Then the ratio of achievement to intelligence is in definite relation to intelligence—a negative relation. It is this same tendency to adapt our education to a low level which has prevented a perfect association between intelligence and the various subjects. The relation of one subject to another, irrespective of intelligence, would be zero if there were no other factors except intelligence responsible for the product. After two years of such attempts as an ordinary public school will allow, we have removed many of the causes of disparity and increased the association between potential progress and progress in arithmetic, reading and language. The correlations, correspondence of S.D.’s, and Σ(IQ-EQ)⁄_n_ registered in Tables 9, 10, and 11 give evidence of this as does also the increase in the AccR, an average of the Arithmetic, Reading, Vocabulary and Completion Ratios. (Table 13.) Are the unremoved causes other than intelligence unremovable? These causes might be, besides the unreliability of tests and the common elements in the tests, the specialized inheritance we have considered, ethical qualities of endurance, ambition, initiative and industry or a general factor. The correlations between Arithmetic Ratios and Reading Ratios and the other intercorrelations of Subject Ratios will yield us an index of how much of this remaining disparity is due to specialized inheritance. These intercorrelations for all years are embodied in Table 13. The partial correlations of quotients when intelligence is rendered constant will be found in Table 6. These intercorrelations, and the partials as well, give an indication of some general factor other than intelligence since the _r_’s irrespective of intelligence are uniform and all are positive. Only the correlation of arithmetic with vocabulary, intelligence being rendered constant, goes to zero. Though this might be due in part to common elements in the tests, it is more likely that there is another factor in operation. Inheritance of specific abilities could not have this uniform effect on the correlations. These correlations all being positive and the _r_’s being very uniform, both correlation of ratios and the partials, makes the interpretation of specialized inheritance of ability extremely unlikely. The correlation of Arithmetic Ratios with Reading Ratios is higher in 1920 than that of Vocabulary Ratios with Reading Ratios. It leaves the possibility that the unremoved factors are inherited ethical differences or that they are a “general educational factor.” The negative correlation of AccR with intelligence, however, being as high as these positive remnants of interrelation, would tend to make more probable an interpretation of this as a remnant of disparity, intelligence accounted for, which is entirely due to the organization of our schools.