We have developed a heminested PCR (polymerase chain reaction) method, performed on single cells, for the analysis of the most common cystic fibrosis (CF) mutation (delta F508). As a quality control, the polymorphic exon 2 of the HLA DQA1 locus was co-amplified from the same cell. With a non-radioactive reverse dot-blot assay, the genotype of these two loci could be determined. Experiments on 98 single fibroblasts, heterozygous for the CFTR and the DQA1 locus, showed that amplification of either locus could be obtained in 97 per cent of the cases, but only 90 per cent showed heterozygosity for CF, 75 per cent showed heterozygosity for DQA1, and 74 per cent showed heterozygosity for both CF and DQA1. Contaminations detected only after DQA1 typing occurred in 3 per cent of our samples. Error rate calculations based on our experimental PCR data indicate that single blastomere diagnosis would lead to unacceptable errors, i.e., an affected fetus, in less than 1 per cent of the cases. The risk of undetected crossing-over or the dubious results that crossing-over could generate, would make isolated polar body diagnosis at the present time very difficult. The combined approach of PCR on polar bodies followed by confirmation of the diagnosis on blastomeres, however, should give a solid base for preimplantation diagnosis of monogenic disorders.