Threshold model analysis of dystocia in dairy cattle when progeny information is limited

Progeny of 374 sires were sampled from the National Association of Animal Breeders data base to study the effect of analysis procedures and of limited progeny information on predictors of sire merit for calving ease. Situations of limited progeny information were achieved by evaluating sires by using their first 50 progeny, a random sample of 50 progeny, and a random sample of 500 progeny. Sires were evaluated by threshold model and BLUP procedures. The designed situations of limited progeny information did not result in a clear difference between procedures. Both procedures had a rank correlation of .57 between sire solutions from the first 50 and 500 progeny; the rank correlation between evaluations based on random 50 and 500 progeny was .70 and .69 for the threshold model and BLUP, respectively. Major changes in rank from random 50 to 500 progeny tended to occur among those sires with few or no calving records from matings to heifers in the random sample of 50 progeny. The BLUP sire solutions were partitioned, and a positive correlation between 500 progeny sire solutions and the partition for parity of dam indicated that easy calving progeny had sires that tended to be mated to heifers and that difficult calving progeny had sires that were mated to cows in later parities.