When interacting with young children, adults often self-repeat their own utterances that vary in sequences of adjacent utterances called variation sets (VS) (Küntay and Slobin 1996). These repetitions benefit children's linguistic development because they emphasize form and meaning. This paper analyzes the use of VS during group interaction and from a multimodal point of view. Sixteen teachers were video-recorded during interaction with two-year-old children in Spanish nursery schools. Results show that the use of VS is particularly frequent in these settings and that they are typically combined with gestures. Teachers directed their VS more often to a group of children than to a single child, and those VS directed to the group were more often accompanied by gestures than the VS directed to individuals. The results also show the influence of group size on children's responses. The study thus sheds new light on our understanding of child-directed speech (CDS), as well as the need to adapt the speech to single children during interaction in large groups.