To compare motor and sensory capabilities for fusion of vertical disparities at different angles of horizontal vergence in healthy humans. METHODS. Eye movements were recorded from both eyes of 12 healthy subjects using three-axis search coils. The stimulus was a cross (+) (3.4 X 3.2°, vertically and horizontally, respectively)presented to each eye with a stereoscopic display. Vertical disparities were introduced by adjusting the vertical position of the cross in front of one eye. The disparity was increased in small increments (0.08°) every 8 seconds. Viewing was denned as "near" if there was a horizontal disparity that elicited 6° to 15° convergence, depending on the subject's capability for horizontal fusion; viewing was denned as "far" at 1° convergence. Maximum motor (measured), sensory (stimulus minus motor), and total (motor plus sensory) vertical fusion were compared. RESULTS. In 9 (75%) of 12 subjects the maximum total vertical fusion was more in near than in far viewing. The three who did not show this effect had relatively weak horizontal fusion. For the entire group, the motor component differed significantly between far (mean, 1.42°) and near (mean, 2.13°). Total vertical fusion capability (motor plus sensory) also differed significantly between far (mean, 1.68°) and near (mean, 2.39°). For the sensory component there was no difference between between far (mean, 0.268°) and near (mean, 0.270°). As vertical disparity increased in a single trial, however, there was a small gradual increase of the contribution of the sensory component to vertical fusion. CONCLUSIONS. Vertical fusion capability usually increases with convergence. This increase is caused primarily by an increase in the motor component. There is a gradual but small increase in the sensory component as target disparity slowly increases.