Pitching the head while rotating (PWR) combines periodic activation of the semicircular canals and the otoliths to generate pitch and roll eye deviations and continuous horizontal nystagmus. Monkeys were tested after individual pairs of semicircular canals were plugged and single units were recorded in the vestibular nerve while the animals were sinusoidally pitched 20-40 deg about a spatial horizontal axis with 5- and 16-s periods and simultaneously rotated about a spatial vertical axis at 30-120 deg/s. As previously shown, the steady-state horizontal response disappeared after plugging the vertical semicircular canals, but was maintained when the lateral canals were plugged. When the left anterior and right posterior canal (LARP) pair was left intact, the steady-state response depended on the axis about which the pitching took place. When the axis was normal to the LARP plane, there was no steady-state response. When the pitching axis was perpendicular to the LARP normal, the response was maximal. Firing rates of otolith units were approximately in phase with pitch position, and the addition of rotation about a vertical axis did not change the response. Lateral canal units did not have a steady-state modulation during pitch or constant velocity rotation. During PWR, they oscillated at twice the pitch frequency. This corresponded to the frequency at which the canal was maximally activated as it aligned with the plane of rotation. The amplitude of modulation increased proportionally to rotational velocity, but the phase remained the same. These characteristics were unchanged during roll while rotating (RWR), which induces little continuous nystagmus. Anterior and posterior canal units were maximally excited near pitch-velocity maxima and minima, respectively, during pure pitching. During PWR, however, the phases of both components simultaneously shifted toward each other and toward being in phase with otolith units. The peak excitation tended toward a forward-pitch position when the rotation was to the ipsilateral side, and toward a backward pitch position when the rotation was to the contralateral side. With 120-deg/s rotation during a 16-s pitch period, the phase difference between anterior and posterior canal units was as small as 17 deg. These data support the postulate that the correlation between vertical canal and otolith units is the critical factor in generating continuous unidirectional horizontal nystagmus during PWR.