The Japanese Environmental Mutagen Society/the Mammalian Mutagenicity Study group conducted a collaborative study to investigate whether cell nuclei or whole cells might be more suitably used to correctly detect genotoxic chemicals in the in vivo rodent alkaline Comet assay. Four participating laboratories applied four sample processing methods, i.e., three homogenization methods using the usual Potter-type shaft, a customized (loose) Potter-type shaft, or a Downs-loose-type shaft, for preparing cell nuclei, and the mesh membrane method for preparing whole cells, to the male rat liver. Homogenization with the usual Potter-type shaft clearly produced damage of the cell nuclei and DNA, while the other three methods seemed to provide similar conditions of the tissue samples. The proportion of cell nuclei: whole cells was 80-90%: 10-20% in all laboratories when the samples were prepared by homogenization using a Downs-loose-type shaft or by the mesh membrane method. The %DNA in tail were comparable in both samples among the negative control groups (single oral administration with physiological saline) of all laboratories, and showed an equal degree of increase in both samples of the ethyl methanesulfonate groups (single oral administration at 250 mg/kg) in all laboratories. In conclusion, the homogenization method using a loosely customized Potter-type shaft or a Downs-loose-type shaft, and the mesh membrane method would be equally acceptable for the in vivo rodent alkaline Comet assay.