The main aim of this review was to collate data on food transfer in primates and to compare these data with predictions that stem from various functional explanations of food transfer. First, the data on adult-adult food transfer were summarized. In adults, although patterns of food transfer can often be predicted by dominance relationships, with dominant individuals obtaining food from subordinates, the data indicate that not all food transfer follows this pattern. Researchers have suggested that food transfer may therefore involve trade for other commodities, including grooming, social support, or sexual access, although these hypotheses are not strongly supported by the available data. Also, while studies of chimpanzees, capuchins, and tamarins have indicated that frequencies of food transfer between individuals are often correlated within dyads, such patterns do not necessarily indicate that reciprocity had occurred. Analyzing the costs and benefits of defending a food item may reveal that the costs of harassment provides sufficient explanation for a notable proportion of food transfer events. A number of theoretical perspectives may prove useful in future investigations of food transfer among adults. Patterns of food transfer from older to younger individuals were then reviewed. Infants obtain food from older individuals in a number of primate species, in particular the callitrichid primates. Quantifying the relative contributions of individuals of different age and sex categories to offspring provisioning will potentially increase our understanding of parental and alloparental care in cooperatively breeding species. The current data on callitrichids and chimpanzees highlight the possibility that both nutritional and informational benefits may be gained by infants. In callitrichids, food transfer peaks around the age of weaning and is most likely to involve novel or difficult-to-process items, with evidence suggesting that infants learn dietary preferences via food transfer. In chimpanzees, most food transfer occurs prior to weaning and involve items that are difficult for infants to process by themselves. Studies suggest that interactions between mothers and infants may result in the transmission of tool-using skills. Obtaining a food reward during such interactions may enhance any social learning processes that are taking place. Whether adults actively direct the dietary choices or food-processing techniques of youngsters remains controversial. We hope to have shown that this area of research has undergone key developments and presents considerable scope for further empirical investigation both in primates and in other animal species. © 2004 Elsevier Inc. All rights reserved.