Information from a maize trial (32 single hybrids, 12 triple hybrids, and one double hybrid) repeated in four locations was analyzed using multivariate cluster ana1ysis (average linkage) and principal components analysis. Results proved that the methods are complementary, and result in basicaIly the same classification of the evaluated genotypes. The clustering general tendency between the germ p1asm under evaluation was due to the similarity of the phenotypical traits measured. The lowest degree of association was due to the relationship of the material studied, because the three superior hybrids had a common parent (H-513), or one parentalline from it. The three best hybrids (TTC-17 x POB43 C6 HC232-2#-1-2, AC8149-2-1-1-3 x H-513, H-513 x POOL23 TSR40-2-3-2) had a higher yield than the experiment's general average (5595 kg/ha), and were al so a littIe earlier (male flowering days and female flowering days) than average, but had a similar plant and ear height. Plants and ears of hybrids 2 of lower yield (T-37 x (LRB-14 x 0-539), 28 (D-471 x V-524- 72-1) y 36 (POB INT DENT 5-1-1-2-HC1 x POB 43 C6 HC232-2#-1-2), were al so characterized as being too high or too short.