Performance of an animal-based test of thirst in commercial broiler chicken farms.

Animal-based measures of thirst are currently absent from animal welfare monitoring schemes due to the lack of a well-validated indicator applicable for on-farm use. In the present study, an on-farm test based on voluntary water consumption from an unfamiliar open drinker was validated in a (semi-)commercial setting. To investigate the effect of thirst on water consumption, we subjected 4 flocks of 1,500 broilers to either 0 or 12 h of water deprivation and subsequently measured the amount of water that small subgroups consumed after the deprivation period (first experiment). Broilers that were water deprived before the test drank more than control broilers (P < 0.001). In a second experiment, a similar test was performed using 20 commercial broiler flocks in Belgium and Brazil. After a pretreatment water consumption test, the birds were subjected to 0 or 6 h of water deprivation, and a posttreatment water consumption test was conducted. Only in Brazil, deprived birds drank significantly more than controls in the posttreatment water consumption test (P < 0.001). A tendency for a difference was found in Belgium (P = 0.083). Pre- and posttreatment water consumption was higher in Brazil than in Belgium (P < 0.001). Stocking density and temperature influenced, respectively, the pretreatment and the control's posttreatment water consumption in Brazil, but not in Belgium. These results indicate that the water consumption test is sufficiently sensitive to discriminate between control and 12 h deprived flocks, and in Brazil even between control and 6 h deprived birds. The location of the test within the house did not affect the amount of water consumed in either experiment, suggesting that this variable does not have to be standardized. However, the amount of water consumed by broilers able to drink freely for a long period depended on indoor climatic variables (in Brazil only) and possibly genotype. This suggests that these variables need to be considered when interpreting the test outcome in terms of the thirst level experienced by the broilers.