Following the classic study of Lindemann (1944) on the various forms and possible outcomes of the reaction to bereavement, little systematic work was done in this field until the last decade. Marris (1958), in a thoughtful but uncontrolled study, interviewed 72 widows whose husbands had died at least 11 months earlier, the mean lapse of time between bereavement and interview being two years and two months. Only 14 of his subjects considered themselves to have recovered completely; 31 reported a lasting deterioration in their health, and 57 reported a continuing difficulty with sleep. More recently, Parkes (1964a), in a study of patients admitted to a psychiatric hospital, has demonstrated that the number of patients whose illness followed the death of a spouse was six times greater than expected, suggesting that bereavement must be one causative factor in the development of the illness. He has also shown (Parkes, 1964b) that widows under the age of 65 consulted their general practitioner because of psychiatric symptoms more than three times as frequently during the six months after the bereavement than they had done during a control period prior to the bereavement. The consultation rate for non-psychiatric symptoms increased by nearly a half in both older and younger widows.