Heading estimation has not previously been reported in the vertical planes. This is a potentially interesting issue because although distribution of neuronal direction sensitivities is near uniform for vertical headings, there is an overrepresentation of otolith organs sensitive to motion in the horizontal relative to the vertical plane. Furthermore, thresholds of horizontal motion perception are considerably lower than those of vertical motion which has the potential to bias heading perception. The current data from 14 human subjects (age 19 to 67) measured heading estimation in response to vestibular motion of 14 cm (28 cm/s) over a 360 ° of headings at 5 ° intervals. An analogous visual motion was tested in separate trials. In this study, earth and head vertical/horizontal were always aligned. Results demonstrated that the horizontal component of heading was overestimated relative to the vertical component for vestibular heading stimuli in the coronal (skew) and sagittal (elevation) planes. For visual headings, the bias was much smaller and in the opposite direction such that the vertical component of heading was overestimated. Subjects older than 50 had significantly worse precision and larger biases relative to that of younger subjects for the vestibular conditions, although visual heading estimates were similar. A vector addition model was fit to the data which explains the observed heading biases by the known distribution of otolith organs in humans. The greatly decreased precision with age is explained by the model with decreases in end organ numbers, and relatively greater loss of otoliths that are sensitive to vertical motion.