Ultrasound technology allows in vivo estimation of carcass composition. Successful genetic evaluation of ultrasonic measures depends upon technician certification guidelines and a viable common-endpoint adjustment strategy for field data. Four technicians and three image interpreters ultrasonically evaluated 172 lambs to determine accuracy and repeatability of loin eye area (LEA), backfat thickness (BF), and body wall thickness (BW) estimations. Correlations between ultrasonic and carcass measurements were 0.66, 0.78, and 0.73 for LEA, BF, and BW, respectively. Performance was similar among technicians and interpreters. Mean bias ranged from -1.30 to -2.66 cm2, -0.12 to -0.17 cm, and 0.14 to -0.03 cm, for LEA, BF, and BW, respectively; prediction standard errors ranged from 1.86 to 2.22 cm2, 0.12 to 0.14 cm, and 0.35 to 0.38 cm, respectively. Repeatability standard errors ranged from 1.61 to 2.45 cm2, 0.07 to 0.11 cm, and 0.36 to 0.42 cm for LEA, BF, and BW, respectively. Changes in ultrasonic measurements were evaluated using seven serial scans on 24 growing Suffolk ram lambs. All equations had similar goodness of fit. Equations were tested on other populations, including similarly-managed rams across breeds and years and ewe lambs fed for slower gain. Correlations between predicted and actual measures ranged from 0.78 to 0.87 for BF and 0.66 to 0.93 for LEA in winter-born rams, were only slightly lower in fall-born rams, and ranged from 0.72 to 0.74 for BF and 0.54 to 0.76 for LEA in ewe lambs. Of the equations tested, linear and allometric forms appear best for general use. / Master of Science