Given the worldwide diversity of duck production, genetics aspects of feed efficiency of ducks should be addressed depending on the production (broilers, layers, and “foie gras”) and on the genetic types (Muscovy duck, common duck, and their intergeneric cross mainly used in France for “foie gras” production after force-feeding, i.e. the mule duck). The two main criteria to characterize feed efficiency are the feed conversion ratio (FCR), computed as the ratio of food quantity divided by the output, and the residual feed intake (RFI) which is obtained through a multiple linear regression of feed intake by output and metabolic requirements. The former is a ratio, with undesirable statistic properties. In addition, it is uneasy to disentangle variations in net feed efficiency from variations in FCR due to production traits. The latter is supposed to be, at least phenotypically, independent from the constituent production traits. This is the reason why it gained popularity, even though it requires a thorough analysis of metabolic requirements, as in the overfed mule ducks, where the fat deposition capacity should not be impaired. In the literature, the values found for FCR depend on the genetic type and on the production: FCR≈2.5 at 12 wk. for male Muscovy and FCR >3.2 at 13 wk. for fatty mule ducks; FCR ≈ 1.9 at 42d for Pekin broilers; FCR ≈2.8 for layers. Usually RFI is moderately heritable (h²≈0.25 in layers; 0.3