The investigations by Pearson of children in the regular school classes indicate that there is no important shift with maturity in the frequency of those with different degrees of ability, when the ability is measured either in terms of years of excess or deficiency with the Jaederholm form of the Binet scale or in terms of estimates of ability relative to children of the same age (_166_ and _167_). In both these studies the correlation of ability with age was shown to be almost zero. For tested ability for 261 school children “r” was .0105, P. E. .0417; with the estimated ability, the correlation ratios were for 2389 boys, .054, P. E. .014; for 2249 girls, .081, P. E. .014. Until we have better data this is certainly the most authoritative quantitative answer to the question of the shift with age in the frequency of the same relative degree of mental capacity. The best method of empirically settling this question of the early appearance and constancy of deficiency would be to test the same group of children again after they had reached maturity and find out how many of those who tested in the lowest X per cent. still remained in the same relative position. This is, of course, not possible at present, but it certainly should be done before we are dogmatic as to the permanent isolation of the lowest X percentage at any age. The nearest approach to this sort of evidence is Goddard's three annual testings of a group of 346 feeble-minded children with the Binet scale (_117_, p. 121-131). Among these 109 showed no variation, 123 gained or lost 0.1 or 0.2 year, 18 lost 0.3 or more, and only 96 gained 0.3 or more of a year. With so small a change in absolute tested ability the probability of a change in position relative to normal children seems to be slight. Only one of the 76 who had tested in the idiot group gained as much as a half year in tested age in three years. It is not possible to settle this question of the constancy of the percentage of intellectual deficiency from one life-age to the next by considering the frequency of different ages of children among those who are sent to special classes for retarded pupils. This is evident from the fact that these classes contain a considerable proportion of those who are feeble mentally mainly because of conative disturbances. These would not be detected by our present tests and would not be classed as _intellectually deficient_. In the second place the pupils for the special classes are usually selected mainly on the advice of their teachers, who cannot, of course, without tests select those who are intellectually deficient except by trying them for a number of years in the regular school classes. This means that a smaller percentage of pupils in the special classes at the younger ages is to be expected.