The aim of the present work was to compare the efficiency of the lattice design with Random Block Design (RBD) and the latter with Completely Randomized Design (CRD), for plant height and fresh matter weight (biomass) traits in mate (Ilex paraguariensis A. St. - Hil.). The trial was performed at the Estação Experimental do Canguiri, Universidade Federal do Paraná, in Pinhais, Paraná State. Out of 64 half-sib progenies, 52 came from four different sites, 13 from each site, in the State of Paraná and 12 from a site in the State of Rio Grande do Sul. Each progeny consisted of 54 plants (nine plots with six plants per plot). An 8x8 lattice design model with nine balanced repetitions was used to carry out the trial. The analyses were performed also using RBD and CRD models. For plant height, the estimation of RBD efficiency compared to CRD (Êb/i); lattice compared to RBD (Êl/b); and lattice compared to CRD (Êl/i) were 48,6%; 61,0%; and 139%, respectively. The estimation of the intraclass correlation among plants within plots (c²) was 9,4% and zero for RBD and lattice, respectively. For fresh matter, Êb/I, Êl/b and Êl/i were 14,3%, 81,7% and 108%, respectively and c² were 16,9% for RDB and 3,1% for lattice. These estimations coupled with Snedecor's F tests showed high significance for the two traits, indicating efficiency of RBD in relation to CRD greater for plant height than for fresh matter.
environment heterogeneity; incomplete blocks; heritability; mate
