Nationality: PA

Thesis (Ph. D.)

Three selective procedures were compared in order to identify outstanding S1 lines in maize populations BR-105 and BR-106. One hundred genotypes (plants) were sampled from each population, from which one hundred S1 progenies, intrapopulation and interpopulation half-sib progenies were obtained. These progenies were evaluated during two growing seasons (92/93 and 93/94) in two locations of Piracicaba, Sao Paulo. The best 10 lines for each selection type were identified based on their own performance, totaling 30 lines from BR-105 population and 30 lines from BR-106 population. Each line was crossed to an inbred tester generating 30 testcross progenies and 30 S1 progenies for each population. These progenies were evaluated using an 8 x 8 Lattice arranged in split-blocks with four replications, during the growing season 94/95 at two locations in the Piracicaba region and one location in Santa Cruz das Palmeiras, all in the State of Sao Paulo. The traits evaluated were: ear weight (PE), plant height (AP), ear height (AE), stand (ES), prolificacy (P) and the ratio: ear height/plant height (AE/ AP). The genetic variance components and other genetic parameters estimates, showed that populations differ in relation to the studied traits. It was verified that, as expected, S1 progenies presented genetic and phenotypic variances, and heritabilities superior to those of half-sib intra and half-sib inter progenies, due to the fact that inbreeding increases genetic variability among progenies within the populations. The estimates of genetic correlation among progenies showed in both populations a higher association involving the non-inbred progenies (half-sib intra, half-sib inter) than that involving inbred and non-inbred progenies (S1, half-sib intra) and (S1, half-sib inter). Results showed that both populations would have different gene frequencies, being genetically divergent, constituting heterotic patterns with potential for developing outstanding hybrids. The S1 lines selected from interpopulation selection showed a small tendency of superiority over the S1 selection, which is superior over intrapopulation selection for both populations, being these superiorities associated with a higher prolificacy index. It was also found that performance of testcross progenies was similar for ear weight (PE) and prolificacy (P) in both populations, regardless of selection type used. Thus, each selection type identified different lines which showed similar performance. It suggests that selected S1 and testcross progenies can be considered promissory because of per se S1 progeny performance that, in average, were equal or superior to S1 parental progenies and, with high hybrid combination potential. Testcrosses (S1' BR-106 x 23-2B) were 16,2 % superior to the average of hybrid checks, and testcrosses (S1' BR-105 x 14-4B) were only 3 % inferior to the average of commercial hybrids used as checks. The inbreeding depression of the selected lines were higher on that derived from BR-105 population than on BR-106, except for the R trait. Among selection types, the lines selected through intrapopulation half-sib progenies were less affected by inbreeding depression for PE, AP and AE traits, than the other types of selection. On the other hand, for P, ES and R traits the lower inbreeding depression effect was observed on S1 selection for both populations, indicating the potential of these populations for inbred-hybrids programs.

Portuguese

BR-ESALQ 1996 CAMARGO D f

Thesis Collection

Thesis Supported by CIMMYT