Experimental Design Efficiency and Test Ability on Erva-mate Genetic Improvement

Abstract

The perennial plant selection is based on individual genetic values predicted from phenotypic observations. The more adequate the control of environmental variability, the more accurate the prediction. This paper  aimed to investigate: 1) the test ability by using the  significance of Snedecor F test for block effects; 2) the  spatial variability within blocks by using the plot  intraclass correlation coefficient; 3) the experimental  design efficiency in terms of the local control, by using  the significance of Snedecor F test for block effects and  the plot intraclass correlation coefficient (c2),  simultaneously; 4) the genotype x environmental  interaction within site, by using the genetic correlation coefficient across replications. The used data concerned  to leaf weight of 141 erva-mate half sib families, grown in a randomised complete block design  with 10 replications and six plants per plot spaced 3 x 2 meters. The data were collected at age six in the third leaf harvesting. The design efficiency was high and the  test ability adequate, as a result of the significance of  block effects and low c2 value (0.0799). There was a loss of 3.48% in genetic gain due to the genotype x environmental interaction within site.