Twenty-two unmated female Ayrshire twin cattle, that had initially been maintained for prolonged periods on one of six constant feeding levels until an equilibrium weight was attained, were subsequently moved up to higher constant feeding levels including ad libitum. In all, results were obtained for 44 equilibrium periods mostly of 96 weeks duration.

For controlled feeding levels, the log-log,. regression of equilibrium body weight on food intake, within animals, was 0·999 (s.e. 0·045). For all results, including mature equilibria, the within-animal regression was 1–014. There was thus no systematic change in an individual's equilibrium maintenance requirement per kg body weight in the range from 25%, to 100%, mature. Efficiency of food utilization for equilibrium maintenance was found to be independent of age also, except for a small increase at advanced ages beyond 8 to 9 years.

There were significant differences between animals in equilibrium maintenance efficiency, the genetic coefficient of variation being 6·4%. The most striking result, however, was a within-animal repeatability of 0·7, which meant that almost the same efficiency was re-attained when, after a prolonged period of many years on one constant feeding level, the same animal was allowed to re-establish a new equilibrium on a higher level.

Following prolonged periods of food restriction, the animals showed remarkable capacity to recover in body weight even at very advanced ages, but nevertheless substantial stunting of mature weight did occur.