The objetive of this work was to study competition trials of maize, wheat, soybeans, drybeans, barley, oats, triticale and sorghum conducted in Rio Grande do Sul State and estimate relations between least significant difference and coefficient of variation. Statistics characteristics of 2014 trials were used, such as: experimental design, number of cultivars, replications, error degrees of freedom, coeffïcient of variation (CV) and least significant difference (LSD) obtained from Tukey's test by the porcentage of the experiment mean. The regression equation passing through the origin obtained between LSD and CV as the independent variable resulted in LSD = 2.72xCV, with the angular coeffïcient ranging from 2.47 (barley) to 3.05 (sorghum). These difference indicate that there are deviations of the traditional experimental layouts used by field crop researchers. The LSD is the estatistic more adequate to control experiment quality. This estatistic can be easily calculated from CV values and its magnitude reflects the commercial yield importance of the crop.
quality; error; experimental precision
