Stochastic catastrophe analysis of effects of density on behavior: An example in laying hens

P. Koene

Ethology, Animal Sciences, Wageningen University, The Netherlands

When an individual animal is simultaneously attracted and repelled by a goal or situation, the individual is in an approach-avoidance conflict. Several models have been postulated to describe the dynamics of this type of conflict and its resolution. In a social situation individuals are attracted to each other for cooperation or safety and simultaneously repelled by competition. Our purpose was to try to describe this conflict in a catastrophe model. This class of models includes a number of characteristics: bimodality, divergence, hysteresis and inaccessibility. In rats we showed that individual non-social conflicts give lead to either stable approach or stable avoidance behavior, i.e. bimodality and bistability in behavior).

The effects of area size and group size on behavior of laying hens were measured. Sub samples of 40 Lohman Brown laying hens were tested in 25 crossed combinations of group size (2, 4, 6, 8, and 10 birds) and density (1, 3.2, 10, 17.8 and 31.6 birds/m2). Using behavioral sampling, all vocalizations and overt interactive behavior were measured.

Multiple regression analysis showed that there were strong signi.cant relations of vocal and overt behavior with stocking density and group size. Increase in group size leads to less vocalizations per hen (p<0.004), while an increase in density leads to more vocalizations per hen (p<0.001). Analysis using statistical catastrophe models shows that in most cases the explained variance increased significantly using a cusp model (example: R2 cusp = 0.56 vs. R2 linear = 0.24).

Part of the data shows linear relations and another part shows non-linear relations between vocalizations and density and group size, i.e. bimodality in behavior. In situations of identical parameters of group size and density two opposite behavioral solutions are found. According to Catastrophe Theory these different reactions of the hens are caused by their preceding experience.


Paper presented at Measuring Behavior 2005 , 5th International Conference on Methods and Techniques in Behavioral Research, 30 August - 2 September 2005, Wageningen, The Netherlands.

© 2005 Noldus Information Technology bv