AbstractThis presentation attempts to clarify certain caveats that apply to the use of linkage analysis when the genetics of one of the characters is not completely known. The effects of linkage and association can be difficult to distinguish both within pedigrees and across the general population. Linkage analysis methods are not robust to departures from the assumptions of the model. These assumptions do not allow for the possibility that association may be present which is due to both genetic systems having an effect on one of the phenotypes. Analysis directed towards detecting the presence of such association may be affected by stratification when unrelated individuals are analysed and is confounded by linkage if this is present when related individuals are analysed. Testing for and estimating linkage simultaneously with association using whole pedigree data is advocated and an outline of the required approach is given.