Although relatively small, Michigan's elk (Cervus elaphus nelsoni) herd is highly valued by both hunters and the general public. Elk and red deer (Cervus elaphus elaphus) are highly susceptible to infection with Mycobacterium bovis, the causative agent of bovine tuberculosis (TB), and outbreaks have been documented worldwide. The Michigan elk range lies entirely within counties where TB is known to be enzootic in white-tailed deer (Odocoileus virginianus). Consequently, a project was undertaken to estimate the true prevalence of TB in Michigan's free-ranging elk herd. All elk harvested by licensed hunters during 2002–2004, and all nonharvest elk mortalities examined by the Michigan Department of Natural Resources Wildlife Disease Laboratory from November 2002–May 2005, were screened for gross lesions of TB with samples of cranial lymph nodes and palatine tonsils collected for histopathology and mycobacterial culture. In all, 334 elk were included in the study. Twenty-three elk with gross lesions were considered TB suspects; all were culture-negative for M. bovis. However, M. bovis was cultured from two elk without gross lesions. The sensitivity, specificity, and negative predictive value of the current TB surveillance protocol were 0%, 100%, and 99.4%, respectively, while the apparent prevalence and true prevalence calculated directly from the sample were 0% and 0.6%, respectively. The positive predictive value and the estimated true prevalence of the population were undefined. The poor sensitivity of current surveillance was likely an artifact of its application to a relatively small sample, in order to detect a disease present at very low prevalence. The low prevalence of TB in Michigan elk, and the early stage of pathogenesis of the few infected animals, does not suggest elk are maintenance hosts at the present time.