The genetic variability of cultured stocks of common carp ( Cyprinus carpio) was studied using two types of genetic markers: allozymes and microsatellites. A comparative analysis was investigated between six strains from extensive aquaculture in two French regions (Dombes and Forez) and five strains from the Czech Republic stemming from artificial selection and maintained in the Research Center of Vodnany. Observed heterozygosity from allozyme data varied from 0.003 to 0.029 for French strains and was significantly lower than that observed for Czech strains, which varied from 0.026 to 0.058. The genetic variability of microsatellites for the whole data set was considerably higher than that for allozymes. The mean number of alleles per microsatellite locus was higher for French strains than for Czech strains (respectively, 6.2 and 4.2). Microsatellites are generally considered to be selectively neutral and the reduction of microsatellite allelic diversity in Czech strains could be related to a possible population bottleneck associated with breeding practice. However, the patterns of genetic variation for the two types of marker are possibly biased by the limited number of sampled individuals and must be confirmed in the future by more extensive studies. Significant differentiation between the strains was shown using F statistics, whatever the tested groupings, and no obvious difference was detected between relationships inferred with the two types of genetic markers. A neighbour-joining analysis of genetic distances (chord distance of Cavalli-Sforza and Edwards Am. J. Hum. Genet 19 (1967) 233.) between strains showed that the topologies of the trees for microsatellites and allozymes were globally congruent. However, a more pronounced discrimination between the strains of the two countries was detected by the microsatellite markers which appeared more suitable to infer the effect of the geographical separation of the ancestral strains or different culture histories.