Shortly before and during the harvest of 1990, a series of warnings were issued by concerned international and UN agencies that Sudan would experience a very poor harvest followed by an acute food shortage over the period 1990-91. The 1990 harvest was estimated to be similar to that obtained in 1984. After the very poor harvest in 1984, Sudan experienced a major famine during which deaths may have numbered in the hundreds of thousands. There were fears that this experience might be repeated in 1990-91. By the time of the subsequent 1991 harvest, it was clear to all that a severe food crisis had been experienced. There were severe shortages of water and food and very high malnutrition rates of children were noted by UNICEF across a wide range of areas. Despite these adverse indications, starvation deaths were probably numbered in thousands, rather than hundreds of thousands. Famine mortality, which may include deaths from famine associated disease, was similarly low. The initial predictions, it now seems, may have over-estimated famine mortality almost one hundred times. Several potential explanations of the over-estimate are examined. These include prediction errors, government and donor responses to the drought such as food aid and immunization, and traditional community and household level coping strategies in times of food shortage.