The effects of chronic confinement stress (1, 5 and 10 days) and of periodic blood sampling on somatic growth and the structure and growth of otoliths was studied in Oreochromis niloticus. During the study, the plasma concentrations of cortisol were measured at various times during the application of stress: they were significantly higher in confined fish than in control fish (mean±s.n. 3.40±0.47 v. 1.26±0.62ngml-1, P<0.05) up to 5 days after the start of a 10 day stress period. The somatic growth (standard length, Ls, and mass) was affected by the confinement and by the sampling (from 1621 ± 1.07 to 14.64 ± 1.15 cm for Ls, and from 173.31 ± 33.14 to 110.50±29.48g for mass). But the confinement masked the effect of the sampling on somatic growth. Tetracycline was injected at the start of the experiment to mark the otoliths, and showed that the short and long duration confinements led to a clear check in the pattern of primary increments in the otoliths. The number of primary increments deposited during the resting periods that followed each period of confinement was always less than the number of days that these periods lasted. No relation was found between the duration of confinement and the structure of the resulting checks. These results suggest that there is a disruption in the laying down of primary increments during periods of confinement resulting in an underestimation of their number compared to the actual number of days of growth. These results call into question the use of otolith primary increments as a means of estimating the age of Nile tilapia that have experienced periods of stress.