With the objective of identifying the maximum possible percentage of reduction in the density of plants recommended, of beans cultivars of different habits of growth, without having reduction of the experimental precision experiments during the agricultural years of 2001/02 and 2002/03 were realized. A randomized block design with subdivided split plot and with three replications was used in two agricultural years. In the main parcel three had been randomized to beans cultivars with different habits of growth (Iraí, TPS Nobre and Pérola) and in the sub-parcel, the five simulations of reduction of density of plants in relation to the recommended to each to cultivar (0%, 20%, 40%, 60% and 80% of reduction). The results showed that the suppositions of the mathematical model (variability, variance homogeneity, additivity and normality) had been all satisfied ones, in the five simulations of reduction of density of plants in the two agricultural years. The experimental precision is modified with the changes of reduction levels of density in plants inside and between agricultural years, giving evidence bigger non-reduction experimental precision for assays in the density of plants, that is, with the use of the recommended density to each cultivar (0% of reduction).
Phaseolus vulgaris; density of plants; coefficient of variation; cultivar register
