Recent studies describe families with both elevated plasma HRG levels and thrombosis. In order to study the possibility that allelic variants of the HRG locus are associated with differences in HRG level, we studied linkage between HRG levels and a dinucleotide repeat polymorphism in a Dutch family which was selected on the presence of both thrombosis and elevated plasma HRG levels. No other known risk factors from thrombosis were found in this family. Linkage was calculated between the dinucleotide repeat and the HRG level considering the HRG level as a quantitative phenotype assuming a population prevalence of elevated HRG of 5%. Two classes of HRG levels were defined by a mean and a variance: one class with normal HRG levels and a second class with high HRG levels. Using a mean HRG level of 99% for individuals with a normal HRG level and 145% for individuals with high HRG, a maximum lod score of 4.17 (odds in favour of linkage of 22 000:1) was found at a recombination fraction of 0, indicating linkage. Considering the pedigree, an association was found between the presence of a specific allele (no. 6) of the dinucleotide repeat polymorphism and plasma HRG levels. Family members carrying allele 6 were found to have higher HRG plasma levels compared with family members lacking allele 6 (149% v 109% respectively). We conclude that in this family, linkage is found between the HRG locus and the HRG level, and that a HRG gene coupled to allele 6 of the dinucleotide polymorphism is associated with elevated plasma HRG levels. No evidence was found for a causal relationship between elevated plasma HRG levels and thrombosis in this family. Copyright © 1995, Wiley Blackwell. All rights reserved