Available oxygen and oxygen unloading capacity were determined at the time of cardiac catheterisation, from blood gas analysis, p 50 in vivo, and haemoglobin, in 79 children with cyanotic congenital heart disease, their age ranging from 2 weeks to 14 years. The patients were followed up for six months to find out how many of those without surgical intervention survived. Actuarial analysis showed that the six month survival rate among the patients with low available oxygen or oxygen unloading capacity was significantly lower (24% and 49%, respectively) compared with those with normal indices (90% and 96%, respectively). Both available oxygen and oxygen unloading capacity were superior to arterial oxygen saturation and arterial oxygen content as prognostic criteria. Haemoglobin and haematocrit were of little value in prognosis. Available oxygen appeared to be the best prognostic criterion for survival at six months. It is concluded that available oxygen and oxygen unloading capacity are useful prognostic criteria in children with cyanotic congenital heart disease. These new criteria, when used in conjunction with the conventional methods of assessing the cardiac lesion, may be of considerable help to the physician in evaluating the prognosis of the patient.