Analysis of Yield Stability of Sunflower Genotypes By Some Statistical Models


Thirteen genotypes of sunflower (Helianthus annus L.) were tested by seven statistical models to estimate yield stability. The models were; regression coefficient according to Eberhart and Russel, regression coefficient according to Perkins and Jinks ,regression coefficient according to Freeman, Perkins, ecovalence according to Wric, Shukla variance, stability index according to Bajpai and Probhakaran, and coefficient of variation according to Francis and Kannenberg. The objectives of this study were to compare these models and to study their relationship to yields stability. It was found that all models located same genotypes as a stable except rgrassion coefficient according to Freeman and Perkins. The stable genotypes were Kws, Allstar, Aqmar, and Zahrat Al-iraq, whereas the unstable genotypes were Carlos, Florasol, and Shimows. In spite of that, we cannot recommend Kws and Allstar genotypes because of low yielding, but, Aqmar, and Zahrat Al-iraq were the best genotypes according to their yield and stability. Selecting genotypes based on yield or stability alone is ineffective, therefore, integrating yield and stability will display good criteria for yield trials.