L atest estimates indicate that nutritional deficiencies account for 3 million child deaths each year in less-developed countries. Targeted nutritional interventions could therefore save millions of lives. However, such interventions require careful optimization to maximize benefit and avoid harm. Progress toward designing effective life-saving interventions is currently hampered by some serious gaps in our understanding of nutrient metabolism in humans. In this Personal Perspective, we highlight some of these gaps and make some proposals as to how improved research methods and technologies can be brought to bear on the problems of undernourished children in the developing world.