The aim was to determine whether administration of melatonin would alter the yield and composition of milk from grazing dairy cows in summer. Twelve sets of spring-calving identical twin Friesian cows were used in the experiment. In late-November (late spring), one twin from each set was given slow-release melatonin implants behind the ears (108 mg melatonin/cow). Two further implantations occurred at 4-weekly intervals to maintain increased circulating concentrations of melatonin for 12 weeks. The other twin served as a control. Milk yield and composition were measured twice prior to treatment and then four times over the following 12 weeks. Concentrations of melatonin, prolactin and insulin-like growth factor 1 (IGF-1) were measured in blood plasma twice before treatment and then either seven (melatonin and prolactin) or three (IGF-1) further times during the experiment. Management procedures for all cows were similar and cows grazed a daily pasture allowance of approximately 30 kg DM/cow as their sole feed source. In melatonin-treated cows there was a decrease in mean concentrations of prolactin in plasma, but concentrations of IGF-1 did not change. Melatonin reduced milk yield by 6 weeks after treatment and by the end of the 12-week experimental period milk yield in melatonin-treated cows had fallen by 23%. Melatonin also reduced concentrations of lactose in milk, but increased concentrations of fat, protein and casein, changes that were broadly similar to those that occur in late lactation in seasonally calving dairy cows. Thus, the results suggest that some of the variation in the volume and quality of milk throughout the season in New Zealand dairy systems may be due to changes in photoperiod mediated by increased concentrations of plasma melatonin in association with decreased concentrations of plasma prolactin.