the calcium sulphate was originally laid down as the hydrated form. Later, dehydration might take place, and the resulting decrease in volume, 18 percent, would cause a removal of support. The assumptions to be made are, first, that the conversion of gypsum was slow and continuous and took place during the deposition of the succeeding limestones, and, in fact, of the later layers of sulphate, and secondly, that conversion on any large scale was confined to patches and short belts.
The objections to this theory are that the general rule of orientation of axes is unaccounted for, and that the second assumption is a large one.
The sunken block of the coastal region
With the formation of the African Rift Valley, the crust collapsed along an east-and-west belt between Arabia and Somaliland, and the Gulf of Aden came into being. As elsewhere, step-faulting on a large scale is the expression of this collapse.
The Admiralty chart (1912) of the Gulf of Aden shows a considerable depth (1200 fathoms) in the middle, but, unfortunately, soundings are confined to but a short distance from the Somaliland coast-line. The 100-fathom line runs parallel to the coast at from 3 to 4 miles from it. The few soundings beyond indicate that no bottom was reached at, in general, 120 fathoms. There is hence no possibility of proving the presence of one or more submarine scarps and their rift-faults.
Along the foot of the Al Hills, in latitude 11o 06′, there is a well-defined east-and-west fault with a throw of 6000 feet and a hade near 45o: the lateral displacement must, therefore, be at least a mile. This fault, at any rate, defines one foundered block or belt on its south side–a belt partly dry land and partly submerged. On the south side of the fault is the uplifted mass of the maritime hills, the Al, Aroru, and Afaf.
The existence of faulting, accounting for the long straight scarp of these hills, has been known for some time, but its exact location and character have not hitherto, I believe, been determined. Contrary to what might be expected, the fault-plane does not coincide approximately with the scarp-face: the trace is well in advance of it, over 5 miles in places, showing the extent to which denudation has caused the scarp-face to recede since its initiation.
The fault is naturally compound: and in one or two places it can be seen to be made up of parallel shears. It is probable that it consists of a number of sheets in the form of steps. The conglomerates, unfortunately, cover up any further evidence as to its nature. It was traced from just east of the political boundary along the foot of the Al Hills to the Mana Hills, a distance of 48 miles. Its course is well marked everywhere by a line of hills which reach, in one place, 2094 feet, although this is exceptional. In many places, these hills form jagged scarps of highly inclined Jurassic strata resting on the Inda Ad slates and capped or flanked by the conglomerates on the north side. Elsewhere, the conglomerates border the fault without the intervention of the Jurassic beds, though it is likely that the latter are present everywhere along the line but are hidden by talus slopes.
With reference to the age of the Rift Fault: Gregory (in Wyllie & Smellie, p. 6), on account of the absence of marine Miocene or Pliocene from part of Somaliland, considers that the Gulf of Aden was not occupied by the sea until shortly before Pleistocene times and that the Gulf of Aden faults are late Pliocene.
The marine ‘Miocene’ beds of Bunda Ziada suggest that the sea was present and that the sunken block of the Gulf had come into existence. Stefanini (1925) records beds with fossils of Oligocene and early Miocene types along a zone restricted to the coast of the Indian Ocean at Hafun, Bender Beila, and the mouth of the Nogal, emphasizing their limited extent but not stating their relation to the Eocene on which they rest. However, the Miocene sea may have had a wider extension over the site of what was later the Rift, and, on the occurrence of that fracture, all traces of the beds on the south of the fault may have been subsequently removed.
The colored sands and clays with their thick conglomerates must have been formed against a great scarp, and the syncline with an east-and-west axis along its base shows that the fault producing the scarp was in operation for some time afterwards. It is unfortunate that the age of these beds could not be ascertained. As suggested above, they may belong to the Miocene beds, but more probably they are later.
Grateful acknowledgments are made to the following:- The Secretary of State for the Dominions and Colonies, for permission to publish this paper; Lt.-Col. J. H. Stafford, O.B.E., M.C., R.E., Senior Commissioner; Major J. F. Phipps, R.E., and Lt. R. H. R. Taylor, R.A., for the use of their maps; Lance-Corporal P. E. Crofton, R.E., for drafting copies; the Air Ministry, for permission to use the aerial photographs; Mr. L. R. Cox and Dr. L. F. Spath for the determination of certain fossil species.
I am particularly indebted to Mr. A. G. Brighton for the determination of many species and for much valuable advice on paleontological questions.
IV. LIST OF WORKS TO WHICH REFERENCE IS MADE
EXPLANATION OF PLATES XX-XXV
PLATE XX
Mosaic of air photographs, showing the faults near Marojeh. Scale: 1 inch = 1 ½ miles; the area represented is 32 square miles. The photograph is clearer when viewed inverted, and should be examined in conjunction with text-figure 1, p. 270.
PLATE XXI
Fig. 1. Synclinal basin near Buran. View from the western end of longer axis.
-
-
-
- Aerial photograph of two s3rnclinal basins, Hormo.
-
-
PLATE XXII
Fig. 1. Ravines in the Marojeh Limestone, Marojeh.
-
-
-
-
- Erosion of strata, Buran.
-
-
-
PLATE XXIII
Composite table of strata, Buran to the northern scarp of the Al Hills.
PLATE XXIV
Geological section along the 49th meridian of longitude from the coast near Bunda Ziada to latitude 10o 11″, Somaliland. Horizontal scale: 1 inch = 3 miles; vertical scale, 1 inch = 6000 feet.
PLATE XXV
Geological map of the country along part of the Anglo-Italian boundary, Somaliland, and some adjacent parts of British Somaliland. Scale: 1 inch = 6 miles, approx.
PLATE XX
PLATE XXI
Royal Air Force official. Crown copyright reserved.
PLATE XXII
PLATE XXIII
PLATE XXIV
PLATE XXIV
DISCUSSION
Prof. J. W. GREGORY, in congratulating the Author on this important contribution to the geology of East Africa, referred to the marked contrast between the agreement of Eastern and Western Somaliland as to the Jurassic succession, and the dissimilarity with regard to the Cretaceous. The evidence indicates that both areas were under the same geographical conditions in Jurassic time, but the eastern area in the Cretaceous Period was affected by movements that did not affect Western Somaliland. Similar Lower Cretaceous occurs near Mombasa, and although that has no known higher Cretaceous, some occurs in Socotra. The Lower Miocene appears to correspond with the Findi Isa beds discovered by Mr. Hobley, from which a much larger fauna has been collected by Miss McKinnon Wood. He congratulated the Author on his determination of the nature and extent of the Aden Gulf Rift Fault, and trusted that, when the fossils had been determined, additional light would also be thrown on the date of the faulting.
Mr. L. R. Cox said that he had worked through part of the Author’s collection of fossils with Mr. A. G. Brighton. The Jurassic succession agrees on the whole with that established by previous workers at Bihendula, although the Upper Jurassic fauna is less in evidence. Doubt had been cast upon the correctness of assigning the lower portion of the Bihendula beds to the Middle Jurassic (that is, Bajocian’-Bathonian), as all the ammonites that had been found at Bihendula were of Upper Jurassic age. In the speaker’s opinion, although no ammonites had been found in the lower beds of the series, the lamellibranchs alone afford clear evidence that the Middle Jurassic is present. The Middle Jurassic faunaos is remarkable for the abundance of the lamellibranch genus Heligmus, which, so far as we know, is confined to the Bajocian, Bathonian, and Callovian. This genus occurs rarely in France and other European localities and recently has been recorded from Madagascar by Barrabe. In Transjordania, Sinai, and Somaliland, however, it is represented by several peculiar species, one rather like a Rhynchonella in appearance, others much flatter and very similar to plicated oysters. The peculiar oyster, Ostrea costellata (Douvilld), is another useful index-fossil of the Middle Jurassic, and is known to occur in Sinai and Arabia as well as in Somaliland.
The speaker was glad that the foraminifera studied by Mr. Brighton had confirmed his suspicion that the Lower Miocene is present in the coastal series near Bunda Ziada, although apparently, it is not easily distinguishable from the Eocene in the field. The presence of Chlamys (A~quipecten) pallium (Linnd) indicates that either Miocene or more recent beds are present; but the most useful index-fossil of this horizon is a small leaf-like oyster, characterized by rounded folds radiating to the margins from an elongate, cylindrical attachment surface, which corresponds to growth on a small twig. This oyster had been found by Miss M. McKinnon Wood at all the three Lower Miocene localities in Kenya Colony from which she had collected and among those collected by Mr. G. M. Stockley in Pemba. The speaker had identified it provisionally with the Linnean species, Ostrea folium, which has not been found living in East African waters, although it occurs in other Indo-Pacific localities. The Lower Miocene form, however, is always smaller than the living one, and very constant in size; since it seems to be so useful a stratigraphical index, it may, therefore, be worth distinguishing by a separate name. An available name is 0. longifolium (1922, Samml. Geol. Reichs-Mus. Leiden, n.s. vol. i, p. 481 & pl. lxi, figs. 95-97), described as occurring at the very same Lower Miocene horizon in Java.
The AUTHOR said that, since no questions had been put, there was little to add. He wished, however, to draw attention to some rock-specimens exhibited:– (a) wind-modified stones from the coastal plain at Berbera, which showed three quite distinct types of etching on different faces; and (b) a block of gypsum from the colored sands near Bunda Ziada. This gypsum is a bed 12 inches thick, with a dip of a few degrees, projecting as a band: the crest of the small scarp thus formed for a length of over 20 feet, has been so sculptured by rain as to reproduce in miniature a drainage-system of zigzag watershed, cwms, and corries as analyzed by Prof. Mart.
NOTES
[1] These beds are inverted in the sequence (Wyllie & Smellie, 1925, p. 2) and inverted and transposed in the legend to plate I (op. cit.).
[2] Described by L. R. Cox (1931).
