Abstract
Plants of two varieties of rice which were quite different in plant colour and yielding ability were planted in the same plots after the manner of mixing each other as shown in Fig. 1. These plots were replicated six times (Fig. 2). The width between rows was 10 inches and the plants spaced five inches apart within rows. From these plots, plants surrounded by those of the same variety (AS), plants in rows adjacent to another variety (ND), plants grown between two of another variety in 10 inches (LD) or in 5 inches apart (SD), and plants surrounded by those of another variety (AD) were individually harvested. The grain yield in grams, the number of spikes and the average weight of a panicle were recorded on plant basis in each plot (Tab. 1). The analysisses of variance were calculated from the nonweighted plant values (Tabs. 2∼4). From Tab. 2, it is seen that the mean square for varieties, and varieties x treatments, compared with error mean square, exceeded the 1% point and those for treatments the 5% point. It is plain, therefore, that there were significant differences in average yielding ability and that these varieties reacted in a differential manner at some treatments. These differential response of the varieties to some treatments was shown to result from the changes in the weight of a spike under different treatments (Tab. 4) and not the number of spikes per plant (Tab. 3). Average grain yields in Tab. 1 sbowed that those of the low-yielding variety became lower by the influence of neighboring high-yielder, and vice versa. It was concluded that the yield of spaced single plants in a population would be variably affected by the surrounding plants, and it would, therefore, be illegitimate to make plant selections on a basis of the yield of single plants in segregating generations.
