The relationship between income and mortality is explored by examining mortality in each income decile and income source (from earnings, government or capital). Swedish individual level income data was analysed for approximately 6.5 million adults. The quality of our data is unprecedented for this type of study, in terms of size and completeness of population coverage and death registration. The results suggest that inequalities in mortality are marked even in Sweden, one of the affluent countries where the effects of health inequalities are assumed to be lowest worldwide. The only income source that was associated with beneficial outcomes for all population groups was earnings. Welfare payments, often associated with illness, are associated with higher mortality, particularly for men. Capital income (our 'wealth' indicator) generally reduces the risk of mortality but increases the risk for some younger groups.