Elevated blood glucose levels are the hallmark of type 2 diabetes as well as a powerful risk factor for development of the disease. We conducted a genome-wide search for diabetes-related genes, using measures of glycemia as quantitative traits in 330 pedigrees from the Framingham Heart Study. Of 3,799 attendees at the 5th Offspring Study exam cycle (1991--1995), 1,461, 1,251, and 771 men (49%) and women provided information on levels of 20-year mean fasting glucose, current fasting glucose, and HbA(1c), respectively, and 1,308 contributed genotype data (using 401 microsatellite markers with an average spacing of 10 cM). Levels of glycemic traits were adjusted for age, cigarette smoking, alcohol and estrogen use, physical activity, and BMI. We ranked standardized residuals from these models, created normalized deviates from the ranks, and used the variance component model implemented in SOLAR (Sequential Oligogenic Linkage Analysis Routines) to evaluate linkage to normalized deviates as quantitative traits. We found peak evidence for linkage to 20-year mean fasting glucose levels on chromosome 1 at approximately 247 cM from p-telomere (pter) (multipoint logarithm of odds [LOD] 2.33) and on chromosome 10 at approximately 86 cM from pter (multipoint LOD 2.07); to current fasting glucose levels on chromosome 1 at approximately 218 cM from pter (multipoint LOD 1.80) and on chromosome 10 at approximately 96 cM from pter (multipoint LOD 2.15); and to HbA(1c) levels on chromosome 1 at approximately 187 cM (multipoint LOD 2.81). This analysis of unselected European Caucasian pedigrees suggests localization of quantitative trait loci influencing glucose homeostasis on chromosomes 1q and 10q. Findings at approximately 187--218 cM on chromosome 1 appear to replicate linkage reported in previous studies of other populations, pointing to this large chromosomal region as worthy of more detailed scrutiny in the search for type 2 diabetes susceptibility genes.