Reading comes with a clear binocular advantage, expressed in shorter fixation times and fewer regressions in binocular relative to monocular visual presentations. Little is known, however, about whether the cost associated with monocular viewing derives primarily from the encoding of foveal information or in obtaining a preview benefit from upcoming parafoveal text. In the present sentence reading eye tracking experiment, the authors used a novel dichoptic binocular gaze-contingent moving window technique to selectively manipulate the amount of text made available to the reader both binocularly and monocularly in the fovea and parafovea on a fixation-by-fixation basis. This technique allowed the authors to quantify disruption to reading caused by prevention of binocular fusion during direct fixation of words and parafoveal preprocessing of upcoming text. Sentences were presented (a) binocularly; (b) monocularly; (c) with monocular text to the left of fixation; (d) with monocular text to the right of fixation; or (e) with all words other than the fixated word presented binocularly. A robust binocular advantage occurred for average fixation duration and regressions. Also, while there was a limited cost associated with monocular foveal processing, the restriction of parafoveal processing to monocular information was particularly disruptive. The findings demonstrate the critical importance of a unified binocular input for the efficient preprocessing text to the right of fixation.