Three independent experiments involving a total of 288 SENCAR mice were used to study the effects of 60-Hz magnetic fields on the growth and development of skin tumors. Given the constraints imposed by the experimental design, the results did not support a role for magnetic fields as a tumor co-promoter. This negative finding could also be interpreted to mean that the SENCAR mouse skin tumor model was not sensitive enough to detect the action of a weak co-promoter. The two-stage (initiation/promotion) model was used to assess the genotoxic potential of magnetic fields because it had been widely used to evaluate chemical carcinogens. This model, however, lacks the sensitivity to detect all but the most potent direct-acting carcinogens, and the tumor response to the action of low doses of promoter results in large random fluctuations in tumor incidence, yield, and multiplicity. The need to limit tumor incidence in the sham is a necessary condition to ensure that a magnetic field-induced effect on tumorigenesis would have a reasonable chance of being detected. This requirement, and the variability in tumor development between and within experiments, increases the level of uncertainty in the system and makes a weak response to the magnetic field difficult to detect and interpret.