Lower Saprolite Research Articles

Regolith-Hosted Rare Earth Element Mineralization in the Esperance Region, Western Australia: Major Characteristics and Potential Controls

A number of regolith-hosted REE occurrences have recently been discovered in the Esperance region in southern Western Australia. This paper summarizes major characteristics of REE mineralization and discusses contributing factors and potential controls. The main aim is to explain why there is a lack of highly sought-after ion-adsorption-clay-type REE deposits across the region despite the presence of the regolith-hosted REE mineralization on a regional scale. Local mineralization mostly occurs as continuous flat-lying enrichment “blankets” within the residual regolith developed over Archaean–Proterozoic granite gneisses and granitoids with elevated REE content. The enriched horizon is commonly located in the lower saprolite and saprock and is accompanied by an overlying REE-depleted zone. This distribution pattern, together with the data on HREE fractionation and the presence of the supergene REE minerals, indicates chemogenic type enrichment formed by supergene REE mobilization into groundwater, downward transport, and accumulation in the lower part of the weathering profile. Residual REE accumulation processes due to bulk rock volume and mass reduction during weathering also contribute to mineralization. It is proposed that climate and groundwater chemistry are the critical regional controls on the distribution of REEs in the weathering profile and on their speciation in the enrichment zone. Cenozoic aridification of climate in southwest Australia heavily overprinted pre-existing REE distributions in the weathering profile. Acidic (pH < 4), highly saline groundwaters intensely leached away any relatively weakly bound, adsorbed or colloidal REE forms, moving them downward. Dissolved REEs precipitated as secondary phosphates in neutral to alkaline environment at lower Eh near the base of the weathering profile forming the supergene enrichment zone. Low denudation rates, characteristic of areas of low relief under the arid climate, are favourable for the preservation of the existing weathering profiles with REE mineralization.

Nickel isotope fractionation during intense weathering of basalt: Implications for Ni output from continental weathering

Nickel isotope fractionation during continental weathering is not well constrained due to the limited research on Ni isotopes during weathering of mafic rocks such as basalt. This study investigated two weathering profiles (strongly and extremely weathered) derived on basalts in tropical China. The strongly weathered profile was dominated by kaolinite and divided into three layers with depth (soil, saprolite and rock). The extremely weathered profile was mainly composed of Fe- and Al-(oxyhydr)oxides and kaolinite, and included a special ferricrust layer, iron-rich duricrust with ferruginous nodules, between the soil and saprolite. We analysed the Ni speciation (e.g. Fe-oxide phases and residual phase) and isotopic composition of soil, ferricrust, saprolite, rock and water samples in the weathering profiles. Pore water and groundwater were collected underlying the extremely weathered profile, and the size-distribution of particulate Ni (1 kDa–450 nm) in water samples was analysed.Nickel in both profiles was mainly hosted in the residual phase (>71% of total Ni) which included primarily secondary silicates and Al-(oxyhydr)oxides. Weathering intensity controlled the mineral compositions and physico-chemical properties of the weathering profiles as well as Ni isotopic variations. Nickel redistribution (∼30%) and isotope fractionation (Δ60Niregolith-rock = −0.08‰ to 0.07‰) were slight in the strongly weathered profile due to the dominance of secondary silicates in Ni speciation. For the extremely weathered profile, significant Ni depletion (∼67%) and isotope fractionation (Δ60Niregolith-rock = −0.25‰ to 0.32‰) were related to the transformation of secondary silicates into Fe- and Al-(oxyhydr)oxides during lateritic weathering. The positive correlation of δ60Ni and Ni concentrations indicated that Ni isotope fractionation was controlled by Ni adsorption and incorporation of secondary minerals. In the lower saprolite, the calculated isotope fractionation between Ni remaining in the profile and leaching to the solution (Δ60Nisolid-solution = −0.47‰) indicated that Ni isotope fractionation was mainly controlled by the preferential adsorption of isotopically light Ni on Fe- and Al-(oxyhydr)oxides. However, in the soil and ferricrust, acidic pH (∼5.0) might reduce Δ60Nisolid-solution to −0.28‰ by releasing the isotopically light adsorbed Ni, leading to the predominance of incorporated Ni in Fe- and Al-(oxyhydr)oxides. Specifically, the highest δ60Ni value (0.27 ± 0.05‰) and Ni concentration in the upper saprolite indicated that the isotopically heavy Ni leached from the soil and ferricrust was scavenged by precipitating secondary silicates. Finally, the estimated isotopic composition of Ni leached from the extremely weathered profile (δ60Nioutput = 0.02 ± 0.15‰) was close to that of pore water (0.06 ± 0.01‰) where Ni was mainly present in the particulate phase. Compared with groundwater (δ60Ni = 0.33 ± 0.04‰) dominated by dissolved Ni, Ni output from the profile likely consisted of isotopically heavy dissolved Ni and light particulate Ni. Overall, this study provides a comprehensive understanding of Ni redistribution and speciation in basalt weathering profiles, highlighting the significance of weathering intensity in controlling Ni isotope fractionation and its implications for Ni outputs from continental weathering.

Mineralogy and geochemistry of the Morro do Engenho lateritic nickel deposit, Goiás Alkaline Province, Brazil

The Morro do Engenho lateritic nickel deposit results from the action of weathering processes on the alkaline mafic-ultramafic rocks that make up the Cretaceous zoned intrusive body belonging to the Goiás Alkaline Province. Historically, the deposit is classified as a silicate one, and mineralization is present especially in phyllosilicates such as serpentine and chlorite. By identifying Ni-bearing phases and providing the lithogeochemical characterization of weathering profiles, this study aims to provide further insights into the metallogenetic processes and controls acting on the mineralization of the Morro do Engenho lateritic Ni deposit. The following methods were used: petrographic studies, SEM-EDS, XRD and lithogeochemical characterization of the weathering profiles developed on each of the parent rocks that make up the body. The Morro do Engenho intrusive massif is formed by a dunite core, surrounded by a peridotite/pyroxenite zone, an alkaline gabbro zone and more externally by a syenite-nepheline zone. The lateritic profile constitutes eight weathering horizons, stacked one upon another from the bedrock namely: (i) saprock, (ii) lower saprolite, (iii) ferruginous saprolite, (iv) ocher saprolite, (v) plasma zone, (vi) lateritic duriccrust, (vii) silcrete and (viii) topsoil. Most Ni mineralization occurs in the lower saprolite, ferruginous saprolite and ocher saprolite zones. The findings pointed to the presence of Ni in silicate mineral phases, represented by serpentine, chlorite, smectite, as well as in Fe and Mn oxyhydroxides. The depressed area, located between the two major topographic elevations, has the highest Ni enrichments, which according to mass balance calculations can account for up to 40%. Hydrodynamic factors, by dissolution/precipitation of silicates in saprolite horizons of dunite core, are suggested as possible agents in the remobilization of Ni, among other mobile elements, to topographically flat zones. The geometry of the Morro do Engenho deposit show that lateral redistribution by hydrodynamic factors could also account for the mineralization of areas peripheral to the dunite core, composed of pyroxenite and gabbro-derived lateritic profiles. The direction of regional groundwater flow, strongly influenced by topography, would also be an important factor in the remobilization of Ni, as well as the mechanical transport of lateritic materials. Silicate mineralization predominates in the Morro do Engenho deposit, concentrated in Mg hydrosilicates and clay minerals; oxide mineralization also occurs, mostly in Fe and Mn oxyhydroxides. The findings discussed in this paper can broaden the understanding of metallogenetic processes involved in the genesis of lateritic nickel deposits. .

Genesis and 40Ar/39Ar dating of K-Mn oxides from the Stavelot Massif (Ardenne, Belgium): Insights into Oligocene to Pliocene weathering periods in Western Europe

The occurrence of Ordovician Fe-Mn-rich sediments (Ottré Formation, Salm Group) in the Stavelot inlier (NE Ardenne, Rhenish Massif, Belgium) has drawn the attention of geoscientists for more than one century, as they contain natural “coticule” whetstones and numerous weathered Mn deposits. Two locations expose the Fe-Mn-bearing rocks: the Chevron area to the north is located at an altitude of 230–380 m, and the Salmchâteau at 480–570 m. The Mn-bearing sediments have undergone greenschist facies metamorphism in relation to the Caledonian and Variscan orogenesis, and more recently weathering processes, transforming the early mineral assemblages (mainly rhodochrosite and/or kutnohorite, spessartine, muscovite, chloritoïd, andalusite, hematite) into secondary Mn oxides,—hydroxides and—oxyhydroxides (cryptomelane, nsutite, lithiophorite, manganite, pyrolusite), clays (mainly kaolinite) and Fe oxides (goethite). In this paper, we investigate the weathering processes of these Mn-rich rocks by i) new mineralogical, petrographical and geochemical data, and ii) new 40Ar/39Ar geochronology of K-Mn oxides.Petrographical and geochemical data indicate that rhodochrosite (and/or kutnohorite) is the main source of Mn, from which Mn oxides have precipitated. Spessartine and other Mn-silicates have poorly contributed to the Mn mobilization. Pure Mn oxides, namely manganite, nsutite, and pyrolusite, successively replace rhodochrosite under increasing O2 conditions due to the weathering fluids, while the formation of cryptomelane and then lithiophorite in veins is associated to the recombination of K+, Li+ and Al3+ released from the phyllosilicate matrix of the host slate, probably in relation with sligh pH changes. This multistage character of weathering processes is well identified in the Salmchâteau area, where weathering conditions are stronger than in the Chevron area because of the elevation. This also positions the Mn levels close to the roll front within the lower saprolite and the fissured horizon, respectivelly. Such differences in the weathering grade of these areas determined a weathering-type deposit in the Salmchâteau area and a carbonate-hosted sedimentary deposit in the Chevron area. Weathering deposits are accompanied by an enrichment in Mn and associated transition metals (As, Co, Cu, Ni). Dating of pure cryptomelane samples corroborate the multistage character of the ore and yield at least four weathering stages: Chattian (26.33 ± 0.17 Ma), Aquitanian (22.2 ± 0.6 and 20.0 ± 0.2 Ma), Serravalian-Tortonian (11.79 ± 0.10 to 10.12 ± 0.18 Ma) and Pliocene or younger (5.30 ± 0.76 and 4.11 ± 0.57 Ma). A relatively continuous weathering of the Massif since the late Oligocene, or older weathering phases, cannot be excluded as the upper part of the saprolite (and laterite) have been removed. As climatic conditions evolve from tropical to temperate by the Oligocene, the uplift of the Ardenne is probably the main factor triggering the development of weathering mantle, together with the increasing seasonality and high precipitation. These ages with others obtained in supergene ores of Western European massifs define a long-lasting weathering wave since the late Oligocene until the uppermost Neogene. The absence of older periods in the 40Ar/39Ar record of these deposits indicates that the upper part of the weathering mantle was recently stripped and limit their economic interest.

Importance of 3-D regolith-landform control in areas of transported cover: implications for geochemical exploration

Two-dimensional (2-D) regolith-landform mapping has gained greater importance and wider acceptance and usage in the minerals industry over the last 20 years. This 2-D method is generally adequate in residual areas but is a poor indicator of the complexity of the deep regolith and thickness of transported cover in depositional environments. This paper highlights the importance of three-dimensional (3-D) regolith-landform control as an aid to: (a) identify major controls on metals dispersion; (b) assist anomaly interpretation; and (c) selection of suitable sample media in areas of transported cover from the Yandal greenstone belt (YGB; 260 × 40 km) of the Yilgarn Craton. It represents the first fully integrated study of the regolith on a single greenstone belt. This greenstone belt has been intensively drilled and its regolith units systematically logged. The YGB has less than 10% fresh and weathered rock outcrop and c . 90% of the area is covered by deep regolith consisting of weathered bedrock (locally up to 160 m) under sediments (up to 100 m thick). The application of 3-D mapping in the YGB is restricted to areas of precise and accurate subsurface regolith data-sets comprising c . 50,000 drill-holes covering c . 70% of the belt. The data-sets for the 3-D mapping comprise transported cover, oxidized saprolite, lower saprolite and bedrock and gold geochemistry. The regolith surfaces show that the variability of transported cover, the depth of the oxidized saprolite and the lower saprolite are controlled by the palaeorelief, lithology, mineralization and deformation. The nature and thickness of transported cover determines the suitability of a particular sampling medium. Thus, prototype derivative maps showing residual areas and areas with varying thickness of transported cover (e.g. <5 m, 5–20 m, >20 m) are required for devising a suitable sampling strategy program in depositional environments. Three-dimensional relationships between regolith and gold distribution both at regional and deposit/prospect scales have provided information that may indicate the size of geochemical dispersion and suggest sampling strategies in different transported cover domains. It also allows testing of possible vectors to mineralized sources for anomalies in transported materials. An important feature of the palaeosurface reconstructions below deep transported cover is its considerable relief which shows striking anomalies at a number of deposits and prospects in the YGB. This reinforces the importance of precisely defining this palaeosurface through 3-D mapping as it may host the best direct geochemical indicators of buried mineral systems in areas of deep cover.

Regolith formation rate from U-series nuclides: Implications from the study of a spheroidal weathering profile in the Rio Icacos watershed (Puerto Rico)

A 2m-thick spheroidal weathering profile, developed on a quartz diorite in the Rio Icacos watershed (Luquillo Mountains, eastern Puerto Rico), was analyzed for major and trace element concentrations, Sr and Nd isotopic ratios and U-series nuclides (238U–234U–230Th–226Ra). In this profile a 40cm thick soil horizon is overlying a 150cm thick saprolite which is separated from the basal corestone by a ∼40cm thick rindlet zone. The Sr and Nd isotopic variations along the whole profile imply that, in addition to geochemical fractionations associated to water–rock interactions, the geochemical budget of the profile is influenced by a significant accretion of atmospheric dusts. The mineralogical and geochemical variations along the profile also confirm that the weathering front does not progress continuously from the top to the base of the profile. The upper part of the profile is probably associated with a different weathering system (lateral weathering of upper corestones) than the lower part, which consists of the basal corestone, the associated rindlet system and the saprolite in contact with these rindlets. Consequently, the determination of weathering rates from 238U–234U–230Th–226Ra disequilibrium in a series of samples collected along a vertical depth profile can only be attempted for samples collected in the lower part of the profile, i.e. the rindlet zone and the lower saprolite. Similar propagation rates were derived for the rindlet system and the saprolite by using classical models involving loss and gain processes for all nuclides to interpret the variation of U-series nuclides in the rindlet-saprolite subsystem. The consistency of these weathering rates with average weathering and erosion rates derived via other methods for the whole watershed provides a new and independent argument that, in the Rio Icacos watershed, the weathering system has reached a geomorphologic steady-state. Our study also indicates that even in environments with differential weathering, such as observed for the Puerto Rico site, the radioactive disequilibrium between the nuclides of a single radioactive series (here 238U–234U–230Th–226Ra) can still be interpreted in terms of a simplified scenario of congruent weathering. Incidentally, the U–Th–Ra disequilibrium in the corestone samples confirms that the outermost part of the corestone is already weathered.

Micromorphology, mineralogy and geochemistry of lateritic weathering over serpentinite in south-east Cameroon

A laterite profile developed on serpentinite in Lomié, south-east Cameroon, was studied to understand its structure and genesis. Micromorphological, mineralogical and geochemical characteristics were studied, using thin sections, SEM-EDS, X-ray diffraction and chemical analyses. In the lower saprolite, magnesite is dissolved and serpentine is gradually replaced by kaolinite, without loss of the lithic fabric (orientation of kaolinite, distribution of magnetite). This nearly iso-volumetric weathering, at the scale of the sample, can be considered as in situ replacement of serpentine by kaolinite. Iron oxide hypocoatings formed along planar voids, and also smectite formed from solution in the void system. Congruent dissolution of kaolinite starting in the mid saprolite eventually led to a collapse of the fabric and to a chaotic accumulation of fragments of the hypocoatings. In the overlying layer, this fabric is even more strongly expressed, and iron oxides are more concentrated. The weathering profile shows a severe loss of Mg and Si, and an accumulation of Al, Fe, Ti and some minor and trace elements. The overlying layers have fabrics and components different from those recognized for the saprolite, such as clastic quartz grains and iron oxide nodules with various lithic fabrics, pointing to an admixture of allochthonous materials. The appearance of mobile elements (e.g. Rb, Sr, Pb) and certain stable elements (Zr, Nb), as well as the removal of serpentinite-related elements clearly show the mixing with materials derived from micaschists in the vicinity. The presence of a discontinuity in the profile is also confirmed by variations in iron oxide mineralogy. At least three lithological discontinuities were identified.

Geochemistry and mineralogy of the regolith profile over the Aries kimberlite pipe, Western Australia

Residual regolith over the Aries kimberlite pipe, situated in the central Kimberley Block of Western Australia, consists of lateritic duricrust, mottled zone, upper saprolite, lower saprolite and saprock. Duricrust and mottled zone are dominated by goethite and kaolinite while saprolite consists mostly of smectite with minor amounts of mica and serpentine. Most kimberlite indicator minerals except chromite have been weathered. Kimberlitic rock fabric is preserved in the regolith up to the top of saprolite. Chromium, Ni, Co, Cu, Nb, Zr, Ti, P and rare earth elements (REE) are significantly enriched in the mottled zone and duricrust relative to the saprolite and fresh kimberlite. Vanadium, Sb, As, Mo and U, derived from the sediments and basaltic rocks around the pipe, are also enriched. Goethite partly hosts Cr, Ni, and Co, and Nb and REE are associated with anatase. The clay fractions of the mottled zone and saprolite contain significant quantities of Cr, Co, Cu, Ni, Zr, Ga, Ti, Nb, Zn and REE. A suite of six trace elements comprising those associated with ultramafic rocks (Cr and Co) and felsic rocks (Nb, La, Sm and P) can readily distinguish the kimberlitic laterite from that derived from ultramafic or felsic rocks.

Confined groundwater zone and slope instability in weathered igneous rocks in Hong Kong

In Hong Kong, slope failure in weathered granite and volcanic rocks has been studied extensively, but the impact of hydrogeological conditions on slope stability has received much less attention. It is customary in Hong Kong to assume that the hydraulic conductivity ( K) of weathered igneous rocks decreases with depth or as the rock mass becomes less weathered. Generally studies of groundwater for determining slope stability treat the saprolite above the rockhead as an aquifer and the rockhead as an impermeable boundary. This paper examines direct and indirect aquifer hydraulic information scattered in various sources and demonstrates that such a model may be inappropriate. Evidence is cited to show that, at least in some places, a relatively high hydraulic conductivity ( K) zone (HKZ) may exist at depth, either in the lower saprolite or at the rockhead. The completely decomposed igneous rocks, if significantly kaolinized, may have low permeability and behave as a confining zone for the HKZ. The groundwater flow regime in such an HKZ may be confined, resulting in a higher rise in water pressure in response to rainfall than might otherwise be expected. These high pressures may result in a significant reduction in slope stability. An examination of the cross-sections of some well-known landslides in Hong Kong reveals a chair-shaped rockhead profile. Numerical models coupling saturated and unsaturated subsurface flow are used to investigate the saturation process in the presence of an HKZ in slopes in response to typical rainstorms. The simplified Bishop's method is used to calculate changes in the factor of safety for slopes during such rainstorms. The saturation process and pressure distribution in a slope are complicated by the presence of an HKZ. An HKZ that is blocked at the toe represents the worst hydraulic condition for slope stability. The paper concludes that, although slope stability studies in Hong Kong have tended to focus on perching within the regolith, a critical hydraulic boundary condition may exist locally at the base of the regolith.

Multi-element dispersion within salt-lake environments: case study of the buried Hannan South gold deposit, Western Australia

Prior to mining, the Hannan South deposit contained 0.26 Mt at 4.98 g t −1 Au confined within a weathered, highly saline water-saturated regolith beneath a salt lake. There was no surface expression, although the host rock sub-cropped at 3–4 m depth. The water table seasonally fluctuates between 0 and 10 m depth. The rock is a highly altered and structurally disrupted basalt, now a sulphidic epidote-rich unit. This unit is truncated at 80 m depth within sub-vertically interlayered Archaean felsic sedimentary and intermediate volcanic rocks. No fresh host rock is preserved. There is a sharp sub-horizontal iron redox front at 25–30 m depth between upper and lower saprolite. At this front, Eh conditions are sufficient to oxidize dissolved Fe 2+ from weathered Fe-bearing minerals in the lower saprolite, to Fe 3+ with the production of strong acid solutions (pH 0.5–2). Acid-enhanced weathering forms the upper saprolite as the front descends through the profile. As the pH increases, Fe 3+ precipitates as Fe oxides, mainly goethite, above the front. Primary Au containing 3–5 wt% Ag is hosted within pyrite and accessory sulphide minerals of As, Cu, Co, Pb, Bi, W and Zn in the saprock. The progressive oxidation of the sulphides has released soluble ore-associated elements, including Au as a thiosulphate complex (Au–Ag(S 2 O 3 ) 2 3− ). This complex is metastable and has precipitated fine-grained secondary Au containing 3 wt% Ag around pseudomorphs of the host sulphides. The other elements are widely dispersed but are largely confined to the host rock in the lower saprolite due to limited groundwater flux. Poorly defined acid sulphide and subtle Mn redox fronts within the lower saprolite have little influence on trace element dispersion. Silver is the most strongly depleted. In the upper saprolite, descent of the redox front over time has led to leaching of Au under acid, oxidizing conditions. Gold has probably dissolved as organic or chloride complexes at the front, dispersed laterally down-gradient and precipitated as pure crystals (99.9 wt% Au) in a sub-horizontal supergene enrichment blanket just above the front. Copper, Ag, Co, Pb and Zn are also strongly leached from the acidic saprolite above the iron redox front and are dispersed and depleted with increasing groundwater flux in the upper profile. In comparison, As, Sb, Bi, Mo and W, which form oxyanions, are relatively immobile and are preferentially retained in goethitic zones. These buried multi-element halos hosted by goethite provide near-surface geochemical exploration targets within otherwise barren upper saprolite.

Gold mass balance in the regolith, Mystery Zone, Mt Percy, Kalgoorlie, Western Australia

Mineralized rocks at the Mystery Zone gold deposit, Mt Percy, are overlain by an almost complete lateritic regolith over 60 m thick, within which there are significant variations in Au concentration. This includes an apparent Au depletion in the upper saprolite and clay horizons and enrichment in the surficial residuum and soil. Gold leaching up through the regolith has been quantified by mass balance calculations. In the lower saprolite, 55% of the Au has been leached from the porphyries, but Au is essentially immobile in ultramafic rocks. Higher in the regolith, leaching trends are similar over both lithologies, with 74–90% depletion in the clay saprolite, and 84–95% depletion in the plasmic and mottled clay horizons. Within the surficial lateritic residuum and carbonate, Au concentrations increase to significantly greater than bedrock grade, but the calculations show there is little, if any, absolute Au enrichment, compared to the fresh rock. Silica is progressively depleted through the regolith, with more leaching over fuchsitic ultramafic rocks. Aluminium tends to remain stable in the saprolite. However, Fe shows contrasting mobility during weathering of the different lithologies. For ultramafic rocks, there is a strong absolute Fe depletion in the clay saprolite, with minor depletion in the overlying mottled and plasmic clay. In contrast, within weathered porphyries, Fe is absolutely accumulated in the most regolith horizons, possibly representing reprecipitation of Fe dissolved from adjacent ultramafic rocks. Results from this investigation agree well with a previous Au mass balance study based on isovolumetric calculations. However, differences higher in the regolith may be due to partial collapse causing errors in earlier calculations.

The mobility of gold in tropical rain forest soils

The complex mineralogical and chemical system of a tropical rain forest weathering profile at the Ashanti mine, Ghana, was characterized to evaluate the role of inorganic and organic processes with the potential to mobilize gold in the rain forest soils. The concentration of gold has been measured in bulk saprolite, soil samples, and in pore-water extractions from the saprolite and soil horizons from the Sansu prospect at Ashanti. Gold concentration in the ground waters varied from below the detection limit (0.01 mu g/l) to a maximum of 84.4 mu g/l at the water table. Average gold concentration in the soil pore waters is 3.6 mu g/l. The gold in the pore waters is present as either soluble gold or as suspended particles less than 0.45 mu m in size.Despite the presence of significant gold concentrations in the soil pore waters, gold concentration in soils over the auriferous shear zones is up to 400 times greater than over the nonauriferous country rocks suggesting a limited gold mobility. Similarly, a close correlation between gold held in the solid soil fraction and in soil pore waters reveals very little in the way of supergene gold enrichment. Conversely, the gold is dissolved and reprecipitated over a range of a few centimeters. The high concentration of gold in the soil pore waters at Ashanti is a reflection both of the high gold content of the hypogene ore and of the efficiency of supergene fluids in the tropical rain forest environment to leach the gold but not retain it in a mobile form. Solubility calculations indicate that gold complexing in the soils appears to be dominated by AuOH(H 2 O), Au(CN) 2 (super -) , Au(S 2 O 3 ) 2 (super -3) , Au(NH 3 ) 2 (super +) , and a gold fulvate complex. Selective extraction of gold from organic and inorganic soil fractions reveals that the element is largely concentrated as native gold. A significant proportion (up to 28%) is concentrated in the light organic fraction of the near-surface soils which is dominated by fulvic acid. The role of gold adsorption by clays and hydrous Fe-Mn oxides appears to be small with less than 1 percent of the gold being present in an easily exchangeable form.Gold appears relatively immobile in the lower saprolite with gold values in solution being below the detection limit (0.01 mu g/l). Higher up in the saprolite, gold appears to be dissolved by hydroxide and thiosulfate complexes in the upper saprolite and by fulvate, cyanide, and hydroxide complexes in the soils. Due to the in situ nature of gold dissolution and reprecipitation, even short-lived species such as thiosulfate and cyanide can play an important role in the supergene chemistry of gold in the Ashanti soils. Halide complexes are not considered important owing to their low concentration in the ground waters (Cl, <9 mg/l) and the unrealistically high Eh-low pH conditions required. Humic acids are also not considered important in the dissolution of gold as they are only sparingly soluble in the pH range of the Ashanti soils (4.2-6.8).

Gold mass transfer during lateritic weathering under equatorial rainforest conditions

At Dondo Mobi (Gabon), under equatorial rainforest conditions, gold dispersion haloes form from weathering of auriferous lisvenite and spreads laterally within Proterozoic schistand Archean gneiss-derived surficial layers (Fig. 1 ). Within the lower saprolite, by eye gold is closely associated with auriferous parent rock. From petrological study of host rocks and gold particles themselves, it has been demonstrated that the surficial dispersed visible gold consists of residual gold particles inherited from the lisvenite source, and that