THE MULTIPLE (PART) LACTATION MODEL FOR T HE EVALUATION OF DAIRY-CATTLE .2. ASPECTS IN THE ESTIMATION OF BREEDING VALUES

Aim of the present study was to examine possible combinations of trait s fora multiple trait evaluation for dairy cattle. The estimation of b reeding values via the BLUP-method can be interpreted as applying a se lection index when Generalized-Least-Squares estimators of the subcell s are used for the calculation of deviations from contemporary average s. Therefore the superiority of a BLUP multiple trait prediction can n ot be evaluated using the classical theory of a selection index. A maj or aspect in this regard is the improvement of the data structure from additional relations through covariances between traits. This was exa mined using real field data. Based on the estimation of variance compo nents for the multiple (part)lactation model applied in Germany and on analyses using parts of the data from the routine evaluation possibil ities for a reduction of the number of traits were derived. Furthermor e, different economical weights for the single parts of lactations wer e examined. The analysis of the results from the routine evaluation of the RLN using data from 1400 A. I. bulls showed that no apparent diff erence exists between the inheritance of genes influencing the second and third lactation. The actual model comprising 5 (part)lactations wa s defined as a standard for comparisons with variants of this model co ntaining only three or four (part)lactations. From an estimation of br eeding values using field data the insignificance of the third lactati on for bulls and cows was confirmed by very similar rankings after an estimation of breeding values with and without consideration of the th ird lactation. The culling rates within the first lactation (5 % in pa rt two and 4 % in part three) are rather low so that a separation of t he first lactation into three parts seems to be unnecessary. However, a problem is caused by the current practice of using incomplete record s from part three of the first lactation for evaluation purposes witho ut correcting for the number of days in milk. The results show that th e third part of the lactation as currently defined without a loss of i nformation can be combined with the second part. This is still valid w hen a higher economic value is assumed for parts two and three as comp ared to part one of the first lactation.