Visual directions of the two stimuli in Panum's limiting case with different interstimulus and convergence distances confirmed the predictions from the reformulated Wells—Hering's laws of visual direction. In experiment 1, six observers each converged on the midpoint of the interstimulus axis at 30, 60, and 90 cm from the eyes and adjusted a probe on the fixation plane to be in the same visual direction as that of each stimulus. Visual direction of the far stimulus was always nonveridical whereas that of the near stimulus was veridical only when its retinal disparity was small. In experiment 2, three observers each converged on the intersection of mid-sagittal plane and (a) the frontoparallel plane of the near stimulus, (b) that of the midpoint between the two stimuli, or (c) that of the far stimulus. The midpoint of the interstimulus axis was 60 cm from the eyes. Visual direction of the far stimulus was veridical only with convergence at the far plane. Visual direction of the near stimulus was veridical with convergence at the near plane, and also, only when its retinal disparity was small, with convergence at the two other planes.