The choice of locality for plant breeding when selecting for high yield and general adaptation

Summary The choice of localities for initial yield testing of selections in a plant breeding programme is often arbitary. We describe a method that allows a more objective approach based on the correlation between the performance of varieties in a locality and their performance over a large area. The method was checked using data from the Western Australian Department of Agriculture regional wheat variety testing programme, and it was found possible to identify combinations of localities which produced significant predictions over both years and genotypes.