In this paper we determine the limits and accuracy of quantitative reverse transcription (RT)-PCR using a modification of the original protocol. The quantification of mRNA with this procedure requires a preliminary estimation of the target molecule (TM) concentration, established from experiments with an internal control molecule (ICM). A definitive quantification is then attained from serial dilutions of the reverse transcription reaction. The success of this latter step is dependent on maintaining an equivalent number of TM and ICM in the reaction. The purpose of our study was to evaluate the influence of the deviation between the TM and the ICM on the result. We show here that we can control the accuracy of the assay by fixing the limit of the TM/ICM ratio. Indeed, when the TM/ICM ratio is between 0.66 and 1.5 (i.e., the difference between TM and ICM is 1.5-fold), the final result has an error of approximately 10%. Exceeding this limit produces errors approaching 60%, as in the case of TM/ICM = 2. When the above conditions are respected, a difference as small as 20% between two samples can be determined with an accuracy of 95%.