ROLE OF VARIATION IN REPLICATED OBSERVATIONS IN VARIETALEXPERIMENTS

摘要

An approach has been proposed to deal with a situation in the varietal experiments where the response of the varieties is statistically non-significant, even though the varietal means are sufficiently apart. The varieties involved in the experiment have its own genetic set-up and this leads to variation in the replicated observations on the hybrids, thus inflating the Error Variance in ANO VA as well as the Coefficient of Variation (C V) of the experiment. It is demonstrated through the yield data ofNon edible oil crop varietal trials (14 Hybrids and 2 Open Pollinated Varieties) that the yield variability in the hybrids may not be uniform (consistent) and thus an attempt to study the varietal performance on the basis of varietal averages may not beappropriate. Instead, the analysis of only the identified consistent varieties, as outlined in the approach, may give validity to the comparisons.