Through use of a compartmental model, we simulated the measurement of cardiac output and distribution by means of radioactively labeled microspheres in fetal lambs with weights between 0.5 and 3 kg. A systematic error in measured cardiac output caused by artifactual changes in blood volume caused by the injection and withdrawal of fluids during the procedure was less than 5% for fetal weights greater than or equal to 1 kg but increased for fetal weights less than 1 kg and when hypovolemia was simulated at all fetal weights. Sensitivity analysis disclosed no significant effect of changes in vascular resistance. We examined the effects of recirculation of microspheres and found no significant increase in error in the measured value of cardiac output due to 20% recirculation of all spheres entering a single isolated organ system; however, errors between 7 and 14% were observed with simulations of 20% recirculation in more than one compartment simultaneously. Recirculation also introduced significant errors in the measured distribution of cardiac output in certain cases. The effect on the measured cardiac output of a temporary change in the true cardiac output was dampened by the artifactual blood volume changes mentioned above and the fact that the measurement is a time-weighted average. We also evaluated four different experimental designs. We conclude that the microsphere technique provides a remarkably reliable means of quantifying cardiac output and individual organ flow in the fetus. The nonrandom errors inherent in the procedure examined in this study are of the order of 10%, which is likely to be less than the moment-to-moment variation in the true cardiac output.