The “Grès polymorphes” group of the Manika plateau are outcrops of relatively fine-grained, more or less hard, massive rock, varying in colour from light grey to pink and brownish yellow, depending on the degree of surface weathering.

Transmitted-light petrographic analysis showed an equant facies, composed solely of quartz grains of fairly similar size, relatively fine to medium, with generally subangular to subrounded contours and uniform extinction.

These monocrystalline grains show discontinuous diagenetic feeding as an intergranular siliceous cement; however, they do not show a ferruginous border around them. It is therefore a purely quartz rock, with no carbonate or feldspars, which fits into the petrographic classification of quartz arenites ( Figure 2 ).

The virtual granulometric study by linear counting of grains under the microscope not only made it possible to specify the average grain size of this sediment (Mz ranging from 120 to 300 µm), which is more or less identical to that of the grains of the so-called “pseudoquartzite” facies defined by [19] and the “Sables ocre” ( Figure 3 ); it also made it possible to calculate the granulometric parameters listed in Table 1 .

These sediments are composed exclusively of quartz clasts, with a sandy grain size, making up 94.7% to 99.0%, compared with only 1 to 5.3% of silt-clay particles. The sandy grains are predominantly sub-rounded, as in the overlying Sables ocres, despite their contours being marked by diagenetic growth halos.

The average grain size (Mz) of these sediments places them in the category of medium-to fine-grained sandstones. The sorting index (S₀) from [24] shows that these are medium to moderately well sorted sediments (S₀: 0.51 to 0.81) characterised by a platycurtic grading curve (K_G = 0.758 to 0.952), which probably indicates that the sediments were supplied by successively varied stocks. The skewness index (Sk) of the granulometric curve reveals values varying between −0.019 and 0.082, suggesting an almost symmetrical curve ( Figure 4 ).

The semi-logarithmic cumulative curve ( Figure 5 ) is similar to those of the overlying Sables ocres [2] [19] ; it has a quasi-sigmoidal appearance tending towards a more or less regular parabolic shape.

The shape of this curve, and in particular its relatively steep slope, would be characteristic of a stock of more or less homogeneous sands, moderately well sorted and transported in a fluvial environment [25] , [26] less agitated [27] . Figure 6 also suggests, according to [28] , that the sediments were transported in a fluvial continental environment.

The “Grès polymorphes” were certainly deposited in a warm environment, with deposition interrupted by recurrent wetter periods attested by the interstratification of alluvial and lacustrine sediments, particularly towards the north of the Congo Basin [21] .

While [4] considers that the “Grès polymorphes” are derived globally from the silicification of lacustrine limestones, [19] believes that these sandstones were formed simply by the silicification of the overlying sands. This silicification is part of an early and complex diagenetic process that is thought to be common at the base of “Sables ocre” where water circulation was once very intense [2] . According to [29] , this process affects the initially carbonated cements of these sandstones and is what sometimes gives them the appearance of millstone.

We observed the concretion of the “Sables ocre”, especially in their superficial part ( Figure 7 ).

However, this local phenomenon seems to take place under specific conditions due to the infiltration of humic acid solutions from the surface horizons of soils forming on these sands.

The chemical composition of the major elements in the “Grès polymorphes” of Manika is given in Table 2 . Examination of this table shows extremely homogeneous contents for the different samples.

The relatively high SiO₂, content varies from 94.98 to 97.20%, with a mean value of 96.11%; the Al₂O₃ content varies from 1.33 to 1.77%, with a mean value of 1.63%. The mean SiO₂/Al₂O₃ ratio, which is an index of chemical maturity, is high at around 59. The Fe₂O₃ varies between 0.25 and 0.87%, with a mean of 0.46%, which is relatively high enough to occasionally rubylize these sediments. K₂O and Na₂O show fairly low levels, varying respectively between 0.04 and 0.10% for K₂O and 0.03 and 0.06% for Na₂O; their mean content being respectively 0.07 and 0.04%.

However, there is no linear correlation between K₂O and Na₂O, nor between each of these elements and Al₂O₃, which seems to be due to the absence of feldspars in these sandstones and also shows that the concentration of Al₂O₃ is controlled by a tiny proportion of neoformed clays, probably kaolinite.

The mean Na₂O/K₂O ratio of 0.65 indicates significant leaching of Na in comparison to K, and therefore fairly high chemical maturity of these sediments according to [30] . Similarly, the average Fe₂O₃/K₂O ratio of 6.72 indicates an almost total absence of feldspars and micas in these sediments. The SiO₂/(Al₂O₃ + K₂O + Na₂O) ratio, which also constitutes an index of chemical maturity according to [31] , varies from 47.25 to 66.85 with a mean value of 55.20, again suggesting a high level of chemical maturity in these sediments. The high chemical maturity of these sandstones can be explained by their recycled nature and by the fact that they were produced in a highly altering paleoenvironmental context, which was probably humid and warm, under the control of a less pronounced relief that favoured strong weathering of the source rocks and over a fairly long transport period.

The geochemical classification diagrams of [32] ( Figure 8(A) ) and [33] ( Figure 8(B) ) classify these sediments as quartz arenites, certainly because of their low concentrations of Al₂O₃.

The chemical composition of trace elements in Manika “Grès polymorphes” is provided in Table 2 . The elements are categorized respectively into the large ion lithophile elements (LILE), high field strength elements (HFSE), rare earth elements (REE) and transition elements groups. For all these groups of elements concentrations are low and strongly clustered around the mean values.

It is difficult to speculate on the mineralogical significance of the elements present in such low concentrations. The least we can say is that the main mineralogical support mechanism for elements in the LILE and REE groups would be adsorption onto kaolinite, the only phyllite mineral presumed to be present in this leached context.

On the other hand, the relatively consistent Zr values (13.36 - 22.13 ppm) compared with all the other elements in the HFSE group could be explained by the occurrence of zircon granules as the mineralogical support phase for this element.

As far as transition trace elements are concerned, Cr appears to be the only element whose average content (122.73 ppm), higher than that of Archean cratonic sandstones [13] , seems to reflect, in addition to its adsorption on clays (kaolinite), a probable occurrence in the state of fine chromiferous spinel granules; this element with mafic affinity is naturally accompanied by Cu, Ni, Sc and V which share the same affinity with it.

The identification of the source material of the sandstones studied was approached using the geochemical discrimination diagram recommended by [23] . This diagram uses the major elements, Al₂O₃, TiO₂, Fe₂O₃, MgO, CaO, Na₂O and K₂O, most of which are relatively mobile to help discriminate effectively between the various origins. On the other hand, incompatible trace elements, which are generally not very mobile, appear to be particularly good indicators not only of the nature of the source rocks, but also of the possible sorting action of the carrier minerals during their transport, and also of the tectonic setting and the evolution of the environment [14] [34] [35] .

During the sedimentary process, the original composition of immobile elements is preserved and is therefore the signature of source materials and of all sedimentary processes, provided that these elements are not carried by heavy minerals whose distribution is generally affected by sorting and recycling over long distances.

It should be noted that the different origins seem to be implicitly associated with specific geotectonic contexts [36] . Thus mafic rocks are often associated with oceanic island arc contexts, intermediate rocks with mature island arcs, felsic rocks with the active continental margin; recycled continental sources are associated either with intracratonic sedimentary basins, or with the passive continental margin, or with recycled orogenic provinces. The diagram in [23] shows that the ‘Grès polymorphes’ of Manika are recycled sandstones ( Figure 9 ) in the context of an intracratonic sedimentary basin, the interior Congo Basin.

It is clear that in the absence of a consistent fine silt-clay fraction likely to act as a mineralogical support for trace elements in these sediments, the identification of source materials is tricky. It should be noted, however, that rare earths elements, Th, Sc and Th/Sc ratios appear to be particularly good indicators of source control in detrital sediments, as their distribution, unlike that of other elements, is not affected by the fractionation of heavy minerals during sediment transport [14] [23] .

The Th/Sc - Zr/Sc diagram in [12] for the composition of these sediments indicates a primary source close to the more characteristic “tonalite-trondhjemite- granodiorite complex” (TTG) rocks ( Figure 10 ) of Proterozoic to Archean age.

Of particular note in this diagram is the alignment of the points along a trend of increasing Zr/Sc ratio, indicating relative fractionation of zircon during sediment transport, confirming the recycled nature of these sandstones.

These are therefore sandstones developed in the context of the interior intracratonic Congo Basin, whose complex structural evolution and sediment succession have been discussed elsewhere by [3] , [4] and especially [37] .

Rare earth element spectra normalized to PAAS ( Figure 11 ) not only show low rare-earth elements abundances in these sandstones compared with the shales (PAAS), which does not allow their potential source materials to be identified, but they also show a slight fractionation of light rare earths elements and weak negative anomalies in Ce (Ce*/Ce: 0.79 to 0.85) and especially in Eu (Eu*/Eu: 0.93 to 0.98), relatively close to those calculated by [13] for Archean cratonic sandstones; this seems to confirm the Archean source of these sediments.

Manika’s “Grès polymorphes” are recycled sandstones resulting from the dismantling of primary sandstones, which were themselves formed by the weathering and erosion of Precambrian craton rocks during various tectonic phases.

The weak negative anomalies in Ce and Eu as well as the average V/Cr ratios of 0.2 well below 2, and Ni/Co of 2.66 below 5 suggest, according to [38] , that these sediments were produced under relatively oxidising conditions.

Normalised to the average content of Proterozoic and Phanerozoic cratonic sandstones [13] , the polymorphic Manika sandstones appear to be severely depleted in incompatible trace elements (LILE, HFSE) more characteristic of felsic than mafic rocks. On the other hand, they show concentrations of transition elements that are fairly close to those of the reference sandstones, and even higher in some cases, particularly for elements with a mafic tendency such as chromium, vanadium and cobalt ( Figure 12 ), suggesting a significant contribution from a mafic or intermediate magmatic source to the original sediments. The same is true of the mean values of the La/Sc (9.85) and La/Co (4.63) ratios, which seem to suggest a felsic origin for these sediments; however, the low Th/Sc (0.77) and Th/Co (0.36) ratios clearly indicate a mafic nature for the primary source of these recycled sandstones.

The chemical alteration conditions of source rocks have a definite impact on the composition of the resulting sediments. In order to estimate the alteration of source rocks into clay products, [39] have defined the alteration index with the following molar formula: CIA = [Al₂O₃ /(Al₂O₃ + CaO* + Na₂O + K₂O)] × 100. [39] - [41] interpret this index as a measure of the extent of the conversion of feldspars (the dominant minerals in the upper crust) into clays. For these researchers, AIC values close to 100% correspond to the alteration of source rocks into clay minerals such as chlorite, kaolinite or gibbsite.

Alteration of the source rocks of the Manika plateau sandstones is characterized by very high CIA alteration indices (84% to 91%), suggesting intense aluminosilicate alteration conditions, probably induced by a hot, humid climate and a relatively flat topography, as we have already pointed out.

The A-CN-K diagram ( Figure 13 ) from [39] also allowed us to get close to the approximate nature of the source rock, which for these sediments would be a magmatic source of felsic composition.

The chemical composition of terrigenous detrital rocks is tightly controlled by the tectonic setting of their source, so that detrital sediments from different tectonic settings show geochemical signatures specific to the context of their source rock [22] [23] .

The tectonic discrimination diagram of [22] identifies four main environments, namely the passive tectonic margin (PM), the active continental margin (ACM), the oceanic island arc context (ARC) and the continental island arc context. In the case of the Manika’s “Grès polymorphes”, this diagram ( Figure 14 ) and that of [23] ( Figure 15 ) showed that these sediments were developed in a passive continental margin tectonic context, corresponding here to the Congo interior intracratonic sedimentary basin.

The compositional variability ICV = [(Fe₂O₃ + K₂O + Na₂O + *CaO + MgO + TiO₂)/Al₂O₃] from [42] shows low values, varying between 0.28 and 0.62 with an mean value of 0.42 which is well below 1. These low ICV values indicate that these sandstones are very mature and were developed in a stable intracratonic tectonic environment [43] .