To evaluate the geographical distribution of the mortality from malignant tumours in relationship with exposure to chemical carcinogens in the work place, and to asses the possible association between these questions and the percentage of population employed in certain high-risk sectors, an ecological study in the Valencian Community (VC), Spain was carried out. Age-adjusted mortality rates for the total number of malignant tumours, lung, bladder, lymphomas and leukaemia during the periods 1981-1985 and 1991-1995 were calculated for the 34 geographical areas. The percentage of population in each area working in sectors in which they may be exposed to chemical carcinogens was obtained. The relationship between mortality on-the-job exposure was studied using linear regression methods. Large differences in cancer mortality were seen. In men, the geographical pattern was very stable and reveals a significant association with the distribution of certain high-risk jobs. Statistically significant positive correlations (p < 0.001) were found between cancer mortality and the percentage of the population working in metal, wood and furniture sectors. In contrast, a negative and statistically significant (p < 0.001) correlation was observed between cancer mortality and the percentage of the population working in agriculture. In conclusion, although the variability in cancer mortality in men was significantly associated with some occupational sectors in the VC, caution is needed when drawing conclusions about causation from ecological studies.