Associations have been reported between restriction fragment length polymorphisms (RFLPs) at fibrinogen loci and plasma fibrinogen concentration, in a British study. We have examined a series of unrelated Norwegians. We found no association between plasma fibrinogen concentration and any genotype in either of two fibrinogen polymorphisms examined (one at the alpha-fibrinogen locus, the other at the beta-fibrinogen locus). We have also examined monozygotic twins and evaluated heritability of fibrinogen level by the intraclass correlation coefficient. We arrived at an unimpressive estimate of heritability. With such a low level of heritability, it would have been surprising if we had found an association with a single gene marker in a relatively limited series of people. The reason for the discrepancy between the British and the Norwegian study is unknown. Great care has to be exercised in interpreting disease associations, since with DNA variations being examined at an increasing number of "candidate loci", the risk of finding spurious associations increases with the number of analyses conducted.