The bacterial environment should be taken into account for assessing the quality of fish larvae, besides biochemical and developmental criteria. Detrimental effects of associated flora have been observed on the survival rates of larvae (e.g. Nicolas et al., 1989), and experimental infections were performed on fish larvae by introducing the pathogen either directly into the rearing medium or into live food organisms (e.g. Kusuda et ai, 1986). A challenge test was used for assessing the effect of a food additive on the ability of turbot (Scophthalmus maximus) larvae to resist infection, but that required rearing the larvae separately (Gatesoupe, 1993). Bergh et al. (1992) have proposed an infection test with transfer of eggs into polystyrene dishes, but it was not convenient for larvae with exogenous food source. The present paper was an attempt to develop a new test with transfer of larvae after feeding. Then this test was inserted into a simple survey of the quality of turbot larvae, with a view to drawing up relations between rearing success, associated bacteria and resistance against pathogens. For this purpose, seven batches of larvae were successively reared in 150 1 cylindrical tanks with conical bottoms from day 1 after hatching until day 10. Three tanks were used for batches 1, 3, 5 and 6, while five tanks were used for batches 4 and 7. The larvae of batch 2 were reared in one tank. The rearing temperature was gradually increased from 14 °C to 18 °C by day 5. The salinity was 35%. For feeding the larvae, rotifers (Brachionus plicatilis) were cultured in seawater according to the method already described (Gatesoupe, 1990). The rotifer diet was made up of baker's years, menhaden (Breuoortia tyrannus) oi l (Sigma F8020), soybean lecithin, DL-a-tocopherol (USB), and vitamin premix (Spyridakis et ai, 1988), at the daily rates of 60.8, 7.7, 0.7, 0.015, and 0.7 mg r1, respectively, on a dry matter basis. The larvae were fed with the rotifers from day 3 to day 9, at the daily rates of 20, 50, 75, 100, 150, 200, and 250 rotifers per initial larva, corresponding to the seven consecutive days of feeding. Each day, the rotifers were transferred from their culture medium into clean seawater tanks (200 rotifers m l 1 ) , where they were enriched with spray-dried fish autolysate, menhaden oi l , soybean lecithin, and DL-a-tocopherol, at the rates of 15, 7.7, 0.7, and 0.015 mg T 1 , respectively, on a dry matter basis. The rotifers were continuously distributed to turbot with a peristaltic pump, the flow rate of which was adjusted so that the whole volume was carried through in 24 h. At day 10, the survivors were counted and 30 fish were sampled in each tank,