Male breast cancer incidence was examined in a collaborative study of data from the cancer registries in Denmark, Finland, Norway and Sweden, comprising a total of 1,529 cases diagnosed from 1943 to 1982. Effects of age, time of diagnosis, birth cohort and country of residence were evaluated by using statistical models of the multiplicative Poisson type. The logarithm of incidence increased linearly with the logarithm of age and had a slope of about 5. In Denmark, male breast cancer incidence increased significantly with time, about 1% per year. No significant time effects were observed in the other countries. A significant effect of birth cohort appeared in Sweden only, but there was no particular trend in incidence by year of birth. In relation to Denmark, the risk of male breast cancer was lower in Sweden, Norway and Finland, the relative risk estimates with 95% confidence intervals being 0.82(0.72-0.94), 0.72(0.61-0.84) and 0.53(0.43-0.64) respectively. The variation within Scandinavia is similar for female breast cancer, pointing to common factors being involved in the etiology of breast cancer in both sexes.