1. Genetically lean (LL) or fat (FL) chickens were fed from 28 to 42 d of age on one of 6 diets with different protein contents (from 73 to 208 g/kg). In order to keep a constant amino acid balance the experimental diets were made by diluting a well-balanced protein-rich diet with a protein-free diet. 2. Dietary protein influenced the growth rate of both genotypes similarly. However, maximum weight gain was reached in LL at a lower protein intake than in FL. 3. Regression between total protein gain (body protein + feather protein) or body protein gain and protein intake exhibited significantly different slopes, that of LL being superior to that of FL. 4. At a given protein intake, feather protein gain was also superior in LL to FL. Moreover feather protein, as a percentage of total protein gain, was superior in LL to FL. When the dietary protein fell below 126 g/kg, feather protein represented a higher proportion of total protein gain. 5. Multiple linear regressions of protein intake (as the dependent variable), and body weight and protein gain or weight gain (as the independent variables) suggest that the maintenance requirement for protein is similar in both lines but that the protein efficiency for growth is significantly superior in LL. 6. In a second experiment both genotypes were offered either a single high protein diet (232 g/kg) or a single medium protein diet (186 g/kg) or had free-choice between a high (269 g/kg) and a low protein (145 g/kg) diet. In free-choice feeding, FL chickens selected an overall dietary protein content which was significantly lower (179 v. 200 g/kg) to that of LL. In both genotypes, free-choice feeding led to fatter and less efficient chickens than predicted by the linear regression between adiposity or food conversion and protein content.