Gene action for seed yield and its component traits in linseed were studied using triple test cross analysis. The main objective was to determine the nature and magnitude of genic effects for different biometrical traits under variable environments that could support further improvements of linseed productivity using appropriate breeding methodology. Epistasis was observed for technical height in E1, 1,000-seed weight in both the environments, plant height, 1,000-seed weight and biological yield per plant in combined over environments. Both [i] and [j + l] type of epistatic interactions were significant for all these traits except technical height and 1,000-seed weight in E1, for which only [i] type of interaction was significant. Additive (D) gene action was significant for most of the traits whereas non-additive (H) gene action was significant for only seeds per capsule (E1, E2 & E3), 1,000-seed weight (E2 & E3), technical height (E2) and biological yield per plant (E3). Additive type of gene action was preponderant for plant height (E1, E2 & E3), technical height (E1, E2 & E3), capsules per plant (E3), seed yield per plant (E1 & E3), 1,000-seed weight (E1, E2 & E3), straw yield per plant (E1 & E3), and harvest index (E1, E2 & E3), whereas non-additive type of gene action was preponderant for seeds per capsule (E1, E2 & E3), and biological yield per plant (E3). The presence of additive gene action for most of the traits including seed yield per plant implies that early generation selection may be useful for the improvement of these traits. However, for traits showing both additive and dominance components of variance, heterosis breeding may be useful but chances of exploiting hybrid vigour through hybrid varieties in linseed due to its autogamous nature are bleak at present. The autogamous nature of crop and absence of genetic-cytoplasmic male sterility warrants diallel selective mating/biparental mating or recurrent selection followed by pedigree method of selection for its improvement.