This paper explores how language teachers learn to teach with a synchronous multimodal setup (Skype), and it focuses on their use of the webcam during the pedagogical interaction. First, we analyze the ways that French graduate students learning to teach online use the multimodal resources available in a desktop videoconferencing (DVC) environment to monitor pedagogical interactions with intermediate level learners of French in a North-American university. Then, we examine communicational and pedagogical aspects of this process which involves orchestrating different modalities and deploying various regulations for "semio-pedagogical" purposes. We define semio-pedagogical skills as the capacity to mediate a pedagogical interaction by combining or dissociating modalities (written, oral, and/or video) that are adapted to objectives and to the cognitive requisites of the task. We posit that these skills have to become part of the professional repertoire of future teachers, as they will increasingly be required to exploit the multimodal potentialities of online communication in their teaching. The study draws on screen capture recordings of teacher trainee-student interactions and is completed by semi-directive interviews with teacher trainees (n = 5). It aims (1) to identify the importance of webcamming in the share of the pedagogical range available to teachers and (2) analyze the non verbal dimension of pedagogical communication via DVC. The outcome of this study is the identification of five degrees of utilization of the webcam medium: there a certain gradation in the way webcamming is used (with a more or less significant use of image) when compared to other modalities. The different uses that are identified vary according to the perceived usefulness of webcamming to monitor teaching and to the teacher trainees' capacity to manage different workspaces.