to have corrected the inaccuracies in diagnosis resulting from the old Binet-Simon scale. Some color is lent to this supposition by the work of Rossy 16 who uses "the below twelve-year basis" as diagnostic of feeble-mindedness in adults, the mental age being arrived at by the Yerkes-Bridges scale. He says, "No subject is diagnosed feeble-minded unless he grades below twelve years mental age by the Point Scale."

This lack of definiteness on the authors' part has lead others to take a specific number of points as diagnostic and this seems to grow out of a suggestion by the authors (Note 5 ante), (p. 93). Seventyfive points may be regarded as the upper limit of feeble-mindedness for adolescents and adults. This is the view taken by Woolley (Note 15 ante), who considers adolescents falling below 75 points as mentally defective, and those between 75 and 85 as being borderline cases.

Another interpretation has been taken by Haines17, who uses as the limit for feeble-mindedness 25 per cent below the normal for the individual's chronological age. This is, of course, the I. Q. method using points instead of years. Applying the three and two-year retardation standard as diagnostic of feeble-mindedness to the results as given by Yerkes (computed from Table 3, p. 52 and Table 8, p. 64. Note 5 ante), we have the number and per cent feeble-minded at each age as shown in Table VII. Just as with the Binet-Simon scale, we see here the varying amount of feeble-mindedness at each

TABLE VII.

SHOWING THE NUMBER AND PER CENT FEEBLE-MINDED BY 1, 3 YEAR RETARDATION ABOVE 9 AND 2 BELOW; 2, 4 YEAR RETARDATION

16 Rossy, C. S. First Note on a Psychological Study of the Criminals at the Massachusetts State Prison. Bulletin No. 13 of the Mass. State Board of Insanity. September, 1913, p. 12.

"Haines, Thomas H. The Illinois Medical Journal.

Mental Examination of Delinquent Boys and Girls.
October, 1915.

age, ranging from 8.1 to 0 per cent, disregarding age 15, with Englishspeaking children. With non-English speaking children the fluctuation from age to age is still greater, from 17.4 to 0 per cent. This gives a total percentage of 4.6 for English-speaking and 9.8 for nonEnglish speaking who are feeble-minded, and a total of the two groups combined of seven per cent feeble-minded.

The distribution of feeble-minded on the four and three-year retardation basis is also given in Table VII. This gives a total percentage of 3.05 feeble-minded for both groups and so corresponds in a way to our hypothesis, but the fluctuations of the percentage feeble-minded from age to age are again too great and indicative of a faulty method of diagnosis. Turning to the I. Q. method as used by Haines (Note 17 ante), we show in Table VIII the number of points 25% below the norm for each age or the I. Q. of .75 and the number and percentage of cases falling below this standard. This table is computed from the norms as given by Yerkes (Table 3, p. 52

TABLE VIII.

SHOWING THE Score on YERKES SCALE AT THE I. Q. .75 AT EACH AGE,
NUMBER AND PER CENT OF CASES FALLING BELOW THIS SCORE

and Table 8, p. 64. Note 5 ante). Here again we see the percentages of feeble-mindedness varying from 3.8 to 18, and giving a percentage of 11.5 for the whole group.

Assuming the normal curve of distribution and the five divisions, the various limits for feeble-mindedness, backwardness, normality, etc., on the Yerkes-Bridges scale have been computed by us with the available data in the same way as these limits were computed for the Binet-Simon scale. These points are shown in Table IX. Here again the accumulation of more data on unselected groups would probably alter the limits somewhat, but the additional results would make more certain the limiting points and add to the diagnostic value of the scale. Just as the addition of more unselected cases will

TABLE IX.

SHOWING THE RANGE IN SCORE ON THE YERKES-BRIDGES SCALE FOR THE FIVE GROUPS FOR EACH CHRONOLOGICAL AGE.

tend to correct the norms for each age of the Yerkes-Bridges scale, so the limiting points for each group as computed by us could be corrected from time to time.

Application of the Method to Tests of Delinquents. Although the limiting points on the Binet-Simon and Yerkes-Bridges scales for the five groups suggested by our hypothesis cannot lay any claim to finality, in view of the comparatively small number of cases at each age in reference to such a hypothesis, and in view of the different methods adopted by different workers, we have nevertheless used these limiting points (Tables V and IX) to diagnose groups of delinquents and to compare the diagnoses on this basis with the diagnoses arrived at by seven different authors. The data have been taken from studies of the mentality of delinquents by Kohs 18, Hickman 19, Jennings and Hallock 20, Pintner21, Renz22, Crane, and Haines (Note 17 ante). All of these studies with the exception of the last are based upon the Binet-Simon scale alone. In Haines' study we have the mental level for each child as determined both by the Binet-Simon and the Yerkes-Bridges scales. The actual result for each case is given and this adds decidedly to the value of the study. There are other estimates of the percentage of feeble-mindedness among delin

18 Kohs, S. C. The Practicability of the Binet Scale and the Question of the Borderline Case. Publications of the Research Department, Chicago House of Correction. Bulletin No. 3, November, 1915.

19Doll, E. A. Supplementary Analysis of H. B. Hickman's Study of Delinquents. Training School Bulletin. Vol. XI, 1915, p. 165.

20 Jennings, H. M. and Hallock, R. L. Binet-Simon Tests at the George Junior Republic. J. of Ed. Psych. Vol. IV, 1913, p. 471.

21 Pintner, Rudolf. One Hundred Delinquents Tested by the Binet Scale. Ped. Sem. Vol. XXI, 1914, p. 523.

22Renz, Emilie. A Study of the Intelligence of Delinquents and the Eugenic Significance of Mental Defect. M. A. Thesis. Ohio State University, 1913.

23 Report of the Commission to investigate the Extent of Feeble-mindedness, Epilepsy and Insanity in Michigan, 1915.

quents that are not available for this method, because of the fact that the distribution by mental age is not given. It would have been interesting to take the results of such writers as Faulkner or Hickson, whose estimates of feeble-mindedness among delinquents are unbelievably and absurdly high. The former found 74 per cent. feeble-minded among 480 men in the Kansas State Penitentiary, and the latter 84 per cent morons and 8 per cent borderline out of 245 boys in the Chicago Psychopathic Laboratory.

TABLE X.

SHOWING THE PERCENTAGES OF CASES FEEBLE-MINDED, BACKWARD OR
NORMAL, AS ESTIMATED BY SEVEN DIFFERENT AUTHORS AND AS
ESTIMATED BY THE 3% HYPOTHESIS.

In Table X we give the percentages of feeble-minded, backward and normal cases as arrived at by the respective authors and also as determined by the diagnosis according to the limits deduced from our three per cent hypothesis. The differences between these two estimates of feeble-mindedness are given in the last column. In every case it will be noted that the percentage feeble-minded is lower by our hypothesis than by whatever method the author may have adopted. This may lead one to some such general conclusion as this, that at present we are subjecting our delinquents to much more severe standards of intelligence than we apply to non-delinquents, or that if the percentage of feeble-mindedness among delinquents is such as the studies in question give, then the percentage of feeble-minded among non-delinquents is much higher than we customarily suppose to be the case. In some studies the discrepancy between the two estimates is very large, as in the case of Hickman, Jennings and Hallock, Pintner and Renz. The distribution of Crane's cases is interesting. We have worked these out for our five divisions and as there are 1,187 cases, we have a fairly large number with which to deal. Diagnosed by our method we have the following distribution:

F. M.-11.3%, Backward-23.8%, Normal-44.0%, Bright-16.9%, Very Bright-4.1%. This is a fairly normal distribution, giving about 50% in the middle group, but skewed to one side as we should expect with a group of delinquents.

In the case of Haines' data, we have the possibility of a comparison of percentages of feeble-mindedness derived from the BinetSimon and Yerkes-Bridges scales. Our percentages as derived from the Binet-Simon scale are much higher than those derived from the Yerkes-Bridges scale, and it is interesting to note that the percentages of Haines as derived from the Yerkes agree fairly well with our figures derived from the Binet, but disagree radically with our percentages as derived from the Yerkes scale. This seems to point to some radical difference between the two scales, which it is not the purpose of this article to discuss. It is however worth noting that by applying Haines' method of diagnosis to the unselected children of Yerkes-Bridges, we arrived above at a percentage of 11.5 feebleminded (see Table VIII). If there are 30.2% feeble-minded among the 880 delinquents, then there are on the same basis 11.5% feebleminded among the 670 presumably non-delinquent school children tested by Yerkes and Bridges. A percentage of 11.5 would be admitted as a very high figure for feeble-mindedness among unselected school children. On the hypothesis of three per cent feeble-minded among the school children, we arrive at 6.6% feeble-minded among the delinquents, or nearly twice as many. This is a very low estimate in comparison with the usual opinions concerning feeble-mindedness among delinquents and we would not maintain that it is correct. We do not believe that the number of cases tested by Yerkes and Bridges is sufficient, and this refers particularly to the higher ages, to ensure the reliability of the limiting points which we have computed. Furthermore owing to the lack of cases above age fourteen, we have applied the fourteen year old limits to all cases above age fourteen and these may be too low for individuals above age fourteen.

Conclusion. We offer these diagnostic tables for the BinetSimon and the Yerkes-Bridges scales merely as indicative of a method which might successfully be followed out in order to increase the diagnostic value of intelligence scales. This mode of procedure is, of course, dependent entirely upon the acceptance of some hypothesis in regard to the division of individuals into groups and the agreement of some percentage to denote the number of individuals in each group. In regard to feeble-mindedness in particular, we believe that much. would be gained by the acceptance of a one, two, three or four per cent hypothesis.