Three strains of mirror carp originating from important regions of carp culture in France (Brenne, Dombes, Forez) were compared for growth at 5 weeks, 1 summer and 2 summers of age in ponds using a paternal cross on the same Dombes dams. An internal scaly control was produced by crossing an homozygous inbred strain on the same dams and mixed with the tested strains in all ponds in a proportion of 1:3. No between-strain differences in weight were seen at any time when taking into account the weight of the internal control. However, the internal control weight at 2 summers of age was lower than that of the tested strains (P<0.05). The use of the internal control divides the number of replicates needed to have a good statistical power by 5 at 5 weeks and by 8 at 2 summers, when compared to separate testing without control. No genotype by environment or strain-control interactions were detected. The validity and optimisation of the use of an internal control in strain testing are discussed.