(2004) Moura, Mara Fernandes; Pinheiro, José Baldin; Morais, Lizz Kezzy de; Moura, Nara Fernandes; Aguiar, Ananda Virgínia; Chaves, Lázaro José; Vencovsky, Roland; Zucchi, Maria Imaculada
The magnitude of the interaction G x S was measured by different biometric techniques to draw inferences on
improvement strategies for soybean. Twenty-five highly homozygotic (F9:2) soybean lines were evaluated in three sowing periods.
The trials were set up in randomized complete blocks with three replications. The controls IAC-100, OCEPAR-4, MTBR-45, and
EMGOPA-313 were used. Number of days to flowering and to maturity, grain filling period, stained seeds, weight of a hundred seeds
and the grain yield (GY) were evaluated. The complex part was predominant in the partition of this interaction. In relation to the
selection gain for GY a mean reduction of 10% is expected in comparison to a selection realized in period 2 only with specific
selections per period, admitting a selection intensity of 40%. A reduction of 28% is expected under a selection intensity of 20%. The
selection of more generalist genotypes is possible, although it leads to losses in mean terms.