Author
 |
POWELL, REX |
|
Submitted to: Interbull Annual Meeting Proceedings
Publication Type: Proceedings Publication Acceptance Date: 9/7/1994 Publication Date: N/A Citation: N/A Interpretive Summary: Breeders have struggled to make reasonable decisions from an array of genetic information. Economic indexes are promoted as the solution and lead to proper choices if the index weights are correct for the economic situation. Relationships between national economic indexes for yield data were documented for Canada, Denmark, France, Germany, Italy, The Netherlands, and the U.S.; the impact of using different indexes was examined using international evaluations. Indexes other than from the U.S. either did not include milk or gave it a negative weight, which had the most apparent consequences for comparisons with Italy and The Netherlands. Correlations between indexes were generally high except for correlations for the U.S. with Italy and The Netherlands. The top 100 bulls for each index were identified by country. For each country, differences in numbers of top bulls among indexes were sizeable. Even if numbers of top bulls for a country were similar for different indexes, many different bulls sometimes were included among groups of top bulls. The impact of selection based on the various indexes was examined through a measure of loss from selection on another index. If the correlation between countries was high, loss from selection on another index was small. However, as the correlation decreased, loss increased. The four largest losses from selection on another index involved comparisons of the U.S. with Italy and The Netherlands. These results stress the importance of dairy breeders' using the proper index, which differs from farm to farm as well as from country to country. Progress may depend more on the choice of an index for a country or an individual breeder than on the replacement of national with international evaluations. Technical Abstract: Relationships between national economic indexes for yield data were documented for Canada, Denmark, France, Germany, Italy, The Netherlands, and the U.S., and impact of using different indexes was examined using February 1995 International Bull Evaluation Service evaluations for 48,962 Holstein bulls. Bulls were required to have evaluations for milk, fat, and protein yields and birth years from 1950 through 1990. Indexes were applied to evaluations on a U.S. evaluation basis, and correlations were computed from residuals after fitting birth year. Correlations were generally high (.9 to nearly 1.0) except for correlations for the U.S. with Italy and The Netherlands. Correlations among indexes from Canada, Denmark, and Germany were above .99. Correlations of indexes with evaluations for protein yield on a U.S. evaluation basis ranged from .885 to .987. The top 100 bulls for each index were identified by country with the most reported daughters. For each country, differences in numbers of top bulls among indexes were sizeable. Even if numbers of top bulls for a country were similar for different indexes, many different bulls sometimes were included among groups of top bulls. Difference in mean index for the top 100 bulls selected on one index versus other indexes was calculated for the 14,277 bulls born in 1986 or later and expressed in units of standard deviation (SD) to produce a measure of loss from selection on another index. If the correlation between countries was high, loss from selection on another index was small. However, if the correlation was about .9, loss was .2 to .3 SD. The four largest losses from selection on another index ranged from .4 to .7 SD and involved comparisons of the U.S. with Italy and The Netherlands. |
