Twenty-seven tetraploid (2n = 4x = 28) wheat (Triticum turgidum L.) landrace populations consisting of 2876 entries collected from the north and north-central administrative regions of Ethiopia were studied for variation of eleven phenotypic characters that have two or more classes. Using the frequencies of each phenotypic class within a character, the Shannon-Weaver index (H') was employed to estimate and analyse the diversity on the basis of four regions and five altitudinal gradients. Most traits were polymorphic. However, monomorphism was common in many of the populations for dense spike, long beak, and glabrous glume. On character basis, seed colour, seed shape, and glume pubescence contributed the highest mean diversity indices (H' = 0.99, 0.96, and 0.88, respectively). Spike density contributed the lowest index (H' = 0.24). However, it was the only trait that displayed significant dimerences ( p < 0.01) among regions, lax-spike types being frequent only in Gojam region. Significant differences ( p < 0.05) among altitude groups were detected only for glume colour and beak length. Much of the variance was among populations within region and within altitude group. For the overall characters, significant differences ( p < 0.01) were revealed only among populations within region and altitude group, but not among regions or altitude groups. However, the highest mean diversity indices were obtained in Tigray region (H' = 0.70). and in the altitude range between 2401 and 2600 m.a.s.1. (H' = 0.68). No clear association was detected between overall phenotypic diversity and altitude. The overall index (H' = 0.70) for the north and north-central regions of this country is of the same magnitude as that of previously reported regions. Therefore, before any collection mission is undertaken, we suggest further studies to gain information on the interrelationships among morphological diversity, variation for biochemical markers, and ecogeographical variables.

One hundred and twenty tef Eragrostis tef (Zucc.) Trotter, germplasm lines from eight regions of Ethiopia were field evaluated in two randomized complete blocks at Debre Zeit and Melkassa during the 1995/96, and at Debre Zeit and Alem Tena during the 1996/97 main seasons. The objectives were to assess the extent and pattern of their diversity with respect to 17 pheno-morphic and agronomic traits, and to obtain an insight into the broad sense heritability (H) and genetic advance (GA) expectations of the various traits. Discernible (p < or = 0.05) variability among regions and among genotypes within regions were observed in seven and nine of the traits, respectively. In the combined analysis of variance over environments, genotypes showed substantial (p < or = 0.05) variation in all traits, and genotypes and environments interacted significantly on nine of the traits. The phenotypic and genotypic coefficients of variation varied in that order from about 3% for days to maturity to 28% for grain yield/plant, and from less than 2% for days to maturity to 15% for number of fertile tillers/plant. Estimates of H were lowest for shoot phytomass/plant (17%) and highest for panicle length (75%). The minimum GA expectations were for days to maturity (2%) and the maximum was for number of fertile tillers/plant (23%). About 67% of the gross phenotypic diversity in the germplasm was explained on the basis of four principal components (PCs). Up to 28% of the variance extracted by the first PC was due mainly to variations in the second and first basal culm internode diameters, culm and panicle length, days to panicle emergence, and number of panicle branches and fertile tillers/plant. Largely variations in the length of the two basal culm internodes and grain yield/plant contributed to 19% of the whole variance accounted for by the second PC. Generally, the study showed that there is ample genetic variation in the genotypes evaluated and this presents a valuable base for use in the improvement work.

No abstract

Sixty tetraploid wheat (Triticum turgidum L.) landrace agrotypes collected from the central highlands of Ethiopia and one commercial check cultivar (Boohai) were evaluated at Akaki experimental station for grain yield and 11 other component traits . The objectives were to estimate phenotypic (PCV) and genetic (GCV) coefficients of variation, broad-sense heritability (H) and genetic advance (GA), and to determine the interrelationships among the various traits . Genotypic differences among the agrotypes were highly significant for all the traits considered . Compared to Boohai, the landrace agrotypes were later in days-to-heading (DTH) and maturity (DTM), and had shorter grain filling period (GFP), lower fertility (KS) and lower 1000-kernel weight (TKW) . By contrast, they were superior to Boohai in tiller number (TN), biological yield (BY) and grain yield plant-' (GYP). Intermediate to high order estimate combinations of GCV, H and GA (as % of the mean) were observed for TN, GYP, number of kernels spike` (NKS), harvest index (HI) and TKW . GYP showed a moderate heritability which was higher than GFP, BY and Plant height (PHt) . DTH and DTM were strongly correlated, but both were negatively associated with the rest of the traits except PHt . The negative correlation of DTM with GYP was largely indirect via other characters . PHt had either a weak or negative association with the other traits. TN and TKW were positively correlated with GYP, and had high and intermediate direct effects, respectively. These two traits, however, were negatively correlated and showed a substantial counter-balance effect via one another . It appears that, for the short-term, improvement of the Ethiopian wheat landraces may be possible through indirect selection for TN and TKW or direct selection for grain yield per se. In the long-run, crossing programmes between indigenous and introduced germplasm may be necessary. Introductionturgidum L.) and hexaploid (T aestivum L .) species are grown. The former constitute more than twoWheat (Triticum spp .) is one of the most important thirds of the wheat area in the country and are usucereals cultivated in Ethiopia . Both tetraploid (T ally grown at elevations between 1800-2800 m.a .s .l.,

Sixty tef germplasm populations consisting of 3000 panicle-derived lines from six western and southern regions of Ethiopia were evaluated for 17 pheno-morphic and agronomic traits on pellic Vertisols at Debre Zeit Agricultural Research Center during the 1999 main season. The objectives were to study the extent and pattern of variation of the germplasm with respect to regions and altitude zones, to classify the populations into relatively homogenous groups and to identify the major traits contributing to the overall diversity of the populations. At 75 % similarity level, the 60 populations aggregated into nine complexes of two to 10, with 12 populations remaining un-grouped. Five principal components (PC) extracted about 81 % of the gross variance among the populations. About 40 % of the variance accounted for by the first PC alone resulted largely from variations in diameters of the two basal culm internodes, grain yield and number of spikelets/panicle, shoot phytomass and grain yield/plant, and number of culm internodes. The entire regional as well as the clinal (altitude zone) variation was explained by five and two PCs, respectively. The discriminant analyses depicted about 77 % correct grouping of the 47 populations into nine clusters and about 62 % and 68 % correct origin-based classification of the germplasm in terms of altitude zones and regions, respectively. In general, the study demonstrated the existence of regional and clinal (altitude zone) variation patterns in the tef germplasm populations. The broad trait variation in the germplasm implies ample opportunities for genetic improvement of tef through selection and hybridization.