Reproducibility of joint regression analysis

When carrying out a Joint Regression Analysis (JRA) one linear regression is adjusted per cultivar. The upper contour defined by these regressions is partaken by the dominant cultivars. Comparison of the other ones with the dominant cultivars is used to find out which cultivars are significantly dominated. We may apply, in this context, one-sided t tests and multiple comparison methods (such as the Scheffé and the Bonferroni) for cultivar selection. The multiple comparison techniques render our treatment more robust. To assess the reproducibility of this technique we applied it to two sets of field trials in which the same cultivars of winter rye were used.