Descriure l’ agregat d’àrees petites amb un elevat risc de mortalitat detectat a la zona sud-oest d’Espanya, en particular a les províncies de Huelva, Sevilla i Cadis (HSC) durant el període 1987-1995, analitzar la seva evolució i explorar les seves possibles causes. Mètodes: Estudi ecològic mixt que examina el risc de morir en 2.218 àrees petites d’Espanya durant els períodes 1987-1995 i 1996-2004 mitjançant un enfocament bayesià empíric. Resultats: Les desigualtats geogràfiques en la mortalitat descrites a Espanya durant el període 1987-1995, on destacava una pitjor situació a la zona sud-oest i en particular a les províncies de HSC, no ha millorat en el període 1996-2004. La consistència de resultats per a diverses causes específiques de mort, sexe i períodes temporals, ha evidenciat que l’origen d’aquestes desigualtats geogràfiques és molt probablement una conseqüència de la combinació “tòxica” de factors socials (socioeconòmics, laborals i ambientals). L’exposició conjunta a substàncies tòxiques amb efectes cancerígens, que conformen una part important dels determinants laborals i ambientals de la salut, està probablement jugant un paper rellevant en l’elevada mortalitat observada a HSC. Conclusions: L’important problema en la mortalitat descrit a la zona sudoest d’Espanya entre els anys 1987-1995 contínua present durant el període 1996-2004, tant en homes com en dones. És necessari portar a terme estudis epidemiològics complementaris, així com desenvolupar un pla integral de salut pública (vigilància, detecció i tractament) al sud-oest espanyol basat en el principi de precaució. / To describe the cluster of high-risk mortality small areas detected in the southwest of Spain, specifically in the provinces of Huelva, Sevilla and Cádiz during the period 1987- 1995, to analyze their evolution and to explore their possible causes. Methods: Mixed ecological study that examines the risk of dying in 2.218 small areas of Spain during the periods 1987-1995 and 1996-2004 using an empirical bayesian approach. Results: The geographical inequalities in mortality described in Spain during the period 1987-1995, which featured the worse situation in the south-west and in particular in the provinces of HSC, has not improved in the period 1996-2004.. The consistency of results for different specific causes of death, sex and temporal period, has shown that the origin of these inequalities is most likely a result of the “toxic” combination of social factors (socioeconomic, occupational and environmental). The joint exposure to toxic substances with carcinogenic effects, which constitutes a significant part of the occupational and environmental determinants, is probably playing an important role in the high mortality observed in HSC. Conclusions: The important mortality problem described in the south-west of Spain between 1987-1995 is still present during the period 1996-2004, in both men and women. It is necessary to conduct additional epidemiologic studies, as well as to develop a comprehensive public health plan (surveillance, detection, treatment) in the southwest Spanish based on the precautionary principle.

[Truncated abstract] The species-rich Southwest Australian Floristic Region (SAFR) is a global biodiversity hotspot. Characterised by a Mediterranean-type climate and nutrient deficient landscape, this region is endowed with 7380 native vascular plant species/sub species, of which 49% are endemic and 2500 are of conservation concern. Despite the global significance of this region, there is still only a poor understanding of the factors influencing high diversity and endemism, and especially the population genetic consequences of narrow endemism and naturally fragmented species distribution. Holly leaved banksia (Banksia ilicifolia R. Br.), although widespread through Southwest Western Australia (SWWA), has a naturally fragmented distribution, with generally small populations restricted to swales and wetland fringes with depth to groundwater less than 10 m. As such, it provides an excellent model to better understand the ecological genetic consequences of local endemism, population size and natural population fragmentation . . . Products of wide outcrossing (over 30 km) showed a heterosis effect over local outcrossing, indicating increased ecological amplitude of offspring following interpopulation mating. These results suggest that the breeding and mating biology of B. ilicifolia counters the negative genetic erosion effects of narrow ecological amplitude and small population size. Recent habitat fragmentation, and reductions in population size and increased isolation, is impacting on these processes, but further research is required to assess the ultimate consequences of these genetic effects for population viability.

This project examines small mammal faunas from cave sites in south-west England and south Wales. The aims are threefold: To examine the rapid environmental changes taking place in the Late Pleistocene and early Holocene: To understand the processes by which small mammal remains were deposited in the caves examined: To demonstrate the value of small mammal studies as an archaeological tool. All identifiable small mammal remains from twelve selected sites are listed. Ten of the sites are new material. As the species examined here are seldom exploited by humans, the small mammals provide a record of the past environment unaffected by human selection of particular species, as might be the case in larger mammal assemblages. An examination of possible agents of accumulation is provided for each site to identify any bias introduced by prey selection. Reconstructions of the environment local to each cave at the time of deposition are offered. The evidence provided by the small mammals is related to the archaeological findings from each cave, to demonstrate the effect of human habitation of cave sites on the depositional and post-depositional processes shown by the microfauna. The environmental evidence provided by the study reflects a wider landscape rather than merely the immediate surroundings of the cave, and so gives a basis for human exploitation patterns in the area accessible from the cave. Reconstructions of the ecological mosaics formed by the rapidly changing climate of the period and the topographic variation around the cave sites are provided, demonstrating the potential complexity of the environment in which the humans and other fauna of the period existed. It is hoped that this will encourage archaeologists to look beyond the general division of environmental boundaries in this period, and to examine the local variation in habitat availability and use.

Recent events such as the New Orleans floods and the Japanese tsunami of 2011 have highlighted the uncertainty in the quantification of the magnitude of natural hazards. The research undertaken here has focussed on the uncertainty in evaluating storm surge magnitudes based on a range of statistical techniques including the Generalised Extreme Value distribution, Joint Probability and Monte Carlo simulations. To support the evaluation of storm surge frequency magnitude relationships a unique hard copy observed sea level data set, recording hourly observations, was acquired and digitised for Devonport, Plymouth, creating a 40 year data set. In conjunction with Devonport data, Newlyn (1915-2012) tide gauge records were analysed, creating a data set of 2 million data points. The different statistical techniques analysed led to an uncertainty range of 0.4 m for a 1 in 250 year storm surge event, and 0.7 m for a 1 in 1000 storm surge event. This compares to a 0.5 m uncertainty range between the low and high prediction for sea level rise by 2100. The Geographical Information system modelling of the uncertainty indicated that for a 1 in 1000 year event the level uncertainty (0.7 m) led to an increase of 100% of buildings and 50% of total land affect. Within the study area of south-west England there are several critical structures including a nuclear licensed site. Incorporating the uncertainty in storm surge and wave height predictions indicated that the site would be potentially affected today with the combination of a 1 in 1000 year storm surge event coincident with a 1 in 1000 wave. In addition to the evaluation of frequency magnitude relations this study has identified several trends in the data set. Over the data period sea level rise is modelled as an exponential growth (0.0001mm/yr2), indicating the modelled sea level rise of 1.9 mm/yr and 2.2 mm/yr for Newlyn and Devonport, will potentially increase over the next century by a minimum of 0.2 m by 2100.The increase in storm frequency identified as part of this analysis has been equated to the rise in sea level, rather than an increase in the severity of storms, with decadal variations in the observed frequency, potentially linked to the North Atlantic Oscillation. The identification as part of this study of a significant uncertainty in the evaluation of storm surge frequency magnitude relationships has global significance in the evaluation of natural hazards. Guidance on the evaluation of external hazards currently does not adequately consider the effect of uncertainty; an uncertainty of 0.7 m identified within this study could potentially affect in the region of 500 million people worldwide living close to the coast.

L’ophiolite de Nouvelle Calédonie présente une des plus grandes sections de manteau océanique obductées au monde, offrant une perspective unique pour l’étude des processus du manteau supérieur. Les roches du manteau appartiennent à une séquence ophiolitique “atypique”, dominée par des harzburgites réfractaires avec quelques lherzolites à spinelle et à plagioclase. À l'exception de quelques cumulats mafiques-ultramafiques, la partie crustale de l'ophiolite est totalement absente. Cette ophiolite a été étudiée depuis plusieurs décennies, toutefois sa nature ultra-appauvrie a rendu très difficile une caractérisation géochimique détaillée. La littérature scientifique ne regroupe que quelques données sur les éléments en trace et les données isotopiques sont totalement inexistantes. Dans ce travail de thèse, une étude géochimique exhaustive (éléments majeurs, en trace et isotopes Sr-Nd-Pb) des péridotites et des roches mafiques associées à l’ophiolite a été réalisée. Les péridotites sont des tectonites avec des textures porphyroclastiques. Les lherzolites à spinelle ont 7-8 vol.% de clinopyroxène riche en Na2O et Al2O3 (jusqu’à Na2O 0.5 wt.%; 6.5 wt.% Al2O3), teneur en Fo de l’olivine de 88.5 à 90.0 mol.%, bas valeurs du Cr# du spinelle (13-17), attestant la nature fertiles de ces roches. A l’inverse les harzburgites costituent des roches très réfractaires : ils ne contiennent pas de clinopyroxène primaire et les teneurs en Fo de l’olivine (90.9-92.9 mol.%), le Mg# de l’orthopyroxène et le Cr# du spinelle (39-71) sont élevés. Les spectres de concentrations en REE présentent des caractéristiques typiques de formation dans un environnement abyssal pour les lherzolites à spinelle, alors que les harzburgites ont des spectres en U typiques d’environnement d’avant-arc. Les compositions en REE des lherzolites à spinelle sont compatibles avec un bas degré de fusion fractionnée (8-9%) d'une source DMM, commençant dans le domaine de stabilité du grenat. Au contraire les concentrations en REE des harzburgites indiquent un haut degré de fusion d’une source DMM, en accord avec une fusion hydratée en environnement d’avant-arc. Les lherzolites à plagioclase présentent des microtextures d’imprégnation, des spinelles riches en Cr2O3 et TiO2 et un enrichissement progressif en REE, Ti, Y, Zr. Les modèles des éléments en trace indiquent que les lherzolites à plagioclase résultent des lherzolites à spinelle par séquestration des liquides MORB très appauvris au sein de la lithosphère océanique.Les roches intrusives sont des gabbronorites à olivine avec des compositions très appauvries (87.3≤Fo ol≤88.9 mol.%, 87.7≤Mg# Cpx≤92.2, An Pl=90-96 mol.%). Le haut Mg#, le bas teneur de TiO2 des pyroxènes, la composition en anorthite du plagioclase et le modèle des éléments en trace montrent que les magmas parents des gabbronorites sont des magmas primitifs, très appauvries, formés dans un environnement de subduction. / The New Caledonia ophiolite hosts one of the largest obducted mantle section in the world, hence providing a unique insight for the study of upper mantle processes. These mantle rocks belong to an “atypical” ophiolitic sequence, which is dominated by refractory harzburgites but it also includes minor spinel and plagioclase lherzolites. Upper crust is notably absent in the ophiolite, with the exception of some mafic-ultramafic cumulates cropping out in the southern part of the island. Although the New Caledonia ophiolite has been under investigation for decades, its ultra-depleted nature has made its characterization an analytical challenge, so that few trace element data are available, while isotopic data are completely missing. In this thesis a comprehensive geochemical study (major, trace element and Sr-Nd-Pb isotopes) of the peridotites and the associated intrusive mafic rocks from the New Caledonia ophiolite has been carried out. The peridotites are low-strain tectonites showing porphyroclastic textures. Spinel lherzolites are undepleted lithotypes, as attested by the presence of 7-8 vol.% of Na2O and Al2O3-rich clinopyroxene (up to 0.5 wt.% Na2O; 6.5 wt.% Al2O3), Fo content of olivine (88.5-90.0 mol.%) and low Cr# of spinel (13-17). Conversely, harzburgites display a refractory nature, proven by the remarkable absence of primary clinopyroxene, very high Fo content in olivine (90.9-92.9 mol.%), high Mg# in orthopyroxene (89.8-94.2) and Cr# in spinel (39-71). REE contents show abyssal-type patterns for spinel lherzolites, while harzburgites display U-shaped patterns, typical of fore-arc settings.Spinel lherzolites REE compositions are consistent with relatively low degree (8-9%) of fractional melting of a DMM source, starting in the garnet stability field. Conversely, REE models for harzburgites indicate high melting degrees (20-25%) of a DMM mantle source under spinel faies conditions, consistent with hydrous melting in forearc setting. Plagioclase lherzolites exhibit melt impregnation microtextures, Cr- and TiO2-enriched spinels and REE, Ti, Y, Zr progressive increase with respect to spinel lherzolites. Impregnation models indicate that plagioclase lherzolites may derive from spinel lherzolites by entrapment of highly depleted MORB melts in the shallow oceanic lithosphere. Mafic intrusives are olivine gabbronorites with a very refractory composition, as attested by high Fo content of olivine (87.3-88.9 mol.%), very high Mg# of clinopyroxene (87.7-92.2) and extreme anorthitic content of plagioclase (An = 90-96 mol.%). The high Mg#, low TiO2 concentrations in pyroxenes and the anorthitic composition of plagioclase point out an origin from ultra-depleted primitive magmas in a convergent setting. Geochemical trace element models show that the parental melts of gabbronorites are primitive magmas with striking depleted compositions, bearing only in part similarities with the primitive boninitic melts of Bonin Islands. The first Sr, Nd and Pb isotope data obtained for the New Caledonia ophiolite highlight the presence of DM mantle source variably modified by different processes.Nd-Sr-Pb isotopic ratios for the lherzolites (+6.98≤ƐNdi≤+10.97) indicate a DM source that suffered low-temperature hydrothermal reactions. Harzburgites are characterized by a wide variation of Sr, Nd and Pb isotopic values, extending from DM-type to EMII compositions (-0.82≤ƐNdi≤+17.55), suggesting that harzburgite source was strongly affected by subduction-related processes. Conversely, combined trace element and Sr-Nd-Pb isotopic data for gabbronorites indicate a derivation from a source with composition similar to Indian-type mantle, but affected by fluid input in subduction environment.

Cette recherche doctorale entend caractériser les fluctuations morphologiques des rennes paléolithiques du Sud-Ouest de la France consécutives aux changements climatiques. En effet, le dernier Pléniglaciaire et le Tardiglaciaire (entre 21 500 - 13 000 cal. BP) ont constitué des périodes de contraintes paléoclimatiques et paléoenvironnementales particulièrement fortes, qui ont occasionné, directement ou indirectement, des répercussions sur la taille corporelle des populations de rennes. Les séries analysées dans le cadre de ce travail proviennent tant du registre actuel que fossile. Plusieurs échantillons de populations actuelles (Norvège, Finlande, Groenland) ont ainsi été analysés, afin de tester notre méthodologie et d’envisager des comparaisons morphologiques avec les populations passées. Le corpus paléolithique comporte 29 assemblages naturels et anthropiques du Sud-Ouest de la France, datés majoritairement du dernier Pléniglaciaire et du Tardiglaciaire. La méthode du Variability Size Index appliquée récemment sur le Renne paléolithique posant des problèmes, s’agissant notamment de la détermination du sex-ratio d’un assemblage, du caractère significatif ou non des différences morphologiques -isométriques et allométriques - observées..., nous avons eu recours à une nouvelle méthodologie combinant plusieurs outils statistiques appliqués à un nouvel indice de taille (VSI* : Variability Size Index modifié). Les analyses ostéométriques et statistiques que nous avons effectuées tendent à identifier des fluctuations morphologiques temporelles attestant d’adaptations de l’espèce. D’un point de vue synchronique, nos résultats permettent de rejeter l’hypothèse précédemment soutenue de différences entre les rennes tardiglaciaires des Landes d’une part et du secteur Dordogne/Gironde d’autre part. Les données saisonnières tendent toutefois à indiquer une absence de migrations de grande ampleur des rennes selon cet axe Nord/Sud. L’existence de populations distinctes est toutefois proposée entre les rennes des Landes et ceux du Bassin de l’Aude. Le Sud-Ouest de la France, à la fin du Paléolithique supérieur, aurait été occupé de manière pérenne par des troupeaux de rennes, dont le comportement migratoire reste cependant à déterminer plus précisément. / This doctoral research aims at characterizing morphological fluctuations resulting from climate change in palaeolithic reindeers from Southwest France. Indeed, the Last Pleniglacial and the Last Glacial (between 21 500 - 13 000 cal. BP) constituted periods of particularly strong palaeoclimatic and palaeoenvironmental constraints, which ended, either directly or indirectly, in repercussions on the body size of reindeer populations. The series analyzed within the framework of this work result from both current and fossil registers. Several samples of current populations (Norway, Finland, and Greenland) were thus analyzed, in order to experiment our methodology and envisage morphological comparisons with the past populations. The palaeolithic corpus is composed of 29 natural and anthropological assemblages of Southwest France, dated mainly from the Last Pleniglacial and the Last Glacial. Since the Variability Size Index method recently applied to palaeolithic Reindeer raises issues, particularly relating to the determination of an assemblage’s sex-ratio, to the characterization or not of observed morphological differences –either isometric and allometric–, we decided to use a new methodology combining several statistical tools applied to a new size index (VSI*: modified Variability Size Index). Osteometric and statistical analyses that we performed tend to identify temporal morphological fluctuations, demonstrating adaptations of the specie. From a synchronic standpoint, our results allow to reject the hypothesis previously asserted regarding differences between Last Glacial reindeers from Landes on the one hand and the Dordogne/Gironde sector, on the other hand. Seasonal data however tend to indicate an absence of large-scale migrations of reindeers according to such a North/South axis. Nonetheless, the existence of different populations between reindeers of Landes and those of the Aude basin is proposed. Hence, Southwest France, at the end of the upper Palaeolithic, would have been occupied continuously by reindeer herds, the migratory behaviour of which however remains to determine more exactly.

Por causa da carência de dados sobre a produtividade de tilápias em tanques-rede em escala comercial, produtores e empresários rurais relutam em investir neste sistema de produção. Este estudo objetivou determinar as funções de produção da tilápia do Nilo em tanques-rede; a biomassa econômica de sistemas de produção comercial de tilápias em tanques-rede de diferentes volumes (300 a 400 peixes/m 3 e 500 a 600 peixes/m 3 ); e analisar a influência das condições ambientais no desempenho dos peixes em tanques-rede. Foram estudados 2 casos na região Sudoeste do Estado de São Paulo: (1) represa do Chapadão, 3,3 ha de espelho d’água e profundidade média de 4,00 m, onde foram instaladas nove gaiolas, totalizando 94,50 m 3 de volume útil de produção; e (2) represa da Colônia Nova, 8,8 ha de espelho d'água e profundidade média de 2,60 m, que alojou 27 gaiolas com volume útil total de 235,70 m 3 , ambas localizadas na Usina Paredão, Oriente, SP. De setembro de 2001 a abril de 2002, foram monitorados diariamente o consumo de ração, a mortalidade de cada gaiola e a temperatura da água das represas nos horários das alimentações; pH, oxigênio dissolvido e transparência pelo disco de Secchi foram monitorados quinzenalmente. Foram realizadas biometrias mensais nas gaiolas das duas represas, compreendendo 3% do lote de peixes de cada gaiola. Os peixes foram alimentados com ração extrusada comercial com 32% de proteína bruta, 3 vezes ao dia (09h00m, 13h00m e 17h00m), 7 dias por semana. A taxa de alimentação foi ajustada com base nas biometrias e mortalidade de cada gaiola. Os dados de desempenho foram analisados por análise de variância e regressão. Para a determinação das funções de produção a partir dos dados de campo, foram utilizadas as funções Logística e de Mitscherlich. Não houve diferença de desempenho em relação à biomassa acumulada e peso médio unitário em função do tempo, na densidade de 300 a 400 peixes/m 3, para ambas as represas (P<0,05), possivelmente porque os dois ambientes apresentaram variações dos parâmetros de qualidade de água muito similares e dentro dos limites aceitáveis para a espécie considerada. A densidade de 500 a 600 peixes/m 3 apresentou maior biomassa acumulada e maior eficiência alimentar que a densidade de 300 a 400 peixes/m 3. Não houve diferença significativa (P<0,05) entre o peso médio unitário dos peixes nas duas densidades, mostrando que o aumento na densidade de estocagem não influenciou o crescimento individual dos peixes. Para a densidade de 500 a 600 peixes/m 3, a biomassa que maximizou a lucratividade da produção (biomassa econômica - BE) foi 145 kg/m 3, ponto em que os peixes devem ser despescados ou remanejados. Para a densidade de 300 a 400 peixes/m 3 , a BE foi de 121 kg/m 3. Para a obtenção de peixes com peso médio unitário de 283g, considerando a BE de 145 kg/m 3 , a densidade de 500 a 600 peixes/m 3 possibilitou otimização do espaço e tempo de produção, melhor eficiência alimentar, maior produção por m 3 de tanque-rede, sendo portanto mais rentável economicamente que a de 300 a 400 peixes/m 3. / Cage farming of tilapia is an excellent alternative to utilize lakes, dams and reservoirs inadequate to conventional aquaculture. Low investment on cage or pond farming of tilapia is credited to the lack of reliable production and economical data for commercial scale farming. This study aimed to determine the production functions of Nile tilapia Oreochromis niloticus in cages; the economic biomass at stocking densities of 300-400 fish/m 3 and 500-600 fish/m 3 commercially produced tilapia in cages of different volumes; and the influence of environmental conditions in fish performance in cages. Two cases in São Paulo State south-west region were studied: (1) reservoir Chapadão (Usina Paredão , Oriente, SP), 3.3 ha of surface area, 4.00 m deep, with 9, 94.50 m 3 cages; and (2) reservoir Colônia Nova (Usina Paredão, Oriente, SP), 8.8 ha of surface area, 2.60 m deep, with 27, 235.70 m 3 cages. Feed intake, survival rate and water temperature were monitored daily during scheduled feedings from September 2001 - April 2002. Dissolved oxygen, pH and transparency of water were monitored each 15 days. Fish growth was evaluated by measuring and weighing 3% of fish of each cage. Caged tilapia fed on commercial, floating pellets (32% crude protein) at 0900, 1300 and 1700 h, 7 days for a week. Feeding rate was adjusted based on sample weight and survival rate. Data were analyzed statistically by ANOVA and regression analysis. The Logistic and Mitscherlich functions were chosen to elaborate the production functions. Carrying capacity of both stocking densities 300-400 fish/m 3 and 500-600 fish/m 3 was 200 kg/m 3. No differences were observed in fish performance regarding accumulated biomass and individual average weight over time between both reservoirs for the stocking density of 300-400 fish/m 3 . The stocking density of 500-600 fish/m 3 presented larger accumulated biomass and better feeding efficiency than the stocking density of 300-400 fish/m 3. No significant differences between individual average weight of fish of both densities were observed (P<0.05), meaning that increasing stocking density didn't influence the individual growth of fish. The biomass that maximized the profit to the stocking density of 500-600 fish/m 3 was 145 kg/m 3. Economic biomass (BE) for stocking density of 300-400 fish/m 3 , was 121 kg/m 3. This density had worse feed conversion than 500-600 fish/m 3. Stocking density of 500-600 fish/m 3 , up to individual average weight of 283 g, present many advantages: optimization of space and production time, better feed efficiency, higher fish production per m 3 of cages, and is more profitable than 300-400 fish/m 3.

This thesis considers the relationship between Germany’s South-West African colony and its British South African counterparts (the Cape Colony, Natal, Rhodesia and, after the second Anglo-Boer War, the Orange River Colony, and the Transvaal) between 1883 and 1915. The chapters consider the complex and fraught relationship, including the British Government’s surprise and the Cape Government’s dismay following Germany’s establishment of the colony: the German public’s pro-Boer stance juxtaposed against the German Government’s refusal to intervene during the second Anglo-Boer War; the Cape Government’s dilemmas over whether to aid German South-West Africa (GSWA) during Germany’s quasi-genocidal campaigns against the Herero and the Nama; efforts to cooperate with German South-West Africa despite labour competition during the period of the unification of South Africa; and the period after 1910, when the diplomatic relationship became an affair of the Union of South Africa, which simultaneously pursued protectionist policy for South African trade, and bilateral cooperation concerning the diamond industry, as well as security along the border between 1911 and 1914. Finally, I consider the impact of the outbreak of the First World War, which saw Germany and GSWA offer support for an Afrikaner Rebellion to draw Britain’s attention away Europe and install a friendly government in South Africa, while also offering the Union an opportunity to conquer GSWA as part of its sub-imperial ambitions. Among the enduring themes are the interplay between political, economic and military developments, including border disputes, illicit trade, labour competition, and armed incursions led by non-state actors. In conclusion, I argue that as the idea of a South African federation progressed, it was driven in part by geopolitical factors and the desire to counter German imperialism. The British Government endorsed a South African union in part to create a South Africa strong enough to fend off German geopolitical threats. / Thesis (PhD (History))--University of Pretoria, 2021. / Historical and Heritage Studies / PhD (History) / Unrestricted

Le lac insulaire constitue un objet géographique tout à fait original par sa structure d'isolat. Il est unevéritable réserve d'eau douce plus ou moins vaste, au sein d'espaces insulaires étroits dominés par unenvironnement maritime. D'origine naturelle ou anthropique, ces lacs s'inscrivent dans la problématiquecomplexe de la ressource en eau de ces territoires étroits soumis à des contraintes grandissantes(démographie, urbanisation, pollutions...). Le sud-ouest de l'océan Indien, composé d'îles trèsdifférentes, constitue un territoire privilégié pour une analyse comparative de ces lacs et de leur place ausein des sociétés insulaires. Ce travail a permis dans un premier temps de recenser et de classifier lesprincipales masses d'eau de la zone. A partir d'un échantillon de lacs, leurs spécificités ont étéobservées pour caractériser leur morphologie (cuvette lacustre et bassins d'alimentation), leurcomportement hydrologique et leurs écosystèmes. Cette analyse a permis de souligner la dimensionmontagneuse, tropicale et le poids de l'endémisme pour ces plans d'eau. La typicité de ces lacs apparaîtpar le croisement de ces composantes qui s'ajoutent au contexte insulaire où le stress hydrique imposedes valorisations spécifiques. L'étude des acteurs, des politiques de gestion et des contextes insulaires apermis de mieux appréhender la place du lac dans les sociétés du sud-ouest de l'océan Indien. En effet,ces plans d'eau, longtemps marginalisés par rapport au littoral, deviennent progressivement un atout etun véritable support au développement de ces territoires. Les contrastes de valorisation d'un lac à l'autreimposent une réflexion à long terme pour que les potentialités de ces masses d'eau soient adaptéesavec les besoins durables des populations insulaires. / The island lake constitutes an absolutely original geographical subject with its isolate system. It is a truefresh water reserve, that is more or less wide, inside narrow island areas dominated by a seaenvironment. Those lakes, which are from natural or anthropological origin, are inscribed in the complexproblem of the water resources of those narrow territories, that are subject to growing constraints –demography, urbanization, pollutions... Composed of very different islands, the South West of the IndianOcean makes up a privileged territory for a comparative analysis of those lakes and their place insideisland societies. First of all this work has allowed to list and classify the main bodies of water in the area.From a sample of lakes, their specificities have been examined to characterize their morphologies –lakeside depression and supply basin – their water performances and their ecosystem. This analysis hasallowed to highlight the mountainous and tropical dimensions and the weight of those lakes' endemism.Their typicality appears thanks to the intersecting of the components that can be added to the islandcontext in which the hydric stress imposes specific exploitation. The study of the actors, the managementpolicies and the island situation has allowed to better comprehend the place of the lake in the societies inthe South West of the Indian Ocean. Indeed those lakes, which have long been marginalized compared tothe coast, have progressively become an asset and a true medium to the development of those territories.The contrast of exploitations from a lake to another imposes a long-term reflection so that thepotentialities of those water bodies fit the sustainable needs of island populations.

Between 1916 and 1928 Dr Louis Fourie, Medical Officer for the Protectorate of South West Africa and amateur anthropologist, amassed a collection of some three and a half thousand artefacts, three hundred photographs and diverse documents originating from or concerned with numerous Khoisan groups living in the Protectorate. He gathered this material in the context of a complex process of colonisation of the area, in which he himself was an important player, both in his official capacity and in an unofficial role as anthropological adviser to the Administration. During this period South African legislation and administration continued the process of deprivation and dehumanisation of the Khoisan that had begun during the German occupation of the country. Simultaneously, anthropologists were constructing an identity for the Khoisan which foregrounded their primitiveness. The tensions engendered in those whose work involved a combination of civil service and anthropology were difficult to reconcile, leading to a form of melancholia. The thesis examines the ways in which Fourie’s collection was a response to, and a part of the consolidation of, these parallel paradigms. Fourie moved to King William’s Town in South Africa in 1930, taking the collection with him, removing the objects still further from their original habitats, and minimising the possibility that the archive would one day rest in an institution in the country of its origin. The different parts of the collection moved between the University of the Witwatersrand and a number of museums, at certain times becoming an academic teaching tool for social anthropology and at others being used to provide evidence for a popular view of the Khoisan as the last practitioners of a dying cultural pattern with direct links to the Stone Age. The collection, with its emphasis on artefacts made in the “traditional” way, formed a part of the archive upon which anthropologists and others drew to refine this version of Khoisan identity in subsequent years. At the same time the collection itself was reshaped and re-characterised to fit the dynamics of those archetypes and models. The dissertation establishes the recursive manner in which the collection and colonial constructs of Khoisan identity modified and informed each other as they changed shape and emphasis. It does this through an analysis of the shape and structure of the collection itself. In order to understand better the processes which underlay the making of the Fourie Collection there is a focus on the collector himself and an examination of the long tradition of collecting which legitimised and underpinned his avocation. Fourie used the opportunities offered by his position as Medical Officer and the many contacts he made in the process of his work to gather artefacts, photographs and information. The collection became a colonial artefact in itself. The thesis questions the role played by Fourie’s work in the production of knowledge concerning the Bushmen (as he termed this group). Concomitant with that it explores the recursive nature of the ways in which this collection formed a part of the evidentiary basis for Khoisan identities over a period of decades in the twentieth century as it, in turn, was shaped by prevailing understandings of those identities. A combination of methodologies is used to read the finer points of the processes of the production of knowledge. First the collection is historicised in the biographies of the collector himself and of the collection, following them through the twentieth century as they interact with the worlds of South West African administrative politics, anthropological developments in South Africa and Britain, and the Khoisan of the Protectorate. It then moves to do an ethnography of the collection by dividing it into three components. This allows the use of three different methodologies and bodies of literature that theorise documentary archives, photographs, and collections of objects. A classically ethnographic move is to examine the assemblage in its own terms, expressed in the methods of collecting and ordering the material, to see what it tells us about how Fourie and the subsequent curators of the collections perceived the Khoisan. In order to do so it is necessary to outline the history of the discourses of anthropologists in the first third of the twentieth century, as well as museum practice and discourse in the mid to late twentieth century, questioning them as knowledge and reading them as cultural constructs. Finally, the thesis brings an archival lens to bear on the collection, and explores the implications of processing the collection as a historical archive as opposed to an ethnographic record of material culture. In order to do this I establish at the outset that the entire collection formed an archive. All its components hold knowledge and need to be read in relation to each other, so that it is important not to isolate, for example, the artefacts from the documents and the photographs because any interpretation of the collection would then be incomplete. Archive theories help problematise the assumption that museum ethnographic collections serve as simple records of a vanished or vanishing lifestyle. These methodologies provide the materials and insights which enable readings of the collection both along and across the grain, processes which draw attention to the cultures of collecting and categorising which lie at the base of many ethnographic collections found in museums today. In addition to being an expression of his melancholy, Fourie’s avocation was very much a part of the process of creating an identity for himself and his fellow colonists. A close reading of the documents reveals that he was constantly confronted with the disastrous effects of colonisation on the Khoisan, but did not do anything about the fundamental cause. On the contrary, he took part in the Administration’s policy-making processes. The thesis tentatively suggests that his avocation became an act of redemption. If he could not save the people (medically or politically), he would create a collection that would save them metonymically. Ironically those who encountered the collection after it left his hands used it to screen out what few hints there were of colonisation. Finally the study leads to the conclusion that the processes of making and institutionalising this archive formed an important part of the creation of the body of ethnography upon which academic and popular perceptions of Khoisan identity have been based over a period of many decades.