The Bull Index Problem in the Light of Modern Genetics

Abstract

The genetic principles affecting the validity of bull indexes are reviewed. All the indexes accurate enough to deserve further consideration start with the average production of the daughters as a basis. They differ only in the use and emphasis which they make of the difference between daughters and dams. Differences in herd environment affect all indexes since they enter into the daughter average which is the base of all indexes. They are less important in the equal parent index than in the daughter average if the correlation between herd environment of daughter and of dam is as high as .25. Differences in the average genetic merit of the cows to which the bull was mated are neglected in the daughter average but are discounted in the equal parent index. Errors due to random environment and to the part played by chance in inheritance are reduced by increasing the number of daughters tested. Errors due to dominance and to “nicking” are also thus reduced but not so effectively. Errors due to herd environment and to differences in the average merit of the dams are biased and do not trend toward zero as the number of daughters tested is increased. No index will give infallible results. The sources of error cannot be absolutely eliminated. All that can be done is to minimize them as much as possible. The more one knows about the conditions under which the records were made, the more he can reduce these errors. Even if all other possibilities of error were overcome, there would still remain the sampling error intrinsic in the Mendelian nature of inheritance. The writer recommends the equal parent index (the daughter average plus the increase of daughters over dams) as soundest in principle, simple enough for field use, freest from systematic error, and having a range not very different from that of the actual records of cows. © 1933, American Dairy Science Association. All rights reserved.