The quantification of the mineralogical and geochemical differences between saprolite and parent rock requires an extensive petrological investigation of the unweathered and weathered samples. An increasing proportion of quartz sand is indicated by decreasing Ti / Zr ratios, while the opposite is the case with decreasing sand content. The contents of Ti in fresh weathered slates is 0.6 mass % (n=10, SD = 0.1 %), mainly bound in rutile, which is associated with mica (Mosebach, 1954) and therefore narrowly correlated with Kt (r=0.92+++, n=10). Zr with average concentrations of 213 mg kg-1 (n=10, s=39 mg kg-1) in unweathered slates is bound in the heavy mineral zircon. With increasing contents of quartz sand, the Zr contents also increases, which causes the narrow correlation between Sit and Zr (r=0.97+++, n=10). On the other hand the contents of muscovite illite and the associated contents of rutile decrease with increasing contents of quartz (Sit / Kt: r=-0.93+++, n=10; Sit / Ti: r=-0.91+++, n=10). Therefore, the concentrations of Ti and Zr change oppositely with a change of texture of the slates and the ratios can be used as a marker of parent material homogeneity. From the Ti / Zr ratios of 150 samples of fresh slates and saprolites and determination of textural composition, a classification of textural units was possible (Felix-Henningsen, 1994): Ti/Zr>5silty clay slate,Ti/Zr5–4clayey silt slate,Ti/Zr4–2.8sandy silt slate,Ti/Zr<2.8silty sandstone and greywacke.

Mass losses of saprolites compared to the fresh rock result from the export of elements, released from dissolving minerals, due to leaching with percolation or groundwater. Also, mass gains are possible, which result from precipitation of dissolved elements within a saprolite zone. The determination of volume of fragments of rock and saprolite is performed by weighing fragments, saturated with penetrating oil (e.g. WD 40), under submersion in water. The bulk density in g cm-3 results from the ratio of the weight of the dried fragment (105 ∘C) and its volume. The total mass loss is indicated by the difference of bulk densities (volume weights) of fresh rock and saprolite: b.m.l.=b.d. rock-b.d. saprolite⋅100, where b.m.l. is mass loss (g 100 cm-3) and b.d. is bulk density (g cm-3).

A 100 % mass of rock or saprolite is geochemically defined by the sum of main element oxides (determined by XRF, X-ray fluorescence) plus the loss of ignition in mass % (Table 3).

As weathering of saprolites did not change the volume of rock, the comparison of volume-related element masses of fresh rock and saprolites is the best and only way to identify the real extent of element losses and gains in saprolites. Therefore, the masses of individual element oxides per volume (Table 4) have to be calculated as follows: mEm+LOIm=(mEc+LOIc)×b.d., where mEc is the contents of main element oxide (g 100 g-1 = mass %, Table 3), mEm is the volume mass of the main element oxide (g 100 cm-3 = vol %, Table 4), LOIc is the loss of ignition (mass %), LOIm is the volume mass of loss of ignition (g 100 cm-3) and b.d. is bulk density.

Correspondingly, the masses of trace elements can be calculated and balanced.

The summarized volume masses of main elements and LOI account for the bulk density (g 100 cm-3, Table 4): b.d.⋅100=ΣmEm+LOIm.

The differences between the volume masses of main element oxides and LOI, respectively, in rock and saprolite show the individual mass losses and gains of element oxides (and LOI) (Table 5): ΔmEm=mEm rock-mEm saprolite, where ΔmEm is the mass loss or gain of element oxides (g 100-1 cm-3).

The total mass loss as well as individual mass losses of element oxides and LOI of saprolites (Table 5) in relation to the volume weight of element oxides and LOI, respectively, in the fresh rock (Table 4) shows the relative mass loss of element oxides (and LOI) in % (Table 6): ΔmEm(%)=ΔmEm saprolite/mEm rock⋅100.

The individual mass loss of element oxides and LOI in relation to the bulk mass loss (Table 5) results in the relative element oxide composition of the bulk mass loss in % (Table 7): b.m.l.=ΣΔmEm/b.m.l.⋅100, where b.m.l. is bulk mass loss (= 100 %) and ΔmEm is the mass loss of main elements and LOI (g 100-1 cm-3).

Silica makes the main contribution to the bulk mass losses of elements in saprolites due to desilication, which is also an indicator for a warm humid (sub)tropical climate. Potential sources of Si are silicates and quartz. The silicate mineral in slates with the lowest weathering stability is Fe–Mg chlorite, which shows a SiO2 : Al2O3 molar ratio of 1.99 (n=11, after selective extraction with 2 N HCl, 80 ∘C), while kaolinite, which was newly formed from chlorite, shows a ratio of 2. Therefore, a quantitative neo-formation of kaolinite from chlorite did not lead to desilication. Furthermore, near the basis of saprolites, the neo-formation of smectites with a SiO2 : Al2O3 molar ratio of 5.12 (Weaver and Pollard, 1973:67) after weathered chlorites indicates that the dissolution of fine grained quartz must have mainly provided the high concentration of dissolved silica. The contents of quartz was determined by XDA (X-ray diffraction and absorption) according to the procedure described in Till and Spears (1969). The contents of quartz (mass %) is converted into mass per volume (g 100-1 cm-3) and is isovolumetrically balanced (Table 8).

The results show that dissolution and leaching of up to 18 % of the original quartz content in the fresh slate largely contributed to the loss of silica and the bulk mass loss of saprolites.

The contents of kaolinite as the neo-formed mineral after Fe–Mg chlorite (and in part smectite) was determined according to Islam and Lotse (1986). The contents of kaolinite were calculated from the amount of extracted Al after dissolution of the clay fraction in 0.5 N NaOH (80 ∘C, 3 min).

All clay minerals were determined by X-ray diffraction (XRD) of oriented specimen of the natural clay fraction <2 µm, dispersed in water, before and after saturation with ethylene glycol, K+, fumigation with DMSO (dimethyl sulfoxide) and heating (1 h 450 and 550 ∘C).

Well-ordered kaolinites can be identified by a peak-triplet at 20–25∘ 2θ in X-ray diffraction diagrams (Cu radiation). Such kaolinites occur as white monomineralic precipitates and filling of veins in semi-terrestrial saprolites. Kaolinites of paleosols and saprolites (references in Felix-Henningsen, 1990, 2015), which mainly developed topotactically during the alteration of primary silicates, belong to the less ordered type of b-axis disordered kaolinites and fireclay minerals. These fractions can be distinguished by the expansion after intercalation with DMSO. As a result of fumigation of an oriented XRD clay specimen (70 ∘C for 72 h) with DMSO, well-crystalline and b-axis disordered kaolinites, which are abundant (>90 %) in saprolites, expand from 0.72 to 1.12 nm (Fig. 8).

On the contrary, the clay fraction (>60 mass %) of the paleosol shows an increasing proportion (up to 90 %) of fireclay from the saprolite to the former topsoil. Fireclay does not expand after treatment with DMSO (Fig. 8). Therefore, DMSO intercalation allows the sources of kaolinites (from saprolites) and fireclay (from paleosol) in sediments to be identified, and the results support the geomorphological reconstruction of landscape development.

The profile, 25 m deep, exposes a terrestrial saprolite from fresh and nearly fresh slate to the upper, strongly weathered zones. The individual zones are interfingered over distances of many metres. Interfingering is a typical phenomenon in deep weathered slates because the more or less steep inclined to vertical cleavage planes and tectonic joints cause differences in permeability and weathering intensity between jointed and massive parts of the slate.

In the lowermost saprolite zone (Cj5, Table 9), the black-grey colour of the slate changes to an olive colour along cleavage plains and joints as a consequence of oxidation, which increases to the upper saprolite zones (Cj4, Table 9). The loss of primary coaly bituminous organic matter follows the penetration of air into the rock matrix and indicates an increase of porosity in the wake of dissolution of minerals and export of dissolved elements. This is confirmed by mass losses (Fig. 14) as well as the loss of mechanical stability. Within the olive-coloured parts, inner crystalline oxidation transformed the primary chlorites to chlorite–vermiculite mixed layer (m.l.) minerals, which occur beside kaolinite. The Stephan Schmidt KG mines this saprolite zone for use as a greenhouse substrate. The colour of the higher saprolite zones (Cj1–Cj3, Table 9) changes to brown, light and dark red and purple-red, as all iron-bearing silicates weathered completely, and free iron oxides, which precipitated as goethite and hematite, were formed.

Leaching of elements by percolation water under oxidizing conditions caused mass losses of 25–30 % of the original rock mass (Fig. 12) and a loss of stability of the saprolite, which changed to a soft, friable material. Especially Si, Al, bases and to a minor extent also Fe contributed to the mass loss. The loss of dissolved SiO2 and bases was a consequence of weathering of silicates and leaching of dissolved elements under warm (sub)tropical climatic conditions, as the mobility of silica depends only on the temperature of the weathering solution. Because of the desilication the amorphous Si fraction of the saprolite is impoverished relatively to the contents of amorphous Al compounds (Table 11). The losses of total Al2O3 (Fig. 12), dependent on acid conditions (pH<4.5) and complexation by dissolved organic matter, increased above the lowermost saprolite zone, as the buffer capacity decreased after the leaching of bases. This also shows that Al, released from the primary silicates, was not quantitatively captured in neo-formed kaolinite. Restricted to the lowermost saprolite (Cj5, Fig. 12), Fe2O3 contributes to the mass loss as intermittent waterlogging above the massive fresh slate supported the mobilization of iron under reducing conditions. In the upper saprolite zones (Cj1–4, Fig. 12) the mass loss of iron does not increase any more as the oxidizing conditions caused the formation of immobile oxides. The iron balance (Fig. 13) shows that FeII, bound within the chlorite minerals of the fresh slate, was successively oxidized from Cj5 upwards to Cj1. A first step of inner-crystalline oxidation was followed by the release of Fe ions from chlorites and precipitation as pedogenic (free) iron oxides. Their contents rise continuously from the fresh rock to the uppermost saprolite zone with increasing intensity of weathering of silicates (Table 11). Mn oxides are irregularly distributed as their better solubility under acid conditions promoted diffusive redistribution in phases of water saturation of the saprolite matrix under weak reducing conditions and accumulation on joint plains under oxidizing conditions. A total of 50 % of the original volume weight of heavy metals, also mainly bound in chlorite minerals as a substitute for Fe, was leached from the saprolite according to their relative solubility under acid conditions (Fig. 13).

The clay mineral association (Table 12) shows a transformation of chlorites of the fresh slate to chlorite–vermiculite mixed-layer minerals within the olive-grey parts along joints of the lowermost saprolite zone as well as in the olive-grey saprolite. This shows that the weathering of primary silicates as well as leaching of silica and bases is accompanied by the oxidation of the primary coaly bituminous organic matter of the slates.

As the saprolite, exposed in 70 m higher altitude at the Ölkaute excursion site (see Sect. 5.3), was formed under groundwater saturation, the saprolite of the Töpferkaute developed obviously during a younger stadium of deep weathering, perhaps during the Miocene. The tectonical uplift of the Hintertaunus during the Upper Oligocene caused the lowering of the groundwater table and erosion of the Cretaceous to Paleogene saprolite. Therefore, deep weathering continued under terrestrial conditions during the middle Miocene warm humid period.

The saprolite is covered with periglacial layers up to several metres thick, which are the parent material of the Holocene Stagnic Luvisol. The analytical data are displayed in Tables 10–12. The basal layer consists of material from the uppermost saprolite zone, redistributed by solifluction, covered by a middle layer of loess and an uppermost main layer of loess mixed with pumice of the Late Pleistocene Laacher See eruption (Table 9). Therefore, the texture of the middle and upper layer is rich in coarse silt, typical of the large proportion of loess. The basal layer shows high clay contents as a consequence of intensive periglacial congelifraction of redistributed saprolite material and the admixture of substrate of a periglacially reworked paleosol rich in kaolinitic clay, which is indicated by the high total amounts of kaolinite in the clay fraction (Table 12). They also contain weakly weathered three-layer silicates, which in the main and middle layers derive from loess. In the basal layer, they originate from redistributed material of less weathered, lower saprolite zones, which were exposed by denudation in upslope areas. Due to congelifraction, cryoturbation and lateral transport the material was admixed to solifluction layers.

Since 1866, white, bleached saprolite has been mined in the Ölkaute quarry, near the village of Eisenbach, as raw material for the ceramic industry. The exposed saprolite from silt slate (Table 13) presents the transition of the bleached horizon (mCewj), up to 10 m thick, to a grey saprolite (mCorj) and further down a black-coloured saprolite (mCrj). In both zones the undisturbed rock structure of the slates is preserved, although the saprolite is strongly weathered, friable and can be disintegrated manually. According to the pedogenic iron forms (Table 15) and clay mineralogical composition (Table 16) differences between the saprolite zones exist.

The white colour of the mCewj zone (Figs. 15 and 16) proves that the dissolution of all iron-bearing silicates and the formation of kaolinite, as well as the leaching of bases, silica and metal ions must have occurred under reducing conditions as a result of saturation with migrating ground water. Therefore, Fe and Mn ions, released from silicates, remained mobile and were completely leached with a lateral groundwater flow. The decay of the primary organic matter in the bleached horizon, which until today is still present in the lower saprolite zone with sustainable reducing conditions, is a consequence of oxidation after the lowering of the groundwater table. This went along with the tectonic uplift of the Hintertaunus area and regression of the nearby sea since the Upper Oligocene. In addition, arid climatic phases existed during the Upper Oligocene and Lower Miocene and led to the formation of silcretes in Tertiary river sediments of the Rhenish Massif, as demonstrated in the Ölkaute. The deep exsiccation of the saprolite must have contributed to the oxidation of the upper zones of the saprolite.

Bulk mass losses are clearly higher within the bleached and reduced zones (Fig. 14) than in the terrestrial saprolite of Eisenbach Töpferkaute (Fig. 12). Because of leaching under reducing conditions, iron and other metal ions are strongly depleted (Fig. 17). Therefore, iron oxide contributes more than 25 % to the bulk mass loss. The relative proportion of SiO2 in the bulk mass loss increases from the lower to the uppermost saprolite zone and indicates the increase of desilication. As only the illite minerals weathered to kaolinite while muscovite remained stable, the losses of K2O are rather low.

A former erosion gully, several metres deep (Fig. 16), which extended from the former plantation plain downwards to a valley, cut through the bleached saprolite and was filled with soil material of a red-white mottled Plinthosol after the soil has been undercut by linear erosion of the soft saprolite. It broke down and was mixed with fragments of the bleached saprolite. The paleosol material has a clay content around 70 mass % (Table 14), typical of Cretaceous–Paleogene fersiallitic paleosols. The reason for such a high clay content is the weathering of muscovite remaining stable within the saprolite and the transformation to kaolinite (65–70 mass %) and illite (30–35 mass %). Up to 50 % of the kaolinite fraction consists of fireclay. The negligible contents of fireclay within the saprolite shows that the formation of fireclay is typically bound to fersiallitic and ferrallitic paleosol horizons, which were subject to strong desilication.

From the characteristics of the soil sediment and the analytical data the processes of soil formation, such as acidification, weathering of silicates and desilication and neo-formation of kaolinite, can be concluded. The isovolumetric balance shows that the volumetric contents of Fe2O3 in the soil sediments match those of the fresh slate (Fig. 14). Iron obviously was not leached from the paleosol. This could be a consequence of permanent oxidizing conditions within the soil zone, which did not allow an enhanced mobility of iron. The fact that the underlying saprolite is nearly free of mineralogically bound iron and completely depleted in free iron oxides leads to two hypotheses: a.

If bleached saprolite became the parent material of the soil after phases of erosion of previous soils (e.g. due to Paleogene climate changes or tectonic uplift), the contents of iron oxides must have been supplied by precipitation of iron in the capillary seam of the groundwater. This seems to be less possible, as the saprolite below is completely depleted in iron, which means that the iron concentration of the groundwater was extremely low.

The western wall of the saprolite open-cast mine cuts a bank of fluvial sediments, superimposing the soil sediment fill of the former erosion gully and the adjacent bleached saprolite. The sediments consist of partly silicified and cemented pure quartz sand with gravel of well-rounded milky quartz, which derived from erosion of less weathered primary quartz veins within the saprolite. Less rounded boulders of Taunus quartzite, up to 40 cm in diameter, indicate a long distance transport of the fluvial sediments because the next sources of Taunus quartzite are located more than 3 km away in a south–south-east direction. They must have been sedimented in the Upper Oligocene Vallendar river system on the Paleogene planation plain, before the tectonic subsidence of the fault blocks and the incision of the Late Tertiary and Quaternary valleys, separating the planation plane around Ölkaute from the High Taunus. The isovolumetric balance of elements (Fig. 14) shows a strong absolute accumulation of silica, indicating silification as a consequence of (semi-)arid climate conditions, which existed throughout Middle Europe in the transition from the Upper Oligocene to the Lower Miocene.

The uppermost layers, 2 m thick, above the Tertiary fluvial sediments consist of periglacial layers, organized in basal, middle and main layers (Table 13).

The basal layer is free of loess but rich in clay (Table 14) and consists of a more or less brown to white loam, which derived from congelifraction of the saprolite. During thaw periods the porous saprolite was saturated with water, which caused rapid congelifraction during the frost periods. The disintegration of kaolinite aggregates (booklets), which tend to develop as pseudomorphoses of weathered silicates, severely enhanced the clay content. Admixture of saprolite zones and paleosol material increased the contents of pedogenic iron. The contents of amorphous silica and aluminum correlate with the contents of clay (Table 15). The middle layer is dominated by loess but contains also a proportion of material of the basal layer, rich in clay. The pedogenic SBt horizon characterizes a Stagnic Luvisol. Micromorphological analyses show that the majority of the clay cutans were destroyed by periglacial frost pressure and therefore occur as fragments and coatings as zones within the loess matrix, apart from recent voids. Thus, a pre-Holocene formation of a relic Bt horizon is probable. The main layer consists of loess mixed with pumice of the Laacher See eruption and shows clay accumulation in its basal part and impoverishment of clay in the upper part as a consequence of Holocene weathering and clay migration.