effects of agronomic treatments can change differentially in relation The presence of T 3 E complicates the interpretationto environmental changes, producing a treatment 3 environment in- of the results and confounds the observed average perteraction(T 3 E). The aim of this study was to find a parsimonious formance of the agronomic treatments with their truedescription of the T 3 E existing in the 24 agronomic treatments values. Thus, significant resources in agronomy researchevaluated during 10 consecutive years by (i) investigating the factorial are devoted to studying and interpreting T 3 E throughstructure of the treatments to reduce the number of treatment terms replicated (or unreplicated) multienvironment trials.in the interaction and (ii) using quantitative year covariables to replace The standard analysis of variance for multienvironment the qualitative variable year.