We have used the Polymerase Chain Reaction (PCR) to amplify a 798bp fragment of the gene for human Apolipoprotein B (Apo 8). that contains sequences coding for the putative LDI-receptor binding domain. 5pg genomic DNA from 10 individuals was amplified1 using 3Omer oligonucleotides spanning bases 9599-10397 (inclusive) of the Apo 8 gene. 30 rounds of amplification were carried out using 5U of Taq Polymerase (Anglian Biotech.) per sample, in a buffer containing : 67mM Tris-HC]. (pH 8.8), 6.7mM MgC.12,16.7mM (NH4)2SO4, 10mM 0 mercaptoethanol, 6.7pM disodium FDTA, 4mg/mi BSA, 10Z Dimethylsulphoxide and 330pM (each) dATP, dCTP, dGTP, dTTP, under the regime: 2mins a 950C, 1min a 550C, 5mins a 700C. Amplified DNA was digested with EcoRi and Scal (Anglian Biotech.) and force-cloned into FroRT/Smal cut M13 mplO (Amersham) using standard techniques. At least 10 clones from each subject were purified. Clones were sequenced using the Sequenase kit (USR Jnc.) and analysed on 8Z denaturing polyacryamide gels. Initially the sequence of one clone per individual was determined.Out of the total of 8000 bases sequenced (10 individuals), 22 differences were detected (Table). No clone was identical to the published sequences2 Since any genuine base change should be present in approximately half the clones analysed (assuming the individual to be heterozygous), we subsequently analysed all 10 clones from each individual. None of the initial differences found were present in any of the other clones, although all of them were reproduced upon resequencing of the original clones. rhis implies that all the base differences seen were artefacts generated by the PCR. The most common changes found were from A to 6 and from T to C. 17/22 (77Z) of the changes noted were associated with a run of bases of the same sequence (Table). This may be an indication of the mechanism by which the