Proposal of methodology to assess the representativeness of environments for genotypic discrimination
Devanir Mitsuyuki Murakami and Cosme Damião Cruz
In this work, a methodology is proposed for evaluating environment representatitivity through genotypic performance (grain productivity, kg/ha) of commercial corn hybrids and the use of the determination coefficient (R2) as a referential for a likely successful search for representative sites. The methodology is based on the adoption of procedures similar to those adopted in the diallelic analysis, as proposed by Griffing (1956), in which the data evaluated are measurements of dissimilarity between environment pairs. For this, several experiments were assessed during 3 different seasons: a) two in the autumn (alternative) growing season of 1996 and 1997; and, b) one in the summer (normal) growing season of 1996/97. All the experiments were arranged in a randomized block design, with three repetitions. The methodology made possible the identification of that (those) environment(s) presenting greater overall capacity for environmental representativeness and the one which best replaces it (them) during the three seasons. It was concluded that this methodology was effective
for the purposes established as well as being easily interpreted and implemented.