This paper reviews 387 patients who had perfusion scintigraphy over a period of two and a half years. Data are presented concerning patients' symptoms and the use of perfusion imaging. Scintigraphy is shown to be subjective and rather unreliable, as practised in this series. Post mortem correlations suggest that used in isolation, perfusion scintigraphy cannot reliably distinguish emboli from other lung diseases. Use of ancillary tests and ventilation scintigraphy is discussed.