Electromyographic activity of dorsal neck muscles elicited by sinusoidal rotations of the body and head was studied in decerebrate cats over a wide range of rotational frequencies and amplitudes. Rotation of the body with the head held fixed in space elicited a cervicocollic reflex (CCR) in the biventer cervicis, complexus, obliquus capitis inferior, rectus capitis major, and splenius muscles. As stimulus amplitude increased, CCR amplitude increased first rapidly and then more slowly, displaying two linear incremental sensitivity ranges. In contrast, the vestibulocollic reflex (VCR) elicited by whole body rotation had a minimum stimulus threshold below which no response was observed, whereas the vestibuloocular reflex (VOR) saturated at intermediate stimulus intensities. When stimulus frequency was varied, the CCR exhibited second-order dynamic behavior. At frequencies below 0.5 Hz, muscle EMG activation was in phase with peak platform angular deviation in the direction that stretched the muscle, and the gain measured as the percent modulation of EMG activity per degree of rotation remained constant. As frequency increased to 3-4 Hz, response phase advanced by 120 deg or more and gain increased with a slope approaching 40 dB/decade. The data were well-fitted by second-order transfer functions containing two zeros. Both the dynamic behavior of the CCR and its high sensitivity to small stimuli resemble the properties of muscle spindle primary afferents, suggesting that the latter may provide the major input responsible for the CCR. Dynamic properties and gains of the CCR and VCR were quite similar at frequencies between 0.2 and 3-4 Hz. Transfer functions of both reflexes contained two zeros whose time constants were correlated in a population of 11 cats, suggesting that reflex dynamics may be matched to the mechanical properties of each animal's head-neck system. Interaction of the CCR and VCR was studied under two conditions. When the head was driven by a servomotor while the body remained stationary, EMG activation by the two reflexes added linearly to produce a large response. When the body was rotated with the head allowed to counterrotate about the C1-C2 joint, the two reflexes combined linearly in an antagonistic fashion: the CCR acted to oppose head rotations produced by the VCR, thus preventing the ratio of head counterrotation to body rotation from exceeding 0.5. The data indicate that the CCR and VCR behave approximately linearly, both individually and in combination. Acting together, the two reflexes assist each other in preventing oscillation of the head on a stationary body.(ABSTRACT TRUNCATED AT 400 WORDS)