Commercial laying hen chicks experience continuous light for up to 24 h/day in the first week of life. Under these conditions, active chicks disturb, and may direct feather pecks towards resting ones. A study performed by Bristol's research group found that on average, across observations taken at all ages, dark brooder flocks performed significantly less severe feather pecking than control flocks and had a significantly lower percentage of birds with missing feathers. Individual feather condition deteriorated less quickly in the dark brooder group and dark brooded birds ended up slightly heavier than birds from the control treatment. In conclusion Bristol found no detrimental effects of dark brooding on commercial farms and suggest this is a promising approach to reducing problems with feather pecking and generally improving the welfare of commercial pullets. These findings regarding the positive impacts of Dark Brooding have been taken up by industry with two companies adopting dark brooding for their birds.

Previous experimental work with small groups showed that both problems were reduced in chicks brooded by dark brooders (heaters). The current study aimed to extend these small-scale trials by examining the use of dark brooders on two commercial rearing farms. Each farm contributed two identical houses, one of them equipped with dark brooders and the other with regular brooders. The experiment comprised five replicates, each consisting of one dark brooder flock and one control flock (total of 10 flocks). Each flock contained 2000 Columbian Blacktail chicks with intact beaks, which were reared to organic standards. Observations took place three times during the rearing period at 1, 8 and 16 weeks and three of the five replicates were also followed into lay, with observations at 25 and 35 weeks. Bird weights, the evenness of body weight, mortality at the end of rear, feather pecking, the percentage of the flock with missing feathers and individual feather scores were measured, as well as the flock's reaction to a novel object and an approaching human in selected areas of the house. Apart from mortality, which was analysed as a paired t-test in PASW Statistics 18, data were ordered in three (or four) levels (visits within (flock within) replicate within farm) and were analysed using the multilevel statistical software MLwiN 2.25. Treatment and age were entered in the model as explanatory variables.