Nucleotide sequence studies detected a double-point mutation in the genomic RNA segment 4 (nt 871 and 872) of the persistent variant C/AA-pi of influenza C/Ann Arbor/1/50 virus. The 3'-end-points of two distinct PCR primers were positioned exactly at this genome location and thereby adjusted the priming determinant complementary to the varied strain or to its wild-type counterpart. Consequently, positive RT-PCR products strictly referred to one of the two viruses examined, in both cases, using either virion or infected-cell RNA templates. Artificial virus mixtures could easily be distinguished by this method in a subsequent qualitative gel analysis. PCR annealing conditions and control reactions were optimized, for the monitoring of influenza virus isolates throughout multifold passages. Thus, sequence diversity in just two neighbouring nucleotides is sufficient to determine whether or not successful PCR amplification takes place, and this method can be used as a reliable means of virus strain differentiation.