It is generally accepted that leanness of terminal sire sheep breeds should be increased. However, it is very difficult to define formally a selection goal for these breeds. Put loosely, at a given age, we probably require terminal sires with as much extra lean meat as possible, whilst minimising any increase in fat weight. In the past, selection has been mainly on weight for age, but this may actually increase fatness at the age of selection.

Intuitively we expect some in vivo measurements of carcass composition to help in achieving the selection goal. In practice, in vivo measurements of carcass composition in sheep are often imprecise, though the treatment of the animals can affect the relative precision. Ad libitum feeding on a high energy, high protein feed appears to increase the variation in carcass composition of ram lambs compared to those on a lower plane of nutrition (Simm et al, 1985 compared to Cuthbertson et al, 1983). Under these conditions we were able to predict carcass composition reasonably precisely from live weight and ultrasonic measurements (means and original and residual s.d.s for carcass lean weight: 18.30, 1.60 and l.00 kg and for fat weight 13.74, 2.70 and 0.96 kg).