The thumb motion of 10 normal subjects during cellular phone use was measured using a reflective marker detection system to compare the maximum, minimum and range of flexion angles of the interphalangeal (IP), metacarpophalangeal (MP) and carpometacarpal (CM) joints. Two micro-reflective markers 3 mm in diameter were each placed on the dorsal surface of the distal phalanx, basal phalanx and metacarpal bone of the thumb. Three markers were placed on the dorsal hand in order to define the dorsal hand plane. Each subject pushed the 12 keys of a folding cellular phone with an 85-mm-long and 40-mm-wide keypad, sequentially from ‘1’ to ‘#’, and the pushing motion was recorded by six infrared video cameras for 12 seconds, using the VICON 612 system. The mean maximum flexion angle of the MP joint was significantly (p< .05) larger than the CM joint, and the mean minimum flexion angle of the CM joint was significantly (p< .01) smaller than the IP and MP joints. The mean range of motion of the IP joint was significantly (p< .05) larger than the MP and the CM joints. In a comparison of different key-pushing motions, only the CM joint was significantly (p< .05) larger in its range of motion. In conclusion, thumb motion on pushing the keys of the cellular phone was produced mainly by the MP and the CM joints. In addition, the ability to reach keys in different areas of the cellular phone keypad is regulated by changing the flexion angle of the CM joint.