It is to be noted that if a sudden change is found in the percentage of children falling below a certain test standard it is perhaps more likely to mean that there is a change in the difficulty of the tests at that point. For example our Table V shows 1.3% of the nine-year-olds test two or more years retarded, while 18.9% of the ten-year-olds are retarded two years or more. This presumably indicates a change in the relative difficulty of the tests for VII and VIII rather than a change in the frequency of retardation at ages nine and ten. When we turn to Goddard's norms for VII and VIII we find that 81% of the seven-year-old children pass the norm for VII while only 56% of the eight-year-old children pass the norm for VIII. The Jaederholm data (_167_) obtained by applying the Binet tests to pupils in the regular school classes and in special classes for the retarded may suggest a possible influx of intellectual deficiency at about 12 years of age or else “more mental stagnation in the intellectually defective” at this life-age and after. If one were to define intellectual deficiency in terms of the standard deviation of the regular school children, this data suggests that there is a marked increase in the number of children sent to the special classes at 12 years of age who are -4 S. D. or lower. Roughly speaking it amounts to 36 children at 12 years of age, 36 at 13, and 21 at 14, as compared with 11 at 11 years and 13 at 10 years. On the other hand, this may as well mean that intellectual deficiency becomes greater in degree rather than in frequency at these ages. The latter interpretation is adopted by Pearson for the Jaederholm data, so that it is perhaps not necessary to consider this evidence further. On the average the pupils in the special classes fall about .3 S. D. months further behind regular school children with each added year of life from 5 to 14 inclusive. A third possible interpretation of the greater number showing the degree of deficiency measured by -4 S. D. with the older ages should be mentioned. It is possible that 1 S. D. has not the same significance for 5-year-olds as for 12-year-olds. The distribution of abilities at succeeding ages may be progressively more and more skewed in the direction of deficiency. We shall return to this point in Part Two as showing the advantage of the percentage definition over a definition in terms of the deviation. In connection with the Jaederholm data on special classes one should also consider the fact that younger children are not as likely to be detected by the teachers and sent to the special classes. It is possible also that the difference in difficulty of the tests for different age groups is somewhat obscured by using a year of excess or deficiency as a constant unit as Pearson has in treating this data. The bearing of this difference in difficulty was pointed out above for Goddard's data.