This study examined the effectiveness of bilingual subtitles relative to captions, subtitles, and no subtitles for incidental vocabulary learning. Learners' processing of novel words in the subtitles and its relationship to learning gains were also explored. While their eye movements were recorded, 112 intermediate to advanced Chinese learners of English watched a documentary in one of 4 conditions: bilingual subtitles, captions, L1 subtitles, and no subtitles. Vocabulary pretests and posttests assessed the participants' knowledge of the target vocabulary for form recognition, meaning recall, and meaning recognition. Results suggested an advantage for bilingual subtitles over captions for meaning recognition and over L1 subtitles for meaning recall. Bilingual subtitles were less effective than captions for form recognition. Participants in the bilingual subtitles group spent more time reading the Chinese translations of the target items than the English target words. The amount of attention to the English target words (but not to the translations) predicted learning gains.