Two sunflower uniformity trials were conducted at Sakha Agric. Exper. Station during 1996 and 1997 summer seasons. In each trial, 640 basic unit were used. The basic unit was two rows (3.0x0.6m). Seed yield obtained from the adjacent basic units were arranged to obtain the various combinations of plot size and shapes. Thereafter, optimum plot size, plot shape and number of replications required to evaluate sunflow‌er field experiments were determined. As plot size (x) increased the vari‌ance per basic unit (Vx) and coefficient of variability (CV%) tended to decrease. The results obtained could be summarized as follows: 1_ Soil heterogeneity index was 0.739 in 1996 and 0.698 in 1997 with an average of 0.718. 2.Optimum plot size was ranged from 4.159 to 5.095 m2 by Smith method. Using the maximum curvature, the optimum plot size was found to be 11.326 m2. 3.The variance is affected by the plot shape and its direction. This would indicated that, there is fertility gradient within the experiment soil. 4.The relationship between the coefficient of variability (CV.) and plot size (X) were expressed by the following equations: C.V. = 26.81 x -0.8648for the first season C.V. = 28_78 x -0.8609for the second season. 5.Plot shape was important for small size plots. However, long and narrow plots were recomended. 6.Effect of number of replications and plot size on the differenc-es between means were indicated increasing the number of replication was more effective in detecting the differences between means than in-creasing plot size.