The emergence or non-emergence of attack behavior results from interaction between the genotype and the conditions under which the mice are tested. Inbred mice of the same strain reared or housed under conditions do not react the same way; reactions also vary according to the place selected for testing and the different opponents. A factor analysis showed that the attack behavior in non-isolated males, tested in neutral area covaried with high testosterone and steroid sulfatase and low brain 5-hydroxytriptamine (5-HT), beta-endorphin and Adrenocorticotropic Hormone (ACTH) concentration, whereas, for isolated males tested in their own housing cage, it covaried with high testosterone activity and low brain 5-HT concentration. A wide genome scan was performed with two independent populations derived from C57BL/6J and NZB/BlNJ, each being reared, housed and tested under highly contrasting conditions, as described above, and confronted with A/J standard males. Common Quantitative Trait Loci emerged for two rearing/testing conditions. For rattling latency we detected Quantitative Trait Loci on Mus musculus chromosome 8 (MMU8) (at 44, LOD score=3.51 and 47 cM, LOD score=6.22, for the first and the second conditions) and on MMU12 (at 39 cM, LOD score=3.69 and at 41 cM, LOD score=2.99, respectively). For the number of attacks, Quantitative Trait Loci were common: on MMU11 at 39 cM LOD score=4.51 and 45 cM, LOD score=3.05, respectively, and on MMU12 (17 cM, LOD score=2.71 and 24 cM, LOD score=3.10). The steroid sulfatase gene (Sts), located on the X-Y pairing region, was linked, but only in non-isolated males, tested in neutral area for rattling latency, first attack latency, and number of attacks (LOD scores=4.9, 4.79 and 3.57, respectively). We found also that the Quantitative Trait Locus encompassing Sts region interacted with other Quantitative Trait Loci. These results indicate that attack behavior measured in different rearing and testing conditions have different biological and genetic correlates. This suggests that further explorations should be done with standardized tests and, in addition, with a wide range of tests, so as to gain an understanding of the true impact of genes or pharmacological treatments on specific categories of aggressive behavior.