The study was based on an index group of 49 mothers who had had depressive disorders in the post-natal year, and 49 control mothers who had been free from any psychiatric disorder since delivery. Nineteen months after childbirth, the interaction between mother and child was assessed by blind assessors using defined observational methods. Compared with controls, index mother-child pairs showed a reduced quality of interaction (e.g. mothers showed less facilitation of their children, children showed less affective sharing and less initial sociability with a stranger). Similar but reduced effects were seen in a subgroup of index mothers and children where the mother had recovered from depression by 19 months. Social and marital difficulties were associated with reduced quality of mother-child interaction.