Die Bodenkultur - Journal for Land Management, Food and Environment


Prediction of breeding values for honey yield (Apis mellifera carnica) in Austria


The object of this study was to establish a useful method of predicting honey yield breeding values in reconciliation with the possibilities given at the Institute of Bee Research in Lunz am See/Austria. The method of combined selection (LUSH 1947) was taken as a basis. This method is in accordance with a multiple regression including individual and family performances (Formula (1)). The organisation of performance testing in Lunz provides partly repeated measure-ments for queens (coionies). Therefore the method of combined selection was adapted to include repeated measurements within individuals and unequal family sizes (Figure 1, Formula (2) and (3), Appendix).

The effect of including repeated measurement on the prediction of breeding values and the expected correlation between true and predicted breeding values (rAA) was shown (Table 1, Formula (4) to (7)). In honey bee breeding the dam of the drone-producing queens is seen as the male representative ("sire"). Therefore the breeding values of these dams (Formula (8» predicted by the performances of their daughters can be used for selecting the drone-producing queens.

The practical carrying out of predicting breeding values takes place in two steps. First LS-Effects (Least Squares) for the environmental factors year, farm and station within farm are estimated and used to correct the honey yield. Subsequently breeding values for queens and dams ("sires") are predicted with the described methods (Formula (3) and (8)).

Key-words: honey yield, breeding value, index, selection.