Severe acute respiratory syndrome (SARS) first appeared in Guangdong Province, China, in November 2002. Although virus isolation and serology were useful early in the SARS outbreak for diagnosing new cases, these tests are not generally useful because virus culture requires a BSL-3 laboratory and seroconversion is often delayed until 2 to 3 weeks after infection. The first qualitative reverse transcriptase-polymerase chain reaction tests for SARS-coronavirus (CoV) were sensitive and capable of detecting 1 to 10 genome equivalents. These assays were quickly supplemented with quantitative real-time assays that helped elucidate the natural history of SARS, particularly the initial presence of low viral loads in the upper respiratory tract and high viral loads in the lower respiratory tract. The unique natural history of SARS-CoV infection dictates the testing of both respiratory and nonrespiratory specimens, the testing of multiple specimens from the same patient, and sending out positives to be confirmed by a reference laboratory. Commercially available reverse transcriptase-polymerase chain reaction tests for SARS have recently appeared; however, meaningful evaluations of these assays have not yet been performed and their true performance has not been determined. These and other issues related to diagnosis of SARS-CoV infection are discussed in this review.