Demand functions were used in an attempt to estimate the importance of resting in cattle. Furthermore, the sensitivity of the technique was determined by quantifying the effects of variations in the level of prior deprivation (Experiment 1) and reward duration (Experiment 2) on elasticity and intensity of the demand function. In both experiments eight heifers were housed in tether stalls where they could be deprived of lying by attaching a girth strap secured to a rafter above the animal. During a 3-h test period (10:00 to 13:00 h) a panel was placed in front of each heifer and she could work by pressing this panel on Fixed Ratio (FR) schedules for a releasing mechanism to be activated allowing her to lie down (a reward). In Experiment 1 the heifers were deprived of lying twice daily (afternoon and morning) for either two periods of 3 h (2 x 3 h) or two periods of 6 H (2 x 6 h) prior to testing. The FR varied from 6 to 24 and the reward duration was 15 min. In Experiment 2, heifers could lie for either 10 or 20 min per reward at either of the two deprivation levels used in Experiment 1. Thus, the four treatments were (A) 2 x 3 It of deprivation and 10 min per reward, (B) 2 x 6 h of deprivation and 10 min of reward, (C) 2 x 6 h of deprivation and 20 min of reward, and (D) 2 x 3 h of deprivation and 20 min of reward. Here the FR varied from 10 to 50. The demand functions were linear in arithmetic co-ordinates, which means that the elasticity may not be constant across FR-values. In Experiment 1 the heifers earned more rewards the higher deprivation level (P < 0.001). The elasticity of demand was calculated for each FR-value for the two levels of prior deprivation. For 2 x 3 h of deprivation the numerical elasticity increased from 0.07 to 0.35 with increasing FR (P < 0.001). For 2 x 6 It of deprivation the numerical elasticity tended to increase (P < 0.10) from 0.02 to 0.10. The demand was more inelastic after the longest prior deprivation for the highest FR-values (P < 0.05), which indicates that the heifers valued access to rest more after the longest deprivation. In Experiment 2 the heifers earned more rewards after two periods of six hours of deprivation when a reward duration of 10 min was given than when a reward duration of 20 min was given (P < 0.01). The numerical elasticity of the demand functions increased with increasing FR from 0.07 to 0.5 (P < 0.001), from 0.06 to 0.4 (P < 0.001), from 0.03 to 0.2 (P < 0.05) and from 0.04 to 0.3 (P < 0.05), for treatments A, B, C and D, respectively. After the longest deprivation the demand function was less elastic when a reward duration period of 20 min was given compared to a reward duration of 10 min (P < 0.05). These results indicate that the longest reward duration was more valuable than the shortest when the deprivation prior to test was high. The variation in elasticity with prior deprivation and reward duration has to be considered in future studies of the demand for rest in dairy cattle. (C) 2004 Elsevier B.V. All rights reserved.