Among the experimental techniques available to study the genetic variability of RNA virus populations, the most informative involve reverse transcription (RT), amplification, cloning and sequencing. The effects of several aspects of these techniques on the estimation of genetic variability in a virus population were analysed. Hepatitis C virus populations from four patients were examined. For each patient, ten series of data derived from independent PCR amplifications of a single RT reaction were obtained. The sample size of each data set was 10 sequences (in nine series) and 100 sequences (in one series). An additional data set derived from an independent RT reaction (about 10 sequences) performed on RNA extracted from the same serum sample was also analysed. The availability of data sets of different sample sizes allowed the effect of sample size on the amount and nature of the genetic variability recovered to be examined. The repeatability of the data obtained in different amplification experiments as well as from different RT reactions was also determined, together with the best strategy to obtain a given number of sequences by comparing the set of 100 sequences obtained from a single amplification with those obtained by pooling the nine sets of 10 sequences. In all cases, these results confirm the high repeatability of the conclusions and parameters derived from the sets of 10 sequences. These results validate the use of relatively small sample sets for the evaluation of genetic variability and for the estimation of phylogenetic relationships of RNA viruses in population and epidemiological studies.