A large body of educational research has focused on play as one of children’s own activities, however, considerably less attention has been paid to structures and practices associated with joint play between adults and children. This article contributes to this line of research by analyzing adults’ participation in joint play with very young children. The data consist of 10 rich make-believe play cases taken from 150 h of videotaped, naturally occurring interactions in a group care setting. The results show that the ability of adults to build sustained co-participation in play with very young children demands delicately timed observations, initiatives and responses with attuned and coordinated use of gesture, gaze and talk. In all, this study provides one way to study and understand better what adults are doing in practice while they are actively co-participating in play. Pedagogical implications for early childhood education are discussed. © 2017 Elsevier Ltd