Summary. The effect of herbage allowance (20, 30, 40, 50, 60 and 70 kg DM/cow. day) on the consumption of nutrients from herbage and milk production by cows in early lactation, was examined. The experiment was conducted on rainfed perennial ryegrass pastures in September and October 1997 in south-eastern Victoria, Australia. The herbage on offer comprised 64% perennial ryegrass, 21% other grasses, 1% white clover, 5% weeds and 9% dead material on a dry matter (DM) basis. The average pregrazing herbage height was 13 cm, at an estimated pregrazing herbage mass of 3.6 t DM/ha. The herbage on offer was of high quality containing 11.6 MJ metabolisable energy/kg DM, 202 g crude protein/kg DM and 525 g neutral detergent fibre/kg DM. Concentrations of calcium, magnesium, sodium, potassium, phosphorus, sulfur and chloride were 4.4, 2.2, 4.4, 31.2, 3.5, 2.7 and 11.4 g/kg DM, respectively. As daily herbage allowance per cow increased, dry matter intake increased curvilinearly (P<0.01) from 11.2 to 18.7 kg DM/cow. day. This was associated with a decrease in utilisation of herbage from 54 to 26% and an increase in milk production from 25.9 to 29.1 kg/cow. day. The cows on all treatments grazed for less than 8.7 h/day. The increase in intake was achieved by an increase in the rate of herbage intake from 1.5 to 2.2 kg DM/h for herbage allowances of 20 and 70 kg/cow.day, respectively. Irrespective of herbage allowance, cows selected a diet that was approximately 10% higher in in vitro dry matter digestibility and 30% higher in crude protein than that in the herbage on offer. The neutral detergent fibre content of the herbage selected was lower (P<0.05) than that on offer. The herbage consumed contained more (P<0.05) magnesium, potassium and sulfur, the same amount of calcium and phosphorus and less (P<0.05) sodium and chloride than the herbage on offer. For rainfed perennial pastures in spring, herbage allowance is an important factor in determining voluntary feed intake and production of dairy cows. To achieve 30 L from herbage, without supplementation, high herbage allowances are required. The increase in herbage intake, with increasing herbage allowance, resulted from an increase in rate of dry matter intake and not an increase in grazing time. No relationship was evident between herbage allowance and the selection differentials for in vitro dry matter digestibility, crude protein and neutral detergent fibre. Selection differentials for rainfed perennial pastures in spring are similar to those reported for irrigated perennial pastures in northern Victoria in spring and autumn. When determining nutrient requirements it is important to consider the interaction between herbage intake and nutrient concentration in the herbage.