OBJECTIVE: To conduct a full genome search for genes potentially influencing two related phenotypes: body mass index (BMI, kg/m 2 ) and percent body fat (PBF) from bioelectric impedance in men and women. DESIGN: A total of 3383 participants, 1348 men and 2035 women; recruitment was initiated with hypertensive sibpairs and expanded to first-degree relatives in a multicenter study of hypertension genetics. MEASUREMENTS: Genotypes for 387 highly polymorphic markers spaced to provide a 10 cM map (CHLC-8) were generated by the NHLBI Mammalian Genotyping Service (Marshfield, WI, USA). Quantitative trait loci for obesity phenotypes, BMI and PBF, were examined with a variance components method using SOLAR, adjusting for hypertensive status, ethnicity, center, age, age 2 , sex, and age 2 Â sex. As we detected a significant genotype-by-sex interaction in initial models and because of the importance of sex effects in the expression of these phenotypes, models thereafter were stratified by sex. No genotype-byethnicity interactions were found. RESULTS: A QTL influencing PBF in women was detected on chromosome12q (12q24.3-12q24.32, maximum empirical LOD score ¼ 3.8); a QTL influencing this phenotype in men was found on chromosome 15q (15q25.3, maximum empirical LOD score ¼ 3.0). These QTLs were detected in African-American and white women (12q) and men (15q). QTLs influencing both BMI and PBF were found over a broad region on chromosome 3 in men. QTLs on chromosomes 3 and 12 were found in the combined sample of men and women, but with weaker significance. CONCLUSION: The locations with highest LOD scores have been previously reported for obesity phenotypes, indicating that at least two genomic regions influence obesity-related traits. Furthermore, our results indicate the importance of considering context-dependent effects in the search for obesity QTLs.