Forty pigs (half each of females and entire males) were used in an ad libitum feeding experiment from weaning to 33 kg live weight (LW). In period 1 to 16 kg LW, the pigs had either food L (134 g crude protein (CP) per kg) or food H (278 g CP per kg) alone. The aim was to delay the growth of pigs on L and create two groups of pigs with different body compositions. At 16 kg LW, eight pigs were killed and their empty bodies chemically analysed. The pigs given L had 1·06 (s.e. 0·08) kg more lipid and a lower protein: ash ratio in their empty body than pigs on H; growth rates, from 9 to 16 kg LW, were 386 (L) and 591 (H) (s.e.d. 17)g/day. Both aims were thus achieved.
In period 2 the pigs previously given L (no. = 16) or H (no. = 16) in period 1, were given L alone (no. = 4), H alone (no. = 4) or a choice between L and H (no. = 8) to 33 kg LW when the choice-fed pigs were killed and their empty bodies chemically analysed. The objective of the experiment was to test the idea that pigs previously given a food deficient in CP, when given a choice between foods, one of high and one of low, protein content will select a diet of a composition that allows them to correct the effects of the previous mis-feeding. The pigs given L alone in period 2 grew more slowly, and ate more per day, than those given H alone, with no effect of the food given in period 1. Of those given H alone in period 2, the pigs previously given L grew faster, and were more efficient, than those previously given H.
The protein contents (g CP per kg food) of the diets selected by the choice-fed pigs were significantly affected by both sex (228 for males v. 181 for females; s.e.d. 15) and the food given in period 1 (233 for L and 175 for H; s.e.d. 15) with no significant interaction between these factors. The pigs from L grew faster (1039 v. 780 (s.e.d. 70) g/day) had a higher daily food intake (1420 v. 1319 (s.e.d. 71) g/day) and were more efficient (0·736 v. 0·595 (s.e.d. 0·034) g gain per g food) than those from H. At 33 kg LW lipid weights were similar for the two period 1 treatments (L had 0·40 (s.e. 0·54) kg more lipid than H); females had 1·79 (s.e. 0·54) kg more lipid than males.
It is concluded that, where pigs are given a choice between foods of low and high protein contents, they will select a diet of a composition such that the effects of previous mis-feeding are corrected. Giving pigs a choice between an appropriate pair of foods allows them to select a diet of a composition suitable for their needs.