An investigation was made of the effect of various surface-finishing operations on the fatigue strength of two titanium alloys, RC 130B and Ti 140A. The types of finish studied included rough-machined, machined and mechanically polished, cold-rolled electropolished, and ground surfaces. Through the use of microhardness measuring techniques it was found that the different finishing methods introduced varying degrees and varying depths of cold work in the surface layers of the metal. Cold-rolling produced the highest hardness in the surface layer while grinding gave the lowest value; in one case the ground surface appeared to be slightly softer than the inner core metal. In general, the fatigue strength for lifetimes exceeding 2 × 107 cycles was found to vary according to the hardness of the surface layer with the highest hardness corresponding to the greatest fatigue strength. Roughness of the surface was also found to influence fatigue strength but to a much lesser degree than hardness. As a first approximation the relationship between surface hardness, surface roughness, and fatigue strength could be expressed by an equation of the form Z = KX−a Yb where Z is fatigue limit, X is surface roughness, Y is hardness, and K, a, and b are constants of the material. Data from the present study were evaluated in terms of this equation and the results are presented in the form of a nomograph. Results of tests made on notched specimens appear to indicate that titanium is less notch-sensitive in fatigue than was previously reported. Early reports of extreme notch sensitivity may have resulted from comparisons made between notched and smooth specimens having quite different surface preparations in the test section.