Abstract:
A uniformity trial in brinjal (Solanum melongena L.) was conducted at the Main Campus of the Kerala Agricultural University, Vellanikkara during the third crop season, 1980. Observations on yield, number of fruits, primary branches and height of each plant was recorded at time of harvest in small units each of size 1.08m2 (2rows x 2 plants at 60 x 45 cm spacing). The variability of each plot size and shape was determined by calculating the coefficient of variation. It was observed that an increase in the plot size in either direction decreased the CV. But the decrease was more rapid along N-S direction. Long and narrow plots yielded lower CV than approximately square plots. The observed relation between plot size and variance was in conformity with the Fairfield Smith’s variance law. At larger plot sizes, the regression line showed a tendency to curve down although negligible. The optimum plot size observed through Smith’s method and maximum curvature method was almost the same. From the above considerations, a plot size of 8.64 m2 (9.6m x 0.9m) was found to be most advisable for conducting most of the field experiments in brinjal. The efficiency of the plot decreased as the size of the plot was increased. There is a general decrease of block efficiency with increasing block size. More compact blocks of the same size show a higher efficiency. Blocks laid out perpendicular to the direction of fertility gradient removed largest variation. The number of replications and total area of land required to give 5% SE of the mean were calculated for the same number of plots per block smaller plots require more replication but less total area than larger plots. But increasing the number of replication rather than plot size was found more advantageous for a fixed experimental area. The fertility contour map of the field revealed that the land is not very homogeneous as far as the fertility pattern is concerned. By assuming arbitrary values of the cost proportional to the number of replications and the cost proportional to the total area per treatment, the optimum plot size for field experiments in brinjal was computed using a linear cost function.
