The use of canonical analysis for selection within a population of spring barley

A selection method is described involving the application of canonical analysis to data for a number of measured traits in a composite population of spring barley grown in two seasons. Lines were selected which fell within the Euclidean space defined by the range, for a number of standard cultivars, on each significant canonical variate. Alternatively, limits were fixed for an acceptable range for each trait and transformed to canonical variates. Lines were selected which fell nearest to the centroid of this acceptable space. The method might be used as an alternative to the selection index method where difficulty is often encountered in assigning appropriate economic weightings to the traits measured.