We developed and utilized a multipoint variance components method to test for linkage between a disease trait and markers on chromosome 5 in the simulated data provided in GAW10 Problem 2. We demonstrated that the discrete trait variance components method recovers unbiased estimates of quantitative trait locus (QTL) location and reasonable estimates of effect size. We also showed that dichotomization of (a continuous trait such as) Q1 diminished the power to detect linkage compared to direct analysis of Q1, and that extended pedigree analyses provided superior power to detect linkage compared to those in nuclear families.