The materials used for the present study consisted of sixteen flax genotypes (G) which were evaluated over six environments (E), 2 seasons (2004/05 and 2005/06) x 3 locations (two locations at El-Fayoum Governorate and one location at El-Beheira Governorate), Egypt.             The analysis of variance revealed highly significant differences among genotypes, environments and G x E interaction for all studied traits, indicating a wide range of variation among genotypes, environments and these genotypes exhibited differential response to environmental conditions. The significant variance due to residual for all characters indicated that genotypes differed with respect to their stability suggesting that prediction would be difficult, which means that mean performance alone (mean yield) would not be appropriate. Interaction component of variance (σ²ge) was less than the genotypic variance (σ²g) for all characters except each of straw weight per plant, seeds number per capsule and fiber fineness. This means that genotypes differ in their genetic potential for these traits. This was reflected in high heritability and low discrepancy between PCV and GCV values for these traits. These results indicate the possibility of using these characters as selection index for improving both of straw and seeds weight per plant.               The criterion, yield stability (YS_i) statistic indicated that S.31/3/2  (local strain classified as dual purpose type) was proved to be superior in yield and stability for most characters studied as well as Daneila (introduction classified as fiber type) was stable for fiber characters (straw yield, fiber yield, fiber fineness and fiber percentage). Therefore, the two genotypes (S.31/3/2 and Daneila) maintained mean performance advantage across nearly all the environments sampled by maintaining high level of the previous mention traits and they are recommended to be released as commercial stable high yield cultivars and/or to be incorporated in breeding program for producing stable high yield lines.                Phenotypic (r_{p < /sub>) and genotypic (rg) correlation coefficients indicated that straw weight per plant was significantly positively correlated with each of plant height, technical stem length, 1000-seed weight and oil percentage. Also, plant height exhibited positive correlation with technical stem length, indicating possibility of using both of plant height and technical stem length as selection index for improving straw weight per plant. Seed weight per plant, exhibited positive association with each of capsules number per plant, 1000-seed weight and oil percentage. Also, the relationship between capsules number per plant and 1000-seed weight was significant and positive in direction, indicating possibility of using both of capsules number per plant and 1000-seed weight as selection indices for improving seed weight per plant. Fiber percentage was positively correlated with fiber fineness, indicating that selection for genotypes had high fiber percentage and high fiber fineness is possible.}