The dye dilution method for measuring cardiac output is based on injecting rapidly a known quantity of a dye at one site into the circulatory system, and withdrawing blood at a distal site for determination of a concentration curve of the dye. Flow (Q) is calculated by the formula: (formula: see text) where m is the amount of dye injected, c mean concentration of dye and t the time of the concentration curve without recirculation. In recent years the only dye used has been indocyanine green (cardiogreen) which has its absorption maximum in the infrared part of the spectrum (at 805 microns) - where oxyhaemoglobin and reduced haemoglobin transmit light equally. Several densitometers for cardiogreen have been developed. The Christian Michelsen Institute densitometer used in our laboratory was found to give very accurate measurements (error less than +/- 2%) of blood flow in model experiments, for flows ranging from 2 to 12 l min-1. The more modern densitometers are usually equipped with computers. The cardiogreen method is probably one of the most accurate methods to study cardiac output during exercise. The error of a single determination of cardiac output values at rest and during exercise is less than +/- 5%. The method does not allow measurement of 'beat to beat' changes, and requires a cardiac output which is stable for approximately 10 s during exercise and 30 s at rest. It has been extensively used in our laboratory to study changes in central haemodynamics in essential hypertension at rest and during exercise, and also to study the haemodynamic alterations induced by anti-hypertensive agents.(ABSTRACT TRUNCATED AT 250 WORDS)