Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
GEOLOGY OF GHANA 1 1. THE STRATIGRAPHIC SUCCESSION OF GHANA Era Period/Epoch System/series/information Quarternary-Recent Unconsolidated clays and sands of lagoon, delta and littoral areas. Caenozoic Tertiary-Eocene Partly consolidated red continental deposits of sandy clay and gravel. Cretaceous Apollonian formation (Upper cretaceous, i.e. Cenomanian – Campanian) Mesozoic Jurassic Amisian formation (upper Jurassic-lower cretaceous) Devonian Sekondian series (middle Devonian – lower cretaceous) Accraian series (early or middle Paleozoic Devonian) Cambrian Voltaian system (late Proterozoic to early paleozoic, i.e. to 300 – 1000 m.y.) Upper Precembrian Buem formation Togo series Dahomeyan system. Age uncertain (middle to Proterozoic late Precambrian, probably reactivated at about 550 m.y. before present) Middle Precembrian Tarkwaian system (Age unknown, possibly 1650 – 1850 m.y.) Birimian system (Age uncertain, approximately 1800 – 2100 m.y.) 2 The geological provinces of Ghana Ghana can conveniently be divided into five geological domains or provinces on the basis of age, tectonics and lithologic characteristics of the supracrustal rocks. These are: (1) The western unit which lies at the eastern margin of the Precambrian West African Shield or Craton, (2) The southeastern unit which is at the southeastern part of the country belonging to the Precambrian Mobile Belt and (3) The flat lying central unit made up mainly of the sediments of the Voltaian system (4) The coastal basins and (5) Tertiary to Recent deposits. The Western Unit Almost 45 per cent of Ghana’s territory belongs to the Shield area. This part consists of lower Proterozoic volcanics and flyschoid metasediments of the Birimian system. The Birimian was deformed, metamorphosed and intruded by syn- and post-granitoids during the Eburnean orogeny which occurred about 1800 million years ago. In elongated basins which follow the northeasterly trending Birimian belts, middle Proterozoic molasses type sediments of the Tarkwaian system were deposited. The Ghanaian shield area has therefore two main rock systems i.e. the Birimian with its associated granitoids intrusives and the Tarkwaian. Supracrustals: the Birimian and Tarkwaian The very thick and extensive sequence of metamorphosed sediments and volcanics that dominates this age province is called the Birimian after the Birim region in southern Ghana where the rocks were first described in detail. Metamorphic grades range from greenschist to almandine-amphibolite facies in these rocks, which are an important source of diamonds and manganese ores. A much smaller and more scattered group of supracrustals, mainly shallow-water sediments, is called the Tarkwaian (after the town of Tarkwa in southern Ghana, were they are gold-bearing). Although greenschist to almandine-amphibolite facies metamorphism is recorded from these rocks, in many places they are described as being 3 ‘hardly metamorphosed’. A consensus has not yet been reached about the stratigraphy of the Birimian and its relationship to the Tarkwaian. Geologists working in Ghana on the one hand, and Ivory Coast and Upper Volta on the other, have arrived at different conclusions. It will be necessary to deal with the two views separately and then to examine reasons for the differences between them. The Birimian in Ghana In Ghana there is a long-established subdivision into Lower Birimian, dominated by metasediments, and Upper Birimian, dominated by greenstone-type metavolcanics. The lowest parts of the succession are primarily phyllites and greywackes. These change upwards to phyllites and weakly metamorphosed tuffs, greywackes and feldspathic sandstones, and the sequence appears to pass conformably into the Upper Birimian, although a local unconformity has been recorded in places. Some of the phyllites contain pyrite, and finely divided carbonaceous matter is present in most of them. Silicification is common among the phyllites, particularly towards the boundary with the Upper Birimian. Quartzites, calcareous rocks and conglomerates are rare, but the conglomeratic horizons contain fragments of granitic and other rocks believed to be derived from older basement. The Upper Birimian consists chiefly of metamorphosed basaltic and andesitic lavas, now hornblende-actinolite-schists, calcareous chlorite-schists and amphibolites (the greenstones). Pillow structures indicating sub-aqueous eruption of the original basaltic lavas are frequently observed. Minor intrusions f mafic rocks cut the volcanics and there are small ultramafic bodies in some places. Smaller amounts of rhyolitic and dacitic lavas and tuffs are also recorded, and subordinate metasediments include phyllite, greywacke, quartz-sericite-schists and mica-schists, as well as grits and conglomerates at the base of the succession. Bands of gondite (quart-spessartite rock) and manganiferous phyllite occur within the greenstones towards the top of the Upper Birimian, commonly associated with tuffs, silicified argillites (hornstones) and chert. The rocks are dark and finely banded, and the presence of carbonaceous matter in them suggests deposition in still waters deficient in 4 oxygen. Mn-rich horizons also occur at stratigraphically lower levels in the Upper Birimian and have been found in uppermost Lower Birimian as well. Because the rocks are tightly folded and commonly sheared and fractured, it is not easy to establish stratigraphic successions and estimated thicknesses. The total thickness of the Birimian in Ghana may be of the order of 10000 to 15000m. The Tarkwaian in Ghana The main occurrences of Tarkwaian sediments occupy two generally synclinal belts surrounded by Upper Birimian metavolcanics, about 270 km apart and smaller occurrences of Tarkwaian rocks may occur elsewhere. The sediments are mainly of shallow-water origin, probably fluviatile, and they contain fragments of Birimian rocks. They could have been deposited in separate elongate basins as molassic facies derived from erosion of the Birimian, during later stages of the Eburnian orogeny. In Ghana, there is in places a strong angular unconformity between the Birimian and Tarkwaian, and the Tarkwaian appears in general to be less strongly deformed and metamorphosed than the Birimian. However, neither the unconformity nor the difference in intensity of folding and metamorphism can be convincingly demonstrated at all outcrops. The overall synclinal structure of the Tarkwaian in the vicinity of Tarkwa itself is relatively simple, with fairly open folds having a northeasterly plunge and northwesterly dipping foliation, though the intensity fo folding increases to the northwest. In the Bui syncline, Tarkwaian beds have been considerably fractured and more strongly folded, being overturned in places. According to the Ghanaian ‘school’, therefore, Upper Birimian metavolcanics occupy the cores of the rather broad Birimian synforms, with granitic rocks (partly basement) in the intervening antiforms. Tarkwaian rocks form the central parts of two synformal belts, and in the larger Tarkwa syncline an unconformity is implied by the fact that the Tarkwaian lies mainly against Upper Birimian metavolcanics in the west, but overlap onto Lower Birimian in the west. Although the Birimian and Tarkwaian of Ghana may have been studied for longer and in more detail than in adjacent countries, they form only a small part of the Proterozoic domain and are not typical of the whole of it. 5 The Birimian and Tarkwaian in Ivory Coast and Upper Volta In northeastern Upper Volta (and southwestern Niger), the supracrustal belts consist mainly of greenstone sequences and volcanosedimentary successions similar to those in Ghana, though they are in general smaller. These are sometimes called Type I belts by francophone geologists, to distinguish them from the Type II belts which occupy most the western part of the Proterozoic terrane, that is, Ivory Coast (and northern Guinea). These are mainly sediments of shallow-water origin, among which quartzites, mica-schists, metagreywackes and metaconglomerates are important lithologies. Calc-silicate rocks also occur, and volcanic rocks are subordinate, ranging from mafic through intermediate to acid compositions. In Ivory Coast and Upper Volta, the greenstone facies (upper Birimian of Ghana) is generally considered to be either older than or broadly contemporaneous with the predominantly sedimentary facies (Lower Birimian of Ghana). These interpretations are based partly on the occurrence of volcanic pebbles in conglomerates of the sedimentary formations. Mica-schist pebbles are also found in some intraformational conglomerates, implying that there was deformation, metamorphism and erosion between the deposition of Lower and Upper Birimian sequences in at least some places. The ‘Tarkwaian’ also occurs in Ivory Coast and Upper Volta, as scattered sequences of molasses-type sediments, with or without conglomerates. Their relationship to the Birimian is less simply defined than in Ghana, because the two have been interfolded to produce very complex structures. In consequence, sediments of Tarkwaian type seem to be regarded as lateral facies variations within the main Birimian sedimentary facies: a series of local shallow deposits of roughtly similar character, occurring at various stratigraphic levels, but characterised by unconformable relationships with underlying rocks. Near Bondoukou, for example, the rocks are stated to overlie both Birimian and granite; dips are low but the rocks have suffered heavy crushing, possibly as a result of thrust movements. 6 A possible compromise In Ghana, the volcanic rocks are interpreted as representing very widespread igneous activity following the accumulation of a thick sedimentary sequence. In Ivory Coast, however, the volcanics are seen as having been erupted along zones of weakness near the borders of depositional basins, so that some greenstones are older than the metasediments, others more or less of the same age, and the Tarkwaian rocks are simply lateral facies variants within Birimian metasediments. Outcrops and communications are generally poor throughout much of the region, and geological mapping is still mainly at reconnaissance level. Foliation of the supracrustals is almost everywhere steep and parallel to original sedimentary layering, and sedimentary structures that would give way-up indications are not always easy to find. Structures are often complex and, where different groups of rocks have been folded together, their relative ages and original relationships are difficult to unravel – unconformities can be obliterated for example, as they have been in this domain. Under these circumstances, geologists working in different regions could well reach conflicting conclusions about the stratigraphy and correlation of rock units. However, it seems likely that there are more fundamental geological reasons for the different interpretations. There are far more Birimian volcanics in the eastern half of the Baoule-Mossi domain than in the west, and there is no a priori reason why they should all be contemporaneous over the whole of this vast region. Volcanism was obviously more sporadic and scattered in the west (Ivory Coast, northern Guinea), and could have broken out at any time. The more voluminous volcanic activity in the east could have begun earlier in the north (eastern Upper Volta, southern Niger) than in the south (Ghana). An explanation along these lines offers a plausible way of reconciling the contrasted Birimian stratigraphies that have been established in different parts of the Proterozoic domain. Granitic rocks Most of the granites in the region fall into two main groups: (a) Large syntectonic batholithic granites have been designated the Cape Coast type in Ghana and the Baoule type in Ivory Coast and Upper Volta. They are 7 generally concordant with regional structures and are often foliated. Many are two-mica granites, though biotite- and hornblende- bearing varieties are also common. Granodioritic compositions predominate, along with K-rich microcline-phyric adamellites. Those larger concordant granitic masses may have metamorphic aureoles where they intrude the supracrustals, but elsewhere they are often migmatitic round their margins and exhibit structural characterisitics of the surrounding rocks. They were probably derived in part at least by remobilisation of older Liberian basement, perhaps in part also by granitisation and partial melting of Birimian metasediments. Pegmatite facies occur throughout these concordant batholiths, sometimes reaching several meters across, and microgranite and aplite veins are also common. (b) Smaller discordant and typically unfoliated late-tectonic to post-tectonic granites, oftern with subcircular outcrops, are designated the Dixcove type in Ghana and the Bondoukou type in Ivory Coast and Upper Volta. They are less abundant that the older syntectonic granites and have a wider compositional range: from hornblende- and biotite-beairng granites to diorites, monzonites and syenites. In Ghana, there is some controversy as to whether any of the Eburnian granitic intrusions are post-Tarkwaian. Granites adjacent to Tarkwaian sediments (as near Konongo in Ghana) appear to have structurally and metamorphically influenced Birimian rocks but not the Tarkwaian. Aplitic veins related to these granites are found cutting Birimian but not Tarkwaian rocks, and granite pebbles occur in Tarkwaian sediments. On the other hand, pegmatites have been found cutting the Tarkwaian in other places. However, there are basic to intermediate intrusive rocks among the Tarkwaian sediments, so there was clearly some igneous activity in progress when they were deposited. Igneous rocks of similar composition and age probably occur elsewhere. The largest basic intrusions in Ghana, for example a norite body about 20 km long near Tumu in the north of the country, could also be a manifestation of late Eburnian magmatism. Some of the dykes referred to earlier also belong to this phase, for in places they are older than post-tectonic granites and have been affected by Eburnian metamorphism. 8 Correlation and Geochronology Because of the different stratigraphic interpretations, correlation throughout the BaouleMossi domain can only be rather general. Over most of the area the Tarkwaian is generally regarded merely as a local variant of the Birimian, which can accordingly be simply subdivided into Upper and Lower. An initial phase of deformation and metamorphism along with emplacement of Baoule-type syntectonic foliated granites has been identified as affecting only Lower Birimian rocks. This is called the Eburnian I phase. Pebbles of mica-schist and igneous rocks in Upper Birimian conglomerates were derived from the erosion of rocks formed in this deformation. The Eburnian II deformation and metamorphism was accompanied by early emplacement of some more syntectonic Baoule-type granities and a later episode of post-tectonic intrusions, including the Bondoukou-type granites. In Ghana, with a three-fold subdivision (Upper and Lower Birimian, and Tarkwaian), this scheme is less easily accommodated. The Eburnian I could be correlated with the unconformities identified locally between Lower and Upper Birimian and Eburnian II would the pre-date the deposition on Tarkwaian sediments. These can in any case be regarded as a molasses facies, eroded from the deformed Birimian supracrustals, and themselves deformed in the waning phases of the Eburnian event. The evidence that most granite emplacement was over by the time the Tarkwaian sediments in Ghana were deposited is certainly consistent with this interpretation. Whether or not the subdivision into Eburnian I and II events can be justified on the available field evidence, it is not yet possible to use geochronological data to define it. Numerous age determinations have been made on a variety of minerals and rocks, including metasediments (mica-schists), metavolcanics (acid lavas), granites and migmatites, by K/Ar, Rb/Sr and U/Pb methods. They provide a consistent pattern of ages for the Eburnian thermotectonic event throughout the area, falling within or close to the range of 2150-1950 Ma. It is likely that the climax of deformation, metamorphism and granite emplacement occurred at around 2100 Ma, i.e. in the later part of Lower Proterozoic time. In terms of the Eburnian I – II, which presumably overprinted most of the structures and other evidence of the earlier Eburnian I event, estimated to have occurred at around 2300 Ma. 9 In Ghana, folding and metamorphism of the Tarkwaian probably occurred after this climax. The sediments contain granite pebbles, but are not themselves intruded by granites. however, a pegmatite cutting Tarkwaian rocks has given an age of 1650 Ma, which is taken as dating the final phase of deformation and igneous activity in this region (it must post-date the basic sills in the Tarkwaian as these were folded and metamorphosed with the sediments). As already noted, there are no undisputed relict Liberian ages recording the presence of a pre-existing basement in this region, except in the transitional zone to the cratonic nucleus, the Sasca domain. Evidence for such a basement remains circumstantial, therefore, and it must be concluded that pre- Birimian rocks were substantially homogenised and had their radioactive ‘clocks’ reset by the Eburnian event. Plate tectonics and the Eburnian event If the transition from Archaean to Proterozoic represents a major change in the patterns of global tectonics, this change is not immediately obvious among the Precambrian terranes of West Africa, where the basement-supracrustal-granite association is remarkably similar throughout, from Archean to Pan African domains. So far as the Lower Proterozoic Birimian region is concerned, the only significant difference from the Archaean lies in the area underlain by supractrustals, which is much greater in the Birimian – though even this could be simply a function of the level of erosion. Otherwise the differences are not great. There are some lithological contrasts, such as the relative proportions of iron-rich and manganese-rich rocks among the metasediments. These apart, the broad distribution of Birimian lithologies approximates to a mirror image of the Archaean supracrustal belts. Greenstone-dominated (Type I) belts characterise the eastern half of the Baoule-Mossi domain, metasediments-dominated (Type II) belts the western half. The regional north to north-east trend is characteristic of most Birimian belts, as it is in the Archaean, and the swing to northwesterly trends in the north of the Archaean nucleus is also seen in the adjacent Birimian belts of northern Guinea. Plate tectonic models have been proposed for Lower Proterozoic terranes in other parts of the world, such as the Canadian Shield, where basement-supracrustal-granite associations 10 can also be identified. The Lower Proterozoic of West Africa is insufficiently known for similar models to be proposed with any confidence, but some general observations are possible. The Sassandra mylonite zone may approximate to a crustal suture along the eastern margin of the cratonic nucleus. At least some of the granitic areas presumably represent areas of older continental crust, within and between which were ensialic and ensimatic volcanosedimentary basins, now represented by the supracrustal belts. The abundance and diversity of Birimian suypracrustal belts and the great size of the Eburnian domain suggests that the Eburnian event was the culmination of many minor collisions involving the aggregation of continental blocks and the closure of numerous back-arc and inter-arc basins. These may have been predominantly ensimatic in the east, judging form the distribution of Type I and II supracrustal belts. If the Birimian continental crust formed by mechanisms of the kind summarised earlier on in the text, one important constraint is the direction of subduction zones marginal to the Archaean nucleus. These must have been directed away from the margin, or the Liberian domain would probably have been more extensively remobilised in the Eburnian. Alternatively, the Birimian supracrustals were mainly deposited in ensialic basins developmed over a large area of continental crust and deformed mainly by differential vertical movements that involved relatively little crustal shortening. Indeed, it is entirely possible that some Birimian belts originated in this way, and geophysical data have been used as a basis for models dominated by rifting. Such a mechanism does not preclude the operation of place tectonics, for the rifted basins could have formed above shallowly inclined subjection zones, as in the Andean belt of western South America. Whatever the tectonic mechanism proves ultimately th be, however, it should be remembered that the present limits of the Archaean and Proterozoic age provinces do not represent the extent of ancient crust in West Africa. There is good evidence that both the Liberian and Eburnian events influenced the crust further east, in the Pan African age provinces. Together with the existence of numerous presumed pre-Birimian relics in the Baoule-Mossi domain, this supports the contention that much of the West African continental crust was in existence at the end of the Liberian, even though the different 11 segments were not necessarily in their present-day position. Another powerful argument for the existence of ancient continental crust in this region is the occurrence of diamonds in Birimian metasediments. The diamonds can only have come from kimberlites emplaced in stable and relatively cool continental crust, and such crust must have been present here about 2500 Ma ago. Economic potential in the Lower Proterozoic rocks of West Africa This region is a more promising metallogenic province than the Archaean cratonic nucleus, principally because of the vast area of supracrustal rocks preserved. Those of major importance are gold, manganese, diamonds and bauxite. Most of the mineralisation is in Type I Birimian greenstones and the Tarkwaian rocks, or in soils and gravels above these formations. It is strongly tectonically controlled, most deposits lying within or perpendicular to the regional structural grain. Gold The major auriferous zones of West Africa are located in Ghana, known as the Gold Coast till independence in 1957. Ti has been estimated that the Gold Coast provided towthirds of Africa’s gold production between the late fifteenth and mid-nineteenth centuries, and Ghana is still a major gold producer. Annual production is around 400000 to 500000 oz (c. 12000-15000 kg), though it was of the order of 10,000-11000 kg in the early 1980s. Total reserves are estimated to exceed 5000 tonnes of exploitable gold, and annual production is planned to reach 50000 kg by 1990, and to rise further thereafter. Most of the primary gold deposits are located along the Lower-Upper Birimian boundary on the west side of the Tarkwa syncline, where three presently active mines are situated (Prestea, Obuasi and Konongo). Other occurrences in similar settings are on the east side of the Tarkwa syncline at its northern and southern ends, and further west in a belt extending from Sewum in the south to beyond Bibiani in the north. Primary gold occurs elsewhere in Ghana too, and in Ivory Coast (e.g. Hire), Upper Volta (e.g. Poura) and in southwestern Mali, both in the main Birimian area and in the Kenieba inlier (Kalana), and there is also gold in eastern Senegal (Sabodala), where the Kenieba inlier extends into it. In Ghana, the major primary gold lodes are associated with 12 persistent and deep-seated shear zones that may have been partly controlled by local unconformities between Lower Birimian phyllites and Upper Birimian greenstones. The country rocks in general comprise metamorphosed carbonaceous and manganifeerous argillites, tuffs and greywackes, along with basic to intermediate igneous rocks. The primary gold occurs in quartz veins and lenticular reefs and also in some of the tuffaceous and argillaceous rocks. It is accompanied by sulphides, especially arsenopyrote but including pyrite and pyrrhotite, chalcopyrite and bornite, and a little galena and sphalerite. There is up to 10% silver in the gold and this helps to pay for refining the ore. Primary gold in the Birimian of other parts of West Africa occurs in generally similar settings to those in Ghana, and significant deposits are reported in Ivory Coast. The gold is probably of syngenetic volcano-exhalative origin, related to greenstone volcanism and associated sedimentation, remobilised during the Eburnian metamorphism to become concentrated and localised along major shears. Although there is no direct evidence of any spatial or temporal relation to granitic intrusions, emplacement of the granites may have provided an additional heat source for mobilisation of the gold into veins. The vertical extent of the primary gold deposits is known to be several hundred meters, but difficulties of mining have meant that exploration below about 100 m has not been extensive. It is clear that this is worth while, as ore in the large Obuasi mine did not become especially rich until depths of 300 m were reached. In Ghana, the greatest activity in gold exploration and mining occurred in the late nineteenth and early twentieth century but very few records were kept. It is virtually impossible to relocate earlier prospects, which might have been abandoned as uneconomic, when poor communications made several ore bodies of various sizes unprofitable. They were rich ores (20-30 ppm Au) and would be highly profitable today, with improved technology and the high price of gold. The deposits need not be large in terms of total tonnage. For example, the reopened Kalana mine in Mali has reserves variously estimated at between 24 and 40 tonnes of gold in ore averaging 30 ppm Au. Initial production is 400 kg per year, rising to 2000 kg annually in the long term. At the Poura mine in Upper Volta, when production ceased in 1966 there were still some 1.5 million tonnes of ore left, averaging around 15 ppm Au 13 and containing some 22 tonnes of gold in total. This mine is also being reactivated. With annual production of the order of 1 tonne of metal, such deposits have substantial lifetimes and exploration may in any case extend their reserves. Prospecting continues for gold in several parts of the Birimian domain. Gold in the Tarkwaian Sedimentary gold is found in several places in the so-called Banket conglomerates near the base of the Tarkwaian, mainly along the eastern margin of the syncline, decreasing rapidly in abundance towards the west. It is associated with heavy minerals, including rutile, zircon and plentiful detrital haematite (often as much as 20%, and up to 60% in small pockets), which makes the Tarkwaian a major feature on aeromagnetic maps, though there is not enough to be economic. The heavy minerals and the gold are concentrated in the more mature and better sorted matrix-poor gravel horizons within the Banket. These gravel layers are not stratigraphically equivalent form place to place, but they are quite extensive, with ore grades remaining consistent over hundreds of meters. Diagenesis during folding of the Tarkwaian has resulted in recrystallisation of haematite and the introduction of some pyrite. The distribution of the gold and sedimentological studies of well preserved features such as graded and current bedding suggest that the sediments represent piedmont fan systems of braided and meandering stream channels, with an upland source lying mainly to the east. Although gold might be expected to occur in similar quantities in the Bui syncline and other occurrences of Tarkwaian rocks (e.g. Bondoukou in Ivory Coast), this has not been found to be the case. Only traces of gold have been discovered, and there are no concentrations of haematitic black sands either. Alluvial (placer) gold Modern stream channels near primary and secondary gold districts in the Birimian and Tarkwaian contain placer gold. The Ofin River system currently provides most of the placer extracted by dredging in Ghana, the main operation being at Dunkwa. In addition there are eluvial deposits, beach sands, terrace deposits and older Pleistocene stream sediments containing alluvial gold. Many such occurrences have been prospected and worked in the past and probably still are. There are notable alluvial 14 concentrations in southeastern Liberia, northeastern Guinea (Siguiri) and southwestern Ivory Coast, on the Sassandra River systems, where the Ity gold deposit is also situated. Manganese In Ghana manganese ores were discovered in 1914 at Nsuta in southern Ghana, southeast of Tarkwa. It was the richest manganese ore discovery in the world at that time. In addition to metallurgical-grade ore, the deposit contained large quantities of unique battery-grade ore called nsutite, which is almost pure MnO2 and could be used in dry cells without processing. There are manganese ores elsewhere in Ghana, but the Nsuta deposit is by far the largest and is the only one exploited. Production in both 1980 and 1981 was over 220 000 tonnes, and was a little over 175000 tonnes in 1983. The ore occurs over a range of low hills in the Nsuta area and the manganese originates mainly from Upper Birimian manganiferous phyllites or their more highly metamorphosed equivalents, gondites (quartz-spessartite rocks). These primary rocks are uneconomic as sources of manganese. Deep weathering on a Tertiary peneplain resulted in supergene enrichment to produce almost pure manganese oxides. These constitute the rich ore and are approaching exhaustion. There appear to be large tonnages of carbonate ore, however, averaging 30-35% manganese, but evidently sufficiently enriched in places to give estimated reserves of some 10 million tonnes of about 48% Mn, which is still economically workable. Manganiferous rocks outcrop frequently along both sides of the Tarkwa syncline and many of the deposits are in Upper Birimian rocks, though some in northwestern Ghana appear to be near the Lower-Upper Birimian boundary, which is also the locus of most of the primary gold mineralisation. The rocks comprise black slates as well as phyllites and gondites, along with basic and acid volcanics, including tuffs. The sediments are often markedly carbonaceous. The oxide ores have nodular or columnar and radiating forms near the surface, grading down into layered ores, where carbonate becomes more important. The main ore minerals are pyrolusite, nsutite, psilomelane and rhodochrosite, along with rhodonite and spessartite, and pyrite is common in the carbonate ores and intercalated argillaceous rocks. 15 The carbonate ores show signs of sedimentary layering, but it is not certain whether they are primary or derived by some form of chemical replacement from the manganiferous phyllites and gondites. The origin of the manganese remains undecided, though the frequent association with volcanic rocks suggests that it may have had a submarine volcano-exhalative origin, formed on an ancient sea bed much in the way that manganese-rich muds form on the floors of present-day oceans. Manganese and gold often occur together in the Ghanaian Birimian, which suggests the possibility that manganese could be used as a pathfinder element in exploration for gold. Very similar manganese ores are found in southern Ivory Coast (Mokta and the Blafagueto hills), northeastern Upper Volta (Tambao) and in southwestern Niger. They were also derived by secondary enrichment from Birimian phyllites, gondites and carbonates (locally associated with rhyolites) during the Tertiary. They occur as residual cappings on low hills and ridges, as at Nsuta. The Tambao deposits have been estimated to contain some 13 million tonnes of 54% oxide ore and a similar amount of carbonate ore. Exploitation of this deposit will require the construction of a rail link to Ouagadougou some 360 km away, and concentrates would then have to be transported a further 1000 km or so, to the port at Abidjan. As a consequence, production may not commence till about 1990. There appear to be no plans at present to exploit other known manganese ores formed from Birimian rocks. Diamonds Alluvial diamonds have been found in Birimian rocks in Ghana and in Ivory Coast. Production comes mainly form industrial dredging, and to a minor extent from smallscale licensed digging of surface eluvial and alluvial placers in the overlying and adjacent soils and sediments. Ghana is the major source of diamonds produced from Birimian rocks, followed by Ivory Coast. Annual production of diamonds in Ghana was about 11.5 million carats during the 1979s, but by 1983 this had fallen to less than 340 000 carats. The diamonds are mostly small industrial stones, though with a significant proportion of gem quality. They come mainly from the Birim field of southeastern Ghana, the largest single diamond-producing area in West Africa. Here the immediate source of the diamonds is a band of Lower Birimian conglomerates (which may be coarse 16 turbidite deposits) about 200 m wide and some 80 km long. Diamonds also occur further west in Ghana and have been reported from southeastern Ivory Coast as well, though these are generally smaller than the Birim diamonds. In the small Bonsa field, about 20 km south-west of Tarkwa, diamonds occur in conglomerates at the base of the Tarkwaian. As the diamonds occur within Birimain and Tarkwaian rocks that are some 2000 Ma old, the source must be even older. No kimberlites have so far been found in Ghana. They may have been destroyed during the Eburnian deformation or metamorphism or they may lie concealed beneath the Voltaian sediments to the north; alternatively they are simply not exposed and so have not been located. The decrease in size of diamonds in the Birimian away from the major Birim field suggests that the source lay in or near this region and the diamonds were transported from it to the other fields, although this would require unusually large-scale transport for a mineral as dense as diamond. Kimberlites occur in other parts of West Africa, notably in Sierra Leone and Liberia, which are important diamond producers, but also in northern parts of Ivory Coast and Ghana, in Guinea and in southwestern Mali. However, these kimberlites cannot have been the source of alluvial diamonds within the Birimian as they are much too young, being mostly of Mesozoic age. Bauxite As with manganese, these deposits owe their formation to tropical weathering during the Tertiary and Quaternary, but as the host rocks are Birimian greenstones and aluminous phyllites and Eburnian granites, they are dealt with here. Although there must be large deposits in several parts of this extensive terrane, it appears tha the only presently exploited bauxite deposits in this age province are in Ghana, where there are plans for considerable expansion of the industry. The open-pit mine at Awaso has an annual production capacity of 300 000 tonnes. It has large reserves of high-grade ore, and production in 1980 was nearly 250 000 tonnes, most of it exported, though this fell to less than 200 000 tonnes in 1981, and by 1983 it was down to only 70 000 tonnes. Major deposits at Kibi, north-west of Accra, are also exploited, and there is a smelter at Tema in the Volta basin, with a capacity of 200 000 tonnes per year. It is intended ultimately to 17 produce year from bauxite to supply the smelter, to which a rolling mill will eventually be added. Iron ores Although on a global scale Proterozoic iron formations are generally thicker and more extensive than those of the Archaean, iron ores are a great deal more widespread in the Archaean uncleus of the West Africa craton than in the Lower Proterozoic Birimian terrane, where their place seems to have been taken by manganiferous sediments. The large Faleme deposit in easternmost Senegal may be of this type, however. It contains some 400 x 1106 tonnes of magnetite ore averaging 45-50% Fe, and 100 x 106 tonnes averaging 62-65% Fe. Such a deposit could support an annual production of some 10 million tonnes for 20 years, but this must await construction of links to the main Bamako-Dakar railway as well as preparation of the mine site. There appears to be a continuation of this deposit across the border in southwesternmost Mali, where some 150 x 106 tonnes of haematite ore have been proved, grading between 36 and 67% Fe. Banded iron formation rocks in southwestern Ivory Coast – for example the Monogaga deposits near Sassandra – have been classed as being of Birimian age, but these deposits occur within the reactivated part of the cratonic nucleus. They are therefore older than Birimian and belong with the Archaean supracrustals. Deposits of Lower Proterozoic iron ores occur in norites and gabbros in various places. They are probably all Magmatic segregations, dominated by titaniferous magnetite, sometimes with significant vanadium enrichment, as in northern Upper Volta. Near Takoradi, in southern Ghana, a deposit of such ores is estimated to be 8 km long and a few hundred meters across, and samples have yielded 55% Fe and 12-22% Ti. An iron and steel smelter is under construction at Takoradi and the aim is ultimately to produce some 150 000 tonnes of steel annually, from 1 million tonnes of ore. Other minerals Small deposits of other minerals and metalliferous ores are found in many places throughout the Birimian terrane, and only a few of these are shown on Figure 4.3. There is a substantial deposit of lateritic nickel and cobalt near Koudougou in Upper Volta, 18 estimated at between 30 and 70 million tonnes grading at 1.5% Ni and 0.1-0.3% Co. similar laterites occur in Ivory Coast and in southwestern Niger, but their size and extent is not known. The Birimian greenstones contain various combinations of copper, lead, sinc, antimony and silver in many places for example, near Poura and the Liptako region of Upper Volta, in southwestern Niger, near Kenieba in southwestern Mali, and in Ghana. Chrysotile asbestos occurs in serpentinites among the greenstones of southeastern Ghana. Molybdenite has been found in Ghana, Upper Volta and southwestern Niger, probably associated with Eburnian granites. Pegmatites associated with granites also yield tin, niobium, tantalum, beryllium and lithium in various proportions, for example in southern Upper Volta, south-east of Bamako in southwestern Mali and near Tera in southwestern Niger. Lithium-bearing (spondumene) pegmatites also occur near Saltpond, about 90 km west of Accra. Tin was mined in Ghana during World War II and tantalite, assaying 60% Ta2O3 in places, has been extracted at Issia in Ivory Coast, from placer deposits derived from pegmatite. Uranium showings have been found close to Kedougou in eastern Senegal, though these may be in younger sediments. Marble deposits have been found in some places, the one at Tiara in Upper Volta being reported to contain some 60 000 m3 of workable reserves. Placer deposits of rutile have been exploited along the coast west of Abidjan in Ivory Coast (Grand-Lathou), presumably derived from mafic and/or pegmatitic rocks in the Birimian, but otherwise generally similar to those occurring in Sierra Leone. As in other Precambrian areas, there is an abundance of materials for constructional and related purposes, including aggregate and crushed rock from granite and migmatite and pegmatitic feldspars (including kaolinise, e.g. near Saltpond in Ghana) for ceramics. In general, however, as the areal proportion of Birimain supracrustals far exceeds that of the other two main age provinces in West Africa, the scope for discovery of valuable minerals is correspondingly greater. THE PAN AFRICAN OF WEST AFRICA - THE EASTERN DOMAIN Introduction In this large eastern Pan African domain, there are many low-grade supracrustal belts whose size and general NNE-SSW trend is similar to that of supracrustals in the cratonic 19 nucleus. However, they are largely confined to a broad belt in the western half of Nigeria, except for scattered outlying ridges of mainly quartzitic rocks nearer to the craton margin, in Benin an southern Togo. The basement has a history of reactivation going back at least to the Liberian and it experienced its last major reactivation in the Pan African. The supracrustals have been strongly deformed, being almost everywhere isoclinally folded with a steep foliation that parallels the trend of the belts. Metamorphism is generally in the greenschist to amphibolite facies. A large area on the eastern side of the craton is overlain by mainly flat-lying sediments of late Proterozoic (Infracambrian) to early Palaeozoic age that occupy the Volta Basin. This major sedimentary basin demonstrates the important distinction between stabilised and reactivated crust. Most of the platform sediments are contemporaneous with deformed and metamorphosed supracrustals of the Buem and Togo Formations. These are thought to lie on Dahomeyan basement in and near the thrust zone which forms the margin of the West African craton (the Togo belt). The Togo belt The Togo belt comprises supracrustal sediments and volcanics of probable late Precambrian to early Phanerozoic age, deformed and partly metamorphosed by powerful northwesterly directed thrusting of Dahomeyan basement rocks onto the West African craton. Their underformed lateral equivalents in the Volta Basin (the Voltaian) will be considered in more detail in Part II. The rocks of the Togo belt are also widely held to be of the same age (i.e. Katangan) as most of the supracrustal belts further east in the main part of the Pan African domain. Closest to the craton is the Buem Formation, also called the Thiele Unit, which forms a band of generally flat country about 15 km across on average, with scattered small hills. It defines the eastern limit of the Volta Basin. The rocks constitute a southestward dipping sequences dominated by clastic sediments, mainly sandstones and siltstones, shales and mudstones being subordinate. There are some massive cherts (silexites of French writers), also limestones, dolomites and sedimentary ironstones. Conglomeratic horizons near the base of the succession have been interpreted as glacial tillites. Volcanics interstratified with sediments in the succession include rocks of both alkaline 20 and calc-alkaline affinities. At least some of them were erupted under water, for pillow structures are preserved. In many places throughout the belt, schistose and massive serpentinites, some of them chromite-bearing, have been tectonically emplaced along the numerous thrust planes that cut the succession. They probably represent slices of upper mantle. There are also cross-cutting dolerites, which may be of late or even post-Pan African age. Rocks of the Buem Formation are largely unmetamorphosed. Deformation is mainly the result of thrust tectonics with much imbrication, the stacking of inclined thrust sheets one upon the other and the overall dip is to the south-east. Two generations of thrust movements have been recognised. Mylonites and cataclastic rocks are common and the thrust zones are often marked by brecciated silexites and serpentinites. There has been much repetition of the succession by the thrusting, and this has combined with generally poor exposures to prevent unequivocal elucidation of the structures. In Ghana, for example, it has been commonly supposed that the Buem Formation is strongly folded into asymmetric overturned structures with southeasterly dipping axial planes. This interpretation places the volcanics near the top of the succession and yields estimates for the total thickness of these largely unmetamorphosed rocks of around 3600 m. An alternative view is that the appearance of strong folding is illusory, due to pinching and swelling of sandstone horizons resulting from rapid lateral facies variations. Where younging directions can be determined the beds are the right way up, and although small chevron folds with steep axial planes are common in fine-grained beds, massive sandstones are not folded. This interpretation would place the volcanic rocks towards the base of the succession rather than at the top, and is in accord with the findings of geologists in Togo and Benin. It also leads to a maximum possible figure for the thickness of the Buem Formation of 40, 000 m, including 5000 m of volcanics, but this makes no allowance for repetition of the sequence by thrusting, and the true thickness must be very much less than this. The Togo Formation immediately to the east of the Buem Formation is called the Akwapimian in Ghana and the Atacorian or Atacora Unit in Togo and Benin. It occupies an irregular 5-50 km wide strip bordered on the west by thrust contacts against 21 the Buem Formation, on the east by thrust contacts with the Dahomeyan basement. It includes the Atacora range in Benin, the Togo Mountains and the Akwapim range in southern Ghana, where the Buem Formation wedges out and the Togo Formation defines the eastern boundary of the Volta Basin. There are two principal lithologies, one psammitic, and the other pelitic. Quartzitic sandstones and quartzites contain conglomeratic layers, and some quartzites are ferruginous. Phyllites and mica-schists, including the Kande-Boukombe Series, also have conglomerate horizons, as well as basltic volcanics, now metamorphosed to greenschists. Marble is recorded north of Boukombe. Tectonically emplaced slices of basement rocks are mainly gneisses but include eclogitic and granulite facies rocks of the nearby high-grade Dahomeyan, as well as elongate lenses and pods of serpentinite. The Togo Formation rocks are not only more highly metamorphosed than those of the Buem Formation, they are also considerably more deformed. Both metamorphism and deformation increase towards the south-east. Minor folds are isoclinal with axial planes that dip to the south-east and are generally parallel to lithological layering. Major folds are upright, with only slight tendency to a southeasterly dip. The thrust movements have been correlated with these folds, though the thrust planes appear themselves to have been folded by a later phase of deformation. The relative age of the Buem and Togo Formations remains a matter of speculation and debate. They may be lateral equivalents, but there seems to be growing agreement that the Togo Formation is the older. At first sight this is at variance with the relationships in, which shows that in this eastward-dipping belt the Togo Formation overlies the Buem, which implies that it should therefore be younger. The Togo Formation was probably unconformable on the Dahomeyan, however, and was brought up from deeper stratigraphic levels by the thrust faulting, to lie upon the younger Buem rocks. However, some comments on relationships between it and the Togo belt are appropriate here. In brief, the Togo Formation is correlated with the Lower Voltaian, partly on the grounds of similar lithologies, both groups of rocks being dominated by alternations of sandstones and shales (quartzites and phyllites or schists), with occasional conglomerates and limestones (marbles). In addition, there is evidence that in southeastern Ghana, where the Lower Voltaian of the Kwahu Plateau abuts against the Togo Formation of the 22 Akwapim range, there is a progressive increase in intensity of folding and metamorphism eastwards from Lower Voltaian into Togo Formation rocks. Similar relationships obtain at the northern end of the Togo belt, where the virtually undeformed Kirtachi Quartzite represents the Lower Voltaian in southwestern Niger. The Middle Voltain lies unconformably on the Lower Voltaian with a basal conglomerate that can be correlated with conglomerates of the Buem Formation. Both sets of conglomerates have been interpreted as tillites, possibly equivalent to the Infracambrian tillites of the Rokelide-Mauritanide belt west of the craton and in the Taoudeni Basin (Part II). The overlying sediments of the Middle Voltaian include shales, siltstones and sandstones, and limestones, dolomites and cherts, lithologies similar to those of the Buem. There are no volcanic rocks, but among the arenaceous facies are greywackes of possible volcanogenic origin. The Upper Voltaian has no equivalent in the Togo belt, and it is believed to be a molasses deposit formed by erosion of Buem and Togo Formation rocks, following their deformation and uplift in the Pan African event. Interpretations of gravity data across the Togo belt suggest that the deep structure here approximates to a mirror image of that on the west side of the craton. The eastern domain The eastern domain of the Pan African rocks is only a small part of the huge expanse of country that lies between the West African and Congo cratons and is underlain by rocks affected by the Pan African thermotectonic event. Much of this great region probably consists of reactivated older crust, for while Pan African ages (in the 650-450 Ma range) are recorded from rocks throughout it, both Liberian and Eburnean areas have been obtained from several localities. The Pan African event was the last reactivation to affect this whole region, which is now as tectonically stable as the craton itself and forms an integral par of the African shield. The basement complex There seem to be no fundamental differences between the basement in this region and in other parts of West Africa. Granulite facies rocks are most abundant close to the margin 23 of the craton, immediately to the east of the Togo belt, in south-eastern Ghana and in Togo and Benin. Gneisses with garnet, pyroxene and scapolite occur among more ordinary quartzo-feldspathic biotite and hornblende-bearing varieties. Eclogites (highpressure garnet-pyroxene rocks chemically equivalent to basalt) have been recorded from among large masses of mafic gneisses that include amphibolites and pyroxenites and contain much garnet. The high grade rocks are generally considered to have been brought up from deeper crustal (and upper mantel) levels by the westward thrust movements that gave rise to the deformation of the Togo belt. Further east, in Nigeria, granulite facies rocks in the basement are confined to charnockite bodies, which are generally associated with granites and are probably of igneous origin. Here the basement is dominated by quartzo-feldspathic biotite- and hornblende-bearing gneisses, schists and migmatites. Metamorphism is generally in the amphibolite facies, as indicated by the occurrence of index minerals such as garnet, sillimanite, kyanite and staurolite in rocks of suitable composition. Intercalated among the gneisses and migmatites are numerous supracrustal relics. They have been collectively termed the Older Metasediments in Nigeria – and they are likely to include remains of supracrustal belts of the Liberian and Eburnian cycles. Most obvious are prominent ridge-formaing quartzites, sometimes sillimanite-or kaynite-bearing and often micaceous, grading into muscovite-quartz-schists. They may be structurally complex and this helps to distinguish them from the more simply folded quartzites of the Younger Metasediments in the low-grade supracrustal belts, which are also generally free of sillimanite or kyanite. There is a metamorphosed banded iron formation, rich in magnetite and haematite, south-east of Kabba in southern Nigeria, where it is an important source of iron ore; smaller occurrences of such rocks are known from the basement in northwestern Nigeria. Also among the Older Metasediment relics are lenses of more or less dolomitic marble, especially in the vicinity of Kabba and Jakura in southern Nigeria (where some of them are quarried, and numerous small thin sheets and lenses of calc-silicates that are probably metamorphosed impure limestones. Amphibolites are widespread throughout the basement, ranging in size from small disrupted remnants in the gneisses and migmatites up to larger layers and lenses a few 24 hundred meters long. They probably represent igneous intrusives and volcanics of basic to intermediate composition. Small masses of ultramafic rocks (mostly serpentinite and talc rocks) have also been recorded; at least some of them probably emplaced tectonically along deep fracture zones during basement reactivation. As in other parts of the West African Precambrian, structures in the basement appear relatively simple on a regional scale. The often steeply dipping foliation is parallel to litholigical layering almost everywhere, and the overall structural trend is NNE-SSW over wide areas, parallel to that of the low-grade supracrustal belts. Local diversions from the regional trend occur in several places, and in eastern parts of Nigeria, especially north of the Benue Trough, the regional trend swings to more nearly ENE-WSW. Poor exposures mean that minor structures are not commonly seen and they are difficult to interpret, especially those produced by plastic flow of semi-molten rock. The supracrustals In their overall aspect, the numerous low-grade supracrustal belts are similar to those elsewhere in West Africa: they are synclinorial, the rocks are characterised by tight to isoclinal folding and steeply dipping foliation, and boundaries with the basement (where exposed) are gradational, faulted or sheared. No unconformities have yet been positively identified. However, they are mainly confined to a broad NNE-SSW zone in the western half of Nigeria, where they are referred to as the Younger Metasediments. This reflects their principal difference from supracrustal belts on the craton; there are few volcanic rocks among them. The Katangan belts are dominated by argillaceous (politic) lithologies, represented by phyllites, muscovite-schists and biotite-schists. Quartzites are important in several belts, where they form prominent strike ridges. A few belts have ferruginous and banded quartzites that resemble banded iron formation rocks, and small occurrences of spessartite-bearing quartzite have also been found. Conglomeratic horizons occur in some of the belts, and marbles and calc-silicates are not uncommon. Other lithologies among the Younger Metasediments represent minor volumes of igneous rocks: amphibolites were probably contemporaneous lavas or minor intrusions, while serpentinites and other ultramafic rocks were probably emplaced tectonically, along deep 25 fractures during deformation of the supracrustal belts. There are also small amounts of acid metavolcanics of dacitic to rhyolitic composition. Metamorphic grades are variable among the supracrustals. Amphibolite facies rocks are commoner than greenschist facies rocks in the southern belts, where staurolite and almandine-garnet are common constituents of the schists. These index minerals are also found in the northern sector, but more rarely, and most of the supractrustal phyllites here are in the greenshist facies. Although it is possible to group the supracrustal belts on the basis of common lithologies, such a grouping can have no implicatiosn about their relative ages. As on the craton, it is diffcult to establish correlations between separate supracrustal belts. The quartzites and mica-schists that form concordant lenses in the basement of Benin (the Badagba Quartzites) may be older metasedimentary relics, but it is more likely that they are infolded outliers of the Togo Formation and therefore of much the same age as the Younger Metasediments. They are isoclinally folded and often have strongly sheared contacts with adjacent Dahomeyan rocks. Major tight to isoclinal folds are often traced out by prominent ridges of resistant rocks among the Younger Metasediments, especially quartzites, and these show up well on aerial photographs. When the closures of such folds are examined in the field, they are invariably obscured by intense shearing parallel to the axial plane foliation. Minor structures are relatively abundant and well exposed in the supracrustal belts, however, and analysis of these has revealed as many as four phases of deformation in some belts. Evidence of an early phase of thrusting and recumbent folding has been found in parts of Nigeria. It has been all but obliterated by the later NNE-SSW structures that characterise the supracrustal belts. However, abundant quartzites in part of the most northwesterly belt, near Zuru, are unusual in that they display a predominance of east-west trending structures with shallow dips. These may represent an earlier folding episode in the supracrustal belts, preserved because of the greater structural competence of quartzites compared with that of phyllites and schists. The Anka belt in northwestern Nigeria stands out from all of the others in having fluviatile sequences of conglomerates, sandstones and siltstones, as well as acid volcanics and minor intrusives that have largely escaped the main effects of Pan African 26 deformation and metamorphism, except along relatively narrow shear zones, where there was tight folding, with stretching and flattening of pebbles. Otherwise, the rocks are mostly only gently folded and sedimentary strucrutes and igneous textures are commonly well preserved in them. It would seem that the deformation was largely fault-controlled as in parts of the Rokelide belt relationships with the phyllites which make up most of the Anka belt are not known because of poor exposure. However, the conglomerates contain fragments of phyllite, boulders of quartzite and blocks of granite that are over a metre across. The late-to post-orogenic volcanic and intrusive acid rocks of the Anka belt are mainly dacitic and rhyolitic lavas and pyroclastics, and dykes and plugs of microtonalite. Particularly striking is a calderalike structure near Maradun, in which beded agglomerates and tuffs, some of them waterlaid, have a concentric inward dip. A relatively undeformed and unmetamorphosed volcanosedimentary series comprising spilitic greenstones with pillow structures, acid lavas, sandstones, greywackes and conglomerates with pebbles of besement and volcanic rocks has been described from just to the northeast of Badagba in southern Benin, and appears to be generally similar to the Anka belt. In the north is the Dahomeyan inlier of Zinder, where basement gneisses and migmatites occur along with supracrustal schists and quartzites. Scattered outcrops of shists and quartzites are also known for the Daura area near the Nigeria-Niger border north of Kano, and these provide evidence of a generally NNE-ward continuation of Younger Metasediments beneath younger sediments of the Chad Basin. South-east of the Benue Trough, the highlands of the Mambilla Plateau extend into Cameroun and are formed mainly of basement gneisses and granitic intrustions. The northwestern half of Cameroun is part of the Pan African domain of West Africa, while the southeastern half belongs to the Congo craton, which is of Liberian age in this region, though there is evidence of both Eburnian and Pan African reactivation along the margin. Supracrustal belts of schists and quartzites in northern and eastern Cameroun and western Chad are assigned by some authorities to the Pan African, by others to the craton. Granites 27 There are many syntectonic to late-tectonic intrusions, mainly rocks of granitic compositions, but including diorites and syenites. They were emplaced into both basement and supracrustals during or just after the main Pan African deformation. In Nigeria they are called the Old Granites to distinguish them from the Jurassic Younger Granites of the Jos Plateau (Part III). They range in size from small subcircular crosscutting stocks to large elongate concordant batholithic bodies, predominantly of granodioritic composition. A distinctive adamellitic variety with microcline megacrysts up to 5 cm and more in size is called the Porphyritic Older Granite in Nigeria. The smaller discordant intrusions are more variable in composition, as they are in the Eburnean terrane. They include the dioritic and syenitic varieties, though most are granites, and there are some unusual types, such as a potash- and magnesia-rich basic syenites associated with pyroxenites, found in south-western Nigeria. Occasional gabbros also occur, one of the largest being in the north-west, south of Zuru. Among these late- to post-tectonic intrusions, there are ring complexes not unlike those that typify the Jurassic Younger Granites (Part III). Good examples are near Toro (granite and diorite) and south of Maru (granite and syenites). The intrusions are most abundant in the vicinity of the supracrustal belts and in the central part of Nigeria. This is the result of geological mapping, which has historically been concentrated in these regions for economic reasons. More recent mapping is filling some of the gaps. Charnockites (pyroxene-bearing acid to intermediate igneous rocks) are also associated with Older Granites in some parts of the basement, notably in two main areas: in southern Nigeria, east of Ibadan, and in the north-east, where there is a distinctive fayalite-bearing variety called bauchite, named after the type locality, Bauchi. The mineralogy and petrology of these rocks suggest an origin as high-temperature anhydrous melts deep in the crust, prior to formation of the Older Granites with which they are associated. Late- to post-tectonic dykes of basalt and dolerite have been recorded from many parts of the Dahomeyan basement, in places as far removed as southeastern Ghana and north-east Nigeria. Some bear evidence of incipient metamorphism and deformation and are undoubtedly late Pan African, but radiometric data confirm that there are unaltered dykes of this age as well. However, there are also dykes related to the Younger Granites (Part III) and it is not always possible to tell which is which. 28 Fractures, fault and mylonites Major shear zones are marked by discontinuous ridges of mylonitised and silicified rocks and lenses of vein quartz. An early phase of shearing and silicification developed virtually parallel to foliation, most commonly between basement and supracrustal belts. Later transcurrent faulting was mostly at a small angle to foliation, trending between NNE-SSW and NE-SW. Dextral movements amounting to a few tens of kilometres have been recoreded in a few places – the best example is near the Anka belt. Regional studies of fracture systems in the Pan African domain of both southern West Africa and the Hoggar region have revealed a conjugate pattern of NE-SW dextral and NW-SE sinistral faults, which cut earlier mylonitic shear zones. The pattern is consistent with an overall east-west directed stress system. Two long NNE trending zones of mylonitised basement rocks, several kilometres across, extend north from Zungeru on either side of the Birnin Gwari belt. These are the Zungeru mylonites, which are thought to have originated as a major thrust between basement and supracrustals. The mylonites have been recrystallised and folded at least four times, along with the adjacent supracrustals. It must be recorded, however, that these mylonitic rocks were originally mapped as metasediments and metavolcanics, and that some workers still support this identification. Further south, gently dipping quartzites of the Effon Psammite Formation (part of the supracrustal belt near Ilesha, may be in thrust-faulted contact with basement gneisses. Correlation and geochronology Relict Eburnean and Liberian ages obtained from a number of localities in the Dahomeyan basement, by Rb/Sr and U/Pb methods, provide further evidence of its long history and polycyclic evolution. Apart from these relict ages, most Rb/Sr and all K/Ar age determinations on basement, supracrustal (including those of the Togo belt) and Older Granites, yield Pan African dates in the 650-450 Ma range. The simple conclusion is that deformation and metamorphism of the supracrustals occurred during the pan 29 African thermotectonic event, which also reactivated large tracts of older continental crust. The structural and metamorphic uniformity among the supracrustal belts of Younger Metasediments throughout the region supports such an interpretation, allowing for rather higher metamorphic grades in the south of Nigeria compared with the north. However, Kibaran ages (c. 1100 Ma) have been obtained by Rb/Sr techniques at a few localities in the basement, and from supracrustals of the Maru belt. At least two Older Granites have given ages of around 750 Ma. There is therefore some geochronological evidence for a Kibaran thermotectonic event affecting the Dahomeyan basement and possibly an older generation of supracrustals. Earlier structural studies in some supracrustal belts led to a similar conclusion, but they have not been fully corroborated by later research. It may also be relevant that marine transgressions into the Volta and Taoudeni Basins commenced about 1000 Ma ago. These transgressions could have been triggered by an orogenic event east of the craton. Whatever status is ultimately accorded to the Kibaran event in this domain, however, there is no doubt that it was vastly overshadowed by the Pan African, the waning stages of which appear to have been marked by the intrusion of basalt and dolerite dykes. One of these gave an age of about 480 Ma, which probably dates the time of emplacement, for the rock is unmetamorphosed. Correlation with the Hoggar North of the area covered in Figure 6.1 the Pan African domain disappears beneath the Mesozoic and younger sediments of the Ilullmedden and Chad Basins. It reappears nearly a thousand kilometres to the north, in the Hoggar massif of southern Algeria, with its two southward projections, the Air of northern Niger and the Adrar des Iforas of northeastern Mali, along with the eastern part of the Gourma. The rocks of this region are generally better exposed than in the south and they have been the subject of intensive research over many years. There is a correlation between the northern and southern parts of the eastern Pan African domain in West Africa. The northern part is sometimes called the Tuareg shield. The most obvious feature is that the Tuareg shield can be divided into three main blocks, 30 separated and cut by major faults and shear zones. Only the central Hoggar-Air segment can properly be correlated with the southern Pan African domain, which it resembles in several respects. Also called the Suggarian belt, It is a polycyclinc terrane, characterised by elongate low-grade metasedimentary belts of probable Upper Proterozoic age in an older (at least Eburnian) gneiss-migmatite basement, the whole reactivated and intruded by granites around 650-600 Ma ago. Kibaran ages (c. 1100 Ma) have also been recorded, though (as in Nigeria) their significance is not clear. There are mainly acid metavolcanics among the metasediments, which are considered to have accumulated in ensialic troughs, the boundaries of which are in part represented by major N-S shear zones. To the west lies the Pharusian belt, comprising two N-S zones of late Proterozoic rocks deformed and metamorphosed in the Pan African. They are separated by a narrow elongate block of cratonic crust, which has granulite facies rocks in it and may be as old as Archaean (c. 3000 Ma). This is the In Ouzzal block, now bounded by faults and shears zones, and believed to be a thrust block of lower curst. In the western Pharusian zone there are remnants of an older platform succession including quartzit4es and stromatolitic limestones, which are comparable to those found in the Taoudeni and Volta Basins (Part II). However, the main Pharusian succession comprises volvanci and volcanosedimentary rocks that include andesites and dacites, greywackes, turbidites, conglomerates and politic facies. These rocks are all of Upper Proterozoic age and were deformed in the Pan African (c. 620 Ma). They were metamorphosed to greenschist and amphibolite grades, with the development of giant Alpine-type nappes and thrust sheets, the overall direction of transport being westward, towards the craton. There are ultramafic masses along the cratonic boundary, which is also associated with positive gravity anomalies. Considerable syntectonic to late-tectonic granitic plutonism occurred in the Pharusian belt (650-530 Ma) and some major N-S shear zones developed. The ‘serie pourpree’ of the Pharusian belt is regarded as a molassic deposit occupying late-stage graben, and is at least partly of Lower Cambrian age (the ‘greenstones’ of the Pharusian belt are sometimes called the ‘serie verte’). South-west of the main Pharusian belt is the Gourma region, in the east of which is another zone of Pan African deformation and metamorphism. This is a complex region, 31 for the great embayment into the West African craton appears to be related to the development of a large rifted basin extending well into the craton. The overall WSWENE trend of the Gourma Trough is marked by positive gravity anomalies which provide some of the evidence for the two branches. The shape of the basin is further defined by the distribution of sediments within it. The total thickness of those sediments is up to 8000 m, thnning southwestward, where the trough narrows and shallows into and on to the craton. Conglomerates and sandstones at the base area are overlain by silts and shales, and then by carbonates, including stromtolitic facies and resembling those of the lower part of the Taoudeni Basin succession (Part II). A phase of clastic sedimentation followed, prograding towards the north-east and filling the trough. The sediments at the deeper northeastern end of the graben were powerfully affected by the Pan African orogeny, with the formation of high-grade metamorphic rocks and southwesterly directed nappes and thrust sheets. An Eburnian inlier at Bourre is interpreted as a horst block, a large cratonic slice thrust upwards from deeper levels during the deformation. It consists of amphibolite facies Birimian metasediments intruded by anEburnian granite (c. 2080 Ma). Ultramafic bodies which may be ophiolite fragments are encountered in this region, as they are further north. Deformation and metamorphism are much diminished southeast of the thrust zone, that is where the sediments rest upon cratonic basement, and here they are overlain unconformably by late Infracambrian sandstones. These are not a molasses deposit, however, for they are part of the main Taoudeni Basin succession and show evidence of northeastward progradation (Part II). East of the central Hoggar-Air block is an apparently older terrane, with largely undeformed and unmetamorphosed platform sediments and molasses deposits, the whole intruded by granites. Relatively little appears to be known about this region, but it is identified as part of an East Saharan craton. It may be substantially younger than the West African craton, however, for sparse age data suggest that it was not stabilised until about 730 Ma ago, only a short time before the main phase of Pan African activity. Pan African plate tectonics east of the craton The Pan African event in the Tuareg shield is seen by the majority of those who have worked there as a fully developed continental collision orogeney involving the West 32 African and East Saharan cratons, and on a scale comparable with that of the Himalayan mountain chain. The Pharusian belts are believed to represent the site of a late Proterozoic ocean basin, in which deep-water sediments accumulated off the margin of the West African craton, and in which volcanic island arcs developed. The In Ouzzal block was presumably originally one or more microcontinental fragments in this ocean. The Gourma Basin is interpreted as an aulacogen, an intracontinental rift extending inland from the continental margin, i.e. southwestward from the margin of the craton. The time of its initiation is place at 850-800 Ma ago, as the late Proterozic ocean basin east of the craton. The evidence for a major continental collision event in the Pan African is less overwhelming in the Dahomeyan domain. Nonetheless, the whole of the eastern craton boundary is marked by a zone of positive gravity anomalies, and ophiolite fragments have been identified at intervals along it. In addition, the regional fracture pattern (sec. 6.3.4) is consistent with a plate collision model for the whole of this eastern Pan African domain, and it seems logical to identify the Togo belt as a southern extensing of the Pharusian belt. It is also possible that the Togo-Benin-Nigeria swell, and perhaps the central Hoggar-Air block as well, represents aggregates of continental fragments, along with island arcs and inter-arc basins, some ensialic, others ensimatic, swept together by subduction and collision to form a coherent continental block accreted on to the West African craton margin. Such a model could perhaps account for a Pre-Pan African (Kibaran) episode of deformation, metamorphism and reactivation. The relative merits of different interpretations of the Pan African event in West Africa can be endlessly debated, but do not justify further discussion here. Economic potential of the eastern Pan African domain The Pan African terranes in West Africa contain fewer large mineral deposits than the older Birimian and Archaen rocks of the craton. There seem to be two main factors responsible for this. First, the proportion of mafic to ultramafic rocks among the supracrustals is relatively small. Secondly, the polycyclic nature of the basement has resulted in large tracts becoming barren of useful mineralisation, although the Pan African reactivation did cause enrichment of some elements in a few places. In general, there is a considerable variety and abundance of relatively minor mineral deposits that 33 offer scope for the operation of small-scale mining enterprises, producing minerals, both to fill local needs and for export. Gold was extracted for many years from several parts of the supracrustal belts in the western half of Nigeria. Recorded total production from the Nigerian goldfields is about 300 000 oz (c. 9000 kg) since the early 1920s, but the real total must be much more than this. At the height of mining activity in the early 1930s, annual production was of the order of 30 000-40 000 oz, which is less than a tenth of the annual production from the Ghana goldfields. Gold mining in Nigeria declined during the after World War II and remains at a low level. Most of the gold production was from alluvial and eluvial placers in river systems draining supracrustal belts in three main areas: around Maru and Anka in the north-west, extending southward to beyond Zuru; in the belts west of Kaduna; and in the south-west in the Ife and Ilesha region. The primary gold is in the form of veins, stringers, lenses, reefs and similar bodies of quartz, quartz-feldspar and quartz-tourmaline rock in both supracrustals and basement. The veins range in thickness from several centimetres to a few meters, and in length up to several hundred meters, often displaying lenticular or pinch-and-swell (boudinage) structure. They are invariably steeply inclined and may occur as isolated bodies or as parallel or en echelon vein systems. The veins are commonly associated with fractures and shear zones, and are mainly concordant with regional foliation trends, but often also cross cutting. There is no systematic correlation between gold mineralisation and the type of country rock, no obvious association of gold-bearing veins with Older Granites or particular lithologies among the supracrustal rocks. However, it is noteworthy that the main areas of gold occur near major faults and shear zones. In addition, a very crude spatial zonation has been recognised: in the south the veins tend to be of pegmatitic type (notably the Iperiando vein near Ilesha), further north are quartz veins carrying sulphides along with the gold (mainly galena, but also variable amounts of pyrite, arsenopyrites, and possibly chalcopyrite), and further north still, simple gold-quartz veins predominate. The primary gold veins are all small and were mostly worked at the surface, only a few being large enough to justify underground mining. Only relatively few of these primary gold deposits were found, however, and it is likely that a high proportion of the placer 34 gold has come from small disseminated pods and veinlets with low gold contents that became concentrated by alluvial processes. Gold is also known from a few places in the Togo belt, associated with the thrust zones. The largest deposit is at Perma in northern Benin, where there has been some mining of alluvial and eluvial concentrations. These were last mined in 1956, but recoverable reserves are estimated at 3000 kg, and production may start again in the near future. A broad belt of tin-tantalum-niobium-bearing pegmatites extends northeastwards for some 400 km from near Ife to the Wamba-Jemaa area just south-west of the Jos Plateau. They are emplaced more or less conformably into a variety of rock types including Younger Metasedimentary mica-schists, quartzites, amphibolites and calc-silicate rocks, as well as basement gneisses and migmatites with their Older Metasediment relics. Many of the mineralised pegmatites are massive bodies that have pronounced pinch-andswell (boudinage) structures, the swellings being loci of intense albitisation of the feldspars and commonly of rich mineralisation. Most pegmatites in the Nigerian basement are dominated by quartz and microcline, commonly accompanied by varying proportions of oligoclase, biotite, muscovite and tourmaline, and they are barren of economic minerals. In general, it is only where late-stage Na-rich hydrothermal solutions have produced either secondary albitisations of the feldspars or quartz-mica greisens, that mineralisation can be expected. In addition to the main economic minerals, cassiterite, columbite and tantalite, there is a host of accessories, including scheelite, beryl, apatite, monazite, the Li-rich mica lepidolite, black and pink-green tourmaline and gem quality blue gahnite (Zn-rich spinel). Field relations indicate that emplacement of the pegmatites was part of the Pan African Older Granite cycle, but it must be emphasised that there is commonly no obvious spatial relationship with granitic intrusions. This is borne out by age determinations. Rb/Sr measurements on pegmatites give ages of around 550 Ma, but nearby granites can be as old as 650 Ma, that is, up to 100 Ma older. Isotopic data indicate a considerable crustal involvement in the generation of these rocks. K/Ar measurements on the micas of some pegmatites have given ages as young as Mesozoic in some cases. This has been ascribed to a reheating episode during emplacement of the Younger Granites. It is noteworthy that the pegmatite belt cuts at a high angle across the regional grain of 35 the Pan African rocks, even though individual bodies are more or less conformable with it. The reasons for this distribution pattern are still not clear, but it does appear to be real and not an artefact of geological mapping. Mineralised pegmatites have been found outside the belt, but they are small and not especially rich. It may be merely a coincidence that the trend of the belt is parallel to that of the Benue Trough. Most of the ores are taken from eluvial and alluvial placers, but the pegmatites themselves are also worked. The pegmatites provide nearly all of the tantalum production in Nigeria (never more than about 20 tonnes annually, and in general averaging about 5 tonnes per year), but only a small proportion of the tin and niobium, the remainder being produced from the Younger Granites of the Jos Plateau (Part III) The possibility that vein-type uranium mineralisation is associated with basement rocks and Older Granites has stimulated exploration from time to time, but so far with little success. The discovery of such deposits in the Poli area of northern Cameroun, east of the Mambilla Plateau, however, suggests that others may be found; uranium showings have also been reported from the Tessalit region of the western Adrar des Iforas in Mali, and from some localities in Nigeria. Iron ores are the only other metallic minerals that can be presently considered of economic interest in this age province. In some supracrustal belts (e.g. the Maru belt) there are banded quartz—haematite rocks, sometimes magnetite-bearing, associated with garnet—grunerite—schists, amphibolites, phyllites and quartzites. They resemble the banded iron formations (itabirites) of the Archaean terrane, but are much smaller and leaner, with average grades not exceeding 40% Fe at best. A certain amount of supergene secondary enrichment has occurred in the surface weathering zone, but this is not deep, and the deposits cannot be regarded as a promising economic prospect. Similar considerations may apply to deposits of banded iron formation type in rocks of both the Buem and Togo Formations in the northern part of the Togo belt. A deposit estimated to contain some 95 million tonnes of iron ore of unspecified grade has been reported near Bandjeli in the west. Further east, in the region round Dako, magnetite— haematite (—limonite) ores grading 40—45% Fe are associated with quartzites and micaschists, some of which contain chlorite, garnet and epidote. There are several separate deposits each containing some hundreds of thousands of tonnes of ore. 36 However, there are richer and purer itabirite-type ores interbedded among basement gneisses and forming prominent ridges near Okene (south-east of Kabba) in southern Nigeria. These are probably Older Metasediment relics and they could be as old as Archaean (and perhaps can be correlated with the iron ores of Liberia and Guinea. There are four such ridges, of which Itakpe Hill, some 15 km north-east of Okene, has been investigated in most detail. It has the form of an isoclinal fold oriented WNW—ESE and closing to the south-east, with an overall southerly dip. The ores are massive, banded (gneissic or schistose), depending partly on composition and partly on the degree of recrystallisation that the rocks have experienced. The main ore minerals are magnetite and haematite. At least 150 million tonnes of 35—50% Fe have so far been proved. The rocks are very quartz-rich in some places, but undesirable impurities such as S and P are absent. These ores will provide raw materials for the Nigerian iron and steel industry, which started production early in 1982, at Aladja, using natural gas as fuel. There is also a major steel complex at Ajaokuta (on the Niger River) and others are being built. Two rolling mills were commissioned in 1983. Initial planned capacity is of the order of 1 million tonnes per year, rising eventually to around 5 million, using ores from Liberia and Brazil as well as from Okene. Output in 1983 totalled over 700000 tonnes of iron and steel products. Magnetite—quartzites in the Kribi region of southern Cameroun resemble the Okene ores. Grades are around 40% Fe, though secondary haematite enrichment has raised this to 70% in places. Deposits of similar type, though smaller, are believed to occur in other areas of the Pan African basement. Manganese does not occur in any quantity in the Pan African domain. Ten million tonnes of ore have been reported near Bayega in Togo, and in Nigeria some very small lenses of gondite-type ore have been found, with limited supergene enrichment. There must be other deposits but probably nothing on the scale encountered in Birimian rocks. Bauxite is another mineral that is so far not known from the Pan African terrane, except for a small deposit at Mt Agou in Togo, where it appears to be associated with a charnockitic complex. Rutile is found as alluvial placers in and near the Togo belt and occasional small concentrations have formed in other basement localities, probably derived from 37 otherwise barren pegmatites. So far as other ferrous and base metals are concerned, the overall prospects in the Pan African domain are generally similar to those in the older rocks of the craton. There are many small and scattered concentrations, but none of significant size has so far been discovered. Small chromite deposits occur among the ultramafic masses associated with the Togo belt and at the adjacent Dahorneyan basement, brought up from deeper levels by thrusting. In Nigeria minor concentrations of chromite occur in serpentinites in supracrustal belts and emplaced along shear zones in the basement. Such bodies might be more favourable as sources of asbestos, talc and magnesite, occurrences of which have been found in a few places, though so far only in small amounts and generally of poor quality. Nickel anomalies are associated with these ultramafic masses, but the prospects for significant concentrations of this metal are not promising. An occurrence of copper and vanadium is recorded at the northern end of the Togo belt, associated with basic rocks, and copper showings are reported in south-west Benin. Otherwise, these metals, along with lead and zinc, are virtually absent from the Pan African terrane. Small amounts of chalcopyrite occur in some amphibolites and altered gabbros and there is galena in some gold-bearing veins, but all these showings are small and disseminated. In addition to asbestos, talc and magnesite that have afready been mentioned, industrial minerals occur in various places. Kyanite is known from at least one locality, and there may also be concentrations of sillimanite, especially among the higher- grade aluminous schists of the southern belts in Nigeria. Mica is extracted from some of the pegmatites and coarser micaceous schists. In a few places, granitic rocks have been altered by hydrothermal fluids forming deposits of almost pure white kaolinire, the largest of which so far found is at Kankara and has been extensively worked. There are numerous occurrences of marble, mainly Older Metasediment relics in the basement and mostly in southern parts of Nigeria, but also in Togo and Benin. The deposit at Jakura is exploited partly for the construction industry, partly to provide fluxing material for the iron and steel complexes. Some of the marbles are fairly pure and 38 calcitic, but others are dolomitic and unsuitable for cement manufacture. As elsewhere, an abundance of crushed rock and aggregate for construction purposes is available throughout the basement areas, particularly in places where the superficial weathering and alluvial deposits are thin and bedrock comes to the surface, notably around inselbergs. The Precambrian of West Africa — synthesis and review The geological picture It is worth reiterating a number of generalisations that appear to be valid for the whole of the Precambrian terrane in West Africa, irrespective of age. They apply to rocks that span some 2000 Ma. (a) There is a tripartite division into basement, supracrustals and granitic and related intrusions. (b) Overall structural trends fluctuate about a generally NNE—SSW direction. Structures are commonly upright, with steep dips and tight to isoclinal folding. Flat-lying structures are found in places, but they do not appear to be common. (c) The relative proportions of areas underlain by supracrustals and basement are smaller in Archaean and Pan African domains than in the Lower Proterozoic (Birimian) domain. The proportion of different supracrustal lithologies varies both within and between the different domains: volcanics predominate in the western Archaean and eastern Birimian provinces, whereas sediments prevail elsewhere. (d) Metamorphic grades are generally similar throughout the Precambrian, with amphibolite to granulite facies rocks characterising the basement, greenschist to amphibolite facies more typical of the supracrustals. There are often steep metamorphic gradients between basement gneisses and migmatites and supracrustal belts of schists and phyllites. (e) There is no conclusive evidence of any signilicant difference in the volumes of syntectonic to late-tectonic intrusions (mainly granitic) emplaced during each major thermotectonic event, between one age province and another. Nor are there obvious signs of compositional trends in space or time, either within or between the different age provinces. 39 (f) Major fractures and shear zones, many with an overall NNE—SSW trend, appear to be a feature of the whole region. They may represent deep-seated lines of weakness, repeatedly rejuvenated during successive thermotectonic events. As the rocks have similar metamorphic grades, they formed under comparable conditions of temperature and pressure, which in turn implies comparable depths of burial. The comparison can only be approximate, for Archaean thermal gradients were steeper than those of the late Proterozoic, because there were more radioactive heat-producing elements in the crust. Comparable metamorphic grades were presumably reached at shallower depths in the Archaean than in the late Proterozoic. Nonetheless, the original depth of burial of the Precambrian rocks was probably between about 10 and 30 km, and this represents the amount of erosion that must have occurred since their formation. Yet at the present time, Archaean rocks more than 2500 Ma old are exposed at about the same level of erosion as Pan African rocks about 500 Ma old. The implication is that erosion rates must have been very high during and immediately after each thermotectonic event or orogeny, and declined greatly thereafter. This is supported by evidence of very high initial rates of erosion in the present-day Alps and Himalayas. The tectonic patterns If similar geological relationships result from similar causal mçchanisms, it seems likely that some form of plate tectonics was already in operation during Archaean times. The Pan African mobile belt is similar in age to early Phanerozoic orogenic belts in other parts of the world, for which generally accepted plate tectonic models have been proposed. If the Pan African thermotectonic event was the result of a ‘plate tectonic orogeny’, then the older terranes were presumably formed by similar processes. Even so, the precise nature of those processes remains a field ripe for speculation. For example: (a) The rates of movement of Archaean lithospheric plates must have been substantially greater than at present, and the sizes of those plates probably smaller, because the Archaean Earth was a good deal hotter than the present- day one. However, at least in the later part of the Archaean, the thickness of continental crust and the styles of deformation do not appear to have been greatly different from what they are now — for instance, horizontal thrusts and fold movements have been recognise in some Archaean terranes 40 (other than the Kasila belt of Sierra Leone). (b) It is possible to recognise three main kinds of orogenic belt in Phanerozoic terranes. There are gradations between these types, and features of each can be discerned in different parts of the West African craton and the adjacent Pan African belts. (c) It has also been proposed that in other parts of the African continent, Pan African terranes represent an aggregation of island arcs and continental fragments swept together by subduction and accretion into a single continental block. If that were to prove valid for the eastern domain of the Pan African in West Africa, then by implication it could also hold for the older rocks of the craton, because of their overall similarities to this Pan African terrane. While palaeomagnetic data do not appear to permit the opening and closure of large ocean basins during the Precambrian, they do not preclude the development of small ensimatic troughs and back-arc basins. They are also consistent with oblique rather than ‘head-on’ collisions between crustal blocks and fragment and island arcs. Major shear zones would form in this way, but it must be borne in mind that they can also develop as a result of direct collision, as seen in the present-day Himalayan mountain belt. (d) An additional variable is the rate and angle of subduction beneath island arcs and continental margins, which can affect patterns of metamor phi rejuvetiation and related magmatism in the overlying crust. (e) Finally there remains the question of the rate of growth of continental crust through geological time. The evidence of repeated reactivations and generation of granitic magmas by crustal remelting that seems to characterise the Precambrian of West Africa supports the prop ositio that most of the world’s continental crust was in existence by the end of the Archaean. It does not constitute proof, how ever and there is still much debate about the mechanisms of crustal growth and the relative importance of juvenile contributions from the mantle versus recycling of crustal material. The mineral deposits Contrasted and conflicting ideas about the possible role of plate tectonics during the Precambrian remain of academic interest in relation to the distribution of mineral deposits in West Africa. The most important mineral deposits in the Precambrian of West Africa 41 as a whole are iron and manganese ores, gold and diamonds. The first two owe their origin more to accidents of marine geochemistry than to tectonic forces, the manganese ores requiring in addition special weathering conditions to make them exploitable. Gold is an elusive element prone to redistribution by hydrothermal processes and subsequent concentration into alluvial placers and its ultimate origin is not easy to determine. Diamonds come from kimberlites that originate at great depths, but the diamonds in the Birimian are found in sediments and their origin cannot yet be related to any primary source region. Bauxite is another ore whose occurrence and distribution is due mainly to favourable climatic conditions acting on a variety of rock types. Plate tectonic models have been used to explain the distribution of different kinds of mineral deposits, particularly of base metals such as copper, lead—zinc and tin, but few predictive models exist. Rock associations are still the best guide to prospecting. Thus, tin is not found in ultramafic rocks, chromium is not found in granites. Similarly, coal and petroleum and phosphates are not sought in igneous and metamorphic terranes. They are to be found among unmetamorphosed and largely undeformed sediments, such as those that occupy the great sedimentary basins of West Africa. 2. GENERAL GEOLOGY OF BIRIMIAN AND TARKWAIAN ROCKS Birimian Supergroup The term "Birimian" was introduced by KITSON in 1918 to describe rocks from the River Birim valley in the Atewa-Kibi range in Ghana. The Birimian as a stratigraphic unit was established in the 1st Geological Sketch Map of the Gold Coast and Western Togoland at the scale of 1 : 500,000, which was published by KITSONin 1928. Birimian rocks form a substantial part of the Man shield as defined by BESSOLES (1977) (also called Leo shield by ROCCI et al. 1991), which occupies the southernmost third of the West African craton. The West African craton has remained stable for nearly 2.0 Ga. It is bounded on the east and west by mobile belts, mainly of Panafrican age. The Man shield itself comprises a western domain consisting of Archean rocks of Liberian (ca. 2.75 Ga), Leonian (ca. 2.95 Ga) and pre-Leonian (ca. 3.1 Ga) age (WRIGHT et al. 1985), and an eastern domain (Baoule-Mossi domain or Birimian/ Eburnean province) 42 composed chiefly of Birimian rocks of Paleoproterozoic age, which were affected by a major tectonothermal event, the Eburnean, around 2.1 Ga. The formation of Birimian rocks and Eburnean intrusives marks a major Paleoproterozoic juvenile crust-forming event, which is loosely referred to as the Eburnean orogeny. Modern geochronological studies in Burkina Faso, Cote d'Ivoire, Ghana, Mali, eastern Mauretania and Senegal indicate that these rocks were formed over a maximum time interval of c. 2.25 to 2.05 Ga (TAYLOR et al. 1988, 1992, OUCHAMI et al. 1990, LIEGEOIS et al. 1991, BOHERet al. 1992, HIRDES et al. 1992, DAVIS et al. 1994, HIRDES et al. 1996). Much of the West African craton including the Birimian domain is covered by generally flat-lying, undeformed sediment of Neoproterozoic to Paleozoic age. There is very little well established and documented information on the basement beneath the Birimian in the Man shield. Only minimal isotopic evidence for the involvement of Archean or very early Proterozoic crust in the genesis of Birmian volcanic and/or Eburnean plutonic rocks exists (TAYLOR et al. 1988, 1992, ABOUCHAMI et al. 1990, LEUBE et al. 1990, BOHERet aI. 1992). Two notable, but local exceptions exist, each close to a Birimian/ Archean boundary: the Winneba granitoid in the southeast of Ghana (TAYLOR et al., 1988, 1992) and granitoid plutons in south Guinea (BOHER et al., 1992), which have Nd model ages of c. 2.7 to 2.4 Ga. For many years, up to about 1985, two schools with almost contradictory concepts on Birimian stratigraphy existed: the anglophone and the francophone schools. JUNNER (1935, 1940) established a Birimian stratigraphy in Ghana comprising a lower series consisting of slates, phyllites, greywackes, tuffs and minor lavas, together with schists and gneisses derived from these rocks, and an upper series consisting of greenstones, mainly metamorphosed basic and intermediate lavas and pyroclastic rocks with some hypabyssal igneous rocks and intercalated bands of phyllite and greywacke. This two-fold subdivision of the Birimian in Ghana into older metasediments and younger metavolcanics was further chronostratigraphically subdivided by DABOWSKI (1972) and TRASHLIEV (1972). Francophone geologists, too, broadly distinguished two units of Birimian rocks in the francophone countries of West Africa: 43 AUBOlN(1961) and TAGINI(1971) -trying to assign an Alpine geosynclinal model to the Birimian of Cote d lvoire -defined eugeosynclinal and miogeosynclinal basins, the eugeosynclinal ones being filled by a group of volcanic and volcano sedimentary formations, often of ophiolitic nature, at the bottom overlain by a group of flysch and molasse. BESSOLES (1977) -disagreeing with the presence of eugeosynclines, miogeosynclines and ophiolites -distinguished the following two units instead: Type 1 units which were said to have been deposited in deep troughs with high length/width ratios, and to consist of greenstones, metabasalts, volcaniclastic rocks, quartzites, phyllites, and gondites, overlain by flysch-type sediments. Type 2 units were said to have been laid down in broad shallow basins and to be composed of predominantly felsic volcanosedimentary sequences, sericite schists, greywackes, quartzites, meta-conglomerates and calc-chlorite schists, with greenstones and flysch being practically absent. In general, francophone geologists proposed a stratigraphic sequence for the Birimian in their working areas which is reverse to the one established in Ghana, maintaining that Birimian sediments were younger than Birimian volcanic rocks. Anglophone and francophone concepts agreed that, in principal, two major types of granitoids intrude the Birimian rocks of West Africa: Large, concordant, foliated plutons which are spatially associated with Birimian sedimentary terranes, and relatively small, discordant, massive intrusives which occur in Birimian volcanic terranes. The former were called "Cape Coast type" in Ghana and "Baoule type" in francophone countries, and the latter were termed "Dixcove type" and "Bondoukou type", respectively (JUNNER1940, RNOULD1961). Cape Coast-/Baoule-type granitoids were believed to be older than Dixcove-/Bondoukou-type granitoids. The advent of sophisticated lithogeochemical and isotopic techniques and the resulting modern petrogenetic work in West Africa during the mid-1980s produced a considerable amount of new information and modified concepts in both francophone and anglophone West African countries. However, so far it has only led to partial harmonization of the francophone and anglophone views with respect to Paleo proterozoic geology of West Africa. 44 Important new contributions by work in trancophone countries include the following: LEMOINE et al. (1985) and TEMPlER (1986) introduced an "orogenic cycle" independent of and prior to the Eburnean orogeny, the "Burkinian", said to embrace the time span between 2.4 and 2.1 Ga (later changed to 2.4 to 2.15 Ga). Rocks affected by the Burkinian event were termed Dabakalian. The Dabakalian formations are represented by small, high-grade metamorphic terranes within the Birimian/ Eburnean terrane of the Man shield. However, recent dating work by BOHER et al. (1992) using the Sm-Nd, RbSr and U-Pb zircon methods yields ages between 2.19 and 2.14 Ga for Dabakalian gneisses, virtually identical to ages of Birimian rocks in Ghana (HIRDES et al. 1992, DAVIS et al. 1994). MILESI et al. (1989) -partly following the classical Ghanaian concept (see above) distinguished an older volcanosedimentary and sedimentary supergroup (B 1 or Lower Birimian), which was affected by a collision event (D 1) as expressed by thrusting, and a younger volcanic supergroup (B 2 or Upper Birimian), affected only by two later phases of transcurrent deformation, D 2 and D 3. The major D 1 tectonic episode, said to have occurred around 2.112-2.100 Ga (FEYBESSEet al. 1989), was considered to be one of the principal criteria for distinguishing B 1 from B 2. The D 2 phase was placed between 2.096 Ga and 2.073 Ga, and consequently the duration of the entire Eburnean deformation said to have been 40 million years (FEYBESSE et al. 1989). In an attempt to harmonize the above versions, VIDAL et al. (1992) used the terms "Dabakalian" and "B 1" as well as "Birimian s.s." and "B 2" synonymously, despite differing age ranges. Subsequently, VIDAL& ALRIC (1994) conceded that the definition of the Dabakalian is narrower than that of the Birimian B 1, "including high-grade metamorphism, rotational deformation, granitization, and sedimentary and volcanic formations". On the other hand, these authors use the terms "Burkinian" and "D1" as synonyms. VIDAL et al. (1992) suggested that "the upper Birimian (B 2) mafic volcanics" rest on pre-existing, continental (Dabakalian or B 1) basement. et al. (1990) and ABOUCHAMI BOHERet al. (1992), on the other hand, noted the lack of evidence for inherited Archean or very early Proterozoic crust in the central and eastern parts of the craton, as well as the juvenile character of Birimian tholeiitic basalts. The latter authors accepted the existence 45 of two Paleoproterozoic cycles (Burkinian and Birimian), however, and proposed that the first took place between 2.2 and 2.15 Ga, and the second between 2.12 and 2.07 Ga. With respect to the depositional environment/tectonic setting of Birimian rocks, ABOUCHAMI et al. (1990) suggested that Birimian volcanic belt basalts are Proterozoic analogues of modern, intraplate ocean-floor-flood-plateau basalts, such as those in the Nauru basin of the south-western Pacific Ocean. BOHER et al. (1992) preferred a model in which, after formation of intraoceanic plateaus, subduction and build-up of island arcs take place; these finally collide with Archean continental nuclei. VIDAL& ALRIC (1994) claimed that Birimian rocks of the Haute Comoe area represent a rift-type or transtensional back-arc basin which formed within Burkinian crust. MORTIMER(1990, 1992) -noting that trace element signatures of Toumodi volcanic rocks resemble those of modern island arc basalts -and LEAKE(1992) proposed for the Fetekro greenstone belt and associated lithostratigraphic units a model involving oblique collision of a younger island arc/marginal basin sequence with a root zone of an older island arc. By extension of this model, they considered the entire Baoule-Mossi domain to have developed by the amalgamation of discrete island arcs and arc-related basins by processes comparable with modern terrane accretion tectonics. In addition, MORTIMER (1992) stated that field evidence from the Fetekro belt suggested that the Burkinian "rests on shaky ground". Recent contributions to Birmian geology by geologists in anglophone West African countries -mainly Ghana -include the following: JUNNER'S(1935, 1940) classical subdivision of the Birimian Supergroup into Lower Birimian (mainly dacitic/rhyodacitic volcaniclastics, wackes, argillites and chemical sediments (LEVEE& HIRDES1986)) and Upper Birimian (mainly tholeiitic basalts with intercalated volcaniclastics) was reinterpreted by LEUBEet aI. (1990). These authors showed that Lower Birimian and Upper Birimian, Le. the sedimentary/volcaniclastic assemblage (sedimentary basins) and the volcanic assemblage (volcanic belts), formed quasi-contemporaneously as lateral facies equivalents. Sediments were derived from the volcanic belts, either as detrital sediments in the form of turbidites, or by means of volcaniclastic eruptions discharging tephra and ash which -depending on grain size -were 46 deposited proximally to the volcanic cones or more distally within the sedimentary basins, thus forming different sedimentary facies. HIRDES& LEUBE(1989) pointed out that gold mineralization in Birimian rocks in Ghana is strikingly concentrated in 10-15 km wide corridors along volcanic belt/sedimentary basin boundaries, and ascribed this phenomenon to the presence of chemical sediments as well as widespread and intensive tectonization in the belt/basin transition zones. Structural data by EISENLOHR& HIRDES(1992) and BLENKINSOP& SCHMIDTMUMM (1994) suggest that Birimian volcanics and Birimian sediments (as well as the overlying clastic-sedimentary Tarkwaian Group) were deformed in a single, progressive deformation event involving NW-SE crustal shortening. It consisted initially of a regional, low-strain phase which subsequently culminated in the local formation of high-strain zones such as thrusts and shears. Recent precise U-Pb mineral dating of rocks from the Ashanti belt fixes the age of the tectonothermal event at about 2100 to 2090 Ma in that particular region, but it is assumed that crustal shortening and deformation affected the Birimian terrane in Ghana diachronously and progressively from SE to NW (HIRDES et al. 1992, DAVIS et al. 1994, HIRDES& NUNOO1994, OBERTHOR et al. 1994). Contrary to long-held views, Dixcove granitoids (termed "belt granitoids" by LEUBE et al. (1990) were found to be 60 to 90 million years older than Cape Coast granitoids ("basin granitoids") by HIRDES et al. (1992) and DAVIS et al. (1994). These authors obtained precise U-Pb ages of 2179 and 2172 Ma for belt granitoids of SW Ghana, whereas basin granitoids yielded 2116, 2090 and 2088 Ma. The authors suggest that belt granitoids and belt volcanic rocks are coeval, whereas basin granitoids are interpreted as having intruded towards the end of deformation, i.e. are late-kinematic. Single-zircon-grain dating of a volcaniclastic wacke sample from the Kumasi basin demonstrates that Birimian sediment was mainly derived from (volcanic belt) terranes where rocks with crystallization ages of c. 2185-2155 Ma were predominant (DAVIS et al., 1994). This age range encompasses previously determined ages for belt-type granitoids (see above). The presence of a 2135Ma zircon, however, could suggest that 47 1. either minor volcaniclastic activity and/or belt granitoid generation, accompanied by sedimentation in the adjacent basins, continued -though volumetrically unimportant -up to c. 2135 Ma, or 2. that deposition of Birimian sediments began at least 20 million years after the main crystallization period, although the sediments were largely derived from the 2185-2155 Ma old belt rocks. Various alternative models have been proposed with respect to the tectonic setting of Birimian rocks in Ghana. Based on the striking parallelism and even-spacing of volcanic belts in Ghana, occasional evidence for slight crustal contamination, as well as no evidence favouring ophiolithic suturing, LEUBE et al. (1990) and HIRDES et al. (1993) put forward a model involving small-scale, equidimensional, parallel and contemporaneous convection cells in the upper mantle which led initially to very pronounced attenuation of an Archean sialic protocrust as well as to subsequent formation of rifts and linear spreading ridges where N-MORB tholeiitic lavas erupted. Subsiding basins probably in excess of 100 km wide between the spreading ridges and volcanic belts were filled with volcaniclastic and detrital sediment originating from the volcanic ridges. Later, NW-SE crustal shortening led to accretion of the relatively competent volcanic belt terranes as well as to intense deformation (isoclinal folding, shearing, and thrusting) in the intervening, incompetent basin sediment terranes. Partial melting of basin sediment towards the end of deformation produced the latekinematic basin granitoids. In brief, the above authors suggested an evolution of the tectonic setting of the Birimian/Eburnean in Ghana from an intracratonic-rift to oceanic-spreading and finally to an accretion-collis ion-related setting. POHL & CARLSON(1992), emphasizing the sparse isotopic evidence for earlier Archean crust in most Birimian and Eburnean rocks as well as the occasional presence of calc-alkaline andesite-dacite-rhyolite sequences, prefer a model in which most Birimian volcanic belts represent island arc complexes and some are extensional back-arc basins. They point to the existence of a possible ophiolite succession in northern Ghana (this being the only evidence so far for suturing typical of Phanerozoic Benioff zones). For the crustal-shortening period of the Birimian, POHL& CARLSON (1992) invoke closure of 48 an oceanic basin -situated between the Man shield and the Nigeria craton -where multiple arc-arc collisions or, in the west, arc-continental platform collisions took place. They claim that the modern paradigm is the compression of the Indonesian-Melanesian Archipelago and the intervening basins by the northward migration of the Australian plate towards the Asian continent. SYLVESTER & ATTOH (1992) maintain that the trace element chemistry of Birimian volcanic belts in Ghana is comparable to that of Archean greenstone belts, and that intermediate calc-alkaline units show the high Ba/La and Ta depletion characteristic of similar rocks in modern subduction environments and distinct from those formed in intraplate settings. They suggest that Birimian volcanic rocks probably originated as immature island arcs built on oceanic crust. DAVIS et al. (1994) stress the remarkable similarity in regional geological patterns and relative internal age relationships between the Paleoproterozoic Eburnean orogen of Ghana and the late Archean Kenoran orogen of the southern Superior Province in Canada. In this district of Canada, detailed and extensive structural and isotopic work suggests that deposition of volcanogenic-turbiditic basin sediments was synchronous with accretion of arcs and micro-continental fragments against a growing continental mass. More specifically, it is proposed that (in the Superior Province) formation of the major sedimentary basins post-dated the main episode of exposed volcanism in adjacent volcanic belts by several tens of millions of years and was a result of uplift of allochthonous terranes during collision. OBERTHUR et al. (1994) and VIDAL& ALRIC (1994) emphasize that -taking into account the vast area of at least 1600 x 1000 Ian underlain by Paleoproterozoic rocks on the West African craton -far-reaching extrapolation of local situations appear at least questionable for this geologically relatively underexplored terrane. This is - for example corroborated by recent findings of HIRDES et al. (1996) in Cote d'Ivoire. These authors demonstrated that the Birimian/Eburnean province -seen up to now as an entity -is, in fact, divided into an eastern and a western subprovince: The eastern subprovince, which covers Ghana, eastern Cote d'Ivoire and probably many parts of Burkina Faso, typically displays c. 2150-2185 Ma belt volcanism 49 and coeval belt plutonism (HIRDESet al. 1992). Evidence for 2100 Ma extrusive volcanism is absent. The western subprovince (e. g. central Cote d'Ivoire (LEAKE1992), western Mali (LIEGEOIS et al. 1991), and probably Guinea) is characterized by younger, c. 2105 Ma old volcanic belts and coeval belt plutons. Supracrustal and intrusive rocks of the 2150-2185 Ma time span are at least locally present as various gneisses which previously had been termed "Dabakalian". It is very well possible that two or more of the various scenarios outlined above, or a combination of some of them, will eventually describe the Paleoproterozoic crustal evolution of the West African craton most appropriately. For example, there seems to be fairly well established evidence that both island-arc settings as well as spreading-related, oceanic environments existed in different parts of the Baoule-Mossi domain during the formation of Birimian/Eburnean rocks. Also, there is some indication of the presence of older continental crust, at least in the marginal parts of the Birimian terrane. Investigations concentrating on differences between neighbouring volcanic belts as well as variations along strike within individual belts seem to be the most promising way of resolving some of the apparent discrepancies outlined above. It is hoped that the Explanatory Notes and geological maps presented here will make one of the many contributions necessary to achieve this aim. Tarkwaian Group The term "Tarkwaian" was first used in Ghana to designate a sequence of coarse clastic sedimentary rocks (conglomerates, sandstones, arkoses, and some subordinate argillites) which overlies Birimian rocks, and for which the type locality is situated in the vicinity of Tarkwa town in southwest Ghana. At Tarkwa, Tarkwaian rocks are characteristically predominantly rudaceous to arenaceous erosion products of Birimian terranes and were laid down in a fluviodeltaic to lacustrine depositional environment. They are separated from the underlying rocks of the Birimian Supergroup by an unconformity documenting an episode (probably short-lived) of block faulting and uplift. Various interpretations of the stratigraphic position of the Tarkwaian exist. Most workers in Ghana regard the Tarkwaian as a distinct stratigraphic unit that is younger than the 50 Birimian rocks, in contrast to BESSOLES (1977) and HASTINGS (1982), who consider the Tarkwaian as the uppermost unit of the Birimian. The latter concept contradicts JUNNER et al. (1942), who show that Tarkwaian sediments truncate Birimian rocks in the Tarkwa area, and that there is a difference in strike of 20°-35° between the two rock series. Tarkwaian sediments play only a very minor role in the mapped area, and therefore a comprehensive review of the Tarkwaian Group is not attempted here. Instead, we refer to WHITELAW (1929), HIRST (1938), JUNNER et al. (1942), VANEs (1967, 1968), SESTINI (1971, 1973), HIRDES et al. (1988), KLEMD et al. (1993) and HIRDES & NUNOO (1994). The geological map of Ghana by BATES(1955) shows Tarkwaian rocks only in the Ashanti, Bui and Bole-Navrongo belts. Field work and literature research demonstrate, however, that Tarkwaian sedimentary rocks are present in all five volcanic belts in Ghana. They are developed in the central portions of the belts; an exception being the Sefwi belt, where they occur on the eastern margin. Tarkwaian sediments were laid down in depositories with high length/width ratios probably inter-montane grabens or depressions formed by rifting. There is no evidence that the individual depositories were ever linked. The degree of preservation of sediment fill in the various Tarkwaian depositories is highly variable. This is partly due to regional differences in erosion level; it is best in the Ashanti belt, where the Tarkwa Au mine is in operation, and in the Bui belt. At Tarkwa, rocks of the Tarkwaian Group are stratigraphically subdivided into four series (JUNNERet al. 1942). Formation and infilling of the depository at the type locality took place during the time span between 2132 and 2116 Ma (DAVIS et al. 1994). Birimian and Tarkwaian rocks were both deformed in a single deformation event at around 2100 Ma (EISENLOHR & HIRDES1992, DAVIS et al. 1994). Sequence of geological events in the Paleo Proterozoic of Ghana The sequence of geological events in the Paleoproterozoic of Ghana, on the basis of geological field relationships and U-Pb high precision ages, is depicted in Table 2. Ages 51 originate from TAYLOR et al. (1988, 1992), HIRDES et al. (1992), and DAVIS et al. (1994, 1995). The given age brackets pertain to central and south-western Ghana. It is assumed that in Ghana formation of Paleoproterozoic crust took place by means of progressive accretion of volcanic arcs and ocean plateaus onto an Archean continental mass located in the SE (DAVIS et al. 1994, HIRDES& NUNOO1994), and it follows that some of the processes outlined above (e.g. deformation) affected the Ghanaian terrane diachronously. Table 2: Sequence of geological events in the Paleoproterozoic of Ghana 2086-2073 Ma hydrothermal alteration in basin granitoids (possibly related to gold mineralization) 2116-2088 Ma emplacement of syn-to late-tectonic basin plutons 2100-2090 Ma peak of deformation (NW-SE crustal shortening) and regional metamorphism 2132-2116 Ma formation of Tarkwaian depositories and deposition of Tarkwaian sediments 2135-2132 Ma uplift, erosion 2135 Ma latest igneous activity in volcanic belts, final stage of sedimentation in adjacent sedimentary basins 2190-2155 Ma formation of bulk of belt volcanics, volcaniclastics and associated synvolcanic belt plutons, and contemporaneous sedimentation in adjacent sedimentary basins 2245 :t4Ma age of oldest detrital zircon in Tarkwaian sediment (source unknown) 2590 Ma minimum age of continental crustal contribution to Winneba (Sm-Nd model) granitoid (areally restricted to small portion of SE Ghana) Remarks on the Ashanti belt At present, Ghana is the second largest gold producer in Africa (after South Africa). Of Ghana's gold output, by far the most comes from the Ashanti volcanic belt (Fig. 1), which stretches for approximately 250 km from Axim town in the south to Konongo town in the 52 north before it disappears under a flat-lying Voltaian cover of Infracambrian age. The economically most important mines are concentrated in the transition zone formed by the NW side of the Ashanti belt and the SE side of the adjacent Kumasi basin. The first-class gold deposits of the Ashanti belt are Birimianhosted, structure-controlled quartz veins and arsenopyrite/pyrite bodies, such as those at the Ashanti Goldfields Mine at Obuasi, which is the largest gold producer in West Africa. Its documented gold production since 1898 exceeds 7001. A second, economically less prominent gold ore deposit type in the Ashanti belt is represented by the Au-bearing quartz-pebble conglomerates of the Tarkwaian Group, which unconformably overlies much of the Birimian of the Ashanti belt, in particular the central part. A characteristic feature of the Ashanti belt, as compared to other Birimian volcanic belts in Ghana, is the predominance of volcaniclastic rocks and the relatively sparse development of lava flows. This, in conjunction with the almost complete preservation of the overlying Tarkwaian rocks, suggests that the Ashanti belt represents a crustal segment characterized by a relatively shallow erosion level (LEUBE&HIRDES1986). The Ashanti belt as a whole represents a synform 1989). Its boundaries (EISENLOHR with the adjacent sedimentary basins (Kumasi basin in the west and Cape Coast basin in the east; Fig. 1) are tectonic, i.e. are formed by the Prestea-Obuasi-Konongo high-strain zone and the Inchaban fault/thrust zone, respectively. In general, on the NW side, where numerous gold mines and prospects are situated, sediments of the Kumasi basin have been thrust obliquely over Ashanti belt rocks (including the Tarkwaian). No notable gold mines occur on the south-eastern flank of the Ashanti belt, but the area is poorly exposed and highly underexplored. Individual Rock Units of the Project Area This section characterizes the individual rock units outlined in the legend of the Sekondi and Axim geological sheets. The same legend is used for both sheets, despite the fact that not all rock units occur on both sheets. Additional information which cannot be incorporated into this compilation may be obtained from the original mapping reports. 53 Supracrustal rocks The term supracrustal refers to any rock that was deposited on basement rock on the Earth's surface, and includes metamorphic equivalents. The investigated southernmost sector of the Ashanti belt is characterized by a relative abundance of belt granitoid intrusions; there is clearly a predominance of intrusive rocks over supracrustal rocks in the project area. Mapping of supracrustal rocks of the area is based on the lithofacies concept by LEUBE& HIRDES(1986), LEUBE et al. (1990) and HIRDESet al. (1993) instead of the chronostratigraphic approach proposed by JUNNER(1935, 1940) and MILESI et al. (1989). All rock types discussed below, except for the Sekondian sediments of Devonian/Carboniferous age and the Apollonian sediments of Cretaceous age, have been subjected to metamorphism of at least greenschist facies; however, the prefix "meta" is omitted. Volcanic rocks of the Birimian Supergroup Within the scope of tbese Explanatory Notes, the term "volcanic rock" is restricted to lava flows and their subvolcanic equivalents; volcanic ejecta are discussed under 3.1.2. All volcanic rocks in the project area and their subvolcanic equivalents occur in the Ashanti volcanic belt, which embraces the dominant part of the two mapped sheets 0402A and 0403B. In the mapped southern part of the Ashanti belt, volcanic rock is frequently interbedded with volcaniclastic and other sediments, in particular on the western flank. Volcanic rocks in the investigated area form three major NE-SW trending lobes or branches which are separated by large, probably composite, belt granitoid plutons (see Section 3.2.1).These are referred to here as Butre branch, Cape Three Points branch and Axim branch. A fourth lobe of volcanic rocks, which is shown on previous geological maps of Ghana (e.g. BATES1955) towards the east of Inchaban town, could not be confirmed by the present mapping campaign; instead occasional amphibolite pods in granodioritic gneisses were encountered. Twenty-four samples of volcanic rock, originating from all three greenstone branches in the southern Ashanti belt area, were analyzed by means of X-ray fluorescence analysis. 54 The SiOz versus Zr/TiOz and Zr/TiOz versus Nb/Y plots of WINCHESTER & FLOYD (1977) (Fig. 2) show that about 40 % of the investigated sample suite are basalts and c. 60 % are andesites. Only one sample displays felsic composition. 3. THE SEKONDI SERIES The Sekondi Series consists mainly of sandstones and shales with conglomerates, pebble beds, grits and mudstones resting with major unconformity on a complex of granites, gneisses and schists. West of Takoradi the underlying rock is a coarse, porphyritic hornblende-granite of the Dixcove type, while in the Cape Coast area it is a fine-grained biotite-granite of the Cape Coast type, with abundant intrusions of muscovite-pegmatite, microgranite and quartz. Elsewhere, the underlying rocks are biotite— and hornblendegneisses, schists aid granulites which are probably metamorphosed and partly granitized Birrimian rocks of pre-Cambrian age. Fossils are not common and have so far been found only in the Takoradi Sandstone and Shales. As a result of Cox’s work (1946) the Series is now regarded as of either Devonian or Carboniferous age with a slight balance in favour of Devonian (p. 33). The subdivision of the Sekondi Series is based entirely on differences in lithology, but fortunately there are several very characteristic beds which persist over the whole area and their recognition is fairly easy in spite of a complex and discontinuous structure. The general classification based on the occurrence at Sekondi-Takoradi is given below:- 55 S6 Sekondi Sandstone. Thickness (ft) (b) Upper—Pebbly argillaceous and feispathic sandstones and conglomerates. 1, 000 (a) Lower—Massive quartzose sandstones and grits with subordinate shales and mudstones. S5 Efia Nkwanta Beds. (c) Upper—Thin bedded siltstone, shale, shaly sandstone, and some coarse sandstone, with nodules, bands, and lenses of chert .. .. .. .. .. .. .. .. 85 (b) Middle—Friable sandstone, both well bedded and massive, with interbedded mudstone and shale .. .. .. 315 (a) Lower—Cross-bedded, soft, fine-grained, pale purple, pink, grey, green, and cream sandstone .. .. .. .. .. 300 S4 Takoradi Shales. Black and grey carbonaceous shales, sandy shales, and shaly sandstone, with interbedded grit and fine-grained sandstone, and with nodules of siderite and pyrite .. 650 S3 Takoradi Sandstone. (b) Massive and bedded friable ferruginous sandstone with coarse grained beds, brecciaconglomerate, and interbedded shales .. .. .. .. .. .. 500 (a) Thin-bedded, brittle, micaceous sandstone with sandy shale and some clay shale .. 100 S2 Elmina Sandstone. Chocolate and purple felspathic micaceous sandstone, with coarse sandstone, conglomerate, shale, and mudstone near the base .. .. .. .. .. 1,000—1,200 S1 Ajua Shales. Varved shales, sandy shales, and sandstones containing scattered boulders and pebbles with a coarse boulder bed at the base . . .. MAJOR UNCONFORMITY Hornblende-granite of the Dixcove type. Biotite-granite of the Cape Coast type. 56 .. .. .. . . 140—200 Biotite- and hornblende-gneiss, schist, and granulite (metamorphosed Birrimian). The lithological descriptions and measurements of thicknesses are based on observations in the western area where the most complete succession can be seen. Between Kafodidi and Cape Coast the complexity of faulting makes it impossible to assess thicknesses accurately. Lithologically, however, there is little variation from east to west, except that, east of Komenda, a strong development of sandstone and quartz-breccia in the Takoradi Shales makes it impossible to differentiate between these and the Takoradi Sandstone, and the two formations are classed together under the single term Takoradi Beds (S3-4). B Description and Petrology 1. Ajua Shales The Ajua Shales, which are the lowest formation of the Sekondi Series, rest directly on the underlying crystalline rocks and are better developed on the west than on the east. They form a series of isolated outcrops, denuded of the overlying Sekondi sediments, to the west of Takoradi; and a series of strips overlain by the Elmina Sandstone runs from Ajua in a north-east direction to a point beyond Kwesimintim. A similar strip stretches from Inchaban to Aboadi, and another occurs two or three miles north-west of Komenda. The formation is not known east of Komenda, and it is believed that it pinches out so that the Elmina Sandstone oversteps on to the crystalline rocks. The Ajua Shales are well exposed at various places along tile coast west of Takoradi, where they tend to form rocky points such as those at Ajua, Pumpuni (Kokobo Point), Achreboanda, and Asemkau (Adoblo Rock). Exposures occur in the Hwini River, south of Kwesimintim, and in the anti-malarial drains east of Kwesimintim. The occurrences in the Inchaban and Komenda areas are not exposed naturally and are known only from test pits and topographical features. Typically, the Ajua Shales consist of thin-bedded, black or dark grey shales with arenaceous laminations and beds of grit, the formation becoming more sandy towards the top. Underlying the typical shales is a basal series, up to 12 or 15 feet thick, of boulder beds, conglomerates, shales and sandstones. Scattered pebbles and boulders occur throughout the formation. 57 At Asemkau and Adoblo Rock the basal beds of the Ajua Shales can be seen resting unconformably on the underlying hornblende-granite with diorite dykes (Plate X A). Here, the base of the sediments consists of a bed of large boulders of the granite and diorite, rounded or sub-angular, up to 2 or 3 feet in diameter, and derived from the underlying rocks. The boulders can be seen in all stages of relation to the parent rock: some still attached, some separated but obviously still in situ, but most of them lying on the surface a short distance from their original position. The boulder bed varies in thickness and in concentration of boulders, but it is usually only one boulder deep and therefore 1 to 2 feet thick. In places it is absent altogether, especially on the higher parts of the old surface. The boulder bed is set in a matrix of Sekondi sediments, generally a very coarse sandstone or a fine or coarse conglomerate, but in places a shale or massive sandstone. The very coarse sandstone is derived from the underlying granite and was probably formed as a detrital sand on the old land surface prior to the deposition of the sediments. The conglomerate contains rounded and subangular pebbles, mostly of Dixcove granite with other igneous rocks, Birrimian greenstone, quartz, and quartzite. Generally, the pebbles in the coarse conglomerate range only up to a few inches in diameter, but there are some boulders over a foot across. Above the conglomerate lies some 6 feet of thin-bedded, fine-grained, rather shaly sandstone, greyish yellow in colour, and containing lenses of massive sandstone; and this is overlain in turn by a further bed, 6 feet thick, of massive, fine-grained, grey sandstone with very coarsegrained bands. The conglomerate occurs only in the troughs of the undulating surface; elsewhere the granite is overlain directly by sandstone. Typical Ajua Shales are exposed at Ajua, Pumpuni, and Achreboanda, and at Adoblo Rock they can be seen resting on the basal beds. They consist of thin-bedded black or dark grey shales with arenaceous laminations. The shales are quite fissile and split along the argillaceous partings, making the rock appear more shaly than it really is. A section across the bedding shows the true nature of alternating sandy and clayey bands. The laminations are frequently sufficiently regular, numerous, and consistent in lateral extent to suggest the rhythmic repetition of seasonal varves formed in still water by deposition from glacial melt waters. At most places, however, the laminations show strong ripple marking and irregular cross bedding with small lenses of both argillaceous and 58 arenaceous material. At some horizons the argillaceous laminations occur as disconnected lenses, or may “wash out “ altogether so that the rock becomes a fine grit with only subordinate shaly matter. These grits consist of very angular quartz grains with abundant fairly fresh felspar—mainly soda-plagioclase with some microcline—and chlorite. The arenaceous facies become more dominant towards the upper part of the formation as at Ajua Point. In some places the shales are rather calcareous. A bed of fine-grained grey nodular sandstone, 6 feet thick, occurs at approximately 15 feet above the basal beds, and near the top of the formation are various beds of sandstone, probably lenticular, interbedded with the shales. They are not sufficiently well exposed to enable a succession to be worked out, but they include a fine-grained resistant cream sandstone, a pinkish grey fine-grained thin-bedded sandstone with large micas and small flattened mudstone pebbles, and a greenish grey argillaceous sandstone. A section giving the measurements of the beds which occur at Asemkau is given on p. 43. The most interesting feature of the Ajua Shales is the occurrence throughout the formation, though apparently confined to the shaly facies, of coarse clastic fragments of rocks of all types, including granite, diorite, quartz-porphyry, quartz, Birrimian greenstone and tuff, Tarkwaian quartzite, and sandstone. They vary greatly in size from the grade of a coarse sand to boulders over 1 foot in diameter, and in shape from rounded to sub-angular and angular. The great majority of them are angular and show little sign of prolonged water wearing, the surface of the pebbles of igneous rocks being commonly pitted and weathered. They occur along well-marked horizons in poorly sorted assemblages, the bases of the pebbles and boulders being on one plane, suggesting that they have been deposited together on the surface of the accumulating varved shales. The varves are compacted under the larger pebbles and boulders, showing that the surface was depressed by their deposition. The pebbles and boulders occur in the coastal sections of the Ajua Shales and in the Kwesimintim area, but they have not been detected in test pits in the Inchahan and Komenda areas. The Ajua Shales are strongly ripple marked, especially the shaly facies. Most of the ripple marks are perfectly symmetrical with concave troughs and cuspate crests, typical of those formed by the oscillatory movement of waves in water which is free from currents. There are many good examples of interference ripple marks formed by two sets 59 of oscillatory ripples more or less at right angles to each other. The strike of the ripples is very variable, ranging from 72° to 177o, or from east-north-east to south, the average being about south-east. Some of the ripples show a slight but distinct asymmetry indicating occasional gentle currents. Junner (1939a, p. 15) deduced from these that the sediments came from the north-east. Another characteristic feature of the varved shales is the common occurrence of intraformational disturbances. Zones varying in width from half-an-inch to 1 foot or so of crumpled, contorted and overfolded shales lie between undisturbed beds. The disturbed zones are frequently, but not always, associated with the bands of pebbles. The Ajua Shales, where they outcrop on the coast and are constantly washed by the sea, tend to form resistant rocky promontories such as those at Ajua, Pumpuni (Kokobo Point), Achreboanda, and Adoblo Rock. Inland, however, the rock is less resistant, probably because the felspars are attacked by humic acids from the overlying vegetation. Where the strata are inclined, the outcrops of Ajua Shales are indicated by well-marked valleys flanked by hills of crystalline rocks on one side, and the basal ridge of the Elmina Sandstone on the other. Examples of this topography occur north of Kwesimintim, between Inchaban and Aboadi, and north of Komenda. Where the shales are horizontal or gently dipping, as at Ejan, they become lateritized and more resistant, forming low hills. The total thickness of the Ajua Shales at Ajua, according to calculations based on dip and width of outcrop, is about 140 feet, and in the Kwesimintim area, probably over 200 feet. Elmina Sandstone The Elmina Sandstone is, with the possible exception of the Sekondi Sandstone, the thickest formation of the Sekondi Series, being probably over 1,000 feet thick. Because of its homogeneity and frequent exposure, it is also one of the best known and characteristic rocks of the coast line between Takoradi and Cape Coast. It extends over large tracts of country and stretches of coast line at Takoradi, Efia, and south of Inchaban, and it is more extensive at Komenda, at Amisanb, and along the coast either side of Elmina. It is a uniform, hard, massive, medium-grained, felspathic sandstone with a characteristic chocolate or chocolate-purple colour, which is due to the pink felspars and the dark 60 brown limonitic cement. The colour persists in depth, and is not a product of weathering as has been postulated by some authors. This is proved from a borehole which was sunk at Efia Nkwanta to a depth of 325 feet (p. 38), the rock retaining its chocolate colour throughout. In places, where the iron oxide has been leached out during weathering processes, or is present in the ferrous state, the rock becomes green or speckled with pink and green, owing its colour to chlorite and ferrous iron. The quartz and felspar grains are angular and well cemented together with iron oxide, probably limonite. The felspars are fairly fresh and are mainly soda-plagioclase, with a little microcline and occasional orthoclase. The rock also contains large dark and white micas, and chlorites after mica, which at some horizons are very abundant, making the rock flaggy. Heavy minerals include rutile, topaz, apatite, magnetite, ilmenite, and zircon. On the whole it is poorly bedded, well jointed, and strongly cross-bedded. The cross-bedding shows a consistent direction of current flow from the north-east. Towards the base of the formation the rock tends to become coarser-grained, and the chocolate sandstone is interbedded with coarse and fine conglomerates, shales, and mudstones. The coarser conglomerates contain well-rounded pebbles up to 3 inches in diameter set in a chocolate sandstone matrix, and derived from Upper Birrimian greenstone, granite, quartzmica-schist, and quartz. The finer conglomerates are more quartzose and have a more limonitic matrix. The shales and mudstone are chocolate in colour when fresh. At the top of the formation the sandstone becomes thin-bedded and somewhat shaly, but retains its felspars, in sharp contrast with the overlying shales and thin-bedded quartzose sandstones of the Takoradi Beds. From Gold Hill eastwards the uppermost part of the Elmina Sandstone consists of an interesting breccia. It contains ill-graded angular pebbles and boulders, up to several feet across, of Elmina Sandstone set in a matrix of the same material, both clearly having been derived from the underlying rock. On the shore at the east end of Gold Hill, where the bed is well exposed at low tide, many of the boulders can be seen practically in their original position (Plate XI A). There are a few small boulders and pebbles of biotite-gneiss. The boulders of Elmina Sandstone and the matrix of the breccia are both greenish in colour, as is the top few feet of the underlying sandstone. The Elmina Sandstone weathers to a white kaolinitic sandy clay, which becomes 61 lateritized only to a small extent. The argillaceous facies weather under favourable conditions to a pure kaolin. A good example was detected in a test pit just south of Ankaful prison camp, where a particularly pure kaolin is overlain by 15 feet of superficial sands, and passes down into increasingly fresh chocolate mudstone and sandstone. The iron has been leached from the clay by circulating water and protected from lateritization by the superficial cover of sand. In some places kaolin is also formed in fault gouges. On the shore, where the rock is continually washed by the sea, it is hard and resistant, forming low rocky off-shore reefs (Plate X B). Inland, the Elmina Sandstone gives rise to a characteristic low-lying topography with gently sloping low hills, contrasting with the higher ground formed by the other members of the series. The depth of weathering is not great and there are many outcrops of rock which form characteristic rounded masses. The soil is thin and poor, resulting in a savannah type of vegetation of grass and scattered trees. It does not make good farming country. The interbedded conglomerates and shales at the base of the Elmina Sandstone form a distinct ridge where the rocks are inclined. This generally forms a well-marked feature contrasting with the valley formed by the underlying Ajua Shales. It is difficult to estimate the thickness of the Elmina Sandstone since, except perhaps in the Inchaban area, there is no unbroken succession from the Ajua Shales to the Takoradi Sandstone, and there are no intermediate marker horizons permitting a composite measurement. At Inchaban the thickness is probably about 1,000 to 1,200 feet. Takoradi Sandstone Junner (1939a, p. 14) referred to this formation as the “highly friable sandstones,” and in his 1939 classification he included them with the Takoradi Shales. Previously, however, he had put them separately in his “group D” (p. 10). Since they are predominantly sandstones rather than shales, and almost as thick as the Takoradi Shales, and also because they form very distinctive topographical features, the present author proposes to separate them under the name “Takoradi Sandstone.” The term “Takoradi Beds” is used where, as discussed below (p. 19), it is desired to refer as a whole to these two somewhat similar and related formations. 62 The actual junction between the Elmina Sandstone and the Takoradi Sandstone is observable at several places, such as at the west end of “ Windy Ridge,” Takoradi, on the shore near Abrobeano, and at Gold Hill. Although the contiguous beds of each formation are of similar lithology (shaly beds at Takoradi and Abrobeano, and breccia and conglomerate at Gold Hill), the contact is distinctly marked in the fresh rock by the abrupt cessation of felspars above it. The absence of felspars is a characteristic feature of the Takoradi Sandstone and the Takoradi Shales. The base of the Takoradi Sandstone consists of approximately 100 feet of grey shales, sandy shales, shaly sandstone, and thin-bedded, fine-grained, brittle, micaceous, cream and pink sandstones. The sandstone bands are only a few inches thick, and many of them contain abundant worm tracks and worm casts. The shales at Gold Hill are fossiliferous, yielding poorly preserved brachiopods, lamellibranchs, and fish remains. At Gold Hill a quartzose breccia-conglomerate with a highly lirnonitic matrix underlies the shaly beds, and forms the base of the Takoradi Sandstone, resting on the breccia at the top of the Elmina Sandstone. Above the basal shaly beds lies the typical facies of the Takoradi Sandstone. When fresh this is a massive, medium-grained, cream coloured, highly friable sandstone. The mineral content is predominantly quartz, the grains being angular but very well sorted (Plate V). There is also abundant zircon, red and yellow rutile with apatite, magnetite, tourmaline, and P glauconite. At many horizons the grains are very coarse, and pebbly bands and conglomerates are common. The pebbles are almost entirely of glassy or white quartz. Bands of grey shale and sandy shale are interbedded with the sandstone but, like the conglomerate beds, are probably lenticular and non-persistent. Towards the top of the formation the sandstone becomes well bedded and the shale bands more frequent. The sandstones display many perfect “text-book” examples of subaqueous cross-bedding, with the bounding planes of the cross-bedded units parallel to the true bedding, and the cross-bedding planes dipping consistently and steeply to the south-west (Plate XII A). When fresh, the massive and bedded sandstones are cream coloured, but with increased weathering the rock becomes limonitized, turning yellow, pink, purple, and dark rusty brown. The limonite is probably leached from lower levels by circulating ground water and redeposited at the surface. Where the leaching is complete, the rock is white and 63 extremely friable. The effect of the limonitization is to indurate the rock and make it very resistant to weathering. At Gold Hill a natural cave penetrates the hard, highly linionitic face of the cliff and opens into the softer leached zone of the rock. The limonite tends to be deposited along bedding planes or in irregular concentric bands. Where the limonitization has gone far, the rock becomes differentially eroded and the surface assumes a strikingly sculptured appearance, emphasizing in strong relief the crossbedding and concentric banding. This is particularly well displayed in the cliffs at Gold Hill (Plates XI B, XII A). The Takoradi Sandstone, being resistant to weathering, forms hills. Where the rocks are inclined, as is usual, the hills become prominent scarp ridges, strewn with limonitic sandstone boulders. They are generally wooded and, contrasting with the adjoining grasscovered low-lying Elmina Sandstone country, they are a familiar feature of the Sekondi Series landscape. Examples occur at Efia, Anunkwari (2 miles south of Inchaban), Abrobeano, and Elmina. The total thickness of the Takoradi Sandstone is 600 feet. Takoradi Shales The Takoradi Shales bear certain similarities to the Takoradi Sandstone and the two formations undoubtedly belong to the same general phase of deposition. In the absence of any known distinctive faunal horizons, and the rarity of any sections exposing the passage from one formation to the other, it is difficult to define exactly the position of the junction between them. Whereas the lower of the two formations is predominantly sandstone, the upper one is essentially shaly, and at Epawano headland (three-quarters of a mile west of Kafodidi) there is certainly a clear line of demarcation between the two. The typical Takoradi Shales are hard, compact, black or very dark grey, fissile shales or sandy shales, rich in carbonaceous matter. The colour lightens considerably on weathering. They are exposed at New Takoradi and at Cox’s Hill, Takoradi.* At the base of the formation the rock is a rapidly alternating succession of thin-bedded micaceous sandstone and grey shales, not unlike those at the base of the Takoradi Sandstone. They can be seen in the cutting at the top of the road leading down to the main harbour 64 entrance. Towards the top of the formation a series of hard, grey-green grit bands, generally some 6 inches thick, but in places increasing to several feet, are interbedded with the shales, together with lenticular beds of fine-grained pale grey sandstone up to several feet thick. These are exposed along the coast line at Esupon. Boulders and fragments of medium-grained friable sandstone similar to the TakQradi Sandstone occur on the surface at a few places inland, but the beds from which they are derived have not been seen in situ. The characteristic feature of the Takoradi Shales is their mineralization. Their carbonaceous content has already been mentioned, and in addition the black shales at Takoradi and Poasi contain thin bands of brittle bitumen and traces of oil. A slight oily smell can sometimes be detected if the shales are heated or struck-with a hammer. Pyrite occurs in several ways : as microscopic disseminations giving the shales a greenish tinge, as visible “pepper-dusting,” in individual and aggregates of cubic crystals, and as spherical nodules up to an inch in diameter. Marcasite may be present also. Atmospheric action on exposed rock surfaces oxidizes the sulphides to the rather unusual sulphate minerals jarosite {K2Fe6(OH)12(S04)4) and halotrichite (FeSO4.A12(S04)3.24H2O) (Junner, 1939a, p. 14), probably through the formation of sulphuric acid. Jarosite occurs as yellow incrustations on old exposed faces, particularly near the sea. It is apparently confined to the Takoradi Shales and occurs at Takoradi, Esupon, and Kafodidi. Halotrichite (iron alum) is rapidly formed on freshly exposed surfaces, and crystals will grow on hand specimens brought into the laboratory, but it does not remain for long in exposed places as it is soluble in water. It occurs along the bedding planes as pale yellow and white crystal aggregates with a fibrous and asbestiform habit. The fibres are at right angles to the bedding and the disruptive effect of the crystal growth forces the bedding planes apart. Towards the top of the formation are large discoidal nodules of compact, finely granular, grey siderite or clay ironstone. The nodules are up to 1 foot in diameter and have a surface coating of red-brown limonite. Calcite and gypsum occur as thin veinlets traversing the shales and siderite nodules, and along the bedding planes. The Takoradi Shales, like the base of the Takoradi Sandstone, are fossiliferous. The upper part at Esupon yields lamellibranchs, brachiopods, and gastropods, and the middle part at Takoradi has yielded fish remains (for details see section IV, p. 32). While most of 65 the carbonaceous matter has no recognizable structure a great deal clearly consists of plant tissue. The sandy beds are commonly rich in concretionary matter suggesting, and possibly derived from, worm casts, worm tracks, and corals. The formation lateritizes readily and its consequent resistance to erosion gives rise to hills which are wooded and similar to those formed by the Takoradi Sandstone. The beds at the junction of the two formations are easily erodecL probably because of their lack of homogeneity. Consequently, two parallel ridges separated by a valley are generally formed, one of the Takoradi Sandstone and the other of the Takoradi Shales. Whereas the surface of the former is strewn with boulders of limonitic sandstone, the latter is covered with fragments of lateritized shale or sandy shale, this generally being the only means of distinguishing the two formations inland. This double ridge feature can be seen in Takoradi at, for example, the cricket ground which lies in the valley between Chapel Hill (Takoradi Sandstone) and the hospital hill (Takoradi Shales); or at Efia Nkwanta, where the road to Adiembra runs along the valley between the hospital hill (Takoradi Shales) and a ridge to the south-west composed of Takoradi Sandstone. The feature is also seen at Anunkwari and Kafodidi. The thickness of the Takoradi Shales is estimated to be 650 feet. * At the time of writing, Cox’s Hill is being removed to make room for the new extensions to Takoradi Harbour. Takoradi Beds The term “Takoradi Beds” is used when it is impossible or undesirable to distinguish between the Takoradi Sandstone and the Takoradi Shales. There are beds, both shales and sandstones, of which the lithological characters are very similar in both formations, and it is clear that they all belong to one general phase of deposition. It seems likely that Junner considered this to be the case since he included ‘the “ highly friable sandstones” with the Takoradi Shales (p. 11). Moreover, such fossil evidence as is available suggests that similar faunal assemblages occur in both formations. (Fossils are not known elsewhere in the Sekondi Series.) It is therefore, on occasion, convenient to have some term embracing both formations. More important is the fact that east of Komenda it is impossible to distinguish one from the other owing to changes in lithology and, except at Elmina, 66 absence of good exposures. The wooded double ridge feature exists and, north of BrenuAchinuni and Ankwana, the rocks forming it demonstrably lie between Elmina Sandstone and the Efia Nkwanta Beds. But from surface indications, and exposures in Elmina town, it is apparent that the rocks corresponding stratigraphically to the Takoradi Shales are rich in sandstones and quartz conglomerates typical of those in the Takoradi Sandstone farther west. The eastern extension of the Takoradi Shales thus shows a progressive change towards a more arenaceous facies and the clear distinction between the two lithological types is lost (p. 26). East of Komenda both formations have been mapped together under the single term “Takoradi Beds.” In the Ankwana area the thickness of the Takoradi Beds is estimated to be approximately 1,200 feet, which is equivalent to the total thickness of the two constituent formations at Takoradi. Ella Nkwanta Beds The Efia Nkwanta Beds occupy an area south of Takoradi, an area around Poasi, and a broken strip extending inland from Asamang (near Ekuasi, Sekondi). They also appear on the coast at Kafodidi, Ampeni, Brenu-Achinum, and Cape Coast, together with blocks at Esamang (near the Elmina by-pass) and Kwapro. They derive their name from the coastal village of Efia Nkwanta or, as it is now called, Nkontompo, 1 miles west of Sekondi Lagoon.* They are characterized by bright colours and an interesting variety of rock types, and are divided into a lower, middle, and upper division. The lower division also includes a series of transition beds connecting the Takoradi Shales with the typical sandstone which comprises the Lower Efia Nkwanta Beds. The total thickness of the Efia Nkwanta Beds is about 700 feet. * Before the construction of Takoradi Harbour and the present motor road between Sekondi and Takoradi, Efia Nkwanta, as its name implies, was the point where the path to Efia joined the old Axim hammock road along the coast. To-day Efia Nkwanta is, strictly, still the junction for Efia but is now on the motor road half a mile inland, although the name generally refers to the surrounding district, not clearly defined, but including 67 Sekondi hospital. The coastal village is properly called Nkontompo. There are, in fact, no Efia Nkwanta beds at Efia Nkwanta as it is known to-day. Lower Efia Nkwana Beds. The type locality of the transition beds between the Takoradi Shales and the typical Lower Efia Nkwanta Beds is at Esupon beach, where a practically continuous section is visible. There is a continuity of deposition and no clear line of demarcation, although the end members of the transitional series are quite distinctive. The topof the Takoradi Shales comprises hard, compact, dark grey shales with beds of greenish grit and nodules of siderite. The base of the transition beds is taken to be a prominent bed of pale grey very finegrained sandstone with a distinctive “floury” weathered surface. Above this, similar sandstones are thicker and more numerous, and interbedded shales more sandy and less predominant until the typical sandstone of the Lower Efia Nkwanta Beds is reached. Sideritic nodules occur in the transition beds. The sandstone of the Lower Efia Nkwanta Beds is, perhaps, the most distinctive of all the members of the Sekondi Series and is well displayed in the second railway cutting east of Efia Nkwanta level crossing, and also on the coast either side of Kafodidi. It is a very fine-grained, soft, cross-bedded, and ripple marked sandstone with a characteristic “floury” texture on weathered surfaces. The principal colours of the rock are “pastel” shades of pale grey, pink, purple, and green, but the cross-bedding and ripple marking is picked out by differential colour banding in dark mauve, brown, and cream.* The average grain size of a typical sample is about .13 mm. and the grains are very well sorted (Plate V). Although the rock is very fine-grained, material of clay grade forms less than l per cent of the total. The grains are poorly cemented and principally of fairly angular quartz with zircon, rutile, sphene, tourmaline, magnetite, and apatite. The cross-bedding is on a large scale and of the irregular type generally associated with eolian deposits (Plate XII B). The cross-bedded units are up to 4 feet thick and the bounding planes are curved and inclined in every direction. Most of the curves are concave upwards, but some of them are sigmoidal or undulating. The cross-laminations are closely parallel with the lower bounding surface and their slopes therefore correspondingly variable. Towards the top of the formation the cross-bedding becomes regular and shows a consistent current direction from the north-east. 68 Middle Efia Nkwanta Beds. The base of the Middle Efia Nkwanta Beds is marked by a rich red, brittle, nodular, and irregularly fractured argillaceous siltstone. It is a distinctive and persistent bed some 14 feet thick and is recognizable at Sekondi, Esupon beach, and at BrenuAchinum. Above this, the middle division is a series of alternating sandstones and shales or mudstones, the individual beds varying in thickness from about 2 feet to 20 feet. They are well displayed at the oil storage installation at Poasi and in the railway cutting just south-west of Asamang level crossing. (For details of these sections see pp. 48, 49.) The sandstones are quartzose and friable, and are not unlike similar beds in the Takoradi Sandstone, but in colour they are generally a brighter pink or orange and are not so well graded as the Takoradi Sandstone (Plate V). The argillaceous beds are grey or purple with greenish bands and patches. The Middle Efla Nkwanta Beds are about 315 feet thick. * Owing to the mode of weathering, the colours will “smudge “ when the rock surface is rubbed, thus enhancing the illusion of a pastel drawing. Upper Efia Nkwanta Beds. The upper division of the Efla Nkwanta Beds consists of well-bedded, purple, pink, grey, and green shales and siltstone with some mudstone, finegrained sandstone, and a few coarse-grained beds. Some of the sandstones are similar in appearance to the Lower Efia Nkwanta Beds but without cross-bedding. They are exposed at Takoradi Harbour, Nkontompo, Asamang, Ampeni, and on the ridge just west of Cape Coast Lagoon. The characteristic feature of the beds is the presence of grey and white chert.* The chert occurs either in thin bands, usually about a quarter of an inch or more thick, or as rounded nodules. In either case the mineral is distributed along the bedding planes. At Nkontompo some of the sediments are sun-cracked and the chert occurs as spheroidal nodules in the infillings of the cracks. At Takoradi Harbour, behind the railway offices, the beds show intraformational disturbances, possibly as the result of slumping. Both the disturbed and undisturbed beds contain bands of chert, but in the former the chert has clearly been contorted together with the beds in which it occurs. The disturbed beds show folding and overfolding, but the chert, being hard and incompetent, is fractured. It is a useful aid to mapping since it remains unaltered on the surface after the rock has completely weathered and enables the beds to be traced where there are no 69 exposures. At Brenu-Achinum surface accumulations have yielded a red-banded jasper, but the beds from which it was presumably derived are not exposed. The thickness of the Upper Efia Nkwanta Beds is about 85 feet. * Previous authors have generally alluded to this mineral as “chalcedony,” which implies the pure, translucent form. It is in fact opaque and, as seen as under the microscope, full of quartz and impurities. Junner originally referred to it as “chalcedony” (1939a, p. 13) but later as “chert “(1940, p. 27). Sekondi Sandstone The present writer has separated the Sekondi Sandstone into an upper and lower division. The upper division is predominantly feispathic and pebbly; the lower is quartzose and without pebbles, but there are interbedded argillaceous rocks in both groups. The formation occurs in two extensive areas, one embracing Sekondi and its environs, and the other a large tract astride the Elmina by-pass and extending to within 2 miles of Cape Coast. A third very small block occurs at Ankaful village, 6 miles north-west of Cape Coast. Lower Sekondi Sandsione. The chert-bearing rocks which comprise the Upper Efla Nkwanta Beds are overlain by a thick, massive, cross-bedded, orange-pink sandstone. Both formations are exposed in the railway cutting at Asamang. Lunn (1932, p. 14) considered the sandstone as the upper part of the Efla Nkwanta Beds together with, probably, all the rocks now classed as the Lower Sekondi Sandstone. It contains abundant scattered fragments of chert, which are frequently in the form of angular flakes, irregularly orientated and apparently broken from thin beds of chert such as occur in the Upper Efla Nkwanta Beds. The present writer regards them as detrital in origin and derived from the Upper Efla Nkwanta Beds, thereby implying emergence and consequent erosion of these beds. There is, however, no visible indication of an unconformity in the Asamang cutting but, as explained below (p. 27), erosion probably took place elsewhere as a result of stream rejuvenation. There was certainly a change in the conditions of deposition, and for this reason the writer considers that all rocks above the chert beds 70 should be included in the Sekondi Sandstone. Detrital chert is a common constituent of both upper and lower divisions of this formation. The rocks immediately overlying the orange-pink sandstone are not observable, but from the topographical features it is assui’ned that they are generally argillaceous. The next exposed rocks in the succession are a series, some 90 feet thick, of greenish shales weathering to red and containing some bands of sandy and silty shales. They are exposed in a road cutting immediately north of the bridge carrying the new road from Sekondi to Adiembra over the railway (p. 50). They underlie a second bed of massive sandstone which forms the uppermost part of the Lower Sekondi Sandstone. This is a mediumgrained, very massive, quartzose sandstone of a pale greyish white colour, weathering to yellow or pink (Plate XIII A). It is notparticularly well graded (Plate V). Its mineral content is principally quartz with associated zircon, apatite, magnetite, and red and yellow rutile. Although it is fairly friable it acquires a hard crust, which makes the rock extremely resistant to weathering. As a result it tends to form bare, towering, rocky outcrops over 20 feet high as at Adiembra and at several points to the north-west. These make a contrast with the usual soil and vegetation covered topography formed by the rest of the Sekondi Series. This bed of sandstone probably crops out also at Brunibuma, on the Elmina by-pass, but is more weathered and rather better bedded. At Adiembra the sandstone shows secondary silicification along fissures, possibly by hot solutions. In the railway cutting at Adiembra Hak, fissures have been widened and worn smooth by aqueous erosion, but the channels are now infilled with detritus. The total thickness of the Lower Sekondi Sandstone is approximately 650 feet. Upper Sekondi Sandstone. The typical rock forming the Upper Sekondi Sandstone is a soft, argillaceous, feispathic sandstone with pebbles. It is generally of a chocolate or pink colour, the former speckled with pink and green minerals similar to the Elmina Sandstone but without the prominent micas. In addition to quartz and felspar, subordinate minerals include magnetite, ilmenite, zircon, and a little rutile. There is a clearly defined contact between the lower quartzose and upper felspathic pebbly sandstones which can be seen in the new Sekondi— Adiembra motor road, 250 yards due east of Adiembra Halt. The whole group is very lenticular, with marked local variations between soft argillaceous 71 sandstone without pebbles, and coarse conglomerates, there being every intervening grade of pebbly sandstone. The pebbles are generally well rounded and consist mostly of white quartz with lesser amounts of greenstone, green quartzite, phyllite, and small fragments of chert. The rock, both pebbly and non-pebbly facies, is poorly graded (Plate V), but in the fresh rock the pebbles tend to lie along clearly defined bedding planes (Plate XIII B). There are abundant subordinate bands of shale and sandy shale at various horizons in the Upper Sekondi Sandstone, generally grey, chocolate, or pink in colour (p. 50). The Upper Sekondi Sandstone is sufficiently resistant to weathering to form hills which occur in irregular groups. North of Elmina, where the formation is faulted on the north and south between Elmina Sandstone, the group of hills covered with secondary bush is a prominent feature of the landscape. The pebbly sandstone forms rugged cliffs on the coast east bf Esikado. The weathering of the pebbly sandstone is effected by the kaolinization of the felspars and subsequent removal of the resulting fine-grained clay material. The pebbles remain at the surface as an eluvia] gravel. This is the origin of the so-called “ Sekondi gravels “which are common on the hills in Sekondi and Esikado. They are not terrace or plateau gravels since they are confined to areas underlain by the Upper Sekondi Sandstone and occur at varying heights above sea-level. The thickness of the gravels varies directly as the abundance of pebbles in the immediately underlying sandstone, being practically absent above pebble-free facies of the rock. The pebbles, therefore, undergo a minimum of lateral transportation except where they occur in the beds of streams draining the Sekondi Sandstone. Where exposed sections are thick enough, e.g. at the entrance to the Sekondi European cemetery, the gravels can be seen to pass down, generally imperceptibly, into weathered pebbly sandstone. The gravels are either unbedded, or poorly bedded in the same sense as the underlying Sekondi Sandstone (due allowance being made for surface creep). 4.THE VOLTAIN BASIN The Volta Basin has a more or less concentric distribution of sediments, because of its overall gently synclinal form. The oldest sediments outcrop round the margins, the youngest occupy a roughly central position. 72 The Lower Voltaian, the Dapango—Bombouaka Group, is dominated by massive crossbedded feldspathic sandstones. It is responsible for the high ground of the Kwahu Plateau in the south, at the base of which the uneven pre-Voltaian surface can be seen in places. It has been suggested that the basement hills to the south of the plateau may be exhumed pre-Voltaian topography. The group as a whole is correlated with the Togo Formation on the east of the Volta Basin. The Lower Voltaian is virtually flat-lying throughout most of the basin, but becomes relatively intensely folded as the Togo belt is approached. Sediments of the Pendjari Group (Middle Voltaian) generally rest with slight angular unconformity on the Lower Voltaian, and in some places they fill erosional channels that have cut through the older beds, so that they rest directly on the basement. They form much of the low ground dominated by the Oti Plains of northern Ghana and Togo and they are also known as the Oti Formation in Ghana. The basal conglomerate of the Pendjari Group is interpreted as a tillite (in Ghana it may be represented by the Akroso Conglomerate). It contains boulders up to more than a metre across, subangular to rounded, some with striated, polished or pitted surfaces. It rests on a striated pavement at one locality, where the grooves are oriented 100o and suggest that the sense of movement of the ice was towards the east. The tillite is succeeded by a variety of sediments to including carbonates, often brecciated or slumped, locally barite-bearing and partly stromatolitic, along with silexites and silicified argillites. The argillaceous sediments are locally enriched in calcium prosphate which is of considerable economic importance in places. This association, known as the triad by French workers is also found in many parts of the Taoudeni Basin and provides an important stratigraphic marker. It is generally not much above 100—200 m thick in total, wherever it is found, and can be appreciably less in places. The rest of the Pendjari Formation is dominated by shales, siltstones and sandstones, glauconitebearing in places. The Middle Voltaian is generally correlated with the Buem Formation, on account of the overall lithological similarities between the two groups of rocks. The Lower Voltaian sediments represent a marine transgression—regression cycle on the craton, while the Middle Voltaian records a glacial event followed by prolonged marine incursion and subsidence of the basin. The thicker Middle Voltaian sediments probably 73 represent a subsiding continental shelf environment. In the eastern parts of its outcrop, adjacent to the Togo belt (Fig. 9.1), the Pendjari Group has been deformed into generally NNE—SSW trending asymmetric folds with southeasterly inclined axial planes. The Obosum Formation (Upper Voltaian) is thickest and coarsest in the south-east. The conglomerates contain pebbles of granite and other igneous rocks, as well as quartzite fragments, and sedimentary structures show the direction of transport to have been from the south-east. There is general agreement that the Obosum beds are a molasse deposit formed by erosion of the Togo belt following its uplift in the Pan African event. Sedimentation may have continued until Devonian times, though there is no direct evidence of this. However, as some of the conglomerates and sandstones have been identified as fluvioglacial in origin, these sediments may also contain a record of the late Ordovician glaciation that is documented in the Taoudeni Basin. Geochronology of the Volta Basin Direct evidence of the age of these rocks is confined to a few radiometric age determinations and sparse palaeontological data. However, they support the inferences based on correlations with the Togo belt and the eastern Pan African domain on the one hand, and with the Taoudeni Basin successions on the other. The Middle Voltaian (Pendjari Group) has given ages of 620 Ma (K/Ar on glauconite) and 640 Ma (Rb/Sr on illite), indicating a late Precambrian age, which is consistent with evidence from stromatolites and microfossils. The age of the tillites is estimated at around 675 ± 25 Ma which places them in the Vendian, or uppermost Precambrian, similar to the likely age of the inferred glacigenic sediments at the base of the Rokel River Group. In summary, the age range of the Lower Voltaian (equivalent to the Togo Formation, Sec. 6.2), is taken to be about 1000—700 Ma, and of the Middle Voltaian (equivalent to the Buem Formation, Sec. 6.2), about 675—570 Ma (cf. Table 9.1). The Upper Voltaian Obosum Formation is a molasse to the Pan African and sediments contributing to it are believed by some workers to range from Ordovician up to Devonian or even younger, a total span of about 450—320 Ma. 74 The Voltaian Group There are few detailed geologic descriptions of sediments of the Voltaian Group sequences; subdivision of the group is difficult due to poor exposure and the lack of laterally persistent lithological marker beds or fossils. The group has generally been divided into three formations, each separated by an unconformity marked by a tillite. The Lower Voltaian Formation consists of a massive to cross-bedded arkosic sequence. The Middle Voltaian Formation consists of a flyschoid sequence, and the Upper Voltaian Formation consists of a molasse sequence (Affaton et al., I 980). The Lower Voltaian Formation unconformably overlies the Birimian Supergroup. A radiometric age of 993 ± 62 Ma from the lower part of the Lower Voltaian Formation gives the approximate period for the beginning of sedimentation of the group. The calculated age was based on a three-point regression line and no analytical data were published (Clauer, 1976, quoted in Cahen et al., 1984). The deposition environment for the Lower Voltaian Formation is likely to have been shallow marine or fluviatile. It reflects a fairly stable tectonic setting throughout the depositional area. The Middle Voltaian Formation overlies the Lower Voltaian Formation with a slight angular unconfonnily. f ong1omnematc beds interpreted as tillites form the basal part of the Middle Voltaian Formation (Petlers, 1991). Shales, siltstones and sandstones, glauconitic in part, constitute the principal lithologies of the Middle Voltaian Formation. K—Ar dating of glauconite from a borehole core at Tibagona yielded a Vendian age of 600±20 Ma (Bozhko, 1969, 1974, 1984, quoted in Bozhko, 1994); the presence of abundant early Vendian microfossils confirms this age. The Upper Voltaian Formation is divided into a lower and an upper unit. The lower unit consists mostly of dirty-yellow, fine-grained, thinly bedded, micaceous, feldspathic quartz sandstones with subordinate argillite intercalations. The upper unit consists of white whitish-yellow, massive, fine- to medium-grained, cross-bedded arkosic and quartzose sandstones. The Upper Voltaian Formation occurs as scattered outcrops in the central part of the Voltaian Basin, with an average thickness of about 400 m. The Upper Voltaian Formation constitutes the molasses deposits formed in the Cambro-Ordovician period by the erosion of some horizons in the Pan- African mobile belt (Grant, 1969). 75 4. THE VOLTA BASIN The Volta Basin has a more or less concentric distribution of sediments, because of its overall gently synclinal form. The oldest sediments outcrop round the margins, the youngest occupy a roughly central position. The Lower Voltaian, the Dapango—Bombouaka Group, is dominated by massive crossbedded feldspathic sandstones. It is responsible for the high ground of the Kwahu Plateau in the south, at the base of which the uneven pre-Voltaian surface can be seen in places. It has been suggested that the basement hills to the south of the plateau may be exhumed pre-Voltaian topography. The group as a whole is correlated with the Togo Formation on the east of the Volta Basin. The Lower Voltaian is virtually flat-lying throughout most of the basin, but becomes relatively intensely folded as the Togo belt is approached. Sediments of the Pendjari Group (Middle Voltaian) generally rest with slight angular unconformity on the Lower Voltaian, and in some places they fill erosional channels that have cut through the older beds, so that they rest directly on the basement. They form much of the low ground dominated by the Oti Plains of northern Ghana and Togo and they are also known as the Oti Formation in Ghana. The basal conglomerate of the Pendjari Group is interpreted as a tillite (in Ghana it may be represented by the Akroso Conglomerate). It contains boulders up to more than a metre across, subangular to rounded, some with striated, polished or pitted surfaces. It rests on a striated pavement at one locality, where the grooves are oriented 100o and suggest that the sense of movement of the ice was towards the east. The tillite is succeeded by a variety of sediments to including carbonates, often brecciated or slumped, locally barite-bearing and partly stromatolitic, along with silexites and silicified argillites. The argillaceous sediments are locally enriched in calcium prosphate which is of considerable economic importance in places. This association, known as the triad by French workers is also found in many parts of the Taoudeni Basin and provides an important stratigraphic marker. It is generally not much above 100—200 m thick in total, wherever it is found, and can be appreciably less in places. The rest of the Pendjari Formation is dominated by shales, siltstones and sandstones, glauconitebearing in places. The Middle Voltaian is generally correlated with the Buem Formation, on account of the overall lithological similarities between the two groups of 76 rocks. The Lower Voltaian sediments represent a marine transgression—regression cycle on the craton, while the Middle Voltaian records a glacial event followed by prolonged marine incursion and subsidence of the basin. The thicker Middle Voltaian sediments probably represent a subsiding continental shelf environment. In the eastern parts of its outcrop, adjacent to the Togo belt, the Pendjari Group has been deformed into generally NNE— SSW trending asymmetric folds with southeasterly inclined axial planes. The Obosum Formation (Upper Voltaian) is thickest and coarsest in the south-east. The conglomerates contain pebbles of granite and other igneous rocks, as well as quartzite fragments, and sedimentary structures show the direction of transport to have been from the south-east. There is general agreement that the Obosum beds are a molasse deposit formed by erosion of the Togo belt following its uplift in the Pan African event. Sedimentation may have continued until Devonian times, though there is no direct evidence of this. However, as some of the conglomerates and sandstones have been identified as fluvioglacial in origin, these sediments may also contain a record of the late Ordovician glaciation that is documented in the Taoudeni Basin. Geochronology of the Volta Basin Direct evidence of the age of these rocks is confined to a few radiometric age determinations and sparse palaeontological data. However, they support the inferences based on correlations with the Togo belt and the eastern Pan African domain on the one hand, and with the Taoudeni Basin successions on the other. The Middle Voltaian (Pendjari Group) has given ages of 620 Ma (K/Ar on glauconite) and 640 Ma (Rb/Sr on illite), indicating a late Precambrian age, which is consistent with evidence from stromatolites and microfossils. The age of the tillites is estimated at around 675 ± 25 Ma which places them in the Vendian, or uppermost Precambrian, similar to the likely age of the inferred glacigenic sediments at the base of the Rokel River Group. In summary, the age range of the Lower Voltaian (equivalent to the Togo Formation, is taken to be about 1000—700 Ma, and of the Middle Voltaian (equivalent to the Buem Formation, about 675—570 Ma). The Upper Voltaian Obosum Formation is a molasse to 77 the Pan African and sediments contributing to it are believed by some workers to range from Ordovician up to Devonian or even younger, a total span of about 450—320 Ma. 5. THE ACCRAIAN The Sedimentary rocks of Accra are exposed on the beaches and cliffs and existed in the Devonian period of over 350 million years ago. The Accraian can be sub-divided into three formations; with the oldest at the bottom. 1. Upper Sansatone –shale Formation 2. Middle shale Formation 3. Lower sandstone Formation. Lower Sandstone Formation: The rocks are essentially sandstone with subordinate amounts of grits, breccias and pebble beds and shales. The sandstones are at places thinly bedded with shale partings, but generally they are thickly bedded and at places fairly massive with current –bedding and fossil ripple-marks. Middle Shale Formation The rocks of the middle formation are essentially shales, but thin limestones and thin sandstones may be found. The sandstones become prominent towards the upper part of the formation. Fossils such as trilobites (extinct Marine Crustaceans), Lamellibranches, Gastropods and Brachiopods have been collected from the shales. Paleontologists therefore ascribe the Middle Devonian age to the shales. Small sacle folding is seen at places. Upper Sandstone-Shale Formation The Formation consists of sandstones and shales often interbedded in thin strata; but at places the sandstone beomes thicker. 78