1. Juvenile fish were sampled by electrofishing using the point abundance sampling method from August to November 1991 in the Lower Rhône River. Variations in fish abundance between point samples were partitioned according to three spatial scales of habitat heterogeneity. 2. The compartment scale, differentiating channel and ‘dike fields’ (an area marginal to the main channel partially enclosed by an artificial embankment), was associated with variation in water current, temperature and biological productivity. Every juvenile fish species studied was more abundant in the dike fields, using them as a nursery. 3. Within both compartments (i.e. main channel and dike fields), the same five bank types were represented, defining the intermediate scale of habitat heterogeneity. In the dike fields, beaches were selected by most of the species, whereas steep silt banks were more often selected in the main channel. 4. After having removed the variation explained at the two larger habitat scales, only depth and shelter were found to have a significant effect on fish abundance at the microhabitat scale. 5. When compared with a classical single‐scale approach, the multi‐scale approach was found to explain a greater amount of fish abundance variations and to facilitate the interpretation of observed patterns by scaling the species’ responses to habitat heterogeneity. The difference between the amounts of variation explained was proposed as a measure of how much habitat selection is hierarchical.

Optimal life-history models generally predict that the reproductive e¡ort of iteroparous organisms may increase with age, as their expectation of future reproduction decreases. The population of three-spined sticklebacks (Gasterosteus aculeatus) in the Camargue (Rhoª ne River Delta, France) is annual, all adults dying after their ¢rst breeding season. As the three-spined stickleback is a multiple spawner, we tested the hypothesis that reproductive e¡ort may increase during the breeding season on ¢eld data. From 1987 to 1998, 653 female sticklebacks were collected in the ¢eld during the breeding seasons. The body size, body weight and weights of the liver, gonads and carcass were measured for these individuals. Only gravid females with mature eggs (176 ¢sh) were included in the analysis. Considering the female three-spined stickleback as a capital breeder, the energetic resources available for allocation between soma and gonads were estimated by its body weight. Somatic condition decreased during the breeding season and reproductive e¡ort (gonad weight relative to body weight) increased. These patterns did not vary signi¢cantly between years. These observed variations in reproductive e¡ort during the breeding season can be interpreted as empirical evidence of a trade-o¡ between reproductive e¡ort and expectation of future reproduction.

No abstract

No abstract

In the Camargue (southern France), movements of fish between a canal and two seasonally flooded marshes were monitored continuously by fish traps for 3 years for one marsh and for 2 years for the other. The timing of the entry and exit of the main species was determined together with species annual demographic balance, defined as the difference between the number of fish leaving and the number of fish entering. Only some of the species inhabiting the canal colonized the marshes. Except those species that reproduced in salt or brackish waters, all the species colonizing the marshes reproduced in it. The most abundant of these species were small sized (mosquitofish, sand smelt and three-spined stickleback). The unpredictability of water levels in summer was particularly unfavourable for recruitment and survival of species that breed late in the year, e.g. pumpkinseed sunfish, or which prolonged their stay in the marsh and only attempted to leave just before the connection was broken, e.g. carp. Only the three-spined stickleback always had a positive demographic balance as a result of colonizing the marshes. Adults of this species entered in winter and young-of-the-year left in April. By limiting its stay in the seasonally flooded marshes, the stickleback minimized the risks related to environmental unpredictability. These results suggest that hydrology (temporal variations of water depth) may influence the fish community structure through interspecific differences in survival and recruitment, as a result of temporal variations in the use of seasonally flooded marshes.

Variations in the liver and spleen masses of the eel Anguilla anguilla were analysed in relation to the parasite load of Anguillicola crassus at autopsy (current infection by swimbladder lumen worms) and in relation to the severity of damage observed in the swimbladder (a way of assessing the intensity of past infections). None of these measures of parasite pressure were shown to account for variation in the relative liver mass, either when controlling for somatic mass or eel age. In marked contrast, a significant increase in spleen size was revealed in eels harbouring many lumen worms and also in eels with severe damage in the swimbladder. Splenic enlargement was nearly two-fold higher among severely affected eels (harbouring more than seven lumen parasites and showing severe damage in the swimbladder) than among infectionfree eels (no lumen parasites and no pathological signs in the swimbladder). Several possible hypotheses are reviewed before arguing for an adaptive host response involving the haematological and immunological functions of the spleen. Indeed, among eels with no pathological signs in the swimbladder, the relative spleen mass was positively associated with the mass of lumen parasites, which suggests a hyper-synthesis of blood cells by the spleen in response to the bloodsucking activity of lumen worms. Nevertheless, among eels with no lumen parasites at autopsy, there was still an increase in spleen size in relation to the severity of the swimbladder damage, which also suggests a hyper-synthesis of splenic immune cells (lymphocytes and macrophages) in reaction to damaged tissues and particularly to larvae in the swimbladder wall. # 2004 The Fisheries Society of the British Isles

No abstract

Although the colonisation of coastal rivers on the Atlantic and Mediterranean coast by glass eels, Anguilla anguilla, has been well studied and understood, the colonisation of lagoons by glass eels is much less known. For the first time in the Mediterranean region, the installation of a glass eel fish-pass in Grau de la Fourcade channels in the Rhône delta enabled us to determine which factors could explain the variations in the catches of glass eel entering the Vaccarès coastal lagoon system. Whatever be the procedure chosen, the results of the model were the same: the temperature, the cumulative water discharge from the channel in the 5 nights before the catch (freshwater lure) and time that the drainage pumps were working explained the glass eel catches in the fish-pass in the Grau de la Fourcade. The tide and the cumulative discharge from the channel for only 3 nights before the catch did not seem to have a significant role in explaining catches. These results show that it is important that the lagoons should continue to receive rainfall runoff from their watersheds so that their water levels are high in winter, and that there is a good colonisation by glass eels as a result of a freshwater lure effect, when strong north winds expel low salinity water to the sea.