The results of a further follow up of two cohorts of men identified as manufacturers of fertilisers in the 1961 and 1971 censuses of England and Wales are reported. Both cohorts have now been followed up until 1985 through the National Health Service Central Register. In the 1961 cohort the previous findings are replicated--low mortality from all causes, all cancers, and all circulatory diseases, and no evidence of dose response relation nor excess mortality from cancer at any site. In the 1971 cohort there remain more deaths from cancer than expected but none of the observed deaths from individual cancers differs significantly from expectation by comparison with national mortality rates or rates for other employed men. Evidence of an association between cancer mortality and frequency of exposure to nitrate containing dust is strengthened in the reanalysis of the 1971 cohort after longer follow up, but a negative trend, significant for digestive organ cancers, is observed in the analysis by product type in which manufacturers of straight nitrogen and compound fertilisers might have been expected to be most at risk. Failure to find specifically any excess mortality from gastric cancer and the contradictory dose response relations, weigh against the notion that the nonsignificant excesses of cancers in the 1971 cohort are related to nitrate exposure.