A new model of otolithic stimulation by linear acceleration is presented and compared to previous models, based upon anatomical evidence and on the ability of normal subjects to sense the direction of a linear acceleration vector acting in the coronal plane (roll-tilt perception). There are two basic methods of generating roll-tilt stimuli: 1) tilt-chairs either inside or outside a centrifuge and 2) fixed-chair centrifuges. The present model is based on consideration of the probable otoconial displacement produced by these two different methods of stimulation and the model incorporates a major role for the elastic restoring force of the otolith membrane. When this force is taken into account, and most previous models have ignored it, the model predicts that different patterns of otoconial displacement will be produced in tilt-chair and in fixed-chair centrifuge experiments. The different roll-tilt perception produced by these two methods may be caused by the different otoconial displacement patterns. It is suggested that the elastic restoring force of the otoconial membrane may contribute to space motion sickness.