Independent groups of up to 32 young men exposed in a standard seated posture to one of five conditions of vertical (Z axis) motion for up to 2 h. Exposure was less in the event of vomiting or a volunteer's voluntary withdrawal from the experiment. A control condition, sinusoidal motion of 0.17 Hz and 0.13 G R.M.S., provided the basis for comparison with the remaining four conditions, each produced by the sum of two sinusoids, the fundamental at 0.17 Hz plus the second or third harmonic. The conditions differed in the phase relationship of the fundamental and second harmonic, or in the relative acceleration levels of the two harmonic components. The predicted motion sickness incidence (MSI) for each sinusoid alone was calculated from a previously derived mathematical model and compared with the obtained MSIs. Certain motion conditions provoked unexpectedly high MSIs compared with the control condition. It was found that R.M.S. acceleration is not reliable as the sole predictor of MSI in complex motion. Further data must be obtained before accurate prediction of MSI in broadband motion will be possible.