Selection indices provide useful information to breeders who usually base selection on meansand ranks. Knowledge on selection criteria for improving tef grain yield is very limited. Accordingly, theobjective of this study was to develop selection criteria for improving tef grain yield through indirect yieldcomponent selection via a selection index. Different trait combinations along with three selection methods(based on best linear unbiased predictors (BLUPs), plot basis, and mean of genotypes) were studied attwo locations, Melkassa and Debre Zeit, Ethiopia, in 2003 and 2004, using recombinant inbred lines(RILs) of tef crosses. Selection based solely on grain yield was estimated to be less effective thanselecting based on grain yield, panicle seed weight and shoot biomass. BLUPs showed smaller meanvalues for panicle length, and shoot biomass compared with selection based on individual plots. Selectionbased on individual plots selected higher yielding genotypes than BLUPs method. Higher gains wereobtained from selection based on grain yield, shoot biomass, panicle seed weight, panicle length, andlodging index, followed by selection based on grain yield, shoot biomass, and days to maturity. The useof a selection index made it possible to detect desirable correlated responses in grain yield, shoot biomass,panicle seed weight and panicle length. This suggested that the use of a selection index might be a usefulindirect selection criterion to improve grain yield. Further research is required to validate the results ofthe present study.

Yield; Eragrostis Tef; Selection Index; Tef; Yield Components

Africa; East Africa

Ethiopia

Research Journal of Agriculture and Biological Sciences