Authoritative reviews of the question of whether occupational exposure to beryllium compounds is associated with increased risk of respiratory cancer were published in 1987 and were critical of the quality of the evidence available up to that time. No clear conclusion could be drawn from it as to the carcinogenicity of beryllium to humans. If studies published since 1987 are to lead to a revision of the regulatory status of beryllium compounds they must clearly be of high quality and scientific validity. These studies, as well as the earlier reports, are reviewed here. I argue that the small and inconsistent excess of lung cancer deaths in employees of one or two plants seen in two post-1987 studies is compatible with a number of explanations other than that they are attributable to occupational exposure to beryllium. Specifically, information on cigarette smoking is poor, and the data do not exist to rule out the possibility that the small number of excess deaths results from residual confounding by cigarette smoking patterns in the populations studied. Indeed, excess deaths from emphysema and ischemia heart disease in the same cohort suggest that confounding by cigarette smoking is a more likely explanation of the lung cancer excess than is occupational exposure to beryllium compounds.