"Normal" human subjects were placed in a series of 5 static orientations with respect to gravity and were asked to view an optokinetic display moving at a constant angular velocity. The axis of rotation coincided with the subject's rostro-caudal axis and produced horizontal optokinetic nystagmus and afternystagmus. Wall (1) previously reported that these optokinetic afternystagmus responses were not well characterized by parametric fits to slow component velocity. The response for nose-up, however, was larger than for nose-down. This suggested that the horizontal eye movements measured during optokinetic stimulation might include an induced linear VOR component as presented in the body of this paper. To investigate this hypothesis, another analysis of these data has been made using cumulative slow component eye position. Some subjects' responses had reversals in afternystagmus direction. These reversals were "filled in" by a zero slow component velocity. This method of analysis gives a much more consistent result across subjects and shows that, on average, responses from the nose-down horizontal (prone) orientation are greatly reduced (p < 0.05) compared to other horizontal and vertical orientations. Average responses are compared to responses predicted by a model previously used to predict successfully the responses to post-rotatory nystagmus after earth horizontal axis rotation. Ten of 11 subjects had larger responses in their supine than their prone orientation. Application of horizontal axis optokinetic afternystagmus for clinical otolith function testing, and implications for altered gravity experiments are discussed.