The modified augmented design, recently proposed as a method of adjusting for environmental heterogeneity where large numbers of cultivars are being tested in non-replicated plots, was used in a potato breeding programme at four locations in 1982. Of three methods of adjusting for soil variation, only the row-column and covariance adjustments worked efficiently. Decisions about which adjustment method to use could be based on easily computed mean squares. Changes in ranking due to adjustment of test cultivars were large in three of the seven trials, but application of inappropriate adjustment methods produced large and misleading ranking changes in some trials. Because of ease of use and ability to investigate environmental variation, the design can be recommended for use in early stages of potato breeding programmes.
Read full abstract