An analysis of statistical data available in eight glass eel fisheries distnbuted over the French Atlantic coastline is used to define the upstream migration dynamics of Anguilla anguilla in estuaries. The CPUE (Catch Per Unit of Effort) data reveal three different types of seasonal trends that reflect different migration kinetics. The recruitment abundance is studied compared to the salinity of the estuary. The first two types of trends correspond to years of high recruitment andor of reduced fishing effort. In brackish estuarine zones, the abundances then follow a symmetrical curve, with a peak in JanuaryFebruary. This time course is due to the passage of a wave of migrants. Further upstream, in tidal freshwater zones, the abundances follow an asymmetrical curve which may describe the progressive accumulation of eels between November and March/April. When recruitment is low andlor the fishing effort is high, CPUEs vary during the season following a plateau curve, with values staying more or less constant between November and March. This latter trend represents a balance status between immigration and capture. In this case, abundance indicescan only be established with total catches. The curves are explained by the migration behaviour of glass eel which exhibit a resting phase in brackish zones and an accumulation in tidal oligohaline zones. These results have direct consequences particularly for the determination of recruitment abundance indices.

By-catch of 0-group flounder, Platichthys flesus L. 1758, in the Loire glass-eel fishery was estimated for 1987 and 1988. Samples were taken throughout the estuary but the by-catch mainly occurred in the Central area. This pattern was observed in both surveys despite different hydrological conditions. The overall by-catch was 41 million animals for 1987 and 7 million for 1988. This difference was not only a consequence of a lowered fishing effort, it also suggests a difference in year-class strength. Recause of the great variability of postlarval distribution the accuracy of the estimates was very poor and only orders of magnitude can be discussed.