| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
New Zealand Dairy Board, Wellington
ABSTRACT
Three different approaches were used to estimate repeatability and, correspondingly, three different interpretations are placed on the results. In the first analysis, where a herd variance component was taken into account, year effects were considered to be random, and found to be negligible—the herd-by-year effects absorbing all the year variations. In this analysis, repeatability (0.65) is applicable to any cow, relative to an over-all average for the whole country, and as such is not very useful.
In the first of the within-herds analyses, years were treated as mere replications and repeatability (0.49) is relative to a general herd level, regardless of year.
In the final analysis, year effects were eliminated, and repeatability (0.61) then applies to the deviation of a cow's records from her herd average in the same year. The increase from 0.49 to 0.61, when yearly effects are taken into account, indicates the importance of the within-herd, year-to-year variation, when considering repeatability estimates.
Within-herd estimates of repeatability obtained here are higher than earlier estimates, probably due to the fact that only four years are represented. It is probable that the decrease in correlations as the time interval increases (shown in Table 1), is due to long-term effects which are not likely to be directly associated with a cow's genetic make-up for butterfat production. Therefore, in using repeatability to estimate the "true producing ability" of a cow in a particular herd, there is valid reason for using the higher values based on the short-term analysis.
1 Present address: Dept. of Animal Husbandry, Cornell University, Ithaca, N. Y.
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
