Performance testing in honeybees relies on comparing colony traits of comme
rcial interest in a 1- or 2-year test period. Groups of sister colonies are
tested to determine breeding values and the procedure is gravely distorted
if unrelated queens are included in the sister groups. Maternity was verif
ied in nine sister queen groups from a routine performance test. Individual
workers (n = 40) were taken from 36 queenright colonies and genotyped usin
g four DNA microsatellites, Queen genotypes were derived from the worker of
fspring. The consistency of maternity of the sister queen groups was evalua
ted using the number of alleles per locus and the putative mother queen gen
otypes One group unambiguously included non-sister queens. In our tested po
pulation, the probability of not detecting unrelated queens ranged from < 0
.001 up to 0.005 depending on the putative mother queen genotype. In light
of the high costs and labour of testing colony traits, we recommend includi
ng maternity testings into the routine procedure of honeybee performance te
sting. (C) Inra/DIB/AGIB/Elsevier, Paris.