Local and federal government agencies are often asked to investigate apparent clusters of cancer in communities or workplaces. Often these investigations cannot utilize the methods that have been developed for evaluation of disease clusters because the clusters are too small, and the populations to be studied and the periods of time to be covered are determined in an a posteriori manner. Still, government investigators are called upon to render an official opinion of the apparent clusters. Application of a theoretical approach to cluster analysis must give way to a more pragmatic approach. A review of 61 investigations of apparent clusters conducted by the National Institute for Occupational Safety and Health (NIOSH) during the period 1978-84 showed that most of the clusters contained five or fewer cases and had no plausible occupational etiology. Despite the few clusters that were identified, these investigations generally provided a service to workers and employers who were concerned about occupational cancer.