substitute does not differ significantly between controls and inpatients. The sibship comparisons will be published in detail elsewhere (Birtchnell 1971). Sibship size distribution is similar in early bereaved patients and controls. There are significantly more early mother-bereaved women patients than early mother-bereaved men patients with younger siblings to be cared for, but there is no such difference among the controls. A similar state of affairs does not obtain for father-bereaved men patients. Proportionally more early bereaved patients than early bereaved controls have no older siblings, though the excess is no more marked when siblings of the same sex as the lost parent are considered. Comment Aberdeen and Dumfries are only 200 miles apart and yet show important differences both in the control populations and in the way the patients and controls differ. In both areas the mncidence of early bereavement is significantly greater in psychiatric patients and this excess is greatest for women. The Aberdeenstudy sugests that parent replacement and the composition of the sibship may modify the effects of bereavement. Acknowledgments: This study was carried out during the tenure of an MRC Clinical Research Fellowship. My thanks are due to Dr J A Baldwin