Eight research and teaching centres have co-operated in an experiment involving 132 sows. Six treatments were used, combining three levels of feeding in pregnancy with two in lactation. Pregnancy treatments were High 3−2 kg meal/day, Medium 2·4 kg/day and Low 1·6 kg/day; in lactation either a high level, 4·1 kg plus 0·2 kg/pig (h) or a medium level, 2·3 kg plus 0·2 kg/pig (m) were given. The sows were maintained on the same nutritional regime for three parities although inevitably some sows were lost before the completion of the trial. Sow live-weight gains in pregnancy were directly related to feed intake; in lactation, losses in sow live weight were dependent on gains made in the previous pregnancy and on the level fed while suckling. After the first parity, in which animals fed the low level in pregnancy tended to produce larger litters (though the difference was not significant), there were no differences in numbers born attributable to treatments.

Increases in the feed intake of sows in pregnancy resulted in consistent and significant increases in the weight of the pigs at birth. This increase in birth weight also resulted in an increase in weight at weaning in the second and third parities. Level of feeding in lactation, which had a highly significant effect on sow live weight, had much less influence on weaning weights of pigs. There were no consistent pregnancy × lactation interactions, the effects of treatments in the two phases being additive.

An assessment is made of the relative efficiency of the six treatments and the value of the experiment is discussed in relation to other studies concerned with sow nutrition.