The effects of rearing density and size grading on sex ratios in 30 families of sea bass reared in the same tank from the fertilization stage onwards were investigated. Two extreme density treatments (high versus low) were applied in quadruplicates throughout the labile period for sex determination. An excess of males (88%) was found in all groups. Growth was faster under low density from 49 to 191 days post fertilization (p.f.) (27-10% relative difference in length) and was then equivalent between the two groups until the end of the treatments (414 days p.f.). Density had no effect on sex ratio, suggesting that the high densities usually applied in aquaculture are not involved in the systematic excess of males reported in farmed populations. The fish were genotyped at three to six microsatellite loci. This allowed them to be assigned to the individual breeders used in the mating design. The percentage of females was significantly influenced by both the dam and the sire, highlighting the existence of genetic variation of sex ratio. Repeated size grading performed from 84 to 199 days p.f. had no effect on the sex ratio of the overall population studied showing that sex determination had not been affected by the treatment. However, the highest percentages of females were found in the lead classes from 84 days p.f. Most of the fish in this lead group were descended from one dam and the percentage of females found in the individual families were higher in this class than in the non-graded group. These results show that females start growing faster than males long before morphological sexual differentiation occurs.