This paper reports a study of the complex associations between older people's participation in activities and their wellbeing in later life using data from a national sample of 1,334 Jewish-Israeli retirees. Confirmatory factor analysis substantiated a division of the activities into solitary, formal and informal categories, as postulated by activity theory. The outcome measure, the latent construct wellbeing, was compiled from scores on the 12-item General Health Questionnaire (GHQ-12), a global measure of life satisfaction, and a measure of satisfaction with the use of time. The analysis also examined the influence of socio-economic status, health status and the quality of inter-personal relationships. ‘Social relationship quality’, also a latent construct, was a composite of measures of satisfaction with children, friends and neighbours and a self-rated loneliness scale. Path analysis using structural equation modelling was employed. The results showed that when the quality of social relationships was taken into account, the amount of activity had no independent effect on the respondents' wellbeing. Moreover, it was social relationship quality, a facet of informal activity that has generally been neglected in activity research, that emerged as the most influential variable in the association between activity and wellbeing. Thus, the findings provide empirical backing for the assertion that the quality of social ties matters more than activity participation per se as predictors of a good old age.