The value of the cava-pulmonary anastomosis (Glenn procedure) was assessed by reviewing the entire experience with this procedure in 83 children at The Hospital for Sick Children, Toronto, from 1961 to 1980. The over-all operative mortality rate was 9.6%, but only 2% in the second half of the series. The average follow-up is 9.4 years per patient. The late mortality rate has been 20%. By actuarial analysis, 84% of survivors were alive 9 years after operation; 54% had adequate palliation and 46% had required further operation. Long-term palliation was significantly better when the anastomosis was performed after 5 years of age. A previous arterial pulmonary artery shunt had no effect on either patient or shunt survival. Patients with tricuspid atresia had a better long-term survival rate than, but similar shunt survival to, those children with transposition, single ventricle, or miscellaneous cardiac anomalies. Angiography in 36 survivors showed patent shunts in 35 and no arteriovenous fistulas. Clinical status was not related to superior vena caval pressure or angiographic findings, indicating that Glenn shunt failure is a functional problem. The procedure is most useful for patients over 5 years of age who have complex heart disease not amenable to total repair. It is also on appropriate initial procedure in young patients over 18 months of age for whom direct atrium--pulmonary artery anastomosis is proposed as the future definitive treatment.