AbstractThree silages were produced from the same swards by harvesting grass using either a double‐chop harvester (DC) or a precision‐chop harvester adjusted to produce a long staple length (PL) or a short staple length (PS). The mean particle lengths were 47, 52 and 14 mm respectively.Of each material 250 t was ensiled unwilted and with formic acid additive at the mean rate of 2‐4 litres t‐1of grass. All three silages were well preserved and degree of chopping had no effect on fermentation as indicated by either pH, ammonia N or organic acid concentration.The silages were offeredad libitumto seventy‐two British Friesian cows with a mean calving date of 28 January which were in their second or later lactation, in a randomized‐block experiment from day 8 of lactation until 12 April. In addition twelve animals on each silage received a uniform daily concentrate allowance of 7·2 kg while twelve were offered concentrates according to their individual milk yields but with a mean concentrate allowance over the twelve cows of 7·2 kg d‐1. Degree of chopping had no significant effect on either silage dry matter intake or milk yield with mean intakes during the final 21 d of the experimental period of 9·2, 9·2 and 9·2 kg d‐1and mean milk yields of 25·2, 25·2 and 25·2 kg d‐1for the DC, PL and PS silages respectively. Method of concentrate allocation did not affect either the total yield of milk during the experimental period, 1717 and 1697 kg, or the total lactation yield, 5635 and 5711 kg, for the uniform and yield‐related allocation methods respectively. The butterfat and protein concentrations of the milk were not significantly affected by either the degree of silage chopping or the metho