Résumé: Catch-at-size data of the eastern Atlantic skipjack were used to estimate changes in total mortality from 1969 to 2007. We used a transitional model of mean length that generalized the Beverton-Holt mortality estimator to allow change in mortality rate under nonequilibrium conditions. Then, from homogeneous periods of time, length-converted catch curves were used to access qualitative changes in selectivity patterns for two surface fisheries (the baitboat fishery operating from Dakar, Senegal, and the European and associated purse seine fisheries). To explore the impact of catch on the mortality rate, a Bayesian change-point analysis was conducted on the catch time series to detect concomitant variation between mortality rates and catch. Finally, potential causes of these changes are discussed in relation to the implementation of new fishing technology, such fish aggregating devices (FAD). The general pattern depicted by total mortality is in agreement with previous knowledge on this fishery: a state of complete or practically complete exploitation during the nineties followed by the decrease in nominal purse seine fishing effort in the last decade, combined with the effect of a seasonal moratorium on FAD fishing operations. There was no evidence of a change in selectivity between the two contrasted periods of time considered: 1969-1979 and 1986-1999; the second period being characterized by the introduction of new technology onboard vessels. In contrast, the covariation over time between total catch and mortality rate three years later highlights the effect of the fishing pressure on the stock of eastern Atlantic skipjack.