Restricted selection indexes used in sour passion fruit intrapopulational recurrent selection

Sour passion fruit has great relevance for Brazil by its socioeconomic importance. Although it is one of the most important fruit trees in the Brazilian fruit production, its productivity is still very low due to the lack of genotypes that are homogenous, productive, and adapted to the different regions of the country. This denotes the need for advances in genetic improvement of this fruit species. In this context, the objective of this work was to estimate genetic gains in sour passion fruit progenies under intrapopulation recurrent selection, using different restricted selection indexes: Kempthorne & Nordskog, Tallis, and James; and non-restricted selection indexes: Smith, Hazel, and Mulamba & Mock. Thus, 13 agronomic characteristics were evaluated in 118 full-sib families and three controls. The selection indexes used to obtain the genetic gains were Mulamba & Mock, Smith, Hazel, and Kempthorne & Nordskog, with the assignment of random weights. The Tallis index was applied to the restriction for fruit mass, and the restriction of James for productivity gains. The Genes program was used to analyze the selection indexes. The restricted selection indexes of Kempthorne & Nordskog, Tallis, and James are more suitable for the selection of full-sib families of sour passion fruit than the Smith, Hazel, and Mulamba & Mock non-restricted selection indexes, and can be applied in recurrent selection programs. The results indicate that the addition of the restrictions proposed by Tallis and James are not efficient for increasing gains when compared to the Kempthorne & Nordskog, since they obtained similar gains.