This retrospective cohort study investigated the cancer mortality of 1341 herbicide applicators licensed before 1 January 1980. The cohort was followed for mortality until 1 January 1988 by means of the Dutch system of population registers. For the subjects who died before 1 January 1988 the cause of death was obtained from the Central Bureau of Statistics. The cause of death was obtained for 98.4% of the deceased subjects. The total mortality was lower than expected according to mortality rates of the total male Dutch population. The total number of deaths from cancer was however slightly higher than expected [standardized mortality ratio (SMR) 114]. Mortality from several subtypes of cancer was also higher than expected, that for multiple myeloma being statistically significantly higher than unity (SMR 815, 95% confidence interval 164-2382). Occupational exposure to herbicides appears to increase the risk for multiple myeloma and possibly other malignant neoplasms.