There has been much discussion on acceptable spontaneous mutant frequencies in the mouse lymphoma assay (MLA). This culminated in the International Workshop on Genotoxicity Testing (IWGT) recommended control limits for the microtitre version of 50-170 mutants/10(6) viable cells, which has now been included in the draft Organization for Economic Co-Operation and Development guideline for assays investigating mammalian cell gene mutation at the tk locus. Some of the factors affecting mutant frequency have been investigated. It was shown that when culturing methotrexate cleansed TK⁺/⁻ cells, a spontaneous mutant frequency of ∼100 mutants/10⁶ viable cells was achieved after only 26 doublings. However, after further culturing for ∼6 months the spontaneous mutant frequency only gradually increased. Culturing for this time did not affect the karyotype of the cell in so much as the modal chromosome number remained stable. The spontaneous mutant frequency could effectively be manipulated by cleansing with various concentrations of methotrexate. The necessity for using appropriately heat-inactivated horse serum was confirmed. Finally, following treatment with 4-nitroquinoline-N-oxide, cells did not preferentially survive when plated at high cell densities (1.6 cells plus 2,000 feeder cells/well) versus cells at low density (1.6 cells/well). It was considered that these findings confirm that the dynamics of spontaneous mutant formation in the MLA are complex. However, the karyotype of L5178Y cells is remarkably stable and assuming investigators are using cells with appropriate provenance and good culturing technique, it is clear that the IWGT recommendations are achievable.