To track a slowly moving object during whole body rotation, smooth-pursuit and vestibularly induced eye movements must interact to maintain the accuracy of eye movements in space (i.e., gaze), and gaze movement signals must eventually be converted into eye movement signals in the orbit. To understand the role played by the cerebellar vermis in pursuit-vestibular interactions, in particular whether the output of the vermis codes gaze-velocity or eye-velocity, we examined simple-spike activity of 58 Purkinje (P-) cells in lobules VI-VII of head-stabilized Japanese monkeys that were trained to elicit smooth-pursuit eye movements and cancel their vestibuloocular reflex (VOR) during passive whole body rotation around horizontal, vertical, or oblique axes. All pursuit-sensitive vermal P-cells also responded during VOR cancellation, and the majority of them had peak modulation near peak stimulus velocity. The directions of maximum modulation during these two tasks were distributed in all directions with a downward preponderance. Using standard criteria, 40% of pursuit-sensitive vermal P-cells were classified as gaze-velocity. Other P-cells were classified either as eye/head-velocity group I (36%) that had similar preferred directions during pursuit and VOR cancellation but that had larger responses during VOR x1 when gaze remained stationary, or as eye/head-velocity group II (24%) that had oppositely directed or orthogonal eye and head movement sensitivity during pursuit and VOR cancellation. Eye/head-velocity group I P-cells contained cells whose activity was correlated with eye velocity. Modulation of many P-cells of the three groups during VOR x1 could be accounted for by the linear addition of their modulations during pursuit and VOR cancellation. When monkeys fixated a stationary target, over half of the P-cells tested, including gaze-velocity P-cells, discharged in proportion to the velocity of retinal motion of a second spot. These observations are in a striking contrast to our previous results for floccular vertical P-cells. Because we used identical tasks, these differences suggest that the two cerebellar regions are involved in very different kinds of processing of pursuit-vestibular interactions.