The interaction between the genotype and feeding regimen (GxE) for carcass traits was estimated from data corresponding to 4981 animals under ad libitum (AD; 2557 individuals) or restricted (RF; 2424) feeding. Body weight at slaughter (BWS), carcass weight (CW) and dressing out percentage (DoP) were analyzed by using Bayesian multivariate Linear Animal Models in which records obtained under different feeding regimes were treated as different traits. Animals belonged to a paternal line (Caldes line) selected for average daily gain (ADG) under ad libitum feeding. Therefore, the complete information of the selection process involving data from 134,419 animals was included in the analysis. Kits from the same litter were equally distributed into both treatments ensuring that animals in a cage had a similar initial weight. Model for ADG included the effects of year-season, parity order, number of born alive, and the additive genetic and common litter effects. For carcass traits, the model included parity order, number of kits born alive, batch, size of the animal at weaning, number of animals per cage, age at slaughter, and the additive genetic, common litter and cage effects. Marginal posterior mean (sd) for the heritabilities were 0.16(0.01) for ADG, and 0.34(0.07), 0.28(0.05) and 0.16(0.03) for BWS, CW and DoP under AD feeding, the respective figures under RF were 0.18(0.03), 0.15(0.03), and 0.32 (0.06). Genetic correlations between ADG and carcass traits were positive and moderate except for DoP which was low and negative. The estimated genetic correlation between BWS, CW and DoP under different feeding regimen were different from 1: 0.90(0.06), 0.78(0.09) and 0.93(0.02), respectively; which, jointly with the differences in genetic variances, indicates the existence of relevant GxE interaction for these traits.