ABSTRACT
Initial insight comes about most rapidly by looking at scatter plot matrices for the six attributes and each environment. The order of the attributes in the scatter plot matrices (SPLOM) was based on a greedy close ordering, while their sign was determined from correlation coefficients amongst the attributes for all 58 genotypes. Another way to present other aspects of the same information is to use SPLOMs across environments for each attribute. Here, as elsewhere, when pictures for different environments are similar, there has to be a possibility of showing a main picture and supplementary pictures. The intensity with which the individual plots show a tendency for correlation of the two attributes varies substantially from attribute-pair to attribute-pair and from environment to environment.
