Genetic markers provide a potentially powerful means of identifying the bre
ed of individual animals. In this study diallelic and microsatellite loci w
ere compared for their efficiency in discriminating among cattle breeds. Da
ta were simulated for seven European cattle breeds using allele frequencies
estimated at 20 microsatellite and 30 diallelic markers. Animals were assi
gned to the breed for which their genotype had the highest probability, and
the power of the method assessed by estimating the error rate or proportio
n of animals misclassified. The number of markers required for discriminati
ng among pure, or both pure and crossbreed, animals was investigated using
either randomly sampled markers or markers selected on individual error rat
e. The relationship between individual marker variability and discriminator
y power was also investigated. Microsatellite markers were found to be more
powerful than diallelic markers for distinguishing among the breeds. The m
ost discriminatory markers were those with the highest average heterozygosi
ty and observed number of alleles. The number of markers needed to achieve
a particular error rate could be reduced by selecting markers with the lowe
st individual error rates. Discrimination among both crossbreeds and pure b
reeds required approximately three times as many markers as discrimination
among pure breeds alone.