The objective of this work was to evaluate the efficiency of reciprocal recurrent selection in producing hybrids of popcorn (Zea mays) from superior endogamous progenies, with full-sibling families. The reciprocal recurrent selection program involved the 'Viçosa' and 'Beija-Flor' popcorn populations. Tests of the hybrids S0xS0 and S1xS1 were carried out in lattice design, in the agricultural years 2002/2003 and 2004/2005. The two assays included common commercial controls, allowing for the comparison of the hybrids performance. Productivity and expansion capacity were evaluated. Direct and indirect gains with selection were predicted on expansion capacity assessed in hot air and Metric Weight Volume Tester popcorn makers. The gains observed were calculated to evaluate the efficiency of the reciprocal recurrent selection. The covariance analyses showed genotypic variability in the populations. When compared to the commercial controls, the S0xS0 hybrids were inferior in quality and equivalent in productivity; and the S1xS1 hybrids were equivalent in both traits. The observed gain in expansion capacity was substantial (around 4 mL g-1) and higher than the predicted gain (2.8 mL g-1). The reduction observed in productivity was not relevant. Reduction in the genotypic variability was verified for expansion capacity. The reciprocal recurrent selection method was efficient in producing S1xS1 hybrids superior to S0xS0 hybrids.
interpopulation full-sib families; genetic gain; cryptic hybrids method
