We describe the laboratory-related aspects of a series of 40 completed presymptomatic tests for Huntington's disease, using linked DNA markers. Pedigree structure and marker heterozygosity are shown to be important factors, both in the number of laboratory analyses required to give an informative situation and the residual uncertainty of the final estimate. Specific problems encountered by the testing laboratory are described, with possible ways of avoiding them, and the close links required between laboratory and clinical staff are emphasised.