A number of recent linkage studies have suggested the presence of a schizophrenia susceptibility locus on chromosome 6p. We evaluated 28 genetic markers, spanning chromosome 6, for linkage to schizophrenia in 10 moderately large Canadian families of Celtic ancestry. Parametric analyses of these families under autosomal dominant and recessive models, using broad and narrow definitions of schizophrenia, produced no significant evidence for linkage. A sib-pair analysis using categorical disease definitions also failed to produce significant evidence for linkage. We then conducted a separate sibpair analysis using scores on positive-symptom (psychotic), negative-symptom (deficit), and general psychopathology-symptom scales as quantitative traits. With the positive symptom-scale scores, the marker D6S1960 produced P = 1.2 x 10(-5) under two-point and P = 5.4 x 10(-6) under multipoint analyses. Using simulation studies, we determined that these nominal P values correspond to empirical P values of .034 and .0085, respectively. These results suggest that a schizophrenia susceptibility locus on chromosome 6p may be related to the severity of psychotic symptoms. Assessment of behavioral quantitative traits may provide increased power over categorical phenotype assignment for detection of linkage in complex psychiatric disorders.