The applicability of ribonuclease cleavage at mismatches in RNA:DNA duplexes (RNase cleavage method) for determining nucleotide variant rates has been examined in a Japanese population. DNA segments of various lengths obtained from 4 different regions of a normal and 3 thalassemic cloned human beta-globin genes were inserted into transcription vectors. Sense and antisense RNA probes uniformly labeled with 32P were prepared. When RNA probes of 771 nucleotides (nt) or less were hybridized with cloned DNAs and the resulting duplexes were treated with a mixture of RNases A and T1, the length of products agreed with theoretical values. Twelve possible mismatches were examined. Since both sense and antisense probes were used, uncleavable mismatches such as G:T and G:G which were made from one combination of RNA and DNA strands could be converted to the cleavable C:A and C:C mismatches, respectively, by using the opposite combination. Deletions and insertions of 1 (G), 4 (TTCT), 5 (ATTTT) and 10 (ATTTTATTTT) nt were easily detected. A polymorphic substitution of T to C at position 666 of the second intervening sequence (IVS2-666) of the beta-globin gene was detected using genomic DNAs from cell lines established from the peripheral B lymphocytes of 59 unrelated Japanese from Hiroshima or those amplified by polymerase chain reaction (PCR). The frequency of the gene with C at the IVS2-666 (allele C) was 0.48 and that of the gene with T (allele T) was 0.52. The associations of the 2 alleles were in agreement with Hardy-Weinberg proportions. No contradiction to Mendelian inheritance was observed in the results obtained from 11 family studies. Two new polymorphic substitutions of C to A and A to T were detected at nucleotide positions 1789 and 1945 from the capping site, respectively, using genomic DNAs amplified by PCR. The feasibility of the RNase cleavage method combined with PCR for large-scale screening of variation in chromosomal DNA is discussed.