Purpose: This study evaluated the social facilitation of elderly patients' food intake beyond the presence of mealtime companions by assessing various relationships. The study examined the relationships between patients' intake and (a) the number of interpersonal exchanges with mealtime fellows, (b) the nature of behaviors expressed by the patients themselves and their fellows, and (c) the degree of complementarity between these. Design and Methods: Interpersonal exchanges and intake were observed on repeated mealtime occasions (n = 1,477) nested within 32 geriatric patients (21 women, 11 men; age, M = 78.8 years). Participants' intake was estimated from plate leftovers. Interpersonal behaviors were examined for both participants and patients with whom they interacted in terms of agency and communion dimensions, following the interpersonal circumplex model of human interaction. With the use of multilevel regression analyses, the number, nature, and complementarity of behaviors that participants engaged in and were exposed to on a given meal were computed to test their impact on intake. Results: The total amount of interaction between patients was positively related to intake. The effect was significant for both participants' own behaviors and those to which they were exposed, and it varied with the nature of the interaction; effects were significant in terms of frequency and complementarity for communal behaviors, and complementarity only for agentic behaviors. Effects could only partly be explained by meal duration effects. Implications: The results provide support for the effect of the number, nature, and complementarity of mealtime interpersonal behaviors on the food intake of elderly patients, and they may inspire new approaches to ensure adequate intake in this malnutrition-prone population.