Quadrados latinos obtidos por meio de técnicas de confundimento em ensaios fatoriais

Abstract

Latin square designs have been very useful in agriculture and animal sciences. They are, in general, obtained following algebrical rules from basic operations for positive integers and the m-modular number system. In this way, from an experiment installed in a factorial scheme, it is possible to obtain as many latin squares as the levels of treatments, through the use of a confounding technique. These schemes lead to orthogonal replications of latin squares. The aim of this work was to evaluate this technique in order to facilitate plannings. An example of a 3 x 3 x 3 factorial design without replications is presented.