Estudos sobre tamanho de parcela para experimentos de melhoramento de café

Abstract

A study to determine the most efficient plot size for coffee breeding experiments is reported in the present paper. Yield data of a 25 selected coffee lines experiment collected for 6 successive years were used. Three different methods were used efficiency to detect differences at a 5 per cent level (Harris et al equation), variance components, and taking into account soil variability through Smith formula. For any number of selections maintaining constant the experimental area the single bill plot was shown to be the most efficient. For single bill plot the efficiency increases as the number of varieties or lines increases. When the running costs are considered two bill plots were shown to be recommended.