ABSTRACT
For most attributes, there appears to be more variation among subgroups within groups than among genotypes within groups. In the first stage, the dissimilarity between the individual genotypes within each maturity class was calculated using squared Euclidean distance. This was then converted to a similarity measure by subtracting each value from the maximum distance within each maturity class. The ordering of the groups within each maturity class is based on days to flowering. The ordering of the soybean genotypes within each group follows the original genotype number. Similarities in brackets were not used to add an individual genotype to a group and the procedure was halted before then.
