Includes bibliography. / Dental fluorosis is endemic in the western Bushveld of South Africa. This study investigated the occurrence of fluoride (Fˉ) in groundwater in the area. It was hypothesised that fluoride in groundwater originated from the dissolution of fluorine-bearing minerals, principally fluorite, mica, amphibole and apatite, and that high F- groundwater would be hosted in rocks with a high fluorine (F) content. It was further hypothesised . that groundwater residence time, rare F-bearing minerals, ion exchange reactions and evaporation affected the Fˉ concentration of groundwater. These hypotheses were investigated by analysing data on Fˉ concentrations in groundwater, paired rock and soil samples, and selected minerals. Three hundred and thirty eight samples of groundwater from the field area were analysed for Fˉ by both Fluoride Ion Selective Electrode and High Pressure Ion Chromatography. The results were added to a database of three thousand water samples covering the western Bushveld and together provide the first complete picture of the distribution of high Fˉ groundwater in the area.

Bibliography: leaves 246-273. / The west coast of South Africa has been exploited for diamonds for most of the last century. Although some literature has accumulated over this period, the evolution of the coastal plain since the Cainozoic remains poorly understood. The aim of this thesis is to offer a multidisciplinary study of the evolution of coastal deposits in the Saldanha Bay area, along the southern part of the west coast. The environments investigated comprise lagoons, coastal dunes and shorelines. The ages of the deposits were obtained with radiocarbon dating, strontium isotope stratigraphy and biostratigraphy.

The Buffels River area of central Namaqualand comprises a sequence of Proterozoic volcano-sedimentary gneisses intruded by granitoid orthogneisses. This study concentrates on the mechanisms of granrilite-facies metamorphism in a small area of isofacial rocks. Particular emphasis is placed on the role of partial melting and the infiltration and buffering of metamorphic fluid is examined.

Reassembling continental fragments of Gondwana has been a subject of interest to many since almost the beginning of the last century. As a result, the broad relative position of the major continental fragments and their dispersal history is well understood using marine magnetic anomalies, coastline geometry, surface geology and limited geophysics. Uncertainty still prevails in reassembly of central Gondwana fragments flanking the Indian Ocean. This thesis aims at utilising geophysical constraints to corroborate an fine-tune the reconstruction of these fragments supported by geological evidence.

Bibliography: pages 225-251. / The hypabyssal quarry kimberlite is the most abundant phase at Monastery. Four petrographically distinct varieties are recognized. Mineralogically, the Ql and Q4 kimberlites are opaque oxide-rich serpentine-phlogopite kimberlites, the Q2, a phlogopite-monticellite kimberlite and the Q3, a monticellite-phlogopite kimberlite. The East-end kimberlite is an opaque oxide-rich serpentine-monticellite kimberlite, but is poorly exposed and highly weathered. The breccia kimberlite hosts abundant country rock fragments in a soft serpentinous matrix. It is an opaque oxide-rich phlogopite serpentine kimberlite breccia. The precursor kimberlite dyke associated with the diatreme is an opaque oxide-rich calcite kimberlite. The ultramafic xenoliths at Monastery are predominantly coarse grained and exhibit a high incidence of modal metasomatism. Some textures intermediate between porphyroclastic and granuloblastic were noted. Significant annealing has occurred. Garnet, orthopyroxene and clinopyroxene may have been derived by exsolution from high temperature aluminous orthopyroxenes. Minerals in the peridotites and pyroxenites have similar compositions to those from other localities. Two groups of phlogopite composition have been noted. Wehrlitic rocks have phlogopite similar to that in richterite-bearing peridotites from Kimberley.

Globally, there are significant contrasts, both thermally and chemically, between peridotite xenoliths exhumed from Archean, and post-Archean terranes. Studies of the thermal structure of the lithosphere, in combination with surface heat flow data, suggest that thermal gradients beneath cratonic regions (Archean blocks which stabilized > 2.5 Ga) are lower than those in off-craton regions, commonly attributed to thinner lithosphere in Proterozoic domains. Although this is true to an extent for southern Africa, the contrasts appear less distinct than Archean-Proterozoic lithosphere contrasts elsewhere, and the thermal structure reflects a regional, Mesozoic, disturbances which has been temporally linked to large-scale tectonic processes. Major element P-T results from peridotite xenoliths sampled during Group II kimberlite magmatism (~ 150 Ma), that erupted through the southwestern Proterozoic NamaquaNatal Province record geothermal gradients akin to those of cratonic terranes. In contrast, xenoliths sampled during the younger Group I kimberlite magmatism (~ 80 Ma) within the Namaqua-Natal Province record equilibration temperatures which are displaced to higher values throughout the pressure ranges. This study reports results from peridotite xenolith samples from seven kimberlites and related rocks which erupted within the Eastern Namaqualand and Namaqualand-Bushmanland-Warmbad area. Two of the localities studied here (Melton Wold and Markt) erupted during Group II kimberlite magmatism, and five localities (Hebron, Uintjiesberg, Gansfontein, Hoedkop and Schuitdrift) erupted during the younger Group I kimberlite magmatism. The results build on prior work, which focused on mineral and whole rock major element chemistry, platinum group elements and Re-Os isotope data, and provide an insight into lithosphere formation and modification in the Namaqua-Natal lithospheric mantle through mineral trace element analysis. These samples also provide an opportunity to further investigate the Mesozoic thermal evolution of the NamaquaNatal lithosphere through application of a REE-based thermobarometer. REE diffusion rates are typically 2 – 3 orders of magnitude lower than divalent major elements, making the REE-based thermobarometer a potentially useful tool for probing the contrasting thermal profiles exhibited in samples from the Group II and Group I kimberlites studied here. Major element-based thermobarometry results and the resulting FITPLOT paleogeotherms indicate that at the time of Grp II kimberlite magmatism the Namaqua-Natal lithosphere was ~ 200 km thick, with a 60 km “diamond window”, and had a geothermal gradient of 40 mW/m2 . In contrast, at the time of Grp I kimberlite magmatism ~ 60 km of lithospheric erosion may have occurred, accompanied by a shift in the thermal regime of the Namaqua-Natal lithosphere to a 45 mW/m2 geotherm. REE-based thermobarometry results produce a large amount of scatter in P-T space, even after a rigorous attempt at identifying well equilibrated clinopyroxene and garnet pairs. The presence of carbonatitic and silico-carbonate metasomatic signatures in these samples necessitates caution in the use of the REE-based thermobarometer when applied to xenoliths entrained by kimberlites. It is likely that the scatter observed in the results presented here is due to differing REE partitioning controls in systems containing carbonate in the melt to those of carbonate-free, silicate melts. HREE concentration in reconstructed whole-rocks and olivine Mg-numbers are consistent with 30% melt extraction in a shallow melting regime. Garnet and clinopyroxene trace element signatures indicate a shift in the style of metasomatism in the Namaqua-Natal lithosphere between Group II and Group I kimberlite magmatism. Zr/Hf versus Ti/Eu systematics reflect kimberlitic/silicate style metasomatism in the samples exhumed during Group II magmatism, whereas samples exhumed during Group I kimberlite magmatism reflect a carbonatitic style of metasomatism.

Includes bibliography. / The onshore marine phosphorite and aluminium phosphate deposits of Varswater Quarry (near Langebaanweg), Bomgat (near Hoedjiespunt), Kreefte Bay and Konstabelkop were studied to determine the origin of the various deposits.

Includes bibliographical references. / The Darling and Yzerfontein hypersaline pans are located on the western coastal plain of South Africa. The area has an arid, Mediterranean type climate. The carbonate minerals calcite and dolomite have precipitated in the pans through the evaporation and concentration of ions derived from weathering of bedrock, marine aerosols and coastal rainfall. Five pans were studied, the brine-type pans Rooipan South, Zwartwater North and Zwartwater South, the brackish saline pan Rooipan North, and the coastal Yzerfontein pan. Directly below the surficial halite crust of the three brine-type pans is a black organic-matter rich layer in which bacterial sulphate reduction is occurring (hydrogen-sulphide smell). All of the pans contain disseminated calcite, with only the Zwartwater North and Zwartwater South pans containing a mixture of dolomite and calcite. Yzerfontein and Zwartwater North pans contain small aragonitic gastropod shells (Tomichia ventricosa) which are no longer active in the pans due to the high present-day salinities of the pan waters. Low-magnesian calcite is likely the first evaporite mineral to precipitate in the pans because of its low solubility.

The Man Craton in West Africa, like the Kaapvaal Craton in southern Africa, hosts diamondiferous kimberlites. However, West African kimberlites are commonly micaceous and unusual relative to archetypal South African kimberlites. Petrographically, they appear more similar to orangeites (aka Group II kimberlites), which represent a type of olivine-lamproite. A suite of 14 representative samples from the Cretaceous Tongo dike-01, Sierra Leone have been analysed for their bulk-rock major and trace elements as well as Sr-Nd-Pb isotope geochemistry. The primary objectives of this study are: 1) provide detailed petrographic observations of the dike, 2) classify the dike relative to kimberlites worldwide, 3) constrain the geochemical effects of secondary processes on bulk-rock analyses, 4) provide an estimate of the close-to-primary parent magma composition, and 5) constrain the petrogenesis of these diamondiferous rocks. The major element chemistry of the Tongo dike-01 reflects concentrations that are similar to both archetypal kimberlites and orangeites. Major elements such as SiO2 (~28.20 ± 3.90 wt. %) and CaO (~ 12.50 ± 1.80 wt. %) display archetypal kimberlite concentrations whereas Cr2O3 (~0.20 ± 0.01 wt. %) and P2O5 (~1.65 ± 0.60 wt. %) resemble those that define orangeites. The high abundance of phlogopite in this dike is illustrated by the high bulk-rock concentrations of K2O (~3.03 ± 0.50 wt. %) and Al2O3 (~4.08 ± 1.00 wt. %), similar to those of orangeites. Like the major element geochemistry, the trace element geochemistry of the Tongo dike-01 also displays mixed archetypal kimberlite and orangeite traits. Trace elements such as Nb (~365.0 ± 50.4 ppm) and Y (~18.77 ± 6.60 ppm) possess concentrations that are similar to kimberlites whereas Rb (~160.0 ± 14.8 ppm) and Th (~36.22 ± 5.30 ppm) resemble orangeites. Trace element ratios are no different, ratios such as Ce/Pb (16-82), Ba/Nb (1-8), La/Nb (0.6-1.2) and La/Sm (11-13) resemble those of kimberlites while La/Yb (280-520) are more similar to orangeites. However, unlike major and trace element geochemistry, the Sr-Nd-Pb isotope geochemistry of the Tongo dike-01 solely resembles those of archetypal kimberlites (87Sr/86Sr)i ~0,7039, (206Pb/204Pb)i ~18.88, (208Pb/204Pb)i ~40.02 and (143Nd/144Nd)i ~0.51253 ± 0.00001. Prior to interpretation of primary processes and parent magma composition estimates, the effects of secondary processes were first evaluated. These secondary processes include crustal contamination, ilmenite contamination and olivine entrainment/fractionation. Samples that had experienced these secondary processes were excluded and a suite of unaltered/least contaminated samples was compiled in order to constrain the close-to-primary magma composition of the Tongo dike-01 and interpret primary petrogenetic processes effecting the kimberlites. To determine a representative parent magma composition, a total of six out of the fourteen samples were excluded from consideration. The estimated close-to-primary magma composition for the Tongo dike-01 is suggested to be SiO2 ~28.20 ± 3.90 wt. %, Fe2O3 ~10.20 ± 1.70 wt. %, TiO2 ~1.70 ± 0.30 wt. %, Al2O3 ~4.08 ± 1.00 wt. %, K2O ~3.03 ± 0.50 wt. %, La ~363.12 ± 25.43 ppm, Gd ~15.87 ± 4.20 ppm, Yb ~1.020 ± 0.06 ppm, ( 87Sr/86Sr)i ~0.7039, ( 206Pb/204Pb)i ~18.88, ( 208Pb/204Pb)i ~40.02 and ( 143Nd/144Nd)i ~ 0.51253 ± 0.00001. Although the petrography and major element concentrations are similar to orangeites found in South Africa, the trace element and Sr-Nd-Pb isotope geochemistry of the Tongo dike-01 reflects a kimberlite composition. Thus, the Tongo dike-01 is more consistent with being classified as a relatively rare type of ‘mica-rich' kimberlite rather than orangeite. Kimberlites from around the world derive from the same asthenospheric mantle reservoir and their major element chemistry is controlled by the compositions/mineralogy of the lithospheric mantle assimilated during kimberlite evolution. The similarity of the trace element ratios and Sr-Nd-Pb isotopes of the Tongo dike-01 in this study relative to archetypal kimberlites worldwide strongly implies that the Tongo dike-01 derives from the same asthenospheric reservoir as these kimberlites, although mineralogically the Tongo dike-01 is different and has a different parent melt major element composition. This is interpreted to reflect the contribution of lithospheric mantle material that is mineralogically different to that assimilated by archetypal kimberlites during the ascent of the Tongo dike-01 parent magma through the sub-continental lithospheric mantle (SCLM). In the case of the Tongo dike-01, its primary melt is K2O-rich and must have assimilated more K2O-rich material in the SCLM. Such material is typically present as metasomatic products, e.g., Phlogopite-Ilmenite-Clinopyroxene (PIC) xenoliths observed in South Africa kimberlites. These xenoliths tend to possess abundant phlogopite. Thus, the main difference between the Tongo kimberlite and archetypal SA kimberlites is the fact that Tongo kimberlite assimilated more K2O-rich metasomatised material in the SCLM during its evolution.

Sequence stratigraphy is a branch of stratigraphy that is concerned with how genetically related geological successions are deposited in time and space. This requires the integration of diverse types of datasets (drill core, outcrop, wireline, reflection seismic surveys, etc.) to build robust depositional models, the cornerstones of stratigraphic frameworks. Although the application of sequence stratigraphy has been a successful tool to predict the lithology of geobodies in the petroleum industry, terminology is inconsistently used by the different schools of thought to define stratigraphic surfaces. This has resulted in multiple sequence stratigraphic models that interpret the same data differently. The limited exploration, to-date, and poor dataset quality have impeded the understanding of the geological evolution of the offshore Algoa and Gamtoos Basins in the southern Cape region of South Africa. To reconstruct the main geological events in the area since the late Early Jurassic, we integrated vintage borehole and seismic data as well as key outcrop observations, generated contemporary gross depositional environment models for the basin fill, and tested the applicability of different sequence stratigraphic models. The studied stratigraphic interval formed since the inception of Gondwana break-up, in syn- and post-rift systems that were increasingly dominated by marine processes, especially in the distal hanging walls. Marine incursions are detected in the Upper and Middle Jurassic in the Algoa and Gamtoos Basins, respectively. However, the severely eroded Algoa Basin syn-rift succession, exacerbated by poor data quality, makes it challenging to understand the timing of the marine incursion in this compartmentalized half-graben. Sedimentation within these half-grabens primarily occurred above the hanging walls, whilst the footwalls (i.e., basement highs) formed the dominant sediment source areas. The geological characteristics of the studied synrift succession prevents the application of the depositional sequence stratigraphic or the tectonic system tracts models. Because subaerial unconformities (SUs) in the distal syn-rift sequence are not detectable, a diachronous, northward advancement of the shoreline until the late Valanginian can be postulated. The observations in the syn-rift sequence, which is bound by a basal SU, followed by third and fourth-order transgressive and regressive cycles and a second-order maximum flooding surface at the top, can be explained with a modified genetic sequence stratigraphic model. In the transitional to drift phase interval, from Hauterivian to Holocene, the successions are bound by third-order SUs and their correlative conformities. In the successions without evidence for subaerial exposure in the drift successions, flooding surfaces are used as sequence-bounding stratigraphic contacts, validating the applicability of the genetic and transgressive-regressive sequence stratigraphic models for this upper part of the studied stratigraphic interval. This study reaffirms the notion that while the sequence stratigraphic concept is model independent, sequence models are sensitive to depositional scale and data resolution. Moreover, it also reiterates that sequence boundaries should not be limited to subaerial unconformities, but rather to correlative surfaces that bound genetically related sedimentary successions.