ABSTRACT1. The primary growth and first regrowth of perennial ryegrass were preserved either without additive (NA), or with a mixture consisting of equal volumes of formic acid and formalin at 10·0 and 13·01/t fresh crop for primary growth and regrowth herbage respectively (35 g formaldehyde per kg crude protein) (FF). The silages were given ad libitum to 40 lactating British Friesian heifers and cows, with a supplement of either pelleted ground maize (104 g crude protein per kg dry matter) (LP) or pelleted ground maize and soya bean meal (254 g crude protein per kg dry matter) (HP), at 6·1 and 7·1 kg per head per day for heifers and cows respectively. The primary growth silage was given from weeks 4 to 14 and the regrowth from weeks 15 to 22 of lactation. The silages were supplemented with urea so that the dietary supply of rumen degradable protein exceeded 7·8g/MJ metabolizable energy.2. Silage FF had lower concentrations of both fermentation acids and ammonia-nitrogen than silage NA. The digestibility of the dry matter, energy and crude protein of the diet containing silage FF was significantly less than that of diets containing silage NA when cows were given primary growth material (P < 0·001). But silage treatment had no significant effect on the digestibility of cellulose. When cows were given regrowth silage, silage treatment had no significant effect on the digestibility of dry matter, energy and cellulose, but formic acid/formaldehyde reduced the digestibility of crude protein (P < 0·01). The inclusion of soya bean meal in the supplement significantly increased the digestibility of crude protein, dry matter and energy, but when cows were given the primary growth silage the effect on the digestibility of dry matter and energy was only apparent with silage NA.3. Cows and heifers given silage FF from primary growth consumed 160g/kg more dry matter than those given silage NA (P < 0·01). This effect was not apparent when the animals were given regrowth silage. The effect of formic acid/formaldehyde on the intake of digestible energy was not significant in either period on the trial, but the inclusion of soya bean meal in the pellet significantly increased the intake of digestible energy (P < 0·01).4. Cows and heifers given silage FF from primary growth produced 90g/kg more milk and 50g/kg more solids-corrected milk than those given silage NA, but the differences were not significant and these trends were not apparent when regrowth silage was given. An increase in the crude protein content of supplement increased the milk (P < 0·01) and protein yield (P < 0·05) of cows, but not heifers, when they were given primary growth silage. The effect was less when milk output was expressed in terms of solids-corrected milk since milk fat content tended to fall when cows were given the HP supplement. Treatments had no significant effect on live-weight change.5. It is suggested that the increases in milk output were mediated via changes in the supply of energy rather than of protein.