Investigation of yield and persistence of white clover using cluster analyses

Abstract

A visual growth score method was used to monitor the growth performance of 24 populations of white clover (Trifolium repens) over a three year period. The mean growth scores for each year were used as population x year data for cluster analyses. Cluster analysis methods based on a regression model where genotypes were grouped for similarity of both intercepts and slopes (Method 1) or slopes only (Method 2), were able to differentiate the persistent populations from those that showed a yield decline over time. The deviations from regression mean square was used as an indication of stability. Method 1 was able to cluster the high yielding persistent populations into a separate cluster, thus facilitating the selection of desirable populations. Deviations from regression identified the stable populations from the cluster. In combination with the simple growth score system, this analytical method is a powerful tool for evaluating a large number of breeding families of perennial pasture clovers and grasses for yield and persistence.