Frames are cognitive structures of knowledge grounded in people's social interaction in recurrent sociocultural activities or individual incidences. Discussion concerning the knowledge in frames largely focuses on its linguistic manifestation with regard to roles and role relations in scenes. Little attention has been paid to the scriptal knowledge in conceptual frames. The present study provides empirical evidence that frame knowledge is also readily manifested in gesture in daily conversation. Gesture not only reveals roles and role relations in a scene as distinct from those in speech, but it also collaborates with speech to jointly manifest part of a script of the frame. In addition, in line with the view of prominence, the part of frame knowledge being selected for manifestation in gesture is the speaker's focus of attention on a certain aspect of the scene during speaking. Finally, the temporal patterning of gesture and speech further bears out growth point theory, in that the planning of an utterance involves the interplay of imagery thinking and linguistic thinking.