Polled Hereford sires (n = 47) were divergently selected on published yearling weight (YW) and maternal (MAT) expected progeny differences (EPD) and mated to grade Angus cows to produce 457 calves in five spring calf crops. Sires selected for high and low YW differed by an average of 6.3 kg in YW EPD and those selected for high and low MAT differed by an average of 4.0 kg in MAT EPD based on 1989 EPD values. Calves by high-YW sires were heavier at birth (2.2 kg; P less than .10) and weaning (7.5 kg; P less than .01) and as yearlings (16.4 kg; P less than .01) than calves by low-YW sires and were taller at weaning (1.90 cm; P less than .01). Regressions of calf performance on corresponding 1989 EPD were 1.18 +/- .20 kg/kg for birth weight, .75 +/- .24 kg/kg for weaning weight and 1.79 +/- .42 kg/kg for YW. Expected progeny differences for individual sires were calculated from the data collected in this study and had correlations with published EPD of .53 for birth weight, .37 for weaning weight and .54 for YW. These corresponded to expected correlations based on accuracies of evaluation of .68, .61 and .58, respectively, and yielded estimates of the genetic correlation between performance in the environment of the study and the environment of the purebred herds where the published EPD were derived of .78 for birth weight, .61 for weaning weight and .93 for YW. The very large regression of YW on YW EPD (1.79 +/- .42 kg/kg) may have resulted from bias in published EPD due to culling of calves at weaning in purebred herds. Use of multiple traits analyses to account for such culling is recommended.