Estudio de la heterogeneidad del suelo, del tamano y formas de parcelas y del numero de repeticiones optimas en ensayos de uniformidad en frijol (Phaseolus vulgaris L.) [online]

In Sept.-Dec. 1976 the ICA- Guali bean var. was planted (200,000 plants/ha) at CIAT, (Palmira, Colombia) to study the heterogeneity of the experimental plot; to map fertility trends and to determine correlation coefficients for adjoining plots as measures for these trends; to determine several plot size combinations and no, of replications for differences in value between treatment means; and to determine several optimal plot sizes as related to costos. The plot (36 x 36 m(2)) was divided into 1296 plots of 1 m(2) each, and their production was harvested individually. Adjacent unit plots were grouped into 30 combinations and classified in rectangulars (parallel or perpendicular to the rows) and squares. A standard deviation of 40 g was used as a criterion for grouping plots with homogeneous yields; 7 groups of homogeneous plots (ranging from 80-320 g) were established. The weighted coefficient regression (b), detemined by Federer`s method, was a more reliable measure of soil variability; in this case it was 0.768, indicating a high level of plot heterogeneity. Hatheway`s methodology facilitates the determination of plot size, based on no, of replications and difference in treatment means. In highly heterogeneous soils, plot size had little effect on reduction of variability. Optimal plot size is based on soil variability and on the costs of the experiment in each zone. (CIAT)