Genetic loci for body weight and subphenotypes such as fat weight have been mapped repeatedly. However, the distinct effects of different loci and physiological interactions among different traits are often not accounted for in mapping studies. Here we used the method of structural equation modeling to identify the specific relationships between genetic loci and different phenotypes influencing body weight. Using this technique, we were able to distinguish genetic loci that affect adiposity from those that affect muscle growth. We examined the high body weight-selected mouse lines NMRI8 and DU6i and the intercross populations NMRI8 x DBA/2 and DU6i x DBA/2. Structural models help us understand whether genetic factors affect lean mass and fat mass pleiotropically or nonpleiotropically. Sex has direct effects on both fat and muscle weight but also influences fat weight indirectly via muscle weight. Three genetic loci identified in these two crosses showed exclusive effects on fat deposition, and five loci contributed exclusively to muscle weight. Two additional loci showed pleiotropic effects on fat and muscle weight, with one locus acting in both crosses. Fat weight and muscle weight were influenced by epistatic effects. We provide evidence that significant fat loci in strains selected for body weight contribute to fat weight both directly and indirectly via the influence on lean weight. These results shed new light on the action of genes in quantitative trait locus regions potentially influencing muscle and fat mass and thus controlling body weight as a composite trait.