Two selection herds with Angus cattle were maintained at Waikeria near Te Awamutu, New Zealand, from 1964 to 1981, with experimental calvings comprising 3355 recorded calves. The selection criterion was yearling weight in one herd (the YW herd) and postweaning weight gain in the other (the PWG herd). In the absence of a control herd, genetic responses were estimated using best linear unbiased prediction (BLUP) methods. Annual direct selection differentials averaged 6.45 kg in the YW herd and 5.00 kg in the PWG herd, respectively 0.90 and 0.79 phenotypic standard deviations of selection per generation. The average generation intervals were 3.23 and 2.88 years respectively. Direct responses were estimated to be 1.96 and 0.97 kg/year respectively (both P < 0.01), whereas correlated responses in yearling weight (pWG herd) were 1.52 kg/year and in postweaning gain (YW herd) were 0.68 kg/year. Significant increases were obtained in birth and weaning weights in the YW herd but not in the PWG herd. Paternal half-sib heritabilities ofdirect selection traits were estimated from the selection herds at 0.34 for yearling weight and 0.22 for postweaning gain, using a technique including a numerator relationship matrix. *Deceased.Estimates from 56 of the selection herd bulls used in a progeny test herd were 0.36 and 0.26 respectively. The genetic response calculated from BLUP analyses of yearling weight was less than the preliminary estimate obtained in another selection experiment using Angus cattle at Waikite,New Zealand, carried out with a contemporary control herd.