Operant conditioning techniques were used to measure the feeding motivation (hunger) of pigs restricted to the low food levels used under certain commercial conditions. Six Large White × Landrace boars were maintained in individual pens where they received their daily food allowance. Initially food was available ad libitum; subsequently boars were restricted proportionately to 0·8, 0·6 and 0·4 of their ad libitum intake in a Latin-square design. Finally boars were again offered their previously measured ad libitum intake (treatment 1·0). Hunger was measured by conditioning boars to press a panel 10 times to receive a 7-g reward of food. Operant response rates were measured over 20-min sessions during each of the treatments and at three times of day (pre-feeding, immediately post feeding and 5 h post feeding). Operant response rate was strongly affected by restricting food allowance (means: treatment 1·0, 42 panel presses; 0·8, 212; 0·6, 266; 0·4, 320). There was a signficant interaction between food allowance and time of day (P < 0·001). Treatments did not vary in response rate before feeding. Treatments 1·0 and 0·8 depressed response rate post feeding (P < 0·001) and treatment 1·0 also 5 h post feeding (P < 0·001). Response rates on treatments 0·6 and 0·4 did not vary significantly from one another or with time of day. Restricting pigs to low food levels resulted in a high level of hunger throughout the day. This was so even for food allowances that met current recommended feeding levels sufficient for maintenance and low weight gain (treatment 0·6 equivalent to 1·3 times maintenance). Commercial levels of feeding for the pig breeding population whilst meeting production needs do not satisfy feeding motivation. The hunger resulting from low food levels may act as a major source of stress under certain husbandry conditions such as tether and stall systems for dry sows, where sows are subject to both physical and food restriction.