Neurofibromatosis type-1 (NF-1) is an autosomal dominant disorder, caused by mutations in the NF-1 gene. Mutation analysis in the NF-1 gene is complicated by the large size of the gene, the high mutation rate, and the presence of pseudogenes. By means of the polymerase chain reaction, we have amplified 70% of the NF-1 coding sequence using reverse transcribed mRNA and genomic DNA from 25 unrelated Scottish Caucasian patients. We have used chemical mismatch cleavage analysis and direct sequencing of asymmetrically amplifed PCR products to characterise mutations within the NF-1 gene. Using the above strategy, we detected 10 novel mutations and an intragenic polymorphism with a heterozygosity of approximately 47% in the Scottish population. Of the 10 mutations, 7 are potentially disease causing. They include splice site errors responsible for exon skipping (1721 + 3A to G) and (5749 + 2T to G), small insertions (7485insGG) and (6519insG), a nonsense mutation (R2496X), and missense and silent mutations (G1166D, K1419R, G1404G, S1311S, N1776N). A correlation of the phenotype with the genotype is presented. Thus, in this study we have identified a heterogeneous group of germline mutations, the majority of which are predicted to cause disruption of the protein product, neurofibromin. This approach has therefore proved to be useful for the detection of mutations in the gene for neurofibromatosis type-1, and can be applied to detection of molecular pathologies in general.