The next higher group in intellectual ability is so high as not to require social assistance outside of school. When we ask how large a per cent. we should be justified in placing in this group and separating merely for special instruction in school, we reach a condition which is at present so ill-defined even in the minds of educators that it seems best to fall back on the general advice that our school systems should provide just as nearly individual instruction as the public purse and managing genius can devise. Mannheim, Germany, for example, takes care of 18 per cent. outside of its regular school classes. The ideal is individual instruction for all. School authorities would be justified in providing special instruction for every degree of mental ability, if the cost would not restrict other more important social undertakings. This less degree of retardation in the group needing only school assistance should not, however, be classed as feeble-minded. We shall see later the percentages for which some authorities have considered it already advisable to provide special school instruction. We need not attempt to estimate the size of this group, as it is beyond the limit of feeble-mindedness. The purely conative cases are not taken care of in the above estimates, which are intended for tested deficients. If the conative cases unaccompanied by intellectual deficiency should be regarded as frequent enough to replace those in the social assistance group who ultimately care for themselves, plus those subtracted by the excessive death rate, we would have a total of 1.5% of the general population feeble-minded enough to warrant social care of some sort. About 0.5% might justly be isolated. The reasonableness of this program can be judged by comparison with authoritative estimates now to be reviewed. The problem here is whether this is an unreasonable program for the diagnostician to assume as scientifically justified, remembering that these estimates are for tested deficients at 15 years of age and do not include purely conative cases which might occur above these intellectual borderlines. E. COMPARISON WITH IMPORTANT ESTIMATES