Social class differences in mortality are larger in middle adulthood than at any other time of life. Circumstances over the lifecourse may contribute to these adult social class differences. However, it is only rarely that the lifecourse approach has been applied to mortality studies among persons in their middle adulthood. The aim of this thesis is to disentangle the effects of the living conditions in the parental home and major transition in youth on social class differences in mortality from various causes of death among women and men aged 31-42 at death, and to evaluate whether the effect of the past circumstances on mortality is through latency, accumulation or pathway mechanisms. This thesis (papers II-V) is based on the 1990 census data for all Finnish persons born in 1956-60 linked with death records (4369 deaths) for 1991-98 and with information on lifecourse circumstances from the 1970, 1975, 1980 and 1985 censuses. These aggregated cross-tables are analysed by means of Poisson regression. Parental home had an association with disease mortality from age 20 onwards, indicating a latency effect. However, the direct effect of the parental home on mortality was minor, and therefore the contribution of latency model to differential mortality remained small. An indication for the accumulative effect of disadvantageous social class was found for cardiovascular diseases and alcohol-related causes. The living conditions in the parental home, i.e. the manual class and one-parent family, had an effect on the transitions a person experienced in youth, and thus contributed to the effect youth paths exerted on adult social class differences in mortality from various causes of death. Youth paths had a substantial effect (about 60-90%) over and above the preceding effect of living conditions in the parental home on mortality. The higher mortality in the lower social classes was mainly attributable to disadvantageous educational path. Moreover, both family formation, particularly early marriage in women and staying single in men, and experience of unemployment in youth, had independent effects on class differences in mortality. These results strongly suggest that youth is a ‘sensitive period’ affecting social class differences in mortality in middle adulthood.