ABSTRACT
This chapter applies the robust sphering to the soybean data by considering the raw (laundered) data in which every variable is in its original scale. The first step is an ordering of the attributes from most to least importance. Such a ranking always depends on the population under study and can only be obtained from the research scientists concerned. The best plant breeding population is one that has good genetic variation for the character (attribute) to be improved and none in the rest. It depends on the parents used in the crosses. The soybean genotypes assigned to each of the maturity groups are presented in a table.
