1. Five continuous records of arterial plasma potassium were obtained from three normal subjects during brief periods (5–7 min) of exercise (100 W).
2. In two of these subjects hepatic venous blood samples were withdrawn at 0.5–1.0 min intervals and analysed in vitro for plasma potassium.
3. Arterial plasma potassium rose rapidly at the start of exercise from 3.8 ± 0.3 mmol/l (mean ± sd) to plateau levels of 5.4 ± 0.1 mmol/l.
4. One of the above subjects and a further subject were studied after β-blockade with propranolol. This resulted in an exaggerated rise in arterial plasma potassium during exercise.
5. Hepatic venous potassium measurements indicated that the liver probably had little effect on potassium changes during exercise.
6. The changes in arterial plasma potassium during exercise are rapid and substantial. If transmitted to the extracellular fluid these changes would alter cell transmembrane potential and might as a result alter receptor sensitivity.