This study aims to contribute to a clearer understanding of children’s developing gesture-speech system examining the interrelation between deictic and representational elements of gestural and vocal types. It analyzes the spontaneous productions of six children, observed longitudinally from 12 to 24 months during the transition from one- to two- and multielement vocal utterances. The focus lies on children’s gestural and vocal repertoires, and one- and two-element utterances encoding different information within and across modalities. Results indicate that deictic and representational elements are unevenly distributed in the gestural vs. the vocal modality, and in one vs. two-element utterances, with patterns that differ from those observed in the adult gesture-speech system. In these early stages speech and gesture are interrelated primarily through deictic gestures, and representational abilities appear to be markedly more constrained in the gestural as compared to the vocal modality.