Relative Precision of Some Incomplete Block Designs for Soybean Uniformity Trials and Optimum Sample Size

Abstract

Experiments of the present study were conducted at the Experimental Station, Faculty of Agriculture, Cairo University, Egypt in 1998 and 1999 seasons. Cutler and Crawford soybean cultivars were used. In 1998 and 1999 four uniformity trials were conducted to estimate the relative precision of incomplete block design i.e. simple lattice, triple lattice, square lattice and balanced lattice relative to randomized complete blocks design.During 1998 and 1999, two separate experiments were carried out to estimate the optimum size of random sample for the determination of seed yield/plant and some yield components. Lattices were always more efficient than randomized complete blocks. The relative precision average of the two seasons ranged from 110.95 % to 167.89 % for Cutler and from 123.40 to 166.13 % for Crawford. The two lattices squares designs of 3 x 3 and 5 x 5 exhibited the highest relative precision of 147.12 % and 167.89 % for Cutler and of 166.13 % and 145.45 % for Crawford, for the two forementioned designs in the same order. Consequently, it could be concluded that although lattice square designs resulted in the highest relative precision for soybean, different cultivars reacted differently where lattice square of 3 x 3 and 5 x 5 resulted in the highest relative precision values for Crawford and Cutler, respectively.The results revealed that the optimum sample sizes for Cutler cultivar were 9, 3, 24, 15 and 21 plants for plant height, number of branches/plant, number of pods/plant, seed index and seed yield/plant, respectively. Similar results were obtained for Crawford except for number of branches/plant and seed yield/plant in which 6 and 24 plants were the optimum, respectively.