Recent highly reproducible results of tritium production by deuterium-deuterium (D-D) fusion from gas/solid-state fusion experiments are discussed in terms of a surface fusion mechanism. Theoretical criteria and experimental conditions for improving and optimizing D-D fusion rates in a gas/solid-state fusion device are described. It is shown that the surface fusion mechanism also provides a plausible explanation for the nonreproducibility of the results of electrolysis fusion experiments.