Genotype by environment interaction and grain yield stability of Ethiopian black seeded finger millet genotypes.
Finger millet (Eleusine coracana (L.) Gaertn) is an important cereal widely produced in Ethiopia across diverse agro-ecologies. It is valued by local farmers for its ability to grow in adverse agro-climatic conditions, where other cereals fail. The yield potential of this crop is in the range of 4-5 tonnes/ha, but the current national average grain yield is far below the potential (2.1 tonnes). Lack of improve varieties which are stable, high yielder and stress tolerant is a major limiting factor to production of this crop in Ethiopia. A field experiment was conducted using twelve black seeded finger millet (Eleusine coracana subsp. coracana) genotypes, including local and standard checks (Degu) at two locations (Bako and Gute) in Ethiopia for three years (2014 - 2016). The objective of this study was to identify stable and high yielding genotypes for grain yield and other agronomic traits among the black seeded finger millet genotypes of Ethiopia. The additive main effect and multiplicative interaction (AMMI) model analysis of variance revealed highly significant (P<0.01) differences between environments, genotype, and Interaction Principal Component Analysis (IPCA-I), but significant variations (P<0.05) for G x E interactions. This indicates that the genotypes performed differently over environments and that the test environments are highly variable. Only the first IPCA-I showed high significance (P<0.01) and contributed 48.39% of the total genotype by environment interaction (G x E). Genotypes BKFM0020, BKFM0006 and BKFM0010, which had high grain yield, but with IPCA value close to zero, indicated the wide adaptability/stability. Similarly, analysis using Eberhart and Russell model revealed that these genotypes were within the relatively acceptable range of regression coefficients (bi), approaching to one (0.742, 0.8176 and 1.0578), and deviation from regression closer to zero (s2di) (0.0385, -0.0661 and -0.0248), respectively. This implied that pipeline genotypes were stable, widely adaptable and high yielders than the other genotypes. Genotype and genotype by environment (GGE bi-plot) analysis also revealed that these candidate genotypes were stable and high yielder. Besides, these genotypes showed resistance to blast disease, which is a threat to finger millet production in the study areas. Therefore, these genotypes were selected as potential candidates for possible release in western Oromia and similar agro-ecologies of the country.
Key words: AMMI, blast, Eleusine coracana