Background. In connection with climate change vary known patterns of environmental influences on the ratio of fatty acids(FA) in oil. Therefore,relevant data of modern geography test. Materials and methods. In work 24 lines and 3 commercial varieties of flax including 3 low linolenic (LL) accessions, grown in the Leningrad and Samara regions were used. FA composition was evaluated by gas chromatography for the ratio of palmitic (PAL), stearic (STE), oleic (OLE), linoleic (ω6,LIO), linolenic (ω3, LIN) acids, ω6/ω3 and iodine number of the oil (IOD). Results. The strongest differences are due to the level of LIN. It is lower in LL and gc-119 from India and higher in 3 lines carrying the gene s1 (deranged anthocyanin biosynthesis). In gc-119, contrast to LL, LIN decrease increase of OLE, instead of LIO. In lines with the gene s1 LIN increase due to the OLE reduction. Contrary to earlier publications the seeds of northern reproduction have more PAL, OLE, less LIN, IOD. 2F ANOVA revealed significant effect of genotype and reproductions place on PAL, OLE, LIN, IOD. LIO is affected only by genotype. Independence of ω6/ω3 is explained by strong abnormity of distribution due to LL. In high linolenic (HL) accessions group both factors influenced all characters except STE. Kruskal-Wallis H test (non-parametric 1F ANOVA analogue) show significant effect of genotype and place of reproductions on ω6/ω3. It reveals the impact of the reproduction place on LIN,no significant effect of genotype on OLE and IOD, which in the case of 1F ANOVA were significant. For characters of HL with normal distribution, comparing of both tests showed that in case of 0,01 p 0,09 conclusions concerning significance may vary,but in cases p 0,01 or p 0,10 they are identical. Conclusion. In our studies the geographical effect is less important than the weather in the year of growing. For abnormal distribution it is desirable to use both statistics and carefully make conclusions about the significance of differences in borderline probabilities.