Comparing behaviour in wild and laboratory strains of the house mouse: Levels of comparison and functional inference.

Many behavioural studies of laboratory house mice have sought functional explanations for their results. This raises the problem of the comparability of behaviour with that of wild house mice from whose naturally selected repertoire the behaviour of laboratory strains has been derived. This in turn requires careful choice of measures to ensure that subtle changes in response during the selection of laboratory strains are taken into account. A comparison of social behaviours showing kin bias in male mice of the CFLP laboratory strain shows that superficial measures based on totals over an arbitrary test period, a widely used procedure, can be misleading because they fail to control for differences in the tendency for CFLP and wild mice to encounter one another and the intensity of social responses when encounters do occur. Such differences may reflect different contexts and consequences of social responses in the two strains.