In this paper we report a study of the uses of mobile video telephony based on the collection and analysis of naturally occurring mobile and Skype video exchanges, with a focus on camera motions. We provide evidence for a set of phenomena characteristic of the organization of ‘video-in-interaction’: a) Mobile and Skype video calls are patterned, often alternating between a ‘talking heads’ arrangement, in which both participants are on screen and facing the camera, and moments in which they are producing various shots of their environment in line with their current interactional purposes. b) Openings occur almost always in the ‘talking heads’ arrangement. c) The video images on either side are produced and expected to be scrutinized with respect to their relevance to the ongoing interaction. d) The ‘talking heads’ arrangement is oriented to as a default mode of interaction, with the implication that when there is nothing else relevant to show, the participants should show themselves on screen. e) In multi-party interactions, the party who is handling the video communication apparatus has an obligation to put other speakers on screen when they talk and the video conversationalists orient to their appearance on screen as making relevant a distinctive participation status. We show how these phenomena all derive from an orientation towards a single maxim, ‘put the face of the current speaker on screen’, which plays a foundational role in the organization of ‘video-in-interaction’ and its articulation with that of ‘talk-in-interaction’.