The catch (landings and discards) at age of roundnose grenadier (Coryphaenoides rupestris) were used to assess the stock in the areas exploited by French trawlers. A pseudo-cohort analysis, corrected for changes in fishing effort in the period before the year which catch at age are analysed, was carried out for the years 1996 and 1997. Age distributions were derived from length measurements and an age length key (ALK) estimated from 1529 otoliths. Although, unvalidated in the size range of adult fish, this ALK looked consistent with validated length at age of juveniles and former age estimates of adult fish. It was also combined with length distributions data from surveys prior to exploitation and from landings in 1990 (early in the history of the fishery) to estimate the total mortality then equivalent to the natural mortality. This poorly converged pseudo-cohort suggested a low annual fishing mortality. However, the behaviour of the stock was further investigated from yield per recruit (Y/R) analysis. Because of a series of gaps in the knowledge of this species (stock(s) distribution, unvalidated ages, …), the estimates of fishing mortality must be considered with caution. The main interest of the analysis is probably to stress the impact of the discards on the potential yield of roundnose grenadier.