Comparison of several sire evaluation methods in dairy cattle breeding

The objectives are discussed of sire evaluation where AI is used but herd sizes are small. In every evaluation scheme, the question arises as to what kind of information should be used, apart from daughter records. Further, we have to consider whether it is worthwhile to wait for the second lactation of the daughters. In sire evaluation, we are nearly always concerned with several traits, and the usual way of getting the aggregate genetic-economic value of an animal is by the application of an index ignoring any correlations between traits. Though that is not optimal, the loss in efficiency is unlikely to be large. Another important point is the determination of the appropriate classification within which the comparison should be made, e.g. should we use herd-year-seasons or herd-years or even group the herds by mean yields. The herd mate method, the contemporary comparison method, the modified cumulative difference method, least squares, regressed least squares and BLUP are discussed, and their relationships to each other are pointed out.