The complete and methodical study of the documents relating to 223 children with a retardation of three years has taught us a number of interesting facts. It is very rare for the cause of the retardation to be single. Usually, several causes were at work simultaneously. Feebleness of mind complicated by illness is noted in 20 per cent. of the cases. Insufficient school attendance (due to other causes than illness), in conjunction with feebleness of mind, is met with in 25 per cent. of cases. If, without taking account of those associations of causes, one enumerates simply the frequency with which each single cause of retardation is mentioned, one obtains the following percentages: Feebleness of mind 50 per cent. Insufficient attendance (without illness) 33 " Illness 25 " Lack of application, laziness 7 " If we admit, as a hypothesis, that the frequency of each of those four principal causes indicates its importance, we shall conclude that laziness very rarely explains a retardation so great as three years, and that the most important factor is undoubtedly feebleness of mind. We should have expected the teachers to give much more frequently the banal reason of lack of application. They have not done so, and these results confirm in a quite unexpected manner the convention according to which every retardation of three years should make one suspect feebleness of mind. It would be interesting to know whether any children really defective in intelligence escape the revelation furnished by our tables. We have put this question in writing to the heads of the schools, and they have notified fifteen children, or 6 per cent., who seem to them to be clearly defective, although without a retardation of three years. On testing the statement, we found that mistakes had been made, and the sole residue of defectives who had escaped our census consisted of three subjects who wanted only a month or a few weeks to have shown clearly a retardation of three years. They were therefore on the border, and such exceptional cases are always to be found when one fixes an exact limit. There is no need to worry about them.