Abstract
A modified computation of the standard error of adjusted yields is de-scribed which leads to a simplification of the test for differences between adjusted treatment means. Extension to expts. of factorial design is also discussed.