With increasing experience in the surgical management of children with congenital heart disease has come an increasing use of prosthetic valves. They have been employed in a wide variety of conditions ranging from simple congenital stenosis or incompetence to an essential part of repair of more complex disorders. As in adult cardiac surgery, there has been considerable debate over the exact prostheses that should be used in individual circumstances. As we point out in this article, experience has varied both geographically and with time. Our own experience at Great Ormond Street, for example, has been of initial satisfaction with results of mechanical valves followed by enthusiasm for, sequentially, the Hancock and the Ionescu prostheses. Now we have turned full circle and predominantly use mechanical valves, primarily owing to the extraordinary incidence of premature, degenerative calcification of bioprosthetic valves in children. The exact rates of degeneration of these valves are variable; thus it is difficult to relate prosthesis survival to the growth rate of the child. The experience with homograft valve replacements is encouraging, since they do not seem to degenerate so rapidly. Valve replacement in children carries a high risk. The younger the child, the higher the risk, with mitral valve replacement being the most dangerous procedure of all. We conclude that valve replacement in children is a procedure that should not be taken lightly and for which, at present, mechanical valves offer the best prognosis, despite the small risk of anticoagulation.