The present report considers the conscious perception of passive horizontal rotations of the trunk, the head, or both, by human observers. It examines in particular how this perception depends on the interaction of canal and neck afferents. Three sets of sinusoidal stimulations (0.2 Hz) were applied to subjects (Ss): Rotations of (1) whole body (pure labyrinthine stimuli, lambda), of (2) only the trunk with the head stationary in space (pure neck stimuli, nu), and of (3) both head and trunk, each with an amplitude and a direction of its own, giving rise to various in-phase and counter phase combinations of lambda and nu.--The Ss were to estimate the magnitude of their turning sensations (psi). In doing so, they were to concentrate either on the rotation of their trunk in space (TS) or of their head in space (HS), or of the head relative to the trunk (HT). The TS and HS turning sensations induced by pure lambda-stimuli were essentially the same as to magnitude and direction. Pure nu-stimulation also led to TS and HS turning sensations. However, the former had the direction of the trunk-to-head (T delta S) deflection, the latter that of the head-to-trunk deflection. The nu-induced HS turning sensation represented an illusion, since the head remained stationary in space. When the lambda- and nu-stimuli were combined, the interaction could be described by a linear summation of their effects. The estimates of TS turning followed the equation psi HS approximately lambda-nu, thus well reflecting the actual TS rotation. The estimates of HS could be described by psi HS approximately lambda+k nu; the term k nu represents the "nu-illusion" contaminating the HS turning sensation. The estimates of HT turning were roughly proportional to nu alone and, therefore, close to the actual HT rotation. We conclude that humans may derive a rather faithful information about trunk rotation from the combined activation of canal and neck afferents, but that the sensation of passive head rotation is contaminated by an (illusionary) contribution from neck afferents. These additive and subtractive modes of interaction have parallels in postural reflexes as well as in neuronal responses that are known from cat.