Progresso genético e análise de trilha na seleção recorrente em milho

Abstract

The objective of this work was to estimate the genetic progress with reciprocal recurrent selection in the maize crop of UFLA, through the evaluation of top cross hybrids from progenies S 2:3 obtained from cycles 0, 3 and 6, as well as obtaining estimates of the genetic components, verifying the maintenance of genetic variability during the cycles. A total of 1152 top cross hybrids were obtained from 192 progenies S 2:3 from the "A" and "B" populations, from cycles 0, 3 and 6 of the reciprocal recurrent selection (SRR) of the UFLA´s corn breeding program. The experiment consisted of 1156 treatments, 192 top cross hybrids belonging to each of the six populations and four controls, in 34x34 square lattice, two replicates and two sites. It was evaluated the dehusked ear productivity: estimated by weighing the dehusked ear of each plot, total plot productivity in kilograms. With this data, the individual and joint analyzes were carried out through the mixed models approach. Initially, two models were compared, one considering homogeneity of variances for progenies within cycle and populations and the other considering heterogeneous variances by AIC and BIC criteria. For the two environments studied, the model considering heterogeneous variances was higher, obtaining lower values for the AIC and BIC selection criteria. Considering the model adjusted for heterogeneous variances, genetic components (genetic variance and heritability) were estimated for each population in each individual environment and for multiple environments. The genetic progress was estimated based on the average progenies productivity selected in each cycle. After six cycles of recurrent reciprocal selection, the genetic progress was 8%. The genetic variance values were stabilized, slightly increasing tendency, proving the existence of genetic variability in the populations, evidencing the possibility of success with the selection in future cycles.