Six healthy female volunteers twice undertook an experiment with different types of clothing leaving the arms and legs covered or uncovered at $24{pm}0.5^{circ}C$ and $50{pm}5%$ RH to study how different thermal stimulation to the distal extremities during the afternoon could modulate circadian parameters of body temperature rhythm. One type of clothing consisted of long-sleeved shirts and full-length trousers (Type I, 989 g, 0.991 clo); the other type consisted of half-sleeved shirts and knee-length trousers (Type II, 750 g, 0.747 clo). Subjects wore Type I or Type II clothing during the afternoon (14:00 h - 19:00 h), and Type I clothing during the evening (19:00 h - 22:30 h) and the night sleep (22:30 h - 06:00 h). Rectal temperature and skin temperatures at the arm and leg were measured continuously. Results were as follows: 1) The circadian amplitude of rectal temperature tended to be greater, and the acrophase was significantly earlier when wearing Type II rather than Type I clothing. 2) The circadian nadirs of skin temperatures of the arm and leg were significantly lower and the amplitudes were significantly greater with Type II clothing. In addition, the acrophase and bathyphase of the circadian rhythm of arm skin temperature were significantly earlier with Type II than Type I clothing. 3) The amplitude of rectal temperature was related closely with that of arm and leg skin temperature. These results suggest that a slightly cool thermal stress during the afternoon to the arms and legs exerted by wearing half-sleeved shirts and knee-length trousers induces a greater amplitude and a phase advance of the overt circadian rhythm of body temperature.