Introduct ion During the last 2 to 3 yr, numerous articles have been published concerning the reemergence of tuberculosis in the U.S. This resurgence was a result of a number of complex factors, among them homelessness and other socioeconomic factors, the human immunodeficiency virus (HIV) epidemic, and hospital and other group-setting outbreaks of tuberculosis. Further complicating these outbreaks were frequent delays in laboratory repotting of smear, culture, and/or susceptibility test resuits. As a consequence, an important question was raised as to how one could provide a more rapid diagnosis for tuberculosis given the urgent need. Because of the obvious demand for a reliable and rapid means of diagnosing tuberculosis for public health and therapeutic reasons, the discovery of the polymerase chain reaction (PCR) about 10 yr ago, with its potential ability to exponentially copy virtually any targeted sequence of DNA in an afternoon, is particularly attractive. Since PCR's discovery, a number of other approaches to amplification have also been introduced to detect Mycobacterium tuberculosis (MTB) directly in clinical specimens; some of these amplification formats are listed in Table 1. Regardless of the format, these amplification assays show promise as rapid, sensitive, and specific methods to diagnose tuberculosis. Despite the significant promise these assays hold for improved laboratory diagnosis of tuberculosis, a number of issues must be considered by clinical microbiologists before these assays are introduced into the laboratory routine. Key issues include expense, the assays' exact role in the diagnosis of mycobacterial infections, and their clinical utility. In this article, I would like to discuss important issues concerning these assays and then suggest some possible immediate and future applications for them. However, if these issues are to be fully appreciated, a brief summary of data published to date is prerequisite.