1. Activity was recorded from abdominal (expiratory) and phrenic (inspiratory) nerves during natural vestibular stimulation in multiple vertical planes and the horizontal plane in decerebrate cats. Vestibular stimulation was produced by rotating the head in animals whose upper cervical dorsal roots were transected to remove inputs from neck receptors; the upper airway and carotid sinus were denervated, and the vagus nerves were transected to assure that the head rotations did not elicit visceral or pulmonary inputs. 2. The plane of head rotation that produced maximal modulation of respiratory nerve activity (response vector orientation) was measured at one or more frequencies between 0.05 and 0.5 Hz. The dynamics of the response were then studied with sinusoidal (0.05-2 Hz) stimuli aligned with this orientation. In some animals, sinusoidal horizontal rotations of the head at 0.5 and 1 Hz or static head tilts in the pitch and roll planes were also delivered. 3. Typically, maximal modulation of abdominal nerve outflow was elicited by head rotations in a plane near pitch; nose-up rotations produced increased outflow, and nose-down rotations reduced nerve discharges. The gains of the responses (relative to stimulus position) remained relatively constant across stimulus frequencies, and the phases were consistently near stimulus position, like regularly firing otolith afferents. Static nose-up tilt produced elevated abdominal nerve activity throughout the stimulus period, providing further evidence that pitch-sensitive otolith receptors contribute to the response. Horizontal head rotations had little influence on abdominal nerve discharges. 4. The abdominal nerve responses to head rotation were abolished by chemical or aspiration lesions of the medial and inferior vestibular nuclei, which is concordant with the responses resulting from activation of vestibular receptors. Transections of axons arising from bulbospinal neurons in the ventral respiratory group, which are known to be the predominant source of expiratory signals to the spinal cord, reduced but did not abolish the vestibuloabdominal reflex. Thus it is likely that nonrespiratory neurons also participate in generating this response. 5. Nose-up pitch of the head; and in particular large (50 degrees) static tilts, produced small increases in phrenic nerve activity. Ear-down tilt and horizontal rotation of the head produced no responses in the phrenic nerve. 6. The existence of vestibular inputs to some respiratory motoneurons suggests that the vestibular system has influences on muscles in addition to those typically considered to have antigravity roles, and participates globally in adjusting muscle activity during movement and changes in posture.