A primer-directed nucleic acid amplification reaction, commonly known as the polymerase chain reaction (PCR), has been adapted to the identification of bluetongue virus (BTV) RNA. The protocol described in this article is designed for the detection of a purported globally-conserved, serogroup specific nucleic acid sequence within the BTV genome. Due to the double-stranded RNA composition of the BTV genome, the original polymerase chain reaction protocol has been modified to include chemical denaturation and reverse transcription steps that allow selective amplification from this unique template molecule. The amplification procedure yields a 210 base pair product when tested on samples of RNA from the prototypic strains of U.S. BTV serotypes 2, 10, 11, 13, 17. The amplification product is tentatively identified by agarose gel electrophoresis of the PCR samples. Final positive identification of the amplified BTV-specific product is determined by Southern blot hybridization of the PCR samples. RNA samples from uninfected cell culture controls and from cell cultures infected with two serotypes of epizootic hemorrhagic disease of deer (EHDV) yielded negative results. Preliminary experiments to quantify the threshold of sensitivity of this protocol, as described here, indicate positive detection of the BTV target RNA sequence at a level of less than 2 fg, or the amount of target sequence derived from less than 7500 viral particles. While this falls short of the theoretical limit of sensitivity of the PCR, the enhanced threshold of sensitivity of the PCR reaction compared to standard nucleic acid hybridization methodology suggests that rapid detection of BTV RNA directly within clinical specimens may be feasible with further refinements. Potential methods of enhancing this reaction are discussed.