Effect of population size and yield level in selection of diagnosis and recommendation integrated system (DRIS) norms

Abstract

Abstract A data bank comprising over 8000 observations of yield and tissue composition was assembled from a wide range of locations and conditions. DRIS norms were developed from parts of the data bank representing different sources and chronology in its development. The effects of population size and yield level used to discriminate between high and low yielding subpopulations on the DRIS norms are studied. The results indicate that the best banks are those which are large, random and have a substantial number of high yield observations.