Aloe marlothii is a winter-flowering succulent that is widespread in the savanna biome of northern and north-eastern South Africa. Plants grow up to 8 m in height and are commonly found on rocky north-facing slopes. Nectar production occurs through a 24 h period with flowers producing copious amounts (c. 250 µl) of dilute nectar (c. 12%). This abundant nectar supply, that is available for a 5-10 week period during June-August, is utilised by numerous opportunistic avian nectarivores. At a study site in Suikerbosrand Nature Reserve, 60 km south-east of Johannesburg, at least 59% (38 species) of birds recorded during census transects fed on nectar; throughout the range of A. marlothii at least 85 species feed on nectar. This diversity surely far exceeds the number of species ever recorded feeding on nectar of a single plant. During the flowering period an influx of birds at the aloe forest occurred, with an overall increase in abundance and diversity. Pollinator exclusion experiments supported the hypothesis that A. marlothii is pollinated by generalist birds; specialist nectarivores are possibly excluded as inefficient pollinators by the nectar of low concentration and high volume. Fruit set was higher in plants that had avian visitors and very low when pollinators were absent. Stable carbon isotope analysis of whole blood was used to quantify the importance of nectar sugars for opportunistic nectarivores. During flowering there was an enrichment in the δ13C isotopic signature of whole blood of nectar-feeding birds towards that of nectar (δ13C = -12.6‰). This shift was most prominent in frugivores, insectivores and omnivores that typically fed on a diet depleted in 13C when nectar was not available. The C4 grass seed diet of granivores was similar to the isotopic signature ofA. marlothii nectar, so we were unable to determine to what degree granivores benefitted from nectar. Stable nitrogen isotopes in whole blood may suggest that many nectar-feeding birds shift their trophic position during flowering. However, we interpret these results with caution because of insufficient knowledge on diet-tissue fractionation factors of wild birds and/or temporal changes in vegetation isotopic values. Stable carbon isotope analysis of breath samples was used to show that A. marlothii nectar is a readily available income energy source for nectar-feeding birds. Because A. marlothii nectar is so dilute we expected it to be an important water source for many opportunistic nectar-feeding bird species. There was no correlation between the enrichment of δ13C of breath CO2 (representing metabolised nectar sugars) and the δ18O in breath CO2 (representing a highly evaporated water source in nectar); for most birds the δ18O in breath CO2 was more similar to that of free-standing water sources. However, because our knowledge on the relationship between δ18O of ingested water and body water, and fractionation processes when CO2 is exhaled is limited, we were unable to quantify water obtained from nectar. The sugars of A. marlothii nectar are probably more important, as a food source for opportunistic nectarivores during dry winter months when insect abundance is low, than the water in nectar, because birds are able to source water from other drinking sites. / Thesis (PhD)--University of Pretoria, 2010. / Zoology and Entomology / unrestricted

Magister Scientiae (Biodiversity and Conservation Biology) / Fog and dew interception and utilization by plant canopies remains one of the least considered aspects of vegetation studies at any scale yet the few studies that have been conducted point to their considerable influence on ecological processes and a critical role in modulating climate in southern African arid ecosystems. Their relevance to succulent plant hydrology was investigated in this study.The first study measured stable 18O and 2H isotope ratios in samples of rain, fog and dew water and compared these with those assayed monthly in stem xylem water of six succulent shrub species over a one year period. Negative 18O and 2H ratios were observed in the stem xylem water of all six species signifying a predominance of water derived from fog and dew precipitation which was most conspicuous during the wet winter. This implied that fog and dew are even more important sources of water than rain and corroborated by significant correspondence found between fog and dew frequencies, succulent foliar water contents and quantum yields of photochemistry.The second study monitored variations in stem diameter at 2-hourly intervals in 8 succulent shrub species of diverse growth form over a 9-month period. Two groups of species were distinguished based on whether their daily amplitudes in stem diameter were consistently positively correlated with daily fluxes in vapour pressure deficit, which were indicative of a persistent CAM photosynthetic mode, or intermittently correlated with daily fluxes in vapour pressure deficit, which were indicative of mixed CAM and C3 photosynthetic modes. Among species displaying a persistent CAM photosynthetic mode, high nocturnal fog and dew precipitation amounts corresponded with low daily amplitudes in stem diameter, and vice versa, which pointed to reduced nocturnal stomatal water loss. These patterns, which were indistinct among species displaying mixed CAM and C3 photosynthetic modes, were corroborated by small daily amplitudes in stem diameter also consistently observed in one species displaying a CAM photosynthetic mode in ambient than artificially fog and dew excluded environments.The third study monitored changes in water mass at hourly intervals of quartz gravel substrates with different dwarf succulent species assemblages over an 8-month period.Consistently greater net amounts of water were intercepted daily by quartz gravel substrates containing Agyroderma pearsonii than Cephalophylum spissum plants as well as those without plants. These attributed to a high water repellence of A. pearsonii leaves and less radiation absorbed by the paler silvery to grey-green leaves of A. pearsonii leaves than the dark green leaves of C. spissum resulting in lower leaf temperatures and less water loss by transpiration. Quartz gravel soils devoid of plants intercepted nearly 5-times greater amounts of precipitation contributed by fog and dew than that contributed by rain. These precipitation amounts exceeding the high percentages of total hydrological input contributed by fog and dew reported in other ecosystems.The study concludes that fog and dew are a vital source of water for succulent shrubs in arid South African ecosystems and imply that diminished fog and dew frequencies associated with elevated night time temperatures accompanying global warming could exacerbate plant drought stress.

Thesis (MScConsEcol(Conservation Ecology and Entomology)--Stellenbosch University, 2008. / The Succulent Karoo biodiversity hotspot is threatened by pressure caused by increasing human populations and its associated land use types. Land use is primarily focussed on agriculture, with livestock grazing as a dominant land use in the region. Cultivation is also practiced along the major perennial rivers, and in drier areas, where this largely depends on rainfall. Only about seven percent of the biome is formally protected, and this area substantially under-represents the biodiversity of the Succulent Karoo and does not incorporate key ecological processes and biodiversity drivers. Therefore, there is urgent need for outside reserve conservation initiatives, whose success depend on understanding the ecosystem function of the Succulent Karoo. This study aimed to determine the impacts of heavy grazing, light grazing and cultivation (in a 30-year old fallow field) on assemblages of ground-dwelling and flying invertebrates. Seasonal assemblage changes were also determined. Vegetation structure and composition were determined using the line-intercept method to determine if vegetation patterns explain patterns in invertebrate assemblages. Abandoned fields harbour the lowest number of plant species, and these together with the heavily grazed sizes are dominated by a high cover of Galenia africana (Aizoaceae). Lightly grazed sites have the highest structural complexity, with a high cover of succulents and non-succulent perennials. After the winter rains, annual plants occupy most of the bare ground in heavily grazed and previously cultivated sites. Seasonal changes in assemblages of ground-dwelling and flying invertebrates were determined by sampling during the four seasons at the same localities. Results of pitfall traps sampling for ground-dwelling invertebrates and coloured pan traps for flying invertebrates showed that overall species richness and diversity peaked in spring for flying invertebrates while peaks in richness for ground-dwelling invertebrates were in summer, with no difference in overall diversity. Overall abundance for ground-dwelling invertebrates was highest in summer and lowest in winter. Ground-dwelling invertebrate fauna was dominated by Formicidae and Araneae. Grazing and cultivation lead to skewed community composition of ground-dwelling invertebrates which favours disturbance tolerant and generalist species such as Anoplolepis steingroeveri (Forel).

Thesis (M.Sc. (Environmental Science)) - Rhodes University, 2009.

Magister Scientiae (Biodiversity and Conservation Biology) - MSc (Biodiv & Cons Biol) / Fire as a form of disturbance is unique in the way it impacts upon the environment, acting like a herbivore with a ubiquitous appetite. Consuming both dead and living material, converting complex organic molecules into organic and mineral products, which return to the soil. The role of disturbance has long been considered a driver of diversity within Mediterranean type ecosystems. Recently the interest in soil microbes has been piqued, as the importance thereof has been emphasised, particularly their role in nutrient cycling and the chelation of essential plant nutrients. The occurrence of fire results in several environmental and ecological impacts on soil, as well as the dynamics of the microbial populations present. This study explores the impact of fire as a disturbance on the plant and bulk soil microbial communities of the Succulent Karoo. This was achieved through two sub-studies, in which three different states were studied: unburnt, 7-year and 2-year post-fire. Today microbial profiles are also used as indicators of disturbance, thus many techniques exploring microbial community composition are available.

Transformation in the arid succulent thicket of the Eastern Cape of South Africa in response to unsustainable livestock production has been widespread, with less than 10 percent remaining intact. Transformation in succulent thicket has resulted in large areas of dense thicket (comprising a two-phase mosaic of perennial-vegetated patches separated by animal paths and bare patches) being replaced with a ‘pseudo-savanna’ of remnant canopy trees with a structurally simple field layer of ephemeral and short- lived perennial grasses and forbs. There is an extensive literature describing the transformation of succulent thicket, with many speculative statements about the underlying mechanisms of transformation. The central focus of this study was to improve our mechanistic understanding of transformation in succulent thicket using field experiments. Hopefully these results will set another foundation upon which future management of succulent thicket can be improved and large-scale restoration initiated. This study comprises four themes that are linked to the concept of landscape function. The central premise of landscape function is that functional landscapes have mechanisms that capture and retain scarce resources. Conversely, as landscapes become increasingly dysfunctional, so these mechanisms become disrupted. In succulent thicket, dysfunctio n appears to be linked to the reduced ability to harvest water, cycle carbon and a loss of organic carbon. In this thesis I examined some of the key processes that influence water and organic carbon fluxes: perennial vegetation cover, soil fertility, litter fall and decomposition, and runoff and soil erosion. The experimental design that was used for all this work was a factorial ANOVA based on replicated fenceline contrasts that reflect differences in long-term management history. The main objectives of this thesis were to: quantify the patterns of transformation in an arid form of succulent thicket, including changes in the biomass, cover and structure of the dominant vegetation guilds; test the stability of the transformed succulent thicket ecosystem to show whether it is a new stable state or an intermediate stage in a trajectory towards a highly desertified state where only the ephemeral grasses and forbs persist; describe and compare soil fertility across transformation contrasts, concentrating on changes in the spatial patterns of soil resources and the ability of the soil to harvest precipitation; to compare litter fall and decomposition of leaf material from the dominant plants in intact and transformed succulent thicket; to quantify and compare run-off and erosion from run-off plots in intact and transformed succulent thicket. Transformation and stability I quantified the changes in plant diversity, physiognomy and biomass that occur across transformation contrasts. Thicket transformation results in a significant loss of plant diversity and functional types. There is also a significant reduction in the biomass (c. 80 t.ha-1) and structural complexity of the vegetation, both vertically and horizontally. These results were interpreted in terms of their implications for ecosystem functioning and stability. To test the stability of the transformed succulent thicket I used aerial photographs and ground-truthing to track the survivorship of canopy trees over 60 years in pseudo-savanna landscapes. I also measured seedling establishment in different habitats. I show that the pseudo-savanna is not a stable state owing to ongoing adult mortality and no recruitment of canopy trees. Soil fertility and water status I hypothesised that the above-ground changes in ve getation would be accompanied by similar trends in the pattern and levels of soil nutrient resources and the ability of the landscape to harvest precipitation. I compared soil fertility (organic carbon, available nitrogen and phosphorus), texture, matric potential, and surface micro-topography in the two main micro- habitats on either side of the replicated fenceline contrasts. The results show that intact spekboom thicket has a distinct spatial pattern of soil fertility where nutrients and organic carbon are concentrated under the patches of perennial shrubs, compared to under canopy trees and open spaces. Transformation results in a significant homogenisation out of this pattern and an overall reduction in the fertility of the landscape. The proportion of the landscape surface that would promote infiltration of water decreases from 60 – 0.6 percent. Soil moisture retention (matric potential) also decreases with transformation. I interpreted these patterns in terms of the ability of the landscape to harvest and release water after rainfall events. Litter fall and decomposition Surface litter and soil organic matter are critical components to wooded ecosystems; contributing to several ecosystem functions. The rates of litter fall and decomposition are ratelimiting steps in nutrient cycling and incorporation of organic matter into the soil. The ecological mechanisms behind the collapse of succulent thicket in the face of domestic herbivory are not fully understood, but are believed to include the breakdown of several ecosystem processes, including litter fall and decomposition. I quantified the changes in litter fall and litter decomposition of four of the dominant perennial woody plants (Euclea undulata, Pappea capensis, Portulacaria afra and Rhus longispina) across the replicated fenceline. Litter fall was measured over 14 months using mesh traps. Decomposition was measured over 15 months using a combination of litterbags and unprotected leaf packs. I also quantified soil microclimate during the experimental period; hypothesising that transformation would lead to soil conditions less amenable for biotic activity. Litter fall in succulent thicket was very high for a semi-arid system, comparing more to temperate forests. The leaf-succulent P. afra contributed the largest single component of the total litter production at a landscape scale. The effect of transformation on litter fall was species specific. Deep-rooted or drought-adapted species showed no change in litter yield with transformation; shallow-rooted species showed a significant decrease. There were few significant differences in decomposition rates across the transformation gradient and between litter types. Portulacaria afra litter had the steepest rate of mass loss, and was most affected by transformation. The more recalcitrant (high C:N ratio) leaves of P. capensis remained largely unaffected by transformation. These results indicate the critical role of the perennial vegetation in incorporating organic carbon into the soil. Transformation of succulent thicket leads to a disruption of the flow of carbon into the soil, reinforcing the cycle of transformation through reduced fertility. Rehabilitation of this ecosystem will require the active establishment of species, such as P. afra, that will restart the flow of carbon into the soil. Run-off & Erosion The landscape function model predicts that functional semi-arid shrublands efficiently conserve limiting resources such as water and water-bourn sediments (soil and organic matter). As these rangelands become transformed through unsustainable livestock production, so their ability to conserve resources decreases. The primary determinant of landscape function and conservation of resources appears to be the proportional cover of perennial vegetation. I hypothesised that the switch from a two-phase mosaic dominated by perennial succulent and woody shrubs to a single phase system dominated by an ephemeral field layer would be accompanied by disruption of the mechanisms that conserve resources. Specifically, I tested the hypothesis that transformation of succulent thicket increases runoff volume, sediment concentration of runoff, soil erosion and loss of organic matter at a patch scale (c. 100 m2). Runoff and water-borne sediment were measured from runoff plots established across replicated fenceline contrasts. Data were collected from eight extreme weather events over two years. There were no significant differences between runoff and erosion across the transformation contrast, as each extreme weather event was unique in terms of its runoff response. The transformed runoff plots alone also gave inconsistent results, largely due to differences in the cover of ephemeral forbs and weakly perennial grasses. Runoff and erosion were not predictable from the data across the transformation gradient due to complex interactions between the nature of the above-ground vegetation, soil micro-topography and land use history. The results highlighted the need for longer-term catchment experiments to generate a predictive understanding of the effect of transformation on runoff and erosion in succulent thicket.

Philosophiae Doctor - PhD / Gethyllis multifolia L. Bolus and G. villosa Thunb. are winter-growing, summerblooming,deciduous and bulbous geophytes that grow naturally in the semi-arid succulent Karoo biome of South Africa. Both species grow under full sun conditions and have four distinctive growth phases: a winter (cold and wet) growing phase, leaf senescence phase towards spring, flowering phase during the hot and dry summer months, and fruit and leaf formation phase in autumn. The medicinal uses of this genus (including G. multifolia “Kukumakranka” and G.villosa “hairy kukumakranka”) range from cures for colic, digestive disturbances,teething problems, fatigue, boils, bruises and insect bites, to being used as an aphrodisiac. Gethyllis multifolia is threatened in its natural habitat and is listed in the ‘Vulnerable’ category of the ‘Red Data List of Southern African Plants’ and the ‘IUCN-World Conservation Union List of Plants’. The literature indicate that the habitats of both species are being exposed to drier conditions and is further threatened by the encroachment of invasive indigenous plant species. It is not known to which extent these factors may pose a threat to the existence of both species. The first objective of this investigation was to determine the costs of vegetative and reproductive growth during the seasonal life cycle of the plant, using carbon(C) and nitrogen (N) as a physiological currency. The second objective was to elucidate a functional basis to explain the difference in the conservation status of both species in their natural habitat. Both species were subjected to drought and shading as environmental stresses and the plant physiological performance was investigated via photosynthetic gas exchange. The third objective of the study was to evaluate the antioxidant content (total polyphenol, flavonol/flavone and flavanone content) and antioxidant capacity [ferric reducing antioxidant power (FRAP), oxygen radical absorbance capacity (ORAC) and 2,2'-azino-di-3-ethylbenzthiazoline sulphonate (ABTS) radical cation scavenging ability] of natural populations and plant samples that were exposed to photo-and-drought environmental stresses. This study was done to elucidate the antioxidant profile of plant parts of natural populations as well as providing farmers, traditional healers and pharmaceutical companies with cultivation environmental conditions to enhance the antioxidant properties of the species. This investigation also attempted to isolate and characterize, by means of thin-layer chromatography (TLC) and column chromatography (CC), natural compounds from both species to lend support to the purported antioxidant benefit of both species and to further lend support to claims made by traditional healers of the medicinal potential of the genus. This study, however, did not engage in any in vivo studies or human trials to support published literature of the medicinal benefits of the genus.

The vertical cliff-face habitat is renowned for many specifically adapted plant species that exhibit a high degree of local endemism. Over a period of nine years the succulents and bulbous succulents on cliff faces in South Africa and Namibia were systematically surveyed and documented. Distinction was made between succulents growing on cliffs as part of a wider habitat and those found only on cliffs (obligate cremnophytes). Most major cliff-face habitats in the study area were visited and all plants were documented. A check list and descriptions (including adaptive traits) of the 220 obligate cremnophilous taxa are provided. During the study some 45 new cremnophilous succulent taxa were discovered and named, representing almost 20% of the total and proving that cliff habitats are some of the least studied environments, not only in southern Africa but globally. Among the newly described cremnophilous taxa is the genus Dewinteria (Pedaliaceae). Using stem length, three basic cliff-face growth forms are identified - compact or cluster-forming ‘cliff huggers’, cliff shrublets or ‘cliff squatters’ and pendent ‘cliff hangers’. Compact growth (often tight clusters or mats) is mainly associated with the winter-rainfall Succulent Karoo and Thicket regions, especially Namaqualand. However, further north the same compact growth forms are associated with an increase in altitude such as the Drakensberg Escarpment and other northern mountains. Most pendent growth forms are associated with the eastern and southeastern summer-rainfall regions; a number of smaller pendent shrublets occur on the high quartzitic sandstone mountains of the Western Cape. The degree of specialisation varies from highly adapted (smaller percentage) to less specialised (often eco-forms), and some taxa have no obvious adaptations. This study revealed a general increase in succulence in most obligate cremnophilous succulent species (compared to closely related species in other habitats), a reflection of their xeric habitat, and plants tend to be more compact. Also, there is a shift in reproductive output, including an increase in vegetative reproduction (backup), wind-dispersed seed and enriched flowering associated with certain species. Most obligate cremnophilous succulent plants in the study area have cliff-adapted features, ensuring long-term survival. / Thesis (PhD)--University of Pretoria, 2012. / Plant Science / unrestricted

Namaqualand is a semi-desert area situated in the winter rainfall region of South Africa hosting the world’s richest succulent flora, and is one of only two arid areas to classify as a global hotspot of biodiversity. The Goegap Nature Reserve, east of Springbok, lies in the Upland Succulent Karoo. After all the livestock were removed and the reserve fenced in 1969, a long-term monitoring project was initiated to record possible vegetation changes under the reduced grazing pressure. The first monitoring was done in 1974 when two line transects, of 1 km each, were surveyed. These surveys have been repeated annually covering a period of more than 30 years. When the reserve was enlarged in 1990, five additional line transects were set out in highly degraded areas. The main aim of this study was therefore to determine whether any changes in the vegetation in terms of species composition, species richness and life form richness in the monitored line transects could be detected. Rainfall as possible force driving these changes was also examined. It was found that changes did indeed take place in the vegetation. The species composition revealed notable changes over the entire monitored period, specifically in the two longest established line transects, whereas in the case of the species richness changes were also detected, but to a lesser extent and not notable in terms of an overall increase or decrease in the number of species. These changes are referred to as fluctuations. Rainfall can be regarded as a definite environmental driving force, with the time, intensity and duration of importance, specifically to the annual species composition. Overall, the veld condition improved and continuous monitoring of the line transects on a regular basis should take place to improve our understanding of the complex dynamics of this ecosystem. Because the changes in species composition are reflected in changes in grazing capacity the results are of direct importance to the management of the wildlife on the reserve. / Dissertation (MSc)--University of Pretoria, 2010. / Plant Science / unrestricted

Parmi les lignées des Caesalpinioideae (dans la famille des Leguminosae), l’un des groupes importants au sein duquel les relations phylogénétiques demeurent nébuleuses est le « groupe Caesalpinia », un clade de plus de 205 espèces, réparties présentement entre 14 à 21 genres. La complexité taxonomique du groupe Caesalpinia provient du fait qu’on n’arrive pas à résoudre les questions de délimitations génériques de Caesalpinia sensu lato (s.l.), un regroupement de 150 espèces qui sont provisoirement classées en huit genres. Afin d’arriver à une classification générique stable, des analyses phylogénétiques de cinq loci chloroplastiques et de la région nucléaire ITS ont été effectuées sur une matrice comportant un échantillonnage taxonomique du groupe sans précédent (~84% des espèces du groupe) et couvrant la quasi-totalité de la variation morphologique et géographique du groupe Caesalpinia. Ces analyses ont permis de déterminer que plusieurs genres du groupe Caesalpinia, tels que présentement définis, sont polyphylétiques ou paraphylétiques. Nous considérons que 26 clades bien résolus représentent des genres, et une nouvelle classification générique du groupe Caesalpinia est proposée : elle inclut une clé des genres, une description des 26 genres et des espèces acceptées au sein de ces groupes. Cette nouvelle classification maintient l’inclusion de douze genres (Balsamocarpon, Cordeauxia, Guilandina, Haematoxylum, Hoffmanseggia, Lophocarpinia, Mezoneuron, Pomaria, Pterolobium, Stenodrepanum, Stuhlmannia, Zuccagnia) et en abolit deux (Stahlia et Poincianella). Elle propose aussi de réinstaurer deux genres (Biancaea et Denisophytum), de reconnaître cinq nouveaux genres (Arquita, Gelrebia, Hererolandia, Hultholia et Paubrasilia), et d’amender la description de sept genres (Caesalpinia, Cenostigma, Coulteria, Erythrostemon, Libidibia, Moullava, Tara). Les résultats indiquent qu’il y aurait possiblement aussi une 27e lignée qui correspondrait au genre Ticanto, mais un échantillonage taxonomique plus important serait nécéssaire pour éclaircir ce problème. Les espèces du groupe Caesalpinia ont une répartition pantropicale qui correspond presque parfaitement aux aires du biome succulent, mais se retrouvent aussi dans les déserts, les prairies, les savanes et les forêts tropicales humides. À l’échelle planétaire, le biome succulent consiste en une série d’habitats arides ou semi-arides hautement fragmentés et caractérisés par l’absence de feu, et abrite souvent des espèces végétales grasses, comme les Cactacées dans les néo-tropiques et les Euphorbiacées en Afrique. L’histoire biogéographique du groupe Caesalpinia a été reconstruite afin de mieux comprendre l’évolution de la flore au sein de ce biome succulent. Ce portrait biogéographique a été obtenu grâce à des analyses de datations moléculaires et des changements de taux de diversification, à une reconstruction des aires ancestrales utilisant le modèle de dispersion-extinction-cladogenèse, et à la reconstruction de l’évolution des biomes et du port des plantes sur la phylogénie du groupe Caesalpinia. Ces analyses démontrent que les disjonctions trans-continentales entre espèces sœurs qui appartiennent au même biome sont plus fréquentes que le nombre total de changements de biomes à travers la phylogénie, suggérant qu’il y a une forte conservation de niches, et qu’il est plus facile de bouger que de changer et d’évoluer au sein d’un biome différent. Par ailleurs, contrairement à nos hypothèses initiales, aucun changement de taux de diversification n’est détecté dans la phylogénie, même lorsque les espèces évoluent dans des biomes différents ou qu’il y a changement de port de la plante, et qu’elle se transforme, par exemple, en liane ou herbacée. Nous suggérons que même lorsqu’ils habitent des biomes très différents, tels que les savanes ou les forêts tropicales humides, les membres du groupe Caesalpinia se retrouvent néanmoins dans des conditions écologiques locales qui rappellent celles du biome succulent. Finalement, bien que la diversité des espèces du biome succulent ne se compare pas à celle retrouvée dans les forêts tropicales humides, ce milieu se distingue par un haut taux d’espèces endémiques, réparties dans des aires disjointes. Cette diversité spécifique est probablement sous-estimée et mérite d’être évaluée attentivement, comme en témoigne la découverte de plusieurs nouvelles espèces d’arbres et arbustes de légumineuses dans la dernière décennie. Le dernier objectif de cette thèse consiste à examiner les limites au niveau spécifique du complexe C. trichocarpa, un arbuste des Andes ayant une population disjointe au Pérou qui représente potentiellement une nouvelle espèce. Des analyses morphologiques et moléculaires sur les populations présentes à travers les Andes permettent de conclure que les populations au Pérou représentent une nouvelle espèce, qui est génétiquement distincte et comporte des caractéristiques morphologiques subtiles permettant de la distinguer des populations retrouvées en Argentine et en Bolivie. Nous décrivons cette nouvelle espèce, Arquita grandiflora, dans le cadre d’une révision taxonomique du genre Arquita, un clade de cinq espèces retrouvées exclusivement dans les vallées andines. / Amongst the lineages of the Caesalpinioideae (in the family Leguminosae), one of the largest groups where phylogenetic relationships remains unclear is the Caesalpinia Group, a clade of ca. 200 species, currently considered to comprise between 14 and 21 genera. The taxonomic complexity of the Caesalpinia Group stems from persisting doubts on the generic delimitations within Caesalpinia sensu lato, a group of 150 species that are provisionally classified into eight genera. In order to establish a stable generic classification, phylogenetic analyses of five chloroplastic loci and the nuclear ribosomal ITS locus were carried out on a matrix containing an unprecedented taxonomic sampling of the Caesalpinia Group (~84% of species of this group included), with virtually all of the morphological variation and geographic distribution represented. These analyses allowed us to determine that several genera of the Caesalpinia Group, as currently defined, are polyphyletic or paraphyletic. We consider that there are 26 well-resolved clades that represent distinct genera, and a new generic classification system is proposed, which includes a key to genera, the description of the 26 genera and all species accepted within these groups. A total of twelve previously accepted genera are maintained in this classification (Balsamocarpon, Cordeauxia, Guilandina, Haematoxylum, Hoffmanseggia, Lophocarpinia, Mezoneuron, Pomaria, Pterolobium, Stenodrepanum, Stuhlmannia, and Zuccagnia), whereas two genea are abolished (Stahlia and Poincianella). In addition, two genera are re-instated (Biancaea and Denisophytum), five new genera are described, (Arquita, Gelrebia, Hererolandia, Hultholia and Paubrasilia), and the description of seven genera are emended (Caesalpinia, Cenostigma, Coulteria, Erythrostemon, Libidibia, Moullava, Tara). Our results also indicate that there could possible be a 27th lineage corresponding to the genus Ticanto, but an increased taxonomic sampling is needed to adequately address this issue. The Caesalpinia Group has a pantropical distribution that corresponds almost perfectly to the geographical distribution of the Succulent Biome, but are also found in deserts, grassland prairies, savannahs, and tropical rainforests. On a planetary scale, the Succulent Biome consists of a series of semi-arid to arid habitats that are highly fragmented, and which are characterised by the absence of fire, such as deserts and dry forests. This biome often harbours succulent plant taxa, such as the Cactaceae in the Neotropics and the Euphorbiaceae in Africa. The biogeographical history of the Caesalpinia Group was reconstructed in order to gain insight into the evolution of the flora within this Succulent biome. This biogeographical portrait of this group was reconstructed using molecular dating analysis, diversification rate shifts tests, the reconstruction of ancestral areas using the dispersal-extinction-cladogenesis model (DEC), as well as through ancestral character reconstruction of the biomes and habits. These analyses demonstrate that intercontinental disjunctions between sister species belonging to the same biome are more frequent than the total number of biome shifts across the phylogeny, suggesting that there is a strong conservation of niches, and that it is easier to move than to switch to and evolve in a different biome. Furthermore, contrary to our initial hypothesis, no changes in diversification rates were detected in our phylogenies, even when species switched biomes or evolved a different plant habit, e.g. becoming lianas or herbaceous perennials. We suggest that even when members of the Caesalpinia Group inhabit different biomes, such as savannahs or tropical rainforests, they are still tracking local ecological conditions that are typical of the Succulent biome. Finally, while total plant species diversity in the Succulent Biome does not compare to the diversity found in tropical rainforests, this biome distinguishes itself by a high number of endemic species, distributed in disjunct patches across the world. This species diversity is probably under-estimated and needs to be carefully re-evaluated, as shown in several recent descriptions of new tree and shrub species from the Succulent biome, all published in the last decade. The last objective of this thesis is to examine the species limits in Caesalpinia trichocarpa, a shrub from the Andes that has a disjunct population in Peru, which potentially represents a new species. Morphological and molecular analyses of populations occurring across the Andes, including Bolivia and Argentina, allow us to conclude that the populations in Peru represent a new species, which is genetically distinct and has subtle morphological characteristics that allow it to be distinguished from populations found in Argentina and Bolivia. We describe this news species, Arquita grandiflora, in a taxonomic revision of the genus Arquita, a clade of five species found exclusively in Andean valleys.