Heterozygosity–fitness correlations (HFC) were assessed for a sample of a gilthead sea bream Sparus aurata population. Two hundred and seventy-one fish were genotyped at 22 known and novel microsatellite loci, from which correlations between the multilocus heterozygosity index (IMLH) and various fitness traits (fork length, mass and specific growth rates) were calculated. Significant global HFCs were found in this sample (0·02 ≤r2≤ 0·08). In addition, all the significant correlations found in this work were negative, indicating that heterozygotes had lower fitness than their homozygote counterparts. Marker location could not explain the observed HFCs. Evidence of inbreeding, outbreeding or population and family structuring was not found in this work. The presence of undetected general effects that may lead to the appearance of HFCs, however, cannot be ruled out. These results seem to be best explained by the occurrence of local effects (due to linkage) or even by possible direct locus advantages.