Estratégias de modelagem na análise de experimentos em delineamento de blocos aumentados em testes clonais de Eucalyptus spp.

Abstract

The objective of this work was to compare analyses of experiment strategies when there is a large number of clones and a reduced number of seedlings to be evaluated. Data from girth at breast height of two seasons of evaluation, 30 and 90 months, from a clonal test of Eucalyptus were analyzed in three locations. The experiments were carried out in the augmented block design with 400 regular clones distributed in 20 blocks and with four common clones (controls). Each plot consisted of five plants spaced 3 x 3 meters. The individual statistic analyses were carried out by season and local, a combined one by local at each season and a combined one involving the three locals and the two seasons. Each analysis was carried out according to two models: augmented design (AD) and one way classification (OWC). The variance components, the heritability, the Speaman’s rank correlation and the coincidence indexes in the clone selection at the two models were estimated. It was found that the augmented block design and the one way classification provide similar results in eucalypt clone evaluation. The coincidence indexes between the two models in the clone selection, in general, were high, showing values of 100% in the local combined analyses at 90 months. The Spearman’s rank correlation showed estimates in accordance with the coincidence indexes. It was also checked that the clones by seasons interaction was expressive and the coincidence indexes in clones selection at 30 months with selection at 90 months in the combined by local analyses were from 42% in the OWC and 47% in AD, when a selection intensity of 5% was applied.