A diagnosis of deficiency should not be made until the examiner is able to estimate whether the removal of training or health handicaps would bring the individual above the borderline. So far as known temporary handicaps affect the standard of the test results with groups they should, of course, also be taken into account. On the other hand, it is clear that the borderline which predicts social failure should not be shifted to allow for differences in permanent handicaps whether those be of race, sex or social position. (4) _The Variation among Different Mental Processes._ With our present knowledge the most difficult variation for which we must make allowance at the borderline is the variation from one trait or process to another in the same individual. One phase of it was discussed above under “c.” The investigation of Norsworthy throws light on this question. Summarizing her tests she says: “Among idiots there is not an equal lack of mental capacity in all directions. There is something of the same lack of correlation among the traits measured in the case of idiots as there is with ordinary people” (_159_, p. 68). Again: “The idiots are nearest the central tendency for children in general in the measurements of mental traits which are chiefly tests of maturity, and farther and farther away as measurements are made which are tests of ability to deal with abstract data. They are two and a half times as far from the median for children in general in tests like the genus-species test as they are in tests like the A test or the perception of weight.” Weidensall (_60_) and Pyle (_46_) also compare delinquent and normal individuals for different tests, showing a variation with the sort of mental activity compared. While Norsworthy thus presents evidence of certain specializations of deficiency, she notes, however, that perhaps feeble-mindedness is more typically general than specific and that general deficiency is more important to consider than specific. Even with that test with which her group of retarded and feeble-minded children did best, only 28% of them passed the point which would be excelled by 75% of the children in general. In their worst test only 1% passed this point. It is also to be noticed that those tests in which they most nearly approached ordinary children are for just those simple processes which would be least likely to be of use in the struggle for social existence. As a whole, therefore, there is nothing in her results which shows that any appreciable number of children who were deficient in the average of tested abilities, would have good enough special ability along a few lines to make them socially passable. Indeed, for all that we know at present, the borderline for _passable ability_ in each of our various mental processes might vary quite as much as Norsworthy found, without this variation affecting a prediction of failure based upon the average of a series of tests.