That mental ability, and especially mental deficiency, is primarily a question of inherited capacity rather than training, is now indicated by a number of fundamental objective studies of the correlation of abilities within the family, which have been analyzed to show the relative influence of inborn and external factors. Among these studies Thorndike's investigation of the tested abilities of twins compared with brothers and sisters in the same family is the most objective, and is very convincing (_199_). He has also summarized the evidence so well that it is not necessary to go into the question here (_198_). One of the most important facts is that equal practise under the same conditions increases the difference between individuals rather than makes them more alike. The work of the English biometricians appearing in Biometrika and the monographs from the Eugenics Laboratory is the most important in this field, and cannot be summarized here. It includes family resemblance in both pathological and healthy mental traits (_126_). As compared with these studies the attempt to show that feeble-mindedness is inherited, because many of those in institutions for the feeble-minded are from families showing mental taints, lacks cogency, since we are still uninformed as to what portion of the offspring of parents with and without deficient minds are deficient. Even if 85% of the children in institutions for the feeble-minded have tainted parents this does not mean that we know what percentage of deficient parents have deficient offspring. It is this latter fact that we must know in order to predict the danger of defective offspring from deficient parents. From what we know about the correlation of parents and offspring in mental ability, it is clear that the more deficient are the parents, the more likely it is that their offspring are deficient. Children of morons are, therefore, not so likely to be deficient as are children of parents with lower grades of ability. From the eugenic point of view, it is, therefore, most important first to protect society from propagation by the lowest grades of deficients, provided that all grades of deficients are equally likely to have children when left unrestrained in society. Since mental and moral qualities are probably correlated positively, the same emphasis would be placed on first isolating the lowest grades in order to reduce inheritance of criminality. The eugenic emphasis waits, however, on the discovery whether the greater tendency for the lowest types to be produced by the lowest types is overbalanced by any tendency of deficients or delinquents of lower degrees to be less productive when unrestrained in society.