Multimodal texts have a strong presence in contemporary society, in the mass media and on Internet, for instance. The use of multiple modalities in the language classroom is common, as can be seen in films, songs, images, and many other resources. The purpose of this study is to observe teaching practices with multimodal texts in English language classrooms in order to gain a view of the role of verbal and visual texts in this kind of setting. The focus is on (1) the perception that the teachers have of their choices about what material to use, and (2) their vision of its use multimodal text in relation to teaching resources in general. The theoretical perspective of the study is based on the theory of the grammar of visual design as modeled by Kress and van Leeuwen (1996, 2001). The pedagogical perspective of the study is guided by the work of Unsworth (2001) and Royce (2002), authors that argue for teaching based on a multimodal approach. The study discusses: (1) the use of images in pedadogical material and teacher’s visions of its use, and (2) the notion of making learners and teachers to be able to read multimodal texts, so as to gain multimodal communicative competence.