Cardiac output was measured by the thermodilution method in a patient with a left-to-right shunt undergoing cardiac catheterization. It appeared that the thermodilution method measured systemic rather than pulmonary blood flow. This occurred because of the slow injection of injectate in the presence of a large left-to-right shunt. Theoretical thermodilution cardiac output curves are provided to illustrate the interaction of these two factors when three different durations of injection and four different shunt sizes are used.