The use of experimental techniques in order to increase trial precision is important to detect small yield gain in crops. The aim of this research was to evaluate the potential use of covariance analyses to increase trial precision. Data of grain yield and final plant population from 313 corn (Zea mays L.) cultivar trials were studied. The experiments were carried out under different environmental conditions of the state of Rio Grande do Sul, Brazil, from 1993 to 1996. An HD statistic to compare the precision for use and not use of the covariance analyses was developed. The HD is the amplitude among cultivar mean estimations divided by the LSD value of Tukey test. The highest HD value, the best trial precision is obtained, which permits to detect small differences between cultivars. It was verified that covariance analyses increased HD from 4.7 to 19.5 improving cultivar discrimination yields adjusted by the covariable final plant population.
Zea mays; grain yield; plant population; statistical analysis
