So far as I can discover nobody has directly attacked the specific problem, what percentage of individuals of a given degree of deficiency who are not under supervision, become legally delinquent at some time in their lives. A slight contribution to the empirical study of the problem is made in the reports of the follow-up work in connection with pupils formerly in special classes in the public schools which I reviewed in Chap. IV, f. We have also a telling report by Bullard of the New York Prison Association published by Moore in 1911 (_156_). It follows the records of 85 feeble-minded boys and men 16-29 years of age, paroled from the Elmira State Reformatory in 1904. The whereabouts of 3 were unknown and 2 died. Of the remaining eighty, 31 were arrested again and 6 others violated their parole. One was arrested 19 times in this short period. The best approach to this problem of measuring the potential delinquency among deficients is afforded by Goring's four-fold table for calculating the correlation between deficiency and criminality in the male population of England and Wales (_20_, p. 259). By means of the annual data on first convictions of crime at different ages and the probable length of life among criminals and in the general population he has been able to predict a potential criminality on the part of 7.2% of the general male population. In other words, the best estimate seems to be that about 7 in every hundred males in England and Wales will be convicted of crime at some time in their lives. About 10% of the convicts in England for a series of years have been isolated in prison treatment because of deficiency. If we now also assume with him that 0.46% of the non-criminal population is mentally deficient, we arrive at the table which enables us to determine, on these assumptions, that it is most likely that 63% of the deficients will be convicted of crime at some time in their lives. If instead of taking this estimate of 10% of the criminals being deficient we had taken 20%, then the probability of a deficient individual being convicted of crime would rise to .77. On the basis of our summary of tested delinquents in the last chapter it seems extremely conservative to suppose that 10% of the manifest and potential criminals are as deficient mentally as the lowest 1.5% of the general population. Even with this assumption we find that the chances would be 48 out of a hundred that a person of this degree of deficiency would be convicted of crime. These estimates, I believe, afford a telling argument for the indefinite isolation of at least those who are in the lowest 0.5% mentally on the ground of their potential criminality, independently of any question of the danger to society from the hereditary transmission of the diathesis of deficient delinquency.