This paper investigates impacts, mechanisms and selection effects of prenatal exposure to multiple shocks, by exploiting the unique natural experiment of the Dutch Hunger Winter. At the end of World War II, a famine occurred abruptly in the Western Netherlands (November 1944 - May 1945), pushing the previously and subsequently well-nourished Dutch population to the brink of starvation. We link high-quality military recruits data with objective health measurements for the cohorts born in the years surrounding WWII with newly digitised historical records on calories and nutrient composition of the war rations, daily temperature, and warfare deaths. Using difference-in-differences and triple differences research designs, we show that the cohorts exposed to the Dutch Hunger Winter since early gestation have a higher Body Mass Index and an increased probability of being overweight at age 18, and that this effect is partly accounted for by warfare exposure and a reduction in energy-adjusted protein intake. Moreover, we account for selective mortality using a copula-based approach and newly-digitised data on survival rates, and find evidence of both selection and scarring effects. These results emphasise the complexity of the mechanisms at play in studying the consequences of early conditions.