Total A B C D 2788 Boys 28 308 1441 1011 3489 Girls 33 307 1748 1401 Per Cent of Total .9 9.8 50.8 38.4 This distribution shows a marked tendency for failures in any subject to be accompanied by further failures (38.4 per cent), not only in the subjects for which it is a prerequisite but in subjects closely akin to it. If this tendency to succeeding failures is really dependent upon thoroughness in the preceding subject, then the repetition of the subject should offer an opportunity for greater thoroughness and should prove to be a distinct advantage in this regard. When we compare the percentage of failures above with that in the following distribution, we fail to find evidence of such an advantage in repetition. The continuity of failures by subjects and the ineffectiveness of repetition are pointed out by T.H. Briggs[42] as found in an unpublished study by J.H. Riley, showing that after repeating and passing the subjects of failure, 33 per cent of those who continued the subject failed again the next semester. LATER GRADES IN THE SAME KIND OF SUBJECTS, FOLLOWING FAILURE BUT WITH NO REPETITION Total A B C D 1269 Boys 5 102 639 523 1191 Girls 8 147 669 367 Per Cent of Total .5 10.1 53.1 36.2 Here the same pronounced tendency is disclosed for the occurrence of other subsequent failures in the subjects closely similar. But for this distribution of grades, secured without any preceding repetitions, the unsuccessful result is 2.2 per cent lower than that found for those who had repeated. This group is not so large in numbers as the one above, and undoubtedly there is some distinct element of pupil selection involved, for it is not easy to believe that the repetition should work a positive injury to the later grades. Nevertheless, our faith in the worth of unconditional repetitions should properly be disturbed by such disclosures. _c. The Grades in Repeated Subjects and in the New Work, for the Same Semester and the Same Pupils_ If it is granted that the teachers of the repeaters are equally good as compared with the others, then the previous familiarity with the work that is being repeated might be expected to serve as an advantage in its favor when compared with the new and advanced work in other subjects. But the grades for the new and advanced work as presented below, and the grades for the repeated subjects as presented earlier in this chapter (section 1), deny the validity of such an assumption and give us a different version of the facts. THE GRADES SECURED IN NEW WORK, AT SAME TIME AND BY SAME PUPILS AS THE GRADES SECURED IN THE REPEATED SUBJECTS