By the end of 1991, less than half the amount of relief food requested for North Sudan at the beginning of the year had been delivered. Despite ample evidence of social and economic stress and high rates of child malnutrition, many donors felt that relief needs had been exaggerated, and were unwilling to accept that relief assistance was urgently needed. The feeble response of the main food aid donors is explained initially by the politics of relief in 1990/91, which seriously delayed the launch of the relief operation. These problems were compounded by an oversimplified understanding of famine among some sections of the relief community, and by the orientation of the international relief system to crisis indicators. Toward the end of 1991, donors argued that despite the shortfall in relief assistance there had been no deaths from starvation, and therefore local people had ‘coped’ better than expected. This paper challenges that view by arguing that excess deaths did occur, but went unnoticed and unremarked. Local people's ‘coping strategies’, which supposedly ‘saved the day’, actually had very negative and sometimes fatal c