In the past the research of intelligence tended to try to detect general intelligence from the standpoint of C. Spearman as A. Binet and 1. M. Terman did.It is true that from the viewpoint of Inter-Individual, general intelligence signifies a great deal, but when it is seen from the standpoint of Intra-Individual which analyzes the individual's inner structure, only the detection of general-intelligence cannot accomplish its full purpose.In this sense, 1.J. Cronbach is entirely right in his view that the analysis of structure of intelligence owes greatly to D. Wechsler's Wechsler-Bellevue Scale (1).In measuring intelligence in all its respects, as W. P. Alexander stated, we confirmed that only by general intelligence it was difficult to explain the correlative difference among tests.After that he clarified that in the explanation of intelligence, non-intellective factor "X", "z" as well as general factor "G" must be admitted as important factors (2).H. W oodrow and M. Morris's factor analysis of intelligenee adopting 1.1.Thurstone's centroid method cannot explain intelligence as mere general factor and its result shows 40-60% residual (3) (4).Wechsler, from clinical standpoint, advanced his research on this non-intellective factors by comprising the standpoint of both Binet's general factor and Thurstone's group factor, and he formed the intelligence scale ofW -B, WISC and WAIS.Cronbach criticized this and said: "It is not possible to eat one's cake and have it" (5).I agree with Cronbach's opinion and it is to be regretted that Wechsler's way of thinking has a defect in confusing the standpoint of general factor with that oCgroup factor.It is worthy of special mention, however, that by Wechsler's contribution the intelligence test has come to be applied to the clinics of mental diseases of various kinds.When we synthesize the results that F.M. Gilhooly (6), A.1.Rabin (7), J.R. Schlosser (8), R.M. AlIen ( 9), (10), R. Goldman (11), M.W. McCullough (12), A. Carp (13) and W.L. Hunt ( 14) gained in the research of patients by using W-B, the fact that Information and Vocabulary are stable subtests of "Hold" and that Digit-Span, Picture Completion, Picture Arrangement, Block-Design, Object-Assembly are unstable subtests of "Don't Hold" was nearly ascertained.So, by using the Japanese WISC which contains two branches of "Essential Verbal" hard to reflect, and of "Performance" easy to reflect, and also besides this by using the main intelligence tests made in Japan, viz.Tanaka A, B, Koga, Noken, Katsura, we made factoranalysis of 47 subtests after Thurstone's centroid method.As the result, we found factors of "Hold" and "Don't Hold" and classified our subjects according to Mental Deterioration