Optimum plot size for field experiments on tomato – A statistical investigation (Lycopersicon esculentum Mill.)

Thumbnail Image
Journal Title
Journal ISSN
Volume Title
Uniformity trial on tomato (Variety: Gujarat tomato-1) was conducted during rabi season of the year 2003-04 at the Vegetable Research Station, Junagadh Agricultural University, Junagadh. The fruit yield data of 1296 plots (basic unit 1.20 m x 0.90 m) were recorded and analysed to work out optimum size and shape of the plot using different approaches viz., maximum curvature method (Day, 1920), Fairfield Smith’s variance method (Smith, 1938), and comparable variance method (Keller, 1949). The size and shape of the block was determined, the relative efficiency of different experimental designs also compared and the effective number of replications for field experimentation were also worked out. The results revealed the followings. ( i ) The coefficient of variation (C.V.%) per unit area decreased with the increase in plot size. When this relationship expressed in equation form transforming variable plot size (x) to log scale, the predictability of coefficient of variation was high (R2 = 0.89). The rate of reduction in C.V.% was more with the increase in length of the plot than that with the increase in width of the plot. (i.e. number of rows) Nine units size plot 10.80 m x 0.9 m (a plot of 1 row each of 10.80 meter length) was found optimum for field experiments on tomato. (For convenience in agronomic practices the plot covering 3 rows each of 3.6 meter length can be taken as net plot) When the plots were arranged in the square or nearer to the square blocks, the efficiency of blocks were more as compared to the plots arranged in long and narrow strips The estimates of soil heterogeneity index, b, showed large variation from picking to picking indicating the need of in depth study on casual variables. For any given size of the plot, the block of size 9 plots had comparatively lower C.V.% as compare to other size of block, hence suggested that the block size influences the precision of the estimate. The block shape also influenced the precision. ( ii ) The Balanced Lattice design was comparatively more efficient than Randomized Block design for field experiments on tomato, followed by Simple Lattice design. Smaller plot was more efficient in controlling soil variation. They required minimum three replications to achieve 5 per cent accuracy in any of the block size. But the total area required by smaller plots was much than that by bigger plots. It is therefore better to have smaller plots with more replications. From the present investigation, following recomm-endations could be made for field experimentation/ survey on tomato crop (Variety: Gujarat tomato -1) to be studied under Saurashtra region. The optimum net plot size for field experiments should be 9.72 sq. m covering 3 rows (spaced 90 cm a part) each of 3.6 meter length. This is more suitable than other size and shape. Minimum three replications are necessary for field experiment having 9 units sized plot. The Balanced Lattice design was comparatively more efficient than Randomized Block design for field experiment on tomato.