In this study meal sharing is used as a way of quantifying food transfers between households. Traditional food-sharing studies measure the flow of resources between households. Meal sharing, in contrast, measures food consumption acts according to whether one is a host or a guest in the household as well as the movement of people between households in the context of food consumption. Our goal is to test a number of evolutionary models of food transfers, but first we argue that before (...) one tests models of who should receive food one must understand the adaptiveness of food transfers. For the Ye’kwana, economies of scale in food processing and preparation appear to set the stage for the utility of meal sharing. Evolutionary models of meal sharing, such as kin selection and reciprocal altruism, are evaluated along with non-evolutionary models, such as egalitarian exchange and residential propinquity. In addition, a modified measure of exchange balance—proportional balance—is developed. Reciprocal altruism is shown to be the strongest predictor of exchange intensity and balance. (shrink)