TODOR GUBATOV 1; NIKOLAY TSENOV1; IVAN YANCHEV2
1 Agronom I Holding Ltd, BG-9300 Dobrich, Bulgaria
2 Agricultural University, BG-4000 Plovdiv, Bulgaria
Gubatov, T., N. Tsenov and I. Yanchev, 2017. Correlation between the ranking of winter wheat genotypes by grain yield and stability through various statistical approaches. Bulg. J. Agric. Sci., 23 (1): 92–101
Setting and aim: The environmental conditions have a serious impact on grain yield of cereals. The change in the reaction of a variety yield is reason to look for ways for its evaluation despite his strong GE interaction. The aim of this study was to establish an effective model for the separation of varieties based on the magnitude of the yield and the stability in a wide range of environmental conditions.
Methods: In a typical country locations of testing (5) for four years 24 varieties of winter wheat, were studied. In the most detail the GE interaction by four statistical approaches that are current was analysed. The ranking approach in determining the yield stability and their interactions using the criteria “yield”, “stability” and “yield-stability” is attached. The possibilities of each method to evaluate the effect of GE for establishing the essential differences between varieties were analysed.
Key results: The GE interaction is essential and sets about 1/3 of the total variation of the trait. Its nature is complex and diffi cult to measure. A large share of the nonlinear interaction of the variety by environment is established. The most signifi cant share of the interaction is between season * location (88 %), and the effect of genotype is about 10-12%. The arrangement of the tested varieties through various statistical methods is different, which is a prerequisite for a serious analysis of their grain yield behaviour.
Conclusions: The relationship between the three tested criteria of traits expressed by main types of correlations show that the “yield-stability” has a strong link with the other two criteria. This is reason to believe that each of applied statistical models for effective differentiation of behaviour of the variety in the group. The most relevant to achieving the objective are two commonly used model AMMI and GGE, due to the presence of software packages for their rapid implementation and visualization.