Abstract
1. Two‐way selection was performed in a White Leghorn population for percentage residual food consumption, defined as the percentage deviation of the observed from the expected food consumption. Expected food consumption was estimated using a multiple regression equation including weight gain, egg production and metabolic body size as independent variables. 2. Three selection lines were established from a common population, two low lines and one high line. In addition, one control line was kept. The generation interval was one year and the experiment lasted for 4 generations. 3. The direct selection response was considerable and the realised heritability was close to 0.5. Food intake and food conversion (FCR) were lower in the low lines than in the high line. 4. The correlated selection responses showed no phenotypic or genetic correlations between percentage residual food consumption and egg production, but one of the low lines had a heavier adult body weight than the high line. No undesirable correlated responses were observed.