GENOTYPE BY ENVIRONMENT INTERACTION AND YIELD STABILITY OF MAIZE HYBRIDS EVALUATED IN ETHIOPIA

• Main Author: Issa, Abdurahman Beshir
• Other Authors: Prof MT Labuschagne
• Format: Others
• Language: en-uk
• Published: University of the Free State 2010
• Subjects: Plant Sciences
• Online Access: http://etd.uovs.ac.za//theses/available/etd-09062010-121424/restricted/

The objectives of the study were to evaluate the adaptability of 17 experimental maize genotypes under the maize growing environments of Ethiopia, to select the best hybrids for commercial production, to utilize various statistical procedures for analyzing G x E interactions and yield stability of Ethiopian maize hybrids across nine environments and to indicate breeding strategies for releasing genotypes with adaptation to target environment. Seventeen experimental maize hybrids, including a locally released standard check, were evaluated for grain yield in a mid altitude areas of Ethiopia, for a period of three years (2004 to 2006) across three locations. The genotypes were planted in a completely randomized block design. G x E interactions and variance components were calculated using factorial analyses. Stability parameters were calculated using various parametric and non-parametric methods. While, a cluster analysis was done to classify the different genotypes and environments. The computer programmes, Agrobase Generation II (Agronomix, 2008), SAS (SAS Institute Inc., 2003) and NCSS 2004 (Hintze, 2001) were utilized to perform the different analyses. The mean squares for G x E interaction were found to be highly significant for yield including the variation among the genotypes (G). The Location (L) and Location x Year (L x Y) interactions have the largest components of variance for grain yield, which is an indication of the variation among testing sites and year to year fluctuation of the weather. The experimental maize hybrid PR1 (mean yield =7.14 t ha-1) was found to be the highest yielder followed by PR13 (6.77 t ha-1) .This result shows the potential of the experimental genotypes to be released for commercial production. Highly significant correlations (P<0.01) was obtained among Eberhart and Russellâs deviation from regression (S2 di), Wrickeâs ecovalence (Wi), Shuklaâs ( ) 2 i s stability variance and AMMI Stability Value(ASV). Mean yield and Lin and Binnsâs cultivar superiority performance (Pi) showed high correlations but they showed no correlations with the rest of the stability parameters. The non-parametric measure of Nassar and Hühnâs absolute rank difference S(1) and variance of ranks S(2) was significantly correlated (P 0.05) with ASV, however they did not correlate with the rest of the parametric measures. Based on the different stability parameters, hybrids PR1, PR13 and BH540 were found to be more stable and responsive to favourable environments while SE21 and SE22 hybrids showed their stability in the low yielding environments. Since, ASV showed higher correlations with both parametric and non-parametric measures, and the mean yield, it is recommended to use ASV as a stability parameter in identifying stable genotypes from multi-environment trials. The analysis of variance for the AMMI model indicated highly significant differences between genotypes and environments as main effects and the interaction effect of G x E was also highly significant. The first two interaction principal component axes (IPCA) of the AMMI model together accounted between 34% and 56% of the total G x E interaction sum of squares for grain yield. The AMMI biplot categorized the genotypes PR1, SE4, PR2, PR3, BH540 and PR13 as well adapted to favourable environments with above average yield. Hierarchical clustering using unweighted pair group method with arithmetic average (UPGMA) grouped the genotypes into five and the environments into three clusters. The results from the study of optimum allocation of resources indicated that allocation of two or three replications, three to five locations for three to four testing years will give adequate information for yield estimates, help to minimize trial costs and provide more precise data for variety release decisions in Ethiopian condition. The research also proved that locations-years tradeoffs are more effective than locations/years-replications tradeoffs in getting statistically efficient data from a maize yield trial.