ALLOCATION OPTIMALE DES RESSOURCES POUR DES ESSAIS DE CEREALES AU QUEBEC

Abstract
Yield, height, heading and maturity dates and lodging index data for a selection of cultivars from the cereal (barley, wheat and oats) cultivar tests of the Conseil des productions végétales du Québec (CPVQ) were used to obtain estimates of the variances associated with cultivar × year, cultivar × location and cultivar × year × location interactions. These estimates suggest that the number of replicates in each test could be reduced to three. Reduction from 3 to 2 yr induces a tolerable increase of the variance of a cultivar mean yield. It would probably be sufficient to observe heading or maturity dates and lodging index in two replicates only.Key words: Genotype × environment interaction, cultivar testing