Determination of the location and number of test environments for a wheat cultivar evaluation program

Abstract

Pattern analysis procedures were used to delineate groups of test environments which elicited similar relative yield performance patterns in sets of wheat cultivars. In each of four years there was a tendency for environments from particular regions to be grouped. However, the composition of these groups varied across years. The relative values for the variance component estimates from data for two subsequent years tended to support the conclusions suggested from the pattern analysis. A phased yield evaluation system is suggested. It is derived from a consideration of the lack of consistency of similarity of cultivar response patterns across regions and the need to evaluate large populations in early generations. This system would tend to maximize discrimination among cultivars on the form of their cultivar x environment response patterns.