Socio-economic conditions in childhood and their effect on morbidity and mortality in adult life have been discussed frequently during the last 10 years or so. At least two hypotheses have been discussed. The hypothesis that perinatal and childhood conditions are directly and causally related to morbidity and mortality has been met by a critique, the essence of which is that the relationship could be an indirect one. Those experiencing poor conditions early in life are more likely to experience poor conditions later in life and the association between perinatal and childhood factors and adult health is mediated by adult social factors. Here, this alternative hypothesis is tested on a sample drawn from the Swedish population. The sample was interviewed in 1980-81 and the register of those who consented to the interview included 6298 men and 6397 women aged 16-74 years. This register was linked to the Cause of Death Register for the period 1980-88. 461 men and 289 women died during the follow up.

Firstly it is concluded that childhood socio-economic group is significantly linked to male mortality, when age is taken into account. This could not be demonstrated for women. The variation in general health between childhood socio-economic groups, on the other hand, is significant for women but not for men.

When a number of mediating variables are added, the effect of childhood socio-economic group remains significant for men. It is thus not possible to exclude the possibility that childhood socio-economic conditions have a direct effect on adult health. At the same time, it is suggested here that at least a part of that effect is mediated by social factors in adulthood. That childhood socio-economic conditions may be independently related to adult health does not support the hypothesis of 'biological programming'.