The role of crystalline silica in lung cancer has long been the subject of controversy. In this article, we review the main experimental and epidemiological studies dealing with this problem. Some evidence for a genotoxic potential of crystalline silica has been obtained in the rare in vitro studies published to date. In vivo studies have shown that crystalline silica is carcinogenic in the rat; the tumour types appear to vary according to the route of administration. In addition, an association between carcinogenic and fibrogenic potency has been observed in various animal species exposed to crystalline silica. An excess of lung cancer related to occupational exposure to crystalline silica is reported in many epidemiological studies, regardless of the presence of silicosis. However, most of these studies are difficult to interpret because they do not correctly take into account associated carcinogens such as tobacco smoke and other occupational carcinogens. An excess of lung cancer is generally reported in studies based on silicosis registers. Overall, experimental and human studies suggest an association between exposure to crystalline silica and an excess of pulmonary malignancies. Although the data available are not sufficient to establish a clear-cut causal relationship in humans, an association between the onset of pneumoconiosis and pulmonary malignancies is probable. In contrast, experimental observations have given rise to a pathophysiological mechanism that might account for a putative carcinogenic potency of crystalline silica.