Selective breeding to improve residual feed intake (RFI) in sea bass (Dicentrarchus labrax) is a major goal that would optimize economic gain while minimizing the environmental impact of production. Due to the difficulty in accurately measuring individual feed intake, no selective breeding program has yet been started. In a previous study, we identified a criterion phenotypically related to RFI variations: the loss of weight during feed deprivation, FD. Moreover, an additional composite criterion (CC) integrating both FD and weight gain during subsequent re-feeding (RF) was closely related to RFI, even though the relationship was only close to significance (P=0.06). The aim of the present study was to estimate the heritability of these two traits, in order to complete the analysis of their pertinence as indirect criteria for use in a selective breeding program to improve RFI. We set up a full factorial mating design combining 41 sires and eight dams to produce mixed families offspring, which were all raised in the same tank. At 50g. 1300 individuals were individually tagged and a fin clip was sampled for DNA extraction and parentage reassignment using five to six microsatellite loci, which showed they originated from 261 families. The evolution of individual body weight was recorded during a growth period of three weeks, followed by successive periods of three weeks feed deprivation and three weeks ad libitum re-feeding, repeated twice. Phenotypic and genetic correlations between the two feed deprivation periods or the two re-feeding periods were moderate (r = 0.45-0.51, and 0.71-0.73 respectively for phenotypic and genetic correlations) indicating that FD and RF are relatively repeatable. FD, RF and CC heritabilities were of the same magnitude (0.23 +/- 0.04, 0.19 +/- 0.04 and 0.22 +/- 0.04, respectively), and sufficient to consider the use of such criteria in a future breeding program.