The objetive of this work was to verify whether the results of soil analysis in a covariance analysis of maize seed yield would increase the experimental precision of maize variety experiments. Three trials were conducted consisting of 30 early varieties, 25 considered as super early and nine as normal cycle. They were conducted at the Federal University of Santa Maria and the fertilizer levels were for a seed yield expectancy above six tons per hectare. The experimental units consisted of two rows of five meters length spaced 0.8m in randomized block design with three replications. The soil samples consisted of six sub samples taken at maturity in the rows of each experimental unit. The results indicated that with a covariance analysis, the effect of adjusting the averages seed yields with the soil fertiity indicators was not significant and it was concluded that the covariance analysis did not contribute to increase the experimental precision of the variable seed yield
maize; covariance; experimental precision
