Résumé: The otolith strontium:calcium ratio (Sr:Ca) has been widely used to assess the connectivity between fish populations in ocean, estuarine and freshwater environments as the concentration of Sr in the otoliths is strongly correlated with water salinity. This correlation was tested experimentally in hypersaline conditions by submitting the extremely euryhaline tilapia species Sarotherodon melanotheron heudelotii (Cichlidae), found throughout West African continental waters and commonly used as an aquaculture species, to a strong salinity gradient (15–106). Experimental and control individuals were reared from birth in a closed system at a constant salinity of 10 and injected with oxytetracycline (OTC) to mark the otoliths at the beginning of the experiment. Randomly selected control fish were maintained for 75 days at salinities of 10–20. The remaining experimental fish were acclimated to a salinity above 100 which was reduced by 10 each week to a salinity of 20. The salinity and temperature of the water were recorded every day and the Sr concentrations in the water were measured weekly by solution-based ICP-MS. The fish from the control and experimental groups were sampled weekly and otolith transverse sections were prepared for Sr:Ca measurements by laser ablation ICP-MS. No significant difference in the otolith growth rates after OTC marking was found between the control and experimental groups (ANCOVA, p = 0.63), showing that the experimental design did not affect fish growth. The Sr concentration in the water was closely related to ambient salinity (positive linear regression, R2 = 0.96). For most of the fish tested (~ 80%), the relationship between otolith Sr:Ca and salinity was positive but nonlinear (power law, R2 = 0.77 on log–log plot). However, about 20% of individuals from both the control and the experimental groups showed consistently low Sr:Ca ratios irrespective of the salinity, suggesting that the Sr incorporation into the otoliths in these fish was strongly regulated. This shows that there is high variability between individuals in the regulation of Sr incorporation by a euryhaline species and indicates that otolith datasets for ecological applications should be interpreted with caution.