In this study, the effects of DNase treatment on the specificity of reverse transcription (RT)-PCR have been investigated on different samples, containing RNA, DNA, DNA/RNA and DNA/cDNA. This was to evaluate the possibilities of getting specific results in a direct in situ RT-PCR. All DNA targets in the samples were eliminated after 1 h of DNase treatment. However, some DNA fragments still remained after both 1 h and overnight DNase treatment. When these fragments served as primers for amplification, nonspecific smears resulted. In samples containing small amounts of RNA, the RNA was affected by overnight treatment with DNase. Our conclusion is that the direct in situ RT-PCR is an unreliable method because the necessary DNase treatment induces nonspecific amplification, and no size-separation is possible. We conclude that the best way to perform an in situ RT-PCR is to hybridize with a labeled specific probe after amplification is completed.