Desbaste seletivo em teste de procedências e progênies de Astronium fraxinifolium Schott com base na variabilidade genética

Abstract

The objective was to propose a management plan for provenance and progeny tests of Astronium fraxinifolium Schott. It is based on knowledge of genetic variability in an effort to promote the conservation of their genetic potential for commercial and environmental purposes. The test was originally conducted in 1996 in Selvíria, Mato Grosso do Sul (MS) Brazil using an experimental of randomized complete block design. We applied a total of 30 treatments (progenies) of provenance in Ilha Solteira, São Paulo (SP) and 30 treatments of provenance in Selvíria, MS each having five repetitions, 10 plants per plot, spaced 3m ? 1.5m in alternate plantings with Jacaranda cuspidifolia. The traits DAP (height/diameter), ALT (total height), DMC (average crown diameter) and FOR (stem form) were measured 18 years after planting and sex of individuals. The estimates of variance components and genetic parameters were obtained by employing mixed models RELM/BLUP (Maximum restricted likelihood/best non-vitiated linear prediction). They detected significant differences among provenances for DAP and ALT. Analysis of each provenance indicated there were significant differences between and within progenies for all the traits, except FOR. The heritability coefficient of average level progenies ranged from 0.43 (FOR-MS) to 0.78 (DAP-MS) and individual heritability from 0.08 (FOR-MS) to 0.37 (DAP-MS). The accuracy was high (> 0.65) and the coefficient of genetic variations were higher than 5.6. From the total number 11% and 16% were flowering from SP and MS provenances, respectively; with a predominance of individuals with male flowers. The selection strategy within progenies, with selection of 50% of the individuals based on the DAP trait, revealed low gains. However, this test is more suitable because it may possibly lead to smaller alterations of the genetic variability of the population and reduce the likelihood of crossings between relatives. Therefore, this proposal for selective thinning is most appropriate because the flowering percentage of individuals is low, making it impractical for the determination of sex of total individuals. Consequently, the estimates of the parameters accurately affect size.