Department of Mathematical and Statistical Methods, Poznan University of Life Sciences, 60-637 Poznan, Poland
ZAWIEJA, B., 2014. Distinctness, uniformity and stability trials with new variety of plant (oilseed rape), finding outliers. Bulg. J. Agric. Sci., 20: 1040-1045
Each new variety reported by breeders to registration, has to be tested with regards to its distinctness, uniformity and stability (DUS). It can only be registered if it differs from other existing varieties in at least one characteristic. Unfortunately if the samples include outliers, then often the wrong decision concerning population is taken. Thus such observations should be eliminated from samples. In this study, a few methods for detecting outliers were compared based on the trials conducted in the Research Center for Cultivar Testing in Słupia Wielka in Poland with new varieties oilseed rape. The comparison of statistical method was executed for the one of aspect of DUS (distinctness, uniformity and stability) tests, namely for distinctiveness of the new varieties. As a criterion of choice the best method the numbers of distinguished pairs of varieties was used. The largest number of varietals pairs was distinguished after application the M-estimate method, thus this method was declared as the best.