Abstract
SUMMARY A numerical model of AI selection procedures has been constructed in an attempt to assess the efficiency of current selection methods and to examine possibilities for their improvement. The results indicate that (1) present selection methods are relatively efficient in their utilization of test resources, but much less efficient in their usage of tested bulls. Additional investments in bull testing are likely to profit the dairy industry. (2) The main effects of improved bull utilization are (a) to permit increases in selection pressure with no increase in cost over current selection procedures and (b) to extend the limits to profitable selection. (3) The adoption of alternative stud management systems, in which bulls are slaughtered early in life, are likely to require excessively large investments in semen processing and storage.