Introduction Short lactations are frequently excluded from the analyses of milk records of tropical cattle. The practice leads to overestimates of milk yield and lactation length and decreases the accuracy of selection procedures. The objectives of this paper are to quantify lactation length using unselected records from a sample of Venezuelan dual‐purpose cows, evaluate some of the sources of variation and discuss suitable editing procedures for milk production records in these populations.

Summary The purpose of the study was to determine the relation between Holstein bulls’ proofs for milk in USA, and the survival and body weights of their F1 progeny under lowland grazing conditions in Venezuela. Semen was selected from sires with the highest proofs available locally (class H, n = 23) or with proofs close to the breed average (class A, n = 18). Mean proofs were 934 kg (H) and 237 kg (A), respectively, when semen was purchased (1989–93), falling to 204 and −386 kg when updated in 2001. Dams were Brahman‐type zebus on seven farms, raising the F1 progeny (n = 866) on native or cultivated pasture with only mineral supplements. Records of survival to 8 and 18 months and weights at birth, 8 and 18 months of age were analysed using linear models. Sire class had no effect on survival or weight at birth, but H progeny were slightly heavier than A progeny at 8 (4.0%, p = 0.06) and 18 months (2.6%, p < 0.05). The interaction class × farm was only significant for survival, with H progeny equal or superior to A groups on six farms. The effect of sire within class was highly significant for survival to 8 months and weight at 18 months. Rank correlations using sires’ milk proofs and F1 progeny weights and survival were not significant. It was concluded that the use of Holstein bulls with high milk proofs may lead to slightly higher 18‐month weights in their F1 progeny in tropical grazing systems, covering the additional cost of H semen, but that sire evaluation for weight gain and progeny survival rates may be justified under these conditions.

Summary To determine the effect of USA Holstein bulls’ breeding value for milk yield on the performance of their female F1 progeny in the lowland tropics of Venezuela, semen was chosen from 39 sires, representing bulls with the highest (class H, n = 21) or average (class A, n = 18) proofs for milk, available locally. The mean proofs of the H and A sires were 937 and 237 kg milk, respectively, when semen was purchased in 1989–93, and 212 and −386 kg when updated in 2001. Progeny from commercial zebu dams were kept grazing on three dual purpose farms, where herd average milk yields were 735, 1113 and 3524 kg per lactation. Records on milk yield, lactation length, calving interval, age at first calving and survival rate to first, second and third calving from 222 progeny (476 lactations) were analysed using mixed linear models. The effect of sire class approached significance (p = 0.07) for milk yield and survival rate to second calving, with H daughters superior to A progeny by 41 and 46%, respectively, and no significant disadvantage for any trait. Class × farm interaction was only significant for age at first calving. Rank correlations using USA proofs for milk and the adjusted mean values for bulls with =15 F1 records were 0.72 (p < 0.01, milk yield) and −0.07 (calving interval). Differences between sires, within classes, were highly significant for age at first calving and calving interval. It was concluded that the use of Holstein bulls with high USA milk proofs is probably justified for producing F1 daughters, but that evaluation for growth rate and daughter fertility under tropical grazing conditions may be advisable.