The milk production of 14 Hereford × British Friesian females in their second lactation (cows) and 36 in their first lactation (heifers) was measured at three levels of nutrition either by machine-milking twice a day, without the use of exogenous oxytocin, or by a calf-suckling technique. Machine-milking significantly increased the number of cows drying-off within 150 days of parturition (i><0·05) and significantly reduced milk yields compared with the calf-suckling technique (P< 0·01). The 150-day cumulative milk yields of the heifers were not significantly affected by method of milking, but the coefficient of variation of milk yield was significantly greater in the machine-milked group(P< 0·05). One machine-milked heifer dried-off within the first 150 days of lactation. There was no evidence of an interaction between method of milking and plane of nutrition in either the cows or the heifers. It was concluded that measuring the milk production of beef cows by machine-milking, without the use of exogenous oxytocin, may be unreliable. Factors influencing the usefulness and reliability of the calf-suckling technique are discussed.