Three inbred strains and all their possible F1 crosses were monitored in one familiar and three unfamiliar situations. Their behavior was described by species-specific elements of the ethogram. Genetic variability was demonstrated both for behavioral elements and for factors extracted by Principal Components Analyses (PCA). We studied how the behavior of genetically different fish changed across situations and examined the interrelations both among variables measured in one situation and between those measured in different ones. Behavioral changes across situations proved to be different for certain strains and crosses, that is, genotype-environment interaction was found. The PCA's carried out for the 4 situations separately yielded unlike factor structures. Another PCA, in which all the variables were included, proved that there was correlation among certain variables measured in different situations. In general, it seems that the corresponding behavioral elements do not always represent the same phene in different situations. We discuss how the genotype-environment interactions can be interpreted, try to define behavioral strategies using the extracted factor structures, and construct a model for the organization of Macropodus behavior.