Data in the present study came from two separate beef cattle breeding projects at Iowa State University. Data set-I included growth and carcass information of progeny in the small, medium, and large lines of synthetic cattle. Progeny were bom during the years 1978 through 1990 at Rhodes and McNay research farms. Data were used to evaluate selection practices in the three synthetic lines, to evaluate effects of some crossbreeding parameters on carcass traits, and to estimate genetic parameters and genetic trends for carcass traits. Data set-II included carcass and serially measured live-animal traits collected over a 6-year period (1991-1996). Most of the data came from progeny of purebred Angus and Simmental sires with known expected progeny difference and synthetic females from a previous project. Data set-II was used to study effects of sex and breed on growth and composition of feedlot cattle and to determine the best strategy to adjust serially measured traits to a constant age end point. The overall mean generation interval was 4.11 years. When averaged by line, 1.82, 1.47, and 1.28 generations of selection was made in the small, medium, and large lines, respectively. Mean actual sire index differentials per generation were, 1.28,-.47, and .84o for the small, medium, and large lines, respectively. There was a significant (P < .05) difference in direct additive effect between Jersey, Angus, and the Simmental breeds for most of the carcass traits considered. However, differences in breed maternal, average individual heterosis, and average maternal heterosis were not different from zero (P > .10). Heritability of hot carcass weight, dressing percentage, longissimus muscle area, fat thickness, and percentage of kidney, pelvic, and heart fat in the small line were, .30, .09, .21, .34, and .15, respectively. The respective values in the medium line were, .52, .35, .33, .29, and .07. Heritability values in the large line were in the order of .31,. 18,. 17, .31, and .18, vii respectively. Sire selection based on weaning indices showed a significaat (P < .05) genetic change for some of the carcass traits. It was concluded that index equations designed to improve beef carcasses need to incorporate carcass information in an index. Analysis of serially measured fat thickness, longissimus muscle area, body weight, hip height, and ultrasound percentage intramuscular fat showed a limitation in the use of growth models based on pooled data. Therefore, it was concluded that regression parameters from a withinanimal regression of a serially measured trait on age, averaged by sex and breed, are the best choice in describing growth and adjusting data to a constant age end point. at Rhodes were, 1.28,-.57 and .92