Abstract
Cross-over interaction is always a major concern for plant breeders when recommending a crop cultivar for different agro-ecologies. Hundred flue-cured tobacco recombinant inbred lines (RILs) derived from F-4:5/F-4:6/F-4:7 populations along with three parental checks were evaluated to determine genotype by environment (GE) interaction. The experiments were conducted for three consecutive years (2012-2013, 2013-2014 and 2014-2015) at two different locations i.e., Mardan (plain area) and Mansehra (hilly area), Pakistan using alpha lattice design with three replications in each environment. Six distinct environments were generated in combination of three years and two locations. Results obtained from additive main effect and multiplicative interaction (AMMI) analysis revealed that 46.5% of phenotypic variation in yield was contributed by environmental effects while 45.0% was explained by GE interaction. First four principal components (PCs) were significant and cumulatively explained 89.9% of variation due to GE interaction. For instance, based on AMMI-1 model, genotype G11 was identified as high yielding (2669 kg ha(-1)) followed by G86 (2586 kg ha(-1)) and G28 (2563 kg ha(-1)). Likewise, in AMMI-2 model, G11, G5 and G56 were identified as most stable genotypes. Generally, performance of inbred lines at Mardan was consistent. Mansehra had the most discriminating and erratic environments over years for FCV lines. Genotypes Gil and G86 appeared as high yielding elite tobacco lines possessing dynamic stability. The mentioned FCV lines were superior to standard checks in yield and stability; hence could be recommended for diverse environments. This study puts emphasis on the significance of conducting multi-environment yield trials to screen not only best performing lines but also to reduce breeding cycles for new tobacco cultivars. (C) 2019 Friends Science Publishers