Abstract

                                Six exotic and six indigenous strains of barley were used to produce four partial diallel and two diallel sets. Original and transformed data were analysed following Griffing's method (1956) for diallel sats and Kempthorne and Curnow's approach (1961) for partial diallel sets. The results were interpreted by rank correlations. In general, significant effect of transformation was recorded. Rank correlations between original and transformed data for mean squares due to gea, sca. error and sca/gca ratio indicated considerable effect of transformation in each analysis. When the statis- tical parameters were translated into genetical components, the effect of transformation. became more evident.

Key words :