It remains doubtful if bee breeding can be compared with normal animal breeding even if adapted models are used to approach a realistic h². In a natural population of honeybee colonies the cloud of drones gives a solid basis for an extremely outcrossing system. Ebbersten (1996) makes clear that the depletion of genetic variation within a colony reduces its viability. The actual breeding according to the German studbook regulations for honeybee-breeding (DIB 2002) combined with the use of controlled mating results in a depletion of this kind of genetic variation. If the observed unexpected ranking of the daughter groups are a consequence of this depletion by decreasing the chance to have groups of bees able to react to ‘unexpected situation in the environment’ remains unclear. The literature depicting aspects of honeybee colony life depending on groups of unrelated supersisters is overwhelming (Jones 2004). Selection and breeding is done for the sake of a gain in the prediction of performance. Maybe that depletion of genetic variance in colonies counteracts the expected increased predictive value of selection. The fact, that about 30% of a group of colonies in a selection-program shows a unstable performance from year to year (Englert & van Praagh 1980) can be seen as another fact, suggesting a reconsidering of the way breeding is done with honeybees.
