Four hundred fifty-two yearling steers from two experiments were measured for subcutaneous fat thickness and longissimus muscle area between the 12th and 13th ribs using real-time linear array ultrasound equipment. Ultrasonic predictions were compared to corresponding carcass measurements to determine accuracy of ultrasound measurements. In Exp. 1, 74% of the ultrasonic estimates of fat thickness were within 2.54 mm of carcass values (r = .81) and muscle area was predicted within 6.45 cm2 for 47% of all carcasses (r = .43). Although similar correlation coefficients between ultrasonic and carcass fat thickness were obtained in Exp. 2 (r = .82), estimates were more biased; only 62% of ultrasound estimates were within 2.54 mm of carcass measurements. Improvement in longissimus muscle area estimates was noted in Exp. 2, in which 54% of ultrasonic estimates were within 6.45 cm2 of carcass values (r = .63). The extremes for each trait proved most difficult to predict; fat thickness was underestimated on fatter cattle and muscle area was underpredicted on more heavily muscled steers. Ultrasonic measurements of fat thickness are precise and accurate in determining carcass fat thickness, but muscle area estimates are inconsistent and warrant further investigation.