1. Gestation diets formulated to contain A, 9 or B, 11 % protein, and lactation diets with C, 13; D, 15; E, 17 and F, 19 % protein provided eight treatment combinations. They were given at 2 kg/day in gestation and approximately 5·7 kg/day in lactation to sows at eight centres, which remained on experiment for 3 to 5 parities and produced in total 529 litters. Mean results for parities 1 to 3 are given in this summary.
2. Sows given more protein during gestation (i.e. diet A v. B) gained more weight in gestation (18·3 v. 22·5 kg net gain) but lost more in lactation (—1·7 v. —5·2 kg). Their offspring were significantly heavier at birth (1·31 v. 1·36 kg) and at 3 weeks old (5·28 v. 5·49 kg), but not at 6 weeks old (weaning).
3. Giving more protein in lactation caused linear increases in piglet weights at 3 and 6 weeks, and a linear reduction in creep feed consumption. From a sample of sows whose milk yields were recorded, extra protein appeared to increase milk yield and its protein content. Extra protein for lactating sows also reduced their live-weight loss in lactation, but this effect was partially counterbalanced by smaller gains in gestation.
4. No treatment had consistent significant effects on litter size or on the interval from weaning to conception. There were no significant interactions between gestation and lactation protein concentrations.
5. The generally large differences between centres were in a few cases associated with centre × treatment interactions, but these were not such as to cast doubt on the general applicability of the main findings of the experiment.
6. From this and an earlier experiment the minimum protein level for diets based on barley was concluded to be 11 % during gestation and 15% during lactation. For lactation diets, however, the economic optimum level may be > 15 %.