Résumé
Bird ringing was mainly used originally to study bird migration. However, it was realised as early as 1928 that encounters (recoveries) of dead ringed birds contain information on survival-mortality processes. Lack was the first to investigate survival processes in bird populations based on ringed individuals, with attention to age dependence in survival. Starting from early historical references I review survival studies in bird populations, along three lines of development. The ups and downs of analyses of recoveries: the ups concern the development of models making clear the difference between a table of recoveries and the classical life table of demographers; the downs concern in particular problems of identifiability in age-dependent models. The development of models and analyses of live recaptures, in particular in relation to long-term studies in bird population biology. Reflections on the unique characteristics of bird ringing. One outstanding characteristic is the possibility of ringing birds at birth and in turn of estimating immature survival, which encourages the development of joint analyses of recoveries and recaptures to overcome the limitations of analyses of recoveries or recaptures in isolation. Examples of survival analyses in gulls Larus spp. and Great Cormorants Phalacrocorax carbo are used as illustrations. Strategies for ringing based on these reflections are discussed.