This study was undertaken in the hope of gaining a better understanding of the causation of persistent criminalism, manic-depressive psychoses, dementia præcox, epilepsy, and mental deficiency. A particular object was to determine the relative importance of hereditary and environmental factors in that causation.The material especially gathered for this study consists of 858 siblings of patients suffering from these disorders. Access was also secured, through the kindness of Rosanoff, to 214 pairs of twins, one or both of each pair being affected by one of these disorders. Of these twins, 138 were dizygotic or probably dizygotic, and 76 were classified as probably monozygotic. In addition, data were secured of the number of persons suffering from these disorders to be found in state institutions, and their percentage in the unselected population.Thus, opportunity was afforded to make the following comparisons:1. A comparison as to incidence of mental disorders in three groups having different degrees of genetic relationship to affected subjects: (a) monozygotic twins of affected subjects, (b) siblings of such subjects, and (c) the unselected population, which is a group including individuals of every degree of genetic relationship to affected subjects but composed mainly of individuals without such relationship.2. A comparison as to incidence of mental disorders in oppositesex twins and siblings of sex opposite to that of their respective propositi. This is a comparison of subjects of the same degree of genetic relationship to affected subjects, with sex as a constant factor. Failure of agreement in incidence is postulated as indicating the presence of some causative factor other than heredity.3. Comparisons between siblings of different ages and between propositi and siblings of different ages.4. Comparison of the occurrence of dissimilar disorders within the families studied, to determine the relationships among such disorders.5. Comparison of the incidence of mental disorders in the two sexes, and in siblings of the same sex as their respective propositi as contrasted with that in siblings of the opposite sex. This comparison brings under consideration another aspect of genetic relationship.By means of the first comparison we have found that, no matter what data we have used, whether of affected subjects or disabled subjects or institutionalized subjects, the finding is the same: the closer the degree of genetic relationship to an affected subject, the greater is the tendency to mental disorder. In those twins judged to be monozygotic, the incidence of mental disorders was as high as 93.5 per cent, with 69.8 per cent in state institutions; in siblings the incidence was as high as 17.8 per cent, with 3.5 per cent in state institutions; while in the unselected population 0.565 per cent only were in state institutions. This would seem to justify the conclusion that genetic factors play a prominent part in the causation of the disorders studied.By means of the second comparison we have found that, of 74 pairs of opposite-sex twins, both were affected in 18 cases. This is a ratio of 24.3 per cent.Of 401 siblings of the sex opposite to that of their respective propositi, we found 59 affected, a ratio of 17.1 per cent.Unfortunately, the number of cases of twins available was not sufficient to permit a statistical study of the individual clinical groups. It seems, however, that the difference in percentages is fully accounted for in the mental deficiency group and does not exist in an amount greater than the probable error of our data in the four remaining groups. Accordingly, the second principal finding revealed by our study is that some prenatal factor or factors other than heredity play a part in the causation of mental deficiency; that such factor or factors are more frequently operative in twin than in single births; and that this factor or factors are apparently not operative in the other mental disorders studied.By means of the third comparison, we discovered that we were obliged to make allowance for age of incidence in the calculation of our ratios, for the reason that some of our siblings were too young to evidence some of the disorders, even if likely to be affected by them. After having made such allowance, we found that differences in age, great or small, between siblings and propositi had no appreciable effect.By means of the fourth comparison, we found many siblings affected with disorders dissimilar to those of their respective propositi. We found several families in which there was coexistence of manic-depressive psychoses and dementia præcox; also even a larger number of families in which there was coexistence of epilepsy and mental deficiency.We found, further, a very frequent familial coexistence of epilepsy with migrainous headaches, enuresis, and outbursts of rage. In a similar way we found a familial association between manicdepressive psychoses and cases of cycloid personality.All this bears out the prevailing belief among physicians that some sort of kinship exists between various constitutional mental disorders, which are clinically rather distinct from one another and conventionally regarded as independent entities.By means of the fifth comparison, we found sex to be a factor in the causation of these disorders. Of the 858 subjects investigated, 414 were males and 444 females. Of these 75 of the males (20.4 per cent—with allowance for age of incidence) and 62 of the females (15.2 per cent) were afflicted with a mental disorder. In other words, incidence is, on the whole, greater in males.Sex is, of course, to be regarded as a genetic factor; therefore this finding may be interpreted as affording further confirmation of our first finding.However, the incidence is not always greater for males in the individual clinical groups, for we find a greater incidence of manicdepressive psychoses and of mental deficiency in the female siblings.In comparing siblings of the same sex as their respective propositi with those of the opposite sex, we find that brothers of males are more often affected than brothers of female propositi, while sisters of female propositi are more often affected than sisters of male propositi; in other words, that incidence of mental disorders is higher in siblings of the same sex as affected subjects than in siblings of the opposite sex. This does not disturb our preceding finding that incidence is higher in males, however, since brothers of female propositi are more often affected than are sisters of female propositi.In presenting this paper, the writer wishes to acknowledge the assistance of his Committee on Graduate Studies: Dr. George H. Mount, Chairman, Dr. Osman R. Hull, Dr. Milton Metfessel, Dr. Albert S. Raubenheimer, Dr. Aaron J. Rosanoff, and Dr. John W. Todd.Particular acknowledgment is due to Dr. Rosanoff for permission to use the data of his collection of records of twins with mental disorders (as yet unpublished). I am indebted to my wife, Kathryn A. Humm, for valuable editorial assistance.Dr. D. Welty Lefever kindly reviewed the statistics of the study.For their cooperation in making accessible complete case records of propositi and case records or other valuable information concerning siblings, the writer is indebted to the following: Dr. C. S. Cronin, Superintendent, Pacific Colony; Dr. W. P. Goddard, Surgeon, San Quentin Prison; Dr. Malcolm F. Hebard, Resident Physician, Los Angeles County Farm; Dr. J. C. Johnstone, Resident Physician, Pacific Colony; Dr. Clara Schmidt, Psychologist, Los Angeles City Schools; Dr. Edwin Wayte, Superintendent, Norwalk State Hospital; Dr. George M. Webster, Superintendent, Patton State Hospital; many other officials of public and private institutions; and, finally, the relatives of subjects of the study and many of the subjects themselves.