A new apparatus which measures the blood pressure in the finger continuously and yet not invasively was tested for its usefulness during exercise. It was compared with upper arm measurements in 23 volunteers during prolonged bicycle ergometry. Simultaneously, a pulse plethysmogram was recorded from another finger of the same arm, whereas in six additional volunteers Doppler measurements were carried out on the radial artery. The results show that finger systolic pressure ceased to rise at about 40% of maximal exercise; the difference with the continuously rising systolic pressure in the upper arm becoming significant at 140 W. At the same time the amplitude of the finger plethysmogram became significantly higher than its initial value, indicating distinct cutaneous vasodilation, whereas the volunteers also became hot and started to perspire. However, the radial artery 'flow', deduced from the Doppler measurements, did not change significantly during exercise. It increased sharply and markedly in the cooling down period. Simultaneously with this increase in flow, HR and both systolic blood pressures fell drastically whereas the plethysmographic amplitude remained about stable at its raised level. The results fit in with the idea that a compromise is achieved between the need for muscle activity and the need for temperature regulation. It is concluded that the Finapres functions well during exercise, but that the systolic pressure in the finger is not representative for its more central counterpart during cutaneous vasodilation. It is argued that opening up of AVAs may contribute to this pressure effect.