A number of reports have indicated that RNA recovered from paraffin-embedded tissue can be used as a substrate in the polymerase chain reaction (PCR). Although it is established that RNA in paraffin-embedded tissue undergoes significant degradation, the specific contributions of different fixatives and fixation times to this degradation are not known. Mouse splenic tissue was harvested and fixed immediately for 2, 8, or 24 h in either formalin, Omnifix II, or Carnoy's fixative and then processed and embedded in paraffin. RNA was extracted from deparaffinized cubes of tissue using an adaptation of the technique described by Chomczynski and Sacchi. RNA was reverse transcribed using a random hexamer primed reaction. PCR amplification for cDNAs of the housekeeping genes glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and hypoxanthine phosphoribosyltransferase (HPRT) mRNAs was then performed. Although GAPDH amplification is used routinely on fresh and frozen tissues, we show that the presence of DNA contamination in the RNA preparations limits its usefulness in paraffin-embedded tissue. Amplifiable HPRT mRNA sequences were detected in nine of 12 samples fixed in Omnifix II, in four of 12 samples fixed in Carnoy's fixative, and in none of 12 formalin-fixed samples. Because of primer selection to preclude amplification of genomic HPRT, DNA contamination is not an issue when HPRT is amplified. Thus, HPRT represents the control system of choice for the evaluation of RNA in PET. The techniques described provide a rapid, uniform, and reproducible method of obtaining RNA from PET for molecular analysis, but they indicate limited utility for retrospective analysis of archival tissues.