Commercial and scientific bottom longline catches of alfonsino, #Beryx splendens$, from seamounts of New Caledonia were sampled to study length-frequency distributions. A total of 14,674 fish were measured. CPUE of #Beryx splendens$ on two seamounts is modelled in terms of length and depth. The data show that mean length increases with depth ; this is well described by a bivariate normal model that estimates catch for a given seamount. In addition, the data show that mean length also varies with the depth of the top of seamounts ; this is described by a recursive model that is designed to predict approximate catch for any seamount. The limitations of both models are discussed, particularly with regard to temporal variation. (Résumé d'auteur)
