Abstract:
Sorghum is an important cereal crop grown primarily in arid and semi-arid climates with a
wide range of adaptability to various agro-ecological conditions of Ethiopia. However, the
limited availability of improved varieties and the low production potential of landraces have
constrained sorghum production in Ethiopia and a little information is available on combining
ability, heterosis, and gene action for locally adapted sorghum landraces. Therefore, the
research was undertaken to generate information on combining abilities, heterosis, and gene
action governing the quantitative traits for yield and its related traits. The experimental
materials consisted of fifty-six parents, one hundred eight F1 progenies produced by Line x
Tester mating design, and five standard checks, which makes a total of 169 entries planted by
using simple lattice design with two replications at Melkassa and Miesso in main cropping
season of 2020/2021. The yield and yield-related traits of the combined analysis of variance
revealed highly significant (P≤0.01) variation due to genotypes for all traits over locations,
which indicated the existence of sufficient genetic variability and potential for selection of
genotypes with high performance for agronomic traits. From the mean performance, high grain
yields were obtained from the hybrids ATX2783 x IS38312(6.67t/ha), ATX2783 x ETSL100383
(6.40t/ha), ATX2783 x ETSL100007(6.21t/ha), ATX2783 x ETSL101152(6.16t/ha) and ATX623
x IS38341(5.82t/ha) with the overall value of 4.32t/ha, which was greater than the mean value
of the parents and the checks. The result of line x tester analysis of variance for pooled
combining ability revealed that the mean square due to GCA of lines showed highly significant
(P≤0.01) variation for all studied traits, and also mean square due to GCA of testers showed
highly significant (P≤0.01) variation for days to flowering, days to maturity, plant height, leaf
width, number of productive tillers, panicle length, stay green and thousand kernel weight. The
mean square due to SCA of lines x testers interactions exhibited highly significant (P≤0.01)
variation for all characters studied, except number leaf per plant and leaf width. A total of
twenty-two parental lines showed positive and highly significant (P≤0.01) GCA effects for grain
yield, and the contributions of GCA of lines and testes were 59.22% and 0.04% to the total cross
variances for grain yield, respectively. Whereas, twenty-one and eleven cross combinations
exhibited highly significant (P≤0.01) and significant (P≤0.05) positive SCA effects for grain
yield, respectively, and it contributed 40.74% to a total cross variance for grain yield. Variance
due to sca was higher than gca and the ratio of variance of general to specific combing ability
was less than unity for all the traits studied indicating the preponderance of non-additive gene
action governing in the inheritance of these traits. Based on the SCA effects of grain yield, male
lines were classified into two heterotic groups aligned to the different cytoplasmic systems of
testers. Among the crosses, ATX2783 x ETSL100829 recorded the highest heterosis (242.12%)
over the mid parents, hybrid ATX2783 x ETSL100978 revealed the highest heterosis (191.01%)
over the better parent, whereas ATX2783 x IS38312 recorded highest grain yield with 81.43
and 40.00% heterosis over the best check ESH-5 and Argity, respectively. The information
generated in this study will be used for future sorghum breeding programs and hence, additio
nal information needs to be generated by evaluating the genotypes at multi-locations.