This paper investigates whether exposure to nutritional shocks in early life negatively affects mortality at older ages, using unique individual data on siblings born during and around the Dutch potato famine of 1846/47. The famine provides exogenous variation on the nutritional status of those exposed to the famine. Moreover, the sibling information allows us to control for other usually unobserved environmental factors at the family level that may affect later life health. For instance, family specific biological or socio-economic factors may mitigate or enforce the impact of nutritional shocks early in life and failing to control for this may bias estimates of the effect of the nutritional shock. We use a unique historical dataset that follows siblings in a family from the time of birth until death. This dataset is merged with data on regional food prices and calories available per capita in the nineteenth century in the Netherlands. Moreover, the calories data allows us to assess effects of the diet composition (proteins and proteins from animals intake) in utero and early life on longevity. We show that conditions during both utero and first five years of life are important for mortality patterns, with the latter having a higher impact, and they remain significant even if we correct for the family component. Next, we find evidence for long term effects of nutritional conditions during first years of life on mortality at age 40-60, especially of males, whereas negative nutrition conditions in utero seem to result in very high selection of survivors, which masks the effects on long term mortality. Finally, we find that children born in families of farmers and in families with low social status were particularly strongly affected by exposure to the Dutch potato famine during utero or early life.