During a post-ANOVA mean separation analysis using Duncan's New Multiple Range test for 32 crop varieties from a field experiments (RCBD with 3 replications), if found that the last letter of the alphabet, z, is exceeded, how best can the significant means (p = 0.05) be separated?