This doctoral dissertation explores the benefits of an extensive exposure to L2 television for content comprehension and vocabulary learning with beginner, adolescent EFL learners, through a longitudinal classroom-based intervention. The study also aims at investigating the role of frequency and imagery in word learning, and evaluating the long-term effects this type of intervention may have on learners’ attitudes towards the use of L2 television for language learning purposes. In order to do so, several aspects were examined, including factors related to the intervention, the learner, and the input itself. Four classes of Grade 8 students viewed 24 consecutive episodes of a TV series – spaced over an academic year – under four experimental conditions, with each class being assigned to a different treatment. Two classes viewed the episodes with English [L2] captions, and two with Spanish [L1] subtitles. One class in each language condition received, additionally, explicit instruction on target lexical items. The first study in this thesis looked at the differential effect of captions and subtitles on viewing comprehension, alongside several mediating factors. Results confirmed the higher efficiency of subtitles over captions for content comprehension at this level of proficiency, and the importance of prior vocabulary knowledge when viewing with captions. It was also found that the episodes’ lexical coverage was a strong predictor of comprehension, although no pattern of improvement could be observed over time – even though learners’ perceived comprehension increased by the end of the intervention. In the second study, word-form and word-meaning gains were examined, following a pre- / post-test design. Results revealed that having explicit instruction on vocabulary (i.e. being pre-taught the words through short pre-viewing activities) yielded significantly higher vocabulary gains, and that proficiency played a key role in how learners made use of this type of input. Language of the on-screen text, however, did not emerge as a predictor of gains, although when captions were displayed this tended to lead to higher learning, especially when combined with instruction. Results also indicated that there was a positive correlation between vocabulary gains and comprehension, and that a high percentage of the vocabulary learnt was retained in the long term. The third study focused on the effects of word repetition, spacing, and imagery support. Analysis showed that words with a higher number of encounters or that appeared in a massed condition (i.e. in the same episode) were better learnt. It was also found that words that were image-supported had also higher learning gains. The last study looked into learners’ perceptions and feeling of learning from viewing audio-visual input in English through questionnaires and interviews. Students reported being highly motivated to learn through L2 videos, finding them useful for a number of language aspects, including vocabulary learning, listening comprehension, and matching aural and written forms of the words. Data also showed a shift in viewing habits in the long term, indicating the appropriateness of this type of classroom intervention to foster autonomous viewing at home. Taken as a whole, the results from this dissertation provide evidence that extensive viewing of captioned and subtitled TV series supports comprehension and L2 vocabulary learning. The characteristics of this type of input (i.e. repeated encounters with words, imagery) have been shown to contribute to facilitate language learning. Additionally, EFL learners, at this age and proficiency level, are motivated to learn through this media – in and outside the formal setting.