Includes bibliographical references (leaves 142-160). / The Cullinan kimberlite is a Group I kimberlite and is located in the northeastern region of the Kaapvaal Craton, South Africa. The kimberlite pipe has been dated at 1180 ± 30 Ma and intrudes the Bushveld Igneous Complex (2.05 Ga). This study explores the geochemistry of a suite of one hundred selected diamonds and their associated mineral inclusions. The majority of the diamonds described here are peridotitic (94%) and the remainder are eclogitic. The peridotic inclusions may be further subdivided into harzburgitic and lherzolitic parageneses.

Includes abstract. / An O, H, Sr and Nd isotope study was carried out on the ~ 2059 Ma Bushveld granites and granophyres. A small number of Rooiberg Group felsites were also studied.

Includes bibliographical references (leaves 97-115). / A suite of sulphide-bearing diamonds recovered from the Ellendale 4 and 9 pipes in the Ellendale diamond province of lamproite intrusions in north-western Australia have been investigated for their nitrogen aggregation state and the Re-Os isotope geochemistry of the sulphide inclusions. The Ellendale lamproites, dated at ca. 20 Ma (Allsopp et at., 1985) intrude the King Leopold Mobile Belt just south of the Kimberley craton and are thus an atypical, off-craton diamondiferous locality. The diamonds contain roughly equal proportions of peridotitic and eclogitic inclusions (Hall & Smith, 1984; Griffin et at., 1988; Jaques et at., 1989). The diamonds in this study range in size from 0.26 to 0.92 carats and are dominated, with the exception of one octahedron, by highly resorbed tetrahexahedroida.

This is a study of Phanerozoic carbonatites and related alkaline silicate rocks from Zandkopsdrift in Namaqualand, South Africa (55 Ma), and Marinkas Quellen (529 Ma), and Dicker Willem (49 Ma) in southern Namibia, all emplaced within the Proterozoic Namaqua Natal province. The aims of this thesis are to characterise their mantle source region, including the timing and nature of the metasomatism that affected their sources, and to constrain their petrogenesis, particularly the relationships between the carbonatites and associated silicate igneous rocks. These associated silicate igneous rocks include olivine melilitite, aillikite and alkaline lamrophyre at Zandkopsdrift, nepheline syenite and trachyte at Marinkas Quellen, and ijolite and trachyte at Dicker Willem. At both Marinkas Quellen and Dicker Willem, the trachytes appear to be derived primarily from fenitised country rock. The major and trace element characteristics, enrichment in LREE and other incompatible elements, large Zr-Hf-Ti depletions and high Zr/Hf ratios all appear to have been inherited from a metasomatised mantle source region rather than being the result of residual source mineralogy. The δ18O and δ13C values of carbonate in the study locations vary significantly. The δ13C values (-3.9 to -8.8 ‰) are within the range of mantlederived carbonatites, whereas the δ18O values are often significantly higher (+8.64 to +22.22 ‰, versus SMOW) for “mantle-derived” carbonatites. The higher δ18O values observed are most likely attributable to low-temperature, post-emplacement alteration by hydrous fluids. O-isotope thermometry of the silicate mineral pairs (clinopyroxene, amphibole, and biotite) gives close to magmatic temperatures (≈800 oC), whereas the equilibration temperature of calcite-silicate mineral pairs is lower. The carbonatite and associated igneous rock samples contain unradiogenic Sr and mildly radiogenic Nd isotope compositions below and above Bulk Earth/CHUR values respectively. This suggests that carbonatite magmas were generated from sources with long-lived Rb/Sr lower than, and Sm/Nd higher than, the primitive mantle. In ƐHf(t)-ƐNd(t) space, the carbonatites and associated silicate rocks plot as much as 8 ƐHf units below the terrestrial ƐNd-ƐHf array, indicating mixing of a source with moderate ƐNd and exceptionally unradiogenic Hf isotope compositions. The radiogenic Pb isotope composition of the carbonatites (206Pb/204Pbi ratios from 18.06 to 22.38), is consistent with a source having high U/Pb, akin to the HIMU mantle end member. The radiogenic isotopes of the carbonatites and the alkaline silicate rocks seem most consistent with a dominantly asthenospheric source, with minor contributions from lithospheric sources. There is little evidence supporting the derivation of carbonatites at the three complexes from parental hybrid carbonate-silicate magmas. Rather, the evidence seems most consistent with deriving the carbonatites directly from very low-degree mantle melts, which subsequently become variably differentiated, first by melt-rock interaction in the mantle and subsequently by fractional crystallization. The associated silica undersaturated silicate igneous rocks appear most likely to represent relatively primitive (melilitites) to differentiated (nepheline syenites) melts of metasomatic wehrlites that were formed by carbonatite-peridotite melt-rock interaction.

The Southern Marginal Zone (SMZ) of the Damara Belt, exposed in the Gaub Canyon in central Namibia, consists of fourteen lithotectonic units of high strain amphiholite facies rock with pelagic, hemi-pelagic and clastic sedimentary protoliths. These rocks are intercalated With lenses of metabasite. Regional high-pressure - low-temperature metamorphic conditions (~1O kbar and ~600°C) dominate the Southern and Southern Marginal Zones of the Damara Belt, leading to the interpretation that these tectonostratigraphic terranes formed in an accretionary prism along an ancient subduction margin. The structures in the SMZ are the result of progressive deformation, inferred to have initiated under low-grade metamorphic conditions (D₁) and evolved through prograde to peak metamorphism (D₂), ending in relatively low-temperature retrograde conditions (D₃). Each of the deformation phases is characterised by a foliation. D₁ is associated With pure shear dominated layer-parallel extension characterised by disrupted lithological layering and hedding-parallel foliation S₀+₁. D₃ is defined as deformation related to the formation of an axial-planar S₂ caused by folding of S₀+₁ around F₂ hinge lines. Widespread isoclinal recumbent folding resulted in transposition of these fahrics and the general foliation is thus termed S₀+₁+₂. This composite foliation contains a down-dip stretching lineation L₂. Folding was contemporaneous With top-to-the-SE directed thrusting in D₂ faults and shear zones that are seen to displace D₁ fabric. Fold hinge lines parallel to L₂ suggest D₂ is characterised by non-ideal simple shear. D₃ is defined by a crenulation cleavage S₃, at near right angles to S₀+₁+₂ foliation resulting from NW-SE pure shear shortening. This phase of deformation is also associated with retrograde, reverse faulting that is localised along some of the D₂ shear zones.

Includes abstract. / Includes bibliographical references (leaves 93-97). / This study is aimed at developing a workflow, and ultimately a model, for quantitative interpretation of sonic and seismic data. Measured data collected at the point of logging can be fraught with errors that can lead to wrong interpretation. One of such data is the shear wave velocity which in most cases is collected with the compressional wave velocity. The measured shear wave velocity log may contain errors that are due to drilling conditions, mud invasion etc. It may also contain cycle skips and might contain a lot of missing data and information. It is because of the poor quality of this type of log that has often made well log analysis companies and log interpreters neglect the measured shear wave log and subsequently generate or create an estimated shear wave log which they use for interpretation and modelling to check how the amplitudes vary with increasing offsets, among other uses.The workflow presented in this study considers the effect of working with the measured data, a reprocessed shear wave log and a locally estimated shear wave log. Specific correction procedures for invasion of the logs was done and synthetic seismograms were created for each type after correction for comparison to a 3D seismic data. The results of this study suggest that oil based mud invasion can cause significant problems to sonic logs especially the shear wave log. It also suggests that, if a shear wave log is of low or bad quality, a reprocessed shear wave log would be better for interpretation and modeling rather than a locally calibrated shear wave log or an estimated shear wave log using global predictions. The conclusion is evident from the synthetics generated using the measured shear wave data and the estimated shear wave data.

Summary in English. / Bibliography: leaves 157-163. / Former evaporite horizons in metamorphic terrains have frequently been referred to as a source for highly saline fluids associated with base metal ore deposits in various parts of the world. Data on the nature of these former evaporites and their associated connate fluids, however, is frequently quite sparse, and the link between the ore deposits and the source evaporites is usually intuitive, at best. In an attempt to characterise such evaporite-derived fluids and their source evaporites, a study of former evaporites in the Duruchaus Formation in the Southern Marginal Zone (SMZ) of the Damara Belt was carried out. The lithological units that make up the evaporite sequence within the Duruchaus Formation were mapped and sampled, along with discordant quartz-dolomite plugs intruding at various levels in the Duruchaus Formation stratigraphy, and in the overlying Bakos Group.

Summary in English. / Includes bibliography. / Saline pans are an important component of the hydrologic cycle in arid areas, and are common in South Africa. Natural saline pans on the coastal lowlands of the Western Cape Province, South Africa, were studied to determine the origin and evolution of salts within the pans. Samples of surface, stream and spring water as well as pan sediments were collected from two large coastal pans near Ysterfontein, and eight smaller pans 10-20 km from the sea, on the inland side of the granitic Darling hills.

Bibliography: pages 63-68. / Associated with ESKOM's ash water beneficiation programme is the precipitation of an ettringitic waste from highly alkaline, saline water. The waste is dominated by ettringite (CauAJ4(OH)24(S04) 6.52H20) with calcite (CaC03) as a minor phase (17.7% for the sample used in this study). Apart from the presence of calcite, the ettringite itself is alkaline due to the presence of OH ions. Following a submission that the waste had potential as an ameliorant of acid soil, research into this possibility was initiated. Following the determination of a calcium carbonate equivalent (HCl-CCE) value of 78% using the HCl back titration method of Horwitz (1980), an incubation experiment was initiated using three acid soils of contrasting characteristics: a so-called Silvermine sand, Kranskop A and Kranskop B soils. The effects on soil acidity of ettringitic waste were compared with analytical grade calcite. Soils (50g samples) were incubated with the two alkaline amendments for two weeks, following which pH(KCl), pH(H20) and KCl-extractable acidity were determined. Ettringitic waste led to apparently lower levels of acidity neutralization for corresponding treatments set on an HCl-CCE basis. This difference was minimized with the highly buffered, sesquioxide and organic-rich Kranskop A soil which could be attributed to the greater reactivity of the ettringitic waste with organically-complexed acidity together with the "self-liming" effect of so4 in sesquioxide-rich soils (sensu Reeve & Sumner, 1972). The waste showed progressively less neutralization with Kranskop B and Silvermine soils apparently in response to a decline in buffering capacity of these soils.

Includes bibliographical references. / While a clear and unequivocal model for the formation of diamonds is still lacking, the past few decades have seen a 'revolution' in the scientist's perception of mantle processes, particularly that regarding diamond formation. Kramers (1977) fIrst clearly recognised the ancient origin of diamonds from his studies of composited sulphide inclusions. However, it was not until mid-Archean ages were obtained from syngenetic garnet inclusions in diamond by Richardson et al. (1984), that it became generally accepted that most diamonds were in fact very much older than their kimberlitic hosts, and could therefore not have been precipitated from the kimberlite magma.