Of all the aspects of vibration, acceleration is recognized as of fundamental importance in the riding-qualities question, and the full significance of the problem of measuring acceleration is now also recognized. The author deals with one method, the use of contact-accelerometers, and describes the characteristics of this type of instrument. He then points out that as contact accelerometers were originally designed for vertical motion it was intended to use the upper contact, making full use of the sensitiveness of this contact. However, this very excellence proves to be a detriment and a source of error in the measurement of comparatively large movements. Among other schemes, the use of the lower contact has been suggested as a means for overcoming interference from vibrations of high frequency and small amplitude. While it has been recognized that this use of the lower contact introduces a certain error, it has been generally assumed that this effect on the accuracy of the results is negligible. Actual experience with such instruments, however, indicates that the error may be of some consequence. The author analyzes the effect of gap-clearance in contact-accelerometers.