Age-related variation in demographic rates has significant consequences for population and evolutionary dynamics, and understanding the processes driving such variation is therefore an important aspect of evolutionary ecology. Reproductive performance may vary over the lifetime of an individual, and this may be the result of both variations in reproductive effort and changes in individual competency. For example, increasing experience is likely to have beneficial effects on reproduction during early life, and senescence, or declines in physiological function, may have negative impacts on the performance of older individuals. The rate at which these changes occur can vary dramatically between species, and even between individuals of the same species. However, understanding the causes and consequences of this variation in the rate of ageing is not always straightforward. As well as the individual-level processes described, the phenotypic composition of successive age classes will contribute to age-related variation observed at the population level. Abrupt changes in performance, such as the poor performance of first time breeders, may be obscured if individuals vary in their age at first reproduction. Population-level patterns may also be influenced by selection; for example, the selective disappearance of low quality individuals from older age classes may mask senescent declines in the performance of longer-lived individuals. Moreover, the physiological mechanisms that underpin within-individual changes in performance are not well understood. Unravelling the drivers of such age-related variation requires longitudinal data, following individuals throughout their lives, which presents challenges for the study of natural populations. Albatrosses are among the longest lived vertebrates. In this thesis, I use data from three species of albatross breeding at Bird Island, South Georgia (54°00’S, 38°03’W) to explore age-related variation. Focusing primarily on the wandering albatross, Diomedea exulans, I characterise the relationship between age and various reproductive traits, and decompose the population-level patterns to reveal effects of experience, senescence and terminal effects across the reproductive lifespan of individuals. I then consider foraging behaviour as a proximate driver of changes in reproductive performance in this species. Using tracking data collected over a 20 year period, I find limited evidence for age-related variation in foraging trips taken throughout the breeding cycle. Going one step further, I explore telomere dynamics in the wandering albatross, examining the potential for telomere length to act as a physiological marker of individual state. Finally, I move on to a species comparison, incorporating data from the black-browed (Thalassarche melanophris) and grey-headed albatross (Thalassarche chrysostoma). I compare the population- and individual-level ageing patterns of these three closely related species, and consider these in light of their differing life history strategies.

Jusqu’à aujourd’hui, les traits démographiques du début de vie ont été peu étudiés relativement aux traits adultes chez une grande majorité d’espèces animales. Les jeunes individus sont généralement plus difficiles à étudier en milieu naturel du fait de leur taille plus petite, de leur comportement plus mobile et des forts taux de mortalité qu’ils subissent. Cependant, la bonne compréhension des processus démographiques requiert la prise en compte de l’ensemble des stades de vie. Le manque de connaissance concernant les plus jeunes individus est particulièrement préjudiciable pour les espèces longévives. En effet chez ces espèces, les individus immatures représentent une partie importante de la valeur reproductive totale et ont ainsi une forte influence sur l’ensemble de la dynamique de la population. Ainsi, ces lacunes affectent notre capacité à effectuer des projections démographiques réalistes dans un contexte de changements globaux et plus généralement limitent notre compréhension de l’évolution des traits d’histoire de vie. Afin de pallier à ce manque, cette thèse porte sur l’étude des traits d’histoire de vie en début de vie chez une espèce d’oiseau marin : l’albatros hurleur Diomedea exulans. Chez cette espèce particulièrement longévive, la période d’immaturité peut durer une dizaine d’années. En se basant sur un suivi individuel à long terme et des analyses de type capture-marquage-recapture, nous avons estimé la survie en début de vie ainsi que l’accès à la première reproduction en fonction de l’âge et du sexe. Nous avons également étudié les effets potentiels de facteurs extrinsèques (climat, pêcherie) et intrinsèques (densité dépendance, âge parental) sur ces paramètres. De nombreuses relations entre les taux vitaux de ces jeunes individus et leur environnement ont été décrites. Sur un plan plus théorique, nous avons également pu tester différentes prédictions comme la canalisation des traits d’histoire de vie ou l’existence de différentes stratégies d’histoire de vie au sein d’une même population. Au cours de ce travail, une attention particulière a été portée aux variations des traits démographiques à l’échelle individuelle. Notre étude montre que les taux vitaux d’individus de même âge et de même sexe peuvent être très variables. Nous avons pu associer les traits démographiques en début de vie aux performances futures au stade adulte et révéler certaines causes de ces variations interindividuelles. / To date, early life demographic traits have been poorly studied compare to adult traits in the vast majority of animal species. Young individuals are usually more difficult to study due to their small size, their mobile behavior and the high mortality rate that characterizes their life stage. However, the full understanding of the demographic processes requires the integration of all life stages. This lack of knowledge toward early-life stages is particularly detrimental for long lived species. Indeed for these species, the immature component represents a substantial part of the total reproductive value, having a high influence on the whole population dynamics. Hence, such limitations affect the accuracy of population projections in the context of global change and more generally our understanding of life history trait evolution. In order to fill in this gap, this PhD project focuses on early life demographic traits in a seabird species : the wandering albatross Diomedea exulans. In this very long-lived species, the immature period lasts around 10 years. Based on long-term individual monitoring and capture-recapture analyses, we estimated early-life survival and the access to reproduction according to age and sex. We also investigates the respective effects of extrinsic (e.g. climate, fisheries) and intrinsic factors (e.g. density dependence, parental effects) on these parameters. Many relationships between early-life vital rates and environmental variables have been described. From a theoretical perspective, we tested several predictions such as the canalization of life-history traits and the existence of different life-history strategies within a population. In this work, special attention was paid to demographic variations at the individual scale. Our study showed that vital rates of individuals of the same age and sex could be very different. We have linked early-life demographic traits with subsequent adult performances and identified some causes of these individual variations.