In 100 children with suspected left to right shunts the ratio of pulmonary to systemic flow was measured both by oximetry and first pass radionuclide angiography. The pulmonary time activity curve from the radionuclide study was analysed by the method of gamma variate fits. There was strong correlation between the two techniques; weaker correlation was found when the shunt was at atrial rather than ventricular level. This difference can be explained only by problems with the oximetric rather than the radionuclide technique. Although there are important limitations to the radionuclide method, it is the more precise and less invasive of the two and is to be preferred when the accurate measurement of left to right shunts is required.