When Mentally Retarded Children (MRs), taken into the special class, make improvements in their mental health, their school activities, or academical training, the effect of the improvements should in some way appear in WISC scores. Thirty eight MRs (including 15 pupils of the writer) were followed up for about 2-6 years. Results (1) IQs increased (WISC VIQ 60-66, PIQ 64-67; Suznki Binet IQ 56-59.(2) Performance dominace diminished.(3) Increase of SS (scaled score) was great (more than 1.5) in Comprehension, Digit span, Picture Completion, and Maze. Arithmetic had also a considereble SS increase. A moron boy of 13 years attained 10 marks of RS (rawscore), 7 of SS.(4) Vocabulary and Digit Symbol gave evrey little SS increase.(5) From a 1-year follow up of normal Kindergarten children, we had findings: 1) SS increase was great in Digit Span, Block Design, Object Assembly, Digit Symbol. 2) Comprehension and Picture Completion gave little SS increase. 3) Verbal dominance disappeared.(6) Generally speaking, the less the RS of the 1st test, the more the SS increase was likely to happen.(7) In many cases a single subtest score could distinguish borderlines, morons, imbeciles; while Block Design could less do it.(8) Looking at the individual cases, it was hard for the writer to find systematic relationships between subtest profiles and clinical observations. Interpretations (1) SS increases are assumed to have resulted from following factors: 1) Improvement of mental health, and of verbal activities.(espeially in Compr.). 2) Improvement of academical training.(esp. in Arithm.). 3) Learning effect of the test itself.(esp. in Maze, and in many P tests in the follow up of normal children). 4) A special factor with Digit Span. It was found out elsewhere by the writer thet in MRs of more than 4.5 years of MA, a sudden increase of Span to a certain limit (5 forward, 3 backw.) is likely to be seen, and that between 5-7 years of MA, Span and intelligence has almost no correlation.(2) SS increase by the factors 3) and 4) isof lettle meaning.(8) We may atribute P-V discrepancy to different conditions: emotionally disturbed, hospitallized, etc., but never to any IQ level of WISC. In the process of WISC scaling, it is a logical impossibility that any level of IQ might have one-sidedly P of V dominance. It was once supposed that low IQ might be P dominant, and high IQ vice vice versa. But Wechsler himself lately disproved it. Both tests may equally refer to the g-factor, except that V tests would have more to do with emotional conditions.(4) Here we may classify the mental ability into3groups. A: Habitual skills that have become crystallized, as Cattell expresses it, as the result of earlier learning application of more fundamental general ability to some fields. It reminds us of “Assoziationsverbindung” by Jaspers. B: Ability to adapt to new situations, where crystallized skills are of little help. Fluid intelligence, as Cattell calls it, reminds us of Jaspers's “Aktverbindung” C: A rather mechanical effort of vigilance which is needed in Digit Span and Digit Symbol. Both tests are affected directly by fatigue or intoxication. MRs are deficient decidedly in both of them.(5) Those tests which gave not substantial SS increase show the limitation of MRs in2aspects: B and C. Those which gave much SS increase by factors 1) and 2), indicate the possibilities of MRs in the fields of A.