We have suggested that two sets of factors are important in the development of the manic-depressive personality: 1. The state of ego development when major anxiety-provoking experiences occur; 2. The dynamics of interpersonal relationships between the family members. How much confidence can we place in the findings which led to these hypotheses? In the first place, it should be emphasized that we worked only with patients who actually had a manic-depressive illness, and our descriptions of the "manic-depressive personality" apply to this group at times when they were free from psychosis. For the most part we worked on the depressive phase of the psychosis, and gave only secondary attention to the manic attack. We do not present any data on the much broader group of cyclothymic personalities.Is it legitimate to emphasize the social and psychodynamic factors in the development and maintenance of these disorders over and above the biological and hereditary? We do not deny, of course, that the behavioral manifestations of the illness are mediated by physiological alterations. However, it is doubtful that genetic and biological factors could so influence the development of interpersonal relations as to account for the remarkably consistent dynamic descriptions of the personalities of these patients. Further, the fairly consistent family relationship patterns lend support to a primarily psychogenic theory. This view is further strengthened by the fact that with successful psychotherapy the patient abandons his stereotyped pattern of envious, competitive, manipulative dependency and moves into a more mature, independent relationship.Nevertheless, it is clear that the family pattern as described could not be a sufficient cause for the patient's specific personality development. Not all children in these families developed similar personalities; many of the patterns were also seen in families of schizophrenic patients. Further, it is characteristic of the American middle class generally, especially of upper middle class families, that they have the status consciousness and concern and the upward mobility which we have described for the families of our manic-depressive patients. Obviously, much more research into family patterns is needed, in normals as well as in the various forms of mental disorder, in other cultures as well as in our own, and in the epidemiology of mental illness in the various social classes.Despite these reservations concerning the reliability of our ideas about them, we believe these family patterns may be influential in the genesis of the manic-depressive personality. Our work to date does not permit us to make a precise statement regarding the mode of action and intermediate steps between the early environmental family influences and the manic-depressive illness. However, inferences drawn from reports of patients in psychotherapy have led us to formulate some tentative hypotheses.As the child moves from the helplessness of early infancy to the more self-assertive and active phases of late infancy and childhood, he will be increasingly exposed to the anxiety engendered by the family's needs for prestige and social conformity as well as to the conflictful situation arising out of envy and competitiveness. At this time the primary closeness based upon the identifications of early infancy will have diminished but the more mature closeness based on the ability to relate to others as individuals distinct from oneself will be in the most rudimentary stage. Thus the child could be expected to feel peculiarly alone and consequently vulnerable to any threat of abandonment. This particular sensitivity and vulnerability persists into adult life leading to a central conflict in the manic-depressive. It may also be that the repression of feelings and discharge of tensions via the manipulative exploitative activity on the part of a parent may provide the child with a model which is later modified in his personality to a hypomanic way of life. In any given case there may be a wide variation dependent upon individual attributes of the infant that serve as stimuli to arouse anxieties in the parents. For example, age, sex, or physical appearance of a child may cause the parent to see the child as a hated sibling, leading to specific kinds of overexpectation and demands. Further, the quality of parent-child interaction may be of specific importance. Thus, the mother-child relationship in schizophrenia has been described as symbiotic; while in the manic-depressive we have visualized it as being less close, with the child serving a somewhat mechanistic role as his mother's instrument. Longitudinal studies of child development will make it possible to check the validity of such hypotheses as these. All that we can offer in their support at present is that such reconstructions in the course of psychotherapy appeared to have been instrumental in the development, of a favorable change in the patient.Although we have stressed the importance of environmental influences as mediated through the significant interpersonal relationships, we do not deny that differences in drive intensity or innate ego strength may be thought of as determining what makes one person a successful extrovert and another a manic-depressive. Predictive description would have to go far beyond what is presently possible, in the direction primarily of the identification and quantitative appraisal of the resources of the personality. Possibly the character and degree of reaction to psychotomimetic drugs such as LSD-25 and the current studies of the responses to sensory deprivation might contribute to the problem of the assessment of ego strength—both in its defensive and integrative aspects.Finally, we believe that the methods used in this study have some merit. The seminar made it possible to review the data from a larger number of cases than can usually be reported in psychoanalytic therapy. By virtue of this review we found ourselves steadily narrowing the number of "positive" findings until at last we proposed those which we thought were most consistently present in the 12 patients. These were subjected to further screening by the survey of the larger sample and the schizophrenic controls. This led to a revision of the description of the family relationship patterns and served to highlight those areas requiring more precise definition and investigation. It is to be hoped that further and more sophisticated elaboration of such cross-checking efforts will help organize the great mass of valuable data which lies imbedded in every psychotherapeutic effort.