Two strains of Saccharomyces cerevisiae were tested as probiotics for rainbow trout fry, during the first month of feeding. Each strain was introduced into separate diets, at the rate of 106 CFU g− 1 and their effects were compared with those of a control diet. Two rearing conditions were simultaneously compared, to test the adaptability of the probiotic treatment. From start feeding onwards, the water supply came from either spring or river, resulting in two different temperature ranges, 11-11.5 and 7-8 °C respectively. Growth and development were optimal in spring water, while some delay was observed with colder river water. A slight but significant increase in mortality was also observed in the river group. In all groups, the counts of bacteria associated with trout intestine were maximum 10 days post start feeding (dpsf; 107 CFU g− 1). The counts of probiotic yeast were also maximum at 10 dpsf (104-105 CFU g− 1), but the decrease was slower in river than in spring water. An autochthonous yeast, Debaryomyces hansenii, was also retrieved associated to the intestine of the control group in high numbers after 240 degree days of experiment (104-105 CFU g− 1), while the colonization level was significantly less in trout fed the probiotic diets. The effect of the dietary yeast was observed by assaying the activity of three enzymes in the brush border membrane of the enterocytes: alkaline phosphatase (AP), γ-glutamyl-transpeptidase (GGT), and leucine-aminopeptidase N (LAP). At 10 and 20 dpsf, the trout reared in spring water had higher activities of the three enzymes when they were fed the strain S. cerevisiae var. boulardii, suggesting an earlier maturation of the digestive system in this group, compared with trout fed either the other strain of S. cerevisiae or the control diet. The effect was not observed in trout reared in river water with slower growth. Both S. boulardii and D. hansenii seemed to stimulate digestive maturation in fish, but the natural colonization by D. hansenii was likely too late for trout reared at optimal temperature. The supplementation of trout starter diet with S. boulardii may be particularly useful in fast growing conditions.

Two probiotics were tested as alternative treatment to limit the prevalence of the vertebral column compression syndrome (VCCS) in rainbow trout, compared with a preventive treatment with florfenicol. Either the antibiotic, or a lactic acid bacterium, Pediococcus acidilactici, or a yeast, Saccharomyces cerevisiae var. boulardii, was introduced into experimental diets, which were compared with the control diet without supplementation. The antibiotic caused some mortality during the first week after treatment, likely because of the toxicity of the compound. However, this antibiotic treatment limited the occurrence of VCCS to 3%, vs. 13% of fish affected in the control group. Pediococcus acidilactici provided the same level of protection as the antibiotic, but on condition that the treatment was applied during the 5 months of the experiment, without any adverse effect on survival. A treatment with P. acidilactici only during the first 20 days of feeding was not sufficient to limit VCCS, nor was the treatment with the yeast sufficient. These results were discussed while comparing the microbiota associated with the intestine. The experiment reinforced the hypothesis that pathogenic bacteria are involved in VCCS. The long‐term dietary supplementation with P. acidilactici seemed promising as a preventive treatment against the syndrome, but large‐scale investigation in fish farms will be necessary to confirm its prophylactic relevance.

The effect of stocking density on feed intake, feed utilization and feeding behaviour was studied in groups of rainbow trout submitted to different levels of food accessibility. In the first experiment, 18 groups of 100, 300 or 500 fish (initial individual weight 26 + 0.5 g) were held in 1 m 3 tanks where they were fed: in excess using belt feeders; or by means of self-feeders with a fixed reward level (4 g per trigger actuation); or a reward level proportional to fish density (2 g/100 fish/trigger actuation). Final biomass after 125 days was c. 25, 70 and 100 kg/m 3 in groups of 100, 300 and 500 fish respectively. Growth and feed intake was affected by fish density, feeding technique and reward level. However, feed efficiency, body composition, as well as nutrient retention and loss were unaffected by the treatment. In the self-fed groups, daily pattern of feeding activity was affected by fish density when the reward level was fixed but not when the reward level was proportional to fish density. In a second experiment, groups of 100 or 500 fish (same initial individual weight) had access to one or three self-feeders. Access to multiple self-feeders had no positive effect on any of the studied parameters. The fish used only one of the three feeders at a time, and the three feeders where not used equally. We conclude that feed intake but not feed utilization was impaired by increased density, indicating that density in itself is not stressful to the fish. Food accessibility is the main factor responsible for the observed density ± related decrease in growth.

An 84-day experiment assessed the combined effects of two fresh water quality levels (H: 1.71 ± 0.15 mg O 2 l −1 and 0.28 ± 0.01 mg l −1 total ammonia nitrogen (T-AN), L: 5.15 ± 0.07 mg O 2 l −1 and 0.54 ± 0.01 mg T-AN l −1 ) and 3 stabilized stocking densities: 24.8 ± 0.2, 74.2 ± 0.5 and 120.0 ± 0.9 kg m −3 ) on rainbow trout. Fish were fed using demand feeders with rewards proportional to stocking density. Mass increase was significantly affected by water quality and stocking density, being highest in H water and the lowest at 120 kg m −3 . There was no significant difference in final weight between 25 and 74 kg m −3 , but at 120 kg m −3 it was 27% and 19% lower in H and L water respectively than at 25 kg m −3 . Feed intake (FI) from day 0-85 was significantly affected by water quality, 1.5% in H compared to 1.1-1.0 in L, but there were no significant differences in apparent feed conversion (AFC). FI was not significantly affected by stocking density but AFC was impaired, it increased with stocking density. Marked changes in fish morphology and composition were related to water quality: Condition K factor and fillet fat content were significantly higher in H than in L groups. Dorsal and pectoral fin condition was affected by stocking density and water quality: fins were significantly longer and less eroded in L groups and at low stocking density. Physiological measures were within the usual ranges, but differed between treatments. Changes in plasma osmolarity, hydromineral balance (Na + ) and acid base balance (HCO −3 ) showed that fish were more affected by water quality than by stocking density. There was no sign of acute stress in acclimated fish as cortisol and glycemia were similar under all experimental conditions. This study highlights the importance of water quality and feeding conditions when considering the effects of stocking density on fish welfare.

No abstract