Abstract:
The study of genetic variability for key agro-morphological traits is crucial in sorghum
breeding. Although, the national sorghum improvement programs have released a number of
sorghum varieties for the highland areas of Ethiopia, the production and productivity is still
low as compared to global potential (6 ton/ha ). Huge amount of sorghum germplasm exists
in the country, but a little information is available on genetic variability, heritability and
genetic advance for the studied germplasm. The current study was designed to generate
information on genetic variability, to assess the associations among traits for yield and yield
related traits and to identify traits that best explain genetic variation and to group the
genotypes based on morphological traits. A total of 36 sorghum genotypes were evaluated in
simple lattice design with two replications in 2021 main cropping season at Chiro and Hirna
West Hararghe. Combined analysis of variance showed highly significant difference (p<0.01)
among the genotypes for all studied traits. Genotypic coefficient of variation (GCV) ranged
from 3% for leaf area to 29.65% for the harvest index, while phenotypic coefficient of variation
(PCV) ranged from 3.2% for leaf area to 30.58% for the same trait, harvest index. Broad sense
heritability ranged from 42% for panicle width to 99% for plant height. The maximum genetic
advance (GA) was recorded for above the ground biomass (179) and minimum (1.54)
fornumber of productive tillers. GAM ranged from 4.87% for leaf area to 61.19% for harvest
index. Highly significant correlations of grain yield with harvest index (r = 0.93) panicle
width (r=0.89), date to maturity (r=0.77), and thousand grain weight (r=0.59) were recorded
at genotypic level. The cluster analysis grouped the genotypes under five clusters. The
maximum inter cluster distance was observed between cluster III and cluster V (212.24).
Therefore, the maximum amount of heterosis is expected from the crosses with parents
belonging to cluster III and cluster V. In cluster mean analysis, cluster I manifested highest
value for half of the studied traits and this cluster might be used as parental lines in sorghum
improvement programs for target area. Principal component analysis revealed that the first
seven principal components with eigenvalues greater than unity explained 75.00% of the total
variation. Based on the current results, some promising genotypes exhibited relatively high
performance for yield and yield related traits, which could be exploited for variety
development in breeding program. However, to have a better recommendation further
research is needed for more than a year across multi- location
