Effects of seven breeds of cow's sire and 12 breeds of cow's maternal grandsire on preweaning performance of crossbred cows and their calves were examined in data from two experiments conducted at the Roman L. Hruska U.S. Mean Animal Research Center. Data included 1,836 records over three to five parities for 516 cows by 143 sires and by 307 maternal grandsires. The statistical model fitted effects of calf sex, parity, cow birth-breeding year or cow-calf birth year, the breed effects and their interactions. Deviations of breed of sire or equivalent grandsire effects on each trait from the mean for Hereford x Angus cows ranged from -1.6 to 5.5 kg (P less than .001) for calf birth weights, -15 to 1% (P less than .001) for calving difficulty, nonsignificant for preweaning calf mortality and -2 to 27 kg (P less than .001) for calf weaning weight. Deviations were nonsignificant for conception rate and calves weaned per cow exposed to breeding, but -2 to 40 kg (P less than .001) for calf weight weaned per cow exposed for breeding, -7 to 78 kg (P less than .001) for cow weight and -20 to 2% (P less than .001) for body condition score. The advantages of Holstein and Brahman cross over Hereford x Angus cows of 23 and 13% in weight of calf weaned/cow-breeding exposure must be compared with the expected greater feed requirements from 7 or 8% heavier cows and at least 50% higher milk production, which emphasizes the need to include input measures and costs in breed evaluation schemes.