1. In experiment 1, 10 laying hens were given the choice to eat food pellets from any of 4 food bowls illuminated by overhead, incandescent luminaires at <1, 6, 20 or 200 lux. During a trial hens were allowed to eat for 5 min. After each minute had elapsed (from the start of eating) the light sources were extinguished and the illuminances re-assigned to the food bowls in a random manner. Each hen received two trials, one where the food was freely available and another where it was hidden in a sand and gravel mix. 2. The hens chose to eat for most time in the brightest (200 lux) and least in the dimmest (<1 lux) environments for both free and hidden food (free: 5.9, 10.5, 10.4, 15.7s for increasing illuminance; Hidden: 5.5, 9.8, 9.1 and 15.7s. 3. In experiment 2, 9 hens were trained to peck at either an illuminated or unilluminated panel to access a food reward behind a guillotine door for 3 s. Five hens were trained to peck the illuminated panel to access food brightly lit (200 lux) or the unilluminated panel to access food dimly lit (<1 lux); 4 hens were trained vice versa. The flock was then divided into three groups of three, and three treatments imposed on each group in a Latin-square arrangement. In treatment 1, one peck at either panel allowed access to the chosen light environment (F1:F1). In treatment 2, 5 pecks were required to access food brightly lit on a variable ratio, but only one to access food dimly lit (ratio V5:F1). In treatment 3, the variable ratio was increased to V10:F1 to access food in the light. 4. Over 40 trials for each hen, the mean number of attempts to eat food in the light (where the panel which allowed access to food brightly lit was pecked at least once) was 34.5 for F1:F1, 12.1 for V5:F1 and 8.5 for V10:F1. The mean number of food rewards taken in bright light was 34.5, 3.1 and 1.8, respectively. For both variables, the difference between F1:F1 and V5:F1 was significant but not between V5:F1 and V10:F1. By interpolation of the 'attempts' data, it was estimated that hens would work 2.3 times harder to gain access to food brightly lit than for food dimly lit. 5. In experiment 3, the influence of the same illuminances applied over a food bowl as in experiment 1 (<1, 6, 20 or 200 lux) on the number of pecks/min, food consumed/min, food consumed/peck and the force of pecks was examined. 6. The amount of food consumed was lowest in the dimmest environment (3.1 vs 7.5, 7.4 and 7.1 g/ min for increasing illuminance, respectively); as was the number of pecks (35.6 vs. 125.0, 123.1, and 125.4 pecks/min respectively for increasing illuminance). The amount consumed per peck did not vary significantly with illuminance. The mean peck force showed a trend to be lowest in the dimmest environment (5.3 vs. 6.6, 7.0 and 6.6 N respectively, for increasing illuminance). 7. Overall, the hens showed a preference and appeared motivated to eat in bright as opposed to dim light. The hens were unwilling to eat at low illuminances although the 'efficiency' of eating (g/peck) was not impaired significantly. These data may have implications for novel lighting systems and those where hens are required to eat in the dark or in very dim light.