for the statistical method to be carried out with greater perfection to enable us to measure the services rendered by special instruction. It is probable that the special school would render greater services than the ordinary school, because it has greater advantages: teachers experienced in the training of defectives, a curriculum better fitted to the aptitudes of the latter, and, most important of all, the possibility of individual instruction. Let us stop here. In the meantime this is all that we can say with regard to the organisation and control of special education. If we were to attempt to go farther we could do so only on _a priori_ grounds. The time has come for experiment. The new classes which are being formed in Bordeaux, Paris, and elsewhere, must be carefully watched. We shall grope, we shall make attempts, certainly we shall commit mistakes, which will not matter very much if only we have the mind to recognise them and the courage to put them right. The essential thing is for all the world to understand that empiricism has had its day, and that methods of scientific precision must be introduced into all educational work, to carry everywhere good sense and light. FOOTNOTES: [14] In trying to explain this success, one must, no doubt, take into account the comparatively advanced age of the children, the probable leniency of the examiners, and, above all, the fact that the ill-balanced subject is a moral rather than a mental defective. APPENDIX [N.B.--Throughout the Appendix Roman numerals refer to ages--_e.g._, IV. 2 = second test for children of four years.] Part of the interest of this work on defective children consists in the fact that in it we find the origin of those ideas and investigations which culminated in the formation of the Binet-Simon Scale of Intelligence, now so widely known throughout Europe and America. The ideal that Binet set before himself was the formation of a scale which should measure intelligence in something the same way as the foot-rule measures height. The first difficulty was the unit. If we regard intelligence as the power to cope with a situation, we see that this power in a general way increases with the experience of the child, or, we may say, with his age. A child of nine should have more intelligence than one of eight, a child of eight than a child of seven, and so on. We may suppose, then, that there is a normal intelligence for each age just as there is a normal height for each age, although in the first case, as in the second, many children fall below and many rise above the standard. It is clearly by no means so easy to establish a norm for intelligence as for height, nevertheless, the method should be the same; that is, we should begin by finding out what the intelligence of children of different ages actually is, and from these results we should derive averages which might be used as norms.