A case/control study of workers at facilities in China known to have a history of heavy silica (14808607) exposure, was conducted in order to assess the role of silica in the etiology of lung cancer. A total of 316 men with lung cancer and 1352 comparison workers from 29 mines and factories (including pottery factories, and clay, tungsten, iron/copper and tin mines). A detailed quantitative exposure matrix was developed using information available from work histories, industrial hygiene records, and a monitoring program specially designed for the investigation. A job title dictionary was developed. Information of confounding exposures such as radon (10043922), arsenic (7440382), polycyclic-aromatic-hydrocarbons (PAH), and tobacco use was obtained. Odds ratios (OR) analysis showed that 30% of the cases were from tungsten mines, 28% from tin mines, 25% from iron/copper mines, and 18% from pottery factories. Information on history of tobacco smoking was obtained. The highest risk associated with cumulative total dust and respirable silica was in tin miners followed by pottery workers. Tungsten miners had decreased risks, and iron/copper workers showed some increase with rise in total dust, but not with respirable silica. Although exposure to PAH and silica correlated among pottery workers, adjustment for this confounder slightly raised rather than lowered the OR for silica exposure. Among tin miners, exposure to arsenic was significantly associated with risk of lung cancer. Both arsenic and PAH exposure were correlated with silica concentrations in these workers, so that adjustment for these confounders was not possible. No clear trend with regard to radon was observed. A significant excess of silicosis occurred among iron/copper and tin miners with lung cancer, but not among pottery workers or tungsten miners, even though these two categories were subjected to heavier silica exposure. The authors conclude that while there is some evidence for an association between silica exposure and lung cancer, it is difficult to dissociate the effects of confounders. Furthermore, inconsistencies exist in the association between silicosis and risk of lung cancer.