Bekele Endashaw (Professor)Assefa Kebebew (PhD)Jifar Habte2018-07-022023-11-092018-07-022023-11-092008-07http://etd.aau.edu.et/handle/123456789/5337Nineteen released varieties and two local landraces of tef from Jimma area were planted in quadruplicated completely randomized block design at Jimma Agricultural Research Center during the 2007 main cropping season to assess the extent of genetic diversity and association among different traits of tef. Data were collected on 19 quantitative traits including days to heading and to maturity, grain filling period, plant height, panicle length, culm length, number of fertile tillers per plant, number of spikelet and panicle branches per main panicle, first and second basal culm diameters, harvest index, single plant phytomass, grain and straw yield, thousand seed weight, head smudge disease incidence and severity, and lodging index. Similarly, qualitative trait data were recorded on panicle form, and seed coat and lemma color. The analysis of variance showed highly significant (P<0.01) genotype differences for all quantitative traits and at (P<0.05) for traits such as culm length and number of fertile tillers per plant which depicted merely significant differences among the test genotypes. The range computed for all of the traits revealed wide range of phenotypic variation in the genotypes studied. The estimated values for phenotypic variance were higher than genotypic variance for all the traits. Likewise, the phenotypic (PCV) and genotypic coefficients of variability (GCV) showed wide variation among the different traits. The GCV ranged from 2.87 to 31.97 % and the PCV ranged from 4.53 to 38.74 %. The highest and lowest heritability estimate was observed for single plant straw yield (93.13%) and number of fertile tillers per plant (16.87%), respectively. Similarly, expected genetic advance ranged from 0.029 for harvest index to as high as 48.04 for number of spiklets per panicle. On the other hand, GA as percent of the mean ranged from 3.09% for days to heading to 54.35 % for lodging index. Single plant grain yield showed positive and significant correlation with all traits except days to grain filling period, harvest index, head smudge disease incidence and severity, and lodging index at both phenotypic and genotypic level and number of fertile tillers per plant at genotypic level. Lodging index showed negative phenotypic and genotypic correlation with all traits except number of fertile tiller per plant, harvest index and disease severity and incidence. Disease index similarly showed significant correlation with about 60% of the quantitative traits of tef under consideration. The cluster analysis grouped the tef genotypes into four clusters of one to 13 genotypes per cluster at about 60% similarity level. The first five principal components with eigenvalue greater than one accounted for 85.2% of the entire genetic variation in the tef genotypes used in this study. Moreover, the contribution of each of the first five principal components is 43.9%, 20.9%, 8.4%, 6.2%, and 5.0%, respectively. Overall, the study revealed variation among the tef genotypes with respect to phenotypic traits. However, future research in this area should focus on data supported by DNA analysis in order to augment the results based on morphological data. As the two landraces used in this study are a mixed population, they should be purified using appropriate breeding procedure based on their data on qualitative traits. DZ-01-974 was the best performing one with respect to most important traits considered in this study. Therefore, an adaptation trial should be conducted on this variety to recommend it for the area if it repeats its good performance. Key words: Tef, genotypes, varieties, landraces, cluster, principal components, quantitativeenTefgenotypesvarietieslandracesclusterprincipal componentsquantitativeThesis