We tested the hypothesis that first-trimester exposure to acute food deprivation is a risk factor for schizophrenia. A sharp and time-limited decline in the food intake of the Dutch population following a Nazi blockade in 1944 to 1945 created a unique if tragic natural experiment to test this hypothesis in three regions of Holland (west, north, and south). In the west, or famine region, birth cohorts exposed to severe food deprivation (an average daily ration under 4200 kJ) during the first trimester showed a substantial increase in hospitalized schizophrenia for women but not for men. Relative risks for women were 2.17 for "broad" and 2.56 for "restricted" schizophrenia. Moderate food deprivation during the first trimester (average daily ration under 6300 kJ) was not associated with increased risk of schizophrenia in the famine region. In the north and south regions, numbers were smaller and there was no exposure to severe famine. Birth cohorts exposed to moderate food deprivation during the first trimester showed a trend toward increased risk of schizophrenia for women. These findings give plausibility to the proposition that early prenatal nutrition can have a gender-specific effect on the risk of schizophrenia.