The Starr-Edwards ball valve prosthesis is generally the standard by which other cardiac valve substitutes are compared. This report reviews information pertaining to several prostheses--the Beall mitral valve and the Bjork-Shiley, Braunwald-Cutter, Lillehei-Kaster and Smeloff-Cutter aortic and mitral valves--considered by some to have specific advantages over the Starr-Edwards valves. Hospital and late mortality rates after valve replacement are comparable for the four aortic valve prostheses reviewed and depend more on patient selection than on the specific prosthesis utilized. Extensive clinical experience with the Bjork-Shiley aortic valve indicates that this prosthesis offers hemodynamic advantages over ball valve prostheses, especially in patients with a small aortic root. Clinical experience with the Lillehei-Kaster pivoting disc prosthesis has been less extensive, but this model provides theoretical hemodynamic advantages similar to those of the Bjork-Shiley aortic valve prosthesis. Problems associated with cloth wear and the unexpectedly slow rate, in man, of tissue ingrowth into the fabric of the Braunwald-Cutter aortic valve prosthesis have been discouraging, although this prosthesis has been associated with a very low thromboembolic rate in patients receiving anticoagulant therapy. The Smeloff-Cutter aortic prosthesis is hemodynamically similar to the Starr-Edwards prosthesis and has been proved to be a reliable and durable aortic valve substitute over the past several years. Mortality after mitral valve replacement is also largely influenced by factors other than prosthetic valve design. On the basis of postoperative data, the five mitral valve prostheses reviewed do not appear to have substantial hemodynamic differences. For patients with a small left ventricular cavity the low profile prostheses, such as the Beall, Bjork-Shiley and Lillehei-Kaster, may be advantageous. Most available evidence indicates that patients receiving aortic or mitral valve prostheses should be given anticoagulant therapy postoperatively.