La plasticité comportementale et les interactions entre organismes sont connues pour avoir une influence sur l’utilisation de l’espace, notamment la sélection de l’habitat de chasse. Dans l’Atlantique Nord-Est, les phoques gris et veaux marins vivent en sympatrie. Ces deux espèces sont réparties dans différentes colonies où les interactions entre organismes (liées à la taille de la colonie et à la présence simultanée ou non des deux espèces) ainsi que la disponibilité de l’habitat (et donc des proies) varient localement. L’objectif de cette thèse était d’étudier l’utilisation spatiale et la sélection d’habitat de chasse à l’échelle locale de ces deux espèces dans des situations contrastées de dynamique de colonies. Pour cela, plusieurs individus de différentes colonies situées en Irlande, en Écosse et en France ont été équipés de balises GPS/GSM. Les données récoltées ont permis de caractériser leurs trajets en mer ainsi que d’identifier les zones de chasse permettant par la suite la sélection d’habitat. Contrairement aux veaux marins, qui ont réalisé des déplacements restreints aux alentours des colonies, la plupart des phoques gris se sont déplacés entre différentes colonies. Ainsi, l’étude de la sélection d’habitat à l’échelle individuelle pour les individus se déplaçant entre différentes colonies a mis en évidence l’influence de la plasticité comportementale et suggéré celle de la personnalité de l’animal. A l’échelle de la colonie, chez les deux espèces, les résultats ont mis en évidence l’influence de la disponibilité de l’habitat (et indirectement de la disponibilité des proies) ainsi que des interactions intra-spécifiques (liée à l’effet de densité dépendance) sur l’utilisation spatiale et la sélection de l’habitat de chasse. Cependant, la sélection de l’habitat de chasse était principalement liée à la distance au reposoir et la bathymétrie. Ces deux paramètres sont liés aux caractères de chasseur à place centrale et benthique des phoques. Cette étude a également mis en évidence, pour les sites où les deux espèces étaient présentes, une certaine ségrégation spatiale et des différences de sélection d’habitat entre les deux espèces. / The behavioural plasticity and the interactions between organisms are known to influence foraging habitat selection and spatial usage. In the Northeast Atlantic, greyand harbour seals live in sympatry. These two species are spread into different colonies where interactions between organisms (linked to the colony size et and the presence of one or both species) and habitat availability (linked to the prey distribution) vary locally. The objective of this PhD was to study the spatial usage and foraging habitat selection of these two species at the local scale, in contrasted situations of population dynamics. Several individuals were tagged with GPS/GSM tags in different colonies located in Ireland, Scotland and France. Telemetry data was used to characterize the seals’ trips at sea and to identify their foraging areas in order to model foraging habitat selection. While harbour seals only performed trips restricted around their haulout sites, grey seals moved between colonies. Studying habitat selection at the individual scale, for individuals moving between colonies, highlighted the influence of behavioural plasticity and suggested the influence of individual personality. At the colony scale, for both species, results highlighted the influence of habitat availability (indirectly linked to prey availability) and intra-specific interactions (linked to density dependence effects) on spatial usage and foraging habitat selection. However, the foraging habitat selection was mainly explained by the distance from the last haulout and the bathymetry. These two parameters are linked to the central place forager and benthic feeder characteristics of the seals. Furthermore, this study also highlighted, for study sites where both species occur, a spatial segregation and differences in foraging habitat selection between grey and harbourseals.

Les champignons mycorhiziens à arbuscules (CMA), classés dans le phylum Glomeromycota, ne peuvent pas être facilement identifiés par la morphologie de leurs spores et leurs mycélia à l'intérieur ou à l'extérieur des racines de leurs hôtes. Ce problème fondamental d'identification rend l'étude de leur diversité, en particulier dans leur habitat naturel (sol et racine) extrêmement difficile. Les gènes ribosomaux ont été largement utilisés pour développer des amorces spécifiques et en inférer des arbres phylogénétiques. Cependant, ces gènes sont très polymorphes et existent en plusieurs copies dans le génome des CMA, ce qui complique l’interprétation des résultats. Dans notre étude, nous avons étudié le polymorphisme intra- et inter-spécifique du gène β-tubuline, présent en faible nombre de copies dans le génome des CMA, afin d’obtenir de nouvelles séquences nucléotidiques pour développer des marqueurs moléculaires. Les gènes β-tubuline amplifiés à partir de l'ADN génomique de cinq espèces du genre Glomus ont été clonés et séquencés. L’analyse des séquences indique un polymorphisme intraspécifique chez trois espèces de CMA. Deux séquences paralogues très variables ont été nouvellement identifiées chez les G. aggregatum, G. fasciculatum et G. cerebriforme. Aucun polymorphisme n’a été détecté chez les G. clarum et G. etunicatum. Toutes les séquences montrent la présence de deux introns hautement variables. La majorité des substitutions ont été localisées dans les exons et sont synonymes à 90%. La conservation des acides aminés suggère un niveau élevé de sélection négative sur le gène β-tubuline et nous permet de confirmer que les CMA représentent un ancien groupe fongique (400 million d’années). L’analyse phylogénétique, réalisée avec vingt et une séquences nucléotidiques du gène β-tubuline, a révélé que les séquences des Glomaceae forment un groupe monophylétique bien supporté, avec les Acaulosporaceae et Gigasporaceae comme groupe frère. Les séquences paralogues nouvellement identifiées chez les G. aggregatum et G. fasciculatum n'ont pas été monophylétiques au sein de chaque espèce. Les oligonucléotides ont été choisis sur la base des régions variables et conservées du gène β-tubuline. Le test PCR des amorces β-Tub.cerb.F/ β-Tub.cerb.R a révélé des bandes spécifiques de 401 pb pour les séquences paralogues du G. cerebriforme. Deux paires d’amorces ont été développées afin d’identifier les séquences du groupe nommé Tub.1. Les tests PCR nous ont permis d’identifier certaines séquences du groupe Tub.1. Une paire d’amorce β-Tub.2.F/ β-Tub.2.R nous a permis d’identifier certaines séquences paralogues du groupe nommé Tub.2. L’analyse d’autres gènes combinée à celle du gène β-tubuline permettra le développement de marqueurs moléculaires plus spécifiques pour l’identification de CMA. / Arbuscular mycorrhizal fungi (AMF) are classified in the phylum Glomeromycota. This fungal group cannot be easily identified using morphology of their spores and hyphae inside or outside plant-roots. This fundamental problem renders the study of their diversity extremely difficult, particularly in their natural habitat (soil and roots). The ribosomal genes have been widely used to develop specific primers and to infer phylogenetic trees. However, these genes are highly polymorphic and exist in multiple copies in the genome of the AMF. This leads misinterpretation of the results. In our study, we analysed the intra- and inter specific polymorphism of the β-tubulin gene, which is present in low copy number in AMF genome, to obtain further nucleotide sequences to develop molecular markers. β-tubulin genes were sequenced from genomic DNA of five Glomus species. PCR amplified β-tubulin genes were cloned and sequenced. Intra- and inter-specific polymorphism analyses indicate the polymorphic nature of β-tubulin gene in three AMF species. Two highly variable β-tubulin paralogs were newly identified in G. aggregatum, G. fasciculatum and G. cerebriforme. No evidence for the presence of paralogs was found in G. clarum and G. etunicatum. All AMF sequences show the presence of two introns, which are highly variable. The majority of substitutions are located in the exons and 90% are synonymous. The conservation of amino-acids level suggests a high level of negative selection acting on the β-tubulin gene and allows us to consider that the AMF represent an ancient fungal group (400 million years). A phylogenetic analysis, carried out with twentyone β-tubulin nucleotidic sequences, revealed that the sequences of Glomeraceae form a highly supported monophyletic group with Acaulosporaceae and Gigasporaceae as sistergroup. The newly identified paralogs from G. aggregatum and G. fasciculatum were not found to be monophyletic within each species. Oligonucleotides were designed in the conserved and variable regions. PCR amplification test of β-Tub.cerb.F / β-Tub.cerb.R revealed specific bands of 401 bp for the G. cerebriforme paralogs. Two pairs of primers were developed to identify sequences of the group named Tub.1. PCR tests have allowed us to identify certain sequences of Tub.1 group. A primer pair of β-Tub.2.F / β-Tub.2.R we identified some sequences paralogous group named Tub.2. Analysis of other genes combined with that of β-tubulin gene enable the development of more specific molecular markers for AMF identification.

The south coast of South Africa represents the extreme western end of the range of the Indo-Pacific humpback (Sousa chinensis, plumbea type) and Indo-Pacific bottlenose dolphins (Tursiops aduncus), which are both confirmed to range as far west as False Bay (Jefferson & Karczmarski, 2001; Hammond et al., 2008). Individual ranging behaviour for both species however is not well resolved. Recent genetic analyses suggest that animals currently considered as plumbea type Sousa chinensis (Reeves et al., 2008) may be a separate species, Sousa plumbea (Mendez et al., 2013). In South African waters less than 1000 adult humpback dolphins (Sousa chinensis, plumbea type hereafter “humpback dolphin”) may comprise the entire population (Karczmarski, 1996), while all estimates suggest the bottlenose dolphin (Tursiops aduncus, hereafter “bottlenose dolphins”) population is relatively large, numbering thousands of animals (Cockcroft et al., 1992; Reisinger & Karczmarski, 2010). Both dolphin species are exposed to variable levels of anthropogenic impacts throughout their range including vessel traffic, chemical pollution and habitat degradation associated with coastal development. This thesis describes the results of a study investigating: 1) the environmental and anthropogenic factors which influence the habitat use of humpback and bottlenose dolphins in two adjacent bays on the southern Cape coast, South Africa – Mossel Bay and Vlees Bay; 2) the abundance of humpback dolphins using Mossel Bay and 3) the interaction of these two dolphin species with white sharks, and the influence this has on dolphin group sizes and habitat use in Mossel Bay. Both land-based and boat-based survey platforms were used in this study with land-based data collected during dedicated watch periods at sites in Mossel Bay (n = 6) and Vlees Bay (n = 4) between February 2011 and March 2013, with a focus on humpback and bottlenose dolphins. A surveyor’s theodolite was used at these sites to collect positional data on animals, while behavioural data were collected through direct observation. Boat-based photographic identification surveys were used to collect data on the presence of individual humpback dolphins in Mossel Bay between April 2011 and November 2013. White shark data from Mossel Bay between February 2011 and March 2013 were provided from boat-based chumming surveys for the collection of photo-ID data from the Master’s thesis of Rabi’a Ryklief, based at Oceans Research. Data were analysed using ANOVA’s, Tukey honest significance tests and generalised additive modelling (Wood, 2006) in programme R, while capture histories of humpback dolphins were analysed with RMark (Laake, 2013) using POPAN open population models (Schwarz & Arnason, 1996) and Huggins heterogeneity closed capture models (Huggins, 1989; Chao et al., 1992). Humpback dolphins socialised over sandy beach habitats in both bays, while feeding/foraging occurred over reef systems in Mossel Bay and off fine grained sandy beach habitats in Vlees Bay. Humpback dolphin resting behaviour was observed at a very low frequency and occurred in all of the primary habitat types in Mossel Bay, while in Vlees Bay resting was only observed over reefs. Bottlenose dolphins in both bays preferentially used wave cut rocky platform habitats for feeding/foraging and resting while socialising occurred in the vicinity of estuaries in Mossel Bay and fine grained sandy beach habitats in Vlees Bay. Higher sighting rates were recorded in the control site, Vlees Bay, than in Mossel Bay for both dolphin species. The largest reverse osmosis desalination plant commenced operations in the sheltered corner of Mossel Bay in October 2011 and discharged approximately five million litres (Ml) of effluent per day (between October 2011 and February 2012) and 18 Ml per day in March and April 2012. In Mossel Bay higher sighting rates of humpback dolphins occurred in the period before desalination began while bottlenose dolphin sighting rates were highest after active desalination decreased to once per month (May, 2012). During the period of peak brine discharge in Mossel Bay, sighting rates were highest for both species in Vlees Bay. Even after desalination operations decreased the sighting rate of humpback dolphins remained low. The operation of the desalination plant at full capacity in Mossel Bay may have led to reduced use of this area by both humpback and bottlenose dolphins. Key habitats in Mossel Bay for both dolphin species are shared with great white sharks (Carcharodon carcharias hereafter “white sharks”) and focus around the three estuaries and their associated near-shore reef systems. The presence of predatory white sharks may limit the time dolphins spend in a specific habitat and influence the number of animals within groups, with larger humpback dolphin groups at sites with high shark utilisation. Both dolphin species had lower individual sighting rates during periods when white shark abundance peaked. Large group sizes of humpback dolphins at Seal Island, and of bottlenose dolphins at Hartenbos and Tergniet, combined with increased rates of travelling and decreased resting and socializing suggest that these areas may pose the largest threat to dolphins due to the variety of shark size classes’ present, especially larger sharks. Closed capture models generated within year population estimates ranging from 48 to 97 individual humpback dolphins (2011: 97, 95% CI: 46 – 205; 2012: 48, 28 – 81; 2013: 68, 35 – 131) while open population modelling produced a ‘super-population’ estimate of 116 animals (95% CI: 54 – 247) using Mossel Bay. During the study 67 humpback dolphins were individually identified with 94.3 % of the individuals in good quality photographs distinctively marked. Fewer humpback dolphins may be present on the south-east and southern Cape coast, including between Algoa Bay and Mossel Bay, than initially thought (Karczmarski, 1996), as definite links exist between Algoa Bay and Plettenberg Bay (Smith-Goodwin, 1997), and Plettenberg Bay and Mossel Bay (this study). The Gouritz River mouth (21º 53' E; Ross, 1984) and De Hoop (20º 30' E; Findlay et al., 1992) were previous suggested as the western limit of this species, but within the last 20 years knowledge on the extent of their range has been greatly improved, and range extension of this species may be occurring to the west with animals present as far west as False Bay (18º 48' E; Jefferson & Karczmarski, 2001). Due to the vulnerability of this species and their wide ranging behaviour, conservation plans need to be implemented on a wide scale to ensure protection of these animals from human impacts throughout their range. A concerted effort is required to further establish the population links between the various locations on the southern Cape coast that these animals frequent. / Dissertation (MSc)--University of Pretoria, 2014. / Zoology and Entomology / MSc / Unrestricted