L’origine des éléments volatils, éléments avec des températures de condensation très basses comme l’eau, l’azote, le carbone et les gaz rares, sur Terre et sur les autres planètes telluriques reste mal comprise. Comprendre comment ces éléments sont arrivés sur Terre permettra de mieux appréhender les mécanismes de formation du système solaire. De par leur caractère inerte, les gaz rares (He, Ne, Ar, Kr, Xe) constituent des traceurs uniques des sources d’éléments volatils. L’étude de la composition en gaz rares du manteau terrestre s’avère donc essentielle afin de déterminer l’origine de ces éléments. L’objectif de ces travaux était de mesurer précisément la composition des isotopes stables et non-radiogéniques des gaz rares dans le manteau à partir de verres basaltiques. Ces derniers étant très souvent contaminés par l’air, des techniques d’analyse des gaz rares ont été mises en place pour s’affranchir de cette contamination. Des échantillons du volcan Fernandina du point chaud des Galápagos ont été étudiés par ablation laser. Les résultats montrent que le rapport 20Ne/22Ne des bulles est en moyenne de 12,65 ± 0,04 (1σ), ce qui est proche de la valeur estimée pour le pôle d’implantation du vent solaire (~ 12,7). Cette étude suggère donc que l’implantation du vent solaire sur les grains précurseurs de la Terre serait à l’origine des éléments volatils légers du manteau.Des données d’échantillons de popping rocks, aussi analysés par cette technique, indiquent que le manteau supérieur est hétérogène à petite échelle et pourraient impliquer le recyclage de gaz rares dans le manteau, y compris He, Ne, lors de la subduction. Un nouveau protocole d’analyse du Kr et Xe a été développé. Les résultats montrent les plus forts excès en 124Xe, 126Xe et 128Xe mesurés pour le manteau et suggèrent une source chondritique. Les isotopes fissiogéniques du xénon ont permis de dater le début du recyclage des éléments volatils dans le manteau aux alentours de 3 Ga. / The origin of volatiles, elements with low condensation temperatures such as water, nitrogen, carbon and noble gases, on Earth and other terrestrial planets is still misunderstood. Determining how these elements were delivered to the Earth will allow a better understanding of the processes of solar system formation. Due to their inertness, noble gases (He, Ne, Ar, Kr, Xe) constitute unique tracers of volatiles sources. Studying the noble gas composition of the Earth’s mantle is hence critical to unravel volatile origin. The aim of this PhD was to measure precisely the composition of non-radiogenic, stable noble gas isotopes in the mantle from basaltic glasses. The latter being very often contaminated by air, analytical techniques were thus set up to overcome this contamination.Samples from Fernandina volcano from the Galápagos hotspot were studied with laser ablation. The results show that the mean 20Ne/22Ne ratio in the vesicles is 12.65 ± 0.04 (1σ), close to the estimated value of the solar wind implanted end-member (~ 12,7). This study suggests that light volatiles in the mantle would originate from implantation of solar wind on the Earth’s precursors grains. Data from popping rock samples, also analyzed with this technique, indicate that the upper mantle is heterogeneous at small scale and that noble gases, included He, Ne, could be recycled into the mantle in subduction zones. A new protocol to analyse Kr and Xe was set up. The results show the highest measured excess in 124Xe, 126Xe and 128Xe for the mantle and suggest a chondritic source. The xenon fissiogenic isotopes allowed us to constrain the onset of volatile recycling in the mantle to around 3 Ga.

The effectiveness of the Environmental Impact Assessment process has been questioned by its critics both locally and internationally, as there is a perception that EIA process is merely a rubber stamping exercise. The objective of this study was to determine whether or not the relevant provincial authorities in South Africa have issued EIA refusals and if so what the main reasons for refusal were. Both Basic Assessment and full EIA processes were considered. Access to the EIA refusals from the various provincial environmental departments and environmental consultants was limited. Only seventeen EIA refusals were received after extended requests over a 12-month period, after which each of these were analysed. The reasons for the EIA refusals encountered in this study have been categorised into seventeen sub-classes relating to the following environmental issues: site location, socio-economics, land use/zoning, lack of justification, Spatial Development Framework (SDF), biodiversity, incompleteness of information, legislation discouraging development, visual/noise impacts, lack of alternatives, services issues, cumulative effects, groundwater, waste, specialist studies, gross non-compliance and air pollution. It is important to note that an EIA application could potentially have more than one screening trigger, and therefore it is possible that the percentages explained in this study can add up to more than 100%. The highest number of the EIA refusals’ screening triggers (8 of 17 = 47.06%) were found to be due to the transformation and rezoning of undeveloped or vacant land, and 5 of 7 (71.4%) of those particular EIA refusals were attributed to applications for residential development. Biodiversity and ecological sensitivity of the site location, as well as construction of infrastructure were next on the scale, with three (17.65%) EIA refusal screening triggers each. Finally, concentration of animals for production and storing and handling of hazardous substances both had two (11.76%) screening triggers. Only one EIA refusal did not include any substantive reasons for refusal and was refused on purely procedural grounds. The lack of justification of the development, lack of technical information and inadequate alignment with future spatial planning also constituted reasons for negative authorisations. From the results it was evident that although it is usually the procedural issues that hinder EIA, this study encountered many substantive issues, making up the majority of the reasons for EIA refusal here. This goes against international opinion that EIAs are usually turned down due to lack of adherence to process. Other findings from this study of particular interest include that no database is maintained for the number and reasons of EIA refusals that are processed, only for those that are authorised. It was also found that there were provinces that have never issued an EIA refusal. Furthermore, it was interesting to note that the reasons given in the findings for the analysed EIA refusals did not necessarily correlate with the screening triggers. / Thesis (Master of Environmental Sciences)--North-West University, Potchefstroom Campus, 2012.

The effectiveness of the Environmental Impact Assessment process has been questioned by its critics both locally and internationally, as there is a perception that EIA process is merely a rubber stamping exercise. The objective of this study was to determine whether or not the relevant provincial authorities in South Africa have issued EIA refusals and if so what the main reasons for refusal were. Both Basic Assessment and full EIA processes were considered. Access to the EIA refusals from the various provincial environmental departments and environmental consultants was limited. Only seventeen EIA refusals were received after extended requests over a 12-month period, after which each of these were analysed. The reasons for the EIA refusals encountered in this study have been categorised into seventeen sub-classes relating to the following environmental issues: site location, socio-economics, land use/zoning, lack of justification, Spatial Development Framework (SDF), biodiversity, incompleteness of information, legislation discouraging development, visual/noise impacts, lack of alternatives, services issues, cumulative effects, groundwater, waste, specialist studies, gross non-compliance and air pollution. It is important to note that an EIA application could potentially have more than one screening trigger, and therefore it is possible that the percentages explained in this study can add up to more than 100%. The highest number of the EIA refusals’ screening triggers (8 of 17 = 47.06%) were found to be due to the transformation and rezoning of undeveloped or vacant land, and 5 of 7 (71.4%) of those particular EIA refusals were attributed to applications for residential development. Biodiversity and ecological sensitivity of the site location, as well as construction of infrastructure were next on the scale, with three (17.65%) EIA refusal screening triggers each. Finally, concentration of animals for production and storing and handling of hazardous substances both had two (11.76%) screening triggers. Only one EIA refusal did not include any substantive reasons for refusal and was refused on purely procedural grounds. The lack of justification of the development, lack of technical information and inadequate alignment with future spatial planning also constituted reasons for negative authorisations. From the results it was evident that although it is usually the procedural issues that hinder EIA, this study encountered many substantive issues, making up the majority of the reasons for EIA refusal here. This goes against international opinion that EIAs are usually turned down due to lack of adherence to process. Other findings from this study of particular interest include that no database is maintained for the number and reasons of EIA refusals that are processed, only for those that are authorised. It was also found that there were provinces that have never issued an EIA refusal. Furthermore, it was interesting to note that the reasons given in the findings for the analysed EIA refusals did not necessarily correlate with the screening triggers. / Thesis (Master of Environmental Sciences)--North-West University, Potchefstroom Campus, 2012.

Biodiversity considerations form an essential component of Environmental Impact Assessment (EIA), especially in areas with both a high biodiversity value and development pressure such as the Cape Floristic Region (CFR) in South Africa. Limited research has been conducted within the South African context on the quality of biodiversity inputs to EIA. The aim of this research was to evaluate the quality of biodiversity input to EIA in the CFR. To address this aim, a customised review package was generated to evaluate the quality of 26 Biodiversity Impact Assessment (BIA) reports in the CFR. The results were then compared with international trends of biodiversity input to EIA in order to show how prevalent such trends are within an area with high biodiversity value. This comparison showed that the quality of biodiversity input to EIA in the CFR generally concur with inadequacies identified in international EIA literature. Typically, significant weaknesses identified during the review were the lack of public participation and an insufficient evaluation of alternatives. Specialists also failed to develop adequate monitoring programmes. Furthermore, a very pertinent limitation was that, in general, assessments are conducted during inappropriate seasons and over insufficient time periods. However, some variations to the international trends are also present within the Region. For example, a particular strength was that a precautionary approach was adopted by most of the specialists to avoid negative impacts on biodiversity. In addition, specialists did not merely focus on lower levels (species and habitats) of biodiversity, but incorporated ecological processes in assessment techniques. The inadequacies identified in this dissertation pose particular challenges for biodiversity management and conservation practices. The development and implementation of mechanisms such as Best Practice Guidelines and improved biodiversity related legislation is proposed to improve biodiversity input to EIA. / M Environmental Management, North-West University, Potchefstroom Campus, 2014

Biodiversity considerations form an essential component of Environmental Impact Assessment (EIA), especially in areas with both a high biodiversity value and development pressure such as the Cape Floristic Region (CFR) in South Africa. Limited research has been conducted within the South African context on the quality of biodiversity inputs to EIA. The aim of this research was to evaluate the quality of biodiversity input to EIA in the CFR. To address this aim, a customised review package was generated to evaluate the quality of 26 Biodiversity Impact Assessment (BIA) reports in the CFR. The results were then compared with international trends of biodiversity input to EIA in order to show how prevalent such trends are within an area with high biodiversity value. This comparison showed that the quality of biodiversity input to EIA in the CFR generally concur with inadequacies identified in international EIA literature. Typically, significant weaknesses identified during the review were the lack of public participation and an insufficient evaluation of alternatives. Specialists also failed to develop adequate monitoring programmes. Furthermore, a very pertinent limitation was that, in general, assessments are conducted during inappropriate seasons and over insufficient time periods. However, some variations to the international trends are also present within the Region. For example, a particular strength was that a precautionary approach was adopted by most of the specialists to avoid negative impacts on biodiversity. In addition, specialists did not merely focus on lower levels (species and habitats) of biodiversity, but incorporated ecological processes in assessment techniques. The inadequacies identified in this dissertation pose particular challenges for biodiversity management and conservation practices. The development and implementation of mechanisms such as Best Practice Guidelines and improved biodiversity related legislation is proposed to improve biodiversity input to EIA. / M Environmental Management, North-West University, Potchefstroom Campus, 2014

This study aims to reveal the geochemical signatures of the basic igneous rocks with well-determined age within the Karakaya Complex in Central and NW Anatolia and also exhibit the relationships between the studied units in terms of geological and petrographical features. The Karakaya Complex comprise a number of tectono-stratigraphic units in the studied regions (the Olukman Melange, the Bah&ccedil / ecik Formation, the Ortaoba Unit and the informally named pillow basalt-limestone association) and the pre-Karakaya basement unit (the Eymir Complex). The basic igneous rocks have been all intensely affected by hydrothermal metamorphism as reflected by the secondary products strongly overprinting the primary mineral phases and most of them exhibit vesicular structures which are filled by mainly calcite. The primary mineral assemblage dominating the basaltic rocks is clino-pyroxene, plagioclase and olivine, whereas secondary phases are characterized by actinolite, pistacite, zoisite/clinozoisite group and chlorite. Kaersutite, as a late stage magmatic mineral, is distinctive for Ti-augite bearing imrahor basalts / on the other hand, the diabase dykes include hornblende as an essential primary phase. The basic rocks are represented by three groups / sub-alkaline, alkaline and transitional. The alkaline samples from imrahor, Hasanoglan, Kadirler and Ortaoba are of Anisian age and akin to oceanic-island basalts (OIB). The sub-alkaline and transitional samples from imrahor and Ortaoba reflect P-MORB features and are younger than the first group. The diabase dykes cross-cutting the Eymir Complex, on the other hand, are too dissimilar, indicating back-arc basin signatures. Based on the data obtained from this study, the Karakaya Complex is characterized by a number of tectonic components (seamount, plume-related mid ocean ridge and back-arc basin) with different ages and origins, which were later amalgamated during the Cimmerian orogeny.

Novel techniques for in-situ lead isotope analysis of melt inclusions using multicollector inductively coupled plasma mass spectrometry, and parallel Faraday cup and ion counter detection have been developed. These provide for measurement of the critical 204Pb isotope, which has been unavailable in melt inclusion Pb isotope studies, but requires on-line determination of 200Hg to correct for 204Hg isobaric interferences. External standardisation allows for effective mass fractionation correction and ion counter gain calibration. An off-line Tau correction was applied and standard glass analyses suggest that this correction provides more accurate and precise results.

The Mount Somers Volcanic Group (MSVG) and Mandamus Igneous Complex (MIC) are the magmatic manifestations of the transition from convergence to extension at the Gondwana margin, which culminated in the separation of New Zealand from Australia and Antarctica. The MIC has been correlated both geochemically and temporally with the Central Marlborough Igneous Province (CMIP). The MSVG and CMIP are located in the Eastern Province of New Zealand. The MSVG is restricted to the Rakaia terrane, whereas the CMIP is restricted to the Pahau terrane. The Rakaia and Pahau terranes are thick accretionary complexes, which were strongly deformed as a result of prolonged subduction at the Gondwana margin. The Pahau terrane is the younger of the two and continued to be deposited and deformed until the abrupt cessation of subduction, which in the Marlborough sedimentary record occurred in the Motuan (100 - 105 Ma). Following the cessation of subduction, after an interval of 2-7 Ma of relative quiescence and subsidence of the Pahau terrane, the MSVG and MIC were erupted/emplaced. The production of MSVG and MIC magmas occurred simultaneously and the activity was of short-lived duration. SHRIMP geochronology yielded crystallisation ages of 97.0 ± 1.5 Ma to 98.0 ± 1.2 Ma from zircons separated from MSVG rhyolites. The SHRIMP ages are within error of the previously published Rb-Sr age for the MIC. The SHRIMP geochronology also confirmed the presence of inherited zircons which yielded ages consistent with their derivation from the Rakaia terrane. Ar-Ar geochronology confirmed the coeval nature of the MSVG and MIC magmatism, but yielded consistently younger ages (94.5 ± 3 Ma for the MSVG and 94.2 ± 1.7 Ma for the MIC). The systematic differences in ages obtained by SHRIMP and Ar-Ar are believed to be method-dependent. The MSVG comprises a calc-alkaline volcanic assemblage, which ranges in composition from basaltic-andesite lavas (SiO₂ = 54.5%) to high-silica rhyolites and ignimbrites (SiO₂ ≤ 78.1%). The MSVG had an original extent of at least 18 000 km². The magmas from the MSVG had high LILE/HFSE, high LILE/REE and moderately high LREE/HFSE which are characteristic of subduction derived magmas. Geochemical modelling suggests that the MSVG magmas were formed from partial melting of a subduction-modified mantle wedge, with high degrees of crustal assimilation. The assimilant had an isotopic composition similar to that of the Rakaia terrane, which is consistent with the geological setting of the MSVG. The MSVG has ⁸⁷Sr/⁸⁶Sri from 0.7055 to 0.7100 and ¹⁴³Nd/¹⁴⁴Ndi from 0.51254 to 0.51230 (ɛNd +0.5 to -4.2), which reflects varying degrees of contamination by Rakaia terrane. Radiogenic isotope modelling suggests that the MSVG end-members were derived from the same parent magma, which evolved through AFC processes from basaltic-andesite to rhyolite. The modelling strongly suggests that assimilation played a lesser role in the petrogenesis of the Malvern Hills magmas than in the petrogenesis of the other units. AFC modelling requires the degree of assimilation to increase as the magmas evolved. Oxygen isotope data are consistent with high degrees of crustal assimilation, and may indicate that the assimilant had higher ¹⁸O characteristics than the Rakaia terrane samples analysed. The MIC is an alkaline suite which ranges in composition from basalt and gabbro to syenite, trachyte and phono-tephrite. The MIC is interpreted to have formed from enriched asthenospheric mantle, with a composition similar to HIMU (²⁰⁶Pb/²⁰⁴Pbi ranges from 19.2 to 20.3). The samples range in isotopic composition from ⁸⁷Sr/⁸⁶Sri = 0.7030 to 0.7036, ¹⁴³Nd/¹⁴⁴Ndi = 0.51275 to 0.51268 (ɛNd +4.6 to +3.3). The range in isotopic composition is due to varying degrees of contamination by Pahau terrane, which reaches a maximum of 25% but in most samples is < 10%. The MIC is contaminated to a much lesser extent than the MSVG which is interpreted to be related to the thinner nature of the Pahau crust in the mid-Cretaceous. The latest phases of activity in the MIC were subjected to lower degrees of contamination which is interpreted to reflect the passage of magmas through pre-existing pathways. The onset of MSVG and CMIP magmatism coincided with the initiation of major rift-related depositional basins, and the eruption of the MSVG is demonstrably associated with normal faulting. The tectonic trigger responsible for the sudden onset of magmatism and rifting in the Eastern Province terranes was the detachment of the previously subducting slab following the cessation of subduction due to the arrival of the Hikurangi Plateau at the margin and the subsequent stalling of the Pacific spreading centre. The capture of the Gondwana margin led to the propogation of extension into the margin by the divergent Pacific plate. The ensuing extension aided the detachment of the subducting slab beneath the Eastern Province terranes. The slab-detachment promoted decompression melting of the sub-lithospheric mantle wedge to produce the MSVG magmas and triggered the ascent of asthenospheric mantle through the slab window, which melted through decompression to produce the CMIP magmatism. The asthenospheric mantle tapped by the slab detachment episode was highly enriched relative to N-MORB and is akin to the similar age HIMU-OIB affinity melts documented from Antarctica and Australia. The short-lived duration of activity is typical of slab-detachment related magmatism which occurs as a passive response to plate reconfiguration. The similarity in geochemistry of the MIC with OIB-affinity igneous centres in Australia and Antarctica implies an enriched mantle domain of large geographical extent. The distribution of relatively small volumes of OIB magmatism is suggestive of a fossil plume component, which was tapped in response to lithospheric extension producing relatively short-lived HIMU magmatism. The same fossil plume component has previously been implicated in the formation of the Cenozoic West Antarctic Rift System and may be responsible for the late Cretaceous magmatism in the Chatham Islands and Tertiary volcanics of the South Island of New Zealand.

The Mount Somers Volcanic Group (MSVG) and Mandamus Igneous Complex (MIC) are the magmatic manifestations of the transition from convergence to extension at the Gondwana margin, which culminated in the separation of New Zealand from Australia and Antarctica. The MIC has been correlated both geochemically and temporally with the Central Marlborough Igneous Province (CMIP). The MSVG and CMIP are located in the Eastern Province of New Zealand. The MSVG is restricted to the Rakaia terrane, whereas the CMIP is restricted to the Pahau terrane. The Rakaia and Pahau terranes are thick accretionary complexes, which were strongly deformed as a result of prolonged subduction at the Gondwana margin. The Pahau terrane is the younger of the two and continued to be deposited and deformed until the abrupt cessation of subduction, which in the Marlborough sedimentary record occurred in the Motuan (100 - 105 Ma). Following the cessation of subduction, after an interval of 2-7 Ma of relative quiescence and subsidence of the Pahau terrane, the MSVG and MIC were erupted/emplaced. The production of MSVG and MIC magmas occurred simultaneously and the activity was of short-lived duration. SHRIMP geochronology yielded crystallisation ages of 97.0 ± 1.5 Ma to 98.0 ± 1.2 Ma from zircons separated from MSVG rhyolites. The SHRIMP ages are within error of the previously published Rb-Sr age for the MIC. The SHRIMP geochronology also confirmed the presence of inherited zircons which yielded ages consistent with their derivation from the Rakaia terrane. Ar-Ar geochronology confirmed the coeval nature of the MSVG and MIC magmatism, but yielded consistently younger ages (94.5 ± 3 Ma for the MSVG and 94.2 ± 1.7 Ma for the MIC). The systematic differences in ages obtained by SHRIMP and Ar-Ar are believed to be method-dependent. The MSVG comprises a calc-alkaline volcanic assemblage, which ranges in composition from basaltic-andesite lavas (SiO₂ = 54.5%) to high-silica rhyolites and ignimbrites (SiO₂ ≤ 78.1%). The MSVG had an original extent of at least 18 000 km². The magmas from the MSVG had high LILE/HFSE, high LILE/REE and moderately high LREE/HFSE which are characteristic of subduction derived magmas. Geochemical modelling suggests that the MSVG magmas were formed from partial melting of a subduction-modified mantle wedge, with high degrees of crustal assimilation. The assimilant had an isotopic composition similar to that of the Rakaia terrane, which is consistent with the geological setting of the MSVG. The MSVG has ⁸⁷Sr/⁸⁶Sri from 0.7055 to 0.7100 and ¹⁴³Nd/¹⁴⁴Ndi from 0.51254 to 0.51230 (ɛNd +0.5 to -4.2), which reflects varying degrees of contamination by Rakaia terrane. Radiogenic isotope modelling suggests that the MSVG end-members were derived from the same parent magma, which evolved through AFC processes from basaltic-andesite to rhyolite. The modelling strongly suggests that assimilation played a lesser role in the petrogenesis of the Malvern Hills magmas than in the petrogenesis of the other units. AFC modelling requires the degree of assimilation to increase as the magmas evolved. Oxygen isotope data are consistent with high degrees of crustal assimilation, and may indicate that the assimilant had higher ¹⁸O characteristics than the Rakaia terrane samples analysed. The MIC is an alkaline suite which ranges in composition from basalt and gabbro to syenite, trachyte and phono-tephrite. The MIC is interpreted to have formed from enriched asthenospheric mantle, with a composition similar to HIMU (²⁰⁶Pb/²⁰⁴Pbi ranges from 19.2 to 20.3). The samples range in isotopic composition from ⁸⁷Sr/⁸⁶Sri = 0.7030 to 0.7036, ¹⁴³Nd/¹⁴⁴Ndi = 0.51275 to 0.51268 (ɛNd +4.6 to +3.3). The range in isotopic composition is due to varying degrees of contamination by Pahau terrane, which reaches a maximum of 25% but in most samples is < 10%. The MIC is contaminated to a much lesser extent than the MSVG which is interpreted to be related to the thinner nature of the Pahau crust in the mid-Cretaceous. The latest phases of activity in the MIC were subjected to lower degrees of contamination which is interpreted to reflect the passage of magmas through pre-existing pathways. The onset of MSVG and CMIP magmatism coincided with the initiation of major rift-related depositional basins, and the eruption of the MSVG is demonstrably associated with normal faulting. The tectonic trigger responsible for the sudden onset of magmatism and rifting in the Eastern Province terranes was the detachment of the previously subducting slab following the cessation of subduction due to the arrival of the Hikurangi Plateau at the margin and the subsequent stalling of the Pacific spreading centre. The capture of the Gondwana margin led to the propogation of extension into the margin by the divergent Pacific plate. The ensuing extension aided the detachment of the subducting slab beneath the Eastern Province terranes. The slab-detachment promoted decompression melting of the sub-lithospheric mantle wedge to produce the MSVG magmas and triggered the ascent of asthenospheric mantle through the slab window, which melted through decompression to produce the CMIP magmatism. The asthenospheric mantle tapped by the slab detachment episode was highly enriched relative to N-MORB and is akin to the similar age HIMU-OIB affinity melts documented from Antarctica and Australia. The short-lived duration of activity is typical of slab-detachment related magmatism which occurs as a passive response to plate reconfiguration. The similarity in geochemistry of the MIC with OIB-affinity igneous centres in Australia and Antarctica implies an enriched mantle domain of large geographical extent. The distribution of relatively small volumes of OIB magmatism is suggestive of a fossil plume component, which was tapped in response to lithospheric extension producing relatively short-lived HIMU magmatism. The same fossil plume component has previously been implicated in the formation of the Cenozoic West Antarctic Rift System and may be responsible for the late Cretaceous magmatism in the Chatham Islands and Tertiary volcanics of the South Island of New Zealand.

Les compositions isotopiques de Sr, Nd et Pb des basaltes d'îles océaniques ont été mesurées de manière systématique depuis 30 ans. Les variations de ces compositions sont classiquement interprétées comme la signature du mélange entre des composants présents au sein des panaches mantelliques. Cependant elles peuvent aussi représenter des interactions entre les magmas et les réservoirs qu'ils traversent au cours de leurs remontées vers la surface. Après avoir mis en évidence les processus pouvant altérer la composition initiale d'un magma et proposer des moyens de s'en affranchir, 2 études locales sont éffectuées (îles de Sao Nicolau et de Sao Jorge, archipels du Cap Vert et des Açores). Ces deux études montrent la part importante des variations d'origine superficielle dans ces deux localités. Ces résultats sont ensuite généralisés à d'autres localités clés de la topologie mantellique et permettent de discuter d'un composant primitif ubiquiste dans les panaches mantelliques