Body compositions of a randombred control population (RBC2), a line selected for increased 16-wk BW (F), a line selected for increased shank diameter (FL), and a line selected for increased leg muscle mass (FM) were compared at 16 wk of age. The F line originated from the RBC2 population and the FL and FM lines were developed as sublines of the F line. The F and FL lines were developed through mass selection for BW and shank width, respectively. The FM line was developed through family selection for leg muscle mass. Selection for BW in the F line has resulted in an increase in the proportion of breast muscle and a decrease in the relative amount of leg bone in comparison with the RBC2 population. Selection for increased shank width in the FL line has increased not only shank weight but has produced concomitant increases in the weight of the tibiotarsal and femur bones. Although the FL line has also shown a correlated increase in BW in association with increases in shank width, the increase in leg bone weight was relatively greater than the increase in BW. Therefore, the relative weights of leg bones in the FL line increased such that they were more similar to the RBC2 population than to the F line. Although the amount of leg muscle increased slightly in the FM line, the relative proportion of leg muscle did not increase significantly. The selection program used was probably ineffective in increasing the proportion of leg muscle because of the high percentage (low selection intensity) of families selected and small family size for estimating family means. There were no differences in the walking ability of F, FL, and FM lines. These selected lines did have poorer walking ability than RBC2 as indicated by higher leg scores for these lines.