An 18-year follow-up for mortality and cancer morbidity was conducted in a cohort of 1,909 men who had started spraying chlorophenoxy herbicides (mixture of 2,4-dichlorophenoxyacetic acid [2,4-D] and 2,4,5-trichlorophenoxyacetic acid [2,4,5-T]) in 1955 through 1971. In all, 384 persons had died during the follow-up, and there was a slight deficit in natural deaths (standardized mortality ratio [SMR] 0.84; 95% confidence interval [CI] 0.75-0.94). By contrast, there was a small, nonsignificant increase in accidental and violent deaths. The overall cancer mortality was slightly less than in the general population (SMR 0.83; 95% CI 0.65-1.02), and not a single case of death of non-Hodgkin's lymphomas (NHL) or soft tissue sarcomas (STS) was detected. With regard to cancer morbidity, the incident cases showed a slight deficit compared to the population figure (standardized incidence ratio [SIR] 0.81; 95% CI 0.67-0.97). One case of NHL was found (2.4 expected with 10 years of latency), but not a single case of STS (0.8 expected with 10 years of latency). While our study does not support the contention that spraying of 2,4-D and 2,4,5-T containing herbicides carries any significant risk of cancer, the medium to low statistical power of the study does not allow any far reaching negative conclusions regarding the carcinogenicity of the agents.